Ss , aes Rial , ciate ade mS eit fei eG + “4 WW BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25, 1973 BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1975 DATES OF PUBLICATION OF THE PARTS No. 1 25 June 1973 No. 2 7 June 1973 No. 3 8 June 1973 No. 4 26 June 1973 No. 5 27 June 1973 No. 6 22 August 1973 No. 7 23 August 1973 No. 8 19 October 1973 No. 9 25 October 1973 Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 5NU No. I. CONTENTS ZOOLOGY VOLUME 25 t 7m I On the cichlid fishes of the genus Pelmatochromis with proposal of a new genus for P. congicus ; on the relationship between Pelmatochromis and Tilapia and the recognition of Sarvotherodon as a distinct genus and II A new species of cichlid fishes of rivers Quanza and Bengo, Angola, with a list of the known Cichlidae of these rivers and a note on Pseudocrenilabrus natalensis Fowler. By E. TREWAVAS (Pl. x). The type specimens and identity of the species described in the genus Lithobius by R. I. Pocock from 1890 to 1901 as Lithobiomorpha). By E. H. Eason Freeliving marine nematodes from the Indian Ocean. By R. M. WARWICK Biology and fine structure of Euglypha rotunda (Testacea : Pro- tozoa). By R. H. HEpDLEy and C. G. OGDEN (Pls. 1-7) A revision of the Haplochromis and related species (Pisces : Cich- lidae) from Lake George, Uganda. By P. H. GREENWoop Preliminary notes on the ontogeny of the frontal body wall in the Adeonidae and Adeonellidae (Bryozoa, Cheilostomata). By P. L. Cook (Pls. 1-3) ; ‘ : : : Some new taxa of recent stalked Crinoidea. By A. M. CLARK (Pls. 1-2) Relationships of the palaearctic lizards assigned to the genera Lacerta, Algyroides and Psammodromus (Reptilia, Lacertidae). By E. N. ARNOLD. Distribution, ecology and evolution of the bellbirds (Procnias, Cotingidae). By D. W. Snow (Pls. 1-2). Index to Volume 25 PAGE 7 t I ON THE CICHLID FISHES OF THE GENUS PELMATOCHROMIS WITH PROPOSAL OF A NEW GENUS FOR P. CONGICUS; ON THE RELATIONSHIP BETWEEN PELMATOCHROMIS AND TILAPIA AND THE RECOGNITION OF _ SAROTHERODON AS A DISTINCT GENUS AND ___I A NEW SPECIES OF CICHLID FISHES OF _ RIVERS QUANZA AND BENGO, ANGOLA, WITH _A LIST OF THE KNOWN CICHLIDAE OF THESE a RIVERS AND A NOTE ON a PSEUDOCRENILABRUS NATALENSIS FOWLER ee oe een a Kio Ate A vee , (NAT: Hist. ” coi a) E. TREWAVAS BULLETIN OF ZOOLOGY Vol. 25 No. 1 ae LONDON : 1973 af . aR ween é Caen, a = ara I ON THE CICHLID FISHES OF THE GENUS PELMATOCHROMIS WITH PROPOSAL OF A NEW GENUS FOR P. CONGICUS; ON THE RELATIONSHIP BETWEEN PELMATOCHROMIS AND TILAPIA AND THE RECOGNITION OF SAROTHERODON AS A DISTINCT GENUS AND Il A NEW SPECIES OF CICHLID FISHES OF RIVERS QUANZA AND BENGO, ANGOLA, WITH A LIST OF THE KNOWN CICHLIDAE OF THESE RIVERS AND A NOTE ON PSEUDOCRENILABRUS NATALENSIS FOWLER BY ETHELWYNN TREWAVAS a Ke Lo WAT. HUST. ~ 5 JUL 1973 amen TED Mi, —SS—_— 0. (Nain SS res os (uy We i Wye ps ARs uw — Es | mm | mm Fic. 10. Scales from between the anal fin and the lateral line in R, Tilapia busumana (roman) and, G, Pelmatochromis buettikoferi (gothic). In T. busumana the circuli are more numerous and close-set than in the drawing. For comparison, H, a flank scale of Hemichromis bimaculatus (gothic). PELMATOCHROMIS, TILAPIA AND SAROTHERODON 17 in line with the changed dentition and presumably diet in adult T. buswmana. It is a feature in which this species resembles Pelmatochromis more than other species of Tilapia. The epibranchial structures Glandular and sensory pads on the roof of the pharynx are present in most (all?) cichlids anterolaterally to the upper pharyngeal teeth. ' When Steindachner proposed the subgenus Pelmatochromis (Gr. pelma, gen. pelmato = the sole of the foot, referring to the shape of the pad) he included two species, P. buettikoferi and P. jentinki, without designating either as type. Regan (1922) used the criterion of page priority to designate P. buettikoferi as type-species, and this is the one in which the pad is not boot-shaped, as Steindachner knew. Pellegrin, Boulenger and Regan eventually considered Pelmatochromis to include also Chromidotilapia Boulenger, 1898 (type C. kingsleyae), finding a gradation bet- ween the extreme shapes of the pad. Thys (1968a) has revalidated Chromidotilapia for species with a boot-shaped pad, since in this group its presence or absence is correlated with the respective absence and presence of microbranchiospines. This seems to be a natural division, although that the correlation is not always functionally or genetically necessary is shown by the fact that Tylochromis and Callochromis, for instance, have both a boot-shaped pad and microbranchiospines. Thys stated (1968a : 374) that the projecting pad is a glandular sac with an opening on each side in front of the upper pharyngeal tooth patches. This I —— | mm Fic. 11. Upper bones of rst and 2nd branchial arches in, A, Tilapia busumana, B, Pel- matochromis nigrofasciatus. The bones are drawn as if detached from the skull and viewed from above. Cartilage dotted. ep 1, ep 2, 1st and 2nd epibranchial; ph 1, 1st pharyngobranchial. The 2nd pharyngobranchial, a small toothed bone, is not shown. 18 E. TREWAVAS cannot confirm by examination with a dissecting microscope for either the boot- shaped or the sessile pad, and Thys gave no evidence for his statement. Dissection of both the hanging and the sessile pads reveals only connective tissue inside, often fat-laden. In the midst of this is a lamina of cartilage and bone that is an expansion of the anterior edge of the second epibranchial (fig. rr). On the oral side the mem- brane is raised into ridges and bosses bearing sensory cells. The histology of these pads has been little studied. Al Hussaini & Kholy (1954) found in the pharynx of Tilapia nilotica columnar cells interspersed with granular cells. Stolk (1957) described secretory cells in the pads of female Haplochrmis multicolor with- out comparing them with the males. Fishelson (1956 : 588, fig. 19) demonstrated both glandular and sensory cells in Tilapia thollont not only between and beside the upper pharygeal tooth-plates, but also on the floor of the pharynx, including the gill-arches. Macroscopically the pad is very similar in Pelmatochromis and Tilapia, but between Tilapia on the one hand and two of the species of Pelmatochromis (and Pt. congicus) on the other there is a difference in the epibranchial structures. In these (P. buettikoferi and P. nigrofasciatus) as also in Chromidotilapia, there is a long series of slender epibranchial gill-rakers on the first arch and the fold of skin (frenum) at the hinge between suspensorium and epibranchial is remote from the epi-ceratobranchial joint. This leaves a free passage for a current, respiratory or food-bearing, to pass between the pharynx and the upper part of the parabranchial chamber, in which direction can only be tested by experiment (fig. 12A). In Tilapia, on the other hand, the epibranchial gill-rakers are few, short and blunt and the frenum is attached near the epi-ceratobranchial joint. The upper gill-fllaments are attached along a line turning inwards behind the frenum, which would obstruct any flow of water between the upper filaments and the pharynx (fig. 12B). 10 mm Tae <7 / . PRESENTED A / : hogy eet” P, Br Printed in England by Staples Printers Limited at their Kettering, Northants, establishment - BULLETIN OF USEUM (NATURAL HISTORY) oh : Vol. 25 No. 2 _ LONDON: 1073 “THE TYPE SPECIMENS AND IDENTITY OF THE SPECIES DESCRIBED IN THE GENUS LITHOBIUS bye Role OCOCK FROM. 1800 .1O 1901 (CHILOPODA, LITHOBIOMORPHA) BY E. H. EASON Bourton Far Hill, Moreton-in-Marsh, Gloucestershire Php 39-83; 58 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Volt25 Nes 2 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, ts issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25, No. 2 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Zool.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 7 June, 1973 Price £1.95 THE TYPE SPECIMENS AND IDENTITY OF THE SPECIES DESCRIBED IN THE GENUS LITHOBIUS iby hal POCOCK FROM 2 —- — Pp amp am — — mp ap a 3-7 — —_ mp amp am — = mp ap ap 8-11 _ — amp amp am = — amp «= ap ap 12-14 —_ m amp amp am — —} amp) ‘ap ap 15 —_ m amp am a a — amp ap — The above is the usual spinulation but, on the 15th legs, DaC may be absent and in the only female with these legs intact VpF, DaT and DpT are present. REMARKS. The history of the generic classification of L. humberti is the same as that of L. pontifex. These two species are very close to one another and there is every reason for including them in the same species-group, in spite of the fact that the female of L. pontifex is unknown. 8. Lithobius stolli Pocock (Figs. 30 to 32) Lithobius stolli Pocock, 1895a: 9, Tab. 1 figs. 10, loa—c. Arenobius (Sowubius) stolli : Chamberlin, 1912 : 178. Sowubius stolli : Chamberlin, 1921 : Io. TyPE LocaLity. Volcan de Agua, Guatemala. MATERIAL EXAMINED. A jar labelled “ Lithobius stolli Pocock, type, 1894.4.1.76 ” contains a tube labelled “‘ Volcan de Agua 9,000-10,000’”’. A female of Lithobius castaneus Newport, 1844, fairly well preserved but with the 15th legs missing. TYPE SPECIMEN. Pocock’s description is of a single female and the above speci- men is the holotype (Reg. no. 1894.4.1.76). DESCRIPTION OF HOLOTYPE. L. castaneus is a well-known Mediterranean species but this specimen differs in detail from Brélemann’s (1930) account and is therefore described below. Size: 27mm long and 3:8mm broad at T.10. Colour: brown with a darker dorsal median band on the more posterior tergites. Shape: almost parallel-sided ; T.1 broader than T.3; body broadest at T.10. Head: rather longer than broad, as broad as T.1; a pair of short faint paramedian sulci run forwards on either side from the posterior marginal sulcus. Antennae: 12:2mm long, of 27 elongate articles, each article being two or three times longer than broad. Ocelli: I + 5, 5, 6, 6, 4 (Fig. 30). Prosternum: with 2 + 2 small teeth; lateral spine short and blunt, stouter than teeth, placed as far lateral to the lateral tooth as the interval between lateral and medial teeth; lateral to the spine the anterior border forms a broad level shoulder before sloping backwards without angulation; median cleft essentially shallow with a minute secondary cleft in the mid-line (Fig. 31). Tergztes: large tergites with posterior angles rounded and posterior borders emarginate ; posterior angles of T.g and 11 squared, those of T.13 with traces of projections ; on T.1, 3 and 5 the marginal sulcus is interrupted posteriorly and curved forwards on 58 E. H. EASON either side; this sculpturing is repeated faintly on T.8, 10 and 12; intermediate tergite quite strongly emarginate posteriorly. Coxal pores: 9, 9, 9, 9; oval to oblong. r4th leg: 9:2mm long; accessory apical claw a quarter the length of principal claw. 15th leg: missing; according to Pocock the accessory apical claw is absent. Gonopod: coxa without trace of medial excavation at base, with two short acuminate spurs separated from one another at their insertions by barely their own breadth; claw simple (Fig. 32). 31 1-Omm 0-5mm Fics. 30-32. Lithobius stolli. Fig. 30. Ocelli. Fig. 31. Dental margin of prosternum, ventral. Fig. 32. Left 2 gonopod, ventral. Spinulation : Ventral Dorsal Cc t P F a Cc t 1p F at I — — mp amp am — — amp ap ap 2-11 _ — amp amp am — — amp ap ap 12 — m amp amp am a — amp ap ap 13 a m amp amp am a — amp ap ap 14 a m amp amp am a — amp ap Pp 15 a I 3 2 I a On the 14th leg VaT and DaF are present on one side only. The numerals indicate the ventral spinulation given by Pocock for the telopodite of the 15th leg. THE LITHOBIUS SPECIES OF R. I. POCOCK 59 REMARKS. Sowubius was erected by Chamberlin (1912) as a monotypic sub- genus of Avenobius, characterized by the absence of posterior tergal projections and the absence of a 15th accessory apical claw, to receive L. stolli. It was later raised to a genus (Chamberlin 1921). Pocock’s illustration of the gonopod of L. stolli shows no trace of basal constriction of the coxa (Pocock 1895a: Tab. I fig. roc) and Chamberlin had, therefore, little justification for including this species in Arenobius and later in the Gosobiidae (see p. 44). Pocock’s illustration is, how- ever, on too small a scale to show the detailed structure of the gonopod and Chamber- lin may have assumed the base of the coxa to be excavated as it is in the majority of neotropical forms: he may also have been influenced by the simple genital claw, another feature he regarded as characteristic of the Gosobiidae, which is clearly shown in Pocock’s illustration. The holotype of L. stolli has been compared with a number of examples of L. castaneus from Europe and North Africa and also with the holotype of L. castaneus Newport in the British Museum (Natural History) (see also Eason 1972) and there is no doubt as to its identity. Sowubius thus becomes the senior synonym of Euporodontus, a genus erected by Verhoeff (1942) with L. castanews Newport as type species. In view of the historical association between Central America and Spain, where L. castaneus is common, artificial introduction seems the most likely explanation of the presence of this species in Guatemala. g. Lithobius provocator Pocock Lithobius provocatoy Pocock, 1891a:152; 1893:122; Chamberlin, 1904: 653; 1920a: 280, Pl. xxix figs. 6-8. Evemobius provocatoy: Chamberlin, 1925 : 408. TYPE LOCALITY. Bermuda. MATERIAL EXAMINED. A jar labelled ‘‘ Lithobius provocator Pocock, type, Bermuda, Challenger’ contains two females, two adult males and an immature male, all more or less mutilated, belonging to Lithobius peregrinus Latzel, 1880 (Reg. no. 1892.12.1.16-20). TYPE SPECIMENS. All the above specimens are syntypes. DEscRIPTION. The specimens are all quite typical and answer to Chamberlin’s (1920a) description of L. provocatoy and Brélemann’s (1930) description of L. pere- grinus. The position of the lateral prosternal spines medial to the respective lateral teeth in peregrinus was not made clear by Brélemann but was adequately described and figured by Matic (1957 : 13, figs. 4 and 5). REMARKS. Pocock (1893) realized that this was not an “ Antillean ’’ species and suspected that it might be a form of the European L. forficatus (Linn.). Chamberlin (1920a) gave a full description of L. provocator based on Bermudan material and later (Chamberlin 1925) erected Evemobius as a monotypic genus, characterized by the arrangement of the prosternal spine and teeth, to receive it. Synonymy of L. provocator with L. peregrinus was first proposed by Jeekel (1963). 60 E. H. EASON Lithobius ethochaetus, which was described from a single female taken at New Orleans in a cargo from Panama (Chamberlin 1938), undoubtedly belongs to L. peregrinus. A record of peregrinus from an inland locality in South Africa (Attems 1928) is further evidence of the extent to which this species, whose natural range is in southern and central Europe, has been artificially spread. 10. Lithobius sydneyensis Pocock (Figs. 33 to 36) Lithobius sydneyensis Pocock, 1891a: 153; Attems, 1914 : 97. Lithobius bermudensis Pocock, 1893 : 126. Walesobius sydneyensis : Chamberlin, 1920b: 77; Archey, 1937: 72, Pl. 16 figs. 1-6. Lithobius avaichensis Brélemann, 1924a : 58, figs. 4-6; 1924b: 187; 1932:55; Matic, 1959: 18, fig. 3; 1968: 71; Demange, 1962: 138; Matic et al., 1967: 192. Walesobius excrescens Attems, 1928: 78. ? Lithobius platensis Gervais : Silvestri, 1899 : 142. ? Chilebius coquimbo Chamberlin, 1955 : 56. ? Andebius callao Chamberlin, 1955 : 57. Type LocaLity. Sydney, Australia. MATERIAL EXAMINED. A jar labelled “ Lithobius sydneyensis Pocock, type, Sydney, Brazier ’’ contains a male and a female, both well preserved, and two mutilated males, one of them immature (Reg. no. 1972.738-740). TYPE SPECIMEN. The well-preserved male, 15 mm long, is here formally desig- nated as the lectotype and has now been placed in a separate tube (Reg. no. 1972.738). DESCRIPTION. Size: 15 to 1g mm long and 2:0 to 2:4 mm broad at T.10. Colour: pale brown. Shape: fusiform; T.r as broad as T.3 or rather broader; body broadest at T.8 or 10. Head: barely broader than long, as broad as T.5. Antennae: two-fifths of body-length, of 26 to 28 slightly elongate fairly regular articles. Ocelli: I + 5,5, 3 (Fig. 33) or, on the left side of the female, I + 6, 5, 4, 2. Prosternum: with 2 + 2 teeth, the medial tooth the larger; lateral spine setiform but stouter than the largest seta, placed posterolateral to lateral tooth; lateral to the spine the anterior border forms an irregular shoulder (Fig. 34). Tergites: posterior angles of T.8 and 10 rounded or blunt, those of T.12 blunt, those of T.14 rounded or blunt ; posterior borders of large tergites moderately emarginate ; posterior angles of T.9, 1m and 13 with triangular projections; intermediate tergite emarginate posteriorly. Coxal pores: 6, 7, 6, 5 or 6, 7, 7, 5 (5, 5, 6, 5 in immature male) ; oval or slightly oblong. 15th leg: about two-fifths of body-length; femur modified in male ; accessory apical claw very small, a fifth to a quarter the length of principal claw. Male secondary sexual characters: the dorsal aspect of the distal end of the 15th femur bears a prominent wart-like outgrowth arising from a shallow circum- scribed depression (Fig. 35). Female gonopod: coxa with two long slender spurs separated from one another at their insertions by about their own breadth; the spurs appear acuminate when viewed laterally but blunt and somewhat serrated apically when viewed ventrally ; claw with a large internal (dorsal) denticle and a much smaller external (ventral) denticle (Fig. 36). THE LITHOBIUS SPECIES OF R. I. POCOCK 61 1-Omm O-2mm 35 Fics. 33-36. Lithobius sydneyensis. Fig. 33. Ocelli. Fig. 34. Dental margin of pro- sternum, ventral. Fig. 35. Left 15th femur of g, posterior (internal). Fig. 36. Left 2 gonopod, ventral. Spinulation of lectotype (male) : Ventral Dorsal Cc t P EF 1 Cc t P FE T 2 — = mp amp am —= = mp ap a 3-9 = = mp amp am = = mp ap ap Io-II — = mp amp am = — amp ‘ap ap 12 = — amp amp am — — amp ap ap 13 = m amp amp am a — amp Pp ap 14 = m amp amp a a = amp Pp Pp 15 = m amp amp a a — amp Pp 2) In addition to the above there are two supplementary posterior prefemoral spines on each 15th leg (Fig. 35); VaF is duplicated on the right 15th leg. 62 E. H. EASON In the female 1VmP, 15VpF, 14DpT, 15DpF and 15DpT are absent; 12VmT, 2DpT and gDaP are present ; the two supplementary prefemoral spines are present on the 15th legs but 15VaF is simple on both sides. REMARKS. Walesobius was erected by Chamberlin (1920b) as a monotypic genus, characterized by the male 15th femoral protuberance, to receive L. sydneyensis. Although Brélemann’s (1924a) description of L. avaichensis from Morocco differs from Pocock’s specimens of L. sydneyensis in the smaller number of ocelli (g to 11), the circular coxal pores, and the presence of only one supplementary prefemoral spine on the 15th leg, examples of both sexes of avaichensis from Gibraltar in the Collection of the Cave Research Group of Great Britain in the British Museum (Natural History) have been compared with those of sydneyensis and there is no doubt that the two forms are conspecific. Walesobius excrescens from Cape Town (South Africa) was described by Attems (1928) as having coxal pores exactly as in the Gibraltan specimens (4, 5, 5, 3; circular), a simple 15th apical claw and the large ocellus anterior to the smaller ocelli (Io + 1): but the 15th accessory apical claw is very small in sydneyensis and was overlooked by both Pocock and Archey (1937), and Attems’ description of the ocelli may well have suffered mistranslation as the arrangement he gave is so unusual as to be barely credible: there is little doubt that sydneyensis and excrescens are conspecific. Lithobius platensis Gervais as described by Silvestri (1899), Chilebius coqguimbo (Chamberlin 1955) and Andebius callao (Chamberlin 1955) cannot be separated from L. sydneyensis by their rather brief descriptions, but their identity is less certain. Synonymy of L. bermudensis Pocock from Bermuda with L. sydneyensis is discussed below (p. 63). This species, as well as occurring in Morocco and Gibraltar, is widespread in southern Spain (Matic 1959, 1968). If we assume that this is its natural range and that it has been introduced to Bermuda, it might also have been spread artificially along the trade-routes to such seaports as Sydney, Auckland (New Zealand), where it was discovered by Archey (1937), and Cape Town. This suggestion is supported by the fact that there are few, if any, indigenous species of Lithobiidae in Austral- asia or South Africa in both of which regions the Lithobiomorpha is represented almost exclusively by the Henicopidae. Lithobius platensis (sensu Silvestri) was recorded by Silvestri (1905) from Quinta Normal (Chile), Buenos Aires (Argentina) and Montevideo (Uruguay), while Chilebius coquimbo and Andebius callao were recorded from Coquimbo (Chile) and Callao (Peru) respectively by Chamberlin (1955). It is possible that these records, all from seaports or coastal localities, also refer to L. sydneyensis ; but an indigenous species of Lithobiidae is known to occur in South America (Turk 1955) and L. platensis (which is probably conspecific with Chilebius coquimbo and Andebius callao) may be another. 11. Lithobius bermudensis Pocock Lithobius bermudensis Pocock, 1893 : 126. TYPE LOCALITY. Bermuda. ————_——____ THE LITHOBIUS SPECIES OF R. I. POCOCK 63 MATERIAL EXAMINED. A jar labelled “ Lithobius (s.s.) bermudensis Pocock, Bermuda, Challenger, 92.12.1.21”’ contains a female of Lithobius sydneyensis Pocock which is shrivelled, deeply pigmented and broken in two. TYPE SPECIMEN. Pocock’s description is of a single female and the above speci- men is the holotype (Reg. no. 1892.12.1.2I). DESCRIPTION OF HOLOTYPE. The specimen is very defective and the following description is necessarily rather incomplete, but it is so close to Brélemann’s (1924a) description of L. avaichensis (=sydneyensis) and the Gibraltan examples of this species which have been examined (see p. 62) that there is little doubt as to its identity. Length: about 14mm. Antennae: broken. Ocelli: 1 + 4, 3, 2. Prosternum: with 2 + 2 teeth and distinct shoulders lateral to the lateral spines. Tergites: some- what distorted but the general shape of the tergites and the posterior projections on T.9, 11 and 13 are asin L. sydneyensis. Coxal pores: 4,5, 5,4; circular. 15th leg: the apical claw has been broken off, but Pocock noted the presence of an accessory claw. Female gonopod: as in L. sydneyensis. Spinulation : Ventral Dorsal Cc eerie F as € t 12) F 4p 14 — m amp amp a a — amp Pp 5) 15 — m amp am a a — amp 5) — In addition to the above there is a single supplementary posterior prefemoral spine on each 15th leg. REMARKS. Pocock realized that L. bermudensis, like L. provocator, was not an “Antillean ”’ species but he was unable to identify it with any species known to him. Chamberlin (1920a) believed it to be an immature stadium of L. provocator (= peregrinus). The fact that it proves to be L. sydneyensis is not surprising as we have already seen how this species seems to have spread far beyond its natural range in the western Mediterranean (see p. 62). Bollman’s (1889) doubtful record of L. lapidicola Meinert from Bermuda may also refer to L. sydneyensis. 12. Lithobius persicus Pocock (Figs. 37 and 38) Lithobius persicus Pocock, 1899 : 399. Alokobius persicus : Attems, 1927 : 245. TYPE LOCALITY. Seir, Persia. MATERIAL EXAMINED. A jar labelled “‘ Lithobius persicus Pocock, type, 1900.1.15.2”’ contains a tube labelled “ Seir 18.viii, coll. by R. T. Gunther”. A male pierced longitudinally by a pin. TYPE SPECIMEN. Pocock’s description is of a single male and the above specimen is the holotype (Reg. no. 1900.1.15.2). 64 E. H. EASON DESCRIPTION OF HOLOTYPE. Size: 19mm long and 2:2mm broad at T.10. Colour: dull brown. Shape: almost parallel-sided; T.1 rather narrower than T.3; body broadest at T.10. Head: distorted but probably broader than T.r1; frontal sulcus very marked. Antennae: 7:0mm long, of 47 articles. Ocellr: I + 5, 4, 3, 2 (Fig. 37). Prosternum: with 2 + 2 teeth; lateral spine slender but stouter than setae, placed immediately posterolateral to lateral tooth; lateral to the spine the anterior border appears to slope backwards without formation of shoulders, but its detailed structure is obscured owing to its being damaged by the pin. Tergites: posterior angles of T.8 and ro rounded, those of T.12 and 14 blunt ; posterior borders of large tergites feebly emarginate ; posterior angles of T.9 obtuse, those of T.11 with feeble projections, those of T.13 with more marked projections (Fig. 38) ; intermediate tergite moderately emarginate posteriorly. Coxal pores: 3, 4, 4, 4; circular; separated from one another by more than their own diameter. 14th leg: with an accessory apical claw. 15th leg: 5:5 mm long; accessory apical claw absent. Male secondary sexual characters: 14th tibia not sulcate although Pocock mentioned a dorsal sulcus on this article; 14th femur with a feeble dorsal sulcus; 15th tibia with a distinct shallow dorsal sulcus. 2-Omm Fics. 37-38. Lithobius persicus. Fig. 37. Ocelli. Fig. 38. 11th to 14th tergites, dorsal. Spinulation : Ventral Dorsal c t 12) F ap Cc t 12 F ap I = —- mp amp amp — -- mp ap —_— 2-11 os —— mp amp amp —_ — amp ap ap 12 —- — amp amp amp — — amp ap ap 13 — m amp amp am a — amp ap ap 14 a m amp amp am a — amp 5) p 15 a m amp amp am a — amp Pp —_— REMARKS. Alokobius was erected by Attems (1926) to receive species such as L. persicus with some modification of the male 14th or 15th tibiae; it was later reduced to a subgenus of Lithobius (Attems 1938) and most authors disregard it altogether. L. persicus is not easily recognizable from Pocock’s description which gives 42 antennal articles, a dorsal sulcus on the 14th as well as the 15th tibia, and square posterior angles on T.11; but no species resembling the holotype at all closely has EME LITHOBIUS SPECIES OF R. I. POCOCK 65 been rediscovered in spite of fairly extensive collecting in Persia and Iraq (Brélemann 1922, 1923; Attems 1951; Chamberlin 1958; Matic 1969). The species probably belongs to the same group as L. laccatus Attems, 1951 and L. easoni Matic, 1969. 13. Lithobius doriae Pocock Lithobius (Archilithobius) doriae Pocock, 1890 : 63. Lithobius doriae : Brdlemann, 1898: 190; Verhoeff, 1937a : 203; Matic, 1970: 37. Lithobius mediterraneus Chalande, 1903 : 221; 1905: 72. Lithobius pilicornis dorviae: Brolemann, 1930: 299, and most subsequent authors. Lithobius doriae uccianensis Verhoeff, 1943 : 15. TyPE Localities. Busalla, Finale and Isoverde, Italy. MATERIAL EXAMINED. A jar labelled “ Archilithobius doriae (Pocock), type, 1889.3.8.25-26 ”’ contains the following two tubes : “ Lith. (Arch.) busallae (Pocock), Finale, O. Thomas (c & p).” An immature female (Reg. no. 1889.3.8.25). “ Lith. (Arch.) busallae (Pocock), Busalla, O. Thomas (c & p)”’ [the word ‘ busallae’ has been crossed out and ‘ doriae’ substituted]. Five males and seven females of L. doriae together with a female of Lithobius castaneus Newport and a female of Lithobius piceus verhoeffi Demange (Reg. no. 1889.3.8.26). The last two specimens have now been separated from those of L. doriae. TYPE SPECIMEN. Pocock’s description was based on a number of specimens from Busalla, one from Finale and one from Isoverde (all in Liguria, Italy), the last of which has not been found. The best preserved specimen from Busalla, a male 17 mm long, is here formally designated as the lectotype and has now been placed in a separate tube (Reg. no. 1889.3.8.26 (part)). DEScRIPTION. In general agreement with Brélemann’s (1930) description of L. pilicornis doriae. Of the variable features, the prosternal teeth are usually 4 + 4 but four specimens have 4 + 5, 4+ 6,5 +5 and 5 + 7 teeth respectively; the posterior angles of T.13 have small but distinct projections, and in a few specimens those of T.11 have very feeble projections; the coxal pores are usually 5, 6, 6, 5 or 5, 7, 7, 5, oblong in the larger and oval or almost circular in the smaller specimens ; on the 15th leg the spines VaC (coxolateral) and VmC are always present; on the 14th leg VaC is present in most specimens and VmC in one only. REMARKS. The subgenus Archilithobius in which Pocock placed L. doriae was erected by Stuxberg (1875) to receive all those species without marked posterior tergal projections : it is now disregarded by most authors but Attems (1926) and Chamberlin (1952) have used the name for two different species-groups neither of which includes L. doriae. L. mediterraneus from the south of France has long been recognized as a synonym of L. doriae. L. doriae uccianensis from Corsica was described as having the coxal pores oval, no more than twice as broad as long, and the spine VmC present on the 14th as well as the 15th leg (Verhoeff 1943) : since both these features are found in one or other of the syntypes from Busalla the sub- species uccianensis should be disregarded. All the features separating L. doriae from L. pilicornis Newport are such as would be expected from its smaller size and, 66 E. H. EASON since the two forms are not known to be truly sympatric, Brélemann was probably correct in regarding them as no more than subspecifically distinct. Pocock must first have named this form L. busallae but the name was never pub- lished. 14. Lithobius birmanicus Pocock (Figs. 39 to 41) Lithobius (Archilithobius) biymanicus Pocock, 1891b : 407. Archilithobius birmanicus : Attems, 1914 : 97. Lithobius (Australobius) birmanicus : Attems, 1938 : 345. Archilithobius birmanicus var. chandellensis Larwood, 1949 : 136, figs. 3a-f. TyPE LocaLities. Tharrawaddy and Palon, Burma. MATERIAL EXAMINED. A jar labelled “ Lithobius (Arch.) birmanicus (Pocock), 89.7.15.75, Tharrawaddy, E. W. Oates (c & p) ’’ contains five males and four females, all mutilated, and one specimen so defective as to be unrecognizable. TYPE SPECIMEN. Pocock’s description was based on a number of specimens from Tharrawaddy and Palon, Burma. No material from Palon has been found and the best preserved specimen from Tharrawaddy, a male 9:0 mm long, is here formally designated as the lectotype and has now been placed in a separate tube (Reg. no. 1889.7.15.75 (part)). DescriPTION. Size: 8-0 to 9-5 mm long and 1-0 to 1:rmm broad at T.10. Colour: dull yellow. Shape: fusiform; T.1 rather narrower than T.3; body broadest at T.8 or 10. Head: barely broader than long, about as broad as T.8 or to. Antennae: almost half body-length, of 20 articles which become progressively more elongate distally. Ocelli: 1 + 3, I (Fig. 39) or 1 + 3, 2. Prosternum: with 44+ 4,4+ 5,5 +50r5 + 6 teeth; lateral tooth or lateral two teeth smaller and placed more posteriorly than others; no lateral spine; lateral to lateral tooth the anterior border curves backwards without angulation (Fig. 40). Tergites: posterior angles of T.8 and ro rounded, those of T.12 and 14 rounded or blunt; posterior borders of large tergites moderately emarginate ; posterior angles of T.g, 11 and 13 obtusely rounded; intermediate tergite feebly emarginate posteriorly. Coxal pores: 2,4, 4,3; 3, 3,433 3) 4, 4,3; 3,4) 4, 4 or 3, 5,5, 4; circular; separated from one another by less than their own diameter. Tarso-metatarsal articulations of anterior legs: with, at most, a faint indication of an articulation. 15th leg: missing in all specimens ; according to Pocock the accessory apical claw is present. Female gonopod: coxa with two or three conical spurs set close to one another ; when there are three the internal spur is much smaller than the others; claw tri- dentate (Fig. 41). Spinulation of lectotype : Ventral Dorsal ¢ t 12) F a (C t Pp F 13 — m amp amp m a — amp p 5) 14 — m amp amp — a — amp Pp — 15th coxa without lateral spine (VaC). THE LITHOBIUS SPECIES OF R. I. POCOCK 67 O-:2mm 05% 39 4| —SS==_ O-Imm Fics. 39-41. Lithobius birmanicus. Fig.39. Ocelli. Fig. 40. Dental margin of prosternum, ventral. Fig. 41. Left 2 gonopod, ventral. Remarks. Attems (1914) and Larwood (1949) seem to have accepted the genus Archilithobius in Stuxberg’s (1875) original sense (see p. 65) but L. birmanicus would not be included in Archilithobius as later emended by either Attems (1926) or Chamberlin (1952). Australobius, on the other hand, was erected by Chamberlin (1920b) to receive an Australian species (A. scabrior Chamberlin, probably intro- duced to Australia from the East Indies) and is characterized by antennae of only about 20 articles, relatively few large ocelli arranged in a compact mass, 3 + 3 or more prosternal teeth, frequent absence of the prosternal lateral spine and usually three or more spurs on the female gonopod. L. biymanicus and the next four species all seem to belong to Australobius which is distributed chiefly in India and south-east Asia. Pocock described the 15th leg of biymanicus with only a single ventral spine each on the trochanter, prefemur and femur, but this reduced spinulation of the 15th leg is hardly credible when associated with the profuse ventral spinulation found on the 14th leg of the lectotype, and Pocock’s description was probably based on an 68 E. H.- EASON abnormal or regenerating leg. Larwood’s naming of a new variety from India, chandellensis, characterized by three ventral spines on the prefemur and two on the femur of the 15th leg and agreeing with the lectotype in other respects (Larwood 1949) was not justified. 15. Lithobius feae Pocock (Figs. 42 to 44) Lithobius feae Pocock, 1891b: 408; Attems, 1914: 97; Silvestri, 1917: 307, fig. I 1-13. Lithobius (Australobius) feae: Attems, 1938: 345; 1953: 152. Burmobius feae : Chamberlin, 1944b : 14. Type LocaLity. Mount Mooleyit, Burma. MATERIAL EXAMINED. A jar labelled “ Lithobius feae Poc.? North Chin Hills (Upper Burma) E. Y. Watson, 93.11.12.3”’ contains a mutilated female. The single female on which Pocock’s description was based was collected by Sig. L. Fea on Mount Mooleyit in Tenasserim (Lower Burma) so that the female in the British Museum (Natural History) cannot be the holotype: nor should it be selected as neotype because the North Chin Hills are too far from the type locality. However, there is no doubt as to its identity and as the only full description of L. feae in the literature, that of Silvestri (1917), was based on specimens from India, the female from the North Chin Hills is described below. DESCRIPTION. Size: 20 mm long and 2-4 mm broad at T.10. Colour: destroyed by partial clearing. Shape: parallel-sided; T.1 broader than T.3, almost as broad as T.10; body broadest at T.8 and 10. Head: broader than long, as broad as T.10. Antennae: 7-5 mm long, of 20 very elongate articles, particularly those of the distal one-third which are two or three times longer than broad. Ocelli: 1 + 3, 2 (Fig. 42). Prosternum: with 9g + 9 teeth; lateral tooth set a little apart from the others and placed slightly more posteriorly, leaving a rather indefinite diastema ; lateral spine minute, setiform, inserted immediately posterior or posteromedial to the diastema ; since all the prosternal setae are missing the size of the lateral spine relative to that of the setae cannot be estimated but it is probably more slender and almost certainly much shorter than the largest of the setae; the whole dental margin projects slightly forwards from the rest of the prosternum (Fig. 43). Tergites: posterior angles of T.8 and ro blunt, those of T.12 and 14 angulated; posterior borders of larger tergites broadly and increasingly emarginate from before back- wards; posterior angles of T.7 with feeble broad projections, those of T.g, 1x and 13 with increasingly long rather blunt projections; intermediate tergite strongly emarginate. Coxal pores: 5,6, 6,6; oval or circular ; separated from one another by about half their own diameter. Tarso-metatarsal articulations of anterior legs: fully developed and functional. r4th and 15th legs: missing ; according to Silvestri (1917) accessory apical claws are present on both. Gonopod: coxa with four conical spurs set fairly close to one another, the internal spur very small; claw simple (Fig. 44). THE LITHOBIUS SPECIES OF R. I. POCOCK 69 1-Omm ee O-5mm Fics. 42-44. Lithobius feae. Fig. 42. Ocelli. Fig. 43. Dental margin of prosternum, ventral. Fig. 44. Left 9 gonopod, ventral. Spinulation : Ventral Dorsal Cc t P EF T Cc t 12 F as I _— —_ — amp m —_— — amp a a 2-7 — —_ Pp amp am — — amp ap a 8 —_— — Pp amp am — — amp ap ap g-12 = = mp amp am — — amp ap ap 13 = m amp amp am a = amp ap ap Lateral spine (VaC) present on 15th coxa only. REMARKS. Burmobius was erected by Chamberlin (1944b) as a monotypic genus, characterized by the presence of a prosternal diastema in addition to the general features of Australobius, to receive L. feae. Burmobius is, therefore, equivalent to Alloporodontius Verhoeff which was erected as a subgenus of Lithobius to receive 70 E. H. EASON a number of Malayan species (Verhoeff 1937b) and was characterized in exactly the same way. But the prosternal diastema is not a striking feature in the specimen of L. feae from Upper Burma and was not noticed by Pocock in the original specimen. There is no evidence of Chamberlin having examined a specimen himself and he must have based Burmobius on Silvestri’s illustration of the prosternum of an Indian specimen which shows a diastema between the four lateral and four medial teeth (Silvestri 1917: fig. I 11). It may be that this intermediate position of the pro- sternal diastema characterizes an Indian subspecies of L. feae but it is more likely that the exact position and even the existence of a diastema are variable in this species. 16. Lithobius weberi Pocock (Figs. 45 and 46) Lithobius (Archilithobius) weberi Pocock, 1894 : 310, Pl. 19 fig. 2. Archilithobius javanicus : Attems, 1907 : 89 (non Pocock, 1894). Archilithobius weberi : Attems, 1914 : 97. Monotarsobius javanicus : Attems, 1914 : 97. Tamulinus japanicus : Attems, 1927 : 244 (typ. error). Lithobius (Australobius) weberi : Attems, 1938 : 345. Lithobius (Tamulinus) javanicus : Attems, 1938 : 345. Lithobius (Australobius) javanicus : Attems, 1953 : 151. Type Localities. Manindjau and Mount Singalang, Sumatra. MATERIAL EXAMINED. A jar labelled “ Lithobius javanicus Pocock, weberi Pocock, type, 1896.10.6”’ contains a tube labelled “ Lithobius weberi Pocock, type, Manindjau (Sumatra) ’’. A male with the 13th, r4th and 15th legs missing. TYPE SPECIMEN. Pocock’s description was based on two males, one from Man- indjau and one from Mount Singalang. The latter has not been found and the specimen from Manindjau is here formally designated as the lectotype (Reg. no. 1896.10.6.92). DESCRIPTION OF LECTOTYPE. Size: 12mm long and 1:6mm broad at T.10. Colour: dull red. Shape: fusiform; T.r as broad as T.3; body broadest at T.8 and 10. Head: broader than long, rather broader than T.5. Antennae: 4:0 mm long, of 20 slightly elongate articles. Ocelli: 1 + 3, 2 (Fig. 45). Prosternum: with 3 + 3 teeth all of much the same size; no lateral spine; lateral to lateral tooth the anterior border slopes gradually backwards without angulation (Fig. 46). cee 45 46 el O:2mm Fics. 45-46. Lithobius weberi. Fig. 45. Ocelli. Fig. 46. Dental margin of prosternum, ventral. | THE LITHOBIUS SPECIES OF R. I. POCOCK 71 Tergites: large tergites with posterior angles rounded and posterior borders moder- ately emarginate ; posterior angles of T.9, 11 and 13 squared; intermediate tergite feebly emarginate posteriorly. Coxal pores: 2, 3, 4,3; circular; separated from one another by about their own diameter. Tarso-metatarsal articulations of anterior legs: with, at most, a faint indication of an articulation. 13th, r4th and r5th legs: missing. = Spinulation : Ventral Dorsal Cc t P F a Cc t P F ay I —_— — Pp a — — — amp a a 2-8 = = p am m —_— — amp = ap a g-II — — Pp am am — — amp p a 12 — — p m am — — amp P P 13th to 15th coxae without lateral spines (VaC) ; dorsal spine (DaC) present on 15th coxa only. REMARKS. Attems (1907) described a single male from Buitenzorg (Java) which he identified as Archilithobius javanicus (Pocock) but there is no doubt that this specimen should have been referred to L. weberi: all Attems’ subsequent references to, and brief descriptions of, javanicus were presumably based on this specimen as there is no evidence of his ever having seen an example of the true L. javanicus. His references to weber, on the other hand, seem to have depended solely on Pocock’s description and records. As in the case of L. birmanicus, L. weberi would only be included in Archilithobius in Stuxberg’s (1875) original sense. Attems (1914), finding that his specimen had fused tarsal articulations on the 1st to 13th legs, transferred “‘ javanicus ”’ (= webert) to Monotarsobius, a genus erected by Verhoeff (1905) to receive those species with fused anterior tarsal articulations and only 1g to 22 antennal articles: at the same time Attems left weber in Archilithobius on the assumption that these articulations are free in that species. He later (Attems 1927) transferred “‘ javanicus’”’ from Monotarsobius to Tamulinus, a genus he had himself erected (Attems 1926) to include those species excluded by Verhoeff’s (1907) revised and more restricted definition of Monotarsobius by the possession of more than 2 + 2 prosternal teeth. He finally (Attems 1953) discarded Tamulinus as a useful division and transferred the species to Australobius (see p. 67) where it certainly belongs. 17. Lithobius javanicus Pocock (Figs. 47 and 48) Lithobius (Archilithobius) javanicus Pocock, 1894: 311, Pl. 19 fig. 3. TYPE LOcALITY. Tjibodas, Java. MATERIAL EXAMINED. The jar labelled “ Lithobius javanicus Pocock, weberi Pocock, type, 1896.10.6”” contains a tube labelled ‘“‘ Lithobius javanicus Pocock, type, Tjibodas”’. A fairly well-preserved male. 72 E. H. EASON TYPE SPECIMEN. Pocock’s description is of a single male and the above specimen is the holotype (Reg. no. 1896.10.6.93). DESCRIPTION OF HOLOTYPE. Size: 7-°0mm long and o-gmm broad at T.8. Colour: dull yellow. Shape: almost parallel-sided; T.1 as broad as T.3; body broadest at T.8. Head: barely broader than long, as broad as T.8. Antennae: 2-8 mm long, of 20 rather irregular articles. Ocelli: 1 + 2, 1 (Fig. 47). Pro- sternum: with 4 + 5 teeth; lateral tooth smaller and placed more posteriorly than others; no lateral spine; lateral to lateral tooth the anterior border slopes sharply backwards without angulation (Fig. 48). Tergites: posterior angles of T.8, Io, 12 and 14 blunt; posterior borders of large tergites very feebly emarginate; pos- terior angles of T.9, 11 and 13 rounded and very feebly projecting posteriorly ; intermediate tergite feebly emarginate posteriorly. Coxal pores: 2, 2, 2, 2; cir- cular ; separated from one another by about their own diameter. Tarso-metatarsal articulations of anterior legs: quite distinct with a change in breadth of articles but no obvious flexion. 15th leg: 2:5 mm long, not swollen or otherwise modified ; accessory apical claw half the length of principal claw. Male genitalia: covered by genital sternite so that the specimen may be immature. OR: 47 48 _eeeEeEEeEeEeEeEEEeEe———————es 0-2mm Fics. 47-48. Lithobius javanicus. Fig. 47. Ocelli. Fig. 48. Dental margin of prosternum, ventral. Spinulation : Ventral Dorsal (Cc t 12) F 4p Cc t P F ar 14 —- m amp m — — <= mp Pp —_ 15 “= m amp m — —_ — mp — —- Spinulation of anterior legs very deficient and difficult to interpret accurately. THE LITHOBIUS SPECIES OF R. I. POCOCK 73 REMARKS. L. javanicus may be no more than subspecifically distinct from L. birmanicus from which it can only be separated by means of the shape of the tergites and the form of the tarso-metatarsal articulations of the anterior legs: the other distinctive features may all be due to immaturity of the holotype and the true status of javanicus will only be decided when adults of both sexes have been examined. It should be regarded, provisionally, as a valid species. 18. Lithobius sculpturatus Pocock (Figs. 49 to 51) Lithobius sculpturatus Pocock, 1901 : 449; 1906: 1041. Monotarsobius ceylanicus Attems, 1909: 21; 1914: 97. Archilithobius sculpturatus : Attems, 1914 : 97. Tamulinus ceylanicus : Attems, 1927 : 244. Lithobius (Tamulinus) sculpturatus : Attems, 1938 : 345. Lithobius (Tamulinus) ceylanicus : Attems, 1938 : 345. Lithobius (Australobius) ceylanicus : Attems, 1953 : 151. TYPE LOCALITY. Madras, India. MATERIAL EXAMINED. A Jar labelled “ Lithobius sculpturatus Pocock, type, 1894.10.24.65-67 ”’ contains the following two tubes : “Madras’’. A well-preserved female answering fairly well to Pocock’s descrip- tion (Reg. no. 1894.10.24.65). “ Kodaikanal”. Two males differing from Pocock’s description and belonging to another species to be described later in this paper (Reg. no. 1894.10.24.66-67). TYPE SPECIMEN. Pocock implied that his description was based on a number of specimens from Kodaikanal (Palni Hills, southern India) and Madras, but he did not select any one specimen as type. The description was clearly based chiefly on the female from Madras which is conspecific with the specimens Pocock (1906) recorded as L. sculpturatus from the Maldive and Laccadive Islands, whereas the males from Kodaikanal are specifically distinct. The specimen from Madras is here formally designated as the lectotype (Reg. no. 1894.10.24.65). DESCRIPTION OF LECTOTYPE. Size: 11-2 mm long and 1:25 mm broad at T.10. Colour: dark brown. Shape: feebly fusiform; T.1 broader than T.3; body broadest at T.8 and 10. Head: broader than long, as broad as T.10. Antennae: 36 mm long, of 20 fairly regular articles, mostly as broad as long. Ocelli: 1 + 2,2 (Fig. 49). Prosternum: with 4 + 4 teeth; lateral tooth smaller and placed more posteriorly than others; no lateral spine; lateral to lateral tooth the anterior border slopes gradually backwards without angulation (Fig. 50). Tergites: posterior angles of T.8 and ro rounded, those of T.12 and 14 blunt ; posterior borders of large tergites moderately emarginate ; posterior angles of T.g obtuse, those of T.11 and 13 squared; intermediate tergite moderately emarginate posteriorly. Coxal pores: 4,5, 5,5 and 3,5, 5,4; circular; separated from one another by their own diameter. Tarso-metatarsal articulations of anterior legs: with, at most, a faint indication of an articulation. 15th leg: 3-8mm long; accessory apical claw half 74 E. H. EASON the length of principal claw. Gonopod: coxa with four conical spurs separated from one another at their insertions by rather less than their own breadth ; internal spur very small; claw short, tridentate (Fig. 51). C88 bees ia 0) 5 ee } O0-2mm ee ey 0-2mm Fics. 49-51. Lithobius sculpturatus. Fig. 49. Ocelli. Fig. 50. Dental margin of pro- sternum, ventral. Fig. 51. Left 9 gonopod, ventral. Spinulation : Ventral Dorsal Cc t 12) 1) it Cc t 1p) BE at I — _— — am m — — amp a a 2 — _ p am m —_ — amp ap a 3-8 = —— p amp m —_ — amp ap a 9 — — p amp m — — amp ap ap 10 — — p amp am — — amp ap ap II — — mp amp am — — amp ap ap 12 a — mp amp am — — amp ap Pp 13 os — amp amp a — — amp op Pp 14 — m amp amp a a —— am psp) P 15 a m amp amp a a — amp p _ FURTHER SPECIMENS. Five males from the Maldive Islands and two females from the Laccadive Islands in the British Museum (Natural History) (Reg. no. 1952.12.11.73~78), on which Pocock (1906) reported, show general agreement with the lectotype but with the following variations : Size: 8-5 to11mm long. Colour: paler. Antennae: one male with only 1g articles. Coxal pores: 2, 3, 3,3; 2, 3, 4,3; 3,3, 4,3 0r 2, 4, 4, 3. 4th and r5th legs: slightly swollen in males. Coxo- lateral spines (VaC): present on 14th as well as 15th legs in one male. Female gonopod: with only three spurs in both specimens which may be immature. LTHRESLIEAOBLUS SPECIES -OF RK. I. POCOCK 75 REMARKS. Pocock overstated the number of ocelli in L. sculpturvatus and also failed to record the single ventral tibial spine (VaT) on the 15th leg. Consequently Attems (1909) described a female from Pont de Galle, Ceylon, which clearly belonged to sculpturatus, as a new species, Monotarsobius ceylanicus, which he later designated as the type species of Tamulinus (Attems 1927). His generic classification of this species was much the same as for the specimen of L. weberi he identified as javanicus (see p. 71). Attems’ references to the species by its correct name seem to have depended solely on Pocock’s description. Like the last four species L. sculpturatus belongs to Australobius. 19. Lithobius shimensis Pocock Lithobius shimensis Pocock, 1895b : 349. Lithobius shinensis : Attems, 1914 : 97 (typ. error). Lithobius (Australobius) shimensis : Attems, 1938 : 345; 1953: I5I. TyPE Locality. Tsu-Shima, Japan. MATERIAL EXAMINED. A jar labelled “ Lithobius shimensis Poc., Tsu-Shima, P. A. Holst, 91.11.15’ contains two mutilated stadia agenitalis of Bothropolys asperatus (L. Koch, 1878). TYPE SPECIMENS. The above two specimens are the syntypes (Reg. no. 1891. I1.15.2A). DESCRIPTION. 7:0mm and 5:5 mm long and answering to Murakami’s (1958) description of the agenitalis II of Bothropolys asperatus. Remarks. Pocock, mistakenly, described adult female gonopods in shimensis and as the multiseriate coxal pores characteristic of Bothropolys and related genera only appear in the more advanced post-larval stadia it is not surprising that Attems (1938), who was guided solely by Pocock’s description, should have referred this _ form to Australobius of which it possesses all the essential features. In fact Austra- lobius does not seem to occur in Japan (Takakuwa 1939, 19414). 20. Lithobius holstii Pocock (Figs. 52 to 54) Lithobius Holstii Pocock, 1895b : 349. Monotarsobius crassipes Holstii : Attems, 1909: 19; 1914: 97; Takakuwa, 1941b: 292, fig. 1; Chamberlin and Wang, 1952 : 183. Monotarsobius takakuwai Verhoeff, 1937a : 193. Type Locatity. Ashinoju, Japan. MATERIAL EXAMINED. A jar labelled “ Lithobius holstii Pocock, Ashinoju (Japan), P. A. Holst, 91.5.16.23. Tsu-Shima (Japan), P. A. Holst, 91.11.15” contains the following two tubes : “Ashinoju (Japan), P. A. Holst”. A well-preserved female (Reg. no. 1891.5. 16.23). 79 BE, i EASON “Tsu-Shima’’. A very mutilated male (Reg. no. 1891.11.15.2B). TYPE SPECIMEN. Pocock made it quite clear that his description was based on the female from Ashinoju and that the specimen from Tsu-Shima was mutilated when he examined it. The former is therefore the holotype (Reg. no. 1891.5.16.23). DESCRIPTION OF HOLOTYPE. Size: 8:0mm long and 1:3mm broad at T.10; the specimen is contracted and its length when alive must have been over 10 mm. Colour: yellow. Shape: fusiform; T.1 narrower than T.3; body broadest at T.10. Head: barely broader than long, as broad as 17.3. Antennae: 3:2 mm long; the right antenna of 19 articles becoming increasingly elongate distally much as described by Verhoeff (1937a) in Monotarsobius takakuwai; the left antenna of 22 articles, all, including the terminal article, about as long as broad. Ocelli: I + 3, 2 (Fig. 52). Prosternum: with 2 + 2 teeth; lateral spine setiform, placed postero- lateral to lateral tooth; immediately adjacent to the insertion of the spine is a distinct node; lateral to the node the anterior border curves backwards without angulation (Fig. 53). Tergites: large tergites with posterior angles rounded and posterior borders moderately emarginate ; posterior angles of T.g, 11 and 13 rounded, not projecting ; intermediate tergite broad and moderately emarginate posteriorly. Coxal pores: 3, 4, 4, 4; circular; separated from one another by about their own diameter. Tarso-metatarsal articulations of anterior legs: fused. 14th leg: moder- ately swollen. 15th leg: moderately swollen; 3-2 mm long; accessory apical claw three-fifths of length of principal claw. Gonopod: coxa with two conical spurs separated from one another at their insertions by about their own breadth and diverging from one another towards their apices ; claw simple but with both internal and external ridges irregular (Fig. 54). Spinulation : Ventral Dorsal (G t 12) F AE (e 16 12 F ai I m a 2 m = == = a a 3-8 — _ — am m — — Pp ap ap g-10 — _ m am m _ -- Pp ap ap II — — mp am m — — P ap ap 12 —_ — mp am am — — amp = ap ap 13 — — amp amp am —_— — amp P P 14 — m amp am m a — amp p —_ 15 — m amp am _— a — amp p — DaC very small on 14th and 15th legs. RemMarKs. The small number of antennal articles (1g-22), 2-+ 2 prosternal teeth, fused anterior tarsal articulations, and the absence of posterior tergal pro- jections place L. hostiit in Monotarsobius as redefined by Verhoeff (1907); but Attems (1909) was not justified in regarding this form as conspecific with L. (Mono- tarsobius) crassipes L. Koch from which it differs in many essential features. Pocock described the claw of the female gonopod as “‘ obsoletely trifid’”’ and this led Ver- hoeff (1937a) to name Monotarsobius takakuwai (from the neighbourhood of Tokyo) THE LITHOBIUS SPECIES OF R. I. POCOCK 77 OOo Oo 52 | — Cl 2 7) 0-2mm Fics. 52-54. Lithobius holstii. Fig. 52. Ocelli. Fig. 53. Dental margin of prosternum, ventral. Fig. 54. Left 2 gonopod, ventral. as a distinct species with a simple female genital claw. According to Takakuwa (1941b), who suggested the synonymy of takakuwai with holstiz, the genital claw shows considerable variation in this species. One of the cotypes of Monotarsobius takakuwai Verhoeff in the British Museum (Natural History) (Reg. no. 1937.9.9.57), a male 11 mm long, is almost certainly conspecific with the holotype of L. holstit. Lithobius palnis sp. nov. (Figs. 55 to 58) TYPE LocaLity. Kodaikanal in the Palni Hills, India. MATERIAL EXAMINED. The two syntypes of Lithobius sculpturatus Pocock from Kodaikanal (see p. 73). TYPE SPECIMENS. Holotype adult male and paratype immature male. B.M. (N.H.) Reg. nos. 1894.10.24.66 (holotype) ; 1894.10.24.67 (paratype). DESCRIPTION OF HOLOTYPE. Size: 10-°8mm long and I-3mm broad at T.10. Colour: brown. Shape: fusiform; T.1 rather narrower than T.3; body broadest at T.8 and 10. Head: broader than long, rather broader than T.5. Antennae: 3:8 mm long, of 20 articles; the basal article and numbers 3 to 7, 11 to 13 and 17 to 19 are as long as broad or slightly transverse; the second article and numbers 8 to 10 and 14 to 16 are distinctly elongate ; the terminal article is very elongate. Ocelli: 1 + 2, 1 on right (Fig. 55); 1 + 3, 2 on left (Fig. 56). Prosternum: with 4+ 4 teeth; lateral tooth set a little apart from the others and placed slightly more posteriorly, leaving a distinct diastema; lateral spine setiform, inserted immediately posterior to the diastema; lateral to lateral tooth the anterior border slopes backwards without angulation (Fig. 57). Tergites: posterior angles of T.8 and ro rounded, those of T.12 and 14 blunt; posterior borders of large tergites moderately emarginate; posterior angles of T.9, 11 and 13, rounded, pyea ad Aj0[0F{ OoIxey ‘OLeIIeNy ‘ejnuy S6gr xafyuod pyea adAj0]0F{ RISIag ‘NaS 66g1 snaisaadg é $61 ‘sneIyy snsouvnasyjnum snidumyos = pyea addAjzoz007T OOIxXa ‘OIoLIONy ‘ouszyIWG F6gr SOAIIOAIDUL Lo6r ‘suia}yy : snatuvavl sniqoyyyiyI4 pf “OU prea Ayyeuorstaoid ad A007] earl ‘sepoqily +t6gr snavupanl pyea adA}0}99'T OOIXa ‘OleIIeNy ‘aurezyIMUGQ S6gr 1yAaqueny LEOI ‘JooyIAA Wanyvyn] sniqosavjouo Wy = pyea ad Ajo[oHL uvdef ‘nfourysy ¢6gr 14sjoy xz, pyea odAjzoz00°T OOIXa ‘OleIIeny ‘ejnuy S6gr wuoupos fe) prea auou euling ‘yiAspooyy yunoyy = 16g avaf Z ‘AON ‘NAS isa) “EF6I ‘JJaoyIAA SisuauUni9IN avlAOp SNIgQOYIV] = : €o61 ‘apueleyy snauvadazipau sniqoyjvy = q thet y1odman siusoayrd sniqoyjvz yo saroadsqns — ad Ayo} aT Ayeqy ‘eyjesngq o6gr apisop (3) prea adAj0[0RL Peula}yenyy ‘OSuvua}eoYV 9p uURIIOA S6gr snquoposap “AON ‘NAS ‘6461 ‘poomieT Sisuazjapubny9a “IRA SNIVUBUAIG SNIQOYINLYIA = pyea edAjzoz00°T euling ‘Appemeleyy, r6gr SnNIWUDULALG “AON ‘NAS ‘I6g1I ‘yooo00g sisuakaupds sniqoyjvy = ad Aj0[0RL epnuog £691 sisuapnutsag snze}s pue AVIpyeA yele} eur Aypesoy adA ayeq saroads yeurmoN 78 ad 4} a; eusIsaqy I Adv 79 POCOCK i THE LITHOBIUS SPECIES OF R. (F6gr ‘000g uoU) Zo6r ‘sua, Vy : snavuvavl sniqoyz1p1Y94 f = prea “AON (NAS ‘bS61 ‘sneryy 2y9712z snaduyogs= prea é SS61 ‘ulpiaquieyD ovyjv2 snigapup = é SS6r ‘ulpiaquieyg oquinbos sniqaziy9 = ¢ 661 ‘1I}SAATIS : SIRAIOD sisuajyzd sniqoyjvyT = “AON ‘NAS ‘Qz61 ‘sul9}}y Suazsasoxa sniqgosajv 4 = “AON ‘NAS ‘bz61 ‘uUeWOOIg SisuayIIDAD sniqoyVT = €6gr ‘yI0I0g sisuapnusaq snrqoynyT = prea "AON (NAS ‘thet ‘yaodmaN snaunjsva sniqoyyyvy = ‘AON "NAS ‘(gZgr ‘yoo “T) sngvsadsv skjodoayjog = “AON “NAS ‘6061 ‘stua}}W SnItuUD]AaI SNIqosADJOUO PY = prea prea “AON "NAS “*g£6r ‘ulploquieyy snjavys0yja sniqoyyvy = ORI ‘[ezyeT snurssasag sniqoyjrT = snyzeys pue AzIpye A ad 430790] adAjo]oF] adA}0}99'T ad A}0]0F] sadAyudg ad 430,097 ad430}390'T sadyuks [elrezyeur od A} 072uSISOq ‘yuOd I ATAVE, eiyeurns ‘ne{puruepy Ppyewmozeny ‘ensy op uvojo,, eyerjsny ‘AsupAs Peuoyeny ‘ensy op uesjo, uedef ‘eunrys-nsy IPU] ‘SPIPPI OOIX9] ‘OloIIoNy ‘urd IWIO epnuliog Aypesoy dA b6gt C6gt 16gI C6gr C6gr IO61 C6gr 16g1 a7yeq 1saqam quvajna stsuadaupas 21048 sisuamnys snqvangqynas qmiayps 40qv9000Ag soroods TeUIWION, 80 E. H. EASON those of T.11 and 13 very feebly projecting posteriorly (Fig. 58); the folds and tubercular elevations on the surface of the tergites noted by Chamberlin (1920b) in Australobius scabrioy are present and most marked on T.10, 12 and 14, but they are no more pronounced than in the other species from India and south-east Asia ; intermediate tergite moderately emarginate posteriorly. Coxal pores: 2, 3, 3, 3; circular; separated from one another by their own diameter. Tarso-metatarsal articulations of anterior legs: fully developed and functional. r4th leg: slightly swollen. 5th leg: slightly swollen; 3:8mm long; accessory apical claw half the length of principal claw. Pints. pes On. Teens 7 55 56 57 Fics. 55-58. Lithobius palnis. Fig. 55. Ocelli, right. Fig. 56. Ocelli, left. Fig. 57. Dental margin of prosternum, ventral. Fig. 58. 9th to 14th tergites, dorsal. Spinulation : Ventral Dorsal Cc t 12 15 ar (e t P F a I — — P a a 2 — — P a _— — _— mp ap a 3-4 — — Pp a m — — amp ap a 5-6 = = P am m _— — amp ap a 7-9 — = — am m — — amp ap a 10 — — — am m — — amp p a II —_ — P am am — — amp p a 12 —_ —_— p am am (a) — amp 5) _ 13 — _— ap amp a(m) a — amp Pp Pp 14 — m amp amp a a — amp op _ 15 — m amp amp a a — amp (p) — Letters in brackets indicate spines present on one side only. THE LITHOBIUS SPECIES OF R. I. POCOCK 81 DESCRIPTION OF PARATYPE. Length: 85mm. Antennae: of 20 articles. Ocelli: 1 + 2, 2. Prosternum: with 4 + 3 teeth; lateral spine and diastema between the lateral two teeth on both sides. Coxal pores: I, 2, 2, 2. Remarks. The presence of a prosternal diastema places the new species in Alloporodontius Verhoeft (=Burmobius Chamberlin) and it is close to Lithobius (Alloporodontius) malayicus Verhoeff from Johore, Malaysia (Verhoeff 1937b). It differs from malayicus in the absence of distinct triangular projections on T.9g, II and 13 and in the position of the prosternal spine and diastema which, in malayicus, are between the intermediate two teeth. There is some doubt as to whether Allopo- vodontius should be retained as a division distinct from Avustralobius (see p. 69), and the exact status of Australobius remains uncertain until more is known of the lithobiid fauna of India and south-east Asia. ACKNOWLEDGEMENTS I wish to thank the Trustees of the British Museum (Natural History) for allow- ing me to borrow specimens and Mr K. H. Hyatt of the Arachnida and Myriapoda Section of the Department of Zoology of the Museum for his personal help and for reading and criticizing my manuscript. REFERENCES ARcHEY, G. 1937. Revision of the Chilopoda of New Zealand. Rec. Auckland Inst. Mus. 2 : 43-100. Attems, C.G. 1907. Javanische Myriopoden, gesammelt von Direktor Dr. K. Kraepelin im Jahre 1903. Mitt. naturh. Mus. Hamb. 24: 77-142. 1909. Die Myriopoden der Vega-Expedition. Ark. Zool. 5 (3) : 1-84. — 1914. Die indo-australischen Myriopoden. Arch. Naturgesch. 80, Abt. A, 4 Heft : 1-398. — 1926. Myriopoda. In Kukenthal, W. & Krumbach, T. (eds.), Handbuch dey Zoologie 4: 1-402, Berlin & Leipzig: de Gruyter. 1927. Myriopoden aus dem nérdlichen und éstlichen Spanien, gesammelt von Dr. F. Haas in den Jahren 1914-1919. Abh. senckenb. naturforsch. Ges. 39 : 233-290. 1928. The Myriopoda of South Africa. Ann. S. Afr. Mus. 26: 1-431. 1938. Die von Dr. C. Dawydoff in franzosisch Indochina gesammelten Myriopoden. Mém. Mus. natn. Hist. nat. Paris, N.S. 6 : 187-353. — 1951. Ergebnisse der Osterreichischen Iran-Expedition 1949/50: Myriopoden von Iran gesammelt von der Expedition Heinz Léffler und Genossen. Sber. dst. Akad. Wiss. Abt. 1, 160 : 387-426. L 1953. Myriopoden von Indochina Expedition von C. Dawidoff (1938-1939). Mém. Mus. natn. Hist. nat. Paris, Série A, 5 : 133-199. Borrman, C. H. 1888. A preliminary list of the Myriapoda of Arkansas, with descriptions of new species. Entomologica am. 4: 1-8. — 1889. Notes on a small collection of myriapods from the Bermuda Islands. Proc. Acad. nat. Sci. Philad. 1889 : 127-129. Broremann, H. W. 1898. Myriapodes d’Ahusquay (Basses Pyrénées). Feuille jeun. Nat. (3) 28 : 187-192, 200-203. | —— 1900. Myriapodes d’Amérique, 1. Mém. Soc. zool. Fr. 13 : 89-131. —— 1922. Myriapods collected in Mesopotamia and N.W. Persia by W. Edgar Evans. Proc. | R. Soc. Edinb. 42 : 54-74. —— 1923. Myriapods from Mesopotamia and Persia collected by Mr. P. A. Buxton. J. Bombay nat. Hist. Soc. 28 : 165-169. 82 E. H. EASON Br6LEMANN, H.W. 1924a. Myriapodes recueillis 4a Larache (Maroc) par M. P. de Peyerimhoff. Bull. Soc. Hist. nat. Toulouse 52 : 54-61. 1924b. Myriapodes du Grand Atlas Marocain. Bull. Soc. Sci. nat. Maroc 4: 184-197. —— 1930. Myriapodes. Chilopodes. Faune Fr. 25 : 1-405. 1932. Tableaux de détermination des Chilopodes signalés en Afrique du Nord. Bull. Soc. Hist. nat. Afr. N. 23 (2) : 31-46. CHALANDE, J. 1903. Contribution a la faune des Myriapodes de France. Feuille jeun. Nat. (4) 33: 221. — 1905. Contribution a la faune des Myriapodes de France. Feville jeun. Nat. (4) 35: 61, 70-72. CHAMBERLIN, R. V. 1904. New chilopods. Pyoc. Acad. nat. Sci. Philad. 56 : 651-657. 1912. New genera of North American Lithobiidae. Can. Ent. 44 : 173-178. — 1915. New chilopods from Mexico and the West Indies. Bull. Mus. comp. Zool. Harv. 59 : 495-541. 1917. The Gosobiidae of America north of Mexico. Bull. Mus. comp. Zool. Harv. 57: 203-255. 1920a. The myriapod fauna of the Bermuda Isles, with notes on variation in Scutigera. Ann. ent. Soc. Am. 13 : 271-303. 1920b. The Myriapoda of the Australian region. Bull. Mus. comp. Zool. Harv. 64: I—269. 1921. The centipeds of Central America. Pyoc. U.S. natn. Mus. 60 (7) : 1-17. 1925. The Ethopolidae of America north of Mexico. Bull. Mus. comp. Zool. Harv. 57 : 383-437. —— 1938. On eighteen new lithobiomorphous chilopods. Ann. Mag. nat. Hist. (11) 2 : 625- 635. 1942. On ten new centipeds from Mexico and Venezuela. Proc. biol. Soc. Wash. 55: 17-24. 1943. On Mexican centipeds. Bull. Univ. Utah biol. Ser. 7 (3) : 1-55. 1944a. Chilopods in the collections of Field Museum of Natural History. Publs. Field Mus. nat. Hist. Zoological series, 28 : 173-216. —— 1944b. Some chilopods from the Indo—Australian archipelago. Notul. Nat. 147 : 1-14. —— 1952. On the Chilopoda of Turkey. /stanb. Univ. FenFak. Mecm. (B) 17 : 183-258. —— 1955. The Chilopoda of the Lund University and California Academy of Science Expedi- tions. Acta Univ. lund. N.F. (2) 51 (5) : 1-61. 1958. On chilopods from Iraq. Ent. Mitt. zool. StInst. zool. Mus. Hamburg 2 : 17-20. CHAMBERLIN, R. V. & Wana, Y. M. 1952. Some records and descriptions of chilopods from Japan and other oriental areas. Proc. biol. Soc. Wash. 65 : 177-188. DEMANGE, J.-M. 1962. Sur quelques Lithobiidae des collections de l'Institut scientifique Chérifien. Bull. Soc. Sci. nat. phys. Maroc 41 : 137-139. Eason, E. H. 1972. The type specimens and identity of the species described in the genus Lithobius by George Newport in 1844, 1845 and 1849 (Chilopoda, Lithobiomorpha). Buil. Br. Mus. nat. Hist. (Zool.) 21 : 297-311. JEEKEL, C. A. W. 1963. On the identity of Lithobius provocator Pocock, 1891 (Chilopoda, Lithobiida). Ent. Ber., Amst. 23 : 84. Kraus, O. 1954. Myriapoden aus El Salvador. Senckenberg. biol. 35 : 293-349. Larwoop, H. J. 1949. Some new Indian Lithobiidae. J. Bombay nat. Hist. Soc. 46 : 133- 139. ; Matic, Z. 1957. Contributions a la connaissance des Chilopodes cavernicoles de Yougo- slavie. Glas. Muz. srp. Zeml. (B) 10 : 9-23. 1959. Biospeologica lxxix : ii contribution a la connaissance des Lithobiidés caverni- coles de la péninsule Ibérique. Archs. Zool. exp. gén. 98, Notes et Revue : 12-25. 1968. Chilopodele din colectia ‘‘ Biospeologica” (Seriile vii—viii). Lucy. Inst. Speol. Emil Racovita 7 : 55-77. —— 1969. Contributo alla conoscenza dei Lithobiidae dell’Iran. Fyragm. ent. 6 : 87-114. THE LITHOBIUS SPECIES OF R. I. POCOCK 83 Matic, Z. 1970. Chilopodi di Toscana, delle Alpi Marittime e delle Alpi Liguri. Fvagm. ent. 7 : 25-38. Matic, Z., DaRABANTU, C. & Cricuic1, M. 1967. Contributo alla conoscenza dei Chilopodi di Spagna e di Malta. Boll. Sed. Accad. gioenia Sci. nat. (4) 9 : 175-199. Murakami, Y. 1958. The life-history of Bothropolys asperatus (L. Koch). Zool. Mag. Tokyo 67 : 217-223. Pocock, R. I. 1890. Contributions to our knowledge of the Chilopoda of Liguria. Annali Mus. civ. Stor. nat. Giacomo Doria 29 : 59-68. 1891a. Descriptions of some new species of Chilopoda. Ann. Mag. nat. Hist. (6) 8: 152-164. 1891b. On the Myriapoda of Burma. Part 2. Report upon the Chilopoda collected by Sig. L. Fea and Mr. E. W. Oates. Annali Mus. civ. Stor. nat. Giacomo Doria 30 : 401-432. 1893. Report upon the Myriapoda of the Challenger Expedition with remarks upon the fauna of Bermuda. Ann. Mag. nat. Hist. (6) 11 : 121-142. . 1894. Chilopoda, Symphyla and Diplopoda from the Malay Archipelago. In Weber, M. C. W. (ed.), Zoologische Ergebnisse einer Reise in Niederldndisch Ost-Indien 3 : 307-404. Leiden. 1895a. Chilopoda and Diplopoda. Biologia cent.-am. 14 : 1-217. 1895b. Report upon the Chilopoda and Diplopoda obtained by P. W. Bassett-Smith Esq., Surgeon, R.N., and J. J. Walker Esq., R.N., during the cruise in the Chinese Seas of H.M.S. ‘ Penguin’, Commander W. U. Moore commanding. Ann. Mag. nat. Hist. (6) 15 : 346-369. 1899. Chilopoda and Arachnida. In Giinther, R. T. (ed.), Contribution to the natural | history of Lake Urmi, N.W. Persia, and its neighbourhood. J. Linn. Soc. Zool. 27 : 399- 406. 1g01. Some new genera and species of Lithobiomorphous Chilopoda. Ann. Mag. nat. ) Hist. (7) 8 : 448-451. 1906. Chilopoda and Diplopoda. In Gardiner, J. S. (ed.), The Fauna and Geography of the Maldive and Laccadive Archipelagoes 2, Supplement 2 : 1041-1045, Cambridge. | SttvEstRI, F. 1899. Contribucién al estudio de los Quilédpodos chilenos. Revta chil. Hist. nat. 3: 141-152. 1905. Myriapoda. In Fauna Chilensis iii. Zool. Jb. (Supplement) 6 : 715-772. 1917. On some Lithobioidea (Chilopoda) from India. Rec. Indian Mus. 13 : 307-314. | StuxBEeRG, A. 1875. Genera et species Lithobioidarum disposuit. Ofvers VetenskAkad. Forh. Stockh. 32 (3) : 5-22. Takakuwa, Y. 1939. Uber Japanische Lithobius-Arten. Trans. Sapporo nat. Hist. Soc. 16: 28-37. — I941a. Weitere ee Lithobius-Arten und zwei neue Diplopoden. Tyvans. Sapporo nat. Hist. Soc. 17: 1941b. Uber ees © japanische Lithobiiden. Tvans. nat. Hist. Soc. Formosa 31 : 292— 299. Torx, F. A. 1955. The chilopods of Peru with descriptions of new species and some zoo- geographical notes on the Peruvian chilopod fauna. Proc. zool. Soc. Lond. 125 : 469-504. Vernorrr, K.W. 1905. Uber die Entwicklungsstufen der Steinlaiifer, Lithobiiden, und Bei- trage zur Kenntnis der Chilopoden. Zool. Jb. (Supplement) 8 : 195-298. 1907. Chilopoda. Systematik. Bronn’s Kl. Ordn. Tierreichs 5, Abt. 2, Buch 1 : 217- 264. —— 1937a. Chilopoden-Studien. Zur Kenntnis der Lithobiiden. Aych. Naturgesch. N.F. 6 : 171-257. — 1937b. Chilopoden aus Malacca, nach den Objecten des Raffles Museum in Singapore. Bull. Raffles Mus. 13 : 198-270. — 1942. Zur Kenntnis mediterraner Chilopoden besonders der Insel Ischia. Z. Morph. Okol. Tieve 38 : 483-525. 1943. Uber Chilopoden der Insel Korsika. Zool. Anz. 143 : 1-20. “eX “~ A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. Arison. Marine Molluscs in the Cuming Collection British Muse (Natural History) described by William Harper Pease. Pp. 96; 14 Pla 1965. (Out of Print.) £3.75. Na WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier deal Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. : Tay1or, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Miner of the Bivalvia. Introduction. Nuculacea-Trigonacea.. Pp. 125; 29 Plé 77 Text-figures. 1969. £4.50. Haynes, J. R. Cardigan Bay recent Foraminifera (Cruises of the R.V. A 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BSg 5NU fos, R. M. WARWICK BULLETIN OF [I MUSEUM (NATURAL HISTORY) ea : Vol. 25 No. 3 LONDON: 10973 FREELIVING MARINE NEMATODES FROM THE INDIAN OCEAN * NAT. HiST. 15 OCT 1973 si, PRESENTED 4 SOG, piste BY RICHARD M. WARWICK Dove Marine Laboratory Pp. 85-117; 14 Text-figures, 2 Tables BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 3 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), wstituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25, No. 3 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 8 June, 1973 Price £1.50 FREELIVING MARINE NEMATODES FROM THE INDIAN OCEAN By RICHARD M. WARWICK * SYNOPSIS Thirteen species of freeliving marine nematodes collected by the Royal Society’s 1963 expedition to the South-east Arabian upwelling region of the Indian Ocean are described. Twelve of these species are new, all being referable to existing genera. Didelta cobbi sp. nov. is closest to D. maculatum Cobb, 1920 and is distinguished from this species by the form of the amphids, and possibly the arrangement of cephalic setae. Pvocamacolaimus papillosus sp. nov. is similar to P. dolichostylum Gerlach, 1953, but has shorter cephalic setae, no subventral buccal teeth and a differently shaped gubernaculum. Sphaevolaimus crenellatus sp.nov. is charac- terized by the posterior position of its amphids, a buttressed cephalic capsule and double- jointed spicules. Siphonolaimus profundus sp. nov. is apparently similar to S. gladiator Wieser, 1956, but differs from it in the setal arrangement on the head, the longer buccal spear and the presence of pre-cloacal supplements in the male. Desmodora masiva sp. nov. is characterized by a short broad perforate cephalic capsule, small amphids, a weakly armed buccal cavity, a long tail and no male supplements. Sabatieria alata sp.nov. is very close to S. furcillata Wieser, 1954, but has a larger overall size, shorter cephalic setae, smaller amphids and a different spicule structure. Hopperia muscatensis sp. nov. is distinct from the type and only other species H. massiliensis Vitiello, 1969, in having longer cephalic setae, a shorter tail and differently shaped spicules. Platycomopsis implicatus sp. nov. is separated from its nearest relative P. dimorphica Mawson, 1956, in the setal arrangement on the head, the smaller amphids, the structure of the spicules and the position of the pre-cloacal supplement. Synonchus alisonae sp. nov. is characterized by the very long filiform tail and relatively long cephalic setae. Specimens of Dayellus dayi Inglis, 1964, conform exactly with the type, but are slightly smaller in many of the measurements. Cvenopharynx fringilla sp. nov. is characterized by the com- bination of a long flagellate tail, spicules with no barbs or other modifications and a guber- naculum with pointed lateral pieces. Mesacanthion southerni sp. nov. is very closely related to M. diplechma Southern, 1914, but comparison with specimens of this latter species from the North Sea reveal differences of specific rank mainly concerning the copulatory apparatus of the male. Mesacanthion avabium sp.nov. is very characteristic: it has long spicules, a paired L-shaped gubernaculum with raised ridges at the distal end and unique accessory organs associated with the vulva in the female. INTRODUCTION Durinc June and July 1963 the Royal Society mounted a cruise on the R.R.S. “ Discovery’ to the South-east Arabian upwelling region as part of the International Indian Ocean Expedition. Details of the work undertaken and stations sampled are given in the Cruise Report (Royal Society, 1963). Benthic samples were taken with a 0-1 m? Smith-McIntyre grab down to depths of 3500 m and undisturbed sediment hauls were subsampled with a corer for meiofauna analysis. These cores are very rich in meiofauna, particularly nematodes, and * Present address: Institute for Marine Environmental Research, 13-14 St James Terrace, Citadel Road, Plymouth. 88 R. M. WARWICK their extraction and analysis on a quantitative basis are proving very time con- suming, so that completion of this work is not expected for some time. However, samples for macrofauna analysis passed through coarser sieves have been found to contain many new and interesting nematodes in the larger size range, and it is these which are described here. The sampling stations where the “ macrofaunal ”’ nematodes were found are listed in Table 1, together with information on the sedi- ment types and the sieve size used. A map showing the positions of the stations is given in Fig. 1, and the species found at these stations are listed in Table 2. 55 60 65 Or = {) | alt 1a 1 | 1 J. 1 it | it 1 4 PAKISTAN Persian 25 / Gulf | q J 4 | MUSCAT AND OMAN 60 _---5059 *5065 20 OF nc 5048 ARABIAN SEA IS ©5021 On 4 "| Fic. 1. Map showing positions of sampling stations. Only species represented by at least one mature male specimen in good condition are described. Specimens have been mounted in glycerine for examination. Curved structures, e.g. spicules, are measured as the curve and not the chord, and head diameters are measured at the level of the anteriormost cephalic setae. Type- material is deposited at the British Museum (Natural History). When more than one specimen is available, descriptions are based on a syntypic series. I am very grateful to Mr A. D. McIntyre of the Marine Laboratory, Aberdeen, for making this material available to me for study. NEMATODES FROM THE INDIAN OCEAN 89 TABLE I Positions and other details of sampling stations Station Depth Sieve Date no. Position (m) Sediment type size 29.6.63 5021 14°55/N-53°55°4'E 2480 Greenish-grey mud o-5 mm 30.06.63 5025 16°41’N—54°01-8’E 1030 Greenish fine sand o-5 mm 5-7.63 5034 17°57°2'N-56°43'5/E 49 Fine sand o-5 mm 10.7.63 5047 19°09°8’N—57°55:1'E 48 Fine muddy sand o-5 mm 10.7.63 5048 19°03’N—58°02:8’E 120 Muddy sand o-5 mm 10.7.63 5050 18°50’N-58°18-3’E 3147 Fine greyish-green mud 2905 u 15.7.63 5059 20°29°5’N-59°09'5/E 660 Greenish mud 295 U 15.7-63 5060 21°03°5’N-59°04'1’E 50 Fine shelly sand o-5 mm 16.7.63 5005 20°32-9’N—-59°55'5'E 3500 Grey-green clay 295 u 19.7.63 5071 22°12:9’N-59°54°3'E 1305 Greenish mud 295 22.7.63 5074 23°57:0’N-66°08-0’/E 182 Muddy sand with shells 0-5 mm and some clay DESCRIPTIVE SECTION Family LINHOMOEIDAE Didelta cobbi sp. nov. (Fig. 2) MATERIAL STUDIED. I ¢ (Holotype); B.M. (N.H.), Reg. No. 1972 : 354. REPRESENTATION. Station 5060. De Man’s Ratios. a b c Body length (mm) Male 106-00 19:91 9:81 4:24 MEASUREMENTS (in mm). Male: Body breadth: 0-040. Oesophagus length : 0-213. Distance of nerve-ring from anterior: 0-079. Distance of excretory pore from anterior : 0-092. Length of longer cephalic setae: 0-014. Length of shorter cephalic setae: 0-003. Length of anterior cervical setae: 0-005. Length of posterior cervical setae: 0-010. Length of amphid + plaque: o-015. Breadth of amphid + plaque: 0:0135. Tail length: 0-432. Cloacal diameter : 0-038. Spicule length : 0-051. Length of gubernacular apophysis : 0-029. DEscRIPTION. The body is filiform, and of almost uniform width between the nerve-ring and cloaca. The cuticle is marked with fine transverse striations. The head bears a circle of ten cephalic setae, of which six are much longer than the remaining four. Between the cephalic setae and the amphids is a circle of four cervical setae, and posterior to the amphids six longer setae (Fig. 2c). The amphids are oval in outline, and are mounted on oval cuticularized plaques. The region where the amphidal nerve joins the centre of the amphid has an opaque granular appearance. The buccal cavity has an anterior thin-walled cup-shaped chamber separated from the posterior more heavily cuticularized chamber by a go R. M. WARWICK constriction. The oesophagus expands only slightly towards its base and is linked to the intestine by prominent cardia (Fig. 2b). The ventral excretory gland extends posteriorly to the level of the cardia. TABLE 2 Distribution of species at the sampling stations Station no. Species Numbers 5021 Synonchus alisonae sp. nov. 4 33, 2 29, 9 juveniles 5025 Platycomopsis implicatus sp. nov. Ig 5034 Mesacanthion southerni sp. nov. Ig, 3 92 Mesacanthion avabium sp. nov. 4 dd, 10 92 Viscosia sp. 19 5047 Dayellus dayi Inglis, 1964 733, 5 22, 7 juveniles Crenopharynx fringilla sp. nov. 733, 9 92, 12 juveniles Mesacanthion southerni sp. nov. 5 66, 2 juveniles 5048 Theristus sp. 3 33, 12) sample badly Cyatholaimidae sp. 29 preserved ; Pavanticoma sp. 7. specimens in Unidentifiable 6 poor condition 5050 Siphonolaimus profundus sp. nov. 4 33, 7 22, 2 juveniles Desmodoridae sp. IQ Sabatiervia alata sp. nov. 2 5d, 5 $2, 4 juveniles Pavamesacanthion sp. ite) 4 juveniles Oncholaimidae spp. Lis 2 juveniles 5059 Desmodora masiva sp. nov. 6 33,7 22 Sphaerolaimus crenellatus sp. nov. 7 33, 20 92 5060 Didelta cobbi sp. nov. Id Dayellus dayi Inglis, 1964 6 3d, 13 22, 7 juveniles Crenopharynx fringilla sp. nov. 266,12 Oncholaimidae sp. I juvenile 5065 Steinervia sp. i (2) Desmodora sp. 2 92 Paramesacanthion sp. 2 99, I juvenile 5071 Theristus sp. I juvenile Sabatieria sp. I juvenile Hopperia muscatensis sp. nov. I 3, 4 99 Oncholaimidae sp. I juvenile 5°74 Procamacolaimus papillosus sp. nov. Ig Crenopharynx sp. I juvenile NEMATODES FROM THE INDIAN OCEAN gI The middle region of the body is devoid of setae. The tail is elongate with its distal two-thirds filiform. Numerous short setae are scattered ventrally im- mediately posterior to the cloaca, and longer setae are present at the junction of the conical and filiform sections of the tail (Fig. 2a). Fic. 2. Didelta cobbi sp.nov. a, Male tail. b, Anterior end of male. c, Lateral view of head. d, Lateral view of spicules and gubernaculum. Male. The spicules are paired, equal and arcuate, with a median list. The gubernaculum bears long caudo-dorsal apophyses. 92 R. M. WARWICK Discussion. Currently four species are referable to this genus; Didelta macu- latum Cobb, 1920, D. scutata Wieser, 1956, D. cascudwm Gerlach, 1956, and D. scutellata Vitiello, 1969a. The last three of these species all have relatively short cephalic setae (less than 8 »). The present species is separated from D. maculatum mainly by the shape of the amphids, which in the latter are much more elongate with the cuticular plate triangular in shape. The lateral cephalic setae are short in D. maculatum but appear to be long in D. cobbi. However, this feature is not at all clear in the holotype, since these setae seem to be rather bent. Family CAMACOLAIMIDAE Procamacolaimus papillosus sp. nov. (Fig. 3) MATERIAL STUDIED. I ¢ (Holotype); B.M. (N.H.), Reg. No. 1972 : 355. REPRESENTATION. Station 5074. De Man’s Ratios. a b c Body length (mm) Male 105'45 916 29:00 3°48 MEASUREMENTS (in mm). Male: Body breadth: 0-033. Oesophagus length : 0-38. Excretory pore from anterior: 0-112. Proximal tip of ventral gland from base of oesophagus: 0-181. Head diameter: o-o12. Length of cephalic setae: o-o12. Amphid diameter: 0-0045. Tail length: 0-120. Cloacal diameter : 0-035. Spicule length : 0-051. Gubernaculum length: 0-027. Distance of pre-cloacal seta from cloaca: 0-023. Distance of posterior supplement from cloaca : 0-035. Dis- tance of anterior supplement from cloaca: 0-28. Length of supplements : 0-021. Description. The body is slender and elongate. The cuticle is marked through- out with rather coarse transverse striation. The mouth is surrounded by six rounded lips each bearing a short, stout, outwardly curving papilla. There are four rela- tively long cephalic setae. The buccal cavity is short and conical, and contains a solid styliform dorsal tooth opposed by a pair of pointed hollow subventral teeth. The amphids have a circular outline which is broken posteriorly for the exit of the amphidal nerve (Fig. 3b). The oesophagus is cylindrical, with no posterior swelling. The excretory pore is prominent and the ventral gland extends well beyond the base of the oesophagus. The tail is conical and bears two large rounded mid- ventral papillae about half-way down its length. Smaller conical papillae are scattered generally over the surface of the tail (Fig. 3a). Male. The spicules are paired, equal and arcuate. They expand in width gradually towards the proximal end, at which a rounded cephalization is set off by a constriction (Fig. 3c). Thin ventral spicular alae are present. The guber- naculum bears a strong rounded dorsal apophysis and triangular projections extending forwards on either side of the spicules. There is a stout pre-cloacal seta and nine strongly cuticularized pre-cloacal supplements. These are tubular, pointed distally and with their cephalate proximal ends again set off by a con- striction. NEMATODES FROM THE INDIAN OCEAN 93 Fic. 3. Procamacolaimus papillosus sp.nov. a, Male tail. b, Lateral view of male head. c, Lateral view of spicules, gubernaculum and posterior supplement. Discussion. Of the four previously-described species Procamacolaimus papillosus sp. noy. is closest to P. dolichostylum Gerlach, 1953. The latter, however, has shorter cephalic setae, no subventral teeth in the buccal cavity and a differently- shaped gubernaculum. Family SPHAEROLAIMIDAE Sphaerolaimus crenellatus sp. nov. (Fig. 4) MATERIAL STUDIED. 5 gd and 5 99 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 350-366 (3 dg and 2 99 measured below). Additional material: 2 gg and 15 99; B.M. (N.H.), Reg. Nos. 1972 : 367-382. REPRESENTATION. Station 5059. DE Man’s Ratios. a be c V%_ Body length (mm) Males 21°47 5°53 9:22 - 2°49 23°73 5°50 9°31 oa 2°42 25°27 5°47 9°40 = 2°35 Females 23°98 5-78 IO‘1I 65°37 2°83 27-11 5°76 10°24 62°57 2-705 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0-116; 0-102; 0-093. Oesophagus length: 0-45; 0-44; 0-43. Distance of nerve- ring from anterior : 0-175; 0-19; 0-18. Distance of excretory pore from anterior : 0-27;? ; 0-26. Head diameter: 0-025; 0-028; 0-025. Length of cephalic setae : 0-004 + 0-003; 0-007 + 0-:0045; 0-007 + 0:0045. Length of longest subcephalic setae : 0-020; 0-024; 0:024. Length of longest cervical setae : 0:0275 ; 0-029 ; 0-033. 04 R. M. WARWICK Amphid diameter : 0-0125; 0:0125; 0-012. Body diameter at level of amphids: 0062; 0:057; 0:056. Tail length: 0-27; 0:26; 0-25. Cloacal diameter : 0-072 ; 0-075; 0-071. Spicule length: 0-121; 0-132; 01125. Gubernaculum length: 0:049; 0°046; 0-045. Females: Body breadth: 0-118; 0-102. Oesophagus length: 0-49; 0-48. Nerve-ring and excretory pore not visible. Head diameter: 0-033; 0-028. Length of cephalic setae : 0:0065 + 0:004; 0-007 + 0-004. Length of longest subcephalic setae : 0-025; 0-025. Length of longest cervical setae : 0-036; 0-037. Amphid diameter :0-:0095; 0-010. Body diameter at level of amphids : 0-063 ; 0063. Tail length:0-28; 0:27. Anal diameter:0-:065; 0-061. Distance of vulva from anterior: 1°85; 1°73. DESCRIPTION. The cuticle bears fine transverse striations, and there is no sign of lateral differentiation. The mouth is surrounded by six lip-flaps which are triangular in en-face view (Fig. 4d). At the bases of the lips are six minute conical labial papillae. More posteriorly there is a circle of ten cephalic papillae; the Fic. 4. Sphaerolaimus crenellatus sp.nov. a, Lateral view of male head. b, Lateral view of female head. c, Male tail. d, En-face view of head. NEMATODES FROM THE INDIAN OCEAN 95 lateral ones are minute and conical and the submedian pairs longer, subequal and setiform. The subcephalic setae are arranged in eight groups with between two and four setae in each group. These setae are of different lengths, as shown in Figs. 4a, bandd. There are ten groups of cervical setae : eight triplets correspond in position with the cephalic setae and are of unequal length; two pairs in the lateral positions just anterior to the amphids are of equal length. Eight files of longish setae (but not as long as the cervical setae) extend backwards from the head becoming progressively shorter until they disappear at about the level of the nerve-ring. Setae are scarce on the remainder of the body, becoming more numer- ous again on the tail, especially that of the male. The mouth opens into a small globular vestible (terminology of Inglis, 1961) which leads by a conical passage bordered by a leaf crown of forty cheilorhabdions to the heavily cuticularized globular buccal cavity. The anterior border of this cavity is indented with six rounded notches. At the posterior end of the buccal capsule there is a ring of buttresses similar to those described by Inglis (1961) for S. anterides. However, the buttresses are smaller than in this latter species, and triangular in shape. In both sexes the buttresses may be small and numerous (Fig. 4a) or less numerous and larger (Fig. 4b). They can only be counted accurately in en-face view (see 21 buttresses in Fig. 4d). Punctation and sclerotization of the buccal capsule are confined to its anterior half, and do not extend to the buttresses (cf. S. anterides where the buttresses are the only sclerotized parts of the buccal capsule). Pos- terior to the buccal capsule the oesophageal funnel is relatively shallow and bears no teeth, but the ventrolateral sectors of the oesophagus project into the funnel as two muscular lobes. The amphids are circular in outline and are situated wholly posterior to the buccal cavity. They are larger in the male than the female, and this is the only sexual dimorphism shown in the head of this species. The oeso- phagus is cylindrical with no posterior swelling, and has a thick cuticular lining. The intestine is very opaque. The tail is of similar shape in both sexes, the distal quarter being cylindrical (Fig. 4c). Male. The spicules are paired and equal. They are divided in the middle by a fine suture, and distally each bears a laterally curving projection. The guber- naculum has a dorsally directed apophysis which for the most part is poorly cuticularized. Because of the opacity of the intestine, the testes are not visible. Female. There is a single anterior ovary, but the opacity of the intestine again obscures details of its structure. ~There is no post-vulvar sac. The eggs are roughly spherical, a typical example measuring o-I12 x o-oggmm. Of three gravid females, all contained only a single egg. Discussion. A key to the species known at that time is given by Wieser (1956), and Freudenhammer (1970) lists species described subsequently. The present species agrees with S. gracilis de Man, 1884, and S. campbelli Allgén, 1927, in having the amphids of both sexes wholly posterior to the buccal cavity. It is closest to S. gracilis on the basis of the relative lengths of the cephalic and subcephalic setae. Although S. gracilis is a very variable species (Lorenzen, 1969, for example describes three distinct forms), none of its varieties has a buttressed cephalic capsule or double-jointed spicules. 96 R. M. WARWICK Family SIPHONOLAIMIDAE Siphonolaimus profundus sp. nov. (Fig. 5) MATERIAL STUDIED. Material in general is rather fragmented. 2 complete gd, t headless § and 2 complete 9° (Syntypes) ; B.M. (N.H.), Reg. Nos. 1972 : 383-386. Additional material: 1 complete 3, 1 complete 2, 1 complete juvenile, 7 head ends, I Q tail end, 1 juvenile tail end, 1 2 middle region; B.M. (N.H.), Reg. Nos. 1972: 387-389. REPRESENTATION. Station 5050. DE Man’s Ratios. a b € V% _ Body length (mm) Males 85°43 19°29 41-24 - 5-98 61-58 16-14 37°44 - 4.68 - - ~ - headless Females 72°83 21-61 34°36 71-19 6-70 60-91 20°30 32:06 71-19 6-70 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0:070; 0-076; -. Oecesophagus length:0:31; 0-29; -. Length of posterior oesophageal bulb: 0-085; 0-087; -. Breadth of posterior oesophageal bulb: 0:042; 0-041; —-. Distance of nerve-ring from anterior : 0-149; 0:134; -—. Head diameter: 0-012; 0-012; -. Length of cephalic setae : 0-008 + 0:0025; 0-008 + 0:0025; —. Length of buccal spear : 0-032; 0:032; —. Amphid width: 0-014; 0-014; —. Body diameter at level of amphids: 0-029; 0-028; -. Tail length: 0°145; 07125; 0-153. Cloacal diameter: 0-046; 0-045; 0-046. Spicule length : 0:079; 0-083; 0-082. Gubernaculum length : 0-025; 0-027; 0-031. Females: Body breadth: 0-092; 0-11. Oesophagus length: 0-31; 0:33. Length of posterior oesophageal bulb: 0-109; 0-105. Breadth of posterior oeso- phageal bulb: 0-044; 0-052. Distance of nerve-ring from anterior:?; 0-145. Head diameter:0-:0125; 0-012. Length of cephalic setae : 0-008 + 0:0025 ; 0:007 + 0:002. Length of buccal spear: 0-029; 0:036. Amphid width: 0-014; 0-014. Body diameter at level of amphids: 0-028; 0-028. Tail length : 0-195; 0-209. Anal diameter: 0-050; 0-050. Distance of vulva from anterior : 4-77; 4:77. Egg length x breadth: 0-29 x 0:08; 0-265 x 0-08. | DeEscriPTIon. The cuticle is marked with fine transverse striations. The head bears a circle of ten cephalic setae, of which four are longer and stouter than the remaining six. Two sets of four short setae are positioned just anterior to and just posterior to the amphids. Setae are scarce on the remainder of the body. The amphids are basically circular in outline, but their margin is considerably expanded along the posterior edge (Fig. 5a), where the amphidal nerve exits. The buccal cavity is typical of the genus, containing a long cuticularized axial spear. The oesophagus is expanded posteriorly into an elongate bulb. The intestine is full of small opaque black granules, as in many other species in the genus. The | NEMATODES FROM THE INDIAN OCEAN 97 tail is conical in both sexes, and in the male bears two files of short subventral setae (Fig. 5c). Male. The male tail is often tightly coiled, and the spicules in all specimens are extruded. The spicules are paired, equal and arcuate. They have a central list which occupies their proximal two-thirds. The gubernaculum has strongly Fic. 5. Siphonolaimus profundus sp.nov. a, Lateral view of male head. b, Dorsal view of male head. c, Male tail. d, Lateral view of spicules, gubernaculum and posterior two supplements. 98 R. M. WARWICK developed dorsal apophyses (Fig. 5d). When the tail is coiled the cuticle for some distance in front of, and a short distance behind, the cloaca is thrown into folds. This obscures the arrangement of the supplements. In the uncoiled tail a series of distinct pre-cloacal supplements is visible. These are small, cup-shaped and weakly cuticularized. They number about 32, are more or less equally spaced, and extend about 5-7 tail lengths in front of the cloaca. Female. The female has a single anterior outstretched ovary. The uterus contains a single large and elongate egg in all specimens examined. Discussion. In the form of the amphids, tail and spicules this species appears to be very similar to S. gladiator Wieser, 1956. However, this species is much smaller, has only four cephalic setae, a different arrangement of cervical setae, a shorter spear and no supplements. Family DESMODORIDAE Desmodora masira sp. nov. (Fig. 6) MATERIAL STUDIED. 3 gd and 3 99 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 390-395 (2 dd and 2 99 measured below). Additional material: 3 gd, 4 99 and one head end; B.M. (N.H.), Reg. Nos. 1972: 396-403. REPRESENTATION. Station 5059. De Man’s Ratios. a b c V%_ Body length (mm) Males 57°40 14°64 10°63 - 2°87 56-98 14°25 10-98 - 3°02 Females 45°42 14°73 10°55 38-84 3:27 40:26 14:10 10:03 41°83 3:06 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0:050; 0-053. Oesophagus length: 0-196; 0-212. Length x breadth of posterior oesophageal bulb : 0-061 x 0-040; 0-064 x 0-045. Head diameter (at level of four long cephalic setae) : 0-032; 0-031. Amphid diameter: 0-0075; 0-0085. Length of four long cephalic setae: 0-008; 0-007. Tail length: 0-27; 0-275. Cloacal diameter : 0:038; 0-042. Spicule length: 0-063; 0-068. Gubernaculum length: 0-019 ; 0:023. Females: Body breadth: 0-072; 0-076. Oesophagus length: 0-222; 0-217. Length x breadth of posterior oesophageal bulb : 0-069 x 0-043; 0-072 x 0-047. Head diameter: 0-031; 0-033. Amphid diameter: 0-008; 0-008. Length of cephalic setae : 00075; 0-007. Tail length: 0-31; 0-305. Anal diameter : 0-031 ; 0:033. Distance of vulva from anterior: 1:27; 1:28. Egg length x breadth: O-II4 X 0-052; -. DEscRIPTION. The cuticle is marked with distinct transverse striations between the posterior edge of the cephalic capsule and a short distance from the tail tip. The head is bluntly rounded. There are six very short conical cephalic setae NEMATODES FROM THE INDIAN OCEAN 99 anterior to the cephalic capsule, and four longer cephalic setae at the anterior border of the capsule. The capsule itself is made of very thick cuticle, and is perforated all over with conspicuous pores. There are small patches of rather opaque sclerotization at the bases of the cephalic setae. The capsule is relatively short, its width being more than twice its height. The amphids are small, and describe a spiral of two turns (Fig. 6a). The buccal cavity is quite small and con- tains a rather indistinct small pointed dorsal tooth. Six files of short setae extend down the length of the body, these being most numerous in the cervical region and on the male tail. The oesophagus is swollen posteriorly into a pyriform bulb, the cuticular lining of which shows no special thickening. The nerve-ring encircles the oesophagus immediately anterior to this bulb. The tail is relatively long and is evenly tapered throughout its length. The cuticle on the tail tip is clear and unmarked (Fig. 6c). Fic. 6. Desmodora masiva sp.nov. a, Lateral view of male head. b, Lateral view of spicules and gubernaculum. c, Male tail. 100 R. M. WARWICK Male. The spicules are paired, equal and arcuate. They are strongly cephalated proximally and pointed distally, with ventral alae (Fig. 6b). The gubernaculum is small and plate-like. There are no pre- or post-cloacal supplements. Female. The ovaries are paired, symmetrical, opposed and reflexed. The eggs are quite large when only a few are present but there may be up to six in each uterus, and they appear smaller when squashed together. The measurement given above represents the maximum size. Discussion. Gerlach (1963) has extensively revised the genus Desmodora. The present species belongs to group IA in Gerlach’s subgeneric key, and can probably be ascribed to the subgenus Desmodora, although the cephalic capsule is shorter than in other members of the subgenus. D. mastra sp. nov. is characterized by the combination of a short broad perforate cephalic capsule, small amphids, a weakly armed buccal cavity, a relatively long tail and the absence of supplements in the male. Family COMESOMATIDAE Sabatieria alata sp. nov. (Fig. 7) MATERIAL STUDIED. 2 $4 and 2 9 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 404-407. Additional material: 3 92 and 4 juveniles; B.M. (N.H.), Reg. Nos. 1972 : 408-414. REPRESENTATION. Station 5050. De Man’s Ratios. a b c V %_ Body length (mm) Males 52°79 10:06 11-71 - 3°22 36-99 SFO 11°37 = 3°07 Females 42-71 9°34 10°31 48-16 2-99 36-10 8-69 10°30 49:28 2°78 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0-061; 0:083. Ocsophagus length:0-32; 0-31. Distance of nerve-ring from anterior:?; 0-148. Head diameter: 0-016; 0-0175. Length of longer cephalic setae : 0005; 0:0055. Amphid diameter: 0-010; 0-011. Body diameter at level of amphids: 0-018; o-o1rg. Tail length: 0-275; 0:27. Cloacal diameter : 0-051 ; 0-050. Spicule length : 0-082; 0-090. Gubernaculum length : 0-030; 0-036. Females: Body breadth: 0-070; 0-077. Oesophagus length : 0:32; 0:32. Dis- tance of nerve-ring from anterior: 0-155; 0-149. Head diameter: 0-019; 0-020. Length of longer cephalic setae : 0:0055; 0-005. Amphid diameter: 0-011; 0-010. Body diameter at level of amphids : 0-020; 0-022. Tail length: 0-29; 0-27. Anal diameter : 0-051; 0-051. Distance of vulva from anterior : 1-44; 1-37. DeEscrIPTION. The cuticle is marked with transverse rows of fine punctations. Laterally the punctations are larger, wider spaced and more irregularly arranged. The mouth is surrounded by six rounded lips each bearing a small conical labial NEMATODES FROM THE INDIAN OCEAN IOI papilla. Anteriorly there are six short conical cephalic setae, and more posteriorly four longer ones. Cervical setae are short and scattered. The amphids of the male describe a spiral of 3-0—3-2 turns, and those of the female 2-8-3-0 turns. The buccal cavity is simply cup-shaped, and bears no armament (Fig. 7a). The basal third of the oesophagus is slightly expanded, but there is no distinct bulb. The tail is relatively long, with its distal half filiform (Fig. 7b). Male. The spicules are paired, equal and arcuate. The proximal ends are only slightly cephalate, and from each a median list extends one-third of the way down the length of the spicule. When the spicules are protruded it can be seen that their ventral alae are produced into subterminal pointed projections near the distal ends (Fig. 7c). The gubernaculum bears a pair of pointed lateral projections in Fic. 7. Sabatieria alata sp.nov. a, Lateral view of female head. b, Male tail. c, Lateral view of spicule. 102 R. M. WARWICK addition to the long round-ended caudodorsal apophyses. There are about 21 very small and inconspicuous pre-cloacal supplements. These consist of simple pores in the cuticle which are rather widely spaced, the anteriormost being approxi- mately two tail lengths in front of the cloaca. Female. The ovaries are paired, equal, opposed and reflexed. Discussion. This species is very close to S. furcillata Wieser, 1954. However, Wieser’s species is smaller, has longer cephalic setae and relatively larger amphids. The spicules of the two species are differently shaped, the median list in S. furcillata being half the length of the spicule. Since species within the genus Sabatieria are generally separated on small but distinct characters, it is preferable to keep these two species distinct although the differences between them are quite small. Hopperia muscatensis sp. nov. (Fig. 8) MATERIAL STUDIED. I ¢ and 4 9° (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 415-419 (only I gf and 2 99 measured below). REPRESENTATION. Station 5071. DE Man’s Ratios. a b c V % _ Body length (mm) Male 43°72 8-74 14-14 - 1-88 Females 35°63 Q:12 13°99 46°49 2:28 31°13 8-50 12°85 45°70 PRG MEASUREMENTS (in mm in order of body lengths). Male: Body breadth : 0-043. Oesophagus length: 0-215. Distance of nerve-ring from anterior: 0-111. Dis- tance of excretory pore from anterior : 0-129. Head diameter: 0-012. Length of longer cephalic setae: 0-005. Amphid width: o-o10. Body diameter at level of amphids: 0-014. Tail length: 0-133. Cloacal diameter : 0-034. Spicule length: 0-087. Length of gubernacular apophysis : 0-032. Females : Body breadth : 0-064; 0-071. Oesophagus length: 0-25; 0:26. Dis- tance of nerve-ring from anterior: 0-129; ?. Distance of excretory pore from anterior : 0-142; ?. Head diameter: 0-012; 0-013. Length of longer cephalic setae: 0:007; 0:005. Amphid width: 0-009; 0-008. Body diameter at level of amphids: 0-014; o-or7. Tail length: 0-163; 0-172. Anal diameter : 0-041 ; 0-042. Distance of vulva from anterior: 1-06; I-or. Egg length x breadth: —; 0057 x 0-050. Description. The cuticle is marked by transverse rows of fine punctations. Lateral differentiation consisting of larger and more widely spaced punctations commences a short distance posterior to the amphids and terminates at the base of the conical portion of the tail. Very coarse lateral dots are present in the cloacal region of the male. The head bears two circles of six labial and six cephalic conical papillae. More posteriorly are four longer cephalic setae. The amphids describe a spiral of 3-1-3-2 turns and are slightly longer than wide. The mouth opens into NEMATODES FROM THE INDIAN OCEAN 103 a small globular cavity, the walls of which appear to be supported by fine cuticular- ized rods. More posteriorly the buccal cavity consists of a long heavily cuticularized tube, at the anterior end of which are three (one dorsal and two subventral) solid triangular teeth (Fig. 8a). A few very short scattered cervical, caudal and pre-anal setae are present, these being particularly numerous on the male tail. The oeso- phagus is swollen slightly towards its base, but there is no distinct bulb. Small cardia link the oesophagus with the intestine. The tail is relatively short, with its distal third filiform. It has a slightly swollen and rounded tip bearing a pair of short terminal setae (Fig. 8b). Male. The spicules are paired, equal and arcuate. The proximal third contains a median list, and there is a dorsal opening near the proximal tip. Each spicule bears a small tooth ventrally near the distal end (Fig. 8c). The gubernaculum has long slightly curved apophyses with rounded ends. Pre-cloacal supplements in the form of fine cuticular pores are present, but are very indistinct and cannot be enumerated with certainty. Fic. 8. Hopperia muscatensis sp.nov. a, Lateral view of male head. b, Male tail. c, Lateral view of cloacal region of male. Female. The ovaries are paired and opposed, but do not appear to be reflexed. The posterior ovary is a little longer than the anterior. In the only female which is gravid, the anterior uterus contains two eggs and the posterior uterus three. 104 R. M. WARWICK Discussion. Hopperia muscatensis differs from the type and only other species, H. massiliensts Vitiello, 1969, on several points, the chief of which are the longer cephalic setae, the different form of the spicules and the much shorter tail. Family LEPTOSOMATIDAE Platycomopsis implicatus sp. nov. (Fig. 9) MATERIAL STUDIED. I ¢ (Holotype); B.M. (N.H.), Reg. No. 1972 : 420. REPRESENTATION. Station 5025. De Man’s Ratios. a b c Body length (mm) Male 63°61 5°84 48-93 II-45 MEASUREMENTS (in mm). Male: Body breadth: 0-18. Oesophagus length : 1-96. Distance of nerve-ring from anterior: 0-47. Head diameter : 0-035. Length of cephalic setae : 0-025 + 0-015. Length of longest cervical setae : 0-027. Width of amphidal opening: 0-012. Tail length: 0-234. Cloacal diameter : 0-105. Spicule length : 0-095. Distance of supplement anterior to cloaca : 0-203. DEscrIPTION. The cuticle is smooth. The mouth is surrounded by three low rounded lips, each bearing a pair of short conical labial papillae. There are ten cephalic setae, six being longer than the remaining four. Just posterior to the amphids on the lateral surfaces of the body are several transverse rows of long stout cervical setae. On the left side from anterior to posterior there are four rows numbering six, three, two and two setae respectively. On the right side there are only three rows, numbering four, three and one setae (Figs. ga and b). On the dorsal surface there are similarly two rows of four and two long setae, but ventrally there are only two very short conical setae and there is no indication that any longer setae have been lost. More posteriorly there are a few very short setae in the cervical region and on the tail tip, but the middle region of the body is naked. The cephalic capsule is very weakly developed. It takes the form of a narrow band of sclerotization of the inner cuticle surrounding the head, and is only visible in optical section where its curves round the sides of the head. There is virtually no buccal cavity; the mouth opening merges imperceptibly into the lumen of the oesophagus. The amphids are pocket-like, with transverse oval openings. The oesophagus is cylindrical. The tail is conical but has a short cylindrical tip. There is no spinerette and no caudal glands are visible (Fig. gc). Male. The spicules are paired, equal and straight with a rather complex folded structure (Fig. 9d). They have funnel-shaped openings at their proximal ends. The gubernaculum is also complex and is folded into a pair of tubes surrounding each spicule. The pre-cloacal supplement comprises a simple mid-ventral papilla pierced by a fine pore from which a duct leads to the interior. Discussion. Mawson (1956) provides a useful key to the species of this genus. Platycomopsis implicatus appears to be closest to P. dimorphica Mawson, 1956. NEMATODES FROM THE INDIAN OCEAN Fic. 9. Platycomopsis implicatus sp.nov. a, Male head from left side. b, Male head from right side. c, Male tail. d, Lateral view of spicules and gubernaculum. 106 R. M. WARWICK This latter species has single long dorsal and ventral cervical setae and the post- amphidal setae are less numerous than in the present species. The aperture of the amphids seems to be very much wider in Mawson’s species. She found it difficult to discern the structure of the spicules, because of the thickness of the tail, but in her drawing she shows the spicule narrowing abruptly near the proximal end, and not widening into a funnel-shaped opening. A final point of distinction between the two species is the position of the pre-cloacal supplement. In P. dimorphica it is situated less than one spicule length in front of the cloaca, whereas in P. implicatus the corresponding distance is more than twice the spicule length. Synonchus alisonae sp. nov. (Fig. 10) MATERIAL STUDIED. 3 gg and 2 99 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 421-425. Additional material: 1 g and g juveniles, one of which is beheaded for an en-face preparation; B.M. (N.H.), Reg. Nos. 1972 : 426-435. REPRESENTATION. Station 5021. De Man’s Ratios. a b c V %_ Body length (mm) Males 58-50 7-18 27:86 II-70 47°08 6:28 15°43 - 10-49 50°15 7°38 13°74 = 10:03 Females 51-60 Tee 14°54 47°07 10°32 54:26 7-01 13°22 40:07 10°31 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth : 0-20 ; 0:22; 0:20. Oesophagus length:1:63; 1-67; 1:36. Head diameter: 0-043; 0°054; 0-052. Length of cephalic setae: 0-016 + 0-014; 0-018 + 0-017; 0-018 + 0-014. Amphid width:0-015; 0-014; ?. Tail length:0-42; 0-68; 0:73. Cloacal diameter:0-13; 0:13; O11. Spicule length:0-175; 0-181; 0-168. Gubernaculum length : 0-095; 0-078; 0-081. Distance of pre-cloacal supplement from cloaca : 0:134; O:II2; 0-103. Females: Body breadth: 0-20; 0-19. Oesophagus length: 1-42; 1-47. Dis- tance of nerve-ring from anterior:0-39; ?. Head diameter: 0-056; 0-054. Length of cephalic setae : 0-017 + 0-013; 0-017 + 0-014. Amphid width: 0-016; 0-015. Tail length:o0-71; 0-78. Anal diameter: 0-10; o-11. Distance of vulva from anterior : 4-92; 4:75. Egg length x breadth: 0-42 x 0-16; -. DESCRIPTION. The cuticle is marked with very fine transverse striations. The mouth is bounded by three rounded lips each bearing a pair of small conical labial papillae. Internally the cuticle of each lip is modified into a small transverse bar from which radiate a series of longitudinal striations, becoming progressively shorter towards the interlabial positions (Figs. rob, c and d). These are similar to the semi-lunar striations (terminology of Inglis, 1964) found in many members of the Enoplidae. There are ten cephalic setae, the submedian pairs being only NEMATODES FROM THE INDIAN OCEAN 107 slightly subequal in length. The cephalic capsule is relatively narrow and its posterior border is roundly lobed between the cephalic setae. The amphids are pocket-like with oval openings. They are displaced slightly dorsally. There are two or three short post-amphidal setae and a few other short setae scattered pos- teriorly to about the level of the nerve-ring. The buccal cavity varies in appearance depending on the angle from which it is viewed. It is basically conical and con- tains a large solid dorsal tooth with a rounded tip. There are no subventral teeth visible. There is no sexual dimorphism shown in the head. The oesophagus is Pin TT Fic. 10. Synonchus alisonae sp. nov. a, Male tail. b, En-face view of head. cand d, Heads of two males. e, Lateral view of spicules, gubernaculum and supplements. 108 R. M. WARWICK cylindrical. The tail in both sexes is filiform with a pointed tip, but varies from 3:2 to 7-1 cloacal or anal diameters in length. Male. The spicules are paired, equal and arcuate, with no marked proximal cephalization. The gubernaculum is of a complex structure. From the lateral view it has a triangular outline, and bears a solid rounded dorsal apophysis (Fig. toe). There is a pair of ventral mamilliform supplements a short distance in front of the cloaca. Rows of subventral setae extend anteriorly from the cloaca about four times the supplement-cloaca distance, and posteriorly for some way down the length of the tail (Fig. 10a). Female. The ovaries are paired, symmetrical and reflexed. The eggs are relatively elongate. Discussion. This species is characterized by the combination of a long filiform tail and relatively long cephalic setae. It appears to be closest to S. filicaudatus (Ditlevsen, 1926), another deep-water species. In Ditlevsen’s species, however, the tail and cephalic setae are much shorter, and the gubernacular apophysis is tri- angular in shape. Family PHANODERMATIDAE Dayellus dayi Inglis, 1964 (Figs. Ira and b) MATERIAL STUDIED. 13 gd, 18 92 and 14 juveniles (2 gg and 2 92 measured below) ; B.M. (N.H.), Reg. Nos. 1972 : 436-481. REPRESENTATION. Station 5047:7 gd, 5 92 and 7 juveniles. Station 5060: 6 3d, 13 9Y and 7 juveniles. DE MAn’s Ratios. a b c V%_ Body length (mm) Males 58-04 8-22 27°53 - 5°92 47°79 8-82 20°19 - 6:26 Females 62:30 10-92 31°74 51-96 8-41 45°79 9°03 26°71 53°67 6°41 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0°102; 0-131. Oesophagus length:0-72; 0-71. Distance of nerve-ring from anterior : 0-275; 0:26. Distance of excretory pore from anterior: 0-058; 0-058. Head diameter: 0-012; 0-013. Length of cephalic setae : 0-014 + 0-0125; 0-016 + 00145. Tail length: 0-215; 0-239. Cloacal diameter: 0-071; 0:075. Spicule length : 0-130; 0-135. Gubernaculum length : 0-043; 0-045. Supplement length : 0-016; 0-018. Distance of supplement from cloaca : 0-036; 0-033. Females : Body breadth : 0-135; 0-140. Oesophagus length:0-77; 0-71. Dis- tance of nerve-ring from anterior: 0-267; ?. Distance of excretory pore from anterior: ?; 0-066. Head diameter: 0-013; 0-014. Length of cephalic setae: 0-014 + 0°0125; 0-015 + 0-013. Tail length:0-265; 0-24. Anal diameter: 0-082 ; 0-060. Distance of vulva from anterior : 4:37; 3°44. Egg length x breadth: 0°165 x 0-090; -. NEMATODES FROM THE INDIAN OCEAN 109 DESCRIPTION. The specimens conform exactly to the description provided by Inglis (1964), except that they are slightly smaller in many measurements, par- ticularly the cephalic setae. The largest number of eggs found in any female is 18, Io in the anterior uterus and 8 in the posterior uterus. DISTRIBUTION. South Africa, South-east Arabia. Crenopharynx fringilla sp. nov. (Figs. r1c, d and e) MATERIAL STUDIED. 9 gd, 10 99 and 12 juveniles. REPRESENTATION. Station 5047:7 gd and 2 99 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 482-490 (only 3 gd and 2 99 measured below). Additional material: 7 99 and 12 juveniles; B.M. (N.H.), Reg. Nos. 1972: 491-509. Station 5060:2 gg andr; B.M. (N.H.), Reg. Nos. 1972 : 510-512. De Man’s Ratios. a: b c V% _ Body length (mm) Males 32°34 4°76 I5'51 ~ 5:66 39°79 4°57 15°47 = SOT, 40:00 4°63 14°74 ~ 5:60 Females 36-22 5°13 15°90 54°14 6°52 34°63 4°63 15°48 56-23 6:58 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth: 0175; 014; 0-14. Oecsophagus length: 1-19; 1-22; 1-21. Distance of nerve- ring from anterior: 0-42; 0-44; 0-44. Head diameter: 0-019; 0-019; 0-o1Q. Length of cephalic setae : 0-018 + 0-014; 0-015 + 0-012; ?. Tail length : 0-365 ; 0:36; 0:38. Cloacal diameter: 0-068; 0-064; 0-073. Spicule length : 0-345 ; 0:36; 0:37. Gubernaculum length: 0-081; 0-078; 0-095. Females: Body breadth: 0-18; 0-19. Oesophagus length: 1-27; 1-42. Dis- tance of nerve-ring from anterior: 0-43; 0-48. Head diameter: 0-019; 0-021. Length of cephalic setae:?; 0-017 + 0-014. Tail length:o-41; 0-425. Anal diameter : 0-073; 0-075. Distance of vulva from anterior : 3:53; 3°70. DEscripTion. The cuticle is smooth. The neck region is strongly tapered so that the head is relatively small. The head structure is exactly as described for other members of the genus by Inglis (1964). The cephalic capsule is lightly built with a lobed posterior margin, and the anterior end of the oesophagus projects anteriorly as three lobes (Fig. tre). The cuticle supporting the six labial papillae which surround the mouth is sclerotized into crescentic areas. The ten cephalic setae are situated at the posterior edge of the cephalic capsule, six being slightly longer than the remaining four. A few scattered setae are present on the anterior end of the body, extending back about one-third of the oesophagus length. Setae are scarce on the middle of the body, but become more numerous again on the tail. The oesophagus is strongly tapered in association with the attenuation of the neck region, and becomes cellular posterior to the nerve-ring. The tail is elongate, with its distal end flagellate (Fig. 11Cc). I1o R. M. WARWICK Va ii NNUAL \NN I AN \\ ue NN KY \\ AN WW iW WW N \\ WN Fic. 11. Dayellus dayi Inglis, 1964. a, Male tail. b, Lateral view of male head. Crenopharynx fringilla sp. nov. c, Male tail. d, Lateral view of gubernaculum and distal end of spicules. e, Lateral view of male head. a4 NEMATODES FROM THE INDIAN OCEAN III Male. The spicules are elongate, paired and equal. They are slightly cephalate proximally and pointed distally. They are transversely striated for most of their length, except for short clear zones at the distal and proximal ends. Their ventral edges are not modified by any barbs or swellings. The gubernaculum has a com- plex folded structure and bears a pair of extensive lateral pieces terminating in sharp points (Fig. 11d). There is no pre-cloacal supplement. Female. The ovaries are paired, equal, opposed and reflexed. One of the females not measured above contains three eggs in each uterus, a typical example measuring 0-216 x 0-097 mm. Discussion. This species is characterized by the combination of a long flagellate tail, spicules with no barbs or other modifications, and a gubernaculum with pointed lateral pieces. C. crassus (Ditlevsen, 1930) and C. afra Inglis, 1964, appear to be most nearly related to C. fringilla, but neither has the above combination of characters. Family ENOPLIDAE Mesacanthion southerni sp. nov. (Figs. 12 and 13a and c) MATERIAL STUDIED. I g and 3 99 from Station 5034; 2 gg from Station 5047 (Syntypes); B.M. (N.H.), Reg. Nos. 1972 : 513-518. Additional material: 3 g¢ and 2 juveniles from Station 5047; B.M. (N.H.), Reg. Nos. 1972 : 519-523. Setae of specimens from Station 5047 are generally damaged. REPRESENTATION. Station 5034:1 g¢ and 3 99. Station 5047:5 gg and 2 juveniles. DE Man’s Ratios. a b c V% Body length (mm) Males 33°91 5°57 14°44 = 3°90 32°10 5°81 14°65 = 337 32°80 ? 14:91 ~ 3°28 Females 25°88 5°37 12°57 50-91 4°40 29°82 5°51 13°70 51:28 5°07 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth : 0-115 ; 0-105; 0-100. Oesophagus length: 0-70; 0:58; ?. Distance of nerve-ring from anterior: 0:23; ?; ?. Head diameter : 0:033; 0:035; 0-035. Length of labial setae:0-012; ?; ?. Length of lateral cephalic setae: 0-057; ?; ?. Length of submedian cephalic setae : 0-048 + 0:039; ?; ?. Tail length: 0-27; 0:23; 0:22. Cloacal diameter: 0-060; 0:055; 0-053. Length of right spicule: 0-32; 0-29; 0-30. Length of left spicule:0:080; 0-067; 0-078. Length of gubernacular apophysis : 0-051; 0:045; 0-040. Length of supplement: 0-040; 0-034; 0-032. Distance of supplement anterior to cloaca : 0-100; 0-098; 0-082. Females: Body breadth:0-17; 0-17. Oesophagus length: 0-82; 0-92. Dis- tance of nerve-ring from anterior: 0:26; 0:27. Head diameter: 0-039; 0-041. Length of labial setae : 0-013; 0-012. Length of lateral cephalic setae: 0-48; ?. 112 R. M. WARWICK Length of submedian cephalic setae : 0-046 + 0-038; 0-038 + 0:033. Tail length : 0°35; 0:37. Anal diameter: 0-062; 0-068. Distance of vulva from anterior : 2:24; 2:60. Egg length x breadth:—; 0:20 x o-II. DESCRIPTION. The cuticle is marked internally with faint transverse striations. The neck region tapers considerably from the base of the oesophagus: it is highly setose in the male (Fig. 13c), but only a few short scattered setae are present in the female. The head is typical of the genus. The three lips are relatively low. The lateral cephalic setae are slightly longer than the longer setae of each submedian pair. The cephalic capsule is punctate and deeply incised, with the cephalic setae situated at its anterior end (Fig. 12a). The mandibles have the usual appearance of two lateral rods united by an anterior curved bar. The three onchia are rela- tively large, and are equal in size. The oesophagus broadens considerably towards its base, but there is no posterior bulb. The tail is more or less evenly tapered throughout its length in both sexes (Fig. 13a). Fic. 12. Mesacanthion southerni sp.nov. a, Lateral view of female head. b, Lateral view of gubernaculum, left spicule and distal end of right spicule. c, Ventral view of gubernaculum, left spicule and distal end of right spicule. Male. The spicules are asymmetrical. The right is elongate and transversely striated for most of its length, except near the distal and proximal ends (Fig. 13a). The left is short and smooth and is built in two halves with a fine suture between them. The proximal half bears a strong dorsal hook at its distal end, and the ee NEMATODES FROM THE INDIAN OCEAN 113 distal half terminates in four outwardly curving teeth (Figs. 12b and c). The gubernaculum is in two halves, each with an elongate posterior apophysis. The supplement is a simple cuticularized tube situated about half the right spicule length in front of the cloaca. Female. The ovaries are paired, symmetrical, opposed and reflexed. The gravid specimen contains only one egg per uterus. Fic. 13. Mesacanthion southerni sp.nov. a, Male tail. c, Anterior end of male. Mesacanthion diplechma (Southern, 1914). b, Male tail (drawn to same scale as a). 114 R. M. WARWICK Discussion. Because of the close similarity of the male reproductive complex, these specimens were at first assigned to M. diplechma (Southern, 1914). The spicular apparatus of this species has been redescribed by Ditlevsen (1934). How- ever, close comparison with specimens of M. diplechma from the North Sea off Northumberland (Station A of Warwick and Buchanan, 1970) have revealed several differences which warrant specific rank. The male tail of M. diplechma, drawn from a North Sea specimen (B.M. (N.H.), Reg. No. 1972 : 538) is illustrated for comparative purposes in Fig. 13b. It will be noted that the right spicule of M. diplechma is much more elongate (0-54 mm in the illustrated specimen), so that the distance between the supplement and cloaca is less than a quarter of the right spicule length. M. diplechma bears a ventral setose papilla near the tail tip, which is absent in M. southerni. Another point of distinction is that the distal tip of the short spicule of M. diplechma bears three teeth (a point on which Southern and Ditlevsen agree), whereas M. southerni has four. Mesacanthion arabium sp. nov. (Fig. 14) MATERIAL STUDIED. 3 46 and 3 99 (Syntypes); B.M. (N.H.), Reg. Nos. 1972: 524-529. Additional material: 1 g and7 99; B.M. (N.H.), Reg. Nos. 1972 : 530- 537: REPRESENTATION. Station 5034. De Man’s Ratios. a b c V% _ Body length (mm) Males 31°25 4:81 18-38 - 6:25 37°00 4°93 16-91 = 5°92 30°42 5:21 16-06 - 5°78 Females 33°42 5:16 16:71 53°07 6:35 38-76 4:81 17°34 52°05 6°59 32°85 4°94 16°85 51°45 6°57 MEASUREMENTS (in mm in order of body lengths). Males: Body breadth : 0-20 ; 0-16; 0-19. Oecesophagus length: 1-30; 1:20; I-11. Distance of nerve-ring from anterior : 0°32; 0:32; 0-28. Head diameter: 0-080; 0-080; 0-082. Length of labial setae: 0-025; 0-023; 0-024. Length of cephalic setae : 0:056 + 0-032; 0:065 + 0:030; 0:058 + 0:027. Length of cervical setae: 0-034; 0-029; 0-032. Tail length : 0-34; 0:35; 0:36. Cloacal diameter : 0-092; 0-090; 0-092. Spicule length: 0:60; 0-61; 0:57. Gubernaculum length:0-126; 0-127; 0-120. Sup- plement length : 0-062; 0-059; 0-058. Distance of supplement anterior to cloaca : 0:23)5)10;225)50222. Females: Body breadth: 0-19; 0:17; 0:20. Oesophagus length: 1-23; 1-37; 1-33. Head diameter: 0-085; 0-085; 0-086. Length of labial setae: 0-024; 0-024; 0-024. Length of cephalic setae : 0-052 + 0:023; 0:055 + 0:027; 0-055 + 0-026. Tail length: 0-38; 0-38; 0:39. Anal diameter: 0-083; 0-080; 0-081. Distance of vulva from anterior : 3:37; 3°43; 3°38. NEMATODES FROM THE INDIAN OCEAN 115 Description. The cuticle is smooth. The head is dome-shaped and the three lips relatively low. The four shorter cephalic setae are only about half the length of the remaining six. The cephalic capsule is deeply incised, with the cephalic setae situated at its anterior end (Fig. 14b). The posterior lobes of the capsule are lightly punctated. The mandibles are similar to those of the previous species, but much more heavily built. Their outer surface is sculptured with a series of a Fic. 14. Mesacanthion avabium sp. nov. a, Male tail. b, Lateral view of male head. c, Ventral view of vulvar organ of female. d, Lateral view of female vulvar region. e, Lateral view of gubernaculum and distal tip of spicules. f, Ventral view of male cloacal region. g, Lateral view of male pre-cloacal supplement. 116 R. M. WARWICK short radiating lines. The onchia are equal in size, relatively smaller than in the previous species. The male bears a pair of cervical setae at the base of each incision of the capsule (i.e. 12 setae in all), and setae approximately the same length as the cervicals extend backwards from the head in six files approximately to the level of the nerve-ring. The paired cervicals and other setae are absent in the female. The oesophagus is cylindrical. The posterior half of the tail is cylindrical (Fig. 14a). In one female the three caudal glands extend 0-53, 0:63 and 0-77 mm in front of the anus. Male. The spicules are elongate, and transversely striated along most of their length. The proximal tips (not illustrated) are non-cephalate. The gubernaculum consists of a pair of L-shaped pieces united in the middle by ventral strips (Figs. 14e and f). The distal ends bear a series of raised ridges. The supplement is a cuticularized tube of rather more complex structure than in most other members of the genus (Fig. 14g). Two subventral files of stout setae are present between the cloaca and the proximal tips of the spicules. They become more numerous near the cloaca. Female. The ovaries are paired, symmetrical, opposed and reflexed. No speci- mens contain fertilized eggs. Transversely oval apparently glandular structures are present ventrally anterior and posterior to the vulva in eight of the ten avail- able specimens. These are very prominent and open by long transverse slits (Figs. 4c and d). They probably aid adhesion to the male during copulation. They may be up to five in number, the distribution in the ten specimens being as follows : ain Or wo WAS g A A B a se i i © POSWEHO Wo Wwe 2 A Ao uw 2 wm ur © © Discussion. This species is characterized by the elongate spicules, the structure of the gubernaculum, the arrangement of cervical and cloacal setae and the organs associated with the vulva. In this latter feature it is unique. REFERENCES AtiGEn, C. A. 1927. Freilebende marine Nematoden von den Campbell- und Staten-Inseln. Nyt. Mag. Naturvid. 66 : 249-309. Cops, N. A. 1920. One hundred new nemas. (Type species of 100 new genera.) Contr. Sct. Nematology, 9 : 217-343. DiTLEvseNn, H. 1926. Freeliving nematodes. Dan. Ingolf Exped. 4 (6) : 1-42. 1930. Papers from Dr Th. Mortensen’s Pacific Expedition 1914-16. LII. Marine freeliving nematodes from New Zealand. Vidensk. Medd. dansk. naturh. Foren. 87 : 201- 242. 1934. Some remarks on the spicular apparatus of Enoplolaimus diplechma (Southern). Vidensk. Medd. dansk. naturh. Foren. 97 : 207-210. FREUDENHAMMER, I. 1970. . Sphaerolaimus uncinatus nov. spec. (Nematoda, Monhysterida) aus der Tiefsee. Vevdff. Inst. Meeresforsch. Bremerh. 12 : 455-461. GerracH, S. A. 1953. Die Nematodenbesiedlung des Sandstrandes und des Kiistengrund- wassers an der italienischen Kiiste. Aycho zool. ital. 37 : 519-640. 1956. Diagnosen neuer Nematoden aus der Kieler Bucht. ieley Meevesforsch. 12: 85-109. 1963. Freilebende Meeresnematoden von den Malediven. II. Kieley Meevesforsch. 19: 67-103. NEMATODES FROM THE INDIAN OCEAN Er, Inciis, W. G. 1961. Freeliving marine nematodes from South Africa. Bull. By. Mus. nat. Hist. (Zool.). 7 : 291-319. 1964. The marine Enoplida (Nematoda) : a comparative study of the head. Bull. Br. Mus. nat. Hist. (Zool.). 11 : 263-376. LORENZEN, S. 10969. Freilebende Meeresnematoden aus dem Schlickwatt und den Salzwiesen der Nordseekiiste. Vevdff. Inst. Meeresforsch. Bremerh. 11 : 195-238. Dre Man, J. G. 1884. Die frei in der veinen Evde und im siissen Wasser lebenden Nematoden dey niederldndischen Fauna. Eine systematische-faunistische Monographie. Leiden. Mawson, P. M. 1956. Freeliving nematodes. Section I. Enoploidea from Antarctic stations. Rep. B.A.N.Z. antarct. Res. Exped. Ser. B, 6 (3) : 39-74. RoyaL Society. 1963. International Indian Ocean Expedition. R.R.S. ‘ Discovery ”’. Cruise 1. South-east Arabian upwelling region. Cruise Report. SOUTHERN, R. 1914. Clare Island Survey. Part 54. Nemathelmia, Kinorhyncha and Chaetognatha. Proc. R. Iv. Acad. 31: 1-80. VITIELLO, P. 1969. Hopperia, nouveau genre de nématode libre marin (Comesomatidae). Téthys. 1: 485-491. 1969a. Linhomoeidae (Nematoda) des vases profondes du Golfe du Lion. Téthys. 1: 493-527. Warwick, R. M. & Bucuanan, J. B. 1970. The meiofauna off the coast of Northumber- land. I. The structure of the nematode population. J. mar. biol. Ass. U.K. 50: 129-146. WIESER, W. 1954. Freeliving marine nematodes. II. Chromadoroidea. Reports of the Lund University Chile Expedition 1948-49. Acta Univ. lund. (N.F.) 50 (16) : 1-148. — 1956. Freeliving marine nematodes. III. Axonolaimoidea and Monhysteroidea. Re- ports of the Lund University Chile Expedition 1948-49. Acta Univ. lund. (N.F.) 53 (13) : I-II5. oS RicHarp M. Warwick Dove Marine LABORATORY CULLERCOATS Nortu SHIELDS, NORTHUMBERLAND + A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. ALIson. Marine Molluscs in the Cuming Collection British Muset (Natural History) described by William Harper Pease. Pp. 96; 14 Plates 1965. (Out of Print.) £3.75. # WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier and ¥ Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. = TayLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy _ of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates 77 Text-figures. 1969. £4.50. Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antur 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. 15 ocT (973 Va PRESENTED 4 Lon,, wot Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 sNU E BIOLOGY AND FINE STRUCTURE OF _ EUGLYPHA ROTUNDA | (TESTACEA: PROTOZOA) R. H. HEDLEY AND C. G. OGDEN BIOLOGY AND FINE STRUCTURE OF BUECLYPHAYROEFUNDA (DESWACEA: PROTOZOA) Vall Miss fos aS aber waisa. * * SJULI973 | Ro PRESERTED x SOc 3 —— BY RONALD HENDERSON HEDLEY AND COLIN GERALD OGDEN Pp. 119-137 ; 7 Plates, 3 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 4 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), ‘stituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Paris will appear at trregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25 No. 4 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 26 June, 1973 Price £1.60 BIOLOGY AND FINE STRUCTURE OF EUGLYPHA ROTUNDA (TESTACEA: PROTOZOA) By R. H. HEDLEY & C. G. OGDEN CONTENTS Page SYNOPSIS 4 5 : 9 . c a 0 . 3 5 121 INTRODUCTION 5 a 6 ° : : 9 9 : 121 Previous work — biology ¢ é : 6 8 d 0 4 122 Previous work — taxonomy ¢ 6 : : 3 : ; 123 MATERIAL AND METHODS . C 7 é ¢ : : : ¢ 123 DISTRIBUTION . 6 0 : 5 5 3 5 . S 125 MORPHOLOGY AND VARIATION. 6 é : , : 6 : 125 REPRODUCTION 0 : 0 : a : 3 6 ¢ : 126 Foop ; 6 S : 0 a 6 : : 0 a : 127 CYTOLOGY 3 : . é q 6 : : 6 6 128 Vegetative stage 0 : ° : : . ° : 128 Siliceous plates and cement. : : F a 3 : 3 129 Pseudopodia_ . j é é c : : , ‘ : 130 Rosette groups . ; Q a ; : A : 0 : 132 Cyst - . : é ‘ a : : : é 133 AcxnowLepGEMEnts 3 a 6 5 : . 6 134 REFERENCES . é - 9 b 0 . ¢ 0 : é 134 SYNOPSIS Euglypha rotunda a fresh-water, moss and soil-inhabiting testacean has been established in clonal culture with a doubling time of between 33 and 45 hours; it is capable of reproducing in monoxenic culture with the bacterium Klebsiella aerogenes (strain NCIB 8017). Full descriptions are given of the non-crystalline siliceous shell, the vegetative stage and the process of reproduction by simple division. A detailed account of cytoplasmic fine structure reveals certain features of special interest : subpellicular microtubules ; two or three contractile vacuoles each with a surrounding spon- giome; a single well-developed Golgi apparatus with associated smooth and coated vesicles, and reserve siliceous plates stacked nearby ; a well-defined region of endoplasmic reticulum, with a high concentration of ribosomes, surrounding the nucleus; the presence of a poly- saccharide cement or glue binding the siliceous plates of the shell; microfilaments in the pseudopodia ; and the occurrence of cross-striated fibrils in the cyst membrane of encysted individuals. The habit of individuals clumping together to form rosette groups with cytoplasmic con- nections between them is described ; the alignment of masses of microfilaments in these regions to form stress areas or desmosome-like structures is noted. INTRODUCTION TuE widely distributed, fresh-water, soil- and moss-inhabiting rhizopods belonging to the genus Euglypha have attracted little attention by cytologists and proto- zoologists since an early study of E. alveolata by Schewiakoff (1888). Most of the 122 Rk. H. HEDLEY, AND C. G. “OGDEN work on these organisms has been concerned with faunal surveys of testate amoebae ; no study of the cytoplasmic ultrastructure has been reported previously. The present account concerns the biology and fine structure of cultured specimens of E. rotunda Wailes, 1911, isolated from collections made in Battersea Park, London, England. Previous work — biology The first general study of the cytoplasm of Euglypha was made by Carter (1865) who compared the internal structures of several protozoa. He observed that in Euglypha the nucleus normally occupies a posterior position, that union between two individuals is common and that there appear to be two types of structure associated with the division — ‘ovules’ (globular, nucleated ‘ cells’) between four and fifty in number, and ‘ spermatozoids’ (granules) both of which are situated close to the nucleus. Several authors (Gruber, 1881; Blochmann, 1887 ; Schewiakoff, 1888 ; Penard, 1890; Popoff, 1912, and Wailes, 1915) have described some aspects of nuclear division or reproduction in FE. alveolata. Schewiakoff (1888) described and illustrated reproduction and nuclear division in E. alveolata. Penard (1890) reported that at conjugation the apertural teeth of two individuals locked together, and suggested that the increase in cytoplasmic volume at division was the result of water uptake. Blochmann (1887) observed copulation resulting in the formation of one extra large daughter cell. The types of reproduction were summarized by Wailes (1915) who suggested that there are three alternative methods: (i) spores which become amoeboid individuals, (ii) budding and (iii) simple division into two animals. Penard (1890) reported that the pseudopodia of Euglypha appear to be stronger than those of naked amoebae and are apparently characteristic for each species. He also considered that the small granules in the body are not excretion granules as previously suggested by Biitschli (1880). The perinuclear cytoplasm was observed to be darkened with Giemsa stain by Popoff (1912). A detailed study of the cyto- plasmic inclusions of E. alveolata by Hall & Loefer (1930) showed that the number of refractile bodies in the granular zone is approximately inversely proportional to the number of reserve shell-plates ; the authors concluded from this that these refractile bodies might be the origin of the reserve plates. Penard (1890) reported that at encystment the cytoplasm is contained behind a membrane within the shell and suggested that this is probably resorbed during excystment. He also described the unusual encystment of a large individual, in which numerous small, nucleated cysts could be seen. The same author (1902 ; 1940) described the encystment of E. brachiata (armata) ; in this species the cyst develops inside a smaller secondary shell which is formed within the parent shell. From a comparison of encystment in several testaceans, Thomas (1962) suggested that the large encysted test formed after copulation enabled the dissemination of the biotypes. In the most recent review of the biology of soil protozoa, Stout & Heal (1967) comment on the ecology and nutrition of some testaceans, including Euglypha. BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 123 In conclusion it appears that apart from a lot of information on the distribution of Euglypha we have a very limited understanding of any aspect of its biology. Previous work — taxonomy Several species of Euglypha were described by Penard (1890; 1902) before the first description of E. rotunda by Wailes (in Wailes & Penard, 1911). This species was considered to be distinct from E. Jaevis by the denticulate apertural-plates, the circular aperture and oval shape of the shell ; from £. strigosa by its size, un- thickened apertural-plates and absence of spines; from E. ciliata by its size and oval aperture and from E. alveolata var. minor by its shell-plates and oval shape as seen in section through the minor axis. It was recorded again by Wailes (1912) from the Shetland and Orkney Islands, who later (Wailes, 1915) redescribed the apertural-plates. In a review of the British Freshwater Rhizopoda fauna, Cash et al. (1915) described fourteen species of Euglypha, including E. rotunda, and recorded some additional localities. Some sizes of E. rotunda from soil in Italy were given by Grandori & Grandori (1934). The variation in shapes of apertural-plates was described by Decloitre (1950), who described the shell-plates and gave further data on size. In a later review of the genus, Decloitre (1962) lists descriptions for numerous specimens including E. rotunda. Recently, Decloitre (1964) reported upon the variation in size between specimens from Ceylon and Tasmania. Two recent reports (Cambar e¢ al., 1964 and Mercier et al., 1964) have described the surface ultrastructure of the shell of EZ. strigosa and a Euglypha sp., from metal and carbon-shadowed preparations. The classification of the family Euglyphidae adopted here is that proposed by Loeblich & Tappan (1961) : Class RHIZOPODEA Von Siebold, 1845 Subclass FILOSIA Leidy, 1879 Order GROMIDA Claparéde & Lachmann, 1859 Superfamily EUGLYPHACEA Loeblich & Tappan, 1961 Family EUGLYPHIDAE Wallich, 1864 ; test hyaline, symmetrical, elongate, composed of rounded siliceous scales, aperture rounded or elongate; one nucleus. MATERIAL AND METHODS Euglypha rotunda was isolated from samples of sphagnum moss, weed and mud collected from ponds in Battersea Park, London, in November, 1967. Initially, rough cultures were obtained by allowing small portions of these samples to stand in a shallow layer of the culture liquid at room temperature, 18-20°C. Agnoto- biotic cultures were kept in small plastic containers on a thin layer of agar agar (I per cent in distilled water), and covered by a shallow layer of culture liquid. A sterilized wheat grain was added to the agar just before it set. The basal culture liquid was a 5 per cent (w/v) solution of soil extract in distilled water, to which 124 R. H. HEDLEY AND C. G. OGDEN nutrient salts were added so that the final culture medium contained 100 mg/l“ sodium nitrate and 15 mg/l-! sodium dihydrogen orthophosphate. Single active animals were isolated to produce clonal cultures. Of twenty such cultures started only four reproduced, three within four days and the other after fourteen days. One clone was used subsequently to produce the working cultures and this clone is now registered (No. 1520/1) and maintained at the Culture Centre of Algae and Protozoa, The Natural Environment Research Council, Cambridge, England. The animals which feed and reproduce readily were sub-cultured approxi- mately every four or five weeks. Cultures made available to Dr H. Netzel (Zoologischen Institut der Universitat Tiibingen) have been used to produce cine-photomicrographs of locomotion and reproduction (Netzel, 1971). Optical microscopy The animals were examined either live, or after fixation, by both phase-contrast and bright-field microscopy. Smears fixed in Schaudinn’s fluid were stained with borax carmine and iron haemotoxylin. Clumps of animals for sectioning were fixed in Schaudinn’s and embedded in agar (Harris, 1965), the sections being stained with either Mayer’s haemalum, toluidine blue or Mallory’s triple stain. Scanning electron microscopy Specimens removed from the cultures were either cleaned by transferring them directly from the culture vessels through several changes of triple-distilled water or, alternatively, first fixed in osmium textroxide or glutaraldehyde and then washed. They were subsequently pipetted with a minimum of water onto a cover- slip, which had been cleaned in acetone and polished with lint-free tissue, to which they adhere quite well on drying. Single specimens for use in the examination of individual siliceous shell-plates were cleaned by immersion in a slightly acidic solution of hydrogen peroxide (20 vol.) for approximately thirty minutes, prior to washing. Suitably clean animals were then placed individually onto small pieces of cleaned cover-slip. To this a drop of concentrated sulphuric acid was added, and evaporated by gentle heating. This liberates the plates from the organic cement material. The prepared cover-slips were then glued to a ‘ Stereoscan’ specimen stub with ‘Silver Dag’, an electrically conductive paint, and coated evenly with ro-15 nm of gold in a coating-unit using the apparatus described by Harris et al. (1972). Specimens were examined in a Cambridge Stereoscan Mk II at either 15 or 20 kV, and the results recorded on Ilford HP3 35 mm film. Transmission electron microscopy The animals were fixed at room temperature for 15 minutes in I per cent glu- taraldehyde in 0:05 M Sorensen’s phosphate buffer with 0-015 M calcium chloride, followed by 10 minutes in 3 per cent glutaraldehyde in the same buffer, and finally post-fixed in 1 per cent osmium tetroxide in 0-1 M Sorensen’s phosphate buffer. After rapid dehydration in ethanol the specimens were embedded in Epon 812. BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 125 Sections were cut with a Du Pont diamond knife on either a Porter-Blum or Reichert ultramicrotome, and stained with a saturated solution of alcoholic uranyl acetate and Reynold’s lead citrate. Some animals for whole-mount examination were placed on Formvar-coated, gold grids in a drop of fresh water. When they had extended their pseudopodia, a Petri dish containing a few drops of 2 per cent osmium textroxide was inverted over the grid for 1-2 minutes. Osmium vapour fixes the animals and pseudopodia instantly. The excess liquid was removed and the grid allowed to dry. Sections mounted on gold grids were used for the detec- tion of periodic acid-reactive carbohydrates, according to the silver methenamine technique of Rambourg (1967). The micrographs were obtained using an A.E.I. EMO6B electron microscope operating at 60 kV and recorded on IIford’s ‘ special lantern contrasty ’ plates. DISTRIBUTION E. rotunda is commonly found in damp and wet mosses, soil, forest litter and standing water. A full list of locality records and references was recorded by Bonnet & Thomas (1960) and Decloitre (1962 ; 1964). The following list illustrates the geographically widespread distribution : EUROPE : Spitzbergen, England, Scotland, Ireland, France, Belgium, Germany, Hungary and Italy. NortH AMERICA: Greenland, Canada and the United States of America. SouTtH AMERICA: Argentina, Colombia, Peru and Venezuela. AFRICA : Angola, Morocco and Madagascar. ASIA: Ceylon, Java, Sumatra and Tahiti. AUSTRALASIA : New Zealand and Tasmania. ANTARCTICA : South Georgia. MORPHOLOGY AND VARIATION The shell is small, varying in length from 34 to 56 wm and having a breadth of 14 to 24 wm (PI. r, fig. A); it is elliptical in shape through both the minor and major axes. The aperture is terminal and circular, having a diameter of 6 to Io ym and is surrounded by between eight to fourteen, evenly spaced, apertural-plates (Pl. 1, fig. D; Pl. 2, fig. B), although the most frequent number of plates is eight, nine or ten. Each apertural-plate is circular in shape, between 4:5 and 5:5 wm long, 3-6 and 4:5 wm wide, 0-59 and 0:82 ym thick (at its widest point) and carries a large median tooth, with either two or three smaller teeth situated on each side (Pl. 1, figs. C & D). The shell-plates range from 5-0 to 7:3 wm in length, 2:9 to 4-5 wm in width and 0-16 to 0:27 wm thick (PI. 1, fig. B), and are arranged regularly in alternate longitudinal rows (Pl. 1, fig. A). Shell-plates overlap to a certain extent at the lateral margins and in the posterior region (Pl. 2, fig. C). At the posterior extremity (Pl. 2, fig. D) the shell appears to end abruptly. 126 R. H. HEDLEY AND C. G. OGDEN An electron probe microanalysis of some shells showed that they had a high silica content and an almost negligible amount of calcium. No evidence of crystal- line material was obtained from electron diffraction examination carried out on both sectioned material, and on shells cleaned of most tissue by sodium hypo- chlorite, separated ultrasonically and washed with distilled water. It is concluded that the shell-plates are composed of a high percentage of amorphous silica. Variation in structure of the shell appears to be confined to the occasional for- mation of an extra large individual with more than the normal number of shell- plates. Such forms appear to result from conjugation of a pair of, or possibly more, mature specimens. The process is similar to that described by Blochmann (r887) and Penard (1902; 1940), in which it appears that the cytoplasm and plates for two daughter-cells are united to form a single shell. This possibly accounts for the specimens that have extra, unevenly spaced apertural-plates. Nevertheless, these animals appear to have a normal arrangement of shell-plates. REPRODUCTION Observations by optical microscopy indicate that the formation of a daughter- cell by means of simple division commences with cytoplasmic movement of the parent to produce a short, thick, pseudopodial trunk. As the apertural-plates of the daughter-cell emerge they are arranged around the outside of the pseudopodial trunk and lie with their teeth opposed to those of the parent. This is followed by rapid movement of the shell-plates from parent to daughter, where they become arranged in a regular pattern. The shell-plates are always added in sequence so that the plates at the terminal region are positioned last. When the plates are in their final position the daughter-cell rocks from side to side. At about this time the first signs of cytoplasmic transfer are seen ; globular vacuoles pass from parent to daughter-cell, and the nucleoles of the nucleus change in density and shape. When the shells are of equal size the granular material of the parent moves to the apertural region and the nucleus becomes indistinct. As the nucleus begins division it is no longer visible due to the rapid movement of granular material between parent and daughter. The plasmalemma and the contractile vacuoles are now visible in the daughter-cell. The rate of cytoplasmic movement then decreases and a nucleus appears in the posterior region of each cell. At the same time a granular band, slightly anterior to each nucleus, becomes evident. There is now a period of apparent inactivity before pseudopodia are extended and independent movement begins. The total time taken for such division is ap- proximately sixty minutes although the shell itself is produced more rapidly, in approximately fifteen minutes. In order to estimate the length of time required to double the population (doubling time), three identical cultures were established and maintained under similar con- ditions. Daily counts of the number of animals observed were recorded, and these have been used to produce growth curves (Text-fig. 1) from which the doubling time has been calculated to be between 1-4 to 1-9 days. BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 127 50,000 20,000 10,000 2,000 POPULATION (Numser OF SPECIMENS /CULT ure| 500 TIME IN DAYS Fic. 1. Growth curves for E. rotunda in culture. The symbols represent three replicate cultures through which curves have been fitted by eye. The calculated doubling times are |] -— 1-9 days; ©——o©x 1-4 days and —-— x 1-85 days. FOOD Although the cultures in the laboratory were normally grown on a mixed bac- terial flora, it is noted that successful cultures can be maintained using a single bacterium as the sole food source. Agar plates, streaked with Klebsiella aerogenes (strain NCIB 8017) and covered with the Euglypha culture medium, were incubated overnight at 25°C to establish the bacteria. The plates were then inoculated with washed Euglypha and incubated in the dark at 20°C. Sub-cultures were made at weekly intervals and under these conditions the Euglypha appeared to grow and reproduce well. 128 R. H. HEDLEY AND C. G. OGDEN CYTOLOGY Vegetative stage The cytoplasm normally occupies the whole of the shell cavity and is enclosed by a plasmalemma (Text-fig. 2, Pl. 3, fig. A). Pellicular microtubules lie beneath the plasmalemma (PI. 6, fig. B) running in an antero-posterior direction. They are not uniformly spaced, but occur often in associated groups of three or four. cao 20 ei \ MI TOCHONDRION | | RESERVE PLATE TONTRACTILE VACUOLE CONTRACTILE VACUOLE \ GRANULES CYST MEMBRANE PIGMENT ZONE’ PSEUDOPODIUM. Fic. 2. Diagram of cyst and vegetative stage of E. yvotunda showing the arrangement of the main organelles. Porter (1966) in a review of microtubules suggested, on the evidence then avail- able, that they might function as cytoskeletal structures. However, the presence of sub-pellicular microtubules would appear to be unnecessary in E. rotunda as the shell delimits the shape of the cell. Ovoid or spherical mitochondria having tubular cristae and a dense granular matrix (Pl. 3, fig. D) are distributed through- out the cytoplasm. The nucleus is usually spherical, between 4:2 and 5-9 wm in diameter, and enclosed within two tripartite membranes (PI. 3, fig. C). The nuclear matrix is finely granular with small concentrations of chromatin scattered through- out. A prominent and densely stained nucleolus, of variable shape, occupies the centre of the nucleus. Cisternae of granular endoplasmic reticulum arise from the outer membrane of the nuclear envelope and occupy a large compact region around the nucleus (Pl. 3, fig. D). Due to the concentration of ribosomes, this region is more heavily stained than the remainder of the cytoplasm. A region of large vacuoles, usually containing electron-dense material, lies immediately anterior to the endoplasmic reticulum. This electron-dense material appears to be leached out in some preparations and the vacuoles appear empty (PI. 3, figs. A & D). This region corresponds to the ‘ pigment zone’ of optical microscope descriptions of some earlier authors (Text-fig. 2). Two or three contractile vacuoles are present at the same level as the large vacuoles, but lie in the endoplasmic reticulum (PI. 3, BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 129 fig. A). They are usually surrounded by numerous vesicles which are continuous with the lumen of the vacuole. Ribosomes appear to be absent from that area surrounding the contractile vacuole which contains the canicular system or spongiome (Pl. 7, fig. A). The spongiome tubules have similar internal and external fibrillar coats to those found in Acanthamoeba castellanii (Bowers & Korn, 1968) and Crithidia fasciculata (Brooker, 1971). The contractile vacuoles discharge directly into the cavity of the shell. There is a single well-developed Golgi apparatus situated adjacent to the base of the concentrated endoplasmic reticulum surrounding the nucleus (Pl. 3, figs. A & D). In section it is U-shaped, the base lying in a channel or groove of the endoplasmic reticulum close to the nucleus, whilst the arms extend upwards into the lateral margins. Both:smooth and coated vesicles are seen to be associated with the margins of the Golgi saccules. The anterior third of the cytoplasm is occupied by numerous food vacuoles which have a single unit membrane and may contain both food organisms and waste material (Pl. 3, fig. D). Siliceous plates and cement There are two distinct types of siliceous plates : firstly, eight to fourteen apertural-plates which are thickened anteriorly at the base of the median tooth, tapered posteriorly and confined to the apertural region ; and secondly, approxi- mately one hundred and twenty shell-plates of uniform thickness which are arranged to form the rest of the shell. The plates overlap each other slightly and are held together with an organic cement or glue (Pl. 4, fig. E). They are composed mainly of amorphous silica which is electron-dense in sectioned material (Pl. 4, figs. A & B). Complete sections of individual plates may be obtained (PI. 4, figs. E & F) although more frequently they fracture in an apparently uniform manner (PI. 4, fig. D; Pl. 5, fig. H). Prior to cell-division reserve-plates are formed in the region of the granular vacuoles and endoplasmic reticulum, but close to the Golgi apparatus. Each teserve-plate is enclosed in a membrane-bound vacuole (PI. 4, fig. F) and is usually situated at the periphery of the cytoplasm (Pl. 4, fig. B). The area between the teserve-plates and the Golgi apparatus contains many vesicles of both the smooth and coated type (Pl. 4, fig. C). Vesicles of a similar nature have been reported to be associated with the Golgi apparatus and scale-containing vesicles in the Haptophyceae — Prymnesium and Chrysochromulina (Manton, 1966; 19672, b), and the coccolithophorids — Coccolithus pelagicus and Cricosphaera carterae (Manton & Leedale, 1969). Although we have no evidence that these vesicles discharge into the reserve-plate vacuoles, their close proximity suggests some correlation. As more plates are formed they become packed in ranks around the nucleus (Pl. 4, fig. A). The reserve apertural-plates are usually found in the most posterior position with their dorsal teeth facing inwards (Pl. 4, fig. D). Sections of specimens fixed with glutaraldehyde, but without the inclusion of heavy metals in this or subsequent procedures, were examined, in an attempt to locate sites of silica deposition. Apart from the siliceous plates, however, no other electron-dense regions were observed (PI. 4, figs. A & B). 130 R. H. HEDLEY AND C. G. OGDEN Previous reports on the formation of siliceous structures in other organisms have dealt mainly with either sponge spicules (Travis et al., 1967; Drum, 1968 ; Garrone, 1969 ; Fjerdingstad, 1970) or the diatom frustule (Desikachary & Dweltz, 1961 ; Drum & Pankratz, 1964; Lauritis e¢ al., 1968). Such structures are formed in similar membrane-bound vesicles to those described here for E. rotunda. The close association of smaller vesicles in E. rotunda which could coalesce with the silicon deposition vesicles are also observed by Lauritis e¢ al. (1968) in the diatom Nitzschia alba. Fine structure studies have shown that sponge spicules in Spongilla lacustris and Haliclona vosea are formed initially on an axial thread of protein (Fjerdingstad, 1970; Garrone, 1969) and an analysis by Desikachary & Dweltz (1961) shows that a small amount of organic material is also present in the diatom frustule. Fjerdingstad (1970) suggests that the silica in the sponge spicule of Spongilla lacustris is arranged in parallel ‘silica units’ of regular structure. Although we have observed similar ‘ units’ in E. rotunda, we consider that this is only the result of imperfect sectioning of hard material. Despite many observations on the formation of siliceous structures in plants and animals very little is known of the source of the silicon or of silicon metabolism. The biochemical and physiological problems involved in silicification in organisms are virtually unexplored. A minimum of organic cement or glue is present to hold the plates in position and maintain the specific shape. The cement is a fine fibrillar material (Pl. 4, fig. E), similar to that found in those membrane-bound vesicles that lie close to the periphery of the cytoplasm (PI. 5, fig. B), mainly in the anterior body-region (PI. 5, fig. H). Such vesicles have been found fused with the plasmalemma and dis- charging their contents into the shell-cavity (Pl. 5, fig. A). The contents of these vesicles stain strongly for polysaccharide (PI. 5, fig. C), using the silver methenamine technique (Rambourg, 1967), as do the organic cement connections between the plates. Variation in the composition of these vesicles is sometimes seen, in which the contents have a regular striated structure and occasionally there is a strongly electron-dense area in the polar regions (Pl. 5, figs. D & E). The striations or bands are 12 nm apart and are separated by a fibrillar region 25 nm wide. Pseudopodia The filose pseudopodia of E. rotunda are usually straight, do not appear to anastomose and radiate from the apertural region (Text-fig. 3 ; Pl. 3, fig. B). Over most of their length the pseudopodia are of uniform thickness, tapering at the extremities. They have a maximum length of 50 wm and a maximum thickness of 2m. The animal moves horizontally on glass, by the extension of unattached pseudopodia, which adhere at the tip and then shorten to exert traction. In this way the animal normally moves at approximately 2 ym per second. The pseudo- podia are extended quickly and apparently at random, but are also capable of rapid withdrawal. In addition to horizontal movement the animal often moves with the shell held in an upright position above the pseudopodia. In this con- dition it appears that the tips of the pseudopodia act as feet, being retracted only BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 131 slightly before moving and reattachment. Viewed from above the shell progresses with an uneven rolling motion, rather than a smooth gliding action. The fine structure of the pseudopodia appears to consist of ground-plasm, limited by a membrane, containing microfilaments. Microtubules are apparently absent, although sections through cytoplasmic extensions within the test do contain micro- tubules. This is consistent with the recent observations on Difflugiella sp. in which microtubules were numerous in pseudopodia within the mouth, but extended for only 1-2 ym into pseudopodia 20-30 pm long (Griffin, 1972). Fic. 3. Diagram of adult E. yotunda showing a typical arrangement of filose pseudopodia, drawn from life. Reports of similar microfilaments have been described from slime moulds (McManus & Roth, 1965) and several amoebae, both naked (Daniels & Breyer, 1967; Pollard et al., 1970; Pollard & Korn, 1971) and with shells (Wohlman & Allen, 1968; Moraczewski, 1970; Griffin, 1972). Microfilaments in the testate amoeba Difflugia have been demonstrated by Wohlman & Allen (1968) to be formed during pseudopod extension, and they suggest that these fibrils participate in forcible pseudopod retraction during cell locomotion. In the naked amoeba, Amoeba proteus, Pollard & Ito (1970) described thick and thin filaments, the latter being associated with an increase in viscosity, although both are required for move- ment. More recently they (Pollard & Korn, 1971) found that the thin filaments were indistinguishable from F-actins obtained from muscle (Huxley, 1963), 132 R. H. HEDLEY AND C. G. OGDEN Acanthamoeba (Pollard et al., 1970) and the slime mould Physarum (Nachmias et al., 1970) ; it was concluded that these structures are involved in motility. Using the drug cytochalasin B, Wessels ef al. (1971) demonstrated that microfilaments are the contractile machinery of non-muscle cells, and suggested that the evidence for this behaviour is overwhelming. Rosette groups Some two or three weeks after clonal cultures are established numerous animals appear linked in rosette-like groups (Pl. 2, figs. A & E). Individuals forming these rosettes are joined by cytoplasmic connections containing mitochondria, numerous vesicles and microfilaments (Pl. 2, fig. F, Pl. 6, fig. A). The majority of these vesicles are small and do not appear to have any characteristic contents (Pl. 6, fig. A). The microfilaments are usually straight although they may be seen to curve when passing through the aperture (Pl. 7, fig. H); they are found both inside (PI. 6, fig. C) and outside the shell (Pl. 7, fig. H). These structures appear to have a binding function, as they form adhesion plaques of concentrated micro- filaments at those points to which they become attached or anchor (PI. 5 fig. H, Pl. 6, fig. B). Using the terminology of Bennett (1969) this would be a ‘ desmo- some’. We make no attempt to classify it further, however, as the types defined by Bennett (1969) refer to individual cell junctions. In the case of E. rotunda the junctions occur between cells sharing cytoplasm and are both internal and external to the shell. The microfilaments in these intracellular connections may be capable of trans- mitting tensile forces through the adhesive plaques which will help to retain the rosette formation, and in this way may be similar to the forces present at desmo- some junctions. Similar attachment areas are reported by Buckley & Porter (1967) from cultured rat embryo cells associated with ‘stress fibers’ (= bundles of microfilaments). They suggested that these ‘stress fibers’ are concerned in stabilizing areas of cellular attachment as well as resisting forces that stretch the cell. Certain morphological changes have been detected in the cytoplasmic contents of those animals in rosette formation. In the mitochondria the intracristate space becomes enlarged and tubular material appears (PI. 5, figs. F & G). Similar struc- tures have been reported previously from encysting Acanthamoeba by Vickerman (1960, 1962), Bowers & Korn (1969) and more recently from encysting Bodo caudatus by Brooker & Ogden (1972). Bowers & Korn (1969) suggest that these features are probably related to starvation rather than encystment. The exocytoic vacu- oles are seen at this time to contain reserve-plates in addition to the normal un- digested materials (PI. 6, fig. D). One unusual feature of these rosette-like groups is that although the evidence suggests either starvation or encystment, bacteria are still seen to be engulfed (Pl. 6, fig. A). Testate amoeba in which individuals clump together and between which the cytoplasm is connected have been reported previously ; Arcella vulgaris (Reynolds, 1939), Difflugia elegans (Erth, 1965) and Difflugiella sp. (Griffin, 1972). Clumping BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 133 without cytoplasmic connections between individuals has also been reported in cultures of other protozoa. Band & Mohrlok (1969) found that the soil amoeba, Mayorella palestinensis, formed clumps alone and with another amoeba, Hartmannella rhysodes, although the latter species did not clump in pure culture. Opinions differ over the factors which induce the formation of such groups in amoebae. Reynolds (1939) considers that this behaviour is related to nutritional factors, and observes that when such groups of Arcella vulgaris are transferred to fresh culture media or spring water they separate. Band & Mohrlok (1969) sug- gest that in Mayorella palestinensis it is temperature dependent and not associated with starvation and encystation. They found no extra- or intracellular changes on examination at the ultrastructural level, but showed that these cells became interlocked. Our observations would indicate that the cause is either starvation or encystation. Cyst Encysted animals occur in cultures usually four or five weeks after inoculation. The cyst is contained within a cyst-membrane which lies close to the shell wall, except in the anterior region where it forms a seal across the aperture of the shell. In mature cysts the membrane retracts so that it lies close to the mid-point region of the shell (Text-fig. 2, Pl. 7, fig. C). The cyst-membrane consists of amorphous fibrillar material, similar to the organic cement material found joining the body-plates. In mature cysts it becomes striated (Pl. 7, fig. D) with striations being visible mainly at positions where the plates do not join or overlap, for example, at the posterior end of the shell (Pl. 7, fig. G). In addition, striations occur in the thickened region across the middle of the shell (Pl. 7, fig. E). An unusual feature of this portion of the cyst-membrane is that during the early stages of encystment, a layer of apparently unconnected membrane is enclosed in the fibrillar material (Pl. 7, fig. F) and is retained even in mature cysts. The striated or banded appearance of the cyst-membrane (Pl. 7, fig. G) pre- sumably reflects a fibrous protein composition. The fibrils are cross-striated, with a regular pattern being repeated every 40 nm (= axial period) and within each of these bands two alternating light and dark intraperiod bands of equal thickness are recognized. The limited detail from the available micrographs precludes further comment but it is noteworthy that cross-striated fibrils of the collagen type — not previously known to occur in protozoans — were reported recently in the rhizopod Haliphysema (Hedley & Wakefield, 1967). During encystment the cytoplasmic volume is considerably reduced (PI. 7, fig. C). There is a reduction in the number of mitochondria and they appear to be found only in the region of the endoplasmic reticulum. The matrix of the endo- plasmic reticulum is reduced in size and appears dense compared with the remainder of the cytoplasmic matrix. The electron-dense vacuoles of the ‘ pigment zone’ are retained (PI. 7, fig. C) but no reserve-plates are visible. As encystment pro- ceeds the plasmalemma becomes progressively invaginated and the pellicular 134 R. H. HEDLEY AND C. G. OGDEN microtubules are no longer found close to it (Pl. 7, fig. B). Thin strands of cyto- plasm connect the young cyst to the cyst-membrane (PI. 7, fig. G), while in older cysts, these appear to be confined only to that portion of the cyst nearest to the unattached cyst-membrane. No apparent changes in the Golgi apparatus or the nucleus have been observed. ACKNOWLEDGEMENTS We should like to acknowledge the considerable technical help provided by Miss J. I. Battershall. Dr C. R. Curds — formerly of the Water Pollution Research Laboratory, Stevenage, and now at the British Museum (Natural History) kindly discussed the typescript and established E. votunda in monoxenic culture. Dr K. M. Towe (Smithsonian Institution) made observations on the siliceous plates, from electron diffraction results. NOTE Since this paper was submitted for publication a brief account of the production of a new daughter-shell by EF. rotwnda is reported by Netzel (1972). These observations were based on the present authors’ original material. REFERENCES Banp, R. N. & Mouriox, S. H. 1969. An analysis of clumping in the soil amoeba Mayorella palestinensis. J. Protozool. 16 : 35-44. Bennett, H. S. 1969. The cell surface: components and configurations. In Handbook of Moleculay Cytology, Vol. 15 : 1261-1293. Amsterdam & London. BLOCHMANN, F. 1887. Zur Kenntnis der Fortpflanzung von Euglypha alveolata Duj. Morph. Jb. 13 : 173-183. Bonnet, L. & THomas, R. 1960. Faune terrestre et d’eau douce des Pyrénées-Orientales. Thécamoebiens du sol. Hermann, Paris. 103 pp. Bowers, B. & Korn, E. D. 1968. The fine structure of Acanthamoeba castellanit. I. The trophozoite. J. Cell Biol. 39 : 95-111. Bowers, B. & Korn, E.D. 1969. The fine structure of Acanthamoeba castellanii (Neff Strain). II. Encystment. J. Cell Biol. 41 : 786-805. Brooker, B. E. 1971. The fine structure of Crithidia fasciculata with special reference to the organelles involved in the ingestion and digestion of protein. Z. Zellforsch. 116: 532-563. Brooker, B. E. & OGpEN, C. G. 1972. Encystment of Bodo caudatus. Protoplasma 74: 397-409. Bucktey, I. K. & Porter, K. R. 1967. Cytoplasmic fibrils in living cultured cells. Pyoto- plasma 64 : 349-380. Butscuir, O. 1880. In Bronn’s Klassen und Ovdnungen des Tierveichs. 1. Protozoa: 1-224. Leipzig & Heidelberg. CamBaR, R., LE Bianc, M., Mercier, M. & THomas, R. 1964. Etude en microscopie élec- tronique des microstructures superficielles de Protistes fossiles ou vivants. C. v. Acad. Sct. (Paris) 258 : 3554-3555- Carter, H. J. 1865. On the fresh- and salt-water Rhizopoda of England and India. Ann. Mag. nat. Hist. (3) 15 : 277-293. BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 135 Casu, J., Wares, G. H. & Hopxinson, J. 1915. The British Freshwater Rhizopoda and Heliozoa. Vol. 3. Rhizopoda. The Ray Society London. 156 pp., 24 pls. DaniEts, E. W. & Bryer, E. P. 1967. Ultrastructure of the giant amoeba Pelomyxa palustris. J. Protozool. 14: 167-179. DectoitrE, L. 1950. Etudes sur les rhizopodes. Feuille Nat. 5: 41-46. DeEcLortrRE, L. 1962. Le genre Euglypha Dujardin. Arch. Protistenk. 106 : 51-100. DEcLoITRE, L. 1964. Thécamoebiens de la XIIéme Expédition Antarctique Frangaise. Terri- toive des Terres Australes et Antarctique Frangaises. Expéditions Polaires Frangaises (Missions Paul-Emile Victor). Publ. No. 259 : 1-47. DesrKAcHARY, T. V. & Dwettz, N. E. 1961. The chemical composition of the diatom frustule. Proc. Indian Acad. Sci. 53 B : 157-165. Drum, R. W. 1968. Electron microscopy of siliceous spicules from the freshwater sponge Heteromyenia. J. Ultrastruct. Res. 22 : 12-21. Drum, R. W. & Pankratz, H. S. 1964. Post mitotic fine structure of Gomphonema par- vulum. J. Ultrastruct. Res. 10 : 217-223. Ertu, M. 1965. Zur Kenntnis der Testaceenfauna der slowakischen Reisfelder. Hydro- biologia 26 : 13-20. FJERDINGSTAD, E. J. 1970. Ultrastructure of the spicules of Spongilla lacustris (L.). Symp. zool. Soc. Lond. No. 25 : 125-133. GarRRONE, R. 1969. Collagéne, spongine et squelette minéral chez l’éponge Haliclona rosea (o.s.) (Démosponge, Haploscléride). J. Microscopie 8 : 581-598. Granpor!, R. & GRaNDoRI, L. 1934. Prime ricerche sui Protozoi dei terreni e delle acque dell’alta Montagna Alpina. Boll. Lab. Zool. agr. Bachic R. Ist. sup. agy. Milano 5 : 1-340. GrirFIn, J. L. 1972. Movement, fine structure, and fusion of pseudopods of an enclosed amoeba, Difflugiellasp. J. Cell Sci. 10 : 563-583. GruBER, A. 1881. Der Theilungsvorgang bei Euglypha alveolata. Z. wiss. Zool. 35 : 431-439. Hatt, R. P. & LoeEFER, J. B. 1930. Studies on Euglypha. I. Cytoplasmic inclusions of Euglypha alveolata. Arch. Protistenk. 72 : 365-370. Harris, R. H. 1965. Simple technique for embedding and supporting delicate biological specimens. Nature, Lond. 208 : 199. Harris, R. H., Martin, B. S. & OGpEN, C. G. 1972. Notes on the preparation of natural history specimens for scanning electron microscopy. Bull. By. Mus. nat. Hist. (Zool.) 24 : 223-228. Heptey, R. H. & WAKEFIELD, J. St. J. 1967. A collagen-like sheath in the arenaceous foraminifer Haliphysema (Protozoa). Jl R. microsc. Soc. 87 : 475-481. Huxtey, H.E. 1963. Electron microscope studies on the structure of natural and synthetic protein filaments from striated muscle. J. molec. Biol. 7 : 281-308. Lauritis, J. A., Coomss, J. & VoLcaNI, B. E. 1968. Studies on the biochemistry and fine structure of silica shell formation in diatoms. IV. Fine structure of the apochlorotic diatom Nitzschia alba Lewin and Lewin. Arch. Mikrobiol. 62 : 1-16. Logsricu, A. R. & Tappan, H. 1961. Suprageneric classification of the Rhizopodea. J. Paleontology 35 : 245-330. Manton, I. 1966. Observations on scale production in Prymnesium parvum. J. Cell Sci. 1: 375-379. Manton, I. 1967a. Further observations on the fine structure of Chrysochromulina chiton with special reference to the haptonema, ‘ peculiar ’ Golgi structure and scale production. J. Cell Sci. 2 : 265-272. Manton, I. 1967b. Further observations on scale formation in Chrysochromulina chiton. J. Cell Sci. 2: 411-418. Manton, I. & LEEDALE, G. F. 1969. Observations on the microanatomy of Coccolithus pelagicus and Cricosphaera carterae, with special reference to the origin and nature of coccoliths and scales. J. mar. biol. Ass. U.K. 49: 1-16. McManus, M. A. & Rotu, L. E. 1965. Fibrillar differentiation in myxomycete plasmodia. J. Cell Biol. 25 : 305-318. 136 R. H. HEDLEY AND C. G. OGDEN Mercier, M., LE Branc, M., THomas, R. & CamBar, R. 1964. Observations, en microscopie électronique, sur la constitution de la theque de quelques Euglyphidae (Rhizopodes tes- tacés). C.v. Acad. Sci. (Paris) 258 : 5967-5968. MoraczEwski, J. 1970. Quelques observations sur l’ultrastructure du thecamoebien : Arcella votundata Play. Protistologica 6 : 353-359. Nacumias, V. T., Kesster, D. & Huxtey, H. E. 1970. Electron microscope observations on actomyosin and actin preparations from Physarum polycephalum, and on their inter- action with heavy meromyosin subfragment I from muscle myosin. J. molec. Biol. 50: 83-90. Netzer, H. 1971. Euglypha rotunda (Testacea) Bewegung und Fortpflanzung. Film E1642/1971. des Inst. Wiss. Film, Gottingen. NetzeEL, H. 1972. Morphogenese de Gahauses yon Euglypha rotunda (Rhizopoda, Testacea). Z. Zellforsch, mikrosk. Anat. 135 : 63-69. PrenarpD, E. 1890. Etudes sur les Rhizopodes d’eau douce. Mem. Soc. Phys. Hist. nat. Genéve 31 : 1-230. PENARD, E. 1902. Faune Rhizopodique du Bassin du Léman. Geneva. 700 pp. PENARD, E. 1940. Protozoaires et psychologie. Archs Sci. phys. nat. 22 : 203-226. PotrarD, T. D. & Ito, S. 1970. Cytoplasmic filaments of Amoeba proteus. I. The role of filaments in consistency changes and movement. /. Cell Biol. 46 : 267-289. PoriarD, T. D. & Korn, E.D. 1971. Filaments of Amoeba proteus. II. Binding of heavy meromyosin by thin filaments in motile cytoplasmic extracts. J. Cell Biol. 48 : 216-219. POoLiarD, T. D., SHELTON, E., WETHING, R. R. & Korn, E.D. 1970. Ultrastructural charac- terization of F-actin isolated from Acanthamoeba castellanii and identification of cyto- plasmic filaments as F-actin by reaction with rabbit heavy meromyosin. J. molec. Biol. 50 : 91-97. Pororr, M. 1912. Uber die geschlechtliche Fortpflanzung von Euglypha alveolata Duj. Arch. Protistenk. 25 : 8-26. Porter, K. R. 1966. Cytoplasmic microtubules and their functions. In Principles of Biomolecular Organization. Eds. Wolstenholme, G. E. W. and O’Connor, M., pp. 308— 345. London. RampourG, A. 1967. An improved silver methenamine technique for the detection of periodic acid-reactive complex carbohydrates with the electron microscope. J. Histochem. Cytochem. 15 : 409-412. ReEynNotps, B. D. 1939. Observations on the clumping of Arcellae into plasmodial clumps. Proc. Va. Acad. Sci. 1938-39 : 29. ScHEWIAKOFF, W. 1888. Uber die karyokinetische Kerntheilung der Euglypha alveolata. Morph. Jb. 13 : 193-258. Stout, J. D. & Heat, O. W. 1967. Protozoa. In Soil Biology, Chap. 6, pp. 149-195. Academic Press, London & New York. Tuomas, R. 1962. Kystes et enkystement chez les thécamoebiens. Bull. Soc. zool. Fr. 87 : 276-280. Travis, D. F., Francois, C. J., Bonar, L. C. & GiimcHeErR, M. J. 1967. Comparative studies of the organic matrices of invertebrate mineralized tissues. J. Ultrastruct. Res. 18 : 519-550. VICKERMAN, K. 1960. Structural changes in mitochondria of Acanthamoeba at encystation. Nature, Lond. 188 : 248-249. VICKERMAN, K. 1962. Patterns of cellular organisation in Limax amoebae. An electron microscope study. Exp. Cell Res. 26: 497-519. Waites, G. H. 1912. Freshwater Rhizopoda from the Hebrides, Orkney and Shetland Islands and Western Scotland; with description of a new species. Scott. Nat. 1912: 59-65. Waites, G. H. 1915. Notes on the structure of tests of freshwater Rhizopoda. J/ R. microsc. Soc. 35 : 105-116. Waitses, G. H. & PENaRD, E. 1911. Clare Island Survey. Part 65. Khizopoda. Proc. R. Iv, Acad. 31: 64 pp., 6 pls. BIOLOGY OF EUGLYPHA ROTUNDA (PROTOZOA) 137 WESSELLS, N. K., Spooner, B. S., AsH, J. F., BRADLEy, M. O., Lupuena, M. A., TayLor, E. L., Wrenn, J. T. & Yamapa, K. M. 1971. Microfilaments in cellular and develop- mental processes. Science 171 : 135-143. Woutman, A. & ALLEN, R. D. 1968. Structural organization associated with pseudopod extension and contraction during cell locomotion in Difflugia. J. Cell Sci. 3: 105-114. Ronatp HENDERSON HEDLEY, D.Sc. British Musrum (NATURAL History) CROMWELL Roap Lonpon SW7 5BD CoLtIn GERALD OGDEN British Museum (Naturat History) CROMWELL Roap Lonpon SW7 5BD Sour PLATE 1 Lateral view of E. votunda showing the arrangement of shell-plates. x 2,300 Individual shell-plates. x 7,500 Lateral view of apertural-plates with two lateral teeth. x 7,500 Aperture of E. votunda with fourteen apertural-plates ; note the thickened region associ- ated with the median tooth and some plates (arrowed) with three lateral teeth. x 5,900 Bull. By. Mus. nat. Hist. (Zool.) 25, 4 PLATE 1 oS op PLATE 2 Light micrograph of animals stained with eosin and Mayer’s haemalum. x 250 Specimen with ten, evenly spaced apertural-plates. x 6,650 View of posterior region showing the arrangement of plates and lateral flattening of shell. x 2,850 Lateral view of posterior region showing the overlap of plates and blunt termination. x 4,280 Group of six animals in ‘ rosette’ formation. x 675 Cytoplasmic connections of ‘ rosette’ formation. x 5,000 Bull. By. Mus. nat. Hist. (Zool.) 25, 4 PLATE 2 ow PLATE 3 Longitudinal section showing the position of contractile vacuole (cv), Golgi apparatus (G) and endoplasmic reticulum (er); note the structureless appearance of the ‘ pigment zone’. xX 5,760 Direct preparation of whole animal showing the arrangement of pseudopodia. x 1,400 Transverse section of normal nucleus, with prominent electron-dense central nucleolus. x 8,640 Longitudinal section of an unusually shaped nucleus (n) lying in the dense endoplasmic reticulum (er) ; Golgi apparatus (G). x 4,320 Bull. By. Mus. nat. Hist. (Zool.) 25, 4 PE AME 3) ley 1S) PLATE 4 Glutaraldehyde fixed, unstained longitudinal section to illustrate the arrangements of siliceous plates in nuclear region ; note that the ‘ pigment zone ’ does not contain electron- dense material. x 4,200 Glutaraldehyde fixed, unstained transverse section showing the peripheral arrangement of reserve plates. x 4,200 Section through part of Golgi apparatus, bottom left, showing proximity of coated vesicles (cves) and reserve shell-plate. x 44,400 Section showing a reserve apertural-plate in close proximity to the Golgi apparatus (G) ; note fracture lines apparently caused by sectioning procedures. x 15,570 Siliceous plate showing the associated amorphous (?) organic cement. x 28,350 Reserve shell-plate in membrane-bound vesicle. x 56,700 Bull. Br. Mus. nat. Hist. (Zool.) 25, 4 PLATE 4 PLATE 5 . Vesicle containing organic cement discharging into shell cavity. x 37,800 . Vesicles containing cement material lying adjacent to plasmalemma. x 56,700 . Section stained with silver methenamine, showing positive reaction by organic cement vesicles for polysaccharide. x 7,780 . Cement body showing regularly striated structure. x 60,400 . Banded organic cement body with dense polar staining. x 56,700 . Examples of mitochondrial degeneration from ‘ rosette ' specimens. x 28,350 . Section through apertural region of a specimen from a ‘rosette’ group showing position of cement containing vesicles; note the concentration of microfilaments (arrowed). x 5,760 % PLATE Mus. nat. Hist. (Zool.) 25, 4 Bull. Br. PLATE 6 Sections from specimens joined in ‘ rosette ’ formation: A. Section through group of animals showing the extent to which cytoplasm protrudes beyond the shell; note the bacteria (b) being engulfed. x 2,800, B. Transverse section showing the position of pellicular micro- tubules (pmt) and cement-containing bodies (em); the arrows indicate concentrations of microfilaments. ™ 7,780. C. Section through apertural region showing microfilaments oriented in straight lines. x 30,090. D. Section of exocytoic vesicle containing siliceous shell. x 30,090. Bull. By. Mus. nat. Hist. (Zool.) 25, 4 PLATE 6 aS ISM METS cS ZN NAT. Hisy. ~ 5 JUL 198 PRESENTED RN ey, 5 ah Gy Lise PLATE 7 Section through contractile vacuole (cv) and spongiome. x 18,900 Section of cyst plasmalemma showing invaginations and pellicular microtubules (pmt). x 29,700 Longitudinal section showing position of cyst membrane (cm) and cytoplasm. x 2,800 Striated cyst membrane on inside of shell-plate. x 55,200 Section through mature cyst membrane showing striated structure. x 30,090 Section through early cyst membrane showing unconnected membrane (arrowed). x 56,700 Junction of shell-plates at posterior of animal with unusually large amount of organic cement behind the plates and cyst membrane; note the striated cyst membrane. x 37,800 Microfilaments curving around apertural-plate ; note the internal microtubules (mt) and concentration of microfilaments at the external surface of the plate (arrowed). x 37,800 Bull. Br. Mus. nat. Hist. PLATE 7 ws A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. ALIson. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75. WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier and = Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. TayLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates 77 Text-figures. 1969. £4.50. Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. AN 3 WAT. HIS}. 8, JUL 1973 | Wa peaches Aes 2 Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 sNU r A REVISION OF THE __ HAPLOCHROMIS AND RELATED _ SPECIES (PISCES : CICHLIDAE) _ FROM LAKE GEORGE, UGANDA _ BULLETIN OF H MUSEUM (NATURAL HISTORY) Vol. 25 No. 5 LONDON: 1973 A REVISION OF THE HAPLOCHROMIS AND REALE DS SPECIES. (PISCES, »CICHLIDAE) FROM LAKE GEORGE, UGANDA ' ; L199 PRESENTED 4 ff ©, 8 BY PETER HUMPHRY GREENWOOD Pp. 139-242 ; 40 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 5 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25 No. 5 of the Zoological series. The abbreviated-titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 27 June, 1973 Price £4.55 A REVISION OF THE HAPLOCHROMIS AND RELATED SPECIES (PISCES : CICHLIDAE) FROM LAKE GEORGE, UGANDA By P. H. GREENWOOD CONTENTS Page INTRODUCTION . : . 5 é : é : : é rag MATERIALS AND METHODS . 4 6 0 : . c 9 2 144 Haplochromis elegans Trewavas. ° ¢ : : : 6 é 145 Haplochromis aeneocoloy sp. nov. 150 Haplochromis nigripinnis Regan : : 155 Haplochromis oregosoma sp. nov. : ees 3 6 4 159 BE scivomis macropsoides sp. nov. oe) dv 1973 | . : ° 162 Haplochromis limax Trewavas . ~) é : : 167 Haplochromis mylodon sp. nov. . 172 Haplochromis angustifrons Blgr. : : c c ; 177 Haplochromis schubotzi Blgr. . 5 é ¢ : 6 . . 183 Haplochromis schubotziellus sp.nov. . : é é 6 : é 189 Haplochvomis taurinus Trewavas ‘ c ‘ : . c : 192 Haplochromis labiatus Trewavas : - : c . 7 ° 196 Haplochromis pappenheimi (Blgr.) . : : : : c : 199 Haplochromis squamipinnis Regan . : 6 é . : 204 Haplochromis petronius sp. nov. : F : c : é - 209 Haplochromis eduardianus (Blgr.) : ° 4 ; : 6 Aug Non-ENDEMIC Haplochromis AND Haplochromis-group SPECIES c 220 Haplochromis nubilus (Blgr.) .- : : . : : 5 | BA Astatoreochromis Pellegrin : : - : ¢ : ‘ eZ 2A Astatoreochromis alluaudi Pellegrin . 5 : : : : 224 Hemihaplochromis Wickler é a 7 é “ 6 ; 22 Hemihaplochromis multicolor (Schoeller) . : : c é = 220 Discussion ¢ c : a Pea Biology of the Lake George Haplochromis species flock ° rez 27, The relationships and history of the Lake George eR species E230 ACKNOWLEDGEMENTS . : : 0 : oO 237) APPENDICES: I : . 6 ¢ 6 6 : . ; 235 II . é F é 3 . 5 0 6 : 238 BIBLIOGRAPHY . : 5 5 5 239 GUIDE TO THE IDENTIFICATION OF THE Haplochromis SPECIES . < 2A INTRODUCTION LAKE GEORGE, smallest of the African ‘Great Lakes’ (text-fig. 1), occupies a virtually square basin of about 270 km? area in the western Rift valley (0°55’ N to 0°05’S, and 30°02’ E to 30°18’ E). Water depth over much of the lake rarely exceeds 2:5 m, although there are some circumscribed areas with depths of up to 4:0 m. Most of the lake is bordered by flat savannah-bush, but to the north and east there are areas of papyrus swamp extending, albeit as narrow fingers, for some 15 km from the lake edge. The principal affluent rivers enter Lake George through 142 P. H. GREENWOOD these papyrus swamps. The Rivers Sibwe, Nsonge and Mubuka arise in the Ruwenzori mountains ; the Mpanga, however, is a westward flowing tributary of the Katonga, a river which also flows eastward to enter Lake Victoria. The shared headwater of this river is a swamp divide (Doornkamp & Temple 1966), apparently impenetrable to all but air-breathing fishes. Much of the shore line is simple, but there are a few deeply indented bays and one steep-sided bay formed from a volcanic crater. Considerable areas of the lake bottom are covered by thick (ca 3m) deposits of flocculent organic ooze, overlying a firm clay substrate. In some places, both in- and offshore, a sandy substrate is exposed or is but thinly overlaid by mud. Two large islands (Kankurunga and Akika) lie close to the western lake shore ; a third (Irangara Island), on the north-western shore, almost occludes the entrance to the lake’s largest and most sheltered bay, Hamukunga Bay. The island shore- lines are varied and include slightly indented muddy bays, short stretches of sandy beach and extensive but narrow fringes of papyrus. For a more detailed description of the lake and a brief outline of its limnological features, reference should be made to Dunn, Burgis, Ganf, McGowan & Viner (1969). In addition to its small size and extreme shallowness, Lake George also differs from the other ‘ Great Lakes’ in being directly linked with another water body, Lake Edward. Connection between Lakes Edward and George is effected through the Kazinga Channel, a 36 km long, river-like passage uninterrupted by swamps or rapids. There is a definite net outflow of water from Lake George into Lake Edward but the current is slight, and on occasion, undergoes wind-induced reversal of flow, at least in the upper layers of water. To what extent the Kazinga Channel allows an actual exchange of fishes between the lakes (or of gene flow between populations of fishes in the lakes) has yet to be determined. Certainly many species of cichlid and non-cichlid fishes are present in both lakes, and the apparent endemism of some Lake George Haplochromis species may well be just a reflection of inadequate collecting in Lake Edward. Nevertheless, some habitats in Lake Edward are not represented in Lake George, and it is almost certain that a few Edward species are absent from Lake George. These absentees include not only species from deep-water habitats but also several from inshore habitats as well. Their absence from Lake George is hardly attri- butable to inadequate sampling because that lake has been intensively collected during the past six years. Fairly comprehensive fish collections have been made recently along the whole length of the Kazinga Channel (see Appendix II). These collections indicate that the Haplochromis species of the channel are exclusively those common to both Lake Edward and Lake George. Surprisingly, even as close to Lake Edward as the Mweya Peninsula none of the inshore-living and apparently endemic Lake Edward species was found in the channel. Clearly, detailed ecological studies will have to be made (particularly at the Lake Edward end of the channel) before this situation is understood. For the moment, however, there seem to be good a priori grounds for believing that, for many species, there is continuity of populations between the lakes. 143 LAKE GEORGE HAPLOCHROMIS SPECIES *(6961) “7y ja uuNG 40z;e poyIpoy ‘aye] oY} JO uorpisod [eoryders008 smoys yosuy “sIOAII yuongye rofeut Aeg oewyny ® s} enriy ©) *s} eBununyuey 2) *s| eieBues] © N dwems snidded [i] ‘ nanBunyey — \ Aeg w 1S eyeysey \ ZA \ a A Za ~~ Zee A \ \ \ \ \ )) plurey e esinanlle / y f e i y / fo s}I pue as10ay oye] —<—= a CRNEN| ‘I OL quvoMd 3 34Vv1 144 P. H. GREENWOOD Little precise information is available on the age of Lake George in its present form, or of its past history. Dating of lake core samples suggests that the deposits are in the order of 3500 years old (unpublished data from International Biological Programme research). The general geological history of the area (as outlined by Doornkamp & Temple 1966) indicates, however, that a lake occupied the present George-Edward basin from at least the later Middle Pleistocene until the later Pleistocene. It seems possible, therefore, that an earlier Lake George ‘ disappeared ’ (through causes unknown) and was later recreated, presumably from the then exist- ing Lake Edward. Certainly the fishes indicate derivation from a common source at some recent time. Ideally, any revision of the Lake George Haplochromis species should be combined with a revision of the Lake Edward species. For a variety of reasons the ideal could not be met ; in particular it has not been possible to get additional material from all areas of Lake Edward. A rather pragmatic reason for undertaking a revi- sion of the Lake George species alone and at this time was the need to provide information for the International Biological Programme team working on the lake (see Dunn ef al. 1969). Since 1967 a group of British and Ugandan biologists has been studying various levels of productivity in Lake George. Three team members have, at different times, been concerned with the ecology and distribution of the Haplochromis species. Since their results will be published in detail, I have con- centrated, in this paper, on the taxonomic problems involved. Only brief outlines of the species’ biology are given, and these may be modified in the light of later research by the I.B.P. team. Because specimens and data for Lake George Haplochromis species now out- number those available for Lake Edward, it seemed inadvisable to attempt any interlake comparisons between samples from species occurring in both lakes. Asa general impression, however, I suspect that interpopulation differences will eventually be detected. MATERIALS AND METHODS Haplochromis species from Lake George were poorly represented in the British Museum (Natural History) collections. Consequently, most of the material on which this paper is based is that obtained during the I.B.P. investigation of the lake. In the only previous revision (Trewavas 1933), Lake George specimens were treated together with fishes from Lake Edward. Because I have not studied Lake Edward fishes in any detail, the synonymies given below include, with few exceptions, only those Lake George specimens actually mentioned in Trewavas’ (op. cit.) paper. Where necessary, however, I have included some Lake Edward specimens. For example, this has been essential when selecting certain lectotypes, or where a misidentification is corrected and, if not included in a synonymy, could lead to zoogeographical misunderstanding. Most of the data on live coloration, distributions and breeding habits were collected personally during several visits to the lake. A lot of this information has been supplemented and refined by the observations of Dr Ian Dunn and Mr James LAKE GEORGE HAPLOCHROMIS SPECIES 145 Gwahaba, the fish biologists of the I.B.P. team. Both these workers have given unstintingly of their time and information, and I am extremely indebted to them. Measurements used in describing the species are those I have employed in other papers on Haplochromis species, viz. : Standard length : measured directly! from the snout tip (including the premaxilla) to the posterior margin of the hypural bones (located by bending the caudal fin at right angles to the body’s long axis). Head length: measured directly! across the head from snout tip to the most posterior point on the opercular bone. 4 Preorbital depth: is the greatest depth of the first infraorbital bone (= lachrymal bone). Interorbital width: is the least distance between the bony (frontal) margins of the orbit. Snout length: measured directly! from the snout tip (ie. the premaxillary symphysis) to the anterior orbital margin. Eye diameter: is the greatest diameter of the bony orbit in the horizontal plane. Cheek depth: is the greatest depth of the muscular part of the cheek (even when this extends below the scale rows) and is measured vertically. Lower jaw length: is measured directly! from the dentary symphysis to the posterior margin of the articular bone (located by opening the lower jaw and finding its point of articulation). Upper jaw length: is measured directly! from the premaxillary symphysis to the posterior margin of the maxilla. Caudal peduncle length: is taken from the posterior margin of the hypurals to a vertical projected from the insertion of the last anal ray. Peduncle depth is the least depth. A character I have used for the first time concerns the so-called pseudorakers on the first gill arch. These structures lie on the anterior (i.e. upper) face of the arch, between the inner and outer rows of true gill rakers. Pseudorakers are localized thickenings of the tissue covering the arch. In gross appearance they resemble true gill rakers, but unlike those structures they lack a bony central core. Vertebral counts do not include the fused first preural and ural vertebrae (which support the parhypural and hypurals). Haplochromis elegans Trewavas, 1933 (Text-figs. 2 & 3) Haplochromis nubilus (part) : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 329 (specimens BMNH Teg. NOS. 1933.2.23 : 288—295 from Lake George). H. elegans (part) Trewavas, op. cit.: 332 (3 paralectotypes, reg. nos. 1933.2.23 : 387-389, supposedly from Lake George, and 4 other specimens, 1933.2.23 : 390-393, also from that lake). NOTE ON THE SyNoNYMY. According to Trewavas (1933), three H. elegans syntypes (all females) are from Lake George, the other syntypical material being from Lake 1 In a direct measurement, one tip of the dividers or calipers is placed at one of the points specified and the other tip is placed on the second point; the distance measured may thus run across the long axis of the fish (as, for example, in snout and head lengths). 146 P. H. GREENWOOD Edward. I have examined the three syntypes (reg. nos. 1933.2.23 : 387-389) and agree with Trewavas’ identification. However, the bottle label, and the Museum register, give the locality for these fishes as Lake Edward and not Lake George. Trewavas (op. cit.) also refers seven specimens from the hypodigm (reg. nos. 1933-2-23 : 390—395) to this species, giving their localities as Worthington’s (1932) stations 522 (Lake Edward) and 613 (Lake George). Six fishes (reg. nos. 1933.2.23 : 390-395) are in a bottle now labelled ‘ Lake George ’ but without any station number quoted. I take these to be the fishes from station 613. Of these specimens, three are referable to H. elegans. The lectotype, an adult male 65:5 mm standard length (BMNH 1933.2.23 : 381), is from Lake Edward. Fic. 2. Haplochromis elegans. Lake George specimen ; a male. DESCRIPTION. Based on 34 specimens, 58-0—72:5mm standard length (but not including the lectotype since it is from Lake Edward). Depth of body 35-7—40-8 (mean, M = 37-6) per cent of standard length, length of head 32-2—35-4 (M = 33:7) per cent. Dorsal profile of head gently decurved or straight, sloping at about 35° to the horizontal ; dorsal margin of eye not entering the line of the profile, but clearly below it. Preorbital depth 11-8—16-5 (M = 14:5) per cent of head, showing very slight positive allometry. Least interorbital width 21-9-27-3 (M = 24:2) per cent of head, length of snout 25-6—32°5 (M = 28-4) per cent, 0-8—-0-9 of its breadth. Eye diameter 28-0-37-0 (M = 33:5) per cent of head (not showing clear-cut allometry in the size range examined), depth of cheek 18-2—24-4 (M = 20°8) per cent. Caudal peduncle 13-8—18-7 (M = 16-2) per cent of standard length, 1-2—-1-5 (modal range 1-2—1-3) times as long as deep. Mouth horizontal or very slightly oblique ; lips somewhat thickened. Length of upper jaw 28-6—34:0 (M = 30-3) per cent of head, length of lower jaw 35:0—40-2 LAKE GEORGE HAPLOCHROMIS SPECIES 147 (M = 37-9) per cent, 1-3-1-8 (mode 1-4) times as long as broad. Posterior tip of maxilla reaching the vertical through the anterior orbital margin, but not quite reaching this level in a few specimens. Gill rakers variable in form but usually rather stout, the lower I or 2 greatly reduced ; 8 or 9 rakers in the outer row on the lower part of the first gill arch. No pseudorakers are developed between the inner and outer rows of gill rakers on this arch. Scales. Ctenoid ; lateral line with 30 (f-2), 31 (£15), 32 (£15), 33 (f-1) or 34 (£1) scales; cheek with 2 or 3 (mode) rows. Five to 6} (mode 5}) scales between the upper lateral line series and the dorsal fin origin, 6—8 (mode 6) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.2), 15 (f.24) or 16 (f.8) spinous and 8 (f.1), 9 (f-21) or 10 (f.12) branched rays. Caudal subtruncate, scaled on its basal half, or a little beyond. Pectoral 26-6—33°6 (M = 30-3) per cent of standard length, 80-0—97:8 (M = go-o) per cent of head. Pelvics with the first two rays produced, especially in adult males. Teeth. The outer row of teeth in both jaws (text-fig. 3) is composed principally of relatively stout, well-spaced, unequally bicuspid teeth ; anteriorly in the lower jaw, the teeth are implanted so as to slope forward at a slight angle. (=) (a) a a ve Va SS) SS Fic. 3. H. elegans. (Left). Premaxillary teeth, left side, viewed from a point slightly anterior of lateral. The teeth are from an anterolateral position in the jaw. (Right). Dentary teeth (right side), lateral in position. Viewed laterally. Scale = 0:25 mm. The major cusp is isoscelene in outline (see text-fig. 3) and very slightly incurved ; the neck of the tooth is slightly flattened in cross-section. Some teeth in each jaw have one margin of the major cusp partly flattened from below the tip so that it appears as a narrow step-like flange adjacent to the minor cusp (cf. H. aeneocolor where the flange is present on most teeth and is more obvious). Posteriorly in the upper jaw there may be from 1 to 5 unicuspid and dagger- shaped teeth ; less often these posterior teeth are tricuspid. There are 34—42 (mean = 38) teeth in the outer premaxillary row. The inner rows (usually 2, less commonly 3 in the upper jaw, and 2 or 3 in the lower) are composed of small tricuspid teeth, often irregularly arranged (particularly in the upper jaw). OstEoLoGy. The neurocranium of H. elegans is typically that of a generalized Haplochromis species (see Greenwood 1962), although the preorbital profile is a little less decurved. 148 P. H. GREENWOOD The lower pharyngeal bone is fairly stout, with its dentigerous area I-I—1-2 times broader than long. The teeth are fine, cuspidate and compressed and are arranged in 24-260 rows. Teeth in the two median rows are somewhat coarser than the others. Vertebral counts for the 30 specimens radiographed are : 28 (f.3), 29 (f.22) or 30 (f.5), comprising 12 (f.5), 13 (f.24) or 14 (f.1) abdominal and 15 (f.2), 16 (f.20) or 17 (f.8) caudal elements. COLORATION IN LIFE. Adult males: ground colour smokey grey overlying bluish- silver. Snout, lips and cheek with a livid iridescence. Belly and branchiostegal membrane dark cinder grey. Dorsal fin with the spinous part sooty, the inter- spinous membrane generally darkest ; lappets black but with a narrow red streak (or spot) at the tip. Soft dorsal with maroon streaks between the rays. Caudal fin with maroon spots and blotches between the rays, and a suffuse maroon flush around the fin margin. Anal dark hyaline (or faintly grey) often dusky at its base and with a pinkish-maroon border. The pelvics are black. Females: ground colour sandy green shading to silvery white on the belly and lower flanks. All fins yellowish-green. Because female H. elegans are not immedi- ately identifiable in the field, these ‘ live ’ colours are in fact ‘ post-mortem ’ colours and should not be considered at all precise. COLORATION IN PRESERVED SPECIMENS. Adult males: ground colour variable but basically grey-brown ; belly and chest dusky, as are, sometimes, the flanks. The flanks are generally crossed by 3—6 faint vertical bars. The branchiostegal membrane is black. Cephalic markings comprise a distinct lachrymal stripe, two bars across the snout and a broader bar immediately behind the orbits; in many specimens there is an even broader, but fainter, bar or blotch transversely across the nape. The lower part of the cheek and the vertical limb of the preoperculum are sometimes dusky. The dorsal fin is dusky, with darker streaks between the spines and rays, or the latter region maculate. Anal fin dusky or indistinctly maculate. Caudal with a dark central area and a light marginal zone. Pelvics are black, and the pectorals hyaline. Females have a greyish-yellow ground coloration, and sometimes very faint traces of 3—6 vertical bars on the flanks. On the head there are slight indications (sometimes just a darker area) of two bars across the snout, and a lachrymal stripe. All the fins are hyaline, the dorsal usually darker than the others ; the caudal is often maculate. Ecotocy. Habitat. Haplochromis elegans is essentially a species of the inshore regions of the lake, especially near papyrus shores or where the bottom is sandy. It rarely occurs in open-water localities or in shallow places where the substrate is mud. Food. Mostly chironomid larvae, although emergent aquatic Diptera are also eaten when available. Breeding. Haplochromis elegans is a female mouth-brooder. All specimens, of both sexes, within the size range studied are adult ; females appear to reach a larger size than do males. In the 15 sexually active females examined, 11 have the right LAKE GEORGE HAPLOCHROMIS SPECIES 149 ovary noticeably larger than the left, 2 have the ovaries equally developed, and 2 have the right ovary a little larger than the left one. Distribution. Lakes Edward and George, and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Within Lake George, H. elegans most closely re- sembles H. aeneocolor (see p. 150), both morphologically and in its habitat preferences. Adult males of the two species are readily distinguished by their coloration, but females and preserved specimens are differentiated chiefly by the fewer and some- what stouter teeth of H. elegans (32-42, M = 38, cf. 40-58, M = 48 in H. aeneocolor), the slightly procumbent anterior dentary teeth-of H. elegans, the well-developed flange on the major cusp of most teeth in H. aeneocolor (see p. 151), the shorter upper jaw of H. elegans (28-6-34:0, M = 30-3 per cent of head, cf. 30:0-37°8, M = 34:9 per cent) and the shorter lower jaw of H. elegans (35:0—40:2, M = 37-9 per cent of head, cf. 38-0-44-0, M = 41-0 per cent in H. aeneocolor). In life the lips of H. elegans appear thicker than those of H. aeneocolor, but this distinction is less obvious in preserved material. Haplochromis elegans shows few specializations in its dental or cranial anatomy, and must be ranked amongst the ‘ generalized’ Haplochromis species. Outside the Lake Edward—Lake George species complex it resembles H. pallidus (Blgr.) of Lake Victoria (see Greenwood 1960). From H. pallidus, H. elegans differs in its adult male coloration, some morphometric characters (e.g. having a shorter snout) and in its overall morphology. The significance of this apparent resemblance will be discussed elsewhere (p. 230) ; however, it should be noted that the resemblance between H. elegans and H. pallidus cannot be shown to be more significant than that existing between it and species of the H. bloyeti complex (see Greenwood 1971). Trewavas (1933) compared H. elegans with H. cinereus (Blgr.) of Lake Victoria, but this comparison is no longer valid now that we have a clearer concept of H. cinereus (see Greenwood 1960). In fact, H. cinereus shows some specialized characters (its dentition for one). These specializations would not be apparent in 1933, because at that time ‘ H. cinereus’ was a dumping ground for several of the generalized Lake Victoria species. Resemblances which I noted between H. elegans and H. velifer Trewavas of Lake Nabugabo (Greenwood 1965b) are somewhat diluted by the greater amount of information now available on H. elegans. For example, the teeth of H. elegans (at least in Lake George populations) have a more acutely pointed cusp, and there are fewer teeth in the outer premaxillary row. There is, of course, a marked difference in the male breeding coloration of the two species. Diagnostic problems arising in connection with H. elegans and species at present known only from Lake Edward (and then very imperfectly known) are virtually identical with those discussed in relation to H. aeneocolor on page 153. STUDY MATERIAL Register number BMNH Locality : Lake George I972.6.2 : 166-167 N.E. lake shore (papyrus) 1972.6.2 : 168-171 Papyrus shore off I.B.P. Laboratory 1972.6.2 : 172-177 Kankurunga Island 150 P. H. GREENWOOD 1972.0.2 : 178-179 Kankurunga Island 1972.6.2 : 180-182 Kankurunga Island 1972.6.2 : 183-206 Kankurunga Island 1972.6.2 : 207-224 Kankurunga Island 1972.6.2 : 225-230 Kankurunga Island 1972.0.2 : 231-236 Kankurunga Island 1972.6.2 : 237-238 Akika Island 1972.6.2 : 239-243 Akika Island 1972.6.2 : 244-260 Akika Island 1972.6.2 : 261-272 IB IPanetby: 1972.6.2 : 273-285 Kashaka Bay 1972.0.2 : 280-2091 Tufmac Bay 1972.6.2 : 292 Close to shore (muddy) Haplochromis aeneocolor sp. nov. (Text-figs. 4 & 5) Haplochvomis nubilus (part): Trewavas, 1933, J. Linn. Soc. (Zool.) 38: 329 (4 specimens, BMNH reg. nos. 1933.2.23 : 296-299). Hototyre. A male, 68-0 mm standard length, BMNH reg. no. 1972.6.2 : 43. The specific name refers to the brassy appearance of adult males. DESCRIPTION. Based on 36 specimens (including the holotype), 58-o—75-0 mm standard length. Depth of body 35-7—41-1 (M = 37-7) per cent of standard length, length of head 32:0—36°8 (M = 34°5) percent. Dorsal profile of head straight or slightly concave, sloping fairly steeply at ca 35°—40° with the horizontal ; dorsal margin of orbit not entering the line of the profile but distinctly below it. Preorbital depth 12-0-18-2 (M = 14:6) per cent of head (not showing any clear- cut allometry), least interorbital width 22-7—29:3 (M = 25°5) per cent, snout length 26-7—31°8 (M = 28-8) per cent, 0-8-1-0 (mode 0-9) of its breadth. Eye diameter 28-6—35:0 (M = 31-4) per cent of head (showing no obvious allometry), depth of cheek 19:0—25-0 (M = 22°8) per cent. Caudal peduncle 12:9-17-4 (M = 15-3) per cent of standard length, 1-2-1-5 (modal range 1-2—1-3) times as long as deep. Mouth angle ranging from horizontal to slightly oblique ; lips somewhat thickened. Length of upper jaw 30:0—37-°8 (M = 34:9) per cent of head, lower jaw 38-0—44-0 (M = 41-0) per cent of head, 1-5-2-1 (modal range 1-6—1-8) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior part of the eye or even to a vertical through the anterior margin of the pupil. Gill vakers of various shapes, from short and stout to relatively slender; the lower I or 2 rakers on the first gill arch are greatly reduced, the upper 2 or 3 often flattened. There are 8 or g rakers on the lower part of the first arch. Pseudorakers are poorly developed. Scales. Ctenoid; lateral line with 30 (f.11), 31 (f.18), 32 (f.4) or 33 (f.1) scales, cheek with 3 (rarely 2) rows. Five to 64 (mode 534) scales between the lateral line LAKE GEORGE HAPLOCHROMIS SPECIES 151 and the dorsal fin origin, 6 (mode) or 7, rarely 54 or 5, between the pectoral and pelvic fin bases. Fic. 4. Haplochromis aeneocolor. Holotype. Fins. Dorsal with 14 (f.3), 15 (f.31) or 16 (f.2) spinous and 8 (f.3), 9 (f.25) or Io (f.8) branched rays. Anal with 3 spines and 7 (f.1), 8 (f.9), 9 (f.20) or ro (f.6) branched rays. Caudal subtruncate, scaled on its basal half. Pectoral 27-0—31-7 (M = 28-9) per cent of standard length, 73-8—93-5 (M = 84:5) per cent of head. Pelvics with the first and second rays produced (the first markedly so), and relatively longer in adult males than in females. Teeth. The outer teeth, although basically of the generalized, unequally bicuspid type, are nevertheless rather distinctive. This is due to the presence of a well- developed, thin flange on that margin of the outer cusp which is adjacent to the minor cusp (see text-fig. 5). Few individuals fail to show flange development on at least the majority of anterior and lateral teeth in both jaws. The flange can be so well developed that the tooth seems to have an expanded and obliquely sloping major cusp (i.e. to be like the teeth of H. limax, see p. 168). Usually the flange is thin and almost transparent, so that there appears to be a dividing line between it and the more substantial body of the cusp itself. Although the flange may be continuous with the occlusal (i.e. distal) part of the cusp (thereby simulating an H. limax-like tooth) it is generally confined to the proximal half or two-thirds of the cusp. In this way a distinct step is developed between the flange and the occlusal tip of the cusp. Apart from the flange, outer teeth in H. aeneocolor are typical bicuspids, with the major cusp having the outline of an isosceles triangle rather than of an equilateral one. The minor cusp is well developed, but its tip is not very acute. The crown of 152 P. H. GREENWOOD an outer tooth has virtually no incurvature, and the neck is a slightly compressed cylinder. The posterior I—4 upper teeth are either compressed tricuspids or are unicuspid and caniniform. \ At " Fic. 5. H. aeneocoloy. Dentary teeth (left side), lateral in position. Viewed laterally. Scale = 0-25 mm. There are 40-56 (mean 48) teeth in the outer row of the upper jaw. In a few specimens, all the outer teeth in the lower jaw are unicuspid, but the upper teeth retain a typical bicuspid form. The inner teeth in both jaws are small, compressed and tricuspid, and are arranged in 2 or 3 (rarely 4) seriesin the upper jaw, and in 2 (rarely I or 3) series in the lower jaw. OsTEOLOGY. The neurocranium is of the typical generalized Haplochromis type (see Greenwood 1962), but with the preorbital profile sightly straighter. The lower pharyngeal bone is moderately stout, its dentigerous area equilateral or slightly broader than long. The teeth are fine, compressed and cuspidate, and are arranged in ca 24-26 rows; the median teeth are not noticeably larger or coarser than those of the lateral rows. Vertebral counts in the 16 fishes examined are 27 (f.1), 28 (f.7) and 29 (f.8), com- prising 12 (f.6) or 13 (f.10) abdominal and 15 (f.4), 16 (fr) or 17 (f.1) caudal elements. COLORATION IN LIFE. Adult males: the flanks, lateral aspect of the chest and belly, lower part of the head, the branchiostegal membrane and the lips are dark sulphurous yellow, with an orange overlay on the operculum. The rest of the flank (i.e. the posterior part) and the caudal peduncle are yellowish-green with a faint bluish overlay, and the ventral aspect of the chest is sooty. The dorsal body surface is dull bronze posteriorly, becoming purple above the flanks, and crimson anteriorly. The snout and anterior dorsum of the head are puce. The overall colour impression gained from a newly caught male is one of brassyness, despite the various colour elements described above. The dorsal fin is dark hyaline on the spinous part (the lappets black), but lighter on the soft part where the margin is crimson. The anal fin is hyaline over the basal third of the soft part, but with the spines and distal two-thirds of the soft part pinkish-crimson ; the ocelli are orange-yellow. The caudal fin is pinkish to red, the colour intensified on the ventral third of the fin and at its posterior angle. The pelvic fins are black, the pectorals hyaline. Male coloration is difficult to describe adequately because the intensity of the various colours is variable and changes rapidly after the fish is removed from water. Some fishes, for example, appear almost black a short while after capture. LAKE GEORGE HAPLOCHROMIS SPECIES 153 PRESERVED COLORATION. Males: the ground colour is essentially like that described for H. elegans (see p. 148), but in H. aeneocolor the dark ventral pigment is more extensive ; in some individuals it covers the entire caudal peduncle and the flanks to a level just below the upper lateral line. Cephalic markings are identical in both species. The dorsal fin is dusky to black; if dusky, the pigment is often concentrated basally so that this region of the fin is almost black. The caudal is more or less uniformly dark, except for hyaline areas on the ventral and posteroventral margin. The anal varies from grey to dusky; the area over the spines is generally black. The pelvic fins are black, the pectorals hyaline. Females have a greyish-silver to greyish-yellow ground colour; the head shows very faint traces of two transverse bars across the snout and an ill-defined, short, lachrymal stripe or streak. The dorsal fin has dark streaks between the rays, especially on the spinous part. The caudal is maculate, usually weakly so, and with the spots most obvious on the centre of the fin; a few specimens have intense maculae distributed over most of the fin. The anal is hyaline as are the pelvics (which may be faintly dusky). Ecoitocy. Habitat. This species is particularly common near papyrus shores, and is rare elsewhere in the inshore region. Apparently it never occurs offshore. Food. Haplochromis aeneocolor seems to be a detritus feeder since plant fragments and insect larvae are predominant elements of its gut contents. Adult insects are, however, also eaten. Breeding. Haplochromic aeneocolor is a female mouth brooder. Of the ro adult females examined, the right ovary is much larger than the left in 6 individuals, slightly larger in 3 and of equal size in 1 fish. All specimens within the size range studied are adult and there is apparently no sexual dimorphism in the maximum size attained. Distribution. Lake George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Until more is known about most Lake Edward species (and especially those not recorded from Lake George) an adequate diagnosis for H. aeneocolor is difficult to compile. For example, H. eduardii Regan super- ficially resembles H. aeneocolor but appears to differ in having stouter, non-flanged and less acutely cuspidate teeth, a shallower body and more rounded (i.e. decurved) head profile. Haplochromis engystoma Trewavas (known only from the holotype and one other specimen) has dental characteristics more like those of H. aeneocolor, but differs in several morphometric characters, especially in its higher (2-0) eye/cheek ratio, shorter lower jaw (34°8 per cent of head) and its strongly decurved head profile. Haplochromis vicarius Trewavas (at least as restricted to the holotype) has an overall superficial resemblance, but differs in having obliquely cuspidate outer teeth, more rows of inner teeth and a larger eye (36-0 per cent of head) ; Poll (1939) has synony- mized H. vicarius with H. eduardii but I doubt the correctness of this decision (see Appendix I). Regrettably, in none of these comparisons could life colours be taken into account because these are unavailable for the Lake Edward species. 154 P. H. GREENWOOD Considering now those species which also occur in Lake George, H. elegans has the closest overall and detailed resemblances with H. aeneocolor (see p. 149). Male coloration is, however, very different in the two species, and there are dental and morphometric differences as well (again, see p. 147). Distinguishing between females of the species is especially difficult, although in the field the more rounded head profile of H. elegans does give some guidance for preliminary sorting. Haplochromis limax Trewavas shows a fairly marked resemblance to H. aeneocolor in its superficial morphology and there is also a certain convergence in dental mor- phology. This arises from the peculiar flange developed on the major cusp of outer teeth in H. aeneocolor (see p. 152). If this flange is hypertrophied, it increases the area of the major cusp and imparts to it an oblique cutting edge. However, if such teeth in H. aeneocolor are closely examined, the junction between flange and main body is apparent, as is a slight indentation on the cutting edge. Furthermore, the flange is much thinner (nearly transparent) than the corresponding margin of a tooth in H. limax. Another distinguishing feature of H. limax is the broader array of inner teeth, and their larger size. Again, male breeding coloration is verv different in the two species. Beyond the confines of Lakes Edward and George, H. aeneocolor, like H. elegans, resembles the generalized Haplochromis species of Lake Victoria, in particular H. pallidus. But, in the absence of any clearly defined specializations in the species involved, little significance can be attached to these resemblances. The peculiar flange formation on the outer teeth of H. aeneocolor is an unusual feature for Haplo- chromis but its recognition as a specialization remains to be confirmed. Certainly it is rarely manifest among Lake Victoria species, but it does occur more frequently (if only as an individual variant) amongst the species of Lakes Edward and George. Finally, and as if to reinforce the generalized nature of H. aeneocolor anatomy, the resemblance between this species and H. nubilus (Blgr.) should be noted. Haplo- chromis nubilus is one of the anatomically and ecologically most generalized species occurring in the Victoria~Edward drainage basin (see p. 221), and in turn shows close affinity with the fluviatile species of east Africa. On all morphometric charac- ters H. nubilus and H. aeneocolor cannot be separated, but male coloration is markedly distinct, the caudal of H. nubilus has a nearly round distal outline, flanged teeth are not found (the teeth are unicuspid in large fishes) and the dorsal head profile is more concave than in H. aeneocolor. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.0.2 : 43 (Holotype) N.E. shore near River Mpanga mouth 1972.6.2 : 44-50 (Paratypes) Kankurunga Island 1972.6.2 : 52-54 (Paratypes) Kashaka Crater 1972.6.2 : 55-63 (Paratypes) N.E. shore 1972.6.2 :64—67 (Paratypes) N.E. shore 1972.6.2 : 68—72 (Paratypes) Kashaka Bay 1972.0.2 : 73-79 (Paratypes) Kankurunga Island 1972.6.2 : 81-84 Papyrus shore off I.B.P. Laboratory SS LAKE GEORGE HAPLOCHROMIS SPECIES 155 1972.6.2 : 85-103 N.E. shore 1972.6.2 : 104—-III Kankurunga Island Haplochromis nigripinnis Regan, 1921 (Text-figs. 6 & 7) Haplochromis nigripinnis Regan, 1921, Ann. Mag. nat. Hist. (9) 8 : 635. Haplochromis nigripinnis : Trewavas, 1933, J. Linn. Soc. (Zool.), 38: 330 (refers to Lake Edward fishes only). Haplochromis guiarti (part) : Trewavas, 1933, op. cit. : 340 (1 of the 3 small fishes from Worth- ington’s (1932) stations 613 and 618, Lake George, viz. BMNH reg. no. 1933.2.23 : 476. HoLotype. A male (probably adult), 64-0mm standard length from Lake Edward, BMNH reg. no. 1914.4.8 : 14. DESCRIPTION. Based on 36 specimens (excluding the holotype), 50:0—68-0 mm standard length. Depth of body 32-3—41-5 (M = 36:8) per cent of standard length, length of head 31°5-35°3 (M = 33:6) per cent. Dorsal head profile straight or gently curved, sloping at ca 35°—40° to the horizontal ; dorsal margin of orbit barely entering the line of the head profile. Preorbital depth 10-0-14:6 (M = 13-0) per cent of head, least interorbital width 20-7-27°5 (M = 24:2) per cent, ratio of interorbital width to eye diameter 1:28-1:75 (M = 1-49). Snout length 24-0-30-3 (M = 27-4) per cent of head, 0-8-0-9 (rarely I-0) its breadth; eye diameter 33-3—40-0 (M = 35-8) per cent, with no detectable allometry ; depth >f cheek 16-3~—22-9 (M = 19°8) per cent. Fic. 6. Haplochromis nigripinnis. Lake George specimen ; a male. Caudal peduncle 15-8—19:8 (M = 17-4) per cent of standard length, 1-3-1°8 (modal range 1-5~-1-7) times as long as deep. Mouth horizontal, lips not noticeably thickened. Length of upper jaw 30:6—36-9 (M = 34-0) per cent of head, length of lower jaw 39:0-47-5 (M = 43:6) per cent, 9 156 P. H. GREENWOOD 1-8—2-2 (modal range I:g—2:0) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior part of the eye or even to one through the anterior margin of the pupil. A noticeable feature of the snout in H. nigripinnis is the size of the anterior open- ing to the nasal laterosensory canal. The opening is as large as (or almost as large as) the nostril. In most other Haplochromis species from Lake George (and apparently Lake Edward also) the opening to this canal is much smaller than the nostril, and is often difficult to locate. The intestine in H. nigripinnis is long (ca 2—24 times total length) and much coiled on itself. Gill vakers on the first arch are, except for the reduced lower 1—3 and the occasional flattened and anvil-shaped upper 1-3, slender and relatively elongate. There are 8-10 (mode 9), rarely 11, rakers on the lower part of this arch. No pseudorakers are developed (see p. 145). Scales. Ctenoid ; lateral line with 30 (f.6), 31 (f.13), 32 (f.14) or 33 (f.3) scales, cheek with 2 or 3 (mode) rows. Five to 64 (mode 54) scales between the upper lateral line and the dorsal fin origin, 6 or 7 (mode), rarely 5, between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.1), 15 (f.28) or 16 (f.7) spinous and 8 (f.7), 9 (f.24) or to (f.5) branched rays. Caudal generally truncate but weakly emarginate in some fishes ; scaled on its basal half. Pectorals 25-8—31-3 (M = 29:7) per cent of standard length, 79°5—94°5 (M = 87-3) per cent of head. Pelvics with the first ray slightly produced. Teeth. The outer teeth in both jaws (except posteriorly in the upper) are slender, compressed and unequally bicuspid (text-fig. 7). The outline of the major cusp varies from equilateral to isoscelene ; all intergrades may occur in one individual or one type of cusp outline may predominate. The crown is slightly incurved. Posterior teeth in the upper jaw are often either unicuspid and slender, or small, compressed and tricuspid. There are 40-60 (M = 52) teeth in the outer premaxil- lary row. Fic. 7. H.nigripinnis. Premaxillary teeth (left), anterolateral in position. Viewed from an anterolateral eT Scale = 0-25 mm. Teeth of the nner rows are small, compressed and tricuspid, and are arranged in I or 2 rows (rarely in 3) in the upper jaw, and 1 or 2 in the lower jaw. The serial arrangement of these teeth is often rather irregular. OstEoLocy. The neuvocranium of H. nigripinnis is identical with that of H. elegans and H. aeneocolor, that is, of a generalized type. LAKE GEORGE HAPLOCHROMIS SPECIES 157 The lower pharyngeal bone gives an impression of being long, slender and fine (especially when compared with the bone in H. elegans or H. aeneocolor). Its denti- gerous surface, however, is about 1-2 times broader than long. The teeth on this bone are fine, slender and cuspidate, and are arranged in ca 34 rows ; teeth situated on the posterolateral angles of the bone are more densely crowded than elsewhere. Vertebral counts in the 13 specimens radiographed are 28 (f.1), 29 (f.9), 30 (f£.2) or 31 (f.1), and comprise 12 (f.2), 13 (f.10) or 14 (f.1) abdominal and 16 (f.9) or 17 (f.4) caudal elements. COLORATION IN LIFE. Adult males: ground colour (including that of the head) is a dark malachite green with a silvery underlay ; the branchiostegal membrane is black. On the head there is a prominent lachrymal stripe and a less intense interorbital bar ; both marks are intensified after death. The dorsal fin is dark grey, with black lappets, and a darker irregular line along the base ; a pink suffusion is visible over the soft part of the fin. The caudal has an overall pink flush except basally, where the membrane is dark hyaline. The anal is black over the spinous part, pinkish elsewhere (the colour intensifying distally) ; ocelli orange-yellow. Pelvic fins are black. Adult females have an overall greyish-silver coloration above a midlateral line, and are chalky white below that level. The upper half of the caudal fin is hyaline but the basal area and lower half of the fin are suffused with pale lemon-yellow ; this pigmentation sometimes extends over the entire fin but even then is most intense on the lower half. Some individuals show dark rather elongate spots along the middle of the caudal. The anal has, distally, similar yellow colour to that of the caudal but it is hyaline basally. The dorsal and pelvic fins both are hyaline. PRESERVED COLORATION. Adult males: the ground colour is dark brown to black below the midlateral line, and to varying degrees above that level as well. When dark pigment does extend dorsally it is generally less intense than on the ventral body. In some fishes up to seven dark, fairly narrow vertical bars extend across the light brown of the upper body ; sometimes there is a longitudinal dark bar extending for a variable length along the upper lateral line scale row. The ventral half of the head is dark brown, the branchiostegal membrane black. A fairly distinct lachrymal stripe is usually visible through the general dark ground coloration of the snout. Two bars (of equal thickness) cross the snout, and often there is a small median blotch above the posterodorsal margin of the orbit. A larger dark blotch crosses the nape anterior to the dorsal fin origin ; this mark seems to be a medial continuation of the first vertical bar of the flank. The dorsal fin is dark, black on its proximal half and dusky beyond ; the lappets are black. The caudal fin has a dark central area basally but otherwise it is greyish. The anal is black on its basal half, and dusky or hyaline distally. The pelvics are uniformly black or blotched black and dusky, the outer half of the fin being the darker part. The pectorals are hyaline. Females have a greyish-silver to greyish-brown ground colour, and are lighter ventrally. In some few specimens very faint indications of vertical bars are visible on the flanks ; such marks are confined to the central flank region and do not extend 158 P. H. GREENWOOD as far as the dorsal or ventral body outline. All fins are hyaline, the dorsal sometimes greyish with dark lappets. Ecotocy. Habitat. Although predominantly a species of offshore areas and the open central part of the lake, H. migrvipinnis is sometimes found within a few feet of the shoreline, especially where the substrate is sandy. Food. As the long and coiled intestine suggests, H. nigripinnis is a vegetarian species. It feeds principally on suspended phytoplankton ; there is no indication from the gut contents of any bottom feeding habits. Like many Haplochromis species, H. nigripinnis is an opportunistic feeder ; insect remains (of both larvae and pupae) are recorded from the gut. Breeding. Female mouth brooding is practised by H. nigripinnis. Of the 8 adult females examined, 5 have the right ovary noticeably larger than the left, 1 has the right ovary slightly the larger and 2 show equal ovarian development. Distribution. Lakes Edward and George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Trewavas (1933) compared H. migripinnis with H. cinereus (Blgr.) of Lake Victoria. The invalidity of this comparison has been commented upon above (p. 149). It is due entirely to there being, at that time, in- sufficient material of either species to permit of precise comparison. Surprisingly, neither Trewavas (op. cit.) nor Regan (1921) compared H. nigripinnis with any other species in the Edward—George complex. Haplochromis nigripinnis is indeed a distinctive species, especially when its coloration, fine dentition, long gut, fine gill rakers and its feeding habits are considered. But, preserved specimens (or live females) have a great similarity with specimens of H. macropsoides, a new taxon described on p. 162. Haplochromis ngripinnis differs from H. macropsoides in having finer outer teeth in both jaws, fewer rows of inner jaw teeth, a longer and more slender caudal peduncle, and in having a relatively larger opening to the nasal laterosensory canal (see above, p. 156). The same character combination (and especially the dental ones) serves to dis- tinguish H. nigvipinnis from such species as H. elegans, H. aeneocolor and H. limax. In all instances, of course, male breeding coloration provides the most outstanding interspecific difference. Outside Lakes Edward and George, the greatest morpho-anatomical (and ecological) resemblances are with H. erythrocephalus Greenwood & Gee, of Lake Victoria (Greenwood & Gee 1969). Both species have, besides a similar gross morphology, a diet of phytoplankton, fine and numerous teeth, slender gill rakers and a long, coiled intestine ; all, of course, correlated characters within each species. Male coloration is particularly different. Male H. erythrocephalus have a bright red head, while the head in H. nigripinnis is dark malachite green (cf. p. 157 above with p. 2r in Greenwood & Gee, op. cit.). There are several other interspecific differences, particularly in the pharyngeal dentition and the neurocranial shape. Skull form in H. erythrocephalus is more like that in the moderately specialized Haplochromis species, and the pharyngeal teeth are finer, more numerous and more densely arranged than in H. nigripinnis. Haplochromis erythrocephalus also has ii LAKE GEORGE HAPLOCHROMIS SPECIES 159 relatively longer and more slender gill rakers. In other words, H. erythrocephalus shows greater specialization for phytoplankton feeding than does H. nigripinnis. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 549-554 Tufmac Bay (trawl) 1972.0.2 : 598-601 Small island north of Kankurunga Island 1972.6.2 : 602—605 Small island north of Kankurunga Island 1972.6.2 : 606 Kashaka Bay 1972.0.2 : 607-611 Kashaka Bay~ 1972.6.2 : 648-654 East side of Akika Island 1972.0.2 : 636-647 Mid-lake ca 5 miles east of Kankurunga Island 1972.6.2 : 805 (figured specimen) Small island north of Akika Island (trawl) Haplochromis oregosoma sp. nov. (Text-figs. 8 & 9) Hororyre. A female, 66-5 mm standard length, BMNH reg. no. 1972.6.2 : 141. The specific name (from the Greek orvego to stretch and soma the body) alludes to the rather elongate form of this species. DEscRIPTION. Based on 20 specimens, including the holotype, 48-0—72°5 mm standard length. Depth of body 30:3—34:3 (M = 32:1) per cent of standard length, length of head 32:0-36°0 (M = 33°8) per cent. Dorsal head profile gently curved, less commonly straight, sloping at an angle of ca 35° to the horizontal ; dorsal margin of eye entering the profile or extending slightly above it. Preorbital depth 10-5-15-2 (M = 13-3) per cent of head (showing ill-defined positive allometry with standard length), least interorbital width 20-0—25:0 (M = 22-3) per cent, ratio of interorbital width to eye diameter 1-5-1°8 (M = 1-7). Snout length 23-5—29-2 (M = 26-1) per cent of head, 0-7—0-9 (rarely 1-0) times broader than long; eye diameter 33-4—41-2 (M = 38-2) per cent, depth of cheek I5:2—20°8 (M = 17-9) per cent. Caudal peduncle 15-g—21-1 (M = 17:9) per cent of standard length, 1-4-2-0 (modal range I-6-1-7) times as long as deep. Mouth horizontal, lips not thickened. Length of upper jaw 28-6—34-7 (M = 31-7) per cent of head, length of lower jaw 38-1-—45:6 (M = 41-9) per cent and 1I:-6—2°3 (mode 2-0) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior margin of the eye. Intestine about 1} times the total length. Gill vakers. The lower 1-3 on the first arch are reduced, the remainder either all slender and elongate or, less commonly, with the upper 2—4 flattened and branched. There are g—11 (mode 10) rakers on the lower part of the first arch. No pseudorakers are developed. Scales. Ctenoid; lateral line with 30 (f.1), 31 (f.3), 32 (f.10) or 33 (f.5), cheek with 2 or 3 rows. Five to 6 (bimodal at 54 and 6) scales between the upper lateral line and dorsal fin origin, 5 or 6 (mode) between the pectoral and pelvic fin bases. 160 P. H. GREENWOOD Fins. Dorsal with 15 (f.11) or 16 (f.9) spines, and 8 (f.5) 9 (f.9) or ro (f.6) branched rays. Anal with 3 spines and 7 (f.1), 8 (f.8) or g (f.11) branched rays. Caudal slightly emarginate, scaled on its basal half or a little further posteriorly. Pectoral fin 25-6—30-0 (M = 28-1) per cent of standard length, 75-0—88-0 (M = 82-7) per cent of head. Pelvics with the first ray slightly produced. Fic. 8. Haplochromis oregosoma. Holotype. Teeth. The predominant tooth form in the ower row of both jaws is a moderately slender, compressed and unequally bicuspid tooth (text-fig. 9). The major cusp in such teeth is produced, isoscelene in outline and slightly incurved. A distinct, step- like flange is sometimes developed on that margin of the cusp adjacent to the minor cusp. Some teeth may have the minor cusp greatly reduced in size. Posterolateral teeth in the upper jaw may be bicuspid like the others, slender unicuspids or com- pressed tricuspids. Tricuspid teeth are occasionally intercalated amongst the anteriorly situated bicuspids in either or both jaws. There are 42-60 (M = 50) outer teeth in the premaxilla. Fic. 9. H. ovegosoma. Premaxillary teeth (left side), anterior in position. Viewed anteriorly. Scale = 0-25 mm. Inner tooth rows in both jaws are composed of small tricuspid and compressed teeth, arranged in a single (rarely double) series. OstEoLocy. Basically, the neurocranium of H. oregosoma is of the generalized Haplochromis type, but differs in having a low supraoccipital crest and a relatively shorter ethmo-vomerine region. LAKE GEORGE HAPLOCHROMIS SPECIES 161 The lower pharyngeal bone is fine, its dentigerous area slightly broader than long (I:I-1-2 times). The teeth are slender and cuspidate, and are arranged in ca 26-30 rows. Vertebral counts in the 12 specimens radiographed are 29 (f.3) or 30 (f.9), comprising 13 (f.5) or 14 (f.7) abdominal and 15 (f.1), 16 (f.8) or 17 (f.3) caudal centra. COLORATION IN LIFE. Adult males: ground colour metallic purple above the midlateral line, shading through iridescent turquoise to silvery on the belly and ventral flanks; a sooty overlay spreads across the chest and belly. The dorsal fin is a dark sooty colour, as is most of the anal except for its scarlet tip ; the anal ocelli are large, near circular in outline, and orange-yellow in colour. The entire caudal fin is scarlet, but the basal fifth may be sooty or solid black. The pelvic fins are uniformly black. Adult females are an overall greyish-silver. The dorsal fin is hyaline, the anal pale yellow, the caudal dark hyaline and the pelvics greyish. PRESERVED COLORATION. Adult males: ground coloration black or intensely dusky, with a silvery underlay ; the chest and midventral aspects of the belly are light dusky. No distinct markings are visible on the head (probably they are obscured by the general dark coloration). Dorsal fin dusky to black, the pigment most intense between the rays. Anal black over its distal two-thirds, dusky beyond. Caudal with variable coloration but always dark over the proximal third ; distally the fin is usually yellowish with a dusky overlay that intensifies between the middle rays. The pelvic fins are black, the pectorals hyaline. Adult females are silvery-grey, shading to silver on the chest and belly; the dorsum and snout are dark grey. No cephalic markings are visible. All fins are greyish-hyaline. Biotocy. Very little is known about the biology of H. oregosoma. Apparently the species is confined to offshore areas of the lake, and it does not occur close to papyrus or other shores. Specimens have been caught over both sand and mud substrata. The breeding habits are unknown ; of g adult females examined, 4 have the right ovary considerably larger than the left one, 4 have the ovaries equally developed and 1 has the left ovary larger than the right one. Individuals less than 55 mm standard length are immature, although males of 56 mm standard length are ripen- ing. Fishes, of both sexes, are fully adult at 60 mm standard length. Haplochromis oregosoma seems to feed on phytoplankton, but as yet too few speci- mens have been examined to establish whether the food is taken in suspension or from bottom deposits. Distribution. Lake George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. From all other Haplochromis species in Lake George, H. oregosoma is distinguished by the following character combination : large eye, numerous and slender gill rakers, slender and elongate body form. Perhaps the species showing most superficial similarity with H. oregosoma is H. nigripinnis ; the characters listed above, together with a difference in eye/interorbital ratio (1-5—1'8, mean 1:7, cf. 1:3-1-8, mean 1°5 for H. nigripinnis) serve to distinguish the species. 162 P. H. GREENWOOD The totality of characters (including skull and jaw form) suggest that H. nigripinnis and H. oregosoma are probably not very closely related. Among the Lake Edward species not recorded from Lake George, there is some resemblance between H. oregosoma and H. engystoma Trewavas. Unfortunately, H. engystoma is known only from the holotype (now in a poor state of preservation) and another, much smaller specimen which may not be a member of the species. Comparing H. engystoma holotype with H. oregosoma, the latter differs in having more gill rakers (10 or I1, cf. 8), straighter teeth, longer lower jaw (38-1-45:6, M = 41-9 per cent head, cf. 34:8 per cent in H. engystoma), and a slightly larger eye (33°4-41:2, M = 38-2 per cent head, cf. 34:8 per cent). No known species from Lake Victoria shows any close resemblance to H. oregosoma. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 141 (Holotype) North end of Kankurunga Island 1972.6.2 : 142-146 (Paratypes) North end of Kankurunga Island 1972.6.2 : 147 (Paratype) Northern tip Kankurunga Island 1972.6.2 : 148-152 (Paratypes) Northern tip Kankurunga Island 1972.6.2 : 153 (Paratype) Kankurunga Island 1972.6.2 : 154 (Paratype) Tufmac Bay 1972.6.2 : 165 (Paratype) In sandy shallows 1972.0.2 : 155-160 Tufmac Bay 1972.6.2 : 161-164 Tufmac Bay Haplochromis macropsoides sp. nov. (Text-figs. 10-12) H. vicavius (part) Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 330-331 (1 of the paratypes, BMNH reg. no. 1933.2.23 : 353 from Worthington’s station 613, Lake George. Trewavas incorrectly lists this station as ‘.. . East shore of Lake Edward ’, but vide Worthington 1932). HototyPe. A male, 76-0 mm standard length, BMNH reg. no. 1972.6.2 : 718. The trivial name refers to the overall similarity between this species and H. macrops (Blgr.) of Lake Victoria (and, so it was once thought, of Lake Edward as well). DESCRIPTION. Based on 30 specimens (including the holotype, and the paratype of H. vicarius, see above), 59:0—77-0 mm standard length. Depth of body 34:9—41:0 (M = 36:5) per cent of standard length, length of head 32:2—37:2 (M = 34:3) percent. Dorsal head profile straight or very slightly curved, sloping at an angle of ca 35°—40° to the horizontal ; dorsal margin of the eye entering the line of the profile or extending slightly above it. Preorbital depth 12-0-15-2 (M = 13°5) per cent of head, least interorbital width 20°5-25:0 (M = 23-1) per cent. Snout length 25-0—30:5 (M = 27-6) per cent, o-8-o-g times broader than long, eye diameter 33:3—39:1 (M = 36-0) per cent, showing very slight negative allometry with standard length, depth of cheek 17°5—23°8 (M = 21-4) per cent. if - J - 7 LAKE GEORGE HAPLOCHROMIS SPECIES 163 Caudal peduncle 13-6—19-3 (M = 15-8) per cent of standard length, 1-0—1-5 (modal range I-2—1-3) times as long as deep. Fic. 10. Haplochvomis macropsoides. Holotype. Mouth horizontal, lips not thickened. Length of upper jaw 31-3—41°5 (M = 34:2) per cent of head, length of lower jaw 39:2—45°4 (M = 41-0) per cent, I-7—2-2 (modal range I-7-I-9) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior margin of the eye or somewhat behind that level. The anterior opening to the nasal laterosensory canal is much smaller than the nostril (cf. H. nigripinnis), and indeed, can be difficult to locate. The intestine is 14-14 times the total length. Gill rakers. Except for the reduced lower 1 or 2 rakers, others on the first gill arch are relatively slender, although some of the uppermost ones may be flat and some lower ones stout. There are 8-10 (mode 9) rakers on the lower part of this arch. Pseudorakers are present, but are poorly developed and small. Scales. Ctenoid; lateral line with 30 (f.2), 31 (f.9), 32 (f.15) or 33 (f.4) scales, cheek with 2 or 3 (bimodal) rows. Five to 64 (mode 6) scales between the upper lateral line and the dorsal fin origin, 6 or 7 (rarely 5) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.4), 15 (f.20) or 16 (f.6) spinous and 8 (f.4), 9 (f.10) or Io (f.16) branched rays. Anal with 3 spines and 8 (f.7), 9 (f.22) or 10 (f.1) branched rays. Caudal truncate or, less frequently, weakly emarginate ; scaled on its basal half. Pectorals 25-7—32°3 (M = 29-4) per cent of standard length, 73:0-95:5 (M = 85:0) per cent of head. Pelvics with the first ray slightly produced. Teeth. The majority of the outer teeth in both jaws are relatively stout, compressed and very unequally bicuspid ; the larger cusp is isoscelene in outline and the crown slightly incurved (text-fig. 11). Posteriorly in the upper jaw the last few teeth are often slender and unicuspid. Some tricuspid teeth may be intercalated amongst the bicuspids anteriorly in either or both jaws. 164 P. H. GREENWOOD al Fic. 11. H. macropsoides. Dentary teeth (right side), anterolateral in position. Viewed anteriorly. Scale = 0-5 mm. As described for H. elegans (see p. 147), some bicuspids can have one margin of the major cusp produced into a narrow flange. There are 42—60 (M = 52) teeth in the outer premaxillary series. The inner tooth rows (2 or 3 in both jaws) are composed of tricuspid, compressed teeth. OstEoLoGy. The neurocranium of H. macropsoides, identical with that of H. elegans, H. aeneocolory and H. nigripinnis, is of a generalized Haplochromis type. The lower pharyngeal bone (text-fig. 12) is moderately fine, with its dentigerous surface ca 1:2 times broader than long. The pharyngeal teeth are slender, com- pressed and cuspidate, and are arranged in ca 32-36 rows. Teeth in the two median rows are not noticeably coarser than their lateral congeners ; teeth situated in the posterolateral corners of the bone are more closely set than elsewhere. Fic. 12. H. macropsoides. Lower pharyngeal bone, occlusal view. Scale = I-o mm. Vertebral counts in the 7 specimens examined are: 29 (f.6) or 30 (f.1), comprising 13 abdominal and 16 (f.6) or 17 (f.1) caudal centra. COLORATION IN LIFE. Adult males: ground colour smokey grey overlying bluish- silver ; snout, cheeks and lips a muted iridescent blue-green. Shortly after death, six faint vertical bars appear on the flanks ; the first and second bars are separated LAKE GEORGE HAPLOCHROMIS SPECIES 165 by a greater distance than that between any of the succeeding bars. Dorsal fin with the spinous part sooty, the interspinous membrane generally darkest in colour ; lappets black but with a narrow red streak or spot at the tip. Soft part of dorsal with maroon streaks between the rays. Caudal fin with maroon spots and blotches between the rays and a maroon flush around the margin. Anal dark hyaline (or faintly grey), dusky at the base but with a pinkish border. Pelvics are black. In every major detail the adult, sexually active male coloration of H. macropsoides is identical with that of H. elegans. The sole colour difference, and that a subtle one, lies in the less brilliantly iridescent blue colour of the snout, cheeks and lips ; in H. macropsoides the colour is more blue-green. More vertical stripes (9) appear after death in H. elegans. I have been able to compare live fishes, of both species, in the same advanced state of sexual activity (i.e. ripe-running) and from the same locality. Except for the slight differences noted, I would consider the coloration to be identical. This is a most unusual situation amongst syntopic Haplochromis species and has not been recorded from the species flock of Lake Victoria. Further comment is reserved until p. 229. Adult females have a golden-silver ground coloration shading to white on the belly. The dorsal fin has sooty lappets and an overall dark coloration except for reddish vertical stripes between the spines. The caudal fin is fairly dusky, with dark red streaks between the middle rays, and traces of red on the upper posterior margin. The anal is pale yellow, with a prominent orange spot occupying the position of an ocellus in a male fish. The pelvics are pale yellow. PRESERVED COLORATION. Adult males have a generally dusky ground colour, darkest on the ventral aspects of the flanks and belly, but greyish-silver on the chest. Atleast 5 dark vertical bars are visible on the flanks ; each bar extends from the dorsal to the ventral body outline. The head is dusky overall, with a distinct lachrymal stripe. Paired trans-snout bars are rarely visible, and then but faintly. No other cephalic markings can be detected (cf. H. elegans, p. 148). The branchio- stegal membrane is black. The dorsal fin is dark grey to dusky, sometimes with a narrow black band along the entire base ; the lappets are black, and the soft part of the fin is darkly maculate. Caudal fin greyish (darkest basally) and maculate but sometimes only weakly so. The anal is black or dark grey along its basal half, hyaline to light grey distally. The pelvics are dusky to black, the pectorals hyaline. Adult females are greyish-silver, the ventral half of the body more silver than grey. No distinct cephalic markings are present. The dorsal and caudal fins are greyish hyaline, the margin of the dorsal often dusky. The caudal is immaculate, but has dark, ill-defined streaks between its rays. All other fins are hyaline. Ecotocy. Habitat. Nota great deal of information is available for this species, probably because in the field it is easily confused with H. nigripinnis and H. elegans. The specimens of H. macropsoides at my disposal are either from lake areas close to the fringing papyrus of islands (especially Kankurunga and Akika) or from more exposed areas offshore from these islands. In all localities the substrate is either sand with a thin mud overlay or organic mud. 166 P. H. GREENWOOD Food. Only 6 of the specimens I have examined contained ingested material in the guts ; in all, this comprised dipteran larvae and pupae. As no other material was present (and particularly no phytoplankton or sand grains) the fishes may have been feeding away from the bottom. Breeding. Haplochromis macropsoides is a female mouth brooder. All 7 of the adult, sexually active females examined have the right ovary much larger than the left one. Distribution. Definitely recorded from Lake George and the Kazinga Channel (where it is scarce). The species probably occurs in Lake Edward as well since some specimens in the British Museum (Natural History), misidentified as H. macrops (Blgr.), a Lake Victoria endemic, are probably referable to H. macropsoides. DIAGNOSIS AND AFFINITIES. As noted above (p. 165) the male reproductive coloration of H. macropsotdes is virtually identical with that of H. elegans. How- ever, the species clearly differ in a number of characters, including dentition (more outer teeth and more inner tooth rows in H. macropsoides) and the larger eye and longer upper and lower jaws of H. macropsoides. From H. nigripinnis, H. macropsoides is distinguished by its deeper caudal peduncle (length/depth ratio 1-o-1-5, modal range I-2—1°3, cf. 1:3—1-8, modal range 1-5-1'7 for H. nigripinnis), broader and more numerous inner tooth rows (3 in both jaws, cf. I or 2) and in having the anterior opening to the nasal laterosensory canal much smaller than the nostril (equal to it in H. nigripinnis). The species also differ in male coloration and, apparently, in their feeding habits (the intestine of H. nigripinnis is much longer and more coiled than that of H. macropsoides, and the former species is known to be a specialized phytoplankton eater). From H. orvegosoma, another large-eyed Lake George species, H. macropsoides differs in body form (depth 30-3—34:3, M = 32:1 per cent of standard length in H. oregosoma, cf. 34:9-41:0, M = 36°5 per cent ; also the caudal peduncle is more slender in H. oregosoma, namely 1:4—2-:0, modal range 1-6—1-7 times longer than deep, cf. I-0-1-5 modal range 1-2~-1-3), in dentition (2-3 inner tooth rows in H. macropsoides cf. 1 in H. ovegosoma), in neurocranial shape and in adult male coloration. At first glance, H. macropsoides resembles H. macrops (Blgr.) of Lake Victoria. Closer inspection shows that the species differ in several characters. For example, the interorbital width is less in H. macropsoides (20:5—25:0, M = 23-1 per cent of head, cf. 26-6—32-2, M = 29:7 per cent in H. macrops), the eye is larger (33-3—39°I, M = 36-0 per cent of head, cf. 28-6—35:4, M = 33-0 per cent in H. macrops), the lower jaw is longer (39:2—45:4, M = 41-0 per cent head, cf. 38-0—42°5, M = 39°5) and the outer jaw teeth are fewer (42-60, M = 52, cf. 40-66, M = 60 in H. macrops). Male breeding coloration differs (cf. p. 164 above with p. 237 in Greenwood 1960). The supposed occurrence of H. macrops in Lakes Edward and George is discussed on p. 232. Haplochromis velifer Trewavas of Lake Nabugabo resembles H. macropsotdes in several respects. However, the species may be differentiated by the shallower preorbital of H. macropsoides (12-0-15:2, M = 13°5 per cent head, cf. 13-8—18°5, M = 16-3 per cent), the shorter snout (25:0—30°5, M = 27-6 per cent head, cf. LAKE GEORGE HAPLOCHROMIS SPECIES 167 29:1-33°4, M = 31-3 per cent for H. velifer) and larger eye (33:-3-39'1, M = 36-0 per cent head, cf. 26-3—33:4, M = 30-6 per cent); male coloration is also quite different in the two species (compare p. 164 above with pp. 321-322 in Greenwood, 1965b). STUDY MATERIAL Register number BMNH Locality : Lake George 1972.0.2 : 718 (Holotype) Kankurunga Island 1972.6.2 : 723-728 (Paratypes) Small island north of Kankurunga Island 1972.0.2 : 734-738 (Paratypes) East side of-Akika Island 1972.0.2 : 739-744 (Paratypes) East side of Akika Island 1972.0.2 : 746-752 (Paratypes) East side of Akika Island 1972.6.2 : 753 (Paratype) Kankurunga Island 1972.0.2 : 754 (Paratype) Kankurunga Island 1972.6.2 : 802—803 (Paratypes) Kankurunga Island 1933-2-23 : 354 (Paratype) Worthington collection 1972.6.2 : 719-722 Small island north of Kankurunga Island 1972.6.2 : 729-733 Small island north of Kankurunga Island 1972.0.2 : 755-750 Kankurunga Island 1972.60.2 : 757-705 East side of Akika Island 1972.6.2 : 766-772 Small island north of Kankurunga Island 1972.0.2 : 773-777 Small island north of Kankurunga Island 1972.6.2 : 778-783 Small island north of Kankurunga Island Haplochromis limax Trewavas, 1933 (Text-figs. 13 & 14) H. elegans (part) Trewavas, 1933, J. Linn. Soc. (Zool.), 38: 332 (1 specimen, BMNH reg. no. 1933.2.23 : 395, collected by Worthington from Lake George [no other locality data given)). ? H. nubilus (part) : Trewavas, 1933, op. cit. (2 specimens BMNH reg. nos. 1933.2.23 : 301-302 from Lake George, are tentatively referred to H. limax). Hototyre. A male 80mm standard length (BMNH reg. no. 1933.2.23 : 243) from Lake Edward. DESCRIPTION. Based on 22 specimens (excluding the holotype), 61-0—84-0 mm standard length. Depth of body 35-4—40-3 (M = 37-9) per cent of standard length, length of head 31-°8—34:4 (M = 33-0) per cent. Dorsal head profile straight (rarely with a slight concavity or a slight convexity), sloping fairly steeply at an angle of ca 40°—45° with the horizontal. Preorbital depth 13-6-18-2 (M = 15-4) per cent of head, least interorbital width 23°3-30°5 (M = 26-0) per cent. Snout 26-5-31-8 (M = 29-0) per cent, 0-8—0-9 (rarely I-0) times broader than long ; eye diameter 28-2-34:1 (M = 31-6) per cent, cheek depth 21-1-26:2 (M = 23:9) percent. Caudal peduncle 13-6-18-1 (M = 15:3) per cent of standard length, 1-0-1-4 (mode 1-1) times as long as deep. Mouth slightly oblique, lips a little thickened. Length of upper jaw 29:2-306-0 (M = 33-1) per cent of head (showing slight positive allometry with standard 168 P. H. GREENWOOD Ss SF ee i eS Fic. 13. Haplochromis imax. Lake George specimen ; a male. length), length of lower jaw 36-0-40-9 (M = 38-3) per cent, 1-3-1-9 (modal range I-4-1-6) times as long as broad. Posterior tip of maxilla reaching or almost reaching a vertical through the anterior margin of the eye. Intestine long, ca 1} to 13 times total length, and much coiled. Gill rakers. Lower 1 or 2 rakers reduced, the remainder relatively slender or with 1 or 2 lower rakers short and stout. (One individual has no reduced rakers, but in this fish the total count is only 7 rakers.) There are 7 (rare) to 10 (mode 9) rakers on the lower part of the first gill arch. The pseudorakers are well developed and are directed medially so that they overlie the true gill rakers of the inner row. Scales. Ctenoid; lateral line with 29 (f.1), 30 (f.9), 31 (f.8) or 32 (f.4) scales ; cheek with 2 or 3 (mode) rows. Five to 7 (rarely), mode 54, scales between the upper lateral line and the dorsal fin origin, 5-64 (mode 6) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.4), 15 (f.15) or 16 (f.3) spinous and 8 (f.1), 9 (f.12) or 10 (f.9) branched rays. Anal with 3 spines and 8 (f.1), 9 (f.18) or 10 (f.3) rays. Caudal subtruncate, scaled on its basal half. Pelvics with the first ray produced. Pectoral 26-4—31:0 (M = 29-3) per cent of standard length, 80-0-98-0 (M = 89-1) per cent of head. Teeth. Although basically the form of outer vow jaw teeth is that of an obliquely cuspidate bicuspid, there is some individual variability, especially in the upper jaw (text-fig. 14). This variability concerns the angle of the cutting edge to the major cusp. In all specimens examined, this edge is most acute in teeth situated postero- laterally on the premaxilla; teeth more anteriorly placed sometimes have the cusp so obliquely truncate that the cutting edge is almost horizontal. The modal LAKE GEORGE HAPLOCHROMIS SPECIES 169 condition, however, is one where the edge is at an angle of about 60° with the vertical. It may be noted that, in this respect, the Lake George fishes differ from the holotype whose teeth are of the more acute type. LY Fic. 14. H. limax. Dentary teeth (left), anterolateral in position. Viewed from a slightly Sa iee position. Scale = 0-5 mm. In the lower jaw, teeth tend to be more uniform with regard to cusp shape, and are like the modal upper jaw teeth already described. The minor cusp, in teeth of both jaws, is very small irrespective of major cusp shape. Posterior premaxillary teeth (usually the last 1-4 of the series) are generally unicuspid or are of the general- ized bicuspid type, that is, with an acutely pointed major cusp. There are 30-54 (M = 46) teeth in the outer premaxillary row. Teeth of the inner series are invariably tricuspid and compressed. There are 4 or 5 (rarely 3) rows anteriorly in the premaxilla, and 3 or 4 (mode) anteriorly in the dentary ; laterally and posteriorly the number of rows, in both jaws, decreases to I. A very distinct interspace separates the outermost row of the inner series from the outer row. OstEoLoGy. The neurocranium is of a generalized Haplochromis type but with the preorbital region more noticeably decurved than in H. macropsoides and H. elegans. Also, when compared with these fishes, the premaxilla of H. limax is more robust ; the dentary, however, is similar in all three species. Comparison of H. limax syncranium with that in Lake Victoria species of similar feeding habits (i.e. scraping epilithic and epiphytic algae) shows that H. limax is more like H. obliquidens Hildg. and H. lividus Greenwood than H. nigricans (Blgr.). In the latter species the preorbital face of the skull is more strongly decurved and the dentary is deeper and more robust. The lower pharyngeal bone in H. limax is moderately stout ; the dentigerous area is ca 1-2 times broader than long. The pharyngeal teeth are fine, compressed and cuspidate, and are arranged in ca 28-30 rows. Vertebral counts in the 6 specimens radiographed are: 28 (f.5) and 2g (f.r1), comprising 12 (f.1) or 13 (f.5) abdominal and 15 (f.4) or 16 (f.2) caudal elements. COLORATION IN LIFE. Adult males: ground colour greyish, with a faint overlay of lime on the caudal peduncle and ventrally on the flanks as far forward as the anal fin. Laterally on the flanks and ventrally on the belly there is a scarlet flush, the intensity and area of which vary with sexual state. In quiescent fishes the flush is the colour of dried blood but it is bright scarlet in sexually active individuals. A similarly coloured flush is developed on the operculum and cheek. The dorsal fin is dark hyaline with deep scarlet streaks between the spines ; between the branched rays the streaks are more precisely demarcated as scarlet 170 P. H. GREENWOOD lines. Anal fin is hyaline with a pink flush ; the ocelli are small and yolk-yellow in colour. The pelvic fins are dusky overall. Adult females have a golden-grey ground colour, shading through silver to white on the lower flanks, belly and chest. The dorsal fin is hyaline but is somewhat dusky along its base. Caudal greyish-yellow, the yellow predominating basally. Anal fin also pale grey-yellow, with small, deep yellow spots in the position of ocelli in males. Pelvic and pectoral fins are hyaline. PRESERVED COLORATION. Adult males: the ground coloration is silvery grey. The flanks and caudal peduncle are crossed by up to seven faint vertical bars, none of which extends ventrally below the level of the pectoral fin insertion. In some individuals a faint midlateral band is visible on the caudal peduncle, and extending forward to about a vertical through the origin of the soft dorsal fin. The chest, in some fishes, is sooty ; the branchiostegal membrane, in all, is black. A pair of parallel bars (of variable intensity) cross the snout ; a faint transverse bar extends across the head immediately behind the orbit, but in most specimens only that part of the bar immediately above the orbit is at all intense and discrete. The dorsal fin is always dark, sometimes sooty, sometimes almost solid black between the rays ; the lappets are black. The caudal fin is greyish, becoming yellow on its ventral third ; in a few specimens there are concentrations of melanophores between the middle rays. The anal is greyish basally, yellowish distally. The pelvics are dusky to black on the outer (i.e. anterior) half, but yellowish elsewhere. Adult females have a coloration similar to that of males, but the ground coloration is somewhat lighter and the pelvic fins are hyaline. Ecotocy. Habitat. The distribution of H. limax is closely correlated with the presence of emergent rooted vegetation, or of other places suitable for the growth of aufwuchs. Haplochromis limax has never been recorded far from the shore line, but the substrata over which it occurs are varied. Food. Aufwuchs, its associated microfauna and macerated phanerogam tissue are the commonest types of ingested matter recorded from the gut. Little of the higher plant tissue is digested ; its occurrence in the gut is probably accidental and associated with the plant-scraping feeding habits of the species. Since sand grains and other inorganic bottom material are sometimes found in the gut, it is presumed that H. limax also feeds by scraping suitable food items from the lake bottom. Breeding. Haplochromis limax is a female mouth brooder. Of the 6 adult females examined, 5 have the right ovary much larger than the left and 1 has the ovaries equally developed. One of the 2 smallest fishes available (both 61-0 mm standard length) is a juvenile, the other is a male with indications of early sexual development. At a standard length of 64 mm, fishes of both sexes are adult. Distribution. Lakes Edward and George. The absence of this species from samples made in apparently suitable areas of the Kazinga Channel is noteworthy and inexplicable (see Appendix II). DIAGNOSIS AND AFFINITIES. No other Haplochromis species in Lake George shows the dental characteristics of H. imax; the male coloration is also highly diagnostic. In Lake Edward, on the other hand, there are two species, H. serridens LAKE GEORGE HAPLOCHROMIS SPECIES 171 Regan and H. fuscus Regan, both with multiseriate inner tooth rows, and obliquely cuspidate outer teeth. Haplochromis limax is distinguished from H. serridens by its straighter dorsal head profile (distinctly curved in H. serridens), the fewer rows of inner teeth anteriorly in the jaws (3-5, cf. 5-8 in H. serridens) and by the presence of a distinct space between the outer tooth rows and the inner series of teeth. From H. fuscus, H. limax is distinguished primarily by the outer teeth having a broader and more obliquely truncate cusp, by the smaller size of the minor cusp on these teeth and by having a truncate (as opposed to rounded) caudal fin. In addition, preserved male H. fuscus are uniformly dark (nearly black) whereas H. limax males are silvery grey. With so few specimens of H. fuscus and H. serridens available for comparison with H. limax it is impossible to evaluate the apparent interspecific differences in some morphometric characters. Data on live colours are not available for H. serridens or H. fuscus. From the little information available, it seems reasonable to consider the three species closely related, with H. serridens the most specialized (at least in its oral dentition). Trewavas (1933) noted similarities between H. limax and H. vicarius. Certainly the outer teeth in many of the H. vicarius specimens available do resemble the H. limax type. But, they are equally like those of H. fuscus (see above). It will be necessary to examine further samples of H. vicarius, and get information on live male coloration, before more definite conclusions can be reached on this possible interspecific relationship. I have compared H. limax with those Lake Victoria species having a similar diet and dental specializations (viz. H. lividus, H. nigricans and H. obliquidens). All three species can be distinguished from H. limax by various characters or character combinations. Tooth form and dental] pattern in H. limax is most like that of H. lividus, but it is by no means identical. The teeth of H. lividus are more slender, their crowns are relatively less expanded, have curved not straight vertical margins, and are more movably implanted. On these characters, H. lividus would seem more special- ized than H. limax. The subequally, or almost subequally, bicuspid teeth of H. nigricans, coupled with the strongly decurved preorbital skull profile, and the relatively massive dentary of this species, all suggest that it belongs to a different lineage from that of the other Lake Victoria algal grazers. These same characters also serve to distin- guish H. nigricans from H. limax. The extreme modification of the teeth in H. obliquidens (see Greenwood 1956b) immediately distinguishes this species from H. limax but does not necessarily rule out a fairly close relationship between the species. The teeth in H. obliquidens seem to be the ultimate expression of a specialization already apparent in H. lividus and H. limax (see Greenwood, of. cit., and above). A fourth Lake Victoria species, H. nuchisquamulatus (Hildg.), has feeding habits similar to those discussed above. However, its teeth retain the basic, unequally bicuspid crown, and are not closely like those of H. limax (see Greenwood, of. cit., and above). 10 172 P. H. GREENWOOD Except for H. nuchisquamulatus, where it is not known, the breeding coloration of these species is clearly different. Finally, comparison should be made with H. annectidens Trewavas of Lake Nabugabo. This species has about the same degree of resemblance to H. limax as does H. lividus. The same can probably be said of H. astatodon Regan of Lake Kivu, but far less is known about intraspecific morphological and dental variability in this species. In brief, H. lividus, H. limax, H. astatodon and, despite its highly specialized teeth, H. obliquidens could well be members of a phyletic lineage. STUDY MATERIAL Registered number BMNH Locality 1933.2.23 : 243 (Holotype) Lake Edward (collected by Worthington) 1933.2.23 : 395 Lake George (collected by Worthington) 1972.6.2 : 112-118 Lake George, various localities 1972.0.2 : 119-123 Lake George, various localities 1972.6.2 : 124 Lake George, Akika Island 1972.6.2 : 125 Lake George, papyrus edge 1972.6.2 : 126-128 Lake George, Kankurunga Island 1972.6.2 : 129-135 Lake George, Kankurunga Island 1972.6.2 : 136-139 Lake George, Kankurunga Island 1972.6.2 : 140 Lake George, no locality 1972.6.2 : 808 (Figured specimen) Lake George, Kankurunga Island Haplochromis mylodon sp. nov. (Text-figs. 15 & 16) Haplochromis ishmaeli (non Boulenger) : Trewavas, 1933, J. Linn. Soc. (Zool.), 38: 334 (both specimens identified as H. ishmaeli are from Lake Edward). Ho.totyre. A male, 85:0 mm standard length, BMNH reg. no. 1972.6.2 : 656. The trivial name, from the Greek, refers to the mill-like crushing dentition of the pharyngeal bones. DESCRIPTION. Based on 21 specimens (including the holotype), 68-0-115-0 mm standard length. Depth of body 36-0-40°5 (M = 38-6) per cent of standard length, length of head 31-:0-35°8 (M = 33-1) per cent. Dorsal head profile with some size correlated variation in outline, being more decurved in larger individuals and almost straight in smaller fishes ; sloping at an angle of 40°-45° with the horizontal at all sizes. Preorbital depth 13-3—20-8 (M = 15-9) per cent of head, least interorbital width 241-28-6 (M = 26-6) per cent. Snout length 28-0-32:2 (M = 30-1) per cent of head, 0-8-o-9 (rarely 1-0) times its breadth, eye diameter 26-0-34:0 (M = 29°8) per cent, cheek depth 20-7-26-3 (M = 22-8) per cent. Caudal peduncle 14-7-18-4 (M = 17-9) per cent of standard length, 1-1-1-5 (mode 1-3) times as long as deep. Mouth horizontal, lips not thickened. Length of lower jaw 35:2—40°5 (M = 37-6) per cent of head, 1-2-1-8 (modal range 1-5—1-6) times as long as broad. Posterior LAKE GEORGE HAPLOCHROMIS SPECIES 173 tip of premaxilla reaching a vertical through the anterior margin of the orbit or a little further posteriorly. Fic. 15. Haplochromis mylodon. Holotype. Gill rakers short and stout, the lower I or 2 reduced ; 7-9 (mode 7) rakers on the lower part of the first gill arch. The median row of pseudorakers on the first arch is well developed but individual pseudorakers are low. Scales. Ctenoid ; lateral line with 31 (f.11), 32 (f.9) or 34 (f.1) scales, cheek with 3 (rarely 2) rows. Five to 7 (mode 53) scales between the upper lateral line and the dorsal fin origin, 6-8 (mode 7) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.2), 15 (f.10), 16 (f.8) or 17 (f.1) spinous and 8 (f.5), 9 (f.13) or I0 (f.3) branched rays. Caudal subtruncate, scaled on its basal half. Pectoral 27-6-33°3 (M = 29°'5) per cent of standard length, 82-0—96-5 (M = 84:1) per cent of head. Pelvics with the first ray slightly produced. Teeth. Except for 1 to 3 unicuspids posteriorly in the upper jaw of most fishes, the outer teeth in both jaws are stout and unequally bicuspid ; the major cusp is almost equilateral in outline, moderately protracted and barely incurved. There are 32-46 (M = 40) teeth in the outer premaxillary row. The zuner teeth are small and tricuspid, and are arranged in I or 2 rows in the upper jaw, and a single (rarely double) row in the lower one. OstEoLoGy. The neurocranium of H. mylodon is virtually identical with that of H. ishmaeli (or H. pharyngomylus) of Lake Victoria (see Greenwood 1960). The shape and size of the facet for the upper pharyngeal bones is strictly comparable in all three species, as is the relative contribution to this facet of the basioccipital and parasphenoid bones. The lower pharyngeal bone is a massive structure (text-fig. 16). Compared with this bone in H. ishmaeli and H. pharyngomylus, that of H. mylodon is slightly less massive. The difference is not nearly so marked, however, as that between the bone 174 P. H. GREENWOOD in Lake Victoria and Lake George populations of Astatoreochromis alluaudi (see Greenwood 1959a, 1965b). The dentigerous area of the lower pharyngeal bone in H. mylodon is slightly smaller than in H. ishmaeli or H. pharyngomylus. Except in the outer rows, and in the posterolateral part of the toothed area, the lower pharyngeal teeth are all massive, stout and molariform (at least in the size range of specimens examined). Even the non-molariform teeth are stout, and are but weakly cuspidate. The extent of ‘ molarization’ in H. mylodon is thus comparable with that found in H. ishmaeli and H. pharyngomylus. Vertebral counts in the 8 specimens radiographed are : 28 (f.1), 29 (f.6) or 30 (f.1), comprising 13 (f.8) abdominal and 15 (f.1), 16 (f.6) or 17 (f.1) caudal centra. Fic. 16. H. mylodon. Lower pharyngeal bone. (a) In occlusal view. (b) In lateral view. From a specimen 110 mm standard length. Scale = 1-0 mm. COLORATION IN LIFE. Sexually active males have a blue-grey ground colour with an iridescent turquoise sheen that is particularly concentrated around the margin of flank scales; the belly is silvery grey, the chest and branchiostegal membrane are charcoal-grey. The head is blue-grey but the interorbital region is iridescent blue and is crossed by 2 dark bars. A prominent and dark lachrymal stripe continues to above the eye, where it expands to form a dark blotch. The dorsal fin is greyish- hyaline, the lappets are dusky, the soft dorsal has an orange-red to red margin, and there are dark red streaks between the rays. The caudal is dark over its proximal third, the remainder being rather dusky but with a red (crimson-lake) margin ; this red colour tends to extend forward onto the lower half of the fin, resulting in a dusky-pink coloration. The anal varies from dark hyaline to dusky pink on the spinous part (the lappets are black) ; the soft part is dark hyaline proximally, crimson to pink distally (the darker shades found nearest the middle of the fin). The anal ocelli are deep yellow. The pelvic fins are uniformly dusky. | | LAKE GEORGE HAPLOCHROMIS SPECIES 175 In general, all colours are more intense in the sexually more active individuals. Sexually quiescent and starting males have a light brassy ground colour with a pale-lime overlay, and are white on the chest and belly. The dorsal fin is dark hyaline, the soft part the darker. The caudal is yellowish basally, hyaline on the distal half, and has a bright orange-red margin that widens on the ventral half of the fin. The anal is pinkish, the pelvics dusky. Adult females are silvery, shading to white on the belly. The dorsal fin is hyaline, as is the caudal which, however, has a yellowish flush on the ventral half and a bright scarlet posterior margin. The distal part of the ventral half of the caudal is also light scarlet. The anal fin is yellowish with a pinkish to scarlet overlay that is particularly intense along the margin of its soft part ; two orange spots occur in the position of the ocelli in males. The pelvic fins are hyaline or faintly yellow. PRESERVED COLORATION. Adult males are brown above the midlateral line, shading to dark brown (almost bitter chocolate) below. Up to 8 rather narrow and fairly faint vertical bars are visible across the lighter part of the flank in some specimens ; ventrally the bars merge with the darker ventral flank coloration. The chest is dusky silver, the branchiostegal membrane dusky below the operculum but lighter between the jaws. The head, except for the operculum, is dark yellow-brown ; the operculum is even darker (i.e. it is comparable with the ventral flanks). There 3s a well-defined and broad lachrymal stripe, and two narrow transverse bars across the snout ; a broad, rather ill-defined band crosses the head behind the posterodorsal margin of the orbits, while another broad but more diffuse band or blotch extends across the nape immediately before the dorsal fin origin. The dorsal fin is dark grey, with short black blotches or streaks between the spines (at least basally). The soft part of this fin is darkly maculate. The caudal is greyish, becoming darker over its basal half. Proximally the anal is dark grey to dusky, the dark area becoming more extensive posteriorly and may occupy as much as the basal half of the fin. The rest of the fin is light yellowish-brown. The pelvics are dusky to black, the pectorals are hyaline. Adult females are yellowish-brown to light greyish-brown dorsally, shading to silvery yellow on the belly and lower flanks. The dorsal and caudal fins are greyish, the anal, pelvics and pectorals are yellowish-hyaline. Ecotocy. Habitat. Haplochromis mylodon occurs near the shoreline over mud and mud-sand substrata. A few specimens have been caught in more open-water localities over a sandy bottom. Food. Within the size range sampled, the diet of H. mylodon seems to consist mainly of gastropods, particularly Melanoides tuberculata. Chironomid larvae are also eaten. Breeding. Haplochromis mylodon is a female mouth brooder. Individuals, of both sexes, less than 75 mm standard length are immature, and a few larger fishes (80 mm standard length) also show no signs of gonadial activity. Females may reach a larger size than the males; the largest male recorded is 90 mm standard length, whereas the largest female is 115 mm standard length. Distribution. Lakes Edward and George and the Kazinga Channel. 176 P. H. GREENWOOD DIAGNOSIS AND AFFINITIES. Amongst the Haplochromis species of Lake George (and, apparently, also of Lake Edward), H. mylodon is immediately recognizable by its massive pharyngeal bones and dentition. On the basis of purely anatomical characters, Trewavas (1933) very reasonably identified Lake Edward specimens of H. mylodon as H. ishmaeli, a species otherwise known only from Lake Victoria. Certainly on such characters it is difficult to dis- tinguish between the two species. However, in life the coloration of adult, sexually mature males is very different (compare p. 174 above with p. 277 in Greenwood 1960). When Lake George specimens are compared with H. ishmaeli from Lake Victoria (see Greenwood 1960) there are, in fact, some slight anatomical differences as well. For example, the cheek is a little shallower in H. mylodon, there are fewer teeth in the outer premaxillary tooth row than in H. zshmaeli and the chest scales are smaller. A shallower cheek but larger chest scales and a longer pectoral fin distinguish H. mylodon from H. pharyngomylus, the other Lake Victoria species with a similar crushing pharyngeal dentition. Once again, adult male coloration provides a ready interspecific difference when live fishes are compared. But, in this instance the coloration is rather less different than in the case of H. ishmaeli and H. mylodon. In brief, H. mylodon, H. ishmaeli and H. pharyngomylus are alike in nearly all morphometric characters and in most anatomical ones as well, but each species has a characteristic male coloration. It is chiefly because of the differences in coloration that I place H. mylodon in a distinct species (and do not include it with H. pharyngo- mylus). The importance of male coloration in cichlid courtship and_ species recognition is such that it would be biologically unsound to consider H. mylodon as anything other than specifically distinct. Poll (1959) described three species (one from Lake Edward and two from nearby localities) with enlarged pharyngeal bones and molariform pharyngeal teeth. Of these species, one, H. malacophagus (from Lake Kibuga, ca 50 km south of Lake Edward), need not be considered in detail. Its pharyngeal bones and dentition are but slightly enlarged and there are other characters which differentiate it from H. mylodon. The second species, H. placodus (from the Molindi River, near Lake Kibuga), has a greatly enlarged lower pharyngeal bone and an almost completely molariform pharyngeal dentition ; in both characters it is comparable with H. mylodon. In overall appearance, too, H. placodus is rather like H. mylodon. The holotype and only specimen (ror mm standard length) differs from H. mylodon in its larger pectoral and nuchal scales (44 between upper lateral line and dorsal origin, 5 between pectoral and pelvic fin bases, cf. 5—7 [mode 54] and 6-8 [mode 7] in H. mylodon), in having a much smaller eye (22-2 per cent of head, cf. 26-34:0, mean = 29°8 per cent) and a shorter pectoral fin (72:3 per cent head length, cf. 82-0—96:5, mean 84:1 per cent). When more specimens of H. placodus are available, its relationships with H. mylodon can be reviewed more critically. But, unless H. placodus holotype is an aberrant individual, it seems unlikely that the two species will prove to be conspecific. The third species, H. pharyngalis, is from the western shore of Lake Edward, at Bugazia. In two of the three syntypes, the pharyngeal mill exhibits a degree of development almost comparable with that of H. mylodon and H. placodus. In the third specimen, however, the bones are not greatly enlarged and only the median | | } LAKE GEORGE HAPLOCHROMIS SPECIES 177 rows of the lower pharyngeal teeth are molariform. Body form in H. pharyngalis is unlike that of H. mylodon, being elongate and slender. The thickened lips and slightly shorter lower jaw of H. pharyngalis give to the face a most distinctive appearance. Several other characters serve to distinguish this species from H. mylodon (and the other species considered here). Outstanding among these diag- nostic features are the minute nuchal and chest scales. I count, in the three H. pharyngalis syntypes, 8 or g scales between the upper lateral line and the dorsal fin origin, and about the same number between the pelvic and pectoral fin bases ; furthermore, scales lower on the chest are so small and thin that, at first sight, this area seems naked. Other diagnostic characters are the low number of gill rakers (5 or 6), the longer snout (34:5 and 35:0 per cent head, cf. 28-0-32-2, mean 30-1 per cent in H. mylodon) and deeper cheek (26-7 and 29:3 per cent of head, cf. 20:7-26°3 per cent, mean = 22°8 per cent in H. mylodon). Because the two H. pharyngalis syntypes measured (82:5 and 88-0 mm standard length) are within the size range of the H. mylodon sample, these morphometric differences are unlikely to be the results of allometric growth. (The third syntype is rather distorted and was, therefore, not measured.) There do not seem to be any grounds for assuming a close or even distant relation- ship between H. pharyngalis and H. mylodon (or, indeed, between that species and H. placodus or H. malacophagus). The peculiarly small nuchal and thoracic squama- tion of H. pharyngalis is, however, characteristic of a species recently discovered in Lake George (see p. 209). The two species also have a similar body form and physiognomy, but the new Lake George species does not have a hypertrophied pharyngeal mill ; their possible relationship is considered below (p. 213). STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 656 (Holotype) Small island north of Kankurunga Island 1972.6.2 : 655 (Paratype) N.E. corner of the lake 1972.6.2 : 657 (Paratype) Over sandy shallows 1972.6.2 : 661-667 (Paratypes) Various localities 1972.6.2 : 668-676 (Paratypes) Various localities 1972.6.2 : 799-801 (Paratypes) Small island north of Kankurunga Island 1972.6.2 : 658 Sandy shallows 1972.0.2 : 677-678 Locality unknown 1972.6.2 : 679-680 50-70 m from bush shore 1972.6.2 : 681 Locality unknown 1972.6.2 : 682 Locality unknown Haplochromis angustifrons Blgr., 1914 (Text-figs. 17-19) Synonymy. Trewavas (1933) gives a full synonymy for H. angustifrons, a species which Regan (1921) had previously synonymized with H. schubotzi Blgr. Trewavas’ redescription of the species (and her synonymy for it) was based entirely on Lake Edward specimens in the B.M.(N.H.) collections, which include 5 178 P. H. GREENWOOD of Boulenger’s syntypes (see Boulenger 1914 and 1915). One of the latter specimens was referred to H. schubotzi by Trewavas (op. cit.), and only 3 of the remaining 4 syntypes were included in her redescription of H. angustifrons. I have examined the fourth and neglected specimen, and can confirm its identity as H. angustifrons. The 4 syntypical specimens have the B.M.(N.H.) register numbers, 1914.4.8 : 25-28. Through the courtesy of Dr K. Deckert (Berlin Museum) I was able to examine 48 syntypes of this species (including the specimen figured in Boulenger 1g14 and 1915). It should be noted that there are apparently 53 syntypes in existence, although Boulenger (1914) originally recorded 56 specimens. As far as I can tell without detailed knowledge of H. angustifrons in Lake Edward, all except 2 of the Berlin syntypes can be referred to this species. I do, however, have some reservations about the identity of a few small specimens in this series. The 2 specimens which I do not consider to be H. angustifrons provide something of a puzzle that may only be solved when a large-scale revision of the Lake Edward Haplochromis species is carried out. Both these fishes are from the Berlin Museum lot number 19778. One, a female 71 mm standard length, appears to be of an H. elegans-like species. The other, a female, 86-0mm standard length, I am tentatively referring to a new species described below (p. 188). In his original description of H. angustifrons, Boulenger (1914) mentions some females as having *. . .einem breite, dunklen, braunen Seitenband vom kiemendeckel zur Schwanzflosse,...’. The 86 mm female mentioned above is the only syntype I examined with such a midlateral band. Haplochromis angustifrons females do not exhibit this colour pattern which is, however, a characteristic of the new species to which this syntypical fish is now tentatively referred (see p. 190 below). As lectotype of H. angustifrons I have chosen the figured specimen, a male 91-0 mm standard length, Berlin Museum number 19118, collected by Schubotz from Lake Edward. At least with respect to Lake George populations of H. angustifrons, the lectotype is unusual in being a male of such large size. In Lake George, adult males are generally much smaller than females. Despite this size discrepancy, the lectotype is a modal H. angustifrons in all morphological characters. Trewavas’ (1933) synonymy of H. angustifrons must now be expanded to include : H. elegans (part) Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 333 (1 specimen from Lake George, BMNH reg. no. 1933.2.23 : 394). H. vicarius (part) Trewavas, 1933, op. cit.: 331 (1 of the 2 paratypes from Lake George, [Worthington’s (1932) station 613], BMNH reg. no. 1933.2.23 : 353). H. schubotzi (part) : Trewavas, 1933, op. cit. : 337 (the 5 small specimens from Lake George, BMNH reg. nos. 1933.2.23 : 409-413, from Worthington’s stations 613 and 627). H. nubilus (part) : Trewavas, 1933, op. cit. : 329 (2 specimens, BMNH reg. nos. 1933.2.23 : 287 and 300, from Worthington’s stations 613 and 618, Lake George). H. guiarti (part) : Trewavas, 1933, op. cit. : 339 (1 specimen, BMNH reg. no. 1933.2.23 : 477, collected by Worthington from Lake George but no station number was given). DESCRIPTION. Based on 41 specimens, 40:0--90:0 mm standard length, all from Lake George. Depth of body 34:3-40°5 (M = 36:2) per cent of standard length, length of head 34°5-38°3 (M = 36-4) percent. Dorsal head profile straight or very weakly convex, LAKE GEORGE HAPLOCHROMIS SPECIES 179 sloping at an angle of ca 35°—40° to the horizontal ; premaxillary pedicels prominent and breaking the dorsal head outline to give the fish a very characteristic ‘ Roman nose’ profile. The upper margin of the orbit just enters the line of the head profile. Preorbital depth 12-5-18-9 (M = 16-6) per cent of head, least interorbital width 17°3-24:0 (M = 20°5) per cent, snout length 25-0-34:0 (M = 29:5); all three proportions show slight positive allometry with standard length. The snout varies from slightly broader than long to a little longer than broad (o-8-1-1), but modally is as long as broad. The eye diameter and the cheek depth both show marked allometry with standard length, the former negatively allometric, the latter positively so. Thus, for these measurements two figures are given, first for fishes < 60mm standard length (N = 12), and second for larger fishes (N = 29). Eye 32-3-37-6 (M = 35-1) per cent head, and 27-8—33-4 (M = 31-0) percent ; cheek 15-3—23-7 (M = 21-3) per cent, and 22-7-29:6 (M = 25-7) per cent. Fic. 17. Haplochromis angustifrons. Lake George specimen ; a female. Caudal peduncle 15:6-20:0 (M = 17:5) per cent of standard length, 1-2-1-8 (modal range I-3-1-5) times as long as deep. Mouth slightly oblique, or horizontal ; lips not thickened. Length of upper jaw 30°2-37°5 (M = 34:4) per cent of head, length of lower jaw 38-8—45-8 (M = 42°5) per cent, 1-5-2-3 (modal range 2-0-—2-2) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior part of the eye or even through the pupil. Gill rakers. The lower r or 2 rakers are reduced, the remainder relatively slender and elongate, although the rakers immediately above the reduced ones may be short and stout. There are 7 or 8 (rarely 10) rakers on the lower part of the first gill arch. Pseudorakers are barely developed ; the tissue between the inner and outer rows of true rakers is slightly thickened and thrown into low and barely discrete pro- jections. 180 P. H. GREENWOOD Scales. Ctenoid ; lateral line with 30 (f.13), 31 (f.22) or 32 (f.6) scales, the cheek with 2 (rare)—4 (mode 3) rows. Four and a half to 54 (rarely 63), mode 5, scales between the upper lateral line and the dorsal origin, 5~7 (rarely 8), mode 6, between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.15) or 15 (f.26) spinous and 8 (f.3), 9 (f.29) or ro (f.9) branched rays. Anal with 3 spines and 7 (f.2), 8 (f.29) or 9 (f.10) branched rays. Caudal truncate to very weakly emarginate, scaled on its basal half or a little further posteriorly. Pectoral 27-4—33-7 (M = 30-4) per cent of standard length, 74:0-89-9 (M = 83:1) per cent of head. Pelvics with the first and second rays somewhat produced. Teeth. The outer vow in the upper jaw usually is composed of both unicuspid and bicuspid teeth (text-fig. 18) the latter sometimes showing every gradation from fully and unequally bicuspid to weakly bicuspid (with the minor cusp virtually absent). As far as I can ascertain, this variability is not size correlated. All bicuspids, like the unicuspids, are slender and compressed ; the major cusp is protracted, slightly incurved and has the outline of an isosceles triangle. Posteriorly in the upper jaw, the teeth are always unicuspid, and unicuspids generally predominate posterolaterally as well. Fic. 18. H. angustifyons. Dentary teeth (left), anterolateral in position. Viewed laterally. aoe = 0-5 mm. In general, the outer tooth row of the lower jaw has greater uniformity of tooth type. Unequally bicuspid teeth predominate. There are 44-66 (mean 56) teeth in the outer row of the upper jaw, the number showing a slight positive correlation with size. Tricuspid teeth predominate in the inner tooth series of both jaws, but many individuals have an admixture of tricuspid and weakly tricuspid teeth, or of uni- cuspids and weakly tricuspids. There are I or 2 (rarely 3) rows in both jaws. OstEoLoGy. The neurocranium closely resembles that in H. schubotzi and the new taxon, H. schubotziellus (see p. 190). It represents a somewhat specialized departure from the basic Haplochromis type; the preorbital region is relatively elongate and gently sloping, and there is an overall reduction in neurocranial width. The lower pharyngeal bone is fine, its outline noticeably elongate and narrow (text- fig. Ig), especially in comparison with the pharyngeal bone of other species in the Lake George flock. The dentigerous area is a little longer than broad (ca 1-1 times), the teeth fine, compressed and cuspidate and are arranged in ca 24-28 rows. Some teeth in the median rows are a little stouter than those situated laterally. Vertebral counts for the 10 specimens radiographed are 28 (f.4) or 29 (f.6), com- prising 12 (f.9) or 13 (f.1) abdominal and 16 (f.5) or 17 (f.5) caudal elements. LAKE GEORGE HAPLOCHROMIS SPECIES 181 COLORATION IN LIFE. Adult males: the dorsum of the head and body is an iridescent violet which shades to turquoise on the midflank and greenish-golden on the belly. Chest and lower jaw are sooty, the branchiostegal membrane dusky to black, and the cheeks greenish-turquoise. Cephalic markings are not always visible, but when developed consist of a prominent, saddle-shaped nuchal bar and two parallel stripes across the snout. Hop 9 Og 9 Ntaeayes Oise One ade Fic. 19. H. angustifrons. Lower pharyngeal bone, in occlusal view. Scale = t-omm. The dorsal fin is dark hyaline with a sinuous black band running the entire length of the fin at a level about one-third of the distance between margin and base. Caudal fin dark hyaline, as is the anal which, however, is black basally and along its margin, and may show a faint pink flush; the anal ocelli are orange. The pelvic fins are black. Shortly after death, traces of 7-10 vertical bars may appear on the flanks and caudal peduncle. Adult females are metallic grey dorsally, shading through silver on the flanks to whitish on the belly. All fins are hyaline or faint yellow ; the caudal is densely and clearly maculate, the spots dark grey and very obvious. (Indeed, this feature is diagnostic for the species in Lake George.) PRESERVED COLORATION. Males: the ground colour is dark brown to black, the dark pigment most concentrated on the snout, cheeks, operculum, belly and lower half of the flanks. The lips are usually lighter than the cheeks, and the thoracic region is lighter than the flanks and belly. Traces of up to 6 narrow, fairly close-set vertical bars are often visible on the flanks. The bars are most distinct dorsally because ventrally they merge with the overall dark coloration for that region. Cephalic markings are not always visible, but when present consist of 2 bars across 182 Pp. H. GREENWOOD the snout, a lachrymal stripe, a large posteriorly directed triangular blotch on the posterior interorbital region, and a broad band across the nape. The dorsal, pelvic and anal fins are dusky to black, the dorsal lappets intensely black ; the pigment on the dorsal and anal fins may be concentrated along the fin base. The caudal is lighter than the others, and has its pigment concentrated between the middle rays. Females are silvery brown, some with faint traces of about 6 ill-defined vertical bars on the flanks. All fins are hyaline, but narrow dark streaks occur between the spines of the dorsal fin, and the soft part of that fin is sometimes weakly maculate. The caudal is invariably maculate, and distinctly so even if the spots are rather pale. (As in live fishes, this feature is a diagnostic one.) Ecotocy. Habitat. Haplochromis angustifrons is essentially an offshore species, and is rarely captured near any type of shore. In its habitat it occurs over both mud and sand substrates, but it seems to prefer the latter. Food. Both planktonic and benthic animals are eaten, of which, respectively, chaoborid and chironomid larvae are the dominant food organisms. Breeding. At least in the Lake George populations there is a very marked sexual dimorphism in the adult size attained. Males are noticeably smaller than females, individuals more than 65 mm standard length are rare, and the smallest male fish examined (40 mm standard length) was sexually active. Females, on the other hand, only reach sexual maturity at a length of about 63-65 mm, and attain a maximum adult size of at least go mm standard length. Some males do reach this size (one is known from Lake George, and the holotype, from Lake Edward, is gi mm standard length) but are rare. In addition to this sexually correlated size disparity, there also appears to be a marked imbalance in sex ratio at all times, but especially during daylight hours when males are particularly scarce (about I in 20 adult fishes). This problem of diurnal sex ratio change (with the concomitant problem of male ‘ migration’), and the apparently real predominance of females at all times and all places, is under active research by the I.B.P. team on Lake George. No further comments can be made at this time. Of the 25 sexually active females sampled, 12 have the right ovary considerably larger than the left one, 6 have the left slightly larger and 7 have both ovaries equally developed. Distribution. Lakes Edward and George, and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Haplochromis angustifrons is immediately dis- tinguishable from other Lake George Haplochromis species by its deep body, distinc- tive “ Roman nose ’ profile and, at least in females, by the clearly maculate caudal fin. The distinctive coloration and small adult size of males are further diagnostic features. Haplochromis angustifrons does not appear to be closely related to any other species in Lake George, nor, as far as can be estimated from known collections, to any species in Lake Edward. The dentition, narrow and elongate lower pharyngeal bone and the body form (especially head shape) distinguish this species from, on LAKE GEORGE HAPLOCHROMIS SPECIES 183 the one hand, the H. elegans—macropsoides complex, and on the other hand, from H. schubotzi and related species. These same characters give H. angustifrons a superficial resemblance to H. empodisma Greenwood of Lake Victoria (see Greenwood 1960), and to H. simpsoni Greenwood of Lake Nabugabo (Greenwood 1965b). These two species are, however, distinguished from H. angustifrons by several morphometric and colour differences. STUDY MATERIAL Register number BMNH Locality : Lake George 1933.2.23 : 287 Collected by Worthington, station no. 618 1972.6.2 : 412 Over sandy shallows 1972.6.2 : 414 Over sandy shallows 1972.6.2 : 420-428 Various localities 1972.6.2 : 432-433 Tufmac Bay 1972.0.2 : 434-437 Tufmac Bay 1972.0.2 : 438-440 Tufmac Bay 1972.0.2 : 441-500 Tufmac Bay 1972.6.2 : 512-515 Between Akika and Kankurunga Islands 1972.6.2 : 516-518 Kankurunga Island 1972.6.2 : 519-526 Small island north of Kankurunga Island 1972.6.2 : 531 Small island north of Kankurunga Island 1972.0.2 : 542-548 Small island north of Kankurunga Island 1972.6.2 : 804 (figured specimen) Between Akika and Kankurunga Islands Haplochromis schubotzi Blgr., 1914 (Text-figs. 20 & 21) Haplochromis schubotzi (part) : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 337 (the 2 specimens 107 and 110mm standard length, collected by Worthington. These specimens were not previously registered in the B.M.[N.H.] collection and are now given the reg. nos. 1972.2.24 : I-2). The five small specimens (52-80 mm standard length) from Lake George which Trewavas (op. cit.) mentions in her description of H. schubotzi are now identified as H. angustifrons (see above, p. 178). The type series of H. schubotzi consists of 5 large males, all from Lake Edward (see Boulenger 1914). One of these fishes (reg. no. 1914.4.8 : 18) is in the collections of the British Museum (Nat. Hist.), the others are in the Berlin Museum. Through the courtesy of Dr K. Deckert I have been able to examine these specimens and thus to select a lectotype for the species. Lectotyre. A male, 118-0 mm standard length, collected by Schubotz from Lake Edward (Berlin Museum number 19116). The three paralectotypes from that museum (also ex Lake Edward) have the lot number 22699. DESCRIPTION. Based on 30 specimens from Lake George, 69:0-125:0 mm standard length. 184 P. H. GREENWOOD Depth of body 33-7-39°3 (M = 36-4) per cent of standard length, length of head 32°5-37°8 (M = 34:9) per cent. Dorsal head profile straight or, less commonly, gently curved, sloping at an angle of ca 35°-40° with the horizontal. Fic. 20. Haplochromis schubotzi. Lake George specimen; a juvenile female. Preorbital depth 16-4—20-9 (M = 19:0) per cent of head, least interorbital width 19:2-25:0 (M = 22-4) per cent; snout length 30-4-39:6 (M = 33-4) per cent, its breadth equal to (mode) or slightly greater than its length. Eye diameter 25:5-32-0 (M = 28-4) per cent of head, depth of cheek 20-7~—26-0 (M = 22:8) per cent. A Lake Edward specimen (the lectotype). From Fic. 21. Haplochromis schubotzt. Boulenger, Cat. Afr. Fishes (1915). eee LAKE GEORGE HAPLOCHROMIS SPECIES 185 Caudal peduncle 16:2-19:2 (M = 17-7) per cent of standard length, 1-2-1-7 (modal range 1-3-1'5) times as long as deep. Mouth horizontal, lips slightly thickened. Length of upper jaw 27-0-33°3 (M = 30°5) per cent of head, length of lower jaw 35-8—44:0 (M = 39:2) per cent, I-4-2:0 (modal range 1-6-1-7) times as long as broad. Posterior tip of the maxilla usually reaching a vertical slightly anterior to un orbital margin, but reaching that level in a few fishes. Gill rakers. A characteristic feature of the first gill arch in H. schubotzi is the well- developed papillose area of tissue immediately preceding the first (i.e. lowermost) gill raker. This raker, and usually the next one, is reduced; the others are well developed and range in form from relatively stout to relatively slender. The pseudo- rakers on this arch are especially well developed. There are 7 or 8 (mode) rarely g gill rakers on the lower part of the first arch. Scales. Ctenoid ; lateral line with 31 (f.10), 32 (f.14), 33 (£5) or 34 (f.1) scales ; cheek with 3 rows (rarely 2 or 4), the scales deeply embedded in the skin. Five to 64 (mode 53) scales between the upper lateral line and the dorsal fin origin ; 6 or 7 (deeply embedded) scales between the pectoral and pelvic fin bases. Fins. Dorsal with 13 (f.1), 15 (f.17) or 16 (f.11) spinous and 8 (f.8), 9 (f.19) or to (f.2) branched rays; anal with 3 spines and 8 (f.15) or 9 (f.15) branched rays. Caudal weakly emarginate, scaled on its basal half. Pectoral 25-6-30-0 (M = 27-6) per cent of standard length, 73-0-87-0 (M = 79-4) per cent of head. Pelvics with the first ray produced in adults, proportionately longer in males. Teeth. Tooth form is loosely correlated with body size. In fishes less than 71 mm standard length, the outer teeth are clearly, but unequally bicuspid, the major cusp is produced and isoscelene in outline. In larger individuals the disparity in cusp size is more marked so that the inner cusp is virtually invisible ; the major cusp seems to be even more protracted and slender. At all sizes, the outer teeth are slender and compressed, with the crown slightly incurved. Irrespective of the fish’s size the posterior and some posterolateral teeth in the upper jaw are relatively more slender than the others and are usually unicuspid or very weakly bicuspid. Elsewhere in this jaw (and particularly in larger fishes) there is usually an admixture of clearly bicuspid teeth, weakly bicuspids and, to a lesser extent, unicuspids. Tooth form in the lower jaw is, on the whole, more uniform. There are 46-62 (M = 52) teeth in the outer series of the premaxilla ; the number of teeth does not show any clear-cut correlation with standard length, but the two largest fishes (123 and 125 mm standard length) do have the two highest numbers of teeth recorded (60 and 62 for the fishes respectively). The inner series in most fishes less than 90 mm standard length are composed of slender tricuspids, but in larger individuals there may be an admixture of tricuspids, weakly tricuspids and weakly bicuspids. All inner teeth are slender, and are generally implanted so as to lie horizontally. There are 3 rows (less frequently 2 or 4) in the upper jaw and 2 or 3 in the lower jaw. OstEoLoGy. The neurocranium of H. schubotzi shows many of the characters seen in the skull of H. riponianus (Blgr.) from Lake Victoria (see Greenwood 1960). In other words, it is a slightly specialized derivative of the generalized skull-type 186 P. H. GREENWOOD seen in, for example, H. elegans. The principal differences lie in the more elongate preorbital region of H. schubotzi skull, and in its straighter and less steeply sloping dorsal profile. The lower pharyngeal bone is relatively slender, with its dentigerous area almost equilateral. The teeth are fine, compressed and cuspidate, and are arranged in from 26 to 30 rows. Some fishes have the teeth in the median rows slightly coarser than the others. Vertebral counts in the 6 fishes radiographed are 29 (f.5) or 30 (f.1) comprising 12 (f.2) or 13 (f.4) abdominal and 16 (f.3) or 17 (f.3) caudal elements. COLORATION IN LIFE. Adult males: the dorsum of the head, snout and body, the operculum and the anterolateral aspects of the flanks have a pinkish to orange- red flush ; the remainder of the body is blue-grey except for the sooty chest, and yellow tinge on the upper part of the caudal peduncle. Lips, lower margin of the preoperculum, the lower jaw and the lateral aspects of the snout are bright iridescent blue, or the cheek may be orange-red. Branchiostegal membrane is black but with traces of iridescent blue over its anterior half. Dorsal fin dark but with reddish to orange streaks between the rays and a faint overall reddish-orange flush on the soft part ; the lappets are black. The caudal is hyaline with a faint red tinge between the rays, especially noticeable on the upper half of the fin. Anal bluish to dusky, the ocelli yolk-yellow. Pelvics dusky to black. Immature males are basically silver-grey, with a faint rose flush on the operculum and anterior flanks, and some iridescent blue on the cheek and lips. Dorsal and anal fins are hyaline, the former with reddish streaks between the spines and rays ; caudal hyaline with a faint pink flush on its ventral half. Adult females have a silvery-grey ground colour, shading to white on the belly. The dorsal, caudal and anal fins are hyaline with a faint yellowish to yellowish-grey flush, the pigment being most concentrated basally. The pelvic fins are hyaline. PRESERVED COLORATION. Adult males: ground colour greyish-brown or greyish- silver above the midlateral line, becoming dusky silver ventrally. Lateral aspects of the belly and the entire thoracic region are dusky. The snout, cheeks and most of the opercular region are dusky or at least darker than the dorsum. Cephalic markings are of variable intensity depending on the basic tone of the head coloration. The lachrymal stripe is generally intense ; the snout is crossed by 2 bars, the upper of which is the wider and is often interrupted medially. On either side of the midline behind the level of the orbits is a dark, near-triangular blotch extending ventrally to the upper orbital margin ; an ill-defined black blotch crosses the nape, anterior to the dorsal fin origin. The dorsal and anal fins are dark grey to greyish-brown, the margin of the soft part pale, the lappets of the dorsal black. The caudal is greyish, either darkest over its basal third or almost uniformly grey-brown. The pelvic fins are black or dusky, the pectorals hyaline. Females are greyish to brown over a silvery underlay, silvery on the lower flanks and belly. Some individuals have very faint traces of transverse barring on the upper part of the body. Cephalic markings apparently are not developed save for a LAKE GEORGE HAPLOCHROMIS SPECIES 187 very faint lachrymal blotch. The dorsal and caudal fins are greyish, the soft dorsal sometimes maculate. The anal is hyaline to yellowish, the pelvics and pectorals are hyaline. EcoLocy. Virtually nothing is known about the feeding and breeding habits of H. schubotzt, and little is known of its distribution within Lake George. Most of the specimens described above came from offshore localities, over sand or muddy sand substrates. Some localities are exposed, others relatively protected. Certainly the species is rarely caught in nets set close to a papyrus margin or close to other emergent aquatic plants; nevertheless it-does sometimes occur in such habitats. Apparently the species is absent from the open waters of the centre lake. The few available records of gut contents suggest that H. schubotzi is insectivorous, but the extent of its dependence on this food source requires confirmation. It is still not known whether or not H. schubotzi is a mouth brooder. Fishes, of both sexes, less than 75 mm standard length are immature, as are some larger fishes (up to 80 mm standard length). Of the 8 adult females studied, 5 have the right ovary considerably larger than the left one and 3 have both ovaries equally developed. Distribution. Lakes Edward and George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Probably the species from Lake George most like H. schubotzt is the new taxon H. schubotziellus. This species is described on p. 188, and its relationship with H. schubotzi is discussed on p. 192. Superficially, H. schubotzi also resembles H. mylodon but is readily distinguished from that species by its unmodified pharyngeal bones, narrower interorbital (19-2- 25:0, M = 22-4 per cent head, cf. 24-1-28-6, M = 26-6 per cent for H. mylodon), and by the presence of the extensive papillose area on the lower part of the first gill arch. In life, male breeding coloration is distinctive. Rather less similar in its overall morphology is H. angustifrons, although small individuals of H. schubotzi could be confused with members of that species. Haplo- chromus schubotzi differs from H. angustifrons in having the inner jaw teeth horizontally aligned, in its shorter upper jaw (27-:0-33°3, M = 30°5 per cent head, cf. 30:2-37'5, M = 34-4 per cent head), in possessing a papillose area preceding the first gill raker (see above), and in the failure of the posterior tip of the maxilla to reach the anterior orbital margin. Again, male coloration is diagnostic. Considering species from outside Lakes Edward and George, H. cinereus (Blgr.) of Lake Victoria shares several characteristics with H. schubotzi. As pointed out before (Greenwood 1960), H. cinereus is, in fact, not the generalized species it was once thought to be by many workers. Its dentition and skull are relatively specialized when compared with the generalized Haplochromis type (Greenwood op. cit.). In these particular characters H. cinereus resembles H. schubotzi, as it also does in having well-developed pseudorakers and a papillose area before the lower gill raker (charac- ters not previously recorded for H. cinereus). Haplochromis cinereus differs from H. schubotz principally in having rather more unicuspid teeth in the jaws (at least when equal-sized fishes are compared) and in having the median teeth of the lower pharyngeal bone noticeably enlarged. The possible relationship of these two species will be considered again later in this paper (p. 233). II 188 P. H. GREENWOOD Gross morphology, neurocranial shape and dental characters are also similar in H. schubotzi and two other Lake Victoria species, H. riponianus (Blgr.) and H. saxicola Greenwood (see Greenwood 1960). The slender lower pharyngeal bone (lacking enlarged median teeth) of H. schubotzi is more like that of H. saxicola than that of H. viponianus. As usual, the coloration of adult males is different and there are morphometric characters distinguishing H. schubotz from the two Lake Victoria species. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 683-685 Bay at the north end of Akika Island 1972.6.2 : 687-692 Bay at the north end of Akika Island 1972.6.2 : 697-698 Bay at the north end of Akika Island 1972.6.2 : 699 Bay at the north end of Akika Island 1972.6.2 : 700-701 Bay at the north end of Akika Island 1972.6.2 : 702-703 Off east shore of Akika Island 1972.6.2 : 704-705 Off papyrus edge of Akika Island 1972.0.2 : 706-707 Locality unknown 1972.6.2 : 708 Locality unknown 1972.6.2 : 709 1m offshore from papyrus edge 1972.6.2 : 710 Over sandy shoal 1972.6.2 : 711-712 Locality unknown 1972.6.2 : 713 Northeast of Kankurunga Island 1972.0.2 : 714-715 Off Kankurunga Island 1972.6.2 : 716 Tufmac Bay 1972.6.2 : 717 Tufmac Bay 1972.6.2 : 807 (figured specimen) Tufmac Bay Haplochromis schubotziellus sp. nov. (Text-figs. 22 & 23) ? H. angustifrons (part) Boulenger, 1914, in Wiss. Evgebn. Deuts. Zentral-Afrika Exped., 1907-1908, Zool. 3: 256-257 (1 of the paralectotypes in the Berlin Museum [no. 19778], a female 86 mm standard length from Lake Edward ; see also under synonymy of H. angusti- frons on p. 178). Hototyre. A female, 76:0 mm standard length, BMNH reg. no. 1972.6.2 : 351. The trivial name (a diminutive) is given because, in the field, specimens of this species are often confused with small specimens of H. schubotz. DESCRIPTION. Based on 28 specimens (including the holotype) 45-0-79:0 mm standard length, all from Lake George. The syntype of H. angustifrons is not in- cluded because its identification as H. schubotztellus is tentative, and it comes from Lake Edward. Depth of body 33-3-38-0 (M = 35:7) per cent of standard length, length of head 32°7-37°3 (M = 35°1) per cent. Dorsal head profile straight or gently curved, sloping at an angle of ca 35°—40° with the horizontal. ae LAKE GEORGE HAPLOCHROMIS SPECIES 189 Preorbital depth 12-5-18-5 (M = 15-6) per cent of head, showing slight positive allometry with standard length ; least interorbital width 17-9-20-8 (M = 19'5) per cent, length of snout 25-0-33-3 (M = 29-2) per cent, 0-8-o-9 times its breadth. Eye diameter 28-8—34-7 (M = 31-8) per cent of head, cheek depth 17-6-24-1 (M = 22:0) per cent, showing very slight positive allometry. Caudal peduncle 13:3-18-5 (M = 16:5) per’cent of standard length, 1-1-1-6 (modal range 1-3-1-4) times as long as deep. Fic. 22. Haplochromis schubotziellus. Holotype. Mouth horizontal, lips very slightly thickened. Length of upper jaw, showing slight positive allometry with standard length, 28-2-37-0 (M = 34:2) per cent of head, length of lower jaw 36-7~45-0 (M = 41-6) per cent, I-4-2-3 (modal range I-7-1-9) times as long as broad. Posterior tip of the maxilla reaching a vertical through the anterior part of the orbit or a little further posteriorly. Gill rakers. The tissue immediately anterior to the first gill arch is but slightly thickened, and is thrown into low, rather ill-defined folds (thus contrasting with H. schubotzi where this area is distinctly papillose and markedly pachydermatous). The first raker, and sometimes the 1 or 2 succeeding it, is reduced ; the other rakers are short and relatively stout, with 7-9 (mode 8) on the lower part of the first arch. Pseudorakers are present but are small and sometimes ill defined. Scales. Ctenoid; lateral line with 30 (f.1), 31 (f.9), 32 (f.15) or 33 (f.1) scales, cheek with 3 (rarely 2 or 4), rows, the scales not deeply embedded. Five to 6 (no distinct mode) scales between the upper lateral line and dorsal origin, 6 (mode) or 7, rarely 5 between the pectoral and pelvic fin bases, the scales not deeply embedded. Fins. Dorsal with 14 (f.1), 15 (f.17) or 16 (f.10) spinous and 8 (£2), 9 (£18) or 10 (f.9) branched rays. Anal with 3 spines and 8 (f.17) or g (f.11) branched rays. Caudal weakly emarginate, scaled onits basalhalf. Pectoral fin 27°7-34'1 (M = 311) 190 P. H. GREENWOOD per cent of standard length, 78-0-94°5 (M = 88-9) per cent of head. First ray of pelvic fin produced, especially so in adult males. Teeth. The predominant tooth type of the outer row in both jaws is a slender, very unequally bicuspid, with the major cusp produced, isoscelene to subequilateral in outline, and fairly strongly incurved (text-fig. 23). Slender unicuspids, and tri- cuspids, also occur in the outer row, and some fishes have all three types of teeth. A nearly constant feature is the presence of at least 1, usually 3, unicuspids at the posterior end of the premaxillary tooth row. Tooth form is less variable in larger fishes, where slender, strongly incurved unicuspids predominate. There are 40-56 (M = 50) teeth in the outer premaxillary row. Fic. 23. H. schubotziellus. Premaxillary teeth (left), anterior in position. Viewed from anterior. Scale = 0-5 mm. The inner teeth are tricuspid and broad, and are implanted so as to lie almost horizontally. Two (rarely 3) rows of inner teeth are found in both jaws. OsTEOLOGY. The neurocranium of H. schubotziellus closely resembles that of H. schubotzi, but the dorsal preotic profile (especially anterior to the midpoint of the orbit) is somewhat more decurved in H. schubotziellus. The lower pharyngeal bone is moderately stout and has an equilateral dentigerous area. The teeth are relatively fine, compressed and cuspidate, and are arranged in ca 20-24 rows. Teeth in the two median rows (especially those in the posterior third of the rows) are a little coarser than the others. Vertebral counts in the 8 specimens radiographed are: 28 (f.2), 29 (f.4) or 30 (£.2), comprising 13 (f.5) or 14 (f.3) abdominal and 15 (f.3) or 16 (f.5) caudal centra. COLORATION IN LIFE. Adult males: the ground coloration is greyish-silver with a faint iridescent blue-green sheen, particularly on the midflank region. The belly and ventral body surfaces are whitish. A fairly distinct, deep blue-black stripe extends midlaterally from the caudal fin base to the posterior opercular margin. The dorsal fin is faintly sooty, with short black blotches along its base ; the lappets are black, the margin of the soft part is red and there are deep red spots between the branched rays. The caudal fin has similar red streaks on its proximal half and a pinkish-red flush distally. The anal is faintly dusky, with a slight pink flush ; the ocelli are orange-yellow. The pelvic fins are black. Females are silver. A prominent black stripe runs midlaterally along the body and onto the caudal fin where it extends nearly to the midpoint. The band is of almost constant depth along the body but tapers somewhat on the fin. The dorsal LAKE GEORGE HAPLOCHROMIS SPECIES 191 fin, proximal half of the caudal and the distal part of the anal are hyaline with a suffusion of pale yellow. PRESERVED COLORATION. Adult males: the ground colour is brownish above the midlateral line, greyish to sooty below ; the thoracic region is greyish to tarnished silver. The lower jaw is yellowish-brown, the branchiostegal membrane black. Body markings are variable, but there is usually a complete dark midlateral band from the opercular margin to the caudal origin or else a band from about the middle of the body to the caudal base (sometimes this band is restricted to the posterior third of the body). Occasionally, a second longitudinal band is present, and follows approximately the course of the upper lateral line. Five to 7 rather faint but broad vertical bars are present on the flanks, and extend from the dorsal fin origin to about the level of the ventral margin of the pectoral fin. Cephalic markings are generally present (but faint) and comprise a lachrymal bar or blotch, a small blotch above and in contact with the posterodorsal margin of the orbit, and 2 faint, narrow bars across the nape. The dorsal fin is greyish to dusky, the soft part generally maculate, the lappets dark or black. The caudal fin varies but usually is dark grey with lighter posterior and ventral margins ; otherwise the entire fin is light except for a central grey basal area. The anal is grey to dusky, particularly over the spinous part and along its distal margin. The pelvics are dusky to black, the pectorals hyaline. Females have a light brown ground coloration shading to silver on the lower flanks and belly. A prominent and broad, dark midlateral band runs from the posterior opercular margin onto the basal part of the caudal fin ; in some specimens it extends to the posterior margin of the fin. This band is generally broken at about its mid- point, or at least is much thinner in that region. A second, but far less definite band runs a little above and parallel to the upper lateral line. The dorsal fin is greyish, often with dark lappets and sometimes with several concentrations of dark pigment along its base ; each blotch extends for a short distance upwards onto the fin membrane. The caudal fin is greyish (and has a continuation of the midlateral body stripe). All other fins are hyaline. Ecotocy. Habitat. The species is widely distributed in Lake George and occurs in most habitats. It is particuarly common in muddy bays and near papyrus- fringed shorelines, but is rarely encountered in the open waters of the midlake region. Food. Very little information is available on the food or feeding habits of H. schubotziellus. The presence in the gut of plant and other organic debris, together with dipteran larvae, suggests bottom feeding, possibly insectivorous habits. Breeding. Almost no data are available on breeding habits. The size range of individuals available for analysis is such that one cannot tell precisely at what length sexual maturity is attained. The three smallest fishes examined (45-48 mm standard length) are immature ; the next smallest fish (66 mm standard length) and all others are adult and sexually active. Of the 6 adult females studied, 4 have the right ovary much larger than the left and 2 have the gonads equally developed. Distribution. Lake George and the Kazinga Channel (and probably Lake Edward as well). 192 P. H. GREENWOOD DIAGNOSIS AND AFFINITIES. The close resemblance between H. schubotziellus and H. schubotzi has been noted already (p. 187). However, the species are im- mediately distinguishable on their coloration, even when preserved. The prominent midlateral band (especially in females) is diagnostic, and also serves to distinguish H. schubotziellus from all other species in Lake George (and probably Lake Edward as well). Compared with H. schubotz, H. schubotzellus has a shallower preorbital (12-5-18-4, M = 15:6 per cent head, cf. 16-4-20-9, M = 19-0 per cent) a shorter snout (25:0-33°3, M = 29-2 per cent head, cf. 30:-4—39:6, M = 33-4 per cent) a longer upper jaw (28-2-37-0, M = 34-2 per cent head, cf. 27-0-33:3, M = 30°5 per cent) and a longer pectoral fin (78-0—94-5, M = 88-g per cent head, cf. 73:0-87-0, M = 79-4 percent). There are also slight differences in the shape of the outer teeth, and in the relative stoutness of the lower pharyngeal bone (H. schubotziellus having a coarser bone with, usually, some teeth in the median rows noticeably coarser than the others). This overall resemblance between the species means that H. schubotziellus also resembles the same Lake Victoria species as does H. schubotzi (see above p. 187). Indeed, the stouter lower pharyngeal bone and somewhat coarser median teeth in H. schubotziellus enhance its resemblance to H. viponianus, although in the latter species the lower pharyngeal dentition is rather more specialized (see Greenwood 1960). In many respects H. schubotziellus bears the same phenetic relationships to H. schubotz as does H. riponianus to H. saxicola (Greenwood, op. cit.). The relationship between H. schubotzi and H. schubotziellus could well be a truly phyletic one. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.0.2 : 351 (Holotype) Kankurunga Island 1972.6.2 : 352 (Paratype) Tufmac Bay 1972.6.2 : 353-355 (Paratypes) Kankurunga Island 1972.6.2 : 356-358 (Paratypes) Kankurunga Island 1972.6.2 : 359-306 (Paratypes) Kankurunga Island 1972.6.2 : 367-372 (Paratypes) Kankurunga Island 1972.6.2 : 373 (Paratype) Kankurunga Island 1972.6.2 : 376-377 (Paratypes) I.B.P. Jetty 1972.6.2 : 378 (Paratype) I.B.P. Jetty 1972.6.2 : 374-375 Papyrus fringe of shore Haplochromis taurinus Trewavas, 1933 (Text-figs. 24 & 25) Haplochromis taurinus Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 336 (description based on Lake Edward fishes only). Hototyre. A female, 135-0 mm standard length, BMNH reg. no. 1933.2.23 : 406 from Lake Edward. DESCRIPTION. Lake George. Based on 12 specimens, 72:0-140-0 mm standard length, all from LAKE GEORGE HAPLOCHROMIS SPECIES 193 Depth of body 30:4-38'5 (M = 36-2) per cent of standard length, length of head 27-9-32°0 (M = 306) per cent. Dorsal head profile variable but usually concave, sloping at an angle of ca 40°-45° with the horizontal. Preorbital depth 13-6-18-2 (M = 15-2) per cent of head, least interorbital width 222-296 (M = 25-4) per cent, length of snout 27-4—33-3 (M = 30-6) per cent, 0-7—-0°9 ofits breadth. Eye diameter 27-8-33-3 (M = 30-6) per cent of head, depth of cheek 23°7-30°0 (M = 27-1) per cent. Caudal peduncle 15-7-19-0 (M = 17-0) per cent of standard length, 1-2-1°5 (modal range, I:2—-I-3) times as long as deep. ss Fic. 24. Haplochromis tauvinus. Lake George specimen ; an adult male. Mouth somewhat oblique, lips thickened ; jaws equal anteriorly or the lower projecting a little. Upper jaw 38-0-42-3 (M = 41-0) per cent of head, lower jaw 43°3-50°0 (M = 46-8) per cent, 1:4-2-0 (modal range 1-6-1-8) times longer than broad. Posterior tip of the maxilla mostly exposed, reaching a vertical through the anterior part of the eye or one through the anterior margin of the pupil. Gill rakers of variable form, from short and relatively stout to moderately long and slender ; the lower 1 or 2 rakers are reduced, the upper 2 or 3 often flattened and anvil-shaped. There are 8-11 (mode 9) rakers on the lower part of the first gill arch. No clearly defined pseudorakers are present on this arch, but the tissue between the inner and outer rows of gill rakers is raised into a distinct ridge with slight but circumscribed thickenings in the position usually occupied by pseudo- takers. Scales. Ctenoid ; lateral line with 31 (f.6) or 32 (f.6) scales, cheek with 3 (mode) or 4 rows. Five to 7 (mode 6) scales betewen the upper lateral line and the dorsal fin origin, 6 or 7 (rarely 8) between the pectoral and pelvic fin bases. Fins. Dorsal with 15 (f.8) or 16 (f.4) spinous and 4g (f.7) or ro (f.5) branched rays ; anal with 3 spines and 8 (f.4) or 9 (f.8) branched rays. Caudal subtruncate, 194 P. H. GREENWOOD scaled on its basal half or a little more. Pectoral fin 23-0-29°8 (M = 27-4) per cent of standard length, 81-5—100-0 (M = 89-3) per cent of head. Pelvics with the first ray noticeably prolonged. Teeth. The outer teeth in H. tawrinus show the form (text-fig. 25) which, in Lake Victoria Haplochromis species, is associated with paedophagus habits (Greenwood 1959b). Also, as in those species, the teeth of H. tawrinus are deeply embedded in the mucosa of the jaws. In both jaws the basic tooth form is similar, namely a cylindrical neck and lower crown, but with a markedly compressed, chisel-like bicuspid upper crown. Upper jaw teeth have a crown in which the minor cusp is distinct and the major cusp is obliquely truncate ; the entire crown is curved in- wards. Lower jaw teeth have the major cusp very obliquely truncate, the minor cusp distinct and the entire crown has a slight but definite outward inclination. The posterior third to half of the premaxilla is edentulous ; the toothed part of the bone carries 32-48 (M = 36) teeth. AQ coed Fic. 25. H.tauvinus. Teeth. (a) Premaxillary teeth (right), anterior in position. Viewed laterally. (b) Dentary teeth (right), anterolateral in position. Viewed from a point slightly anterior of lateral. Scale = 0-5 mm. The zmner teeth in both jaws are small and tricuspid, and are arranged in I or 2 rows. Like the outer teeth, those of the inner rows are deeply embedded in the mucosa. OsTEoLoGy. The neurocranium of H. taurinus is of the generalized Haplochromis type. The premaxilla and dentary are also basically of a generalized type, the pre- maxilla not therefore showing the relative elongation of its ascending process (as occurs in some Lake Victoria paedophages, e.g. H. parvidens). Thus, in all syncranial features H. tauvinus is comparable with the less specialized embryo and larval fish- eating species of Lake Victoria, viz. H. maxillaris and H. obesus. The lower pharyngeal bone is relatively fine, its dentigerous area a little broader than long (ca 1-1 times). The teeth are slender, compressed and cuspidate, with those in the two median rows coarser than the others; there are ca 20-22 rows of teeth. LAKE GEORGE HAPLOCHROMIS SPECIES 195 Vertebral counts in the 7 specimens radiographed are 28 (f.1), 29 (f.5) or 30 (f.1), comprising 13 abdominal and 15 (f.1), 16 (f.5) or 17 (f.1) caudal centra. COLORATION IN LIFE. Males: the live colours of sexually active males are un- known. A juvenile male had similar coloration to that of a female (see below) except that there were faint traces of a rosy flush on the operculum and anterior parts of the flanks (especially intense above the pectoral fin insertion). Other differences noted were that the lower limb of the preoperculum, the cheek and the lower lip were a pale iridescent blue. The dorsal fin had red streaks between the rays, as had the caudal fin where the colour was most intense on the middle of the fin. Two well-defined yolk-yellow ocelli were present on the otherwise hyaline anal fin. Females, both adult and juvenile, have a silvery-grey ground colour shading to white on the belly, and a faint, yellowish overlay on the flanks (more intense in adults than in juveniles). The dorsal fin is hyaline but faintly yellow along its insertion. The caudal fin is yellowish-green over its basal half, pale yellow-green distally. The anal is faintly yellow, and the pelvics are hyaline. PRESERVED COLORATION. Adult males are brownish above, shading to silvery grey (with faint dusky overtones) on the flanks and belly ; the chest is dusky silver, the branchiostegal membrane dusky grey. The head has a well-defined and intense lachrymal stripe continued through the eye and terminating as a blotch above and slightly behind the dorsal margin of the orbit. Other cephalic markings include a dark vertical arm of the preoperculum, and 2 rather faint transverse bars across the snout. A dark area just anterior to the dorsal fin origin is faintly visible. The dorsal fin is greyish, the caudal yellow-brown with traces of dark pigment between the rays (especially those in the middle of the fin). The anal fin is yellowish with a faint dusky overlay, the pelvics are black and the pectorals hyaline. Females are light brownish-yellow above, shading to silvery yellow ventrally. The snout is grey-brown and there is an ill-defined and faint lachrymal stripe which does not extend through the eye to the dorsum. The dorsal and caudal fins are greyish (the former slightly the darker) ; all other fins are hyaline. BioLocy. So few specimens of H. taurinus have been caught that it is impossible to generalize on the biology of the species. Apparently it is confined to inshore regions of the lake, where it has been taken off the papyrus fringe and also over sandy beaches in sheltered areas. Judging from the dentition (see above p. 194) and the widely distensible mouth, H. taurinus, like similarly adapted species in Lake Victoria, feeds on the embryos and larvae of other cichlid fishes (see Greenwood 1959b). This supposition is borne out by the only two guts that yielded food remains. In these there were fragments of larval cichlids, bones of small fishes (of a size compatible with their being from larval fishes) and a fatty, yellow fluid closely resembling yolk. Little information has been collected on the breeding habits of H. taurinus. The two available fishes less than 80 mm standard length are both immature ; all speci- mens of 83 mm standard length and longer are sexually active. Only 2 of the adult females examined have ovaries in an advanced stage of oogenesis ; in both fishes the right ovary is slightly larger than the left one. 196 P. H. GREENWOOD Distribution. Lakes Edward and George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Among the Haplochromis species of Lakes George and Edward, H. taurinus is immediately recognizable by its dentition (see above p. 194) and by its broad and laterally distensible mouth. Outside these lakes, H. tawrinus bears a close resemblance in both general mor- phology and in its dentition to H. maxillaris Trewavas of Lake Victoria (see Green- wood 1959b). Morphometric differences between the species are slight, with the jaws of H. maxillaris being somewhat larger and thus the gape in this species being a little greater than in H. tauvinus. As far as can be told from the colours of juvenile male H. taurinus (compared with both adult and juvenile H. maxillaris) there is also a difference in this character. Taking all anatomical characters into consideration, H. maxillaris is the more specialized species of the two. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 29 Sandy shoal 1972.6.2 : 30 Kankurunga Island 1972.6.2 : 31-34 Papyrus fringe opposite I.B.P. Laboratory 1972.6.2 : 35 Locality unknown 1972.0.2 : 36-37 Locality unknown 1972.6.2 : 38-42 Locality unknown 1972.6.2 : 806 (figured specimen) Kankurunga Island Haplochromis labiatus Trewavas, 1933 (Text-figs. 26 & 27) Haplochromis labiatus Trewavas, 1933, J. Linn. Soc. (Zool.), 38: 335 (holotype and only specimen ; from Lake Edward). Haplochromis labiatus was described from the holotype alone, although two smaller fishes were also mentioned (Trewavas 1933) ; all 3 specimens are from Lake Edward. Only 1 specimen identifiable as H. Jabiatus has been caught in Lake George (from a locality close to the northern shore of Akika Island, in shallow water ca 1-14m deep, over mud, and near sparse stands of the reed Phragmites). The Lake George fish is 87-0 mm standard length and is of indeterminable sex. Depth of body 41-4 per cent of standard length, length of head 32-2 per cent. Dorsal head profile concave above the orbit, sloping steeply at ca 45° to the hori- zontal. Preorbital depth 16-1 per cent of head, least interorbital width 24-3 per cent, length of snout 0-8 its breadth and 28-6 per cent of head length. Eye diameter 28-6 per cent of head, depth of cheek 25-0 per cent. Caudal peduncle 14-1 per cent of standard length, 1-1 times its depth. Mouth horizontal, lips noticeably thickened (but not produced into lobes medially) posterior tip of the maxilla reaching a vertical through the anterior part of the eye. Upper jaw 32:2 per cent of head, lower jaw 35:7 per cent, 1-4 times longer than broad. LAKE GEORGE HAPLOCHROMIS SPECIES 197 Gill rakers. On the lower part of the first gill arch there are 8 rakers of which the lowermost is reduced, the following 2 are stout and the remaining 5 are relatively slender. Pseudorakers are present but weakly developed. Ay A) Uv) Fic. 26. Haplochromis labiatus. Lake George specimen. Scales. Ctenoid, lateral line with 30 scales, cheek with 3 rows. There are 54 scales between the upper lateral line and the dorsal fin origin, 7 between the pectoral and the pelvic fin bases. Scales on the thoracic region are small. Fins. Dorsal with 15 spines and 10 branched rays, anal with 3 spines and 9 branched rays. Pectoral 31-0 per cent of standard length, 96-5 per cent of the head. Caudal scaled on its basal half; the distal margin is frayed but was apparently truncate when intact. Pelvics with the first ray barely produced. Teeth of the outer row in both jaws are stout and somewhat compressed (text-fig. 27). In the upper jaw the two posterior teeth on each side are unicuspid, but the remainder are unequally bicuspid, with the crown vertically orientated. There are 32 teeth in this row. All teeth in the lower jaw are bicuspid, but those located posteriorly are smaller than the more anterior teeth. Anteriorly, the lower jaw teeth are slightly procum- bent, each tooth lying forward at an angle of about 80° to the horizontal. The inner teeth in both jaws are small and tricuspid, and are arranged in 3 rows in the upper jaw and 2 in the lower. The lower pharyngeal bone is moderately stout and its dentigerous area is slightly broader than long (1-1 times). The teeth are compressed and cuspidate, with the posterior teeth of the median rows coarser than the others. The vertebral count is 13 abdominal + 16 caudal centra. (The type has 13 + 17 centra.) Only preserved coloration is known. The body is brownish above and shades to silvery on the ventral flanks, chest and belly; the entire head is brownish. All 198 P. H. GREENWOOD fins, except the pelvics, are hyaline-greyish, the soft dorsal is maculate posteriorly. The pelvics are dusky on the anterior half but hyaline posteriorly. Distribution. Lakes Edward and George; not yet recorded from the Kazinga Channel. as Fic. 27. H. labiatus. Dentary (anterior portion) with anterior and anterolateral teeth in situ (right side, viewed laterally); from the Lake George specimen. Scale = 0-5 mm. COMPARISON WITH THE HOLOTYPE. The holotype is a larger fish (109 mm standard length) and is from Lake Edward. The Lake George specimen differs from the holo- type in several minor ways, but in most details and in its overall morphology it resembles that specimen more closely than it does specimens of any other species. The principal morphometric difference is in the longer jaws of the Lake George fish (35:7 and 32-2 per cent of head, cf. 32-3 and 28-0 per cent for the upper and lower jaws respectively); such a difference, however, is well within the range of variation for these characters in other Haplochromis species. The lips of the holotype are clearly much better developed than are those of the Lake George specimen, and its teeth are predominantly unicuspid, not bicuspid as in the Lake George fish. However, some teeth in the holotype do show indications of avery smalllateralcuspremnant. Both these differences could be attributable to the larger size of the holotype. Certainly the difference in lip development is well within the range of variation encountered in other species with hypertrophied lips and is not necessarily size-correlated. A further dental difference lies in the more clearly procumbent anterior teeth of the holotype. I am unable to comment on the sig- ficance of this character. DIAGNOSIS AND AFFINITIES. As only two specimens are available (and those from different lakes) it is difficult to provide a precise diagnosis. With so few specimens studied doubts might well be raised as to the validity of the species. However, if various dental and morphometric characters are combined, it seems most likely that H. labiatus is a valid species. The teeth, thick lips and short lower jaw (ca 35 per cent of head) in H. labiatus together with its strongly concave profile distinguish the species from H. limax (teeth with obliquely cuspidate major cusps, compressed and relatively slender ; lower jaw 36:0—40-9, mean 38:5 per cent). The concave profile of H. /abiatus is an immediately obvious difference when a comparison is made with H. elegans ; the species also differ in that H. elegans has LAKE GEORGE HAPLOCHROMIS SPECIES 199 a shallower cheek (18-2-24:5, mean 20°8 per cent head, cf. 23-5 per cent), and has obviously bicuspid teeth, even in large specimens. Dental differences like these noted above distinguish H. labiatus from H. aeneo- color (which also has more teeth in the premaxillary, viz. 40-56, mean 48, cf. 32), a longer lower jaw (38-0-—44:0, mean 41-0 per cent head, cf. about 35-0 per cent) and a straight or convex dorsal head profile. Similarly, the dentition and gross morphology serve to separate H. labiatus from other Lake George species. Within this lake the appearance of H. labiatus (and, basically, its dentition) is most like that of H. elegans and H. aeneocolor, but the re- semblances are less close than are those with H. beadlei Trewavas of Lake Nabugabo, and with species of the H. crassilabris species complex in Lake Victoria. Unfor- tunately I have still to resolve satisfactorily the H. crassilabris problem (see Green- wood 1965b). Nevertheless, species of this complex can each be distinguished from H. labiatus. Characters separating H. beadlei from H. labiatus holotype were detailed in Greenwood (1965b). These will be found less trenchant when the Lake George fish is taken into account. More specimens of H. /abiatus must be studied before the relationships, both phyletic and phenetic, of the two species can be settled. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 809 East side of Akika Island Haplochromis pappenheimi (Blgr.), 1914 (Text-figs. 28-30) Tilapia pappenheimi (part) Boulenger, 1914, in Wiss. Evgebn. Deuts. Zentral-Afrika Exped., 1907-1908, Zool. 3 : 254-255 (10 of the 32 syntypes, all from Lake Edward, see note below). T. pappenheimi (part) Boulenger, 1915, Cat. Afr. Fishes, 3: 232-233 (4 of the 6 specimens listed, all from Lake Edward; the figured specimen is not in the B.M. (N.H.) collections but is in Berlin [see below]. The skeleton listed cannot be identified with certainty, but probably it is not this species). Haplochromis pappenheimi : Regan, 1921, Ann. Mag. nat. Hist. (9), 8: 634-635 (all Lake Edward fishes). Haplochromis pappenheimi : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 334 (all Lake Edward fishes). NOTE ON THE TYPE SERIES. Six of the syntypes were deposited in the British Museum (Natural History) ; the other 26 specimens were retained by the Berlin Museum. Regan (1921) reviewed the B.M. (N.H.) material and referred 2 specimens to the new species H. nigripinnis and H. eduardii described in that paper. Regan did not examine the Berlin syntypes, and no lectotype was chosen. The 4 remaining syn- types were considered conspecific by Regan, a conclusion with which I concur. Through the kindness and cooperation of Dr Deckert, I have been able to study the 26 syntypes (including the figured specimen) from the Berlin Museum collections. This series proved to be polyspecific and cannot be fully evaluated until the Lake Edward Haplochromis species are revised. For the moment, however, it should be noted that 12 syntypical specimens are conspecific and are considered to be H. 200 P. H. GREENWOOD pappenheimi (see below). The other specimens, in part, are probably referable to H. nigripinnis (8 specimens Z.M. Berlin, nos. 22693 and 22698), H. eduardit (2 specimens, Z.M. Berlin, no. 22692) and 2 specimens to species as yet undescribed (Z.M., nos. 22695 and 22696). Boulenger’s (1914) original description of H. pappenheimi is quite inadequate by current standards, and thus it is impossible to determine from it the morphological limits of his species. The reason for my deciding that certain specimens are ‘ H, pappenheimi’ is essentially an attempt to avoid unnecessary nomenclatural change. There is certainly a biologically and morphologically valid taxon, occurring in both Lakes Edward and George, whose characteristics are recognizable in 16 of the H. pappenheimi syntypes (4 specimens from the B.M.[N.H.] and 12 from the Berlin Museum (Z.M.B. lot nos. 22689 and 22697). The four B.M. [N.H.] fishes are those on which Regan [1921] based his redescription of the species). It is to these 16 specimens that I have decided the name ‘ pappenheimi’ should be restricted and from which the lectotype should be chosen. If any of the other syntypes was chosen as lectotype, then the name ‘ pappenheimi’ would either fall into synonymy or would replace the name of an already established taxon. Either way, a new name would have to be found for the taxon here considered to be Haplo- chromis pappenheimi. Regrettably, the specimen illustrated in Boulenger (1914, 1915) cannot be referred to the taxon H. pappenheimi as recognized by Regan (of. cit.) or myself. It is a fish of 66:0 mm standard length, probably a female, ZMB no. 22692. As far as I can determine this fish is a specimen of H. eduardii Regan. To avoid the nomenclatural changes that would follow the choice of this fish as lectotype, I have selected for that purpose a specimen from the 16 syntypes showing the diagnostic features of H. pappenheimi, sensu Regan (1921). This fish, a female 73-0 mm standard length (ZMB no. rg1r0) has a characteristically elongate and slender body, and also clearly shows the dental and gill raker characters of the species (text-fig. 28). 1) 4, WMA Ria) AY Rain icaiae Bs os aneantateee Fic. 28. Haplochromis pappenheimi. Lectotype; a Lake Edward specimen. LAKE GEORGE HAPLOCHROMIS SPECIES 201 Description. Based on 20 specimens, 38:5—61-0 mm standard length, all from Lake George. Depth of body 26-6-30-4 (M = 29-9) per cent of standard length, length of head 30°5-34°4 (M = 32:3) per cent. Dorsal head profile straight, sloping gently at an angle of ca 20°-25° with the horizontal. Preorbital depth 12:5-16-8 (M = 14-8) per cent of head, least interorbital width 23°3-29°2 (M = 26-4) per cent, length of snout 1-0-1-1 times its breadth, and 24:0- 30:8 (M = 28-1) per cent of head. Eye diameter 31-5-37:0 (M = 33°9) per cent (not showing any allometry in this size range), depth of cheek 13-4-19°5 (M = 17-0) per cent. Caudal peduncle 19-5-24:5 (M = 22-2) per cent of standard length, 1-5-2-0 (modal range 1-7-1-8) times as long as deep. Fic. 29. H. pappenheimi. Outline drawing of a Lake George fish. Mouth slightly oblique, jaws equal anteriorly. Posterior tip of the maxilla reach- ing a vertical slightly anterior to the orbital margin. Upper jaw 25-0—29-2 (M = 27°5) per cent of head, lower jaw 35°7-41°5 (M = 39:0) per cent, I-7-2°5 (modal range 2:3-2°5) times longer than broad. Gill rakers. The lower 1 or 2 rakers (rarely the first 4) are reduced, the others are long and relatively slender. Occasionally some of the uppermost rakers are flattened, or flattened and bifid. There are 1o(rare)-13, mode 11, rakers on the lower part of the first gill arch. No pseudorakers are present. Scales. Ctenoid; lateral line with 33 (f.7) or 34 (f.12) scales, cheek with 2 or 3 rows. Five (rarely) to 7 (mode 6) scales between the upper lateral line and the dorsal fin origin, 5-7 (mode 6) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.1), 15 (f.10) or 16 (f.9) spinous and 8 (f.4), 9 (f.12) or to (f.5) branched rays. Anal with 3 spines and 7 (f.2), 8 (f.7) or 9 (f.11) branched rays. Caudal slightly emarginate, scaled on its basal half or slightly more. Pectoral fin 22:0-27-5 (M = 25:0) per cent of standard length, 71-0-84:5 (M = 77-1) per cent of head. Pelvics with the first ray slightly produced. 202 P. H. GREENWOOD Teeth. The outer teeth in both jaws are unequally bicuspid, with the major cusp obliquely truncate in most fishes, but especially so in larger individuals (text-fig. 30) ; both the crown and the neck of the tooth are compressed. Occasionally, the teeth situated posteriorly and posterolaterally in the premaxilla are tricuspid or unicuspid ; this difference may be size correlated, with unicuspids commoner in larger fishes (i.e. > 55 mm standard length). aatit Fic. 30. H. pappenheimi. Dentary teeth (right), lateral and anterolateral in position. Viewed laterally. Scale = 0-25 mm. The posterior quarter to third of the premaxilla is edentulous ; there are 28-38 (M = 34) teeth on the rest of the bone. Inner teeth are small, invariably tricuspid, and are arranged in a single row in both jaws. Sometimes in the upper jaw the row is irregular and gives the impression of being double. OsTEoLOGY. The neurocranium of H. pappenheimi departs from the generalized Haplochromis type, approaching that found in H. schubotzi and H. schubotziellus, and in H. guwiarti of Lake Victoria (see above p. 186, and Greenwood 1962). The lower pharyngeal bone is fine, and its demtigerous surface is distinctly broader than long (1:25-1:30 times). The teeth are very fine, slender, compressed and cuspidate, and are closely arranged in from 30 to 34 rows. The posterior margin of the bone is noticeably concave in outline when viewed from above, having the shape of a shallow V. Vertebral counts in the 8 specimens examined are 30 (f.3), 31 (f.4) or 32 (f.1), comprising 14 (f.6) or 15 (f.2) abdominal and 16 (f.4) or 17 (f.4) caudal centra. COLORATION IN LIFE. There appears to be no marked sexual dimorphism in coloration, although the possibility of this occurring cannot be overruled because no sexually active males have yet been examined. However, large and adult males were caught in the Kazinga Channel and these did not differ from females, except in being slightly darker. The ground colour in both sexes is silver, shot with green iridescence above the midlateral line, and whitish on the belly. All fins are hyaline, but in males the lappets of the dorsal fin are black, as are the pelvic fins. The anal fin of males carries from 1 to 3 yolk-yellow ocelli. PRESERVED COLORATION. Ground coloration is greyish-silver, darker (i.e. greyer) on the dorsum and flanks to about the midlateral line. All fins are hyaline to greyish, the lappets of the spinous dorsal black in males, as are the pelvic fins. Distribution. Lakes Edward and George and the Kazinga Channel. LAKE GEORGE HAPLOCHROMIS SPECIES 203 Ecotocy. Habitat. The slender body form and pelagic habits of this species have resulted in relatively few specimens being caught in the nets used by the I.B.P. team. Thus, not a great deal is known about the habitat preferences of H. pappenheimi. The species is apparently confined to upper water levels in offshore regions of the lake (both in bays and in the open lake). An intensive study of the species is now in progress. The use of purse-seine nets and small-mesh trawls should result in many more samples being taken. Food. The few guts examined contained only zooplankton, particularly copepods and cladocerans. Breeding. Most of the specimens on which this description is based are sexually immature, but 2 fishes (females, 60 and 61 mm standard length) show signs of ovarian activity. Obviously, adults of H. pappenheimi in Lake George must reach a larger size (as they do in Lake Edward and the Kazinga Channel). There is, of course, the possibility that adult individuals move out of Lake George, and that the breeding sites are in the Kazinga Channel. To date the only evidence (and that rather flimsy) supporting this hypothesis is the capture, in May 1972, of numerous large adults in the channel, when, using the same gear (a purse seine), only juveniles were collected in the Lake itself. One of the paralectotypes from Lake Edward, a female (72 mm standard length) is brooding young in the buccal cavity, as are 2 females (72-0 and 69:0 mm standard length) in the paralectotypical series of the British Museum (Natural History). Unfortunately it is impossible to sex the other paralectotypes with any certainty. All are apparently female, the largest a fish 92-5 mm standard length. DIAGNOSIS AND AFFINITIES. The species is immediately distinguishable from other Lake George Haplochromis species by its slender body form, head shape, by its long and slender caudal peduncle, and by the shape and distribution of its outer jaw teeth. In life the silvery blue-green coloration is also diagnostic. There is an overall similarity between H. pappenheimi and H. guiarti, a species endemic to Lake Victoria but which was once thought to occur in Lake Edward (Trewavas 1933 ; but see below p. 232). Haplochromis pappenheimi differs from H. guiarti in several characters; for example, the dentition (obliquely cuspidate teeth, contrasted with the usual uni- cuspid and unequally but acutely bicuspid teeth of H. gwiarti ; the fewer teeth in H. pappenheimi, and the fully toothed premaxilla of H. gudarti compared with the posteriorly edentulous bone of H. pappenheimt), a shallower preorbital (12-5-16°8, M = 14°8 per cent head, cf. 16-3-21°5, M = 18-3 per cent in H. guiarti), shorter snout (24:0-30-8, M = 28-1 per cent head, cf. 31-7-37:5, M = 34-4 per cent), larger eye (31°5-37:0, M = 33-4 per cent head, cf. 23-6-29°8, M = 265 per cent in H. guiarti of the same size), and a longer caudal peduncle (19-8-24-5, M = 22-2 per cent standard length, cf. 16-2-20-8, M = 18-9 per cent). Among the Lake Victoria Haplochromis species there is only one other, H. fusiformis Greenwood & Gee, that resembles H. pappenheimi. Haplochromis fusiformis is confined to the deeper waters (90-100 ft) of the lake and is benthic in habits. Like H. pappenheimi it has a partly edentulous premaxilla and is slender bodied. Interspecific differences, however, are well marked and include a shallower body, 12 204 P. H..GREENWOOD deeper preorbital, smaller eye and longer lower jaw in H. fusiformis, as well as the retention of typical bicuspid teeth in that species. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 297-298 Kankurunga Island 1972.6.2 : 317-320 Small island north of Kankurunga Island 1972.6.2 : 328-332 Small island north of Kankurunga Island 1972.6.2 : 333-336 Small island north of Kankurunga Island 1972.6.2 : 337-342 Tufmac Bay Haplochromis squamipinnis Regan, 1921 (Text-fig. 31) Haplochvomis squamipinnis Regan, 1921, Ann. Mag. nat. Hist. (9), 8 : 636 (a single specimen, the holotype, BMNH reg. no. 1914.4.8 : 32, from Lake Edward). Haplochromis squamipinnis : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 338-339 (mostly Lake Edward fishes, but 1 from Lake George). Note on H. mentatus Regan 1925, A PUTATIVE SYNONYM OF H. squanupinnis. Trewavas (1933) noted that the holotype and only specimen of H. mentatus (a Lake Edward species) closely resembled H. squamipinnis. She was unable to examine the holotype and thus did not see fit to formally synonymize the species with H. squamupinnis. During a recent visit to the Museum of Comparative Zoology, Harvard University, I was able to examine H. mentatus holotype (MCZ no. 31523), a small fish 94-0 mm standard length, of indeterminable sex. In all morphometric characters and in most anatomical features this specimen is, as Trewavas suggested, like a young H. squamupinnis. However, I could not find any trace of the minute scales which are closely adherent to the bases of the anal and dorsal fin rays of H. squamipinnis, irrespective of the individual’s size (see below p. 206). These scales, which extend for a short distance onto the fins, are delicate and easily dislodged. Furthermore, there is considerable individual variability with regard to the number and position of the fin rays with which the scale rows are associated. Despite this variation, however, I have yet to examine a specimen of H. squamipinnis in which there is absolutely no trace of fin scales. Thus, it is difficult to assess the significance of their total absence in H. mentatus holotype, especially since in all other trenchant characters the specimen agrees with comparable-sized H. squamipinms. Personally, I would be inclined to consider it either a young H. squamipinnis in which all traces of fin scales rows are lost, or an aberrant member of the species in which these scales failed to develop. DESCRIPTION OF H. squamipinnis. Based on 34 specimens 34-0-202-0 mm stan- dard length, from Lake George (including the specimens collected by Worthington in 1931). Because most characters show some allometry with standard length, the sample has been divided into two groups, viz.: (a) fishes < 120 mm standard LAKE GEORGE HAPLOCHROMIS SPECIES 205 length and (b) fishes > 129 mm standard length. Ranges and means for the various morphometric characters are given accordingly. Depth of body (a) 32:1-37:0 (M = 34:7), (b) 34:7-41-0 (M = 37-4) per cent of standard length, length of head (a) 33-5-37-6 (M = 35-4), (b) 33:9-36-2 (M = 35:0) per cent. Dorsal head profile straight but broken by the prominent premaxillary pedicels, sloping at an angle of ca 40° with the horizontal. Preorbital depth (a) 15-0-19:0 (M = 17-0), (b) 17-8-20-4 (M = 21-9) per cent of head, least interorbital width (a) 19:2-23-0 (M = 21-9), (b) 21-3-25-4 (M = 23:0) per cent ; length of snout equal to or slightly greater than its breadth in fishes of both size groups, but in (a) 30-8-33-8 (M = 32:3) per cent of head and in (b) 31-5-37-1 (M = 34:9) per cent. Eye diameter (a) 23-0-33:0 (M = 27°5), (b) 21-2-25-5 (M = 23-0) per cent of head ; depth of cheek (a) 18-3-27-6 (M = 24:5), (b) 27-8-31-7 (M = 29:3) per cent. For the cheek depth and eye diameter the lowest and highest values respectively relate to the smallest specimen examined. 10mm pee |) Fic. 31. Haplochromis squamipinnis. Lake George specimen. Inset shows squamation on dorsal fin. Caudal peduncle not showing allometric growth, its length 14-7-20-6 (M = 16:8) per cent of standard length, 1-1-1-5 (modal range, 1-2-1-4) times as long as deep. Mouth distinctly oblique, sloping upwards at an angle of ca 35-45 degrees to the horizontal. Jaws equal anteriorly or the lower projecting slightly (the usual condition). Length of lower jaw in (a) 41-8-53-8 (M = 47°5), (b) 47-3-56°6 (M = 51-3) per cent of head, length of upper jaw in (a) 33-3-40-0 (M = 37-0), in (b) 39°3-45°5 206 P. H. GREENWOOD (M = 42-4) per cent, I-1-1-5 (modal range 1-2~1-4) times longer than broad in both size groups. The smallest specimen examined (34:0mm standard length) has, proportionately, the shortest jaws. Posterior tip of the maxilla reaching a vertical through the orbital margin in most fishes, not quite reaching this level in a few, and extending beyond it to below the eye in others ; this variation does not seem to be size correlated. Gill vakers. The lower 1-3 rakers are reduced, the remainder are relatively slender. In some specimens the uppermost I or 2 rakers may be flat and anvil- shaped, and in some individuals all the rakers are short and relatively stout. There are 8-11 (rare), usually 9, rakers on the lower part of the first gill arch. Pseudorakers are poorly developed: the tissue lying between the inner and outer gill raker rows is thickened and slightly produced into ill-defined, low pro- jections. Scales. Ctenoid; lateral line with 31 (f.1), 32 (f.17) or 33 (f.12) scales, cheek with 3, 4 (mode) or 5 rows. Five anda half to 7 (rare), usually 6, scales between the upper lateral line and the dorsal fin origin, 5~7 (mode 6) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.1), 15 (f.21) or 16 (f.9) spinous and g (f.19) or Io (f.12) branched rays. Anal with 3 spines and 9g (f.22) or ro (f.9) branched rays. A peculiarity of both dorsal and anal fins is the presence of short vertical rows of minute scales extending from the body onto the fin (see text-fig. 31). The scales are closely applied to the fin rays, both spinous and branched. Not all rays have associated scale rows, and the distribution of the rows shows considerable individual variability. Most frequently, the entire soft part of the dorsal fin has basal scale rows, as does the entire soft anal. Generally the anterior part of the spinous dorsal is asquamous, the scale rows only beginning at about the sixth or seventh spine. However, the rows may begin as far forward as the fourth spine, or may not appear until the eleventh spine. The scales are easily dislodged and part of this variability may be attributable to damage sustained when the fish is caught, especially if it has been trapped in a gill-net. No specimen was seen in which all the fin scale rows are missing, and the anal fin squamation is usually better preserved than that of the dorsal (another reason for thinking that damage during capture may account for some of the observed variability). The caudal fin is truncate to weakly subtruncate, and is scaled on its basal half or a little more. Pectoral fin length shows no marked allometry with standard length ; however, the relatively shortest fin is found in the smallest fish (34-0 mm standard length). Length of pectoral 23-5-30:0 (M = 28-1) per cent of standard length, 66-6-87-0 (M = 79-9) per cent of head. Pelvic fins have the first ray produced and relatively more elongate in adult males than in females. Teeth. Except for the smallest specimen (34:0 mm standard length), the pre- dominant tooth form in the outer vow of either jaw in fishes of 49-0-202-:0 mm standard length is a slender but strong unicuspid with a slightly to strongly incurved LAKE GEORGE HAPLOCHROMIS SPECIES 207 crown. Some fishes in the size range 75-85 mm standard length have a few weakly and unequally bicuspid teeth interspersed among the unicuspids anteriorly and anterolaterally in both jaws; in one fish, most outer teeth in the lower jaw are bicuspid, although some typical unicuspids occur anteriorly. In fish of all sizes the posterior premaxillary teeth are unicuspid. The 34mm standard length fish has typically unicuspid teeth posteriorly and posterolaterally in the upper jaw, but unequally bicuspid teeth (the major cusp long and slender) anteriorly ; all the lower jaw teeth in this fish are unicuspid and typical. = The number of outer premaxillary teeth is positively correlated with standard length ; in fishes < 118 mm standard length there are 34 (in the smallest specimen) to 70 (M = 48) teeth, and in fishes > 130 mm standard length, 46-80 (M = 60) teeth. Fishes > 90 mm standard length have the inner tooth rows composed of slender unicuspids, although in fishes as long as 100 mm some weakly tricuspid teeth may also occur. At lengths between 70 and go mm, there is usually an admixture of tri- and unicuspid teeth, but in some fishes only tri- and weakly tricuspids are found. In smaller specimens, tricuspids predominate. OsTEOoLOGY. The neurocranium in H. squamipinnis is of a relatively specialized type, both with respect to the basic Haplochromis neurocranial type, and also to the presumed basic skull form in piscivorous species (such as is shown by H. guwiarti, H. victorianus and H. serranus of Lake Victoria, see Greenwood 1962, 1967). It compares closely with the neurocranial type found in Lake Victoria species like H. longirostris and H. mento, species that have deviated from the near basic H. serranus grade (Greenwood, op. cit.). The lower pharyngeal bone is relatively fine and has an equilateral dentigerous area. The teeth are compressed, but strong, with weakly developed cusps and are somewhat sparsely distributed on the bone in ca 18-20 rows. Vertebral counts in the 11 specimens radiographed are 29 (f.3) and 30 (f.8), com- prising 13 (f.11) abdominal and 16 (f.3) or 17 (f.8) caudal elements. COLORATION IN LIFE. Adult males: the ground colour is greyish with a turquoise to blue-green sheen covering most of the flank and caudal peduncle. The dorsal head and body surfaces are dark grey, the chest and belly greyish to greyish-sooty, the branchiostegal membrane sooty. Dorsal fin dusky, darkest along its basal third, the upper margin of this dark area with a gently undulating outline; the lappets are black but there is a narrow, pinkish margin to the soft part of the fin. Deep orange-red spots occur between the posterior rays of the soft dorsal. Caudal fin dark hyaline, with deep red spots and a light red (almost pink) flush over the ventral half of the fin and at its upper and lower distal angles. Anal fin almost completely pink, but with a hyaline area around the yolk-yellow ocelli. The pelvic fins are black. Females and immature males are golden-silver dorsally, shading to yellowish- green on the flanks, and to white on the belly and ventral flanks. Dorsal fin yellowish-green, becoming hyaline dorsally. The caudal is similarly coloured, but 208 P. H. GREENWOOD has dark spots over its proximal two-thirds. Anal fin dark hyaline basally, yellow- green distally ; females often have yellow spots in the position of the ocelli in males, but young males may lack any indications of such markings. The pelvics are hyaline or yellowish, becoming dusky in near-adult males. PRESERVED COLORATION. Adult males: the ground colour is brownish-grey above, shading to silvery grey or dusky silver on the lower flanks and belly ; the chest is silvery or dusky silver. Upper surface of the snout and the lips (at least anteriorly) are dark grey, the lower jaw and branchiostegal membrane dusky grey. The only cephalic marking is an ill-defined lachrymal stripe. The dorsal fin is grey, the lappets black. Basally there is a darker band (in places almost black) which widens over the soft part of the fin so that the basal third to half of the fin is black or very dark grey in colour. The distal part of the soft dorsal is densely and darkly maculate. The caudal is dark grey, most intensely so over its proximal third to half. The anal is grey to dusky grey, with the basal third noticeably darker. The pelvic fins are black. Females and juvenile males are light brown above, shading to silvery yellow on the flanks and belly. The lips, snout and lower jaw are coloured as in adult males, but are slightly paler in some individuals. The lachrymal blotch is poorly defined but is generally visible. The coloration of the dorsal, caudal and anal fins is like that of adult males, but is more variable in its intensity. The pelvics are hyaline or a little dusky (more so in juvenile males) over the distal part of the first 3 rays. Ecorocy. Habitat. At least when adult, individuals of H. squamipinnis are found in all habitats, but are especially common in the offshore open-water areas of the lake. Juveniles may have a more restricted distribution to areas nearer the shoreline. Food. Fishes predominate in the diet of H. squamipinnis at all the growth stages investigated (70-200 mm standard length), and appear to be the sole food source of fishes more than 150 mm standard length. In small fishes insects contribute sub- stantially to the diet. As far as could be determined, the principal prey species are other Haplochromis, but the macerated nature of the gut contents in H. squamipinnis generally pre- cludes accurate identification. Breeding. Haplochromis squamipinnis is a female mouth brooder. All ex- amined individuals less than 120mm standard length are immature; the first indications of sexual activity are found in fishes (of both sexes) in the size range 125-130 mm standard length. Because all specimens larger than 155 mm standard length are females, there may be sexual dimorphism in the maximum adult size attained. Of the 13 sexually active females examined, 5 have the right ovary larger than the left one, 2 have the right ovary slightly larger and 6 have the ovaries equally developed. Distribution. Lakes Edward and George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. The overall appearance of H. squamipinnis with its long and obliquely sloping jaws, its unicuspid teeth, and the peculiar scale rows LAKE GEORGE HAPLOCHROMIS SPECIES 209 on the median fins, immediately distinguishes this species from all known Haplo- chromis species in Lake George. The same criteria would distinguish it from all known Haplochromis species in Lake Edward, but.a rather similar species is now known to occur in that lake (Greenwood, unpublished information). Haplochromis squamipinnis resembles a number of piscivorous Haplochromis species in Lake Victoria, but none of these has scale rows on the fins. Taken in concert, the cranial and dental characters of H. sguwamipinnis place it near H. victorianus and H. serranus (i.e. the ‘ serranus’ group of Greenwood 1967). However, in certain characters, especially skull form, H. squamipinnis approaches the rather more specialized structural grade seen in H. mento and H. longirostris (i.e. of the ‘ prognathus’ group as defined by Greenwood, of. cit.). Despite the general and often particular resemblances between H. squamipinnis and these various Lake Victoria species, it can be distinguished from them by various combinations of morphometric characters. It may be significant, in phylogenetic terms, that the male coloration of H. squamipinnis is not markedly different from that of H. victorianus (see Greenwood 1962, p. 157), and that several Lake Victoria species of the ‘ serranus’ group show basically similar male breeding coloration. Haplochromis squamipinnis also resembles H. venator Greenwood, the sole pisci- vorous Haplochromis species of Lake Nabugabo (Greenwood 1965b). Skull mor- phology in H. venator is rather more like that of H. mento than of H. squamipinnis, and there are some differences in gross body form as well as some specific morpho- metric ones (especially the deeper body and wider interorbital of H. squamipinnis). Haplochromis venator, of course, lacks the scale rows on its dorsal and anal fins. Certainly the similarities between this species and H. squamipinnis are no closer than those between H. sguamipinnis and the Lake Victoria species considered above. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2 : 379 I.B.P. Jetty 1972.0.2 : 380 I.B.P. Jetty 1972.6.2 : 381-386 Kashaka Bay 1972.6.2 : 387-395 Various localities 1933.2.23 : 449 Collected by Worthington 1933-2.23 : 439-443 Collected by Worthington 1933.2.23 : 450-451 Collected by Worthington 1933.2.23 : 452-455 Collected by Worthington 1933.2.23 : 444-448 Collected by Worthington Haplochromis petronius sp. nov. (Text-figs. 32-34) Hototype. A male 85-5 mm standard length, BMNH reg. no. 1972.6.2:1, from Kashaka Crater, Lake George. The trivial name, from the Latin, meaning ‘ of, or pertaining to rocks’, refers to the usual habitat of this species in Lake George. 210 P. H. GREENWOOD DESCRIPTION. Based on 25 specimens, 67-0—88-0 mm standard length (including the holotype), all from Lake George. Depth of body 33:2-38-3 (M = 35-8) per cent of standard length, length of head 31°8-35:°8 (M = 33°8) per cent. Dorsal head profile straight, sloping steeply at an angle of ca 50°-55° with the horizontal. Preorbital depth 14:5-19-3 (M = 17-6) per cent of head, least interorbital width 19:2—25'0 (M = 22-1) per cent. Snout length 0-7 — 0-9 (mode 0-8) of its breadth, 28-0-34:6 (M = 31-2) per cent of head, diameter of eye 25-0-30-°8 (M = 28-0), depth of cheek 24:5-30-2 (M = 27-2) per cent. Caudal peduncle 12-6-16-8 (M = 15-7) per cent of standard length, 1-1-1-4 (modal range 1-2-1-4) times as long as deep. Fic. 32. Haplochromis petronius. Holotype. Jaws equal anteriorly, mouth horizontal, the lips thickened. Posterior tip of the maxilla reaching a vertical through the anterior margin of the pupil, rarely not reaching so far posteriorly. Length of upper jaw 34:2-41:6 (M = 36:8) per cent of head, length of lower jaw 34-8—41-5 (M = 37-8) per cent, I-3-1-7 (modal range I:3-1°5) times its breadth. Gill rakers. The lower 1-3 rakers of the first gill arch are reduced, the others are usually short and stout, but rather slender in some fishes. There are 7 or 8 (mode) rarely g, rakers in the outer row on this arch. The pseudorakers are very well developed and are transversely aligned so as to link the inner and outer rows of true gill rakers on the first arch. Scales. Ctenoid; lateral line with 30 (f.3), 31 (f.4), 32 (f.15) or 33 (f.2) scales, cheek with 3 or 4 rows. Scales on the nape and chest are small, the latter rather deeply embedded. The transition between the larger ventral belly scales and the much smaller thoracic scales is abrupt, and occurs at about the level of the pectoral fin insertion. This abrupt type of size transition is unusual in Haplochromis (see Greenwood 1971). Another unusual character of the squamation in H. petronius LAKE GEORGE HAPLOCHROMIS SPECIES 211 is the presence of a small naked area (equivalent to about 4 or 5 scales) immediately anterior to the insertion of the first dorsal fin spine. There are 5-7 (mode 6) scales between the upper lateral line and the dorsal fin origin, and 7~10 (modal range 8-9) between the pectoral and pelvic fin bases. Fins. Dorsal with 15 (f.3), 16 (f.21) or 17 (f.1) spinous and 8 (f.1), 9 (f.21) or to (f.3) branched rays. Caudal strongly subtruncate, almost rounded ; scaled on its basal third to half. Pectoral fin 26-5-32:4 (M = 28-1) per cent of standard length, 75-0—97-8 (M = 82:0) per cent of head length. Pelvics with the first ray somewhat produced in both sexes. Teeth. In the outer row, the teeth posteriorly in the premaxilla are unicuspid and caniniform ; anteriorly in this jaw and throughout the lower jaw the teeth are stout, unequally bicuspid (often weakly so), and have the crown incurved (text-fig. 33). These teeth are cylindrical in cross-section, with the upper part of the crown compressed. Occasionally, a few unicuspids occur anteriorly in the upper jaw, interspersed among the usual bicuspids. Also, in a few fishes the posterior premaxil- lary teeth are slender and bicuspid, not unicuspid as is usual for teeth in that position. There are 36-50 (M = 42) teeth in the outer premaxillary row. Fic. 33. H. petronius. Premaxillary teeth (left), anterolateral in position. Viewed laterally. Scale = 0-5 mm. Inner teeth are usually tricuspids in fishes < 75 mm standard length (and in a few larger individuals) but most fishes > 75 mm long have an admixture of tricuspid, weakly tricuspid and unicuspid teeth. There are 2 or 3 rows of inner teeth in the upper jaw, and 2 (less frequently 3) in the lower jaw. OstEoLoGy. The neurocranium is of the generalized Haplochromis type (see p. 147) with a moderately decurved preorbital profile. The lower pharyngeal bone is relatively stout, its dentigerous surface equilateral. The teeth are cuspidate and compressed, with those of the median rows noticeably coarser (text-fig. 34). In a few specimens some or all of the median teeth are sub- molariform. There are ca 20-24 rows of teeth. Vertebral counts in the 12 specimens radiographed are 29 (f.10) and 30 (f.2), comprising 12 (f.1), 13 (f.10) or 14 (f.1) abdominal and 15 (f.1), 16 (f.8) or 17 (f.3) caudal centra. COLORATION IN LIFE. Males, adult but not sexually active : the flanks are greenish- yellow, tinged with blue, the blue concentrated along the scale margins ; dorsally 212 P. H. GREENWOOD the colour changes to greenish-violet. The belly and chest are white, the branchio- stegal membrane light grey. The lips are turquoise, and there are strong tinges of turquoise on the basically grey snout and cheeks. ue AK Se < OOO +o peyote S X Sy N' \ ree! Ooo. p afO.8. 9 ooo 0. ‘ Ye Pg Awe ZL evan 2 Y/Y “ch 900,00 oy ara hg 00// 00 ~URO 8» 94 oY A 0\=8 6 0'0/,/ 0 OAS) of 92 0%, as cere) Oo 0 / Fic. 34. H. petronius. Lower pharyngeal bone, occlusal view. Scale = 1:0mm. The dorsal fin is light sooty-grey, with the base greenish-blue, and the lappets red (as is the margin of the soft part) ; the soft part also has a scattering of well-marked, elongate, deep red streaks and spots. The anal fin is also sooty-grey (sometimes flushed with pink or red, perhaps a correlate of sexual activity) with a light-blue area along its base and a narrow scarlet outline to its margins ; there are as many as eight bright orange ocelli on the posterior part of this fin. The caudal is dark grey, outlined in red, this marginal band expanding at the posterodorsal and ventral angles of the fin ; deep maroon streaks occur between the rays. The pelvic fins are black or sooty, the spine and first ray are, however, bluish-white. Females : no live females have been observed. PRESERVED COLORATION. Only males are available, all are adult. The ground colour is uniformly light grey-brown ; in some specimens there are traces of up to 7 dark but faint vertical bars across the flanks and caudal peduncle, those on the flanks extending from the dorsal body outline almost to the ventral body margin. The snout is dark grey to dusky, the lower jaw dusky-grey and the lips pale. The ventral aspects of the cheek and operculum are of variable duskiness, almost black in some fishes but only a dark brown in others. The branchiostegal membrane is greyish- sooty. Cephalic markings comprise a usually distinct and intense lachrymal stripe, and 2 transverse bars, of variable intensity, across the snout. The dorsal fin is dark grey to sooty, the membrane between the last few branched rays maculate. The caudal is grey, maculate distally and dark grey, almost black LAKE GEORGE HAPLOCHROMIS SPECIES 213 basally. The anal fin is greyish to sooty, the pelvics variable, from dusky to black but with the spine and first ray much lighter. Ecotocy. Habitat. Haplochromis petronius is the only Lake George Haplo- chromis species that appears to have a clearly circumscribed habitat. With a few exceptions (see below) the species has been found only in a rocky bay situated im- mediately behind the village of Kashaka. This bay is an old volcanic crater, one wall of which has collapsed and thus connected the crater with the lake. The bay is roughly ovoid in outline, its greatest and least axes being about 1-3 and 0-8 km. At its centre the water is some 6m deep, but around the margins it is between I and 3m. This marginal area has a rough substrate composed of rocks and stones derived from the crater walls. Some plant debris (including dead trees) lies among the rocks which, in places, are also covered lightly by a thin slick of organic mud. The bay is sheltered from all directions and has a relatively narrow entrance. Haplochromis petronius is found only over the marginal rocky area of Kashaka Bay ; nets set in the deeper central area, either at the bottom or floating at the surface, caught no H. petronius, although other fishes (both cichlid and non-cichlid), including species found inshore, were caught. The species is common in catches from its habitat ; other Haplochromis species inhabiting the same region include H. angustifrons, H. elegans, H. aeneocoloy and H. schubotziellus, but none occurs in such abundance as does H. petronius. Outside Kashaka bay, H. petronius is rarely encountered ; it is probably significant in this connection that no other areas of the lake have a similar rock-boulder sub- strate. The very few H. petronius caught in the main lake are either from near the papyrus fringe over a mud-bottom or from over a sandy substrate on an exposed shore facing the lake centre. Food. Insects, both larval and emergent, seem to be the commonest food organisms in the diet of H. petronius, but the sample I examined was small (20 specimens). Breeding. The specimens I have examined are all males, and indeed I have been unable to catch any females, despite intensive fishing in the area. Thus, it would seem that there is a definite segregation of the sexes, and probably at all times of the year, itself an unusual phenomenon amongst Haplochromis species. The smallest fish examined (67 mm standard length) is a ripening male, although another specimen 69 mm standard length is probably immature. Distribution. Known only from Lake George. DIAGNOSIS AND AFFINITIES. Taken in its totality, the appearance of H. petronius is highly characteristic, and readily distinguishes this species from other Haplochromis in Lake George. The small chest and nape scales, the abrupt size transition between the thoracic and ventral body squamation, and the small scaleless area before the first dorsal fin spine are trenchant diagnostic features. However, these particular characters are also found in a species known from Lake Edward, viz. H. pharyngalis Poll (see also p. 177). The overall morphology of H. pharyngalis is also very like that of H. petronius, especially the steep head profile, the horizontal mouth with its thickened lips and the near-rounded caudal fin outline. 214 P. H. GREENWOOD The principal interspecific differences are twofold. First, in H. pharyngalis the lower pharyngeal bone is greatly enlarged (massive in I of the 3 known specimens), and there are several rows of enlarged molariform or near-molariform teeth (see Poll 1939a, p. 46, fig. 26). Second, the nuchal and thoracic scales (particularly the former) are relatively smaller in H. pharyngalis. In addition, there are fewer (i.e. 6) gill rakers in H. pharyngalis, and there is a naked area, about 1 scale row deep, on the ventral margin of the cheek. With only 3 specimens of H. pharyngalis available for comparison and with no information on their live colours, it is difficult to assess the precise relationship between the species. Certainly the differences in pharyngeal bone development (and correlated dental differences) and in the size and shape of the apophysis for the upper pharyngeal bones seem well marked. But, these are of similar nature to those distinguishing the George-Edward populations of Astatoreochromis alluaudi from those of Lake Victoria (see Greenwood 1959a, and especially 1965a). That H. pharyngalis shows a range of pharyngeal bone enlargement (even in only 3 speci- mens), and that H. petrvonius exhibits incipient molarization of the pharyngeal teeth, adds to the impression of close relationship between the species. The question can only be pursued when larger samples of H. pharyngalis are available. An even closer resemblance exists between H. petronius and H. wingatit (Blgr.), a species of the Nile and Lake Albert (see Greenwood 1971 for a revision of this often misidentified species). Again, comparisons are hampered by the small number of specimens available. Only 2 well-preserved and 1 poorly preserved specimen of H. wingatii are known, all much smaller than the smallest H. petronius available, and there are no data on their live colours. Morphometrically, H. wingatii and H. petronius are indistinguishable except for the longer pectoral fin in H. petronius (75:0-97:8, M = 86-4 per cent head length, cf. 65:0-66°5 per cent for H. wingati1) and a higher vertebral count (29 or 30, cf. 28). Two slight interspecific differences are the absence of a naked predorsal area in H. wingatii and the existence of a naked strip below the cheek scales in that species. There is also a difference in the dentition. In the holotype of H. wingatiz (53 mm standard length) the majority of the outer teeth in both jaws are slender, strongly recurved unicuspids. The outer teeth in H. petronius are relatively coarser and, except in the largest fishes, are all bicuspid. Even in large individuals the pre- dominant tooth form is the bicuspid. The two other H. wingatit specimens are smaller and do have bicuspid teeth (as is usual for small individuals of species with a unicuspid definitive dentition). The bicuspids in these specimens are slender (like the few bicuspids of the holotype) and thus are unlike the bicuspids of H. petronius. Resemblances between H. wingatiu, H. petronius and H. pharyngalis are striking, particularly since they involve, principally, details of squamation not found in other species of Lakes Edward, George or Victoria, or even in the fluviatile Haplochromis species of east Africa. Reduced pectoral and nuchal squamation is known from other Haplochromis and Haplochromis-like species, but in all these fishes it is asso- ciated with rheophilic habits (Greenwood 1954; Thys van den Audenaerde 1963). The habitats of the 3 species under consideration are far removed from the torrential, LAKE GEORGE HAPLOCHROMIS SPECIES 215 and thus it is difficult to attribute the resemblances in squamation to environmentally induced parallelism. If, therefore, the similarities are a reflection of phyletic affinity, then H. wingatii, H. petronius and H. pharyngalis would appear to be of a lineage distinct from all other Haplochromis in Lakes Victoria, Edward and George. The implications of this conclusion will be considered later (p. 235). STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.2. : r (Holotype) Kashaka Crater Bay 1972.6.2 : 2-10 (Paratypes) Kashaka Crater Bay 1972.6.2 : 11-21 (Paratypes) Kashaka Crater Bay 1972.6.2 : 22 (Paratype) North end of Kankurunga Island 1972.6.2 : 23 (Paratype) Caught over sandy shallows 1972.0.2 : 788 (Paratype) Kashaka Crater Bay 1972.6.2 : 811 (Paratype) Kashaka Crater Bay 1972.6.2 : 293 Sandy shoal Haplochromis eduardianus (Blgr.), 1914 (Text-figs. 35-37) Schubotzia eduardiana Boulenger, 1914, in Wiss. Evgebn. Deuts. Zentral-Afrika Exped., 1907-1908, Zool. 3: 258-259 (Lake Edward). Schubotzia eduardiana : Regan, 1921, Ann. Mag. nat. Hist. (9), 8 : 639. Schubotzia eduardiana : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 340. All these references are to specimens from Lake Edward; the species was not discovered in Lake George until recently. NOTE ON THE ALTERED GENERIC STATUS OF THE SPECIES. Boulenger (1914) defined the monotypic genus Schubotzia solely on the basis of its dental morphology. The outer jaw teeth are unusual (text-fig. 36). The crown is somewhat expanded relative to the cylindrical neck, has the tip distinctly rounded and strongly incurved so as to lie almost horizontally. In outline (i.e. as a flattened object) the tooth is paddle-shaped. The inner teeth, by contrast, are typical tricuspids but are restricted to I or 2 rows in each jaw. In other characters, Schubotzia does not differ from Haplochromis. The lower jaw, especially the dentary, is deep and stout, and the premaxilla is rather inflated. However, both these characteristics can be seen in other Haplochromis species (e.g. H. nigricans and H. obesus of Lake Victoria), although the condition could not be described as a common one ; it is usually associated with a specialized dentition. The lower jaw of Schubotzia is slightly overhung by the upper jaw, again an unusual condition but one found in some Haplochromis species (e.g. H. xenognathus of Lake Victoria, a species with an unusual dentition, albeit one quite unlike that of Schu- botzia eduardiana ; see Greenwood 1957). The teeth of Schubotzia are outstandingly different when compared with those of other Lakes Edward and George cichlids (text-fig. 36) and especially with those of the 216 P. H. GREENWOOD Haplochromis species known to Boulenger in 1914. But, when the Schubotzia tooth type is seen against the wide range of tooth morphology found within the genus Haplochromis as a whole (or just a segment like the species of Lake Victoria), then it does not seem to be so unusual. In my opinion, the morphological differences separating Schubotzia from Haplo- chromis are relatively slight, and certainly less than those distinguishing Haplo- chromis from the Lake Victoria monotypic genera Hoplotilapia and Platytaeniodus (or those genera from one another). To retain Schubotzia eduardiana in a separate and monotypic genus serves only to hide its close phyletic relationship with Haplochromis. Thus, I would favour classifying this taxon with its closest relatives, that is in the genus Haplochromis. Rosen and Bailey (1963, p. 6) have succinctly stated the pragmatic and theoretical difficulties associated with the generic concept. Particularly they stress the often phyletically misleading results of undue emphasis placed on one or two outstanding morphological differences as generic criteria. I am fully in agreement with these authors’ support for a wider use of the subgenus to indicate morphological divergence without losing sight of phyletic relationships. Reducing the genus Schubotzia to subgeneric status could well meet the requirements of this particular case. However, there is a great need for new and careful consideration of the generic and infrageneric classification of the ‘ genus ’ Haplochromis as currently defined. Such a study must be based on phyletic principles and must test the phyletic integrity of what might well be a polyphyletic taxon. It is my intention to undertake just such a study ; for the moment, I prefer not to establish Schubotzia formally as a subgenus of Haplochromis. Lectotype. A specimen 71:0 mm standard length, BMNH reg. no. 1914.4.8 : 35, from Lake Edward. This specimen is eviscerated and its sex cannot be determined ; the absence of dark pigment on the body suggests that it is a female (see below p. 220). Three paralectotypes are in the collections of the Berlin Museum ; all are from Lake Edward. DESCRIPTION. Based on 20 specimens, 50:5~79:0 mm standard length, from Lake George. A smaller fish, 35 mm standard length is not included in the morphometric section, since it is distorted, but certain features of its dentition are considered on p. 218. Depth of body 30:5-36-7 (M = 33:5) per cent of standard length, length of head 31°5-34:3 (M = 33:1) per cent. Dorsal head profile gently decurved, but sometimes straight, sloping at an angle of 30°-35° with the horizontal. Preorbital depth 11-2-15-4 (M = 13:3) per cent of head, least interorbital width 23:0-28°3 (M = 25:9) per cent. Length of snout o-8-o-9 of its width, 23-0-28-3 (M = 25-9) per cent of head, eye diameter 29:6—36-0 (M = 32-8) per cent, depth of cheek 16-7—22-2 (M = 20-0) per cent. Caudal peduncle 14-7-19°-5 (M = 16-9) per cent of standard length, 1-2-1-7 (modal range 1-3~1-5) times longer than deep. | | | LAKE GEORGE HAPLOCHROMIS SPECIES 217 Mouth horizontal, lips slightly thickened ; lower jaw shorter than the upper when the mouth is closed. Length of upper jaw 30-8—36-0 (M = 32-8) per cent of head, length of lower jaw 29:2-35:I (M = 33-1) per cent, I-o-1-4 (mode 1-1) times longer than broad. Posterior tip of the maxilla reaching a vertical through the anterior margin of the eye, rarely not quite reaching that level. Lib MLM CZ ors ue es \ A Fic. 35. Haplochvomis eduardianus. Lake George specimen ; a male. Gill rakers. The lowermost I or 2 rakers are reduced, the others are relatively slender although as many as 3 of the lower rakers may be relatively stout. The pseudorakers are well developed and prominent, but are short and stout. There are 8 or g (rarely 10) gill rakers in the outer series on the lower part of the first gill arch. Scales. Ctenoid ; lateral line with 31 (f.8), 32 (f.11) or 33 (f.1) scales, cheek with 2 or 3 (mode), rarely 4 rows. Five to 7 (modes at 5 and 54) scales between the upper lateral line and the dorsal fin origin, 6-8 (mode 7) between the pectoral and pelvic fin bases. Fins. Dorsal with 14 (f.1), 15 (f.11) or 16 (f.8) spinous and 8 (f.6), 9 (f.13) or 10 (f.1) branched rays. Anal with 3 spines and 7 (f.2), 8 (f.16) or 9 (f.2) branched rays. Pectoral fin 26:5-32-4 (M = 28-1) per cent of standard length, 80-5—98-o (M = 82:0) per cent of head. Caudal subtruncate, scaled on its basal half. Pelvics with the first ray slightly elongate. Teeth. With one exception in the material studied, the shape of the outer teeth in both jaws is remarkably uniform. Each tooth is unicuspid, with a flattened crown that is almost half the total length of the tooth. The whole crown is strongly in- curved (especially in teeth situated anteriorly and anterolaterally), and its tip is broadly rounded (text-figs. 36 & 37). The neck of the tooth is cylindrical and, compared with that of a typical bicuspid tooth, much stouter. If a tooth were straightened out so that the crown and neck are in one plane, then it would have the outline of a paddle. 218 P. H. GREENWOOD The exceptional specimen mentioned earlier (a fish 50-5 mm standard length) differs from the others only in having a few small tricuspid teeth intercalated among the typical teeth posterolaterally in the upper jaw. y Fic. 36. H. eduaydianus. Premaxillary teeth (right) anterior in position. Viewed from below and a little laterally, to show labial aspect of the teeth. Scale = 0-25 mm. In the smallest available specimen (35 mm standard length), most teeth are like those of larger fishes, but in both jaws there are a few teeth with traces of a small lateral cusp. This minor cusp is not separated from the major one by a distinct gap, asin typical bicuspids. Instead, the demarcation between cusps is more in thenature of a narrow, V-shaped groove. Because a minor cusp is present, the outline of these teeth is not rounded but is rather bluntly oblique. Posteriorly in the lower jaw of this small fish there are a few tricuspid teeth. There are 40-52 (M = 48) teeth in the outer row of the upper jaw; the number not showing any correlation with the fish’s size. Inner teeth (text-fig. 37) in both jaws are tricuspid, with the crown compressed and strongly incurved ; arranged in 2 or rarely 3 rows in the upper jaw and a single (rarely double) row in the lower jaw. Virtually no interspace exists between the outer tooth row and the first row of inner teeth. Fic. 37. H. eduardianus. Dentary teeth (left), anterolateral in position. Viewed from medial aspect to show lingual face of teeth. Scale = 0-25 mm. LAKE GEORGE HAPLOCHROMIS SPECIES 219 OsTEoLoGy. The neurocranium is of the generalized Haplochromis type with a moderately decurved preorbital profile ; that is, a skull of the type found in H. limax and H. petronius rather than of the type in H. macropsoides and H. elegans. The premaxilla is a stout and inflated bone (especially the dentigerous arm) and resembles a more extreme form of the premaxilla found in H. limax. Stoutness and compact- ness also characterize the lower jaw. Both dentary and articular are stout, deep bones, and elements of the suspensorium are so arranged and proportioned that the dentary tip lies behind the vertical through the premaxillary symphysis. Presumably the general stoutness of the jaws is at least partly correlated with the presence of stout teeth. The lower pharyngeal bone is relatively stout, although its teeth are fine (except for a few coarser posterior teeth in the middle tooth rows of larger fishes). There are ca 24-26 rows of teeth arranged over a dentigerous area about 1-3 times broader than it is long. Vertebral counts in the Io specimens radiographed are 29 (f.7) and 30 (f.3), com- prising 13 (f.7) or 14 (f.3) abdominal and 15 (f.2), 16 (f.6) or 17 (f.2) caudal elements. COLORATION IN LIFE. Adult males have a silvery blue-grey ground coloration, shading to white on the thoracic region ; many flank scales have deep red-brown centres, the intensity of the colour and the number of reddened scales correlated with the degree of sexual activity. Dark cephalic markings are always well developed. A thin stripe crosses the snout at the level of the lower orbital margin, and a much broader band runs between the upper orbital margins. Often this upper band is interrupted medially. A broad lachrymal stripe is continued above the eye as a wide blotch, the blotches from each side usually meeting medially. Behind this mark is another, this time in the form of a broad band crossing the nape from an origin at about the level of the upper opercular margin’s midpoint. The dorsal fin is hyaline with bright scarlet lappets and a diffuse scarlet flush between the spines; this flush becomes concentrated into discrete blotches and spots between the branched rays. The anal is dusky on its proximal half, pinkish to scarlet distally ; the almost round ocelli (I or 2) are very large and are yolk- yellow in colour. The caudal is entirely suffused with bright scarlet, although its ground colour is greyish-green. The pelvic fins are dusky but have scarlet streaks between the rays posteriorly on the proximal half of the fin. Females (at all stages of sexual activity): the ground colour is dark silver-grey shading to white on the belly and chest. Cephalic markings are usually absent, but if visible are a faint replica of those in the male (see above). The dorsal fin is hyaline, with a narrow scarlet marginal band on the soft part, and thin red streaks on the lappets of the spinous part ; in some individuals red streaks are present be- tween the spines and rays. The caudal is hyaline with an overall pink flush, some- times intensified to scarlet streaks between the rays and along the posterior margin of the fin. The anal is very pale yellow, with 2 deeper yellow spots (not ocelli) in the position of ocelli in males. The pelvics are whitish-hyaline. PRESERVED COLORATION. Adult males: the ground colour is silver-grey, the flanks crossed by up to 10 vertical bars of variable intensity and definition ; the 13 220 P. H. GREENWOOD first 2 bars lie immediately behind the opercular margin, the last on the caudal peduncle. Dorsally, the bars extend to the body outline and at least anteriorly may extend onto the dorsal fin membrane ; ventrally the bars do not reach much below the level of the pelvic fin insertion. Cephalic markings are very intense. Across the snout are 2 parallel bars, the upper generally twice the width of the lower. A broad lachrymal stripe is present, and there is a well-defined bar or triangular blotch extending towards the midline immediately behind the orbit. Posterior to this mark is another, but strap-shaped one which lies anterior to the dorsal fin origin. This bar appears to be a medial extension of the first vertical bar on the flank. All fins, except the pectorals and pelvics, are yellowish, the basal region of the caudal often dusky. The pectorals are hyaline, and the pelvics black on the outer third, otherwise dusky. Females are silvery grey or yellowish, with all fins yellow or hyaline. No cephalic markings are visible except, in some specimens, for a faint darkening below the eye, that is, in the position of a lachrymal stripe. Ecotocy. Habitat. The species is found in most inshore areas of the lake, over sand and mud bottoms, and in both exposed and sheltered localities. It has also been caught in the near-shore areas of the open lake but not further than about roo m from the nearest land (in this case a small, reed-fringed island). No specimens have been collected in midlake or other distinctly offshore regions. Food. The diet of H. edwardianus is still unknown. Most specimens examined had nothing recognizable except for a few sand grains and a few macrophyte frag- ments in any part of the gut. The highly specialized dentition would suggest equally specialized feeding habits. By analogy with the similarly shaped teeth of Plecodus spp., one might suspect a similar diet of fish scales (Marlier & Leloup 1954). Since the intestine of H. eduardianus is short (about half the length of the body) a vegetarian diet is almost certainly ruled out. As so many (ca go per cent) of the specimens examined have nothing in the guts it seems probable that, whatever the food, it is rapidly digested. Breeding. Haplochromis eduardianus is a female mouth brooder. Sexual maturity is reached at a standard length of ca 55 mm, and both sexes attain the same maximum length. All of the 9 sexually active females examined have the right ovary much larger than the left one, and in some individuals only the right ovary is developed. Distribution. Lakes Edward and George and the Kazinga Channel. DIAGNOSIS AND AFFINITIES. Haplochromis eduardianus is distinguished from all other Haplochromis species in Lakes George and Edward by the morphology of its teeth. This peculiar tooth form led Boulenger (1914) to place the species in a distinct and monotypic genus. Reasons for not accepting Boulenger’s classification are discussed above (p. 215). In essence, I argue that to place H. eduardianus in a monotypic genus (i.e. Schubotzia) is to obscure its phyletic relationships. That the dentition (both oral and pharyngeal) of Haplochromis group cichlids is, in an evolu- tionary sense, easily modified can be seen readily amongst the component species LAKE GEORGE HAPLOCHROMIS SPECIES 221 of the Lake Victoria Haplochromis species flock (see Greenwood 1965c for summary and further references). Indeed, in at least one instance the morphological sequence leading from generalized to highly specialized dentitions is still preserved among the extant species of that lake (Greenwood 1957). There are also examples where intraspecific variability is such that if only the extreme condition was known, and the criteria for generic status were based solely on morphological ‘ gaps’, then the species would have to be accorded generic status (see Greenwood, op. cit., p. 96, with reference to H. xenognathus). The reality of a morphological gap like that between the H. eduardianus tooth form and the shape of the teeth in any other Haplochromis species cannot be denied ; the difficulty lies in attempting to interpret the significance of the gap. The existence within a single species flock of such examples as H. xenognathus (and others where different species bridge a morphological gap) warns against hasty action that could obscure the essence of a phyletic classification ; that is, the demonstration of relationships as well as divergence. Is there any reason to suppose that, phylo- genetically, the current evolutionary end-point seen in ‘ Schubotzia’ eduardiana and Platytaeniodus degeni or Hoplotilapia retrodens (both Lake Victoria monotypic genera ; see Greenwood 1956a) is any different from that represented by H. xenog- nathus? Certainly all three monotypic genera have as their nearest relatives a species of Haplochromis. That such a species can be found for H. xenognathus (Greenwood 1957) but not for Schubotzia, Playtaeniodus or Hoplotilapia may be more a reflection of a past epigenetic situation than of phyletic history. STUDY MATERIAL Register number BMNH Locality : Lake George 1972.6.5 : I-3 Tufmac Bay I972.0.5 : 4 Kashaka Crater 1972.6.5 : 5-7 Small island north of Kankurunga Island 1972.0.5 : 8-13 Various localities 1972.6.5 : 14-20 Various localities NON ENDEMIC HAPLOCHROMIS AND HAPLOCHROMIS-GROUP SPECIES IN LAKE GEORGE Haplochromis nubilus (Blgr.), 1906 (Text-fig. 38) Haplochromis nubilus (part) : Trewavas, 1933, J. Linn. Soc. (Zool.), 38 : 329. Most of the Lake George specimens referred to H. nubilus by Trewavas (1933) were misidentified, but two specimens (BMNH reg. nos. 1933.2.23 : 285-286) from Worthington’s (1932) station 618 appear to be of this species. They are not included in the description given below. Four specimens collected by the I.B.P. team can be identified with certainty since all are sexually active males and their live colours were recorded. These fishes came from a catch made close inshore, near reeds, in a bay of Akika Island. 222 P. H. GREENWOOD A summary of morphometric characters is given below. It is based on 3 speci- mens only because the fourth is distorted and damaged. The latter specimen is, however, used in the description of the teeth and for fin ray and scale counts. S.L. Deptht+ Headt PO% IO0% Snt% Eye% Cheek% Lij% Uj% CPt 67-0 7 Ses tS 27-4 207 2:3 30:3) 2-5 es 4-5 eg 4o 164 7O'5 39°3 341 16°7 27:0 313 29°2 22'9 37°5 37'5 16°5 74°5 37°5 34:2 156 24°77 284 27°4 23°5 391 35°3 161 t+ =per cent of standard length. % =per cent of head. Lj =lower jaw, Uj =upper jaw. Caudal peduncle 1-3~-1-4 times longer than deep. Dorsal head profile sloping steeply at ca 40°—45° with the horizontal, its outline straight except for a marked concavity above the orbital region. Upper margin of the orbit distinctly below the level of the dorsal profile. Mouth horizontal, lips somewhat thickened ; posterior tip of the maxilla reaching a vertical through the anterior margin of the orbit. Jaws equal anteriorly, the lower ca 1-3 times longer than broad. Snout slightly broader than long. Fic. 38. Haplochromis nubilus. A Lake Victoria specimen. From Boulenger, Fishes of the Nile. Gill vakers. Eight or 9 on the lower part of the first arch ; the lower 2 rakers are reduced, the remainder are relatively stout. Pseudorakers are present and discrete, but are rather low and small. Scales. Ctenoid; lateral line with 31 (f.2) or 32 (f.2) scales, cheek with 3 or 4 rows. Five or 54 scales between the upper lateral line and the dorsal fin origin, 6 between the pectoral and pelvic fin bases. Fins. Dorsal with 15 (f.4) spinous and 9 (f.2) or ro (f.2) branched rays, anal with 3 spines and 8 (f.2) or g (f.2) branched rays. Pectoral 26-2—28-2 per cent of standard LAKE GEORGE HAPLOCHROMIS SPECIES 223 length, 76-5-83-3 per cent of head. Pelvics with the first ray variably produced, markedly elongate in one fish. Caudal distinctly subtruncate, almost rounded ; scaled on its basal third to half. : Teeth. Posteriorly in the upper jaw the outer teeth are large and unicuspid. Elsewhere in this jaw, and throughout the lower jaw, the teeth are unequally bi- cuspid and moderately stout, or are an admixture of such teeth with slender uni- cuspids. Bicuspid teeth have the major cusps equilateral in outline and barely incurved ; some teeth have faint indications of a flange on one aspect of the cusp (see p. 151). : There are 40-44 teeth in the outer premaxillary row. Inner teeth in both jaws are small and tricuspid, are arranged in 2 or 3 rows in the upper jaw, and in I or 2 rows in the lower. OstEoLocy. The syncranium of H. nubilus is typically that of a generalized Haplochromis species. The lower pharyngeal bone is relatively slender, its teeth compressed and bicuspid, with those of the median rows slightly coarser (especially posteriorly). There are ca 24-28 rows of teeth arranged on a dentigerous area that is ca 1-2-1-3 times broader than long. Vertebral count is 29 (comprising 13 abdominal and 16 caudal centra) in the 4 fishes radiographed. COLORATION IN LIFE. Adult males have a highly characteristic velvety-black coloration that is virtually uniform over the whole body. The dorsal fin also is black except for a narrow scarlet margin along its entire length, and some scarlet spots and streaks on the soft part. A scarlet flush covers most of the anal fin, although the spinous part may be a little dusky ; the ocelli are yolk-yellow. Proxi- mally the caudal fin is black, and this dark colour may extend along the centre of the fin almost to its margin. The margin is bright scarlet, the colour expanding at the posteroventral margin of the fin. The pelvics are jet black. The live colours of females from Lake George are unknown; in Lake Victoria the body is deep olive-green and all the fins are greyish-green. Preserved males are either uniformly black on the body or the dorsum (above the midlateral line) may be lighter (i.e. a deep brown) and crossed by 4 or 5 rather faint vertical bars. The dorsal fin is black except for a pale (yellowish-orange) margin. Almost the entire anal is pale yellow, although there is a faint and narrow dark band basally. The caudal is dark proximally and between the middle rays, but other- wise it is yellowish. The pelvics are black. Ecotoey. Little can be said about H. nubilus in Lake George. The specimens I examined came from a shallow inshore area close to emergent vegetation. A similar habitat seems to be the preferred one for H. nubilus in Lakes Victoria and Nabugabo (Greenwood 1965b). No data are available on the feeding and breeding habits of the species in Lake George. Lake Victoria populations are female mouth brooders, and have a rather omnivorous diet in whichv laral insects and small Crustacea predomi- nate. 224 P. H. GREENWOOD Distribution. Lakes Victoria, Kyoga, Nabugabo, Edward and George, and in many rivers and streams connected with these lakes. To date no specimens have been caught in the Kazinga Channel. Dracnosis. Morphologically, H. nubilus closely resembles H. aeneocoloy (see p- 154) in nearly all characters, especially morphometric ones. Pseudorakers are present in H. nubilus but not in H. aeneocolor, and the caudal fin is virtually rounded in H. nubilus (but truncate in H. aeneocolor). There are some dental differences between the species but these are not particularly trenchant. However, in H. nubilus flange development on the major cusp is cer- tainly less common than in H. aeneocolor, and the flange, when developed, is less prominent. Also, judging from Lake Victoria H. nubilus, it seems probable that unicuspid outer teeth occur more frequently in this species than in H. aeneocolor. In life, male breeding coloration is certainly diagnostic. ASTATOREOCHROMIS Pellegrin, 1903 This genus is readily distinguished from Haflochromis by the higher number of anal fin spines: 4 or more, usually 5. For a full diagnosis of the genus see Greenwood 1959a and 1965a, b. Astatoreochromis alluaudi Pellegrin, 1903 (Text-fig. 39) Astatoreochromis alluaudi : Trewavas, 1933, J. Linn. Soc. (Zool.), 38: 321 (I specimen from Lake George). Astatoreochromis alluaudi occidentalis Greenwood, 1959, Bull. Br. Mus. nat. Hist. (Zool.), 5 3174-175. Very few specimens of A. alluaudi have been collected from Lake George by the I.B.P. team, despite an intensive fishing effort. Four specimens (45-104 mm standard length) are available for study. Because A. alluaudi is easily recognized by its having 4 or more anal spines and greatly enlarged pharyngeal bones, and because the Lake George fishes do not differ greatly from those described elsewhere (Greenwood 1959a) no full description is necessary. The live colour of Lake George fishes (previously unknown) is identical with that described for the Lake Victoria populations (see Greenwood, op. cit., p. 172). The lower pharyngeal bone and dentition of A. allwaudi from Lakes Edward and George are much less massive than those from comparable-sized fishes in Lake Vic- toria. This led me to describe a subspecies, A. a. occidentalis, for these western lakes (Greenwood 1959a). Subsequent research, however, strongly suggests that the degree of pharyngeal bone development (and of tooth molarization) is probably under direct environmental control (Greenwood 1965a). In other words, the differ- ences between Lake Victoria and Lake Edward-George fishes is not genetically determined. Thus it seems inadvisable to continue recognizing two subspecies. Lower pharyngeal bones and teeth in the 4 Lake George fishes fit broadly into the reductional pattern described for fishes from lakes other than Victoria and Kyoga LAKE GEORGE HAPLOCHROMIS SPECIES 225 (Greenwood 1959a). However, the Lake George fishes seem to have rather more massive bones (and greater molarization of the teeth) than do fishes from Lake Nakavali. In the two smallest Lake George specimens (45 and 46 mm standard length) the bone is only a little less developed than in a comparable-sized specimen from Lake Victoria, and there is equal molarization of the teeth (see fig. 3 top right, in Greenwood 1959a). The 80mm standard length Lake George fish has a bone comparable with that of the 123mm standard length Lake Nakavali specimen figured (op. cit.), but the 104 mm standard length fish has a relatively less massive bone which is comparable with the same specimen from Lake Nakavali. Fic. 39. Astatoreochromis alluaudi. A Lake Victoria specimen. From Boulenger, Fishes of the Nile. The ratio of head length to pharyngeal bone width (measured from tip to tip of the upper arms) is in the range 2-9-3-0 (see Greenwood, of. cit.). Vertebral counts for the 4 specimens are 13 + 16 (f.2), 13 + 15 (f.1) and 14 + 15 (f.1). Ecotocy. Little more can be added to our knowledge of this species in Lake George. The 4 specimens came from different areas of the lake, but a common feature for each locality is its proximity to the shore and the presence of rooted aquatic vegetation in the area. Only fragments of gastropod shells were found in the intestines of the larger fishes (80 and 104 mm standard length) ; the 2 smaller individuals (45 and 46 mm standard length) yielded fragmentary remains of chironomid larvae. HEMIHAPLOCHROMIS Wickler, 1963 The primary generic distinction of this superficially Haplochvomis-like taxon is its reproductive biology (see Wickler 1963). The two morphological characters that separate it from Haplochromis (at least those species occurring in Uganda) are: 226 P. H. GREENWOOD (i) The typical elongate and raised cover to the lateral line canal opening in each pore scale on the body is absent from many of these scales in Hemithaplochromis. Instead there is either a simple pore or seemingly no opening at all; scales in the posterior part of the upper lateral line series and those of the entire lower line are those most often missing a cover. (ii) In males, instead of there being well-developed ocelli on the anal fin, there is a bright orange spot at the posteroventral angle (or tip) of the fin. This character is, of course, associated with the different reproductive behaviour of species in this genus (Wickler, of. cit.). Hemihaplochromis multicolor (Schoeller), 1903 (Text-fig. 40) For a full synonymy of this species, see Greenwood 1965b. Only 1 specimen has been collected by the I.B.P. team. This apparent scarcity of H. multicolor in collections from the lake is probably a reflection of the small adult size attained and the habitats occupied, rather than a true indication of its abundance. Trewavas (1933) does not record H. multicolor from either Lake George or Lake Edward, and the only record from the Kazinga Channel is a few specimens I caught (by dip-netting amongst reeds) near Katungura (unpublished information). The Lake George fish (BMNH reg. no. 1972.6.5 : 21) is 32 mm standard length and was caught near Busatu Island ; its sex is indeterminable. Depth of body 35:9 per cent of standard length, length of head 35-9 per cent. Dorsal head profile very gently curved, sloping at an angle of ca 40° with the horizontal. Preorbital depth 17-4 per cent of head, least interorbital width 30-3 per cent, snout length 0-8 of its breadth and 26-0 per cent of head ; eye diameter 30-3 per cent of head, cheek depth 26-0 per cent. Caudal peduncle as long as deep, 15-6 per cent of standard length. Fic. 40. Hemihaplochvomis multicolor. A Lake Victoria specimen. From Boulenger, Fishes of the Nile. Mouth slightly oblique, lips not thickened. Upper jaw 30-3 per cent of head, lower jaw 34:8 per cent. Posterior tip of the maxilla not quite reaching a vertical through the anterior orbital margin. LAKE GEORGE HAPLOCHROMIS SPECIES 227 Gill vakers all short and stout, the lower 3 a little shorter than the others; 7 rakers on the lower part of the first gill arch. Scales. Ctenoid; 29 in the lateral line series. Many of these scales (particularly in the lower series and posteriorly in the upper series) lack the longitudinal, arched cover to the pore opening, which is represented by a simple pit ; other scales are without any visible opening at all. Cheek with 2 rows of scales; 5 scales between the upper lateral line and the dorsal fin origin, 3 between the pectoral and pelvic fin bases. Fins. Dorsal with 14 spines and 10 rays, anal with 3 and g. Pectoral 23-4 per cent of standard length, 65:2 per cent of head length. Caudal fin rounded, scaled over slightly more than its proximal third. Teeth. Posteriorly in the upper jaw, the outer teeth are very slender and unicuspid. Elsewhere in this jaw, and in the entire outer row of the lower jaw, the teeth are stout and unequally bicuspid. The major cusp is somewhat oblique in outline. There are 34 teeth in the upper jaw outer series. Inner teeth are slender, slightly curved and unicuspid, and are arranged in a single row in each jaw. Vertebral count. Thirteen abdominal and 16 caudal vertebrae. Ecotocy. Nothing is known about the bionomics of this species in Lake George. It seems probable (by analogy with its behaviour elsewhere) that in Lake George H. multicolor lives among submerged plants and in open-water areas within the margin of papyrus swamps (see Greenwood 1965b). Hemthaplochromis multicolor is a female mouth brooder ; adults more than 55 mm standard length are uncommon. Distribution. Widespread in the lakes, rivers, swamps and streams of Uganda, and probably in other regions of eastern Africa as well. The species also occurs in the Nile. DISCUSSION Biology of the Lake George Haplochromts species flock Detailed biological studies of the fishes are being carried out by members of the International Biological Programme team. Some of their results will be available shortly and it is thus appropriate now merely to make some general observations. Unlike the Lake Victoria Haplochromis species, those of Lake George are less closely associated with a particular habitat or substrate type. Nevertheless, any one kind of habitat, for example, offshore open water, the papyrus fringe, protected bays or an exposed sandy shore, has a definable assemblage of species in which a few are numerically dominant. Open offshore waters have the fewest species, and of these only a few are found also in other habitats, and then never as the dominant elements. Species with clearly restricted habitat preferences are: the mollusc-eating H. mylodon, the supposed larval and embryo cichlid-eating H. tawrinus and the grazer on epiphytic and epilithic algae, H. limax. Species caught infrequently 14 228 P. H. GREENWOOD (despite widespread sampling in many habitats) like H. schubotz, H. labiatus, H. nubilus, Hemthaplochromis multicolor and Astatoreochromis alluaudi, may be pre- sumed to have limited habitat preferences. Only Haplochromis petronius can be said to have a truly restricted habitat since it is virtually confined to a rocky crater bay (the only one of its kind in the lake). Even in this example, however, there are exceptions since three specimens of H. petronius have been collected in the main lake (over sand and close inshore). As with the Haplochromis species of Lake Victoria, the most readily observed adaptations seen in the Lake George fishes are those associated with feeding habits. Differences in tooth form and number, pharyngeal bone shape and dentition, and in jaw size and arrangement are the more obvious characters involved in this adaptive radiation. The overall impression gained is of a small-scale Victoria flock. In other words, one in which the major trophic adaptations are developed but to a slightly lower degree of specialization, and with fewer species occupying one trophic niche or showing the same level of specialization. Lakes George and Edward harbour one trophic specialization not found among the Haplochromis of Lake Victoria, namely a species (H. pappenheimt) feeding, as an adult, almost exclusively on pelagic zooplankton. Possibly the absence of an Engraulicypris species from Lakes Edward and George has enabled a Haplochromis to fill this niche. Two Haplochromis species in Lake Edward (H. mylodon and H. pharyngalis) have the enlarged pharyngeal bones and teeth associated with a diet of molluscs. Only H. mylodon occurs in Lake George, but, as in Lake Edward, another mollusc crushing cichlid, A statoreochromis alluaudi, is present. Specialized mollusc shellers (as opposed to crushers) have not evolved in the Edward-George flock. This specialization is well-represented in the H. sauvagei— Macropleurodus bicolor series of Lake Victoria (Greenwood 1957, 1965c). Possibly snails are a less abundant food source in Lakes Edward and George than in Lake Victoria ; qualitative sampling in all three lakes certainly suggests that this is so (personal observations). Another sharp contrast between the flocks of Lake Victoria and George is the presence in Lake George of only one truly piscivorous species (H. sqguamipinnis). Among the inshore species of Haplochromis from Lake Victoria, about 30 per cent are piscivores. The situation in Lake Edward is different again because there are at least three piscivorous species among the known but yet undescribed species from that lake. However, the proportion of piscivores to non-piscivores in Edward is still much lower than in Victoria, probably about one in ten species. A discrepancy is also noticed in the absolute number of larval and embryo fish- eating species. Only one paedophagus species, H. tauwrinus, is known from Lake George (and Edward) whereas at least seven species occur in Victoria. The relative proportion of paedophages to all other species (from all habitats and trophic groups) is, however, less disparate, being ca 6 per cent in Lake George and ca 4 per cent in Lake Victoria. Although there is only one epiphytic and epilithic algal grazer in Lake George (H. limax) as compared with at least four such species in Lake Victoria (Greenwood LAKE GEORGE HAPLOCHROMIS SPECIES 229 1956b), the total for Lakes Edward and George together is three, proportionately a much higher number than in Lake Victoria. It is difficult to generalize about the insectivores and detritus feeders, except to say that in both flocks species belonging to these trophic groups are numerous (Greenwood 1965c). At least one species in Lake Victoria (H. erythrocephalus) and one in Lakes Edward and George (H. nigripinnis) is a specialized feeder on suspended phytoplankton. In Lake George there is a very rapid extinction of incident light so that over much of the lake the photic zone extends merely to a depth of about 60cm. Most Haplochromis species live below this zone or at least spend a great deal of the day in very poorly lit waters. This raises several interesting questions regarding the fishes’ adaptations and behavioural responses to such photic conditions. In par- ticular there are problems associated with breeding behaviour ; there would seem to be insufficient light for potential mates to recognize one another visually. Another difficulty associated with breeding is the nature of the substrate in Lake George. Over much of the lake the bottom is composed of soft, near liquid, organic ooze. This would appear to provide a most unsuitable substrate on which to spawn in the typical Haplochromis fashion. There are, of course, places in the lake where the bottom is hard and where there is reasonably good light penetration. But, no evidence has been collected to suggest that these areas are the only ones used as spawning sites. Brooding females have been found in all habitats, although this does not necessarily imply that the spawning site was near by. It does, however, indicate that for the species involved there are no definite ‘ nursery ’ zones. Field conditions in Lake George virtually preclude direct observations on the fishes’ spawning behaviour, and what information we may get in the future must therefore be indirect, i.e. from aquarium studies. A comparative study of the Lake George species with those from the deep, near-aphotic waters of Lake Victoria would be particularly valuable, especially since the latter species are also faced with a generally soft substrate. Whether or not vision plays an important part in courtship and species recognition, the Lake George Haplochromis species, like those from other and clearer lakes, show distinctive and species-specific male coloration. The Lake George species do, however, differ in one respect. Many females have large, pigmented spots in the same position and of the same colour as the ocelli (or egg dummies, see Wickler 1962) in males. These pigment spots lack the clear border that, in males, makes the spot anocellus. Unfortunately I do not know if female ‘ egg spots ’ are so well developed, or if they are present at all, in Lake Edward populations of the same species. As a footnote to these remarks on coloration it may be noted that no Lake George (or, as so far recorded, no Lake Edward) Haplochromis species exhibits a colour polymorphism like that found in several Lake Victoria species. In these fishes a certain percentage of females has an outstanding coloration in the form of piebald black on silver or black on yellow, or a peppered black, red and orange on a yellowish background. Sex-limited polychromatism is now recorded in at least eight species from Lake Victoria, but none has been found in any Lake George species from among the thousands of Haplochromis specimens examined by the I.B.P. team. 230 P. H. GREENWOOD This absence of polychromatism in the Edward-George flock is rather surprising, not only in view of the flock’s obvious relationship to that of Victoria, but because polychromatism is found in species of Haplochromis from Lake Kivu (see Poll 1939b). (The Kivu Haplochromis are related to those of Lake George in probably much the same way as are those of Lake Victoria.) No explanation is immediately apparent. Species living in such poorly lit waters as those of Lake George might be expected to show certain compensatory hyperdevelopment of various sensory organs, especially the eyes and acustico-lateralis systems. Only the gross morphology of these organs has been investigated so far, and the conclusions reached are equivocal. The cephalic lateral line canals and their openings are not noticeably enlarged. In general, the eye (as measured by its diameter relative to head length) of Lake George species does not seem to be greatly enlarged. A comparison was made between the mean eye diameter in Lake George Haplochromis and their ecological counterpart species from Lake Victoria (the comparisons confined to individuals of the same size, and from species with the same maximum adult size). In the Lake Victoria species examined, the range of mean eye diameter is from 27-0~31-0 per cent of head length, whereas in the Lake George species it is from 28-0—38-0 per cent (the modal range being 31-35 per cent). Haplochromis squamipinnis is excluded from these figures because individuals attain a larger adult size, and eye diameter propor- tions generally show a strong negative allometry with body length. When a com- parison is made between H. squamipinnis and similar-sized individuals of Lake Victoria piscivores, no noticeable difference was noted in eye size (mean eye diameter for H. squamipinnis 23-0 per cent of head, cf. 20-0-26-0 per cent for the Victoria species). These comparisons were extended to include species from the deeper waters of Lake Victoria (see Greenwood & Gee 1969), where light values are probably similar to those found below the upper 30 cm of water in Lake George. The deep-water Victoria fishes have a range of eye diameter between 25-0 and 36-0 per cent of head length, with a modal value at about 31-0 per cent. This compares with values of 27:0—-38-0 per cent (modal range 31-0—35:0 per cent) for the Lake George species, a suggestive similarity. A comparison of the cephalic lateral line canals in these two species-groups was most inconclusive, mainly because of the difficulty in quantifying the characters involved. The relationships and history of the Lake George Haplochromis species From both zoogeographical and historical standpoints, the fishes of Lake George should be considered in conjunction with those of Lake Edward. The two lakes are now interconnected by the Kazinga Channel, they share many otherwise endemic cichlid species, and there is no evidence to suggest that Lake George has ever been in direct connection with any other major water body. Regrettably, it was neither possible to effect a full revision of the Lake Edward Haplochromis species, nor was it feasible to collect in parts of the lake never previously sampled. There is no doubt that many species remain to be discovered in the deeper (i.e. western) parts of the lake (as, for example, was the case in the deep waters LAKE GEORGE HAPLOCHROMIS SPECIES 231 of Lake Victoria). The few recent collections made in Lake Edward (by Dr Dunn of the I.B.P. team), coupled with a brief re-examination of existing collections, show that there are definitely several undescribed species from inshore habitats. Despite the drawback of having to exclude Lake Edward in detail, the material examined, together with that from Lake George enables one to reconsider currently held views on the origin of the Lake Edward—George Haplochromis species flock. Such reflection is very necessary, both in view of the more detailed geological and palaeontological knowledge now available (Greenwood 1959c ; Doornkamp & Temple 1966 ; Bishop 1969) and because of the rather different conclusions I have reached on the interrelationships of the Lake George and Lake Victoria Haplochromis species (themselves extensively revised since Trewavas’ [1933] pioneer work on the Edward-—George species). That the Haplochromis species flocks of Lakes Victoria and Edward—George have a close phyletic as well as a phenetic relationship is beyond doubt. What has still to be determined is whether the Edward—George flock was derived directly from part ~ of the Victoria species assemblage, or whether the two flocks evolved independently, but in parallel, from common ancestral species. Trewavas (1933) believed that Lake Edward ‘...received its Cichlidae, or their not very remote ancestors, from Lake Victoria,...’. This concept has been basic to thinking on the subject ever since (Brooks 1951; Greenwood 1959c ; Temple 1969). Trewavas’ views were influenced mainly by the overall similarity of the Haplochromis species in the two lakes, and by the fact that three otherwise endemic Victoria species were thought to be present in both lakes (see below). At the time of Trewavas’ paper there was little geological evidence available to suggest either the nature or the duration of the route through which the faunal exchange might have taken place. The Rivers Katonga and Ruizi (now with a drainage via swamp divides into both the Victoria and Edward-George basins) suggested a possible passage way, particularly if, in earlier times, the swampy areas were readily passable. Later, Wayland’s (1934) geological and palaeoclimatic hypotheses seemed to support the idea of an aquatic connection between the lake basins (Greenwood 1951, 1959C). The ichthyological evidence once used in support of a Victoria-Edward (and George) interconnection will be reviewed first. On Trewavas’ reckoning there were six cichlid species shared between the lakes, viz. Hemihaplochromis multicolor, Astatoreochromis alluaudi, Haplochromis nubilus, H. guiarti, H. macrops and H. ishmaeli. Furthermore, every endemic Haplochromis species from Lake Edward-George was, in her opinion, closely related to a species from Lake Victoria (the endemic Edwardian monotypic genus Schubotzia eduardianus providing the only clear-cut exception [but see above p. 215)). As noted earlier, the idea of a close overall relationship between the Victoria and Edward-George Haplochromis is still valid (and in many instances is reinforced by new information). I would find it difficult, however, to establish a direct phyletic relationship of an ancestor-descendant kind between each Edward-George species and its Victoria counterpart (the supposed H. guiarti of Lake Edward and H. mylodon excepted). 232 P. H. GREENWOOD It is my opinion that H. ishmaeli, H. guwarti and H. macrops are not present in Lake Edward or Lake George. The fishes once identified as H. ishmaeli are now placed in a new taxon (H. mylodon, see p. 172) and the specimens thought to be H. macrops do not conform with the revised definition of that species (Greenwood 1960), nor are they conspecific with any other endemic Victoria species (see below). The status of the supposed H. guiarti from Lake Edward is difficult to determine without a full revision of the Lake Edward Haplochromis ; no similar species occurs in Lake George. For the moment I can only say that H. guzarti might be the sole example of an otherwise endemic Victoria species occurring in Lake Edward. The importance of determining the identity of Edward * H. gwiarti’ needs no further emphasis. Two specimens identified by Boulenger (1914) as H. macrops were kindly lent to me by the Berlin Museum. A detailed morphometric and morphological study shows that both specimens differ from H. macrops (see Greenwood 1960) in dental and certain proportional characters. One specimen (a female 70mm standard length) can be identified as a specimen of H. nigripinnis. The other (64 mm standard length, probably a female) is of H. macropsoides (see above p. 162). A third specimen (in the British Museum [Natural History], reg. no. 1933.2.23 : 397), identified by Trewavas (1933), has outer jaw teeth with markedly oblique major cusps, quite unlike the acute cusps of H. macrops (see Greenwood, op. cit.). This specimen also differs from H. macrops in several morphometric characters. In all these divergent characteristics, and especially in its dentition, the B.M. (N.H.) fish agrees closely with the type (and some paratypes) of H. vicarius Trewavas, a Lake Edward endemic (see Appendix I, p. 238, for a discussion on the status of this species). Thus, all three Lake Edward fishes formerly identified as H. macrops are now referred to endemic Edward—George species. The identity of Edward-—George specimens previously identified as H. ishmaeli is discussed on p. 176. All the specimens are now included in a new and endemic species from Lakes Edward and George, H. mylodon. Anatomically, H. mylodon is very like H. ishmaeli and H. pharyngomylus of Lake Victoria. The main interspecific difference lies in the coloration of the adult males. In this respect, H. mylodon bears the same relationship to its Victoria counterparts as do certain endemic Haplochromis species of Lake Nabugabo to their counterparts in Lake Victoria (see Greenwood 1965b). It could, therefore, be argued that H. mylodon represents an instance of direct speciation from an H. tshmaeli or H. pharyngomylus- like ancestor that invaded the Edward basin at some time past. Material collected by Worthington from Lake Edward and subsequently identified by Trewavas as H. guiarti is polyspecific. In fact, only a small part of it can be confused with H. gucarti as currently defined (see Greenwood 1962). Of the remain- ing specimens, one resembles H. squamulatus of Lake Victoria, and the others show characters of the H. victorianus—H. serranus species complex in that lake. It must be stressed that none of these specimens is referable to its Lake Victoria counter- part. Preliminary work suggests that in Lake Edward there are, in addition to an H. gwiarti-like species, two other piscivorous species endemic to Lake Edward. LAKE GEORGE HAPLOCHROMIS SPECIES 233 On the basis of preserved material alone, it is difficult to separate the H. guwzarti- like specimens from the true H. guiarti. When specimens are placed side by side, the Lake Edward fishes are distinguishable on the basis of their total morphology, especially head shape. The situation here is quite comparable with that existing between H. mylodon and H. ishmaeli (or H. pharyngomylus) but without the benefit of information on live male coloration. Turning for the moment to the cichlid species which are definitely shared by the lakes. Hemihaplochromis multicolor has such a wide distribution in eastern Africa (including the Nile) that it is irrelevant to this discussion. Its absence from Lakes Edward and George would be of greater significance than its presence. Haplochromis nubilus has a somewhat more restricted range and can definitely be categorized as a species of the Victoria drainage basin. Astatoreochromis alluaudi can also be categorized in this way. Both species, unlike other Victorian Haplo- chromis and related genera, are common in streams and rivers entering the lake, and both penetrate for some distance into papyrus swamps. Taken in its entirety, the ichthyological evidence does not really seem to provide a strong argument in favour of a strictly Victorian derivation for the Edward— George cichlid species. In particular it does not support the idea of derivation from a developed, or partly developed Haplochromis species flock, an idea that I had previously espoused (Greenwood 1959c ; also Temple 1969). The degree of anatomical differentiation between most known Edward-George species and their morphological counterparts in Lake Victoria is sufficiently well marked to suggest that one is observing the results of parallel evolution and not direct speciation in Edward-George from an already specialized invader species. Since both Astatoreochromis alluaudi and Haplochromis nubilus are relatively eury- topic, their presence in both lake basins could mean that they were components of the cichlid complex inhabiting the area prior to lake formation. Possibly, but less likely on ecological grounds, the two species could have gained access to Lake Edward-—George via the Katonga—Mpanga River system. The distribution of the extant non-cichlid fishes in the area contributes little of value to this discussion (see Greenwood 1959c). Only the occurrence of Barbus altianalis in both Victoria and Edward-George argues strongly for some past con- nection between the basins (as it does for a connection with Lake Kivu ; see below). Otherwise, the non-cichlid fishes of these lakes have little in common ; the number of endemic Victoria fishes contrasts with the depauperate but clearly Nilo-Albertine nature of the Edward-George non-cichlid species assemblage (Greenwood, of. cit.). At this point brief mention should be made of Lake Kivu and its small Haplo- chromis flock. Historically, Lake Kivu was derived from a river that once flowed northwards into what is now the Edward-George basin. This river was dammed by the formation of the Bufumbiro volcanic chain, probably during the late Pleisto- cene. As the embryo Kivu gradually filled, it found a new outlet, now the Ruzizi, which drained into Lake Tanganyika. Rapids in the Ruzizi seem to block the passage of fishes (at least northwards) between Lake Kivu and Lake Tanganyika, although 234 P. H. GREENWOOD certain non-cichlids (e.g. Barilius moorii and Barbus pellegrint) are found in both lakes, perhaps as relicts of an earlier, unimpeded river connection. The Haplochromis of Lake Kivu definitely show no relationships with those of Lake Tanganyika, but are distinctly of the Victoria-Edward type. There has been no recent revision of the Kivu Haplochromis species, and data on their live coloration are unavailable ; furthermore, an examination of the type series of two species (personal observations) strongly hints of more species than are currently recognized (Poll 19392, b). Comparing the Edward—George Haplochromis with the Kivu species, on a purely morphological basis, suggests that the Kivu fishes are quite distinct, although showing affinity with Lake Edward—George species (or, in one case, a Lake Victoria species). Of the Kivu species I have studied in detail, H. astatodon Regan resembles H. serridens of Lake Edward, H. graueri Blgr. (at least, that is, one of the types) resembles fairly closely H. schubotzi, and H. paucidens Regan has the general oro- dental specializations of H. labiatus but in many features is more like members of the H. crassilabris species complex in Lake Victoria (see Greenwood 1965b). Haplo- chromis vittatus (Blgr.), too, shows most phenetic affinity with a Victoria species group (especially H. gowersi, a member of the ‘ prognathus’ group in that lake ; see Greenwood 1967) ; it does not closely resemble H. squamipinnis of Lakes Edward and George. The remaining Kivu Haplochromis species (and those still undescribed) I feel less able to comment upon. Haplochromis wittei Poll and H. schoutedeni Poll could be related to either H. elegans or H. aeneocolor of Lake Edward-George, especially the former species, while Haplochromis adolphifrederici (Blgr.), if it is distinct from H. gvaueri, has superficial resemblances to H. schubotzi and H. schubotziellus of Edward and George. As noted earlier (p. 230) sex-limited female polychromatism occurs in at least two Kivu species (H. wittei and H. adolphifrederici) but has not been recorded from any of the Edward—George species. A detailed revisionary study of Lake Kivu Haplochromis species may throw more light on their phylogeny. This would be of great interest because the ancestors of these fishes could have been derived from the proto-George-Edward flock (before the Bufumbiro dam was formed) or could have evolved after that time, from ances- tors living in the river before it was dammed. Since this river originated in the Ruanda Highlands it might well have been populated by different species from those in the westward flowing rivers of the Kenya Highlands which populated the embryo Lakes Victoria and Edward-George. Modern geological studies on the Pleistocene sequence in Uganda also seem to support the idea of parallel evolution in the cichlid species flocks of the Victoria and Edward basins. (See summaries in Doornkamp & Temple 1966; Bishop 1969). Older ideas and temporal sequences based on Wayland’s pluvial hypothesis (1934) are no longer tenable. The formation of Lake Victoria is currently dated at about the later mid-Pleisto- cene, and is thought to be consequent upon the reversal and ponding-back of rivers LAKE GEORGE HAPLOCHROMIS SPECIES 235 that flowed across its present basin into the western Rift lake system (i.e. into a ' proto-Lake Edward-George and Albert). For a summary of the evidence, see particularly Doornkamp & Temple (1966). River reversal was initiated by local uplift along a line nearer the western Rift than the developing Victoria basin. Asa result of this uplift the formerly westward- flowing rivers drained both to the east and to the west, an anomalous situation still persisting. Extensive swamps developed over the watershed, and today these provide an effective barrier to fish dispersal along the rivers. If one accepts the geological evidence, then one must conclude that a lake existed in the western Rift some time before Lake Victoria started to develop as a series of small lakes in the eastern sections of the reversed rivers. There is good palaeonto- logical evidence for the existence of the western Rift lake or lakes from at least Kaiso Formation times (earlier Pleistocene) onwards (Greenwood 1959c). Essen- tially, this fossil record is one of non-cichlid fishes so it throws little direct light on the question of Haplochromis relationships. Judging from the reconstructed topography of western Uganda in the earlier Pleistocene (Doornkamp & Temple, of. cit.) there was a steep escarpment bordering the eastern shoreline of proto-Lake Edward-George. It seems unlikely, therefore, that Haplochromis species could enter this western lake after the formation of Vic- toria. Furthermore, if the species that evolved in the developing Lake Victoria were as stenotopic (i.e. lacustrine) as are their present derivatives, it is highly im- probable that they would spread along the inter-lake rivers (even assuming that such a passage was physically possible). Thus the conclusion seems inevitable that, for all of their histories as lakes, Victoria and Edward-George have been effectively isolated from each other, and that Lake Edward-George is older than Lake Victoria. Since both basins were filled from the same river systems (the old east-west drainage) it is reasonable to assume that their initial fish colonizers were the same. In other words, their present-day Haplochromis species flocks were derived from common ancestral species, presumably of the general- ized type now represented by H. bloyeti (see Greenwood 1971). One Lake George species, H. petronius (see p. 209), does not fit this picture of a close phyletic relationship between the flocks of Lakes Victoria and Edward-George. Nor does it seem to be related to the H. bloyeti stock. As discussed in greater detail above (p. 213), H. petronius shows marked affinities with H. wingatii, a species known from the Nile and Lake Albert (Greenwood 1971). The characters relating these two species (and also H. pharyngalis of Lake Edward ; see p. 214) are not present in any Lake Victoria Haplochromis species. It is unlikely, too, that these characters are products of convergent evolution. To me, the implication is that H. petronius was derived from a different lineage than that of the other species. It cannot, of course, be told if that lineage occurred in the Victoria basin but failed to survive there. Certainly there is no indication of H. wingatii-like species in any of the streams and rivers flowing into Lake Victoria today. That related species appear to have persisted in Lake Albert and the Nile, and also in Lake George, suggests that the ancestor of H. petronius entered that lake from a source other than the old westward draining rivers. The nearest living relative 236 P. H. GREENWOOD of H. wingatii is probably H. desfontainesi, a species now restricted to North Africa (see Greenwood 1971). Perhaps the ancestor of H. petronius (and H. pharyngalis) was a northern rather than an east—west river species, that gained access to Lake Edward-George from the Nile before the lake was isolated from that river by the Semliki rapids (see Greenwood 1959¢). The distribution of Haplochromis nubilus and Astatoreochromis alluaudi in Lakes Victoria and Edward-George, as well as in the small lakes lying between these basins (Trewavas 1933) suggests that the species are remnants of the original species complex inhabiting the old east-west river systems. Trewavas (op. cit.) interpreted the presence of H. nubilus and A. alluaudi in Lakes Nakavali, Kachira and Kijane- balola as possible evidence of the route through which the postulated Victoria to Edward faunal exchange took place. It now seems more likely that the species are fluviatile relicts in those lakes. The absence of other and more typically Lake Victoria or Lake Edward species from these small lakes puzzled Trewavas (op. cit., p- 311). Probably the explanation is simply that these species or their immediate ancestors were never in that area. Elsewhere I have argued (Greenwood 1965c) that the Haplochromis species flock in Lake Victoria represents the amalgamation of several smaller flocks, each evolved in isolation from a common ancestor or, later in the lake’s history, a few common ancestral species. The isolation I envisaged was essentially one of small lakes lying within the area of what is now the basin of a single large lake. The present fauna of Lake Edward-George could be looked upon as another of these isolates but one which, because its basin retained its physical identity, has been given the status of a separate species flock. Phyletically speaking it is perhaps wrong to doso. Rather, one should refer to it as the Edward—George subflock. As matters stand, there is insufficient knowledge of the physical and ecological factors involved in the processes of speciation and adaptive radiation within the Edward-George subflock. Lake George has now been sufficiently well sampled for one to be almost certain that some species occurring in Lake Edward are absent from Lake George. Likewise it is clear that there are many more species in Lake Edward than are currently recorded. (Personal observations on recent collections from Lake Edward.) Collections from the Kazinga Channel show that its cichlid fauna is virtually identical with that of Lake George. That is, the Edward species not recorded from George are also absent from the Channel (see Appendix II). It seems, therefore, that the channel is at least partially a differential species filter between the lakes. The factors inhibiting occupation by certain species (and these do include some from Lake George) have not been discovered. This question is yet another whose solution will depend upon learning more about the ecology of the fishes, especially those from Lake Edward. The unusually complete fossil record for the fishes of Lake Edward shows that throughout the Pleistocene, and well into the Holocene, the non-cichlid fishes were more diverse than at present (Greenwood 1959c). The genera Lates and Synodontis, now absent, were present until local Mesolithic times, and another present-day absentee, Polypterus, persisted into the early Holocene (de Heinzelin’s level N.F.P.R. at Ishango is now dated at ca 8000-10 000 years B.P.). EE EE ee LAKE GEORGE HAPLOCHROMIS SPECIES 237 Depauperization of this Nilo—Albertine fauna was sudden and of a relatively recent date (see Greenwood, op. cit., p. 73). Localized vulcanicity polluting the water (especially of inflowing streams) may have been a major factor in this process. The differential adaptability of species to these altered conditions could account for the fact that some survived while others were wiped out. If the arguments presented above on the origin of the Lake Edward-George cichlids are sound, then these fishes must have survived the environmental hiatus that exterminated several non-cichlid species. There is no evidence that the cichlids or their ancestors reinvaded the lake after the volcanic period, although the time elapsed could have been sufficient for the flock to evolve (see Greenwood 1965b). : Assuming that the Edward-George Haplochromis evolved from mid-Pleistocene fluviatile colonizers implies that speciation and adaptive radiation took place in the presence of such predators as Lates and Hydrocynus. Worthington’s ideas on the inhibitory effects of Lates and Hydrocynus on these processes are well known and well argued over (see Fryer & Iles, 1972, for a comprehensive summary of various view- points in this discussion). The history of the Lake Edward—George Haplochromis species flock now seems to provide an even stronger counter-argument to the Worthington hypothesis than the one presented in my 1959c paper. There, I had assumed that the flock was derived from an at least partly differentiated one (at the species and adaptational levels) invading from Lake Victoria. ACKNOWLEDGEMENTS This paper stemmed from the need to provide the Royal Society—International Biological Programme team investigating Lake George, with a means of identifying the Haplochromis species found there. Thus, I am particularly indebted and grateful to the Royal Society ad hoc Committee for I.B.P./Lake George who so generously approved the expenditure that enabled me to visit Lake George on several occasions. To the team members themselves I am equally indebted. Every member has at some time or other supplied me with data, helped make collections and given me access to their unpublished research ; at all times they have been the most magnifi- cent hosts. Even at the risk of seeming invidious, I must thank especially Dr Ian Dunn and Mr James Gwahaba, the two workers principally concerned with fish research, whose unstinted help has proved invaluable ; and also Dr Dunn, in his capacity as team field leader, for much help. Later Dr George Ganf, as team leader, did much to help me both in the field and with local organization. In the Museum, I owe a great many thanks to my assistant Mr Gordon Howes who has undertaken many jobs to my gain (not least the excellent radiography he has produced), and to Mrs Sharon Chambers for her excellent drawings. To Dr K. Deckert and Dr C. Karrer of the Zoologisches Museum der Humboldt- Universitat, Berlin, I am much indebted for lending me several type specimens. Likewise I am indebted to Dr Max Poll of the Musée Royal de |’Afrique Centrale, Tervuren. To them collectively and individually go my sincere thanks. My thanks 238 P. H. GREENWOOD go also to Mrs M. M. Dick of the Museum of Comparative Zoology, Harvard Univer- sity, for the gracious help she has given whenever I have made use of the M.C.Z.’s collections. Finally, it is a great pleasure to thank Professor W. Banage, Zoology Department of Makerere University, Kampala, who, with his staff, has provided me with hos- pitality and facilities whenever I visited the University. APPENDIX I The status of Haplochromis vicarius Trewavas, 1933 Poll (1939), synonymized H. vicarius with H. eduardii Regan, 1921 on the grounds that the large collection of specimens available to him bridged the morphological gap distinguishing the species. However, Poll seems only to have considered as specifically trenchant the posterior extent of the maxilla, and does not mention the dental characteristics of either taxon. My experience with various Haplochromis leads me to place little importance on the maxillary character, but considerable value on the form of the teeth. The question raised by Poll’s proposed synonymy is complicated by the fact that the type series for H. vicarius is very probably polyspecific. One specimen is labelled ‘Holotype’ although no holotype was formally designated (Trewavas 1933). This fish and at least two paratypes have a distinctive cusp shape to the outer teeth of both jaws (see p. 171), a cusp type that does not occur in the teeth of H. eduardi holotype. Furthermore, I can find no reasons to believe that the tooth shape in H. vicarius holotype represents an extreme variant of the H. eduardii tooth-type (or vice versa). Thus, I would suggest that H. vicarius is specifically distinct from H. eduardit. When the Lake Edward Haplochromis species are fully revised I suspect that other characters will be found to support this separation. APPENDIX II Kazinga Channel fishes During May and June, 1972, collections were made at several places in the Kazinga Channel, particularly in the neighbourhood of Katungura (approximately the midpoint of the channel). Other regions sampled were near the Lake George end of the channel and at Mweya, near the opening into Lake Edward. Small-mesh gill nets and a purse seine were used, and sites near the shoreline and in midchannel were sampled. A list of the species collected in the area around Katungura, with notes on the region of the channel in which they most frequently occur, and a subjective evalua- tion of their abundance, is given below. Haplochromis elegans: common inshore, especially near reed beds; also caught offshore, but is less abundant there. LAKE GEORGE HAPLOCHROMIS SPECIES 239 . aeneocoloy : inshore near reeds ; not very abundant. . nigripinnis : only in midchannel ; rare. . ovegosoma : inshore ; rare. . macropsotdes : inshore and midchannel ; rare. . mylodon : inshore ; very rare. H. angustifrons: mostly from midchannel where it is fairly abundant ; occurs inshore but is rare. H. schubotzi: inshore ; rare. H. schubotziellus : rare in midchannel, even rarer inshore. H. taurinus: midchannel only and then infrequently. H. pappenheimi: abundant everywhere, particularly inshore. Unlike catches of this species in Lake George, those from the channel contained large (II0-130 mm standard length) and sexually active individuals of both sexes. H. squamipinnis : ubiquitous, but in small numbers. H. eduardianus : infrequently caught, and then only by dip-netting among the reeds. Collections made near the Mweya landing were hampered by technical difficulties ; only gill nets, set inshore and in midchannel, were used and then on but one occasion. These yielded specimens of H. elegans, H. aeneocolor, H. angustifrons, H. taurinus, H. pappenheimi and H. squamipinnis, all in small numbers. Because of inadequacies in the sampling methods used at Mweya landing, and since only one collection was made there, this list must be incomplete. It is surprising that the well-sampled Katungura area did not produce any speci- mens of Haplochromis limax, H. nubilus or Astatoreochromis alluaudi. All three species were found in similar habitats in Lake George. There is, of course, a noticeable water flow in the channel, but this alone could hardly be the cause of these particular species’ absence. More probably, their “ absence’ is a reflection of the sampling methods used (and the time available for sampling). Se eae ‘ BIBLIOGRAPHY Bisnop, W. W. 1969. Pleistocene stratigraphy in Uganda. Mem. geol. Surv. Uganda, no. 10. BouLencer, G. A. 1914. Family Cichlidae. In Wiss. Ergebn. Deuts. Zentral-Afrika Exped., 1907-1908, Zool. 3 : 253-259. 1915. Catalogue of the Freshwater Fishes of Africa, 3. London. Brooxs, J. L. 1950. Speciation in ancient lakes. Q. Rev. Biol. 25: 30-60 & 131-176. Dunn, I. G., Burais, M. J., Ganr, G. G., McGowan, L. M. & Viner, A. B. 1969. Lake George, Uganda: a limnological survey. Verh. int. Verein. theor. angew. Limnol. 17: 284-288. Doornxamp, J.C. & Temple, P.H. 1966. Surface, drainage and tectonic instability in part of southern Uganda. Geogrl. J. 132 : 238-252. Fryer, G. & Ites, T. D. 1972. The Cichlid Fishes of the Great Lakes of Africa. Their Biology and Evolution. Oliver and Boyd. Edinburgh. GREENWOOD, P.H. 1951. Evolution of the African cichlid fishes : the Haplochromis species- flock in Lake Victoria. Nature, Lond. 167 : 19-20. — 1954. On two species of cichlid fishes from the Malagarazi River (Tanganyika), with notes on the pharyngeal apophysis in species of the Haplochromis group. Ann. Mag. nat. Hist. (12), 7: 401-414. 240 P. H. GREENWOOD GREENWoopD, P. H. 1956a. The monotypic genera of cichlid fishes in Lake Victoria. Bull. Br. Mus. nat. Hist. (Zool.), 3 : 295-333. —— 1956b. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part I. Bull. By. Mus. nat. Hist. (Zool.), 4: 223-244. 1957. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part II. Bull. By. Mus. nat. Hist. (Zool.), 5 : 73-97. 1959a. The monotypic genera of cichlid fishes in Lake Victoria, Part II. Bull. Br. Mus. nat. Hist. (Zool.), 5 : 163-177. ——1959b. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part Ill. Bull. By. Mus. nat. Hist. (Zool.), 5 : 179-218. 1959c. Quarternary fish fossils. Explor. Parc. natn. Albevt Miss. J. de Heinzelin de Braucourt, 4: 1-80. 1960. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part IV. Bull. Br. Mus. nat. Hist. (Zool.), 6; 227-281. —— 1962. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part V. Bull. By. Mus. nat. Hist. (Zool.), 9 : 139-214. 1965a. Environmental effects on the pharyngeal mill of a cichlid fish, Astatoreochromis alluaudi, and their taxonomic implications. Proc. Linn. Soc. Lond. 176 : 1-10. 1965b. On the cichlid fishes of Lake Nabugabo, Uganda. Bull. By. Mus. nat. Hist. (Zool.), 12 : 313—357- 1965c. Explosive speciation in African lakes. Proc. R. Instn Gt Br. 40: 256-269. 1967. A revision of the Lake Victoria Haplochvomis species (Pisces, Cichlidae), Part VI. Bull. By. Mus. nat. Hist. (Zool.), 15 : 29-119. —— 1971. On the cichlid fish Haplochromis wingatit (Blgr.), and a new species from the Nile and Lake Albert. Revue Zool. Bot. afr. 84, 3—4 : 344-365. ——& Ger, J. M. 1to969. A revision of the Lake Victoria Haplochvomis species (Pisces, Cichlidae), Part VII. Bull. By. Mus. nat. Hist. (Zool.), 18 : 1-65. Maruier, G. & Lereup, N. 1954. A curious ecological ‘niche’ among the fishes of Lake Tanganyika. Nature, Lond. 174: 935-936. Pott, M. 1939a. Poissons. Explor. Parc. natn. Albert Miss. H. Damas (1935-1936), 6: I-73. 1939b. Poissons. Explor. Parc. natn. Albert Miss. G. F. de Witte (1933-1935), 24: 1-81. ReGan, C. T. 1921. The cichlid fishes of Lakes Albert Edward and Kivu. Ann. Mag. nat. Hist. (9), 8 : 632-639. Rosen, D. E. & Batrtey, R. M. 1963. The poeciliid fishes (Cyprinodontiformes), their struc- ture, zoogeography, and systematics. Bull. Am. Mus. nat. Hist. 126 : 1-176. Tempe, P. H. 1969. Some biological implications of a revised geological history for Lake Victoria. Biol. J. Linn. Soc. 1 (4) : 363-373. Tuys Van DEN AUDENAERDE, D. F. E. 1963. Description d’une espéce nouvelle d’Haplo- chvomis (Pisces, Cichlidae) avec observations sur les Haplochvomis rhéophiles du Congo oriental. Revue Zool. Bot. afr. 68, 1-2 : 140-152. Trewavas, E. 1933. Scientific results of the Cambridge expedition to the East African lakes, 1930-1. 11. The cichlid fishes. J. Linn. Soc. (Zool.), 38 : 309-341. Waytanp, E. J. 10934. Rifts, rivers, rains and early man in Uganda. Jl R. anthvop. Inst. 64 : 333-352. WortTHINGTON, E. B. 1932. Scientific results of the Cambridge expedition to the East African lakes, 1930-1. 1. General introduction and station list. J. Linn. Soc. (Zool.), 38 : 99-119. WICKLER, W. 1962. Zur Stammesgeschichte functionell korrelierter Organ- und Verhaltens- merkmale: Ei-Attrappen und Maulbriiten bei afrikanischen Cichliden. Z. Tierpsychol. 19 : 129-164. 1963. Zur Klassification der Cichlidae, am Beispiel der Gattungen Tropheus, Petro- chromis, Haplochromis und Hemihaplochromis n. gen. (Pisces, Perciformes). Senckenberg. biol. 44 : 83-96. LAKE GEORGE HAPLOCHROMIS SPECIES 241 A GUIDE TO THE IDENTIFICATION OF THE HAPLOCHROMIS SPECIES FROM LAKE GEORGE A simple dichotomous key cannot be compiled for these fishes. Intraspecific variability is high and few species can be diagnosed on the basis of single characters. Thus, this ‘ key ’ should be used as a general guide rather than as a means of identify- ing a taxon without recourse to other discriminating characters included in the full species descriptions. It is based on adult and subadult specimens. Haplochromis labiatus is not included here because only one Lake George specimen is known. A specimen of H. labiatus would probably key out to H. elegans, but its dental characters (see p. 198) should prove diagnostic. Morphometric characters are defined ‘on p. 145. Unless otherwise specified, ‘teeth’ refers to the outer row of teeth in both jaws. The Lake George species? Teeth spatulate, the upper half of each tooth strongly incurved (see text-fig. 37) ; lower jaw shorter than the upper. c . H. eduardianus Teeth bicuspid or unicuspid and caniniform, sometimes a mixture of both, and oc- casionally with some tricuspids intercalated ; lower jaw not shorter than upper . A A Scales on the chest very small and deeply embedded (difficult to detect), especially when compared with those on the belly ; a small scaleless area immediately an- terior to the first dorsal spine . 5 3 : H. petronius Scales on chest not deeply embedded, not disproportionately smaller than those on the belly, and readily visible ; no naked area at base of the first dorsal spine : B B.A vertical row of small scales extending onto the fin membrane along the basal part of many (if not all) dorsal and anal fin rays and spines. Lower jaw long (42-57, mode ca 50 per cent of head length, showing positive allometry) and oblique ; teeth usually unicuspid. Adults reach a large size (> 150mm) . H. squamipinnis No small scales extending onto the dorsal and anal fins. Adults rarely more than 115 mm long, modally ca 80mm _. a : Cc C Lower pharyngeal bone massive eB text- fig. 16), most of its teeth strong and molari- form . 0 . H. mylodon Lower pharyngeal Bone nr massive, if maoleiora pharyngeal teeth present, few in number, small, confined to middle row. D D Teeth few in number (32—48, mean 36 in upper ani Stout, deeply embedded ars om tissue (difficult to see) and, although bicuspid, of characteristic shape (see text- fig. 25); gape of mouth manifestly large, lower jaw 43-56, mean 47 per cent of head ; dorsal profile of head concave. Adults reach a standard length of 140 mm H. taurinus Teeth otherwise than above; mean lower jaw length less than 45 per cent of head usually less than 40 per cent . c : : 5 1B; E Depth of body less than 35 per cent of standard length (ace = 31 per cent) ; modal number of gill rakers 10 or 11 (but as many as 13), the rakers slender . F Depth of body usually more than 35 per cent of standard ae modal somites ef gill rakers less than 10 (usually 8 org) . G F At least the posterior third of the premaxilla Parone teeth : “teeth small, flat and of a characteristic shape (see text-fig. 30), 28-38 (mean = 32) in upper jaw. Body colour uniformly silver in both sexes. Body fusiform, its ere 27-31 (mean = 30) per cent of standard length . A 6 ° 6 . H. pappenheimi 1 The two other Haplochyomis-group species are identified as follows: More than 3 (usually 4 or 5) anal spines: Astatoreochromis alluaudi. Many scales of the lateral line series without pores: Hemihaplochromis multicolor. 24 K P. H. GREENWOOD Entire length of premaxilla toothed ; teeth relatively slender (see text-fig. 9), 42-60 (mean = 50) in upper jaw. Males dark, females greyish-silver. Body depth 30-34 (mean = 32) per cent of standard length . : . 4H. oregosoma Teeth in outer row of both jaws with an obliquely truncate cusp (see text-fig. 14), long and movably implanted ; 4 or 5 Saas 2 rows of inner teeth in the upper jaw 6 . . H. limax Teeth brneewise than above, and only 2 or 3 inner rows : (often only I row) c H Usually less than 40 teeth in the upper jaw (34-42, mean = 38); teeth bicuspid, most without a well-developed flange on the major cusp (see text-fig. 3, and cf. text-fig. 5). Upper jaw 28—34 (mean = 30) Pe cent of head (i.e. equal to or less than the eye diameter) . 6 : 5 . H. elegans More than 40 teeth in the upper jam ee Gai mean = 50) F : I Distinct and prominent midlateral dark band running from behind operculum onto the caudal fin; snout length 31-40 (mean = 33) per cent of head H. schubotziellus No distinct midlateral band (or, if a series of short midlateral streaks present, the last not extending onto caudal fin) ; snout length usually less than 30 per cent of head length : ff When fish is viewed laterally, the upper margin of the orbit i is seen 1 to be continous with the dorsal profile, or the eye appears to extend above this line. 4 5 K The upper margin of the orbit lies below the dorsal profile of the head . : L Dorsal head profile sloping smoothly (not obviously interrupted by prominent pre- maxillary pedicels). Preorbital depth 12-15 (mean = 14) per cent of head. Outline of toothed area on lower pharyngeal bone broader than long (see text- fig. 12). Caudal fin not distinctly maculate . H. macropsoides Slope of dorsal head profile interrupted by the prominent premaxillary pedicels. Preorbital depth 13-19 (mean = 17) per cent of head. Outline of toothed area on lower pharyngeal bone noticeably longer than broad (bone appears long and narrow, see text-fig. 19). Caudal fin very distinctly maculate . . H. angustifrons Thickened and papillose area of tissue preceding first gill raker of first gill arch ; pseudorakers between inner and outer row of gill rakers especially well developed and prominent. Snout length 31-40 (mean = 33-4) percentofhead . H. schubotzi No manifestly thickened and papillose area preceding first gill raker (or if tissue in that region slightly thickened, definitely not papillose) ; pseudorakers absent or poorly developed. Snout length usually less than 30 per cent of head : M M Caudal fin with an almost rounded distal margin . D . . A. nubilus P. H. GREENWoop, D.Sc Caudal fin with truncate or weakly subtruncate distal margin. Two species, viz. (i) Most teeth with a well-developed flange on the major cusp (see text-fig. 5). Upper jaw 30—38 (mean 35) per cent of head. Eye diameter 28—35 (mean = 31-4) per cent of head. Lips slightly thickened. Nostril opening much larger than the anterior opening to the nasal lateral line canal. Intestine ca 14 times total body length . : . H. aeneocolor (ii) Few teeth ei a coop on the aioe cusp (Ges foe fig. 7). Eye diameter 33-40 (mean = 36) per cent of head. Lips not noticeably thickened. Opening to nostril of equal size to that of nasal lateral line canal. Intestine long (ca 2—2% times total body length) and much coiled . ¢ 3 . Hz. nigripinnis a Aas on MOSS pea NAT. Department of Zoology hist. British Museum (Natura History) CROMWELL Roap Lonpon SW7 5BD ~ 5 JUL 1973 | es ) x A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. Attson. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75. Wl WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepede, Cuvier and Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. a) TayLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy — of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plat 3 77 Text-figures. 1969. £4.50. ss Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antu) 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. AG MUS % NAT. Hist, ~ 4-5 JUL 75 PRESINED = Printed in Great Britain by ohn Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 sNU PRELIMINARY NOTES ON THE _ ONTOGENY OF THE FRONTAL . AND ADEONELLIDAE (BRYOZOA, a CHEILOSTOMATA) PoE. COOK Vol. 25 No. 6 LONDON: 1973 PRELIMINARY NOTES ON THE ONTOGENY OF THESERONTAL BODY WALL IN THE ADEONIDAE AND ADEONELLIDAE (BRYOZOA, CHEILOSTOMATA) BY PATRICIA L. COOK Pp. 243-263 ; 3 Plates BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 6 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), dnstituted im 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volwmes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25 No. 6 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 22 August, 1973 Price £1.45 PRELIMINARY NOTES ON THE ONTOGENY OF THE FRONTAL BODY WALL IN THE ADEONIDAE AND ADEONELLIDAE (BRYOZOA, CHEILOSTOMATA) By PATRICIA L. COOK CONTENTS Page ABSTRACT . : ¢ : : ¢ : é . : . 245 INTRODUCTION . : c é : : : : : : - 245 OBSERVATIONS ON THE ADEONIDAE . 0 t ¢ 5 : 6 et OBSERVATIONS ON THE ADEONELLIDAE : : a ¢ A - 254 DEVELOPMENT IN THE ADEONIDAE AND ADEONELLIDAE : : . 256 DEVELOPMENT IN SOME OTHER FAMILIES. : : c c » 259 CONCLUSIONS . é a : : : . : : : | 260 ACKNOWLEDGEMENTS . é : S 2 : : ° : e202) REFERENCES. 7 3 é . é 2 : : : > | 202 ABSTRACT Specimens of species belonging to 6 genera of the nominal family ‘Adeonidae’ have been found to exhibit two distinct forms of development of the calcified frontal body wall and the associated apparatus concerned with protrusion of the lophophore. The first type, found for example in Adeona, is interpreted as umbonuloid; the second, found in Adeonella, is inter- preted as cryptocystidean. Both ontogenetic processes are briefly described, and some of the genera previously ascribed to the Adeonidae are provisionally placed together in the family Adeonellidae. INTRODUCTION TuE development of the calcified frontal body wall in the Cheilostomata and its relationships with the apparatus for protrusion of the lophophore have been dis- cussed recently by Cheetham (1968), by Banta (1970, 1971) and by Tavener Smith & Williams (1970). Banta particularly considered some phylogenetic inferences resulting from his observations, and clarified and correlated much previous work, especially that of Silén (e.g. 1942) and Harmer (e.g. 1902, 1957). Both Cheetham and Banta introduced methods and defined criteria which may be applied to future studies affecting the taxonomy and classification of the Cheilostomata (see also Boardman & Cheetham 1969, Boardman ef al. 1970). The observations made here, principally on a few members of one group, the ‘Adeonidae’, emphasize the fundamental nature of the problems which may be recognized as a result of this type of examination, and the wide scope for further study which has arisen from the most preliminary analysis. The protrusion of the introverted lophophore in Bryozoa is effected by the increase in internal coelomic pressure as a result of muscular contraction. In 15* 240 PATRICIA L. COOK the Cheilostomata the type of calcification restricts the region available for mus- cular action to the area of the frontal body wall. In the Anasca, the frontal membrane is that part of the body wall delimiting the internal coleom frontally. Increase in internal pressure is accomplished by the contraction of the parietal muscles which are attached to the frontal membrane and extend to the basal or to the lateral walls. In the Cribrimorpha, the relationships are essentially the same, although the frontal membrane is variously overarched by a lattice of cal- cified spinous structures, leaving a space, the epistege, above it (see Banta 1970). In those Cheilostomata in which the frontal body wall is extensively calcified, two forms of apparatus for lophophore protrusion are known at present. In the first form, an anasciform frontal membrane and parietal muscles are present at an earlier ontogenetic stage than the calcification of the frontal body wall. In the second form, a sac-like proliferation of sub-opercular cells becomes attached to parietal muscles at a later stage than the primary calcification of the body wall. The correlation between the apparatus and the form and ontogeny of the calci- fication is not, however, a simple one. Much work remains to be done, including re-examination of specimens of the specific examples given by both Silén and by Harmer, as preliminary work has indicated that some of these may not, in fact, illustrate the relationships originally ascribed to them (see p. 260). The genera considered here have usually been included in a single family, the Adeonidae (see Bassler, 1953 : G213; Harmer, 1957: 788). With few exceptions they share an erect, branching bilaminar form of growth. They usually possess adventitious and vicarious avicularia, and have sexual polymorphs, those modified for brooding being of a distinct type. Most, but not all, possess a frontal pore, or complex of pores, here considered to be spiramina. The family name Adeonidae was introduced by Jullien & Calvet (1903 : 53), as ‘Adeonidae, J. Jullien, 1903’. Hincks (1887: 150) had used the name several times (pp. 153, 155) in his discussion of Busk’s family ‘Adeoneae ’ (see Busk 1884 : 177), but had not introduced it formally as the name of a family. Both Hincks and Gregory (1893 : 241-244) made interesting observations on the heterogeneity of forms included in the family, especially noting the different types of development of the spiramina or frontal pores. Examination of species nominally classified within 6 genera of ‘Adeonidae’ showed that two major forms of development of the calcified frontal wall were present, and that two forms of protrusion apparatus could either be observed or inferred to be correlated with them. The genera have therefore been separated into two family groups, for convenience. Adeonidae Jullien, 1903. Development of frontal wall umbonuloid (see p. 256). Operculum not sinuate. Marginal pores present. Frontal spiramina present, evanescent or permanent, single or multiporous. Avicularia with acute rostra usually present, adventitious and vicarious. Sexual polymorphs sometimes present, brooding internal. Examples: Adeona, Adeonellopsis, Reptadeonella, Bracebridgia. Adeonellidae Gregory, 1893. Development of frontal wall cryptocystidean (see p. 257). Primary calcified orifice and operculum sinuate. Marginal and usually frontal pores (pseudopores) present. Peristomial spiramina sometimes present. ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 247 Avicularia with acute and spathulate rostra usually present, adventitious and vicarious. Sexual polymorphs sometimes present, brooding internal. Examples: Adeonella, Laminopora. Many other genera, especially fossil examples, which have been placed in the ‘Adeonidae ’, require investigation. Recently, some have been ascribed to other families (see p. 260 for summary). Most of the specimens examined, from the collections of the British Museum (Natural History) (B.M.) and from the United States National Museum (U.S.N.M.) were in the dry state. Many very large, almost complete colonies in the British Museum were collected at least 90 years ago, and the growing edges were damaged and the surfaces covered with dirt. Detached pieces of these and whole specimens of the smaller colonies were soaked in 10 per cent trisodium phosphate solution for 24 hours, washed in water and gently cleaned in dilute detergent with a camel-hair brush, under a low-power microscope. This cleaned and restored much flexibility to the cuticular parts with little resulting damage. The specimens were examined in water after treatment, and differences in appearance noted when they were again dried. Similar comparisons were made between specimens preserved in alcohol and living colonies. Specimens of Adeonidae and Adeonellidae, together with some of species of Microporellidae, Inversiulidae and Umbonulidae, were also decalcified and sectioned. Some sections of Reptadeonella and Escharoides comprising soft and hard parts im situ were also examined. The ontogeny of frontal wall calcification has been directly observed in living specimens of Reptadeonella. Specimens are referred to by the names and numbers under which they are currently stored. Many unnamed specimens of Adeona spp. at the British Museum (Natural History) together with MSS notebooks and drawings by Busk have been examined, but no attempt has been made to assign the material to described species, as a full study, particularly of the Australian reticulate forms, would first be neces- sary. Lack of suitable material has also restricted study of some forms ; for example, it was found impossible to prepare specimens of Bracebridgia pyriformis (the type- species of Bracebridgia) to show the ontogeny satisfactorily. Terms used. Zooid is used for a unit of the colony, autozooid for an individual inferred or observed to have a feeding function. The term sexual polymorph is used for zooids which differ from the majority of autozooids in their calcified parts, and which may also be observed or inferred to have a sexual function, in some cases demonstrably as male individuals, in others as female brooding zooids. Much further work is needed to establish the kind and degree of sexual polymorphism occurring in the Adeonidae and Adeonellidae. The term brooding zooid is here used for sexually polymorphic zooids, which are usually enlarged in a lateral direc- tion, and which have been observed, or can be inferred to have contained embryos. The term lophophore includes tentacles and supporting structures, e.g. the ten- tacle sheath ; ascus refers to a sacciform structure derived from invagination of epidermal cells, and ascopore is restricted to a pore leading through both calcified and adjacent membranous layers into an ascus. The term ‘ epistegal space’ is used to denote the space above the frontal membrane in forms with umbonuloid develop- ment. It is analogous in function with an ascus, but developmentally similar in 248 PATRICIA L. COOK its relationships to a cribrimorph epistege (see Larwood 1962). The term spiramen is here restricted to a pore leading through calcified tissues and adjacent mem- branous layers to a space which is morphologically frontal in position to the oper- culum (see p. 257). OBSERVATIONS ON THE ADEONIDAE 1. Reptadeonella violacea (Johnston). See Harmer, 1957 : 814. MATERIAL EXAMINED. B.M., living (later preserved dry and in alcohol), Ghana, Vernon Bank off Tema, 10m, collected 1968, 1970 and 1972, registered as 1970.2.8.19, 1970.8.2.12 and 1972.I.2.11, respectively. In alcohol, English Chan- nel, 1966.1.10.104 ; dry, Brazil, 1888.4.16.14—17. Zoarium encrusting. Autozooids with a simple frontal spiramen and a distally or distal-laterally directed suboral avicularium. Vicarious avicularia absent. Some sexual polymorphs present which are larger than autozooids, with a wider secondary calcified orifice ; these have been observed to contain embryos and are brooding zooids. Male zooid polymorphs not recognized. Primary zooidal development consists of two major episodes of rapid onto- genetic change, separated by a period of apparent quiescence. The edge of colonies in active growth consists of a zone of uncalcified buds, covered by distended, pale yellow cuticle. The calcification of basal, lateral and distal walls is rapid, and separate zooidal units are observable after 6-8 hours. Peripheral series of 2-4 rows of partially calcified zooids remain in this condition in the laboratory, with little apparent change for as long as 14 days. Under normal conditions this quiescent period may well be shorter, but the great majority of well-preserved colonies in the British Museum shows a zone of partially calcified zooids at the growing edge, and thus some pause in the calcification process seems to be usual. Primary calcification of the frontal body wall is extremely rapid and may be com- plete in less than 12 hours. Calcification commences as a thin lamina which is an inward extension of the proximal and lateral walls. The lamina advances distally and medially beneath the cuticle leaving regularly spaced uncalcified areas. These become the marginal pores at a later stage. In the central area of the membranous frontal wall the parietal muscles and, distally, the operculum are differentiated at this stage. The central part of the membrane then becomes depressed (probably as a result of the contraction of the parietal muscles), and the advancing calcified lamina can be seen to continue its development above the membrane, forming the ‘epistegal space’. The developing lophophore can be seen beneath the membrane at this stage. The membranous frontal body wall thus assumes the same relationship with the developing lophophore and viscera as does the frontal membrane in Anascan species, and the overarching calcified frontal shield assumes the same relationship with the frontal membrane as does the spinous frontal shield in the Cribrimorpha. As calcification proceeds the tissues underlying the cuticle become brightly pigmented. The purple colour is at first concentrated around and above the ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 249 marginal pores; it spreads and darkens in later ontogenetic stages, and remains under the distended cuticle, even when the calcified layers beneath it have reached considerable thickness. As the calcified lamina extends distally and medially over the frontal membrane, it develops paired, medially directed distal processes which meet and fuse centrally above the now fully functional operculum (cf. Pl. 1, figs. 1, 2). Thus a secondary, calcified orifice is formed. This differs from that of the cryptocystidean orifice of, for example, Watersipora, in that it is not in direct contact with, and does not exactly outline, the operculum, which is slightly below it. In Watersipora the primary orifice is calcified and is formed as a result of the contemporaneous differ- entiation of the operculum and the calcification of the surrounding membranous frontal body wall (see Banta 1970). In Reptadeonella the primary orifice is uncalci- fied and is formed by the opening of the operculum. The central foramen formed by the median fusion of the distal calcified processes (see Pl. 1, fig. 2) becomes reduced in size by further, centrally directed calcification, but remains open as a frontal spiramen. It is visible, even in zooids with considerably thickened frontal calcification, as a pit, lined with pigmented tissue beneath distended cuticle. The lophophore and viscera are completely formed before the fusion of the distal processes, and the tentacles have been seen to be partially protruded at this stage. Generally the autozooids are feeding actively while the avicularian chamber is being formed, soon after the formation of the secondary calcified orifice. Brooding zooids have not been seen to protrude tentacles, nor have specifically male zooids been recognized. Thus many details of the degree and nature of sexual poly- morphism are at present lacking. Subsequent ontogenetic changes, although continuous, are apparently far slower than these primary episodes, and consist mainly of the thickening of frontal wall calcification and the development of secondary adventitious avicularia. All these ontogenetic changes may be inferred in well-preserved material by ob- serving zooids at various stages progressively from the growing edge towards the ancestrular region (see Boardman e al. 1970). In specimens preserved in alcohol, the primary membranous frontal body wall, the developing calcified lamina, the operculum and parietal muscles and the fusion forming the secondary calcified orifice can all be seen, although the cuticle is not as distended as in life. In dried specimens the cuticle covering the calcified lamina is shrunken and not easily visible, but the frontal membrane and operculum can be seen beneath the develop- ing calcified frontal shield in some zooids. 2. Adeona sp. (Pl. 1, figs. 1-6). MATERIAL EXAMINED. B.M., dry, ? Australia, 1934.2.10.8, and Port Jackson, Australia, 1886.5.26.1-4. Zoarium erect, bilaminar and regularly reticulate, with round or oval fenestrae. Autozooids with a simple frontal spiramen and a median distal or distal-laterally directed avicularium proximal to the secondary calcified orifice. Vicarious avicu- laria rare. Fenestral kenozooids regularly present. Some polymorphic zooids 250 PATRICIA L. COOK occurring in groups. These are large, with wider secondary calcified orifices than the autozooids, and are inferred to be brooding zooids ; separate male zooids not recognized, At the growing edge there are zooids with basal lateral and distal calcified walls only, and with a shrunken membranous frontal wall. Proximally to these are autozooids with an incomplete curved calcified lamina, or a complete lamina, with a large central pore, beneath which the frontal membrane and operculum can be seen. Even more proximal zooids, at a later ontogenetic stage, show the develop- ment of the frontal avicularian chamber (see Pl. 1, figs. 4, 6). Scanning photo- graphs show the distal median suture in the primary lamina, formed by the fusion of two distal lateral processes, as in R. violacea (see Pl. 1, fig. 3). They also show the progressive thickening of the calcified frontal shield, inferred from the increasing depth of smooth calcification lining the peristome, and the deepening of the pit formed around and above the frontal spiramen. The appearance of these specimens is very similar to that of dry specimens of R. violacea, and is inferred to represent a similar series of ontogenetic changes. 3. Adeona foliacea Lamouroux. See Harmer 1957 : 790. MATERIAL EXAMINED. B.M., in alcohol and dry, Jedan, Aru Is., Siboga Stn. 273, 301A, 371A, 1972.10.1.2,I. Zoarium erect, bilaminar, lobate, and articulated. Autozooids with a single frontal spiramen and a distal-laterally directed suboral avicularium. Vicarious avicularia frequent, marginal, little differentiated from the autozooids, with a frontal spiramen. Brooding zooids inferred to be the larger zooids of the lateral rows. Separate male zooids not recognized. The earliest ontogenetic stages are not present, even in specimens preserved in alcohol. There are indications of a primary lamina in some broken zooids at the growing edge which suggests that the ontogeny resembled that of Adeona sp., above. The later stages, including the derivation of the median suboral avicu- larium chamber which is associated with paired lateral marginal pores, are the same as those seen in Adeona sp., above. The large vicarious avicularia appear to develop in a similar manner to the autozooids, and have a frontal spiramen proximal to the rostrum. Unfortunately, none of the early stages in vicarious avicularian development has been preserved, and decalcified whole mounts show that there are no lophophores present in indi- viduals at the later stages. Embryos have been seen in some of the submarginal zooids and these, at least, are therefore inferred to be brooding zooids. 4. Adeonellopsis meandrina (O'Donoghue & de Watteville) (Pl. 3, figs. 1-3). See O’Donoghue & de Watteville 1944 : 425, pl. 16, fig. 14. MATERIAL EXAMINED. B.M., dry, Cape of Good Hope, 1840.9.30.29; Port Nolloth, South Africa, 1942.8.6.31; Lamberts Bay, South Africa, 1963.9.4.21 ; South Africa, 1968.1.16.125, 126. ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 251 Zoarium erect, bilaminar, foliaceous, convoluted and anastomosing. Auto- zooids with a simple frontal spiramen and paired medially directed lateral avicu- laria. Vicarious avicularia absent. Polymorphic zooids larger than autozooids, with wide secondary calcified orifices, inferred to be brooding zooids. Separate male zooids not recognized. Stages in the development of the primary calcified lamina and frontal spiramen above the frontal membrane, like those in R. violacea and Adeona sp. above, have been seen. The development of the paired avicularia differs in that each chamber is derived separately from a marginal pore. ‘In some zooids only one avicularium develops (see Pl. 3, fig. 3). Large polymorphic zooids are scattered in groups, and are inferred to be brooding zooids. 5. Adeonellopsis sp. (Pl. 2, figs. 1-3). MATERIAL EXAMINED. B.M., dry, Broughton Island, Australia, 1883.11.29.55. Zoarium (inferred from fragments) erect, with wide bilaminar, unbranched, crenulate and anastomosing laminae. Autozooids with a multiporous frontal spiramen, and a single median distal or distal-laterally directed avicularium. Vicarious avicularia large, scattered, directed distally. Brooding zooids inferred to be those larger than autozooids, with a larger multiporous frontal area. The early ontogenetic stages in calcification of the primary lamina above the frontal membrane resemble those described above (cf. Pl. 1, fig. 4 with Pl. 2, fig. rz). After the fusion of the two processes in the distal part of the lamina, zooids at later stages differ in that the frontal spiramen calcifies as a large, porous plate. In the earliest stages finger-like processes protrude into the uncalcified central area. In later stages these fuse, forming a group of irregular pores. Further calcification consists of another series of processes which grow across the pores, producing a stellate effect. The smooth surface of the processes is similar to that of the calci- fication lining the peristome and the inner surface of the avicularian rostra. It contrasts strongly with the rounded blocks of calcification composing the frontal shield, as seen in frontal view (see Pl. 2, figs. 2, 3, and see Banta, 1971 : 168). 6. Adeonellopsis sp. MATERIAL EXAMINED. U.S.N.M., dry, Australia. Zoarium erect, bilaminar, branching. Autozooids with a single, stellate frontal spiramen and single or paired, distally directed suboral avicularia. Vicarious avicularia marginal. Sexual polymorphic zooids not recognized. The earliest ontogenetic stages are rarely preserved, but appear to be similar to those of A. yarraensis, which this species strongly resembles. Use of the scan- ning electron microscope, however, has shown distinct differences in the mode of formation of the frontal spiramen. A pore is left as an uncalcified area as the frontal lamina grows distally, the calcification meeting distally beyond it. More proximally placed zooids show that as the frontal calcification thickens, the pore becomes sunken in a shallow pit, the sides of which are formed of smooth calcifi- cation. A series of finger-like processes grow medially from the sides of the pore, 252 PATRICIA L. COOK at its basal level, i.e. below the secondarily thickened surface level of the frontal shield. Zooids at this stage of development also show paired grooves apparently associated with lateral marginal pores, which protrude into the proximal edge of the secondary calcified orifice, their edges producing a small sinus. Later develop- ment of the grooves in a frontal direction produces the suboral avicularian chambers. The interior of the rostra is lined with smooth calcification, which also extends into the secondary calcified orifice, filling the sinus and forming a shallow, curved pro- jection, similar to that seen in Bracebridgia subsulcata (see below). 7. Adeonellopsis yarraensis (Waters). See Harmer 1957 : 799. MATERIAL EXAMINED. B.M., dry, Uraga Channel, off Tokyo, Japan, 30C, 1972.9.1.1. Zoarium erect, bilaminar and branching. Autozooids with a multiporous frontal spiramen and paired distal medially directed suboral avicularia. Vicarious avicularia rare, marginal and small, directed distally. Large submarginal poly- morphic zooids present, inferred to be brooding zooids. At the growing edge, the development of the proximal part of the curved lamina above the frontal membrane can be seen. Zooids more proximal in position show that the later ontogeny differs from that described above for other species. The lamina continues to grow distally, but does not produce large paired processes which fuse. At the distal edge it leaves a series of small shallow depressions, each with a central slit. Four or five of these are formed and, at the same stage, two medially directed grooves develop leading from the nearest lateral oral marginal pores on each side. The next stage appears to comprise general thickening of the calcification, deepening the grooves which begin to form avicularian chambers. The frontal shield is also extended distally beyond the avicularian chambers and the last developed depression. Concurrently, the depressions, which have become sunken in the surrounding calcification, produce finger-like processes from their walls which grow centrally to produce stellate pores. As calcification proceeds, the whole complex appears to migrate proximally, but this is an illusion related to the thickening of the frontal wall. The base and sides of the depressions are formed of smooth calcification, like that lining the peristome and avicularian rostra, and unlike that of the surrounding frontal shield. 8. Adeonellopsis sp. MATERIAL EXAMINED. U.S.N.M., dry, ‘Albatross’ Stn. 2324, Havana, Cuba, 33 fathoms, Stn. 2336, 23°10’ N, 82°18’ W, 157 fathoms. B.M., dry, Discovery Bay, Jamaica, 200 feet, 1965.8.2.3, Havana, Cuba, 80 fathoms, I9II.10.1.1274, as Porina subsulcata. Zoarium erect, bilaminar, branching. Autozooids with a simple frontal spiramen, and a distally directed suboral avicularium. Vicarious avicularia marginal. Sexual polymorphs not recognized. This species closely resembles the specimen of Porina subsulcata illustrated by Smitt (1873, pl. 6, fig. 136). It may be distinguished from Bracebridgia subsulcata ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 253 (see below) by the constant presence of a simple frontal spiramen, by the distally directed, smaller suboral avicularium, and by the lack, in late ontogenetic stages of a raised ridge of calcification as in B. subsulcata. These closely similar species have been confused together, and the Adeonellopsis seems to be undescribed. No zooids show the earliest stages of calcification; but a few have a simple lamina overarching a depressed frontal membrane. Two zooids show the develop- ment of the frontal spiramen from a notch in the distally advancing lamina. Later ontogenetic stages indicate that the avicularium is associated with one of the lateral marginal pores. The distal part of the peristome is at first raised, but becomes immersed as the calcification of the frontal shield thickens. Later stages include the development of small, variously orientated adventitious avicularia on the frontal. g. Bracebridgia subsulcata (Smitt). (Pl. 2, figs. 4-6.) See Smitt, 1873 : 28, pl. 6, figs. 136-140. MatERIAL EXAMINED. Naturhistoriska Riksmuseet, Stockholm, dry, types of Porina subsulcata Smitt, Florida, 35 fathoms, Pourtales Coll., no. 1837 (for figs. 136, 137) and no. 1829 (for figs. 138-140). U.S.N.M., dry, ‘Albatross’ Stn. 2405, 28°45’ N, 85°02’ W, 30 fathoms. B.M., dry, 13°50’S, 38°46’ W, 32-38 fathoms, 1890.1.30.17-20; ‘Challenger’ Stn. 122, off Bahia, 1887.12.9.735a, 1934.2.16.42, as Adeonella distoma var. imperforata Busk (1884 : 188, pl. 20, fig. 4). Zoarium erect, bilaminar, branching. Autozooids with no obvious frontal spiramen, but with a distal-laterally directed suboral avicularium, its rostrum protruding above the proximal edge of the secondary calcified orifice, which also has a small rounded projection. Vicarious avicularia marginal. Sexual poly- morphic zooids not recognized. Very few growing tips show zooids at the earliest stages of calcification of the frontal shield. In two zooids only, distal processes of the lamina can be seen to have fused distally to form a small pore (PI. 2, fig. 4). In one of these zooids the ridge delineating the proximal edge of the avicularian chamber has been formed, and the frontal spiramen, just proximal to it, is nearly occluded (al A, wie S))o All other zooids inferred to have reached the same or later ontogenetic stages have no sign of a spiramen, which is presumably completely closed by the development of the avicularium and the thickening of the calcification of the frontal shield (see Pl. 2, fig. 6). The periphery of the shield develops as a raised, pyriform ridge of calcification, the central part of the shield remaining relatively depressed (see Pl. 2, fig. 6). In later ontogenetic stages further thickening obscures the ridge, and small, variously orientated adventitious avicularia develop on the frontal. Smitt’s fig. 136 shows zooids with distinct simple frontal foramina. His figured specimen indicates that these are a representation of the depression between the two lateral ridges of calcification in the later ontogenetic stages. Adeonellopsis distoma (Busk), see also Busk (1884 : 187, text-figs. 56, 57), is apparently not closely related to A. distoma var. imperforata. A. distoma has an extensive multiporous frontal spiramen. 254 PATRICIA L. COOK OBSERVATIONS ON THE ADEONELLIDAE 1. Adeonella coralliformis O'Donoghue (Pl. 3, figs. 4-6). See O'Donoghue, 1924 : 55, pl. 4, fig. 24. MATERIAL EXAMINED. B.M., dry, South Africa, 1923.7.20.15, 1962.6.4.13 ; dry and in alcohol, South Africa, 1963.3.20.21, 1968.1.18.5. Zoarium erect, bilaminar, branching. Autozooids with frontal pseudopores and marginal pores. Primary calcified orifice with a distinct rounded sinus. Secondary orifice raised, with paired lateral avicularia directed distally and medially. A bar extends above the orifice extending from each lateral flange of the peristome, forming a small secondary calcified orifice distally and a peri- stomial spiramen proximally. Vicarious avicularia small, marginal, variously orientated. Enlarged marginal zooids inferred to be brooding zooids. This species is very similar in several characters to A. regularis Busk (1884: 186, text-fig. 55, pl. 20, fig. 2), also from South Africa, and a study of the variation of the two forms is necessary. The specimens are well preserved and the growing tips are intact. Zooids at the edge have basal, lateral and distal walls calcified, the frontal wall being entirely membranous. Other zooids in the same horizontal series, or slightly proximal in position, show development of a thin porous lamina of calcification beneath the membrane. In a few zooids the calcification is incomplete, but in the majority has reached the distal end of the zooid. At this stage the outline of the oper- culum and primary orifice within the central, rounded, uncalcified area is visible. The next proximal zooids have a fully formed primary calcified orifice and oper- culum. It is not possible to see the development of an ascus in whole decalcified mounts, but the appearance of the young zooids is very similar to those of Watersipora, Fenestrulina and Onchoporella, in which the rapid growth of the ascus as a sac developing from an invagination of subopercular epidermal cells can be seen. It is therefore inferred that the development in Adeonella consists of calci- fication of a cryptocyst beneath a hypostegal coelom and frontal wall of cuticle and epidermis, as described in Watersipora by Banta (1970), and in Schizoporella by Banta (1971). Scanning electron microscope photographs show that the proximal part of the orifice calcifies in two stages (cf. Pl. 3, figs. 4 and 5). Concurrently, paired lateral oral avicularian chambers are developed, apparently associated with marginal pores, but often showing frontal pseudopores at the bases of their cavities. More proximal zooids show a general thickening of the frontal calcification and the growth of a raised lateral peristome, which incorporates the developing avicularian rostra. Paired processes then grow medially from each lateral peristomial flange, which meet and fuse, forming a bar beneath which the proximal part of the primary calcified orifice is visible. Later ontogenetic stages consist of general thickening of the calcification, immersing the peristome, and producing an apparent migration of the peristomial spiramen proximally. In decalcified sections the relationships of the spiramen and operculum are visible, the retracted tentacles extending to the area behind the operculum. ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 255 N Adeonella polystomella Reuss. See Cook 1968a : 180. MATERIAL EXAMINED. B.M., in alcohol, 5°24'15”N, off Ghana, 1926.12.7.9; ‘Calypso’ Stn. 1, 27°05’N, 17°14’ W, 45-43m, C511, C52D, 1970.8.21 and 8 ; Marche-Marchad Coll., Senegal, 1970.7.8.2r. Zoarium erect, bilaminar, branching and occasionally anastomosing. Auto- zooids with marginal pores and only one series of frontal pores, and a sinuate primary calcified orifice. Paired suboral avicularia very small, directed distally. Vicarious avicularia rare, small, distally directed. Sexual polymorphic zooids probably present. The earliest ontogenetic stages have not been preserved, but near the growing tips a few zooids show a partially calcified, rounded primary orifice. More proxi- mal zooids show the development of the sinus and operculum. Scanning photo- graphs indicate that the primary orifice is calcified in two stages, as in A. coralli- formis. The frontal calcification differs from other species examined in the absence of pseudopores from the central area. Specimen 1970.8.21.1 includes the extensive, encrusting base of a young colony, and these zooids show a single submarginal series of pseudopores. The small oral avicularia appear to develop directly from a pair of lateral marginal pores. The marginal zooids are slightly larger than those in the centre and may be sexual polymorphs, but it is not known whether they are male zooids, brooding zooids, or both. 3. Laminopora bimunita (Hincks). See Harmer 1957 : 820. MATERIAL EXAMINED. B.M., dry, Port Elizabeth, South Africa, 1899.5.1.1346 ; East London, South Africa, 1942.8.6.32 ; Durban, South Africa, 1968.1.16.6 ; in alcohol, False Bay, South Africa, 1962.6.4.11. Zoarium erect, bilaminar, broadly foliaceous. Autozooids with frontal and marginal pores and a distinctly sinuate primary calcified orifice. Paired avicularia with long, acute mandibles, placed at the level of the sinus and directed proximally and medially. Large, distally directed vicarious avicularia rare and scattered, not marginal. Large sexual polymorphic zooids present, some inferred to be brooding zooids. The specimens from False Bay show development of a porous calcified lamina beneath a thick primary membranous layer, presumably consisting of cuticle and underlying epidermis and coelomic tissue. The avicularia develop in association with marginal pores and concurrently with the formation of the primary calcified orifice. No peristomial spiramen is formed in any of these specimens, although it has been reported to occur in this species (see Harmer 1957: 821). Later calci- fication greatly thickens the frontal wall, the primary orifices becoming sunken at the bottom of pits. Large zooids with a wide sinuate primary orifice and relatively small avicularia are present in 1968.1.16.61, either marginal in position or scattered in small groups throughout the colony. It is not known whether some of these zooids are males, 16 250 PATRICIA L. COOK but, as Harmer described (1957 : 821), one type of large zooid has a narrow primary calcified orifice and sinus, the other, presumably the brooding zooid, has a wider primary calcified orifice with a shallow sinus. 4. Laminopora contorta Michelin, see Harmer 1957 : 819. MATERIAL EXAMINED. B.M., dry, ? Cape St Vincent, 1854.11.15.334-3306 (see Harmer 1957 ve locality); in alcohol, ‘Calypso’ Coll. Stn. 14, 14°53'43” N, 23°31'24” W, 25-30m, 1970.8.2, 10, Stn. 26, 15°16’30” N, 23°47'31" W, 50-65 m, 1970.8.2.6, Stn. 75, 16°04'20" N, 22°58’10" W, 45 m, 1970.8.2.5 ; Marche-Marchad Coll., Senegal, 1970.8.2.22. Zoarium erect, bilaminar, irregularly branching, occasionally anastomosing. Autozooids with frontal and marginal pores and narrow, distinctly sinuate primary calcified orifices. Avicularia, if present, small, acute, lateral, directed distally. Vicarious avicularia rare, marginal, small, directed distally. Sexual polymorphic zooids marginal or scattered, slightly larger than autozooids. The frontal calcification is similar in appearance to that of L. bumunita. Zooids with an incompletely calcified primary orifice, and with a porous lamina beneath a thick brownish-purple membrane, are inferred to indicate cryptocystidean development. The external cuticular membrane is distended in specimens pre- served in alcohol. As the frontal calcification thickens, the primary orifice becomes sunken in a pit, as in L. bimunita ; in neither species is a distinct, tubular peristome produced like that in Adeonella. Decalcified sections show the introverted lopho- phore with its tentacles extending up to the operculum and completely filling the space behind it. Scattered large zooids have a wider primary calcified orifice than that of the autozooids and are frequently without avicularia. They are inferred to be brooding zooids, although in view of the possible bisexual nature of similar zooids in L. bimunita, further work is needed to ascertain their function. DEVELOPMENT IN THE ADEONIDAE AND ADEONELLIDAE The developmental series in the species assigned to the genera Reptadeonella, Adeona, Adeonellopsis and Bracebridgia have in common the growth of a calcified lamina above a frontal membrane, in which both the parietal muscles and the operculum are differentiated and functional early in the zooidal ontogeny. This type of growth is similar to that found in Umbonula, Escharoides, Metrarabdotos and in some features in Exechonella, and is termed ‘ umbonuloid’ (see Cheetham 1908 : 58). Umbonuloid development has now been directly observed in living colonies of Reptadeonella, Metravabdotos and Exechonella. In all three cases the early functioning of the parietal muscles, and their apparent contraction during development of the calcified frontal shield has been observed. Umbonuloid growth may perhaps be regarded as the distal extension of a single layer of calcification from the gymnocyst above the frontal membrane. Theoreti- cally this may be derived from a double-layered fold, in which the upper layer does not calcify. It is not known how nearly homologous this may be with some forms, at least, of Cribrimorph development, in which the gymnocystal extensions ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 257 are spinous or derived from spines and must, therefore, be double-layered. Some supporting evidence is provided by study of living specimens of Reptadeonella and Metrarabdotos, both of which have strongly pigmented coelomic tissues. The cuticle in both forms is transparent and in life heavily pigmented tissue is present above the calcified frontal shield and beneath the distended frontal cuticle. This tissue is concentrated around the marginal pores and is present beneath the cuticle of uncalcified marginal buds. It is therefore presumably associated with coelomic tissues and is strongly suggestive of the presence of a coelomic connection at least, probably through the marginal pores (see Banta 1971). Nutritive coelomic tissue is thus present in umbonuloid forms beneath the cuticle but above the frontal calcified layer in a position which is analogous, but not homologous, with the hypostegal coelom of the cryptocystidean forms. The relationships of the calcified layer and the protrusion apparatus in the members of the Adeonidae studied are, however, exactly the same as that found in the Cribrimorpha, and the ‘ epistegal space ’ above the frontal membrane is analogous and may be homologous with an epistege. Thin sections of Ghanaian specimens of Reptadeonella comprising both soft and hard parts im sitw and sections of decalcified species of other umbonuloid genera (including Umbonula, Escharoides and Metrarabdotos) show that cuticle and cellular tissues are sometimes present lining the upper side of the epistegal space formed by the basal side of the calcified frontal shield. As cuticle and presumably coelomic tissues are present on the frontal side of the shield as well, the supposition that the shield represents a partially calcified fold is well worth further investigation. Umbonuloid development differs from that described as ‘ gymnocystidean ’ (see Banta 1970: 53), in that the calcified frontal shield is an extension from the gymnocyst above a membrane, not an extension of the gymnocyst itself at the expense of the frontal membranous area. This last would lead hypothetically to the formation of an ascus by invagination of subopercular cells of epidermal origin as in the ‘ cryptocystidean ’ type of development. The formation of the frontal pore or pore-complex differs entirely from that seen in species of Microporellidae, which apparently have cryptocystidean ontogeny. In whole decalcified mounts of these forms the development of the ascus can be seen, and in sections the ascopore can be seen to lead directly into the ascus. The frontal pore in the Adeonidae leads into the space above the frontal membrane, and, as it is in a position morphologically above the operculum, is a spiramen. It is here defined as a frontal spiramen, to distinguish it from the peristomial spiramen of the Adeonellidae (see below). Although the frontal spiramen complex may be large, especially in some of the sexual polymorphs of Adeonellopsis, it is not yet known whether it is significantly involved in the passage of water into the ‘ epi- stegal space’, during protrusion of the lophophore. Thin sections of Adeona show that the spiramen does not become occluded by calcification during the thickening of the frontal shield. Observations have been made on living Metrarab- dotos, which has no frontal spiramen, but an elongated narrow peristome com- parable with that of Reptadeonella. Protrusion in Metrarabdotos was generally quicker and more frequent than in Reptadeonella, but in neither species was there any noticeable difference in behaviour between them and cryptocystidean forms 258 PATRICIA L. COOK with elongated peristomes, such as Smittina remotorostrata (Canu & Bassler), see Cook 1968a : pl. 8, fig. b, pl. 11, figs. c, d for general notes. The later ontogenetic stages in the Adeonidae comprise the development of oral avicularia and the thickening of the frontal calcification. The single suboral, or paired, lateral-oral avicularia all appear to be derived in association with the marginal pores. Where the avicularium is single and median, as in many species of Adeona and Adeonellopsis, the chamber derives equally from both sides, a ridge forming suborally across the primary lamina distal to the frontal spiramen (see Pl. 1, fig. 4; Pl. 2, figs. rand 5). In Bracebridgia, only one lateral pore is involved. Concurrently, there is considerable thickening of the frontal calcification around the marginal pores and above the primary calcified lamina. The peristomial area also thickens frontally, although the peristome does not usually protrude above the general level of the frontal shield. In Bracebridgia, the distal part of the peristome is at first raised, but in later ontogenetic stages becomes immersed. The frontal spiramen becomes gradually obscured at the bottom of a pit in Adeona and Adeonellopsis, but in Bracebridgia is rapidly and completely occluded. In Adeona and Adeonellopsis in frontal view the position of the frontal spiramen appears to migrate proximally during ontogeny. Further calcification frequently occludes the frontal avicularia, and secondarily developed adventitious avicularia may be formed, with variable orientation, over the frontal shield. There is evidence of a development series of increasing complexity in the ontogeny of the frontal spiramen, ranging from the single simple pore of Adeonellopsis meandrina to the multiporous complex of A. yarraensis. Adeonellopsis spp. nos. 5 and 6 above may show intermediate forms of development in the series, but further observations on other species are needed. The developmental series in the species assigned to Adeonella and Laminopora is inferred to resemble that described as ‘cryptocystidean’ (see Banta 1970). The frontal calcification consists of a distally advancing layer, with marginal pores and frontal pseudopores, beneath membranous layers with presumably an inter- vening hypostegal coelom. No muscles or operculum are fully differentiated until the outline of the primary orifice has been formed by the calcification. Once the operculum has been formed, the ascus is developed by an invagination of sub- opercular cells and the parietal muscles presumably develop in succession proxi- mally as the ascus enlarges (see Harmer 1902). In Adeonella the peristomial area becomes extended and thickened distally and laterally, and raised lateral oral avicularia are frequently formed. The lateral peristome then produces medially directed extensions which meet and fuse in the midline, forming a peristomial bar above the primary orifice. Further lengthening of the peristome by frontal thickening of calcification produces a narrow secondary orifice and a long suboral peristomial spiramen, which, like the frontal spiramen in the Adeonidae, may appear to migrate proximally during ontogeny. A peri- stomial spiramen formed from fused lateral oral processes is known in several genera of Bryozoa, for example, in Gigantopora (see Harmer 1957 : 870). Further calcification may obscure the first formed avicularia, and secondary, adventitious avicularia, variously orientated, may be produced, resulting in an ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 259 appearance, in old, worn colonies, almost indistinguishable from specimens of Adeona at a similar ontogenetic stage. Harmer (1957 : 805) considered that the ‘ ascopore’ of Adeona and the ‘ spira- men’ of Adeonella had fundamentally the same relationships. They both open into a space which is frontal in position in relation to the operculum, but differ entirely in structure, and in space and time relationships within the series of onto- genetic changes forming the calcified frontal body wall in the two families. Both types of spiramen differ from each other and from the ontogeny and relationships of the ascopore of the Microporellidae. Banta (1970) has shown that conclusions drawn from examination of dried material, without comparison with well-preserved specimens showing the full ontogenetic series, are liable to be misleading. Where live specimens can be observed, and their appearance compared with those preserved in alcohol or as dry material, a sequence of inferences can be made with a fair degree of confidence. Banta has pointed out that the late ontogenetic stages in appearance of the crypto- cystidean ascus and the umbonuloid equivalent are very similar. Where a series of developing zooids is visible with a membrane which is left below the front of an advancing calcified lamina, and in which the operculum and parietal muscles are visible early in the ontogeny, umbonuloid development may be inferred. Comparison of stained whole mounts of growing edges of cryptocystidean species of Watersipora, Stylopoma, Microporella and Onchoporella, shows that the groups of cells producing the ascus take up stain differentially, and the developing ascus can be seen. This, together with other correlated obser- vations, is taken to indicate cryptocystidean development. In the absence of cuticle and epidermal layers, as for example in fossil material, it is at present impossible to do more than infer the course and type of calcifi- cation of the body wall by indirect means. Further investigations may provide more character correlations which will be recognizable in fossil specimens, and may lead to more confident inferences about their ontogeny (see Banta 1971 : 171). At present, for the Recent specimens studied, the following correlations may be suggested for those genera here assigned to the Adeonidae and to the Adeonellidae. Where members of a series of zooecia of increasing ontogenetic age can be inferred to develop a pore or pores in a primary lamina which has no pseudopores and which terminates in a non-sinuate, secondary, calcified orifice, and where the avicularian chamber or chambers are not associated with a raised lateral peri- stomial complex, the development is probably umbonuloid. Where members of a series of increasing ontogenetic age can be inferred to develop a porous lamina, terminating in a distinct, sinuate orifice and, in some cases, where a pore is pro- duced by extensions from a raised peristomial avicularian complex, the develop- ment may be cryptocystidean. DEVELOPMENT IN SOME OTHER FAMILIES In the course of this investigation, specimens belonging to several other genera have been examined, notes made on the ontogenetic stages seen, and some types 260 PATRICIA L. COOK of development inferred. Within the ‘Adeonidae’, many fossil genera remain to be examined. For example, the genera Schizostomella and Dimorphocella possess individuals with sinuate orifices and apparent frontal spiramina, together with polymorphic zooids, which may be inferred, by analogy, to have had a sexual function. The characters of these genera are urgently in need of investigation. Several Recent genera, some including fossil species, have already been revised, and the following notes summarize the current position. Inversiula was referred to the Inversiulidae by Harmer (1957 : 955), but replaced in the Adeonidae by Powell (1967 : 339), on the grounds that Vigneaux’s prior definition (1949 : 68) of the family stated that gonoecia (brooding zooids) were absent and that therefore Inversiula was excluded. Inversiula appears to develop in a typically crypto- cystidean manner, the ascus forming from an invagination of a group of cells concurrently with the differentiation of the proximal opercular margin. The plane of the closed operculum, which is hinged distally, is nearly vertical to that of the frontal wall, and the distal wall is very shallow. In decalcified sections, the asco- pore can be seen to communicate directly with the ascus, and is completely separate from the operculum and region of the tentacle sheath, in the same manner as that seen in Fenestrulina and Onchoporella, both members of the Microporellidae. Although modified brooding zooids have been described in one species of Inversrula by Powell (1967 : 340), this would appear to be an interesting case of convergence with the Adeonidae and the Adeonellidae, and I consider that Jnversiula has closest relationships with the Microporellidae. Triporula (synonym Enantiosula) and Anarthropora were referred to the Exechonellidae by Cook (1967 : 965), and inferred to have modified umbonuloid development. The genera Tvigonopora and Metrarabdotos were referred to the Umbonulidae and the Metrarabdotosidae respectively by Cheetham (1968 : 59, 62), who con- sidered them both to have umbonuloid growth. The living specimens of Metrarab- dotos observed have confirmed Cheetham’s conclusions. Other umbonuloid genera mentioned by Cheetham, none of which, however, have been referred in the past to the ‘Adeonidae’, were Umbonula, Hippopleurifera, Hippomenella and Posterula, all placed by him in the Umbonulidae, and Exochella and Escharoides which he referred to the Exochellidae. Some genera referred by Harmer (1957 : 651) to the Ascophora Imperfecta, and thus inferred by later workers to have umbonuloid development, do not appear to display this type of ontogeny of the frontal wall. Some preliminary observations on members of the Petraliidae, Petraliellidae and Celleporariidae indicate that their development is not umbonuloid, but may be similar to the cryptocystidean type. CONCLUSIONS At present there are alternative conclusions as to the composition and relation- ships of the families Adeonidae and Adeonellidae. First, the similarity in colonial structure and polymorphism may be considered to be such that the ontogeny of ate ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 201 the frontal shield and protrusion apparatus is not of phylogenetic significance. Second, it may be concluded that the groups are polyphyletic, exhibiting con- vergence between two or perhaps among several lines of evolution, and parallel development within these lines to a remarkable degree. I regard the testing of the second conclusion as that potentially the more fruitful for future investigations of the Cheilostomata in general. The necessary studies would involve Bryozoan workers in many fields for their completion, and observations on the following subjects would have to be made. The form of the colony is preponderantly erect and bilaminar, but encrusting forms are known. The erect habit, with its positional dimorphism of central and marginal zooids and polymorphs, is known in several other groups, but has been insufficiently investigated. A study of these and of the early astogenetic stages may provide hitherto unrecognized patterns of colonial structure and evolution in many genera. The occurrence of sexual polymorphs, both male and female, also with positional significance, was described and summarized for the ‘Adeonidae’ by Harmer (1957 : 788-821). It is also known in Hippoporidra (see Cook 1968b), and almost certainly occurs in other forms. Sexual polymorphs have probably not been recognized because the polymorphism is not necessarily reflected in the calcified parts of the individuals (see also Hippopodinella described by Gordon 1968), and the study of living specimens and of the lophophore and viscera has been neglected. Although sexual polymorphs have been observed and inferred in the Adeonidae and Adeonellidae, much more work, on living and preserved specimens, is needed to ascertain their detailed morphology and distribution. Preliminary examination of many fossil and Recent species has indicated that there are probably trends in the degree of differentiation of sexual polymorphs with time, which may have taxonomic and phylogenetic significance. The zooids inferred to be brooding zooids appear to be very similar in some representatives of the two families, but have not been certainly recognized in all genera. Similar kinds and degrees of polymorphism are known in other, obviously unrelated, genera, and a thorough investigation may throw light on the kind and degree of convergence which may have occurred in the Adeonidae and Adeonellidae. Avicularia are almost invariably found throughout the two families and where present are consistently without a calcified bar. The vicarious avicularia vary from a relatively undifferentiated type, often with a frontal spiramen, to more complex individuals. Further work on their morphology and distribution, and comparison with similar types, particularly among the Anasca, would be instructive. Electron microscopy will probably reveal considerably more evidence as to the structure and development of the calcified parts of Bryozoan colonies. It is most important that these observations are made in conjunction with those possible with a light microscope, in order that correlations may be made available for workers in the more conventional fields. Taxonomic revisions, involving comparison of type-material with plentiful, well-preserved specimens, together with studies on population variation are urgently required, before any conclusions may be drawn as to the distribution of genera in 262 PATRICIA L. COOK time and space. Finally, after extensive work on fossil and Recent populations, it may be possible to construct a hypothetical phylogenetic model for the evalu- ation of the evolution of the Adeonidae and Adeonellidae, and to relate them to each other and, by extrapolation, to other Cheilostome groups. ACKNOWLEDGEMENTS Many of the ideas and interpretations suggested here were first tested and dis- cussed by the members of the 1969 seminar on Bryozoa at the United States National Museum. I should like to thank the members (Drs R. S. Boardman, A. H. Cheetham, O. Karklins, O. Nye, R. Hinds, R. Singh), and also Dr R. Pohowsky (Geology Dept, Cincinnati University), Dr W. C. Banta (Biology Dept, University of Washington), Dr R. Wass (Geology Dept, University of Sydney) and Dr G. P. Larwood (Geology Dept, University of Durham), all of whom have made helpful comments and criticisms. I am also deeply grateful to Mr W. Pople (Zoology Dept, University of Ghana) for his invaluable help in collecting and culturing living specimens, and to Dr A. H. Cheetham for technical criticism of the first draft of the paper. Finally, I am grateful to Dr A. H. Cheetham, and to Dr A. Andersson (Naturhistoriska Riksmuseet, Stockholm), for the loan of specimens. REFERENCES Banta, W. C. 1970. The body wall of Cheilostome Bryozoa, III. The frontal wall of Watersipora arvcuata Banta, with a revision of the Cryptocystidea. J. Morph. 135 : 37-56. 197t. Op. cit. IV. The frontal wall of Schizoporella unicornis (Johnston). Ibid, 135: 165-184. Basster, R. S. 1953. Bryozoa. In R. C. Moore, Treatise on Invertebrate Paleontology, part G, 1963 : Gi—G253. BoarpMAN, R. S. & CHEETHAM, A. H. 1969. Skeletal growth, intracolony variation, and evolution in Bryozoa: a review. J. Paleont. 43 : 205-233. & Cook, P. L. 1970. Intracolony variation and the genus concept in Bryozoa. Proc. N. Am. Paleont. Convention : 294-320. Busx, G. 1884. Report on the Polyzoa, the Cheilostomata. Rep. Zool. Chall. Exp. 10, 30: 1—xxili, I—216. CueEetHAM, A. H. 1968. Morphology and systematics of the Bryozoan genus Metrarabdotos. Smithson. misc. Collns 153 : 1-121. ——, Rucker, J. B. & Carver, R. E. 1969. Wall structure and mineralogy of the Cheilo- stome Bryozoan Metrarabdotos. J. Paleont. 43 : 129-135. Cook, P. L. 1967. Polyzoa from West Africa. The Pseudostega, Cribrimorpha and some Ascophora Imperfecta. Bull. By. Mus. nat. Hist. (Zool.), 15, 7 : 321-351. 1968a. Bryozoa (Polyzoa) from the coast of tropical west Africa. Atlantide Rep. 10: 115-262. 1968b. Observations on living Bryozoa. Atti Soc. ital. Sci. nat. 108 : 155-160. Gorpon, D. P. 1968. Zooidal dimorphism in the Polyzoan Hippopodinella adpressa (Busk). Nature, Lond. 219 (5154) : 633-634. Grecory, J. W. 1893. On the British Palaeogene Bryozoa. Tvrans. zool. Soc. Lond. 13, 6: 219-279. Harmer, S. F. 1902. On the morphology of the Cheilostomata. Q. Jl microsc. Sci. N.S. 46: 263-350. 1957. The Polyzoa of the Siboga Expedition, Pt. 4. Cheilostomata Ascophora. Rep. Siboga Exp. 28d : 641-1147. ONTOGENY OF THE FRONTAL BODY WALL IN BRYOZOA 263 Hincxs, T. 1887. Critical notes on the Polyzoa, pt. 1. Ann. Mag. nat. Hist. (5), 19: 150-164. JuLiren, J. & Catvet, L. 1903. Bryozoaires provenant des compagnes de 1'Hirondelle (1886-1888). Rés. Camp. scient. Monaco, 22 : 1-188. Larwoop, G. P. 1962. The morphology and systematics of some Cretaceous Cribrimorph Polyzoa (Pelmatoporinae). Bull. Br. Mus. nat. Hist. (Geol.) 6, 1 : 1-285. O’DonocHuE, C. H. 1924. The Bryozoa (Polyzoa) collected by the S.S. ‘ Pickle’. Rep. Fish. mar. biol. Surv. Pretoria, 3 (1922), 10 : 1-63. —— & DE WaTTEVILLE, D. 1944. Additional notes on Bryozoa from South Africa. Ann. Natal. Mus. 10, 3 : 407-432. , Powe Lt, N. A. 1967. Polyzoa (Bryozoa) Ascophora from North New Zealand. Discovery Rep. 34: 199-394. Siten, L. 1942. Origin and development of the Cheilo-ctenostomatous stem of the Bryozoa. Zool. Bidr. Upps. 22 : 433-488. Smitt, F. A. 1873. Floridan Bryozoa, Pt. 2. K. svenska VetenskAkad. Handl. 9, 4: 3-84. TAVENER SMITH, R. & WiLLiAmMs, A. 1970. Structure of the compensation sac in two ascophoran Bryozoans. Proc. R. Soc. B 175 : 235-254. VIGNEAUX, M. 1949. Reévision des Bryozoairés néogénes du Bassin d'Aquitaine. Mém. Soc. géol. Fr. N.S. 28, 60 : 1-153. Patricia L. Cook Department of Zoology British Museum (Natura History) CROMWELL Roap Lonpon SW7 5BD 7 PLATE 1 Adeona sp. ? Australia, 1934.2.10.8, Busk Coll. Fic. 1. Zooecium from near growing edge; showing proximal and lateral frontal lamina, marginal pores and distal communication pores. Fic. 2. Zooecium at a later ontogenetic stage; showing fusion of distal parts of lateral laminae forming median suture, formation of secondary calcified orifice and frontal spiramen. Note early stages in formation of avicularian chamber on either side of median suture. Fic. 3. Same zooecium as fig. 2; showing suture. Fic. 4. Zooecium at a later ontogenetic stage; showing frontal, upward extension of peristome after thickening of frontal shield, and development of avicularian chamber. Fic. 5. Zooecium at a later ontogenetic stage; showing formation of single avicularium proximal to secondary calcified orifice ; note apparent proximal migration of frontal spiramen. Fic. 6. Zooecium at a later ontogenetic stage ; showing formation of avicularian rostrum and condyles. All photographs taken using electron scanning microscope. Magnification: Figs. 1-2, 4-6, x 145; fig. 3, x 370. Bull, By. Mus. nat. Hist. (Zool.) 25, 6 PIA BE 2 Adeonellopsis and Bracebridgia Fics. 1-3. Adeonellopsis sp., Broughton Island, Australia, 1883.11.29.55. Fic. 1. Zooecium near growing edge, showing primary lamina and secondary calcified orifice. Note the beginning of formation of suboral avicularian chamber and subdivision of central spiramen area into multiporous plate. Fic. 2. Same zooecium ; showing arrangement of blocks of calcification round the depressed frontal area. Fic. 3. Same zooecium ; showing frontal spiramen. Note the smooth calcification lining the pores. Fics. 4-6. Bracebridgia subsulcata (Smitt), 28°45’ N, 85°02’ W, 30 fathoms, ‘ Albatross ’ Stn. 2405. Fic. 4. Zooecium near growing edge, showing primary lamina and frontal spiramen. Note that frontal, upward extension of peristome has already taken place, and suboral avicularian chamber has begun to develop. Fic. 5. Zooecium at a later ontogenetic stage ; showing partially occluded frontal spiramen and formation of avicularian chamber. Fic. 6. Zooecium at much later stage; showing frontal suboral avicularium and calcifi- cation of proximal oral process ; the peripheral ridge of frontal calcification is formed. All photographs taken using electron scanning microscope. Magnification: Fig. 1, x 120; fig. 2, x 300; fig. 3, x 600; figs. 4—6, x 140. 2 PLATE 25,6 :) Mus. nat. Hist. (Zool Bull. Br. PLATE 3 Adeonellopsis and Adeonella Fics. 1-3. Adeonellopsis meandvina (O'Donoghue & de Watteville), Cape of Good Hope, 1840.9.30.29, Krauss Coll. Fic. 1. Zooecium near growing edge; showing frontal lamina at an earlier ontogenetic stage than that shown in Adeona, PI. 1, fig. 1. Fic. 2. Zooecium at a later ontogenetic stage; showing formation of frontal spiramen and secondary calcified orifice. Fic. 3. Zooecium at a later ontogenetic stage ; showing development of one of the avicu- larian chambers (frequently paired in this species). Fics. 4—6. Adeonella covallifoyrmis O'Donoghue, South Africa, 1968.1.18.5, O'Donoghue Coll. Fic. 4. Zooecium near growing edge; showing early form of primary calcified orifice, frontal pseudopores and the early formation of avicularian chambers. Fic. 5. Zooecium at a later ontogenetic stage; showing fully formed primary calcified orifice, with complete sinus, further development of avicularian chambers, and early develop- ment of medially directed lateral peristomial processes. Note frontal, upward extension of peristome as frontal shield calcification thickens. Fic. 6. Zooecium at a later ontogenetic stage ; showing sinus of primary orifice, complete avicularian rostra, formation of supraoral bar from fusion of lateral processes, and subsequent formation of secondary calcified orifice and peristomial spiramen. Note apparent proximal migration of peristomial spiramen. All photographs taken using electron scanning microscope. Magnification: Figs. 1-3, x 120; figs. 4-6, x 85. Bull. By. Mus. nat. Hist. (Zool.) 25, 6 PLATE i 2 = f a 7 * : = Phe he us f My - ” \ i ’ 5 ; cS h 6 A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. Arison. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75. ; WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepéde, Cuvier and Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. TayLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates 77 Text-figures. 1969. £4.50. Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. ESENTEL =, PR Oy ae Oay Ww? Gy LY Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 5NU SOME NEW TAXA OF RECENT STALKED CRINOIDEA A. M. CLARK ‘BULLETIN OF 1 MUSEUM (NATURAL HISTORY) aoe Vol. 25 No. 7 LONDON: 1973 SOME NEW TAXA OF RECENT STALKED CRINOIDEA BY AILSA McGOWN CLARK Pp 265-288 ; 2 Plates, 6 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 7 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), «stituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25, No. 7 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 23 August, 1973 Price {1.50 SOME NEW TAXA OF RECENT STALKED CRINOIDEA By AILSA McGOWN CLARK SYNOPSIS Two small families of present-day stalked crinoids are reviewed. The Hyocrinidae is divided into two subfamilies, one including a new genus from the North Atlantic. Another new genus from the same area is added to the Phrynocrinidae, from which the genera Por- phyrocrinus, with a new species from the southern Indian Ocean, and Naumachocrinus are split off as a new family. The range of the bathycrinid Democrinus chuni is extended. New diagnoses are given for the four families mentioned and illustrations of the holotypes of two previously known species. THE small remnant of the class Crinoidea surviving to the present day consists primarily of the free-living Comatulida or feather stars, leaving only about seventy recognized species of stalked sea-lilies. Nearly half of this total is made up by the handsome Isocrinidae, well known for more than 200 years despite their absence from shallow water and distinguished by their relatively large size, with stalks often more than half a metre long bearing whorls of jointed cirri. The present paper is, however, concerned mainly with two of the lesser-known four remaining families — the Phrynocrinidae with four species and the Hyocrinidae with seven. Recent collecting by the National Institute of Oceanography in the North Atlantic has yielded two new genera, one belonging to each of these families and has prompted a review of the included taxa. I am indebted to Mr B. Rowbury of N.1I.O. and also to Professor J. H. Day of the University of Cape Town, from whom came a third new species, this time from an un-named sea mount in the southern Indian Ocean. Family HYOCRINIDAE P. H. Carpenter Hyocrinidae P. H. Carpenter, 1884 : 217-218; Koehler & Bather, 1902: 78; Koehler, 1909 : 264; A. H. Clark, 1912: 272; Gislén, 1939 : 7-17. In his review of this family, Gislén (1939) listed a number of differences between Calamocrinus Agassiz, 1890 and the remaining genera included, namely the presence of five basal plates, a marked angle between the basal and radial rings, arms almost as stout as the radials and branching several times, the brachials short and broad, the distal syzygies separated by more or less numerous muscular joints, the proxi- mal pinnules only about a fifth as long as the arms, the tegmen (or disc) relatively high, extending to about the tenth post-radial ossicle, and the orals relatively small. The angle in the calyx between basal and radial rings appears slight in Agassiz’s drawings of the syntypes of Calamocrinus diomedae (Agassiz, 1892, pls. 2 and 3) though well marked in the additional specimen drawn by Westergren (in Agassiz’s pl. 28). A new member of the family described below, closely related to Hyocrinus, 268 AILSA M. CLARK does have a distinct angle between the two rings, at least in the anterior part of the calyx, and I do not think that this character is very significant. Some of the other differences, such as the relative stoutness of the arms of Calamocrinus in comparison with the radials and the shortness of the brachials, are likely to be correlated with the large size of the type material, calyx height c. 12 mm, whereas most of the other hyocrinids known do not exceed 8 mm calyx height. The relative shortness of the proximal pinnules, I think, is correlated with the fact that the arms branch. However, the presence of five basal plates (presumably primitive), the relatively small oral plates, the numerous muscular joints in the arms in comparison with the number of syzygies and possibly the fact of the arms branching and the very large size of the tegmen I consider are characters of supra-generic weight. Accordingly I am proposing a separate subfamily for Calamocrinus. DIAGNOSIS OF FAMILY HyocriInrpAE. A family of Articulata with a rounded stalk, sometimes slightly polygonal proximally with small tubercles in vertical series, without cirri, numerous proximal columnals discoidal and inflexibly jointed by synostoses in adults, increasing in number by intercalation between several segments, not just above the topmost one, attached to the substrate by an ex- panded terminal plate (when the attachment is known) ; the calyx with thin-walled plates, the sutures between the two rings and between the radials distinct, the three (rarely five) between the basals indistinct or lost by fusion; the arms (or division series) abruptly narrower than the radials in adults, well separated laterally exposing the prominent tegmen (or disc), usually five simple arms (irregular branch- ing normally in only one genus), syzygies usually alternating with muscular joints (except in Calamocrinus where syzygies are fewer and Hyocrinus where they are more numerous) ; only P, and P, undeveloped, the proximal pinnules very long, with gonads lined by large side plates and with distinct cover plates running the whole length of the ambulacra ; oral plates usually enlarged. Subfamily CALAMOCRININAE nov. Diacnosis. A subfamily of Hyocrinidae with the basal ring made up of five similar plates; arms branching several times irregularly, the first axillary not usually before the tenth post-radial ossicle ; muscular articulations considerably outnumbering the syzygies ; tegmen (or disc) very large, reaching up to about the tenth post-radial ossicle in adults ; oral plates inconspicuous. INCLUDED GENERA. Calamocrinus A. Agassiz, 1890, type-species C. diomedae A, Agassiz, 1890 ; monotypic. Subfamily HYOCRININAE Diacnosis. A subfamily of Hyocrinidae with the basal ring made up of either a single fused ossicle or of three unequal plates (the sutures more or less indis- tinct) ; the arms normally unbranched ; the syzygies alternating regularly with muscular joints, with a few exceptions proximally where two successive muscular NEW RECENT SEA-LILIES 269 joints may occur; the oral plates much enlarged, forming a conspicuous cone in the centre of the disc, which is of moderate height, not usually extending beyond about the fifth brachial. INCLUDED GENERA. Hyocrinus Wyville-Thomson, 1876, type-species H. bethel- lianus Wyville Thomson, 1876 ; monotypic. Gephyrocrinus Koehler & Bather, 1902, type-species G. grimaldit Koehler & Bather, Ig02 ; monotypic. Ptilocrinus A. H. Clark, 1907, type-species P. pinnatus A. H. Clark, 1907 ; also including P. antarcticus Bather, 1908 and P. brucei Vaney, 1939. Thalassocrinus A. H. Clark, 1911, type-species T. pontifer A. H. Clark, 1911 ; monotypic. The following new genus is now added : ANACHALYPSICRINUS* gen. nov. Diacnosis. A genus of Hyocrininae with the more proximal stalk segments at least all discoidal in adult specimens, those immediately below the calyx tending to alternate in thickness and projection, with thicker and slightly tubercular-edged plates alternating with thinner intercalary ones; [the distal part of the stalk and its attachment unknown]; the calyx distinctly asymmetrical in adults owing to a marked inclination backwards (i.e. towards the CD interradius containing the anal cone) of the anterior half of the radial ring, although the basal ring is more evenly inverted-conical ; three faint inter-basal sutures just distinct ; arms unbranched, arising almost vertically from the calyx, only Br, with a muscular joint at both ends, elsewhere such joints alternating with syzygies, the brachial formula, I+2,3,4+5,6+7,8+4, etc. the first pinnule, P,, being on the left side of Br, in each case. TYPE-SPECIES. Anachalypsicrinus nefertiti sp. nov. AFFINITIES. These are discussed after the description of the type-species. Anachalypsicrinus nefertiti sp. nov. (Figs. ra—e, 2. Pl. 1) MaTERIAL. ‘Discovery’ Investigations st. 7711/66; 53°11-2' N : 20°5-1' W- 53°11-6’ N : 20°3-9' W (North Atlantic, c. 400 nautical miles west from Northern Ireland), 2432-2380 metres. Holotype [B.M. reg. no. 1972.12.5.1] and three paratypes. Dracnosis. As for the genus. Description. The holotype (Pl. 1 fig. a) is the largest specimen and has the total height of calyx and disc 35 mm, much larger than any hyocrinid previously recorded. Numerical data from it are given in Table 1 with that from the two normal paratypes (the third being badly deformed). * Discovery-lily, from the collecting vessel. AILSA M. CLARK nN “I ° TABLE I Some numerical data from the holotype (1) and two normal paratypes of A nachalypsicrinus nefertiti sp. nov. Measurements are in millimetres ; larger parts of the first two speci- mens were measured with calipers and the naked eye ; the rest with a micrometer eye- piece. The arm measurements are from the anterior arm in each case. I 2 3 Stalk : Length remaining 25 30 22 Number of segments remaining 50 68 76 Number of segments in top 20 mm 43 52 74 Diameter (below flared top) 38 36 Iv Calyx : Diameter of lower end of basals 5:0 5:0 ie Diameter of top of basals 14 14 36 Diameter (maximum) of radials 22 19+ 4°4 Diameter (maximum) of one radial 13 Ls 2°60 Height of basals Io 9 2°5 Height of radials 15 12 33 Total height (including disc) 35 27 75 Post-radial series : Arm length 170 126 22 Breadth at 4 + 5 3°2 3°1 08 Length Br, to 4 + 5 4:0 4°2 3°55 Length of longest pinnules 52+ 35 10 Number of segments in longest pinnules 50+ 40 14 Only a short piece of the proximal part of the stalk remains attached to the calyx of the holotype. It emerges at a slight angle so that the two topmost columnals are incomplete on the posterior side. The sutures between the columnals are very irregular and undulating, often with isolated pockets in notches having some tendency to form vertical series and representing incipient intercalary seg- ments. Just below the basal articulation is a ring of ten tubercles, which alternate with ten hollows in the basals. The very top of the stalk is slightly flared. The basal ring is inverted conical, sloping fairly evenly anteriorly but slightly concave in posterior profile with a small bulge at the base. The three sutures lie approximately in radii B, C and E (see Fig. ra), producing a small plate opposite interradius BC. The radial ring is much more asymmetrical, inclined posteriorly so that in side view it resembles the sloping headdress of the celebrated model of Queen Nefertiti, the posterior profile forming an almost straight line with the basal ring, whereas the anterior one makes an angle of about 140°. There is also some lateral compression so that the maximum diameter antero-posteriorly is distinctly greater than that at right angles to it. The radials curve inwards at the top all round so that the bases of the arms are vertical. There is a slight median convexity up each radial. The individual arms are much narrower than the radials, being separated by more than their own width. The first brachial is partially occluded by the upper edge of the radial, so that it is shorter than the immediately following brachials, though these are short in comparison to the more wedge-shaped brachials that NEW RECENT SEA-LILIES n I radial ring Fic. 1. a-e. Anachalypsicrinus nefertiti gen. & sp.nov. aandb. Holotype and large normal paratype viewed obliquely from below (the position of the stalk shown by cross- hatching) to show the sutures in the basal and radial rings [the scale equals 5 mm]. candd. The same viewed vertically from below (parallel with the axis of the top of the stalk) showing the posterior offset of the radial ring. e. Basal ring of the abnormal paratype similarly. fandg. Diagrams of the basal rings from below of the holotypes of Hyocrinus bethellianus and Thalassocrinus pontifey drawn from descriptions. AILSA M. CLARK n nN Fic. 2. Anachalypsicrinus nefertiti. Holotype. a. Junction of calyx and stalk from the side. b. Posterior edge of tegmen (interradius CD) showing anal cone and base of right posterior arm (C). [The scale equals 2 mm.] follow. Br, is the only proximal brachial with a muscular articulation at each end, otherwise syzygies occur alternately, ie. at I + 2, 4 + 5, 6 + 7, etc., though in the distal parts of the arms occasional instances occur of two consecutive muscular joints. Although some of the arms are broken, there appears to be a size difference between the anterior and posterior ones, the latter being somewhat shorter and more slender. The breadth at brachials 4 + 5 is 3:2mm on arm A but only 2.5mm onarms Cand D. [The paratype, specimen two, has a complete left anterior arm (E) 126 mm long, compared with c. 110 mm for one of its posterior arms. ] The first pinnule (which can be called P,) is on the left side of Br; in each case. An intact one has 38 segments and measures 30 mm, though the immediately following pinnules are markedly longer. All the proximal pinnules have a smooth genital expansion from the third or fourth segment to the twelfth or thirteenth ; this is enclosed partly by the pinnulars and partly by the series of rectangular side plates which stops abruptly at the end of the gonad. The ambulacra are supported by a series of smaller petal-like cover plates, three or four pairs of which correspond to each pinnular. The second to fourth pinnulars bear abruptly projecting lateral flanges. The joint between the first two pinnulars is capable of considerable vertical flexure, whereas those of the remaining segments are almost rigid. The genital area of each pinnule has a low rounded keel dorsally, lacking on the more distal segments. The disc plating is most obvious on the anal cone (Fig. 2b) which is slightly offset in interradius CD towards the right posterior arm, C. The triangular inter- radial areas are sunken between the ambulacra and in each one there are up to 100 small, rounded papillae, projecting to various degrees, perforated individually by the water pores. In the concavity of the high, flange-like oral plate at the apex of each area there are one or two taller papillae and usually at least one capitate NEW RECENT SEA-LILIES 273 papilla or spinelet. Radially the disc ambulacra merge into the arms at about the sixth brachial. There are a number of free-living myzostomes among the pinnules. PaRATYPES. A large piece of stalk, broken at both ends, longer than those attached to the calyces, is included in the sample. It measures 155 mm in length, 3°75 mm in breadth and has 155 discoidal segments, the sutures between them rather irregular. At one end there is a fairly regular alternation of thick segments with intercalary thinner ones but the size evens up along the length of the piece. Very little flexure appears possible. The small specimen has the more distal (i.e. lowest) remaining columnals relatively long, 0-7mm in height and 1-omm in diameter, but the upper segments are shorter, there being an abrupt change in length at about the fiftieth segment from the calyx. The upper extremity of the stalk is much less ornamented and fluted in the small specimen than in the larger ones. The positions of the interbasal sutures (see Fig. ra) agree in the small specimen with the holotype but the normal adult paratype (specimen two in Table 1) and the deformed specimen both have sutures in radii A and D instead of C and E, only the suture in radius B being common to all four specimens. Specimen two also differs from the holotype in the alignment of the maximum diameter of the radial ring. Although this in general slopes posteriorly, it is widest between radius E and interradius BC, almost at right angles to the antero-posterior plane. The small specimen (number three) does not show any distinct asymmetry. As might be expected, the small specimen has the brachials relatively longer than in the others, the first three almost as long as broad (measuring median breadth rather than the expanded articular breadth) and the brachials after the sixth are longer than broad. It is preserved with its arms almost straight and vertical, only the very tips curling over. The total height of calyx and arms is 28mm. It has six or seven pinnules on each side of each arm and the last three in each series are progressively shorter so that each ends about level with the arm tip. No gonads are evident. The deformed specimen has the whole anterior half of the calyx bulging and covered with an irregular mosaic of plates. Only the two arms on the right side are normally developed ; the two on the left are reduced and basally contiguous, one being twisted, while the anterior arm is completely absent. A specimen of Tvichometra was found attached by its cirri to the loose piece of stalk. AFFINITIES. Anachalypsicrinus nefertiti is clearly closely related to the four genera currently included in the subfamily Hyocrininae. In comparison with these it agrees only with Thalassocrinus in having a combination of: separate basals, the second syzygy situated at brachials 4 + 5 and the first pinnule on Br;. The positions of the interbasal sutures may approximate in Thalassocrinus pontifer to those in the holotype of this new species (A. H. Clark’s description is not very precise but there is evidently a suture approximating to radii B and E, though the third is said to be in the posterior interradius, matching with the interradial suture 19 274 AILSA M. CLARK in CD). Ptilocrinus and Gephyrocrinus have fused basals, the second syzygy at 5 + © and the first pinnule on Br,. Hyocrinus does agree in having distinct basals (though with sutures aligned like a mirror image of those of the holotype of Ama- chalypsicrinus nefertiti) but its second syzygy is at 3 + 4 and the first pinnule on Br, ; also the arms of H. bethellianus are much more slender relative to the width of the radials and the brachials are longer, even allowing for the smaller size of the described specimen — top stalk diameter 1:25 mm, while the anus is placed towards the left posterior arm, not the right. The holotype of Thalassocrinus pontifer, from the East Indies (Fig. 3) has the minimum stalk diameter just below the flared top 1°8mm and the calyx c. 5-5 mm high, being considerably smaller than the two largest specimens of Anachalypsicrinus. This might account partially for the almost perfect radial symmetry of Thalassocrinus but the difference in this charac- ter coupled with the more laterally directed arms justifies a generic distinction in my opinion. Geographically, the closest relative is Gephyrocrinus grimaldii, from the vicinity of the Canary Islands and Madeira, the holotype of which has the top stalk dia- meter only 1-0 mm, close to that of the smallest specimen of A. nefertiti. Unfor- tunately, Koehler’s illustrations are poor but G. grimaldii evidently has the arms much less well marked off from the radials, while the ornamentation of the top of the stalk is considerably greater than in the small A. nefertit?, apart from the fused basals and different positions of the second syzygy and first pinnule. Thalassocrinus pontifer A. H. Clark (Fig. 3) Thalassocrinus pontifer A. H. Clark, 1911 : 473-476; 1915, fig. 145. The figure has been made from a rough sketch of the calyx of the holotype and only known specimen. It omits details of the tuberculation at the top of the stalk and on the tegminal plates. The interbasal sutures, described by A. H. Clark as “almost obsolete’ are not shown. The narrow bands at the edges of the radials (shown stippled) are slightly darker than the rest in the preserved specimen and give the impression of being bevelled, though I did not think them depressed as A. H. Clark described. Possibly they indicate recent growth. Family PHRYNOCRINIDAE A. H. Clark Phrynocrinidae A, H, Clark, 1907: 510; Gislén, 1925 (part) : 92. The family was established by A. H. Clark, without diagnosis, to accommodate his new genus Phrynocrinus, type-species P. nudus from southern Japan. This was distinguished by having the stalk terminating below in a lobed attachment disc and above in only one or two discoidal columnals, all those in between being trapezoidal in shape (viewed from the side), broader than high, all joined by flexible synarthrial articulations of which the successive ones are flattened in alternate planes approximately at right angles so that they appear to be paired, the calyx NEW RECENT SEA-LILIES n “I wn Fic. 3. Thalassocrinus pontifer A. H. Clark. Holotype. U.S. National Museum cat. no. 24783. Side view of calyx, adjacent ossicles and tegmen; the basal sutures not shown. The foremost arm is broken at Bry. [The scale equals 5 mm.} very squat and markedly flaring above, the arms diverging from the base leaving exposed extensive interradial areas of the tegmen or disc, relatively few syzygies present and no pinnules developed before the eighth or ninth post-radial ossicle, also branching occurs irregularly after the twelfth post-radial ossicle. The holotype of P. nudus is relatively large, the diameter of the topmost columnals 6 mm. Two further genera, both monotypic, have since been referred to the family, namely Naumachocrinus A. H. Clark, 1912, type-species N. hawatiensis, and Por- phyrocrinus Gislén, 1925, type-species P. verrucosus, from the Kei Islands. Both of these genera have more or less numerous discoidal columnals proximally, the calyx elongated and approximately cylindrical, not appreciably wider than the top of the stalk, also the arm bases closely apposed laterally, continuing the vertical alignment of the radials. In fact, the general appearance is very like that of the Bathycrinidae, though some of the distal columnals with their compressed joints aligned in alternate planes are common also to Phrynocrinus. The arms of Nawma- chocrinus hawatiensis are unknown beyond the first brachial and in Porphyrocrinus verrucosus the termination of the stalk is unknown. However, new material con- generic with P. verrucosus shows that the distal end of the stalk in this genus also terminates in a lobed disc. The bathycrinids are distinguished by the root-like ending of the stalk distally, with numerous irregular branching radicular ‘ cirri’. This precludes the inclusion of Porphyrocrinus and Naumachocrinus in the Bathy- crinidae, if they are to be excluded from the Phrynocrinidae, which I am convinced must be done. It therefore becomes necessary to establish a new family to ac- commodate these two genera. 276 AILSA M. CLARK Since Gislén’s diagnosis of the Phrynocrinidae allowed for the inclusion of Por- phyrocrinus, a modified one follows. Diacnosis. A family of Articulata with the stalk round or elliptical in cross section, without cirri, even the topmost columnals linked by flexible synarthrial joints and only the most proximal one or two sufficiently short in adults as to be termed discoidal (though in immature specimens the uppermost segments are relatively shorter and several may be discoidal), only one new columnal developed at a time immediately below the calyx, the remaining segments giving a moniliform appearance owing to the compressed elliptical joints alternating in alignment, attached to the substrate by an expanded terminal plate; the calyx inverted conical in shape, relatively short and compact, the sutures between the five basals and five radial plates distinct, the two rings not widely dissimilar in height ; the division series following the radials similar in width basally to the radials but diverging abruptly, exposing the large tegmen (or disc) to view, ten arms (possibly more) in the two known genera, most syzygies separated by two or three muscular joints ; pinnules lacking basally from the first two or three possible positions on each side, the more proximal ones not markedly enlarged and with only small rods in the ambulacra, no conspicuous side or cover plates. Phrynocrinus nudus A. H. Clark (Fig. 4) Phrynocrinus nudus A. H. Clark, 1907 : 507-510, fig. 1. Phrynocrinus obtortus Matsumoto, 1913 : 221. A sketch of the calyx of the holotype of P. nudus in the U.S. National Museum shows that in fact there are short vertical interbasal sutures and the individual basals are not really triangular as described by A. H. Clark, though he did note that the angles are blunted. The diameter of the top of the stalk is 6 mm, compared with 4mm in the holotype of P. obtortus Matsumoto, also from southern Japan, and I think that this size difference is sufficient to account for the more obviously pentagonal shape of the basals in the latter. Judging from the material of a new member of the family described below, there may be some variation in the shapes of the columnals and the exact alignment of the successive stalk joints, which characters Matsumoto also used to try to distinguish his nominal species. Taking all this into account, including the geographical proximity of the type-localities, I consider that only a single species should be recognized and accordingly refer P. obtortus to the synonymy of Phrynocrinus nudus. ZEUCTOCRINUS* gen. nov. DiaGnosis. A genus of Phrynocrinidae with only the more proximal columnals markedly trapezoidal in side view with compressed alternating joints, the more distal ones becoming cylindrical (the distal termination of the stalk unknown) ; * From Greek zeuc/os — yoked or joined in pairs, referring to the appearance of the columnals. NEW RECENT SEA-LILIES 277 Fic. 4. Phrynocrinus nudus A. H. Clark. Holotype. U.S. National Museum cat. no. 22601. a. Side view of calyx and adjacent ossicles. b. Junction point of distalmost columnals and terminal stalk plate, showing the absence of a distinct suture at the last constriction, contrary to A. H. Clark’s fig. 1D. [The scale equals 5 mm.] the division series consisting normally of only two ossicles of which the IBr, in adults are proximally contiguous laterally but not in immature specimens, the two ossicles joined by syzygy; brachial syzygies normally at 1 + 2, 3+ 4,7 + 8, 13 + 14 and then probably at intervals of three or four muscular joints ; pinnules not developed in the first three possible positions, the first one being P, on Bry) or thereabouts. TypE-sPECIES. Zeuctocrinus gisleni sp. nov. Zeuctocrinus gisleni sp. nov. (Fig. 5. Pl. 2) MATERIAL. ‘Discovery’ Investigations st. 7711/66; 53°11-2’ N : 20°5-1’ W- 53°11-6’ N : 20°3-9’ W (North Atlantic, c. 400 nautical miles west from Northern Ireland), 2432-2380 metres. Holotype [B.M. reg. no. 1972.12.5.4] and four para- types. Diacnosis. As for the genus. Description. All the specimens have the stalk broken, so that their distal ends and attachments are unknown. The holotype (Fig. 5a, b, Pl. 2 figs. a, b) is the largest specimen, number I in Table 2. It has the topmost columnal only partially developed and almost discoidal but the second one is already more than half as high as its minimum diameter. Together they form a prominent synarthrial tubercle on one side, though the 278 AILSA M. CLARK YES bu SAR ORTHO sith oe f= a — =| ES ae = Bs Fm \ “a Fic. 5. Zeuctocrinus gisleni gen. & sp.nov. aandb. Holotype. a. Calyx and adjacent ossicles. b. Distal most remaining three columnals (numbers 31-34 from top) c-e. Smallest paratype. c. Crown and proximal part of stalk. d. Columnal 52. e. Columnal 59. [Both scales equal 2mm; that on the right applies to c-e.] opposite side is still flat. The next 25 or so segments have their joints symmetrically elongated in alternate planes almost at right angles (in fact the plane of elongation spirals through go° during about six pairs of articulations). This produces a moniliform appearance to the stalk in side view, with the segments forming pairs. The modification in shape gradually lessens distally so that the last few columnals are almost cylindrical. NEW RECENT SEA-LILIES 279 TABLE 2 Numerical data from the holotype (1) and four paratypes of Zeuctocrinus gisleni sp. nov. Measurements are in millimetres. I 2 3 4 5 Stalk : Length remaining 61 50 50 57 58 Number of segments remaining 34 26 40 50 62 Number of segments in top Iomm 7 2. ais} 105 18 2 Length of top 10 segments 16 14 9 4 I'7 Length of lowest remaining segment 2:0 2:3 1-3 1-7, 1-2 Maximum diameter at lower end 3:6 2°5 22 1-3 08 Minimum diameter near top 2:8 2:0 2-0 12 o-7 Calyx : Diameter of lower end of basals 3:6 2:6 24 1:2 O-7 Diameter of top of radials 5:2 4°2 37 24 1:8 Maximum height radially raze) 1°7 I°-7 Id 0-6 Post-radial series : Length IBr, 2°2 1:8 1°5 T°2 08 Length IBr, 2:8 2°4 2:3 I 13 Median breadth IBr series 2:6 I'7 1:7 pare) O-7 Breadth at 3 + 4 2-2 5 Ie7, O-4 Length IBr, to 3 + 4 974 79 7:0 5:0 3:9 The calyx is distinctly asymmetrical (Fig. 5a), lowest in radius E (the left an- terior one viewing orally). The general skin-covering seems particularly opaque in this area and the sutures can only be seen by dissolving it in bleach — and then only very indistinctly. The suture between the basal and radial rings forms a zig-zag line. The interbasal and interradial sutures are approximately equal in height since the latter are cut short by the meeting of the adjacent IBr, inter- radially ; the interradial calyx height is c.2-7mm. The division series and arms are almost cylindrical in cross section. The two ossicles of the division series (in one case four) are joined by syzygy. All the arms are broken by or at the fourth syzygy, which is usually at 13 + 14 but in one case each at 12 + 13 and 14 +15. The previous syzygies are at I + 2, 3+ 4 and 7 + 8 except on the pair of arms based on the IIBr series of four ossicles, which have syzygies at 2 + 3, 4 + 5 and 8 + with a pinnule on the inside of Brg. All the other arms remaining beyond Br, have the first pinnule (P,) on the outside of Br,). It has 9 + (?c.3) evenly tapering segments and measures 5:9 + mm in length (probably + c. 2-0 mm). The tegmen (disc) is joined to the arms to about Br, while the prominent anal cone reaches level with Bry. PARATYPES. Specimen 2 has the top columnal fully developed, forming a per- fect inverted pair to the second one. The successive joints are again aligned at not quite go° so that there is a slight spiralling effect of the joint faces. The calyx again is somewhat asymmetrical but in this case the anterior radius, A, is the lowest. The IBr series are not quite in contact laterally, allowing the radials to 280 AILSA M. CLARK be seen interradially to a height of 3-3 mm. There are slight lateral flanges on the proximal brachials. The longest remaining arm has 21 segments left, measuring 21 mm in length; when intact it was probably about as long again. The breadth at the distal end is I-2 mm as opposed to 1:5 mm at 3 + 4. The positions of the syzygies are very irregular and in only one or two cases does the joint between brachials r and 2 appear to be a proper syzygy, though 3 + 4 is developed on each arm. In one case the next syzygy is not until 13 + 14, followed by 17 +, while another has i + 2, 3 + 4,6 +7, 9+. One arm even has two successive syzygies, 3 + 4 + 5 (as described by A. H. Clark in the holotype of Phrynocrinus nudus), followed by 8 + 9 and 12 +. Py on Bry is normally the first pinnule again, though there is a pinnule on Br, in one case and on Bry or Br, in two others. P, has 14-18 segments, the basal ones stout, though longer than broad and flared at the joints, while the distal segments are attenuated. The length is 6-5-8mm. P,; may bear a small gonad near its base. There are some very slender and often branching rods in the ambulacra. One pinnule is bifurcated near its base. Specimen 3 has several of the topmost columnals relatively short. The calyx is not appreciably asymmetrical. The sutures between basals and radials are distinct, even when wet, owing to a slight bevelling of the edges. The division series are again slightly spaced laterally. All the arms are broken by or at the fourth syzygy; the syzygies remaining are at I + 2, 3+ 4, 7 + 8 and 12 or 13 +, with the first pinnule on the outside of Bry. Specimen 4 has the four topmost columnals discoidal, while the lower ones conversely are much longer relatively than in the larger specimens and are medially constricted with length: median breadth 1-6:1. It too has the sutures in the calyx distinct by virtue of being bevelled at the edges. The maximum height of the basals is c.o-°8mm. No appreciable asymmetry is present. The division series are separate laterally and have slight lateral flanges, continued on to the arm bases. The smallest specimen (Fig. 5c-e) shows these same tendencies even better developed. Up to 11 of the uppermost columnals are discoidal and the distalmost are only just cylindrical, not stouter medially, while the longest segments have length : median breadth as much as 2°5:1. The upper segments are so short that the usual moniliform appearance is not obvious. The division series are widely spaced laterally and there are very marked lateral flanges on them and on the first few brachials. The main ontogenetic changes can be outlined as follows: The younger speci- mens have a great disparity in the relative length of the upper (short) and lower (long) columnals. The basals and radials become relatively broader in larger specimens and the calyx takes on some degree of asymmetry, though it remains to be seen if this has an antero-posterior correlation. The division series are at first widely separated laterally at their bases but gradually approximate, to become contiguous when the basal diameter of the calyx exceeds about 3mm; they have prominent lateral flanges, continued on to the arm bases, but these gradually become obsolete and the ossicles more nearly cylindrical. | | | | NEW RECENT SEA-LILIES 281 AFFINITIES. In comparison with Phrynocrinus, the only other genus included in the family as here restricted, Zeuctocrinus differs in having the more distal columnals much less moniliform and the division series limited, normally to only two ossicles. Family PORPHYROCRINIDAE nov. The justification for establishing this new family is given in the discussion of the Phrynocrinidae above. Diacnosis. A family of Articulata with a rounded stalk, without cirri, with several of the proximal columnals discoidal and inflexibly jointed by synostosis but with new segments formed only between the topmost one and the calyx, the middle columnals elongated and more or less cylindrical (sometimes barrel-shaped or else waisted), with synarthrial joints. which distally become elliptical, alter- nating in alignment, giving a moniliform appearance, attached to the substrate by an expanded terminal plate ; the calyx probably thick-walled, relatively small and more or less cylindrical, not markedly flared at the top, the basals slightly or very much shorter than the radials ; the arms continuing the alignment and width of the radials, closely approximating laterally and obscuring the tegmen (or disc) from lateral view, syzygies alternating usually with muscular joints and pinnules lacking basally from the first two or three possible positions on each side of the undivided arms, none markedly enlarged and all lacking conspicuous side or cover plates. [The syzygies and pinnules unknown in Naumachocrinus, the arms of the only known specimen being broken after the first brachial.] INCLUDED GENERA. Porphyrocrinus Gislén, 1925, type-species P. verrucosus Gislén, 1925 ; hitherto monotypic. Naumachocrinus A. H. Clark, 1912a, type-species N. hawatiensis A. H. Clark, Igi2a ; monotypic. PORPHYROCRINUS Gislén Porphyrocrinus Gislén, 1925 : 91-92. ? Monachocrinus (part; M. incrassatus) Gislén, 1933 : 483-485. The type-species, Porphyrocrinus verrucosus Gislén is only known from a single specimen taken near the Kei Islands (c. 5-5° S : 133° E) in 345 metres (a relatively small depth for a stalked crinoid). It lacks the distal end of the stalk so the form of attachment is unknown. As discussed under the heading of the Phrynocrinidae, I consider that Gislén was mistaken in referring it to that family. The general facies with closely apposed arms arising from the nearly cylindrical calyx without abrupt distinction and the discoidal, rigidly articulated, proximal columnals giving way distally to more or less cylindrical segments (though these are somewhat compressed at the joints in alternating planes), is far more like that of the Bathy- crinidae than of the Phrynocrinidae. 20 282 AILSA M. CLARK Another specimen which I think will prove to be referable to Porphyrocrinus is the holotype (and only known specimen) of the supposed bathycrinid Monacho- crinus incrassatus Gislén, 1933, from St Helena, in which only one ossicle remains above each radial. Gislén discarded the idea of its affinity with Porphyrocrinus because in that genus ‘ the upper part of the stem is not thickened and the distal stem-joints only very gradually reach the type of long slender knobby stem-joints which is so well developed in most Bathycrinidae and all Monachocrini. Besides this, all the Phrynocrinidae [with which he allied Porphyrocrinus] are large and stout forms, while the new species just as the Bathycrinidae is a small and slender form.’ Although the holotype of Phrynocrinus nudus was certainly much larger, the diameter at the very top of the stalk 6mm, the same measurement in the holotype of Porphyrocrinus verrucosus is only 1-5 mm, compared with 0-8 mm in the holotype of Monachocrinus incrassatus. It seems to me premature to assume that the last nominal species is a small one on the basis of a single specimen and I believe that the differences in the stalk could well be correlated with the smaller size. Since the species of Monachocrinus are characterized by having division series and ten arms, it remains for more nearly complete specimens to be collected in the vicinity of St Helena with either arms or division series present beyond the first post-radial ossicle and preferably also the entire stalk to show where its true affinity lies. The possibility that M. incrassatus can be referred to Porphyrocrinus is enhanced by two stalked crinoids collected in the southern Indian Ocean, about mid-way between South Africa and Amsterdam Island, in 400 metres. These have the diameter at the top of the stalk 1-2 and 1:3mm. Although both have the post-radial series broken after the first ossicle, fortunately two loose arms are present with them, indicating that the first ossicle is Br, not IBr,, though it is just possible that an intervening axillary and first brachial have been lost. These two specimens resemble P. verrucosus except that they have more numerous discoidal proximal columnals, the radials are relatively longer, though possibly within the range of specific variation, the first pinnule is on Br,, on both the loose arms, as in only one out of five arms of Gislén’s specimens — the other four having a pinnule on Brg, and the texture of the stalk is smooth throughout, though the verrucose distal part of the stalk of the Kei Islands species may be unnatural. The likeli- hood of these two specimens being conspecific with the holotype of Monachocrinus incrassatus is discouraged by the occurrence in the same haul of the bathycrinid Democrinus chuni — well known from South Africa, which is half-way to St Helena but where Porphyrocrinus has not been taken. It seems necessary therefore to distinguish a new species. Porphyrocrinus polyarthra* sp. nov. (Fig. 6a-g) MATERIAL. Cape Town University no. AFR.A1248I, 36°48’ S : 52°08’ E (southern Indian Ocean, about half-way between South Africa and Amsterdam * Many-joints, referring to the multiple consecutive rigid joints in the distal part of the stalk. NEW RECENT SEA-LILIES 283 Island to the east), 400 metres.t Holotype with a complete stalk [to be deposited in the South African Museum], paratype [B.M.reg. no. 1972.12.5.7] incomplete distally, both lacking the arms after the first post-radial ossicle; also present several detached pieces of stalk and two arms. Dracnosis. A species of Porphyrocrinus with the columnals smooth throughout, up to seven consecutive distal ones rigidly jointed by synostosis, usually immedi- ately above the multilobate terminal plate ; ,numerous proximal columnals dis- coidal (c.17 when the topmost stalk diameter is 1-3 mm); calyx approximately cylindrical, but slightly constricted medially, flaring gently above and below (at least in the holotype and paratype but this may be subject to variation), the inter- basal and interradial sutures distinct, the radials appreciably longer than the basals ; the first pinnule probably P, on Bryp. DEscRIPTION. The holotype is the first specimen in Table 3. The distal end of its stalk terminates in a much expanded plate with finger-like lobes closely applied around a piece of scleractinian coral. The six distalmost columnals are rigidly united by successive synostoses, superficially appearing as an interrupted line like a syzygy. The disarticulated joint faces (Fig. 6g) show about 15 radiating shallow grooves around a large figure-of-eight-shaped concavity. The alignment of the stalk bends round through about go° just above the terminal plate. Two segments with flattened, alternating, flexible synarthrial joints follow, their maxi- mum breadth 21mm. Then come five more consecutive synostosial joints suc- ceeded by bollard-shaped segments about twice as long as their maximum (articular) breadths, the joints slightly elliptical in alternating planes, making up most of the length of the stalk. Towards the top (proximally) the columnals shorten so that the twenty-fifth one from the calyx is about as high as broad and the seventeenth and those above it can be called discoidal. At the very top the stalk broadens again after narrowing. There are about go columnals altogether and the length is c. 150 mm. TABLE 3 Numerical data from the holotype (1) and paratype of Porphyrocrinus polyarthra sp. nov. I 2 Stalk : Number of discoidal proximal segments 17 II Number of segments in top 5mm 22 18 Diameter at extreme top (= bottom diameter of basals) 13 I'2 Minimum diameter just below top 0-9 (oe) Calyx : Diameter at top of basals Diameter at top of radials 14 I°4 Height of basals (maximum) 0:8 0-6 Height of radials (maximum) I'l I'l + This locality is the highest point yet known on the South-west Indian Ocean submarine ridge. 284 AILSA M. CLARK Fic. 6. a-g. Porphyvocrinus polyarthva sp.nov. a-c. Holotype. a. Distalmost part of the stalk attached to coral (stippled). b. Columnals 59 and 60 from top. c. Calyx with adjacent ossicles. d. Proximal part of loose arm from Br,, showing bases of P, and e. P,(tip missing). f. Pg, (tip missing). g. Face of distal synostosial joint from large broken stalk fragment. h. Naumachocrinus hawatiensis A. H. Clark. Holotype. U.S. National Museum cat. no. 29573. Outline of calyx with adjacent ossicles. [The scale equals 5 mm for a-f and h and 2 mm for g.] The calyx is very slender and slightly constricted at the upper end of the basals, though the radials flare out at the top, this even flaring being continued by the first brachials, which are broader than long. The sutures in the calyx are distinct, those between the basal and radial rings making a zig-zag line. Two detached arms were found in the crinoid debris taken at this station which are very likely to have come from this specimen or the paratype. [Although Democrinus chuni was also collected, its arms differ in tapering from Br, and have the first pinnule on Br,.} The first ossicles on both these two arms have a syzygy on the proximal face and their breadth corresponds with that of the distal end of the first post-radial ossicle still attached to the calyx, so I am fairly confident that they can be regarded as Br,s. They expand to a maximum breadth of 1-5 mm at Br,;, having well-marked lateral flanges, and only very slowly taper distally. Both measure c. 40 mm in length. Syzygies and muscular joints normally alter- nate from I + 2, 3 + 4 onwards but one of the two arms has extra muscular NEW RECENT SEA-LILIES 285 joints at 13-14, 46 - 47, 53 - 54 and 58 - 59, though the other only has one such joint at 59-60. The last brachial remaining in both is Br,,, ending in a syzygy and showing no terminal modifications. The first pinnule is on the left side of Br,) and can be regarded as P3. It probably had six or seven segments and a fairly attenuated tip tapering from a stout base; the joint between the first two segments is particularly flexible and this is even more true on the subsequent pinnules (see Fig. 6e and f). PARATYPES. The distal ends of two other stalks are present. One has the lobes of the terminal plate almost meeting around a cylindrical piece of coral and only two of the consecutive distalmost segments jointed by synostosis, though three further ones are similarly jointed after an interval of several synarthrial joints. The other is stouter and has the six distalmost joints rigid, curving through go° as in the holotype. The surface of all these pieces of stalk is quite smooth. The calyx with incomplete stalk is very similar to that of the holotype except for having fewer discoidal proximal columnals. AFFINITIES. The affinities of these specimens and justification for describing them as a new nominal species are recounted in the discussion of the genus Por- phyrocrinus. However, the use of the number of discoidal proximal columnals as a character of specific weight needs some qualification. Gislén (1925) counted nine of the topmost columnals of the holotype of Porphyrocrinus verrucosus as discoidal but I would say that the ninth one was already too thick to be so de- scribed, at least judging from his figure 2, leaving only 8 compared with 11 and 17 in the two slightly smaller specimens of P. polyarthra. Even so, the wide variation in these two and the likelihood that the number of discoidal proximal segments decreases with growth (as it does in bathycrinids and phrynocrinids) combine to suggest that this difference may not be significant. The holotype of P. verrucosus lacked the distal extremity of the stalk beyond the eighty-sixth columnal and Gislén only observed one distal synostosis, between segments 33 and 34. How- ever, it is just possible that more distal ones were overlooked since the last 30 segments had their articulations obscured by the ‘verrucose’ surface texture (though the shape should have indicated any absence from the usual alternating synarthrial jointing). If more nearly complete specimens from the vicinity of the Kei Islands also prove to show multiple distal synostoses and if the peculiar texture of the distal part of the stalk of the holotype of P. verrucosus turns out to be ab- normal, then the distinction of two species will be difficult. Naumachocrinus hawaiiensis A. H. Clark (Fig. 6h) Naumachocrinus hawaiiensis A. H. Clark, 1912a: 195-197; 1915, fig. 129. A figure of the calyx of the holotype in the U.S. National Museum is given here, since it is very small in Austin Clark’s figure. I could not detect the very short interbasal sutures. The basal ring is 0-3-0-4mm high; the decimal point was omitted in A. H. Clark’s description. 286 AILSA M. CLARK Family BATHYCRINIDAE Bather Bathycrinidae Bather, 1899: 922; Gislén, 1924:206-212; 1927:53-54; 1938: 3-14; A. M. Clark, 1970: 13. Although a number of diagnoses for this family have already been given (in- cluding one of my own which was largely comparative with the fossil Bourgueti- crinidae), a modified version for comparison with those for the three preceding families may be useful. Dracnosis. A family of Articulata with a rounded stalk, without true cirri, with some of the proximal columnals discoidal and inflexibly jointed by synostosis but with new segments formed only between the topmost one and the calyx, the middle columnals elongated and more or less cylindrical (sometimes barrel-shaped or else waisted), with synarthrial joints which distally become elliptical, alter- nating in alignment, giving a moniliform appearance, attached to the substrate by irregular, branching, jointed, radicular ‘ cirri’ arising from several of the distal- most columnals; the calyx thick-walled, relatively small and rather variable, ranging from almost cylindrical to inverted conical, the basals either separate or fused together, longer or shorter than the radials, which may also be distinct or occasionally fused with the basals ; the division series or arms following the radials continuing the same alignment, closely approximating laterally and obscuring the tegmen (or disc) from lateral view, syzygies numerous, mostly alternating with muscular joints though proximally there may be several pairs of successive mus- cular joints, pinnules lacking basally from the first two or three possible positions on each side of the arms, none markedly enlarged and all lacking conspicuous side or cover plates. Remarks. This family is normally linked with the fossil Bourgueticrinidae in the suborder Bourgueticrinina on account of the root-like stalk attachment com- mon to both (where known). However, the diagnosis just given is remarkably similar to that provided for the new family Porphyrocrinidae, except for the lobed attachment plate terminating the stalk in the latter (besides minor differences in the calyx). A thorough review of the relationships between the recent and fossil stalked Articulata is needed. Several of the families known only from recent material were omitted from the classification given by Ubaghs (1953). It is to be hoped that their affinities will be resolved by the long-awaited crinoid part of the Treatise on Invertebrate Paleontology. Democrinus chuni (Déderlein) Rhizocrinus chuni Doderlein, 1907 : 14-15, pl. 1 fig. 5, pl. 6 fig. 6. Democrinus chuni : Gislén, 1938 : 26-27; A. M. Clark, 1972 : 146-150, fig. 17. MATERIAL. Cape Town University no. AFR.A1248I, 36°48’ S: 52°08’ E (southern Indian Ocean, about half-way between South Africa and Amsterdam Island to the east), 400 metres. One specimen. NEW RECENT SEA-LILIES 287 Cape Town University no. ABD 11F [‘Anton Bruun’ cruise 7, st. 389C], 30°12’ S : 32°01’ E (off Durban), 1360 metres. One small specimen. The specimen from the southern Indian Ocean has about 100 mm length of stalk still attached to the calyx; probably this represents only about half the total length. The calyx has an extra transverse suture across the basal ring at about three-quarters of its height. The small specimen shows the usual immature inverted conical shape of the calyx. The stalk trifurcates at the twenty-fifth segment, about Io mm below the calyx. RanGe. The first record provides an extension of range for this species, other- wise known from east and south Africa. SUMMARY OF NEW TAXA Page Subfamily Calamocrininae, of family Hyocrinidae_ . : . : . 268 Genus Anachalypsicrinus of subfamily Hyocrininae . é é : . 269 Species Anachalypsicrinus nefertiti . é : : : : : 200 Genus Zeuctocrinus of family Phrynocrinidae . : : : : 270 Species Zeuctocrinus gislent . : : : : : : : a yh Family Porphyrocrinidae, split off from Phrynocrinidae_ . : : 5 ashe Species Porphyrocrinus polyarthra : : é 282 REFERENCES Acassiz, A. 1892. Calamocrinus diomedae, a new stalked crinoid — with notes on the apical system and homologies of Echinoderms. Mem. Mus. comp. Zool. Harv. 17 (2): 1-95, 32 pls. BatueEr, F. A. 1899. A phylogenetic classification of the Pelmatozoa. Rep. Br. Ass. Adumt Sci. 1898 : 916-923, I fig. CARPENTER, P. H. 1884. Crinoidea. 1. The stalked crinoids. Rep. scient. Res. Voy. “ Challenger’, Zool. 11 (32) : 1-440, 62 pls. Crark, A. H. 1907. Two new crinoids from the North Pacific Ocean. Proc. U.S. natn. Mus. 32 : 507-512, 2 figs. 1911. Thalassocrinus, a new genus of stalked crinoids from the East Indies. Proc. U.S. natin. Mus. 39 : 473-476. 1912a. Naumachocrinus, a new genus belonging to the crinoid family Phrynocrinidae. Proc. U.S. natn. Mus. 42 : 195-197. 1912b. Echinoderma of the Indian Museum. 7. Crinoidea. Calcutta : 1-325, 61 figs. — 1915. A monograph of the existing crinoids. Vol. 1. The comatulids. Part 1. Bull. U.S. natn. Mus. 82: 1-406, 513 figs., 17 pls. CrarK, A. M. 1970. Echinodermata. Crinoidea. Marine Invertebrates of Scandinavia. No. 3. Oslo: 1-55, 19 figs. — 1972. Some crinoids from the Indian Ocean. Bull. Br. Mus. nat. Hist. (Zool.) 24: 73-156, 17 figs. DépErtEIN, L. 1907. Die gestielten Crinoiden der Siboga-Expedition. Siboga-Exped. 42a : 1-52, 12 figs., 23 pls. 288 AILSA M. CLARK GisLén, T. 1924. Echinoderm studies. Zool. Bidr. Upps. 9: iv + 316, 355 figs. 1925. Two new stalked crinoids from the Kei Islands. Vidensk. Meddy. dansk. naturh. Foren. 79 : 85-95, 22 figs. 1933. A small collection of crinoids from St. Helena. Vidensk. Meddr. dansk. naturh Foren. 93 : 475-483, 9 figs., 1 pl. 1938. A revision of the recent Bathycrinidae, with a study of their phylogeny and geographical distribution. Lunds Univ. Arssky. N.F. 34 (10) : 1-30, 18 figs. 1939. On the young of a stalked deep-sea crinoid and the affinities of the Hyocrinidae. Lunds Univ. Arsskr. N.F. 34 (17) : 1-18, 12 figs. KorxHter, R. 1909. Echinodermes provenant des campagnes du yacht ‘ Princesse Alice’. Result. Camp. scient. Prince Albert 34: 1-317, 32 pls. Koruier, R. & BaTHER, F. A. 1902. Gephyrocrinus grimaldii, crinoide nouveau provenant des campagnes de la ‘ Princesse Alice’. Mém. Soc. zool. Fy. 15 : 68-79, 4 figs. Matsumoto, H. 1913. On a new stalked crinoid from the Sagami Sea (Phrynocrinus ob- tortus). Annotnes zool. jap. 8 : 221-224. Artsa M, Crark, M.A. Department of Zoology British MusrEumM (NatuRAL History) Lonpon SW7 5BD PLATE 1 Anachalypsicrinus nefertiti gen. & sp. nov. Holotype. b. Large normal paratype. c. Abnormal paratype. d. Small paratype. (All x 1.) © Bull. By. Mus. nat, Hist. (Zool.) 25, 7 >| 4 % %j 7 ? : = = ths PLATE 2 Zeuctocrinus gislent gen, & sp. nov. a. Holotype, showing radius with division series of four ossicles (x 2). b-f. Holotype viewed from opposite side and four paratypes (specimens 1~5 in Table 2) (x 1). _ ia Bull. Br. Mus. nat. Hist. (Zool.) 25, 7 uuu = = — S A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. ALison. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75. 3 WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepéde, Cuvierand Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. a TayLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp.125; 29 Plates 77 Text-figures. 1969. £4.50. es Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. a oe Mitre Ke t ay Lage BNET... HIST 15 OCrig#S Nintoiiag thie « 4 Dey yi>> Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 5NU RELATIONSHIPS OF THE PALAEARCTIC LIZARDS ASSIGNED TO THE GENERA _ LACERTA, ALGYROIDES AND _ PSAMMODROMUS (REPTILIA: LACERTIDAE) ay RELATIONSHIPS OF THE PALAEARCTIC LIZARDS ASSIGNED TO THE GENERA LACERTA, ALGYROIDES AND PSAMMODROMUS (REPTILIA: LACERTIDAE) BY EDWIN NICHOLAS ARNOLD Pp. 289-366 ; 15 Text-figures, 4 Tables BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 25 No. 8 LONDON: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 25, No. 8 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation : Bull. Br. Mus. nat. Hist. (Zool.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 19 October, 1973 Price £3.35 RELATIONSHIPS OF THE PALAEARCTIC LIZARDS ASSIGNED TO THE GENERA LACERTA, ALGYROIDES AND PSAMMODROMUS (REPTILIA: LACERTIDAE) By EDWIN NICHOLAS ARNOLD CONTENTS Page INTRODUCTION . > : i A 2 3 4 : : 292 THE PRESENT CLASSIFICATION OF Lacerta, Algyroides aNnD Psammodromus 204 CHARACTERS ALREADY USED OR POTENTIALLY USEFUL IN CLASSIFYING Lacerta, Algyvoides AND Psammodromus 300 Osteological characters 300 Hemipenial characters 306 Chromosomes ° 6 és 310 External morphology and colouring 310 Hybridization . : ; 5 : 315 RELATIVE TAXONOMIC VALUE OF CHARACTERS 315 Introduction : 3 : 5 315 Ecologically correlated characters 317 Ventral scaling . 318 Collar 320 Dorsal scales : : . . : : . 5 320 Characters distinguishing the subgenus Aychaeolacerta . ; : 320 Discussion 2 5 : : 0 323 Hemipenial characters c : . : : 2 324 RECLASSIFICATION OF Lacerta, Algyroides AND Psammodromus : ; 32 Laceyta Linnaeus 1758 330 Lacerta part I 331 Lacerta part II 334 Algyroides Bibron & Bory 1833 341 Podarcis Wagler 1830 343 Psammodromus Fitzinger 1826 347 Gallotia Boulenger 1916 : - é . : : F : 349 RELATIONSHIPS OF Algyroides, Lacerta, Podarcis, Psammodromus AND Gallotia c 5 c : : ‘ 355 SUMMARY OF PROPOSED SYSTEMATIC CHANGES 357 ACKNOWLEDGEMENTS 358 REFERENCES 5 5 a 3 A 4 6 : 359 APPENDIX I. RELATIONSHIPS OF THE AFRICAN SPECIES oF Algyroides . 363 AppENDIXx IJ. MATERIAL EXAMINED : 4 2 9 : 304 APPENDIX III. VARIATIONS IN PRESACRAL VERTEBRAL NUMBER . z 365 SYNOPSIS It is pointed out that the equatorial and southern African s pecies of Lacerta Linnaeus 17 58 and Algyroides Bibron & Bory 1833 are not at all closely related to the palaearctic members of these genera. The present classification of the palaearctic sp ecies of Lacerta, Algyroides and 292 E. N. ARNOLD Psammodvomus Fitzinger 1826 is discussed and found to be unsatisfactory in several respects. Characters available for grouping species are described, including a number of features (mainly skeletal and hemipenial) not employed before. Using ecological and functional data, an attempt is made to assess the lability of available characters : many of the external and cranial features used in ‘ classical’ lacertid systematics appear to be potentially labile and therefore must be given low comparative weight in judging relationships. Using this comparative weighting, it appears that Algyroides and Psammodromus are natural groups and that Lacerta is divisible into four main sections. Two of these will be retained in Lacerta (as Lacerta parts I and I1) but the others are raised to the status of separate genera: Gallotia Boulenger 1916 and Podarcis Wagler 1830. Proposed systematic changes are listed in full on p. 357. INTRODUCTION AT LEAST fifty-five species are currently assigned to the genus Lacerta.* Of these, forty-nine are limited to a relatively small part of the southwestern Palaearctic region while two others, L. vivipara Jaquin and L. agilis Linnaeus, occur not only in this area but also range more widely in Eurasia (see Fig. 1). The remaining four members of the genus are found in Africa south of the Sahara desert. These are: L. echinata Cope (tropical forests of West Africa from Liberia to the eastern Congo Republic), L. jacksoni Boulenger (eastern Congo Republic, Uganda, Kenya and Tanganyika), L. rwpicola Fitzsimons (Zoutpansberg, Transvaal, Republic of South Africa), and L. australis Hewitt (near Ceres in Cape Province, Republic of South Africa). JL. echinata and L. jacksoni have strong affinities with their Central African neighbours in the genera Algyroides, Bedriagaia and Gastropholis, and L. rupicola seems to be most nearly related to Tvopidosaura. It does not seem possible to retain these species in Lacerta and their precise relationships will be discussed elsewhere. L. australis is apparently known only from the single type specimen (which is not available to me) and is not well enough studied for its real affinities to be clear. But, on geographical grounds, it is not likely that L. australis is closely related to the Palaearctic species of Lacerta for the type locality is sep- arated from the range of this assemblage by a hiatus of over 7000 km. In the rest of this paper, Lacerta will be used to refer only to the fifty-one Palaearctic species. The classification of Palaearctic Lacerta has presented a number of difficulties. Many characters used in their systematics exhibit great intraspecific variation making the delimitation of species-boundaries difficult. Some of this extensive variation is geographic ; for instance, Mertens & Wermuth (1960) list no fewer than 31 subspecies of L. erhardii, 32 of L. pityusensis and 39 of L. sicula and many subspecies of the last have been recognized since (Brelih 1961, 1963, Lanza 1966, Lanza & Borri 1969, Mertens 1966). Many forms also show considerable intra- populational variability especially in colour and pattern. Notwithstanding this, classification at the species level is now fairly stable thanks to the investigations of a long series of workers including Eimer (1881), Bedriaga (1886), Werner (1904), * This figure includes Lacerta cappadocica Werner which has sometimes been placed in a separate genus, Apathya Méhely, and four parthenogenetic forms allied to L. saxicola Eversmann, viz. L. aymeniaca Méhely, L. dahli Darevsky, L. rostombekovi Darevsky and L. unisexualis Darevsky. It is very probable that several other species will eventually be recognized. fl Fic. 1. Distribution of lizards usually assigned to Lacerta, Algyvoides and Psammodvomus. Black: area containing Psammodvomus, European Algyroides and 51 species of Lacerta (49 of which are confined to the region marked). Diagonal hatching: distribution of Lacerta vivipara outside the main range of Palaearctic Lacerta. Broken line: distribu- tion of Lacerta agilis beyond main range of Palaearctic Lacerta and of L. vivipara. Vertical hatching: combined distribution of ‘ Lacerta’ echinata, ‘ Lacerta’ jacksoni and the African species customarily assigned to Algyvoides. Solid black circle: only known locality of ‘ Lacerta’ rupicola. Black triangle : only known locality of ‘ Lacerta’ australis. Boulenger (1905, 1913, 1916, 1920), Méhely (1907, 1909, Ig10), Schreiber (1912) and more recently Mertens (numerous publications mainly in Senckenbergiana Biologica), Klemmer (1957), Peters (1962a, b) and Darevsky (1957, 1966, 1967). Problems still remain in defining acceptable species-groups within the genus, despite several attempts having already been made to divide Lacerta into subunits, and comprehensive schemes of classification having been put forward by Bedriaga (1886), Werner (1904) and Boulenger (1916). The last author divided the Palae- arctic forms of Lacerta into five sections or subgenera: I — Lacerta s. str., II - Gallotia Boulenger, III - Zootoca Wagler, IV —Podarcis Wagler and V —Thetia Gray which is, in fact, a junior synonym of Scelarcis Fitzinger. Méhely (1907, 204 E. N. ARNOLD 1909, 1910), who dealt only with the small, mainly climbing forms commonly known as Wall Lizards separated these into two groups, ‘Archaeolacertae’ and ‘ Neo- lacertae’. But Boulenger (1907, Ig10) was very critical of Méhely’s division and did not recognize it in his own classification when he placed all Wall Lizards in Section IV (subgenus Podarcis). Subsequent investigations, however, have sup- ported Méhely’s division and Podarcis of Boulenger is now divided into the subgenera Podarcis s. str. and Archaeolacerta Mertens, which are respectively more or less equivalent to Méhely’s Neo- and Archaeolacertae. A seventh subgenus, Apathya Méhely, is sometimes recognized for L. cappadocica (e.g. by Mertens 1952). This classification is not entirely satisfactory for the borders of some subgenera are not well defined and a number of species have either never been properly assigned to a particular subgenus or their position has been recently questioned. Further- more, many of the characters by which the subgenera are distinguished are probably quite labile and on their own do not provide an adequate basis for arranging species into natural assemblages (see pp. 315). Such characters may turn out to delimit natural groups but confirmation from other features is necessary. At present, subgeneric classification is mainly based on external features plus a few variations in skull structure. It seems essential to re-assess these characters and, more im- portant, to increase the range of variables available for classifying Lacerta and its allies. This increase should involve not only the absolute number of characters used but also the number of sources from which they come. For this reason, I have investigated the skeletal and hemipenial morphology of these lizards. It seems convenient to discuss the relationships of two other genera in conjunc- tion with Lacerta, namely Algyroides Bibron & Bory and Psammodromus Fitzinger. Algyroides, as presently understood, is like Lacerta in having a disjunct range with four species in Europe and three in central Africa (viz. A. africanus, A. alleni and A. vauereSelli). As with Lacerta, the equatorial species have no close affinities with their Palaearctic congeners and must be removed from Algyroides (see Appendix I). Both Algyrotdes and Psammodromus (four species in northwest Africa and southwest Europe) are usually regarded as being closely related to Lacerta. Algvyroides differs externally from Lacerta only in possessing strongly enlarged dorsal scales ; Psammodromus also has enlarged dorsal scales but differs from both Algyroides and Lacerta in having the collar reduced or absent. This paper consists of four sections: (i) a statement of the present classification (p. 294) ; (ii) a description of the principal characters that have either already been used to classify Lacerta, Algyroides and Psammodromus or that appear to be potentially useful (p. 300) ; (iii) an assessment of the relative importance of these characters (p. 315) and (iv) an attempt to revise the classification (p. 327). THE PRESENT CLASSIFICATION OF LACERTA, ALGYROIDES AND " PSAMMODROMUS Genus LACERTA Linnaeus 1758 Until very recently, the subgeneric allocations of forty (almost 80 per cent) of the species of Lacerta were widely accepted and these are listed below. The characters RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 295 on which the subgenera of Lacerta are based are set out in Table I (p. 296) together with those of Algyroides (Palaearctic forms) and Psammodromus. Subgenus Lacerta s. str. Linnaeus 1758 (type species: L. agilis) agilis Linnaeus 1758 (West and central Europe to central Asia), schreiberi Bedriaga 1878 (Iberian Peninsula), sérigata Eichwald 1831 (Caucasus area ; West Iran), trilineata Bedriaga 1886 (Southeast Europe and Southwest Asia), viridis (Laurenti 1768) (central and southern Europe etc.). Subgenus Gallotia Boulenger 1916 (type species: L. gallotz) atlantica Peters & Doria 1882 (eastern Canary Islands), gadloti Duméril & Bibron 1839 (western Canary Islands, not Gran Canaria), simony1 Steindachner 1889 (Gran Canaria, formerly Roques Zalmor off Hierro, western Canary Islands). Subgenus Zootoca Wagler 1830 (type species: L. vvipara) andreanszkyi Werner 1929 (Atlas Mountains), praticola Eversmann 1834 (Caucasus and northern Balkan Peninsula), vivipara Jaquin 1787 (northern Eurasia). Subgenus Podarcis Wagler 1830 (type species : L. muralis) dugesii Milne-Edwards 1829 (Madeira), erhardii Bedriaga 1882 (southern Balkan Peninsula and Greek Islands), filfolensis Bedriaga 1876 (Malta and nearby islands), hispanica Steindachner 1870 (Iberian Peninsula and North- west Africa), lilfordi (Ginther 1874) (Balearic Islands), melisellensis Braun 1877 (eastern Adriatic region), milensis Bedriaga 1882 (Milos and nearby islands), muralis (Laurenti 1768) (South and central Europe), pityuwsensis Bosca 1883 (Ibiza and nearby islands), sicula Rafinesque 1810 (mainly Italy, Adriatic and Tyrrhenian areas), taurica Pallas 1814 (southeastern Europe), tiliguerta Gmelin 1789 (Corsica and Sardinia), wagleriana (Gistel 1868) (Sicily). Subgenus Archaeolacerta Mertens 1921 (type species: L. bedriagae) armeniaca Méhely 1909 (Caucasus), bedriagae Camerano 1885 (Corsica and Sardinia), caucasica Méhely 1909 (Caucasus), dahli Darevsky 1957 (Caucasus), danfordi (Giinther 1876) (western Turkey and south to Petra), graeca Bedriaga 1886 (southern Greece), horvathi Méhely 1904 (northwestern Jugoslavia), monticola Boulenger 1905 (Iberian Peninsula), mosorensis Kolombatovi¢ 1886 (southwestern Jugoslavia), oxycephala Duméril & Bibron 1839 (southwestern Jugoslavia), vostombekovi Darevsky 1957 (Caucasus), rudis Bedriaga 1886 (Caucasus), saxicola Eversmann 1834 (Caucasus etc.), wnisexualis Darevsky 1966 (Caucasus). Subgenus Scelarcis Fitzinger 1843 (type species: L. perspicillata) perspicillata Duméril & Bibron 1839 (northwestern Africa). Subgenus Apathya Méhely 1907 (type species: L. cappadocica) cappadocica Werner 1902 (central Turkey to northwest Iran). Ajquonbaa., auou ua}O [eULION, 8-9 wor} eorqut 9/971] ‘re[Nsurz9y7 poyerrosuy) pajeayun pur yey uayoO avsviapaq “TT suonjrindod May e uy yensq, euri0N (8) 9 sora deproAo ‘sopsue -Joo1 oyewTxoiddy solrea ynq yeam Ajyensn woryerses peeex uazjo ‘pasrer Ayyens~) oN 21991] 10 ouON [eULION, 9 JOyeIM UOTPPOTIq WIT ynq “3148's vjA2InT sy pezye1I3aS pajaey uazjo ‘pasrey ON quonbar.t Jeur10 Ny oz—o1 (g) Joyeom wOrzeoTIq wt 3Nq ‘13S °s YA2IVT SY SOLIe A, pajeey pur pasrey oN quonberq [eulI0N 8-9 ayeoriqut Ajsuorjs ‘surei3 -ojayjered pourpouy peyer1eS payeey ‘ayeoquir JoyyeI ‘pasreyy Sunok oy} ur parnojoo A];4811q [rey ui9}3ed [esiop ut wistydiounp Tenxag Seporeqn} [eySrpqns SMOI [PUIPNzSu0] Jo Joquinu :Suryeos jeryua, sayeos jo adeys : Suryeos yeryua, TETIOD sayeos [es10qy PIerys yejorred so}yeUIsIeUl aleos preaq (1261 a ur suorydaoxa ‘sax ONG SOLA, SOLA sox jesodwiayesdns 4sa1y ° Ayeos Ayeas Ayeas Ayeas Ayes a (sijzsp -7 owos 4 a[qeue, soseo sour ur ‘auQ auO Eliv@) Ur U0) OM} ATTeNS SPIerys jeseujsog Z saynurds : avapiapaq *T whyzsuvaapun UI Surpus uazjo ‘sautds 4 pure voavas “7 “7 ut soutds BuO] Y}IM syueYaqoyT osuryog wo1f eyep Yyonut) ur ydooxe ‘sajoraqn} —(12sasnp “7 you) sourds paainoox ynq sapo.1oqnyz soutds ‘soporeqny podeys wnqeyjzide peruodimoy podeys-umo1y paainoaz aydunis poedeys-umory peainoax ‘aydung = -umor9 YIM sdy aqoT JO UOI}e} UOUTeUIO-OINTT auou sAeMye jsowly a1qeiiea dUuON jens, jens, Y}00} prosA10}q povundsiy “J pue saseo 2sasnp “7 ut suodaoxa WWyzsuvadpUuDY “T DULOS ysour ut az0[/dur0sut ‘aqyojduroo Ayyensq. ur jdaoxa azaydur05 ajojdui09g ayardur09 apyjoure, reynooeidng uly. aqqerie A UI} IO a}yeIOpoy HOI HOY L SuIapos3so [eIUeIa xoAu0o yey yoor ‘possoidap Ajyensn joor ‘pyinq X9AU09 JOOI xaAu09 xaauoo 4tinq AysryT 3451-03-97 e19poy ‘pring 3481 Apare.y yoor ‘daap ‘aatssepy yoor ‘daap ‘aatsseyy SOTOPF [[NYS (}u9A-jnous) azis wut Sg—¢ wu o6-Sb wrt £9-SE wu + O01zZ—09 turogt-oL = Apoq 3[npe ozeutxoiddy VIAIIDIOIDYIA SUIADPOT 0904007 v1yoyvy “I]S ‘'S DLBQIVT ro} 2 @ (uostreduroo Joy papnyout aie snwospoumuvsg pure sapi04d5)f) DIAIIVT JO vIDUISQNs oy} YsInSurystp 07 pasn Aqsnor Aoid siopoerey) :] ATavy 297 RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS ON ouON pajaey 10 [eUWION (gh€ -d aas) o1-9 Iosuo01}s UdAd dvji9a0 10 “1}S 's vjAIIVT SY quasqe 10 yRAM AIDA IeT[OD payutod pure pajooy ‘ayeoriquyr ‘os1e 7] ON Ayeos auo Ayyensy¢) aTqere, aTqerie, ayo[durog YI} 07 97eIOPOT X9AU09 yoor ‘pjing Aaray-0}-a}e1Opoy wut og—S& SNUOAPOUMDS JT ON snayoasou * py ut ydooxa ‘ouON [eulION, 9 s19appog Se + a[qeiIeA uoryeIIaS pojzutod u93jo ‘pajaoy ‘ayeoriquir ‘osre7] sox Ayeas om} ATrensy, a[qerie arqere,, arqrie,, a}yeIOpOUL 0} UIY TL atqere a urut oL—of sapioasts] y sox ouON payaex Apsuorys ATpensp, 8-9 uorzeoTIq WIT aaa] ‘repnBuez0y7 poyeriasuy) peyeeyun ‘paster A[SU0I}sS JON poyeursieura Apreapo JON sayeos [e1oAes yO MOpuIM juoredsuei 9014} 10 OMT soutds paainoos ofdurts ouON ajyo[duioour ATens~, ur yee yoor ‘passoidap ‘4,Inq ATVYysBrT ull og—oS SOR auON [euION or uorzeoTIq UIT apy ‘repn8uez 97 poyeriosut) peyeeyun ‘paster A[Suo1}s JON ON ayeos ouo jo MopurM juoredsuery OM], soutds poainoas a[duits aUON aqenea ury} Apireq yey uoqjo joor ‘pling 7481]-0}-97e19poW wriut og—-oF Bunod oy} ur parnojoo ATVYsIIq [re], uia}}ed [esiop ut wistydrourp penxas sapo1eqn} [e}ISIpqus SMOI [CUIPN}ISUOT JO Joquinu : Surpeos ~e1zUa A saTeos jo odeys :Surjeos jerzuo,, teTIOD sayeos [es10cT Plerys jeqorsed soyeursrewo ayeos jerodurazeidns 4si1q preAg spjarys jeseujsog (1261 aulyod Woy eyep yonu) urnteyyide peruodruay JO U01} 2} USUeUIO-O1NIT, yq00} prosA103q4 oR][ourey repnooeidns SUIINPO9}SO [eIUeID Soroey [INAS (qu9A—jnous) ozIs Apoq }[npe oyewurxoiddy vdywd S19AD]0IS 298 E. N. ARNOLD The precise systematic positions of the following eleven species are less certain. L. lepida Daudin 1802 (southwestern Europe and northwestern Africa) L. parva Boulenger 1887 (Turkey and Armenia) L. princeps Blanford 1874 (eastern Turkey to southwestern Iran) Boulenger (1916) placed these species in the subgenus Lacerta s. str., but Peters (1962) believed that they had no close relationship with the other forms in this subgenus. Indeed Peters (1961) put L. lepida in the subgenus Gallotia and (1962a) suggested that the affinities of L. parva lie with L. fraasii (q.v.). More recently, Eiselt (1968) has again placed L. princeps close to the members of the subgenus Lacerta s. str. L. fraasii Lehrs 1910 (Lebanon) Boulenger (1920) emphasized the similarity of this species to L. vivipava and even thought that the two forms might be conspecific. He consequently placed L. fraasti and L. vivipara in the subgenus Zootoca. Peters (1962a) disputed this close relationship and instead pointed out the likeness between L. fraasii and L. parva (see above). He suggested that, eventually, a separate subgenus might have to be created for these two species and also discussed their suitability as ancestors of the Podarcis group. L. brandtii De Filippi 1863 (northwestern Iran) This species was regarded by Boulenger (1920) as “‘ one of the most primitive members of the L. muralis group [i.e. Podarcis in Boulenger’s sense] which it con- nects with L. parva”. Wettstein (1951), on the other hand, believed that L. brandtii was connected with L. vivipara and L. fraasii and allocated it to the sub- genus Zootoca. Peters (1962) in his discussion of L. parva states that the possibility of a relationship with L. brandtii should not be dismissed, and Bohme (1971) also believes this. « L. cyanura Arnold 1972 (southeastern Arabia) L. jayakarvi Boulenger 1887 (southeastern Arabia) In his Monograph of the Lacertidae (1920) Boulenger placed L. jayakari in Podarcis as he understood that subgenus, and regarded L. Jaevis as its closest relation. He suggested that both these species might have been derived from ancestral forms close to L. brandtii. Klemmer (1957) regarded the position of Lacerta jayakari as uncertain. The relationships of this form and the recently discovered L. cyanura have been briefly discussed in a recent paper (Arnold 1972). L. peloponnesiaca Bibron & Bory 1833 (southern Greece) According to Boulenger (1920) this species is closely related to L. tawrica which is now placed in Podarcis s.str. Klemmer (1957) regarded L. peloponnesiaca as incertae sedis because it has a very robust skull with heavily ossified temporal areas atypical of the subgenus Podarcis. But Buchholz (1960) pointed out that RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 299 this difference is not clear-cut since the degree of temporal ossification in L. pelopon- nesiaca is very variable. L. derjugini Nikolsky 1898 (Caucasus) Méhely (1909) placed L. derjugini in his ‘Archaeolacertae’ and Lantz & Cyrén (1947) followed this allocation, regarding L. derjugint as a member of the subgenus Archaeolacerta. Other authors list this species as a member of Zootoca, e.g. Boulenger (1920), Mertens & Miiller (1928, 1940), Klemmer (1957). The last author considered that it might link Zootoca with Archaeolacerta. L. chlorogaster Boulenger 1908 (southern Caspian region) L. laevis Gray 1838 (East Mediterranean coastal region) Like L. derjugini these two species were placed by Méhely (1909) in his ‘ Archaeo- lacertae’ and most authors have followed this course. However, Mertens (1957) pointed out that L. chlorogaster is more like the members of the subgenus Podarcis in its skull and body shape and in having keeled dorsal scales and he tentatively transferred it to this subgenus. Similar arguments could be applied to L. laevis which resembles L. chlorogaster in these respects. Genus ALGYROIDES Bibron & Bory 1833 (type species: A. moreoticus) fitzingert (Wiegmann 1834) (Corsica and Sardinia), marchi Valverde 1958 (southeastern Spain), moveoticus Bibron & Bory 1833 (southern Greece and Ionian Islands), A. nigropunctatus (Duméril & Bibron 1839) (eastern coastal area of Adriatic Sea). Genus PSAMMODROMUS Fitzinger 1826 (type species: P. hispanicus) P. algirus (Linnaeus 1758) (southwest Europe and northwest Africa), blancz (Lataste 1880) (northwestern Africa), hispanicus Fitzinger 1826 (southern France and Iberian Peninsula), microdactylus (Boettger 1881) (northwestern Africa). Recent work Very recently, Béhme (1971) has challenged certain aspects of the accepted classification of Lacerta and its allies on the basis of differences in structure of the hemipenial micro-ornamentation (see p. 309). Amongst his principal findings are the following. 1. L. lepida and L. princeps appear to be closer to the subgenus Gallotia than to Lacerta s. str. 2. Archaeolacerta may be an unnatural group as two species (bedriagae and graeca) have a different pattern of micro-ornamentation from the others. 3. Archaeolacerta (with the exception of the two species mentioned in paragraph 2) and Zootoca are not clearly distinguishable and should be merged. 4. Psammodromus and Algyroides are probably not natural groups. 300 E. N. ARNOLD CHARACTERS ALREADY USED OR POTENTIALLY USEFUL IN CLASSIFYING LACERTA, ALGYROIDES AND PSAMMODROMUS In these descriptions, subgeneric names refer only to those species of Lacerta definitely assigned on p. 295. Species of uncertain position are mentioned separately. Osteological characters 1. General shape of the skull (Fig. 2) Most lizards in the genera Lacerta, Algyroides and Psammodyomus have a rela- tively deep, robust skull in which the parietal table has convex lateral edges and is domed in transverse section. But, in contrast, some species of Lacerta and Algyroides have a flattened, delicately built skull and a parietal table with con- cave or straight lateral borders and a flat transverse section. Many intermediates exist between these two extremes. 2. Nasal openings of the skull (Fig. 2) In many species the nasal opening is small. It may expose only the external vestibule of the nasal passage but more usually allows the anterior part of the principal nasal chamber to be seen from above. In intact animals it is possible to palpate the nasal opening which in these cases either does not reach backwards under the frontonasal scute or only extends under its anterior borders. In other forms the nasal bones do not extend so far forward and the exposure of the prin- cipal nasal chamber is much greater so that the septomaxilla is visible from above. In such cases the openings may extend backwards under the posterior borders of the frontonasal scute or beyond. 3. Development of cranial osteoderms Cranial osteoderms (the crusta calcarea) are developed mainly during the period between hatching and maturity and form a continuous layer, closely applied to the dorsal and lateral bones of the skull. They also may develop in previously unossified regions such as the skin over the orbits and in the temporal region. The cranial osteodermal layer is laid down in discrete sections each corresponding to an epidermal scute so that, as the osteoderms increase in thickness, the sutures between the scutes are left as distinct grooves. The development of osteoderms on the skull roof shows a rough correlation with the robustness of the skull; de- pressed, delicately built skulls have relatively weak osteoderms. 4. Supraocular lamellae (Fig. 2) The dorsal surface of the orbit is unossified in hatchling lizards but eventually is wholly or partly occupied by a series of four osteoderms (the supraocular lamellae) corresponding to the four supraocular scales. These begin to develop near the lateral borders of the frontal bones and slowly extend outwards. In adults of most species the supraocular osteoderms cover the whole orbital area but in a num- ber of forms the second, third and sometimes the fourth osteoderms are usually RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 301 1 2 Fic. 2. Principal skull types: dorsal view and cross-profile. (1) Robust undepressed skull with small external nares, complete supraocular lamellae and transversely arched parietal region. (2) Delicately built, strongly flattened skull with large external nares, fenestrated supraocular lamellae and flat parietal region. po = postorbital bone. pf = postfrontal bone. incomplete so that a flexible area of unossified skin remains between them. This occurs in Algyroides fitzingeri and A. marchi and in many Archaeolacerta spp., in Lacerta cyanura, L. (Apathya) cappadocica and some populations of L. (Podarcis) hispanica and L. (Scelarcis) perspicillata, in females of L. (P.) dugesii and in many L. andreanszkyi, L. derjugini and some L. fraasit. 5. Ossification of the temporal region Osteoderms may occur in the skin covering the jaw muscles and bordered by the jugal, postorbital, squamosal and quadrate bones and the lower jaw. Such osteoderms are well developed in adults of most Lacerta s. str. species (not agilis or some populations of ¢trilineata), L. lepida, L. princeps, L. (G.) atlantica, some 302 E. N. ARNOLD L. (G.) simonyi, and L. peloponnesiaca and in Psammodromus algirus. Minor development (particularly along the posterior border of the jugal) may occur in individuals of other species especially in Podarcis (commonest in L. melisellensis and L. tawrica). 6. Pterygoid teeth Teeth are found on the pterygoid bones of all species of the subgenera Lacerta s. str. and Gallotia, in L. lepida, L. princeps, L. peloponnesiaca, L. brandtii, L. jayakari and L. laevis. They also occur in Psammodromus algirus and some Algyroides moreoticus and A. nigropunctatus. Klemmer (1957) has given data for various species in the subgenus Podarcis. In this group, some species rarely have pterygoid teeth, in others they occur in a substantial number of individuals and in L. (P.) milensis and L. (P.) taurica they seem to be almost universally present. In Lacerta as a whole there appears to be an imperfect correlation between presence of pterygoid teeth and robustness of the skull. 7. Postfrontal and postorbital bones (Fig. 2) These bones are separate in the hatchlings of the majority of Lacerta species but fused in the three members of the subgenus Gallotia, in L. (Z.) vivipara and in L. schreiberi. They are also fused in all four species of Psammodromus. In some other forms, these two elements coalesce during life, e.g. L. lepida, L. dugesiti and in old individuals of several other species. 8. Number of presacral vertebrae Presacral vertebral number varies both between species and within them (see Appendix III), the usual range for Lacerta being 25 to 29* vertebrae, for Algyroides 24 to 28 and for Psammodromus 26 to 28. Intraspecific variation includes geo- graphical, sexual and intrasexual differences. In each sex of a given species, most individuals usually have the same number of presacral vertebrae (occasionally two numbers are relatively common), the remainder deviate from this figure by one or occasionally two vertebrae. Typically the total range for each sex of a species is not more than three vertebrae. The commonest presacral vertebral numbers of each sex generally differ, females usually having on average one more vertebra than males (occasionally it is two more) ; 26 and 27 are the commonest figures for males, 27 and 28 for females. The three species assigned to the sub- genus Gallotia differ from this pattern in showing almost no sexual or individual variation in vertebral number. This condition is rare in the Lacertidae as a whole, occurring elsewhere only in a few species of Acanthodactylus. Of the specimens of Gallotia examined in this study (viz. 25 L. atlantica, 35 L. galloti and 27 L. simonyt), nearly all had a presacral vertebral count of 26 irrespective of sex. The exceptions were one male L. galloti and two female L. simonyr all of which had 25 presacral vertebrae. * Two out of 32 female L. (L.) agilis examined had 30 vertebrae as did one out of five female L. (Z.) praticola. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 303 g. Clavicle (Fig. 3) Some individuals of Lacerta, Algyroides and Psammodromus have clavicles that correspond to the condition found in the Lacertidae as a whole. In these the clavicle is greatly expanded near the midline and the expansion is penetrated by a large foramen so that the bone in this region forms a continuous loop (Fig. 3(2)). Cc —_ a ‘SS DIR ANS MK ??) 1 2 Fic. 3. Variations in clavicle, interclavicle and sternum. (1) Clavicle emarginated, inter- clavicle with lateral arms approximately at right angles to its longitudinal axis, sternum with heart-shaped fontanelle. (2) Clavicle unemarginated, interclavicle with lateral arms directed backwards, sternum with oval fontanelle. c = clavicle. i = inter- clavicle. s = sternum. In other individuals the clavicle is emarginated posteriorly so that the loop is inter- rupted (Fig. 3(1)). This interruption may be minor or the whole posterior section may be missing. In Psammodromus, the species of the subgenus Gallotia and pethaps in L. parva and L. fraasii (the small samples available, 17 and 6 respec- tively, do not permit certainty) the loop of the clavicle is complete in all specimens. In most other species both the intact and emarginated conditions occur, although the proportion of each varies from species to species. In a few Lacerta species 304 E. N. ARNOLD only the emarginated condition was observed: these include L. (P.) fiulfolensis, L. (P.) muralis, L. (A.) bedriagae, L. (A.) graeca, L. (A.) oxycephala, L. (S.) per- spicillata and L. brandtit. 10. Sternal fontanelle (Fig. 3) In all species of Lacerta, Algyroides and Psammodromus the sternum usually has a distinct fontanelle penetrating its posterior area, although it may be un- fenestrated in a few individuals of some species. The shape of the fontanelle is somewhat variable in most species but two main patterns are present. The majority of forms nearly always have an oval or round fontanelle while most members of Podarcis and L. peloponnesiaca have a heart-shaped (cordiform) one in which there is a well-developed, posteriorly directed process arising from the anterior border. Occasionally this extends right across the fontanelle to join its posterior margin, thus dividing it in two. Some individuals of L. laevis, L. danfordt, L. andreanszky1 and Algyroides moreoticus also have a sternal fontanelle which approaches a heart- shape but in these species the posteriorly directed process is not well developed. 11. Interclavicle (Fig. 3) This element is cruciform in all species of Lacerta, Algyroides and Psammodromus. The lateral extensions of the interclavicle generally run at about go° to the main axis of the bone although they may be directed slightly forwards or more rarely slightly backwards. L. (P.) dugesii and L. (S.) perspicillata differ in having the lateral extensions clearly directed obliquely backwards in all individuals examined. 12. Pattern of tail vertebrae (Fig. 4) The systematic importance of variations in the caudal vertebrae of lizards in general has been discussed by Etheridge (1967). Three main patterns are present in the Lacertidae all of which occur in the genus Lacerta. Each begins with a proximal series of non-autotomic vertebrae. The number of bones in this proximal series varies from four to seven, males of any species tending to have a higher average number than females. The highest numbers encountered occur in L. fraasit. Apart from the proximal series, the rest of the tail vertebrae are autotomic and the more anterior of these always bear transverse processes. It is these transverse processes which differentiate the three basic patterns (see Fig. 4). Pattern A. All vertebrae with transverse processes have a single pair anterior to the autotomic plane. Pattern B. The more anterior of the vertebrae with transverse processes bear two parallel pairs of which the anterior is longer. One pair lies in front of the autotomic plane, one pair behind. Normally from one to four vertebrae are involved although the number may exceptionally be up to seven. This pattern is restricted to Lacerta and the European species of Algyrozdes. Pattern C. Again the more anterior autotomic vertebrae bear two pairs of transverse processes, one each side of the autotomic plane. The members of the more anterior of these pairs run laterally or occasionally are directed slightly RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 305 ese eS apes oe oe a= ao aoe a5 xe x fl 3 is SS ee a Sie er a a oe nd OS omen al Sam 1 11 2 ty Fic. 4. Principal patterns of caudal vertebrae. (1) Pattern A: vertebrae with single pairs of transverse processes only. (2) Pattern B: some anterior autotomic vertebrae with two, more or less parallel pairs of transverse processes, the posterior pair being the shorter. (3) Pattern C: some anterior autotomic vertebrae with two diverging pairs of transverse processes, the anterior pair being the shorter. to w 306 E. N. ARNOLD forwards. The members of the posterior pair are often blade-like, usually longer than the front ones, and run obliquely backwards. From three to seven vertebrae are involved. The A and B patterns are most similar, sometimes differing only in as little as the latter having secondary processes on one vertebra. Furthermore, they fre- quently occur together in the same population so it seems best to treat them as one class. A or B caudal patterns, or both, are found in all members of Lacerta s. str., most Archaeolacerta, the species of Zootoca, L. (P.) dugesi, L. (S.) perspicil- lata, L. (Ap.) cappadocica, L. lepida, L. princeps, L. jayakari, L. cyanura, L. derjugini, L. laevis and L. chlorogaster ; they are also found in Algyroides. The C pattern occurs in all members of Gallotia, in most members of Podarcis (not L. dugesit), in L. peloponnesiaca, in L. (A.) danfordi (in the subspecies L. d. danfordi and L. d. anatolica but not in L. d. kulzeri), and in Psammodromus. L. parva, L. brandtii and L. fraasii have caudal patterns that are, to some extent, intermediate between the B and C types. L. parva. Two to six vertebrae with double processes (two to four most com- monly). The anterior pair is directed somewhat forwards in many individuals while the posterior pair is directed backwards ; in most cases it is larger than the anterior one but it may only be equal to it or even shorter. L. brandtii. Four to seven vertebrae involved (six being the commonest num- ber). The hind pair of transverse processes diverges backwards as in the C pattern but is usually not longer than the anterior one. L. fraasii. Number of vertebrae with double processes usually two to four. The posterior pair diverges backwards and may be longer or much shorter than the hind pair. It has not so far been possible to find any correlation between the caudal pattern and any functional parameter. The pattern does not appear to be related to the type of locomotion adopted or specialized use of the tail or to the ease with which autotomy takes place. Hemipenial characters The genitalia of many animals provide useful systematic characters ; for instance the baculum of mammals, the genital armature in insects, various features of the turtle penis (Zug 1966) and the hemipenis of snakes (summarized in Dowling & Savage 1960). Although not generally used, the hemipenes of lizards including lacertids can also provide helpful taxonomic information. Members of a species- group often show a similar hemipenial facies that differs from those occurring in other assemblages. The factors responsible for the evolution of different hemi- penial structure in related groups are discussed on p. 324. Differences are not connected with the pattern of copulation for this does not seem to vary much in the Lacertidae. It is unlikely that the differences often function as isolating mech- anisms since most sympatric lizard species appear to maintain isolation primarily by ethological means and species most likely to interbreed, that is ones which are closely related, tend to have a similar hemipenial structure. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 307 Technique. In the overwhelming majority of specimens examined, the hemi- penes had not been everted before preservation, so the following descriptions are based almost entirely on dissections of retracted organs. The ventral surface of the hemipenis is exposed and one lobe and the stem of the organ are opened by a parasagittal incision. This method of investigation is not entirely satisfactory as it is very difficult to make accurate comparative measurements of the uneverted hemipenis, but a good general impression of the morphology of the organ can be obtained by this means. General structure of the hemipenis. The hemipenes of all species of Lacerta, Algyroides and Psammodromus are symmetrically bifurcate with large, usually plicate lobes. The sulcus runs along the hemipenial stem to divide into two branches each of which runs on to one of the lobes. The two lips of the stem sulcus overlap across it, but apically, where each forms the outer lip of one of the lobe sulci, they are usually enlarged and nearly always reflected outwards away from the sulcus. The inner lips of the lobe sulci are typically less developed than the outer ones and are sometimes scarcely apparent especially in the everted organ. However, their degree of development shows some correlation with that of the outer lips. 1. The armature and the arrangement of the lobes in the retracted organ The hemipenes of many lacertid species are apparently different from those of most other lizards in having a clearly defined and often complex supporting struc- ture — the armature (Arnold, in press). This is formed of dense connective-tissue and typically consists of a plate-like structure embedded in the dorsal surface of the retractor magnus muscle. The plate is attached to the cartilaginous region around the sulcus by a series of short connectors and may bear two, often club- shaped bodies, the clavulae, which lie between the retracted lobes. In forms where an armature is present, the retracted lobes are not simple sacs as in most other lizards but are flattened and complexly folded. The lobes are also peculiar in that the retractor magnus muscle does not insert widely over their surface but only along a narrow tract via a tendinous connexion. Clearly armatured hemipenes with folded lobes are found in most lacertid genera but not Takydromus, Platyplacopus, Psammodromus, the European species of Algyroides and most Palaearctic Lacerta species. But L. vivipara and especially L. cyanura and L. jayakari have clearly differentiated armatures with distinct folding of the lobes. Armature development is also found in some L. (Ap.) cap- padocica, which appear to have traces of lobar folding, and to a lesser extent in L. (A.) graeca where the lobes are not folded at all. Some other species of Lacerta have varying amounts of more diffuse connective tissue in the retractor magnus muscle and around the hemipenial lobes ; however, it is not differentiated enough to constitute a clear armature although it is almost certainly homologous with this structure. The wide distribution of armatured hemipenes in the Lacertidae may indicate that it is the primitive pattern within the family and that the con- dition found in other species is a secondary simplification. The sporadic occurrence of armatured hemipenes within Lacerta tends to support this hypothesis. 308 E. N. ARNOLD 2. Lobe proportions (Fig. 5) The hemipenial lobes can each be divided into two parts: an apical section from the apex to the point where the two lobes meet, and the remaining basal section. The relative lengths of these two parts vary although it is difficult to measure them accurately in the retracted hemipenis. The apical section in the majority of species is only about equal to or distinctly shorter than the basal section. In Podarcis (excluding L. dugesii) the apical portion is nearly always longer than the basal part and may at times be twice as long. This is true also of Gallotia (although the two sections may be subequal in L. (G.) simonyz), of L. peloponnesiaca, L. (A.) danfordi and to a lesser extent of L. laevis. The lobes are long too in Psammodromus algirus and in this species they are also rather narrow. 1 2 3 Fic. 5. Variation in hemipenis structure. Diagrams represent retracted organs viewed from below and opened by a parasagittal incision along the stem and one lobe which is spread outwards. Lines at sides indicate relative lengths of apical and basal sections of the lobes. (1) Widespread pattern in Laceyta: small sulcal lips, relatively short apical section, no large papillae. (2) Typical pattern in Podarcis: large sulcal lips, long apical region, no large papillae. (3) Typical pattern in Gallotia : small lips, moderate apical region, long apical papillae. o = outer lip of lobe sulcus. i = inner lip of lobe sulcus. 3. Size of lips bordering the lobe sulci (Figs. 5, 6) The reflected outer sulcal lip varies considerably in size. In most species it is relatively small (Figs. 5(1), 6(r)) but it can be large especially in Podarcis (excluding L. dugesii). L. peloponnesiaca and some L. (A.) danfordi have sulcal lips that are as large as in Podarcis and a lesser degree of lip enlargement occurs elsewhere, for instance in L. laevis, L. (S.) perspicillata and especially L. (Z.) andreanszkyi. In all these forms the inner lip is also large. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 309 Oo Fic. 6. Variation in hemipenis structure. Diagrams represent transverse sections of single retracted lobes just apical to the hemipenial bifurcation. (1) Small-lipped type (most Lacerta spp.). (2) Large-lipped type (Podarcis, etc.). o = outer lip. i = inner lip. 4. Presence of plicae on the lobes In Algyroides, Psammodromus and nearly all species of Lacerta, the lobe surfaces have regular plicae running radially to their apicobasal axes, the only exceptions being L. fraasii and L. parva. These have the lobe surfaces irregularly folded with a series of longitudinal flaps on them. 5. Apical papillae (Fig. 5) The apical region of each lobe is usually irregularly plicate, although in some members of Podarcis there may be a short series of small, blunt tubercles. The three species of Gallotia differ from the rest of Lacerta in having a terminal lobar area of large, pointed papillae, each of which is conical and somewhat flattened. Relatively large, pointed apical papillae also occur in Psammodromus blanct, P. hispanicus and P. microdactylus. 6. Micro-ornamentation of the lobe surface Klemmer (1957) pointed out that the minute projections on the lobe plicae differed in shape between various Lacerta species and were potentially useful as taxonomic characters. Klemmer based his studies on fresh material (which is inevitably in rather limited supply), but it is possible to examine the pattern of micro-ornamentation in hemipenes extracted from alcohol-preserved specimens, even those over a century old provided they were killed during the mating season, which means that this character can be fairly readily surveyed. Variation in the hemipenial micro-ornamentation of lacertids has recently been studied in some 310 E. N. ARNOLD detail by Bohme (1971) and my own observations, using a scanning electron micro- scope, appear to confirm his. Basically there seem to be three main types of micro-ornamentation, although intermediates do occur: (x) simple spikes or recurved spines, (2) tubercles with a ring of spinules at their apex (crown-shaped tubercles), (3) irregular, quite often bifurcated tubercles. As will become apparent, closely related species tend to have a similar pattern of micro-ornamentation but there are a number of excep- tions to this generalization. Chromosomes A review of the rather limited karyological data available for the Lacertidae up to 1968 is given by Gorman (1969). Subsequently Orlova & Orlov (1969), Kupriyanova (1968, 1969) and Gorman ef al. (1971) have published information on several more species. To date all the genera and species that have been investi- gated are Palaearctic ones, viz.: Acanthodactylus, Evemias (subgenera Evemias s. str. and Mesalina), Lacerta, Ophisops, Psammodromus and Takydromus. Nearly all the species examined appear to have the same ‘nombre fondamentale ’ (i.e. number of chromosome arms, Matthey 1949) : in the diploid state this is 38. The commonest diploid formula is 36 acrocentric macrochromosomes plus two micro- chromosomes. This is found in Psammodromus algirus, P. hispanicus and the majority of Lacerta species investigated, viz. armeniaca, agilis, caucasica, chloro- gaster, dahli, derjugini, laevis, lilfordi, melisellensis, muralis, oxycephala, praticola, rostombekovt, vudis, saxicola, sicula, taurica, trilineata, unisexualis and viridis. L. lepida, L. vivipara and L. strigata deviate slightly from the standard pattern : L. lepida has 32 acrocentric macrochromosomes, two metacentric ones and two microchromosomes; L. vwivipara lacks microchromosomes while L. strigata has one of the usual pairs of acrocentric macrochromosomes replaced by a sub- metacentric pair (Orlova & Orlov 1969). L. parva, while exhibiting the standard “nombre fondamentale ’, is peculiar in having only 24 chromosomes in the diploid state: 14 metacentrics and ro acrocentrics. A more extensive survey of the lacertids might well reveal other variants. External morphology and colouring 1. Arrangement of nasal and anterior loreal scales The rostral scute and more frequently the first upper labial scute may contribute to the border of the nostril in Lacerta, Algyroides and Psammodromus, but its greater part is made up by scutes usually termed supranasal (lying dorsal, anterior and sometimes ventral to the nostril) and postnasal (lying posterior to the nostril). The scaling in this region occurs in three basic patterns: (1) Supranasal separated from anterior loreal by a single postnasal. (2) Supranasal separated from anterior loreal by two superposed postnasals. (3) Supranasal contacting anterior loreal over a single postnasal. These three patterns usually occur with a single anterior loreal but in some populations of L. (L.) agilis and in L. (Ap.) cappadocica the RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 311 scaling posterior to the nostril is very variable and there may be either one or two anterior loreals and in L. (Ap.) cappadocica up to three postnasals. In fact the number of postnasals not infrequently varies within species but sometimes it is relatively stable throughout what, on other grounds, appears to be a natural species-group, e.g. there are two postnasals in most members of the subgenus Lacerta s. str. but only one in Gallotia and in the majority of individuals of Podarcis. Because it is so easily observed, postnasal number can be a useful, although by no means infallible, keying character. 2. Lower eyelid The two monotypic subgenera of Lacerta, Apathya and Scelarcis, are partly based on possession of a distinctly transparent ‘window’ in the lower eyelid. In L. (Scelarcis) perspicillata this consists of a single large scale but in L. (Apathya) cappadocica the window is made up of from six to eight smaller elements which have conspicuous dark borders. In fact many small lacertids without eyelid windows appear to be able to see, to some extent, through the lower eyelid which is often translucent enough in living animals for the pupil and iris to be visible when the eye is closed. In a number of species there is a series of enlarged scales in the centre of the eyelid and in L. dugesii these are sometimes as large as those found in L. (Ap.) cappadocica. Indeed the principal distinctive eyelid features of this last species are the distribution of pigment (concentrated at the scale edges) and increase in transparency rather than gross structure. The selective forces promoting the development of eyelid windows and perma- nent spectacles (in which the upper and lower eyelids are fused) are probably varied. Walls (1934) thought that spectacles protect the eyes from mechanical damage, but Williams & Hecht (1955) suggest that pigmented windows may be important in cutting down the intensity of incident light and point out that many species with eyelid windows keep their eyes shut when basking. Another possi- bility is that eye closure reduces water loss from the corneal surface. This may be important in small heliothermic species with relatively large eyes which cus- tomarily bask in hot sun. The potential total water loss from the corneal surface may be quite large compared with the overall water balance of the animal. Fon instance, Reichling (1957) suggests that corneal transpiration may account for over 20 per cent of total transpirational loss in Lacerta agilis (a fairly small-eyed species) at normal activity temperatures. Keeping an adequate liquid film on the cornea might also present problems in these conditions. Certainly eyelid windows and spectacles are quite widespread in small heliothermic lizards, e.g. Cabrita, Eremias, Holaspis and Ophisops (Lacertidae), Gymnophthalmus (Teiidae), Platy- saurus (Cordylidae), many skinks and diurnal gekkonids. 3. Supratemporal scales and their relationship to the parietal scutes (Fig. 7) The shape and number of the supratemporal scales are very variable in Lacerta. In Lacerta s. str., L. princeps, L. lepida, L. simonyi there are usually two (or three) large, relatively deep scales, whereas in Podarcis the supratemporals are small, 312 E. N. ARNOLD po 2 Fic. 7. Differences in relationship between the outer margin of the parietal scale and the underlying bones of the skull. (1) Parietal scale margin running along outer edge of postorbital bone. (2) Parietal scale margin close to postorbital—postfrontal suture. po = postorbital bone. pf = postfrontal bone. shallow but again of about equal depth. In L. (G.) galloti and L. (G.) atlantica, the anterior supratemporals are much shallower than the posterior ones while in L. (S.) perspicillata and some Podarcis the supratemporals are scarcely distinct from the small temporal scaling. Most other species of Lacerta and Algyroides have a large anterior supratemporal which is both larger and deeper than those following it. In Psammodromus the supratemporals may all be about the same depth or there may be some tendency for the anterior ones to be slightly narrowed (P. algirus) or enlarged. Variation in the relationship between the anterior supratemporals and the parietal scute was used by Méhely (1909) to define his two Wall Lizard groups (equivalent to the subgenera Podarcis and Archaeolacerta). In the former the outer parietal border is usually convex whereas in Archaeolacerta it is supposedly emarginated by the first supratemporal. In fact, it is sometimes difficult to decide if the parietal is emarginated and it is often easier in preserved material to see differences in the relationship of the anterior outer border of the parietal to the underlying bones of the skull. In the great majority of Psammodromus, Podarcis, L. peloponnesiaca, L. (S.) perspicillata, L. (Z.) andreanszkyi, L. (Z.) vivipara, L. (G.) atlantica and L. (G.) galloti the anterior border of the parietal scute with the anterior temporal runs along the edge of the skull table (i.e. along the outer edge of the post- orbital bone) or very close to it. In the remainder of Lacerta (with rare exceptions) RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 313 and in Algyroides, the parietal border either runs along the suture between the postorbital and postfrontal bones (if present) or fairly close to it or alternatively may run forwards from the outer edge of the postorbital to approach the postorbital— postfrontal suture anteriorly. Méhely (1909) suggested that the peculiar arrangement of supratemporals in L. (G.) galloti and L. (G.) atlantica is a result of the original anterior supratemporals being incorporated into the parietal, the present anterior supratemporals being originally ordinary temporal scales. Some L. (G.) galloti have a cleft in the edge of the parietal in the position where the suture between the first and second supra- temporals would be expected to be. Possibly the parietal shields of at least some of the other forms with small apparent supratemporals also incorporate the original supratemporals. L. (Ap.) cappadocica, L. cyanura and L. jayakari are peculiar in having all the supratemporals resting at least partly on the skull table whereas in other species some of the more central supratemporals are off it. L. jayakari is also singular in the adults having the posterior supratemporals diagonally elongated so that the anterior part of each lies above the posterior section of the one in front of it. 4. Dorsal scaling Shape. Dorsal scales can be flat or raised, keeled or smooth, juxtaposed or imbricate. In Lacerta the dorsal scales are always smaller than the upper caudals but in Algyroides and Psammodromus they are nearly as large or considerably larger. In these two genera they are also strongly imbricate and keeled and are also often pointed. Scale shape may sometimes correlate with the type of habitat occupied by the species concerned (see p. 320). Number. The number of dorsal scales in a transverse row at mid-body is a common parameter in lacertid systematics. In Lacerta the number varies from as low as 25 in some L. (Z.) vivipara to over 110 in some L. (G.) galloti, and the ranges in Algyroides and Psammodromus are about 21 to 31 and 15 to 28 respectively. Peters (1961) used the high dorsal count of L. lepida compared with its supposed relatives in the subgenus Lacerta s. str. as evidence that it should really be placed in Gallotia (where two of the three species also have fine scaling). However, there is a trend in Lacerta for small species and ones from moist habitats to have lower counts than large forms and ones from more arid regions. As both the fine-scaled members of the subgenus Gallotia and L. lepida are large and found in quite dry habitats, their high scale numbers might be a result of convergence. 5. Preanal region Most species have a single large preanal scute bordered anteriorly by one or two semicircles of smaller scales. But in some cases the preanal is small or divided and bordered by up to four semicircles of scales. In the preanal region there is again some tendency for the larger species and those from arid regions to have higher numbers of scales. 314 E. N. ARNOLD 6. Collar Lacerta and Algyroides both have a well-defined collar. That is a backwardly directed transverse skin-fold on the lower surface of the neck, just anterior to the lateral arms of the interclavicle. The outer surface of this fold is covered by a transverse series of large plates. The posterior margins of the plates may form a regular, continuous line (collar smooth) or they may project backwards to form a free serration (collar notched). In Psammodromus a collar is either absent or at best very poorly defined. 7. Ventral scaling Shape. There is considerable variation in the shape of the ventral scales and in their degree of overlap among the species of Lacerta, Algyroides and Psammo- dromus. As with dorsal scaling there is a marked correlation with habitat type (see p. 318). Number of longitudinal rows. Most species have six longitudinal rows of ventrals but eight is not uncommon, and ten is the usual number in L. (S.) per- spicillata, L. princeps, L. (G.) atlantica, Psammodromus algirus and some popula- tions of L. lepida. L. (G.) galloti and L. (G.) stmonyi have between 10 and 20 ventral rows. 8. Keeling on subdigital lamellae The subdigital lamellae of nearly all Lacerta species are flat or tubercular, but those of the L. (Ap.) cappadocica examined have a single distinct central keel ; L. cyanura has a double row of keeled lamellae under each toe and the lamellae of L. parva may also occasionally bear two faint ridges (Lantz & Cyrén 1939). This character is sometimes stated or implied to be an adaptation to locomotion on loose surfaces, but its distribution in many families of lizards does not fit this hypothesis since many rock dwelling species also possess keeled lamellae, e.g. L. (Ap.) cap- padocica, L. cyanura, Platysaurus (Cordylidae), various members of Agama and Uromastyx (Agamidae) and many skinks (e.g. Mabuya laevis, M. quinquetaemata and M. sulcata). D. Western has made the more likely suggestion (personal com- munication) that the function of these keels is to reduce the area of contact with hot substrates and thus reduce heat flow. The toes of Psammodromus hispanicus are also distinctly bicarinate, while those of P. microdactylus and some P. algirus are more weakly keeled. g. Caudal scale whorls One of the characters used by Méhely (1909) to distinguish the subgenus Avchaeo- lacerta from Podarcis is the degree of variation in the length of the caudal scale whorls. In Archaeolacerta the whorls are alternately long and short whereas in Podarcis they are said to be subequal. In fact the two conditions are not clearly separated and most members of Algyroides, Psammodromus and Lacerta show at least some alternation in the length of successive scale whorls and proper assess- ment of this feature thus requires accurate measurement of several whorls on each RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 315 individual lizard. This makes the character very time-consuming to assess and, as there is also considerable intraspecific variation, it appears to be of rather limited value except as a subsidiary character in identification. 10. Colour and pattern Colour and pattern are very variable in Lacerta and to a much lesser extent in Algyroides and Psammodromus but two aspects in particular seem useful in defining species groups: these are the degree of sexual dimorphism in the dorsal pattern and the presence or absence of bright colour on the belly. Sexual dimorphism in dorsal pattern is well marked in most populations of Podarcis and L. peloponnesiaca and occurs also in some populations of Lacerta s. str. and Gallotia. Most other species of Lacerta either lack any sexual dimorphism in dorsal pattern or have it only slightly developed. This is also true of Psammodromus and most species of Algyroides (not A. moreoticus). Bright, often transient ventral colouring occurs in at least some of the breeding individuals of most populations of Lacerta, Algyroides and Psammodromus. In the majority of these both the throat and belly are brightly coloured and the two areas may contrast with each other. However, in the subgenus Lacerta s. str., and in L. jayakari, L. princeps and L. lepida there is no bright colouring on the belly. This is true also in many populations of L. (P.) szcula, some L. (P.) pelopon- nesiaca (according to Buchholz 1960) and some L. (P.) milensis. Other individuals of the last species have bright colouring restricted to the second longitudinal rows of ventrals from the mid-line. Breeding males of Psammodromus algirus also have a pale belly. Other aspects of coloration are discussed elsewhere. Hybridization If two species are able to produce viable (although not necessarily fertile) hybrids, then a relatively close similarity in genetic material and therefore relationship is suggested. A number of cases of hybridization have been reported within Lacerta and most of these have been summarized by Mertens (1950, 1956, 1964, 1968, 1972). The various known or assumed crossings are shown diagrammatically in Fig. 8. It will be seen that in nearly all cases hybridization has been within and not between accepted subgenera. Figure 8(a) involves only members of Lacerta s. str. ; Fig. 8(b) only members of Podarcis and Fig. 8(c) only members of Archaeolacerta plus L. derjugini which has been assigned to both Archaeolacerta and Zootoca (see p. 299). RELATIVE TAXONOMIC VALUE OF CHARACTERS Introduction The distribution of many characters customarily used for defining intrageneric groups in Lacerta (see Table I) does not always correlate with that of the ‘new’ osteological and hemipenial features introduced here, even although the latter tend to correlate well with each other. For instance, it is generally accepted on the 316 E. N. ARNOLD trilineata w a viridis ————— schreiberi agilis | c c strigata tiliguerta b) Ee : w muralis erhardii c| w* melisellensis |< ' c : sicula —_——__——_ pityusensis w wagleriana graeca erivieeey oxycephala ¢) derjugini caucasica Ww Ww dahli —"<-> saxicola unisexualis c Ww = armeniaca c rudis Fic. 8. Known or assumed hybridization between species of Lacerta. c indicates hybrid- ization reported from captive animals. w indicates hybridization assumed from morpho- logically intermediate animals caught in areas of sympatry. * indicates personal observation, For further explanation, see text. evidence of the characters usually employed that L. dugesii is a member of the subgenus Podarcis but its skeletal and hemipenial characteristics do not support such a relationship. Similarly the lizards now separated into the subgenera Archaeolacerta, Lacerta s. str. and Zootoca are all very similar in many features of their hemipenes and postcranial skeletons which may suggest a closer relation- ship than is usually acknowledged. This conflict of evidence makes it necessary to assess what relative taxonomic weight should be given to the various characters available. One factor that seems especially important in making this judgement is the comparative lability of the characters concerned. Obviously, beyond certain RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 317 limits, the more labile a feature is, the less likely it is to indicate real relation- ships. Among the factors suggesting that a feature is highly labile are the following. (xt) High intraspecific variability in a substantial number of the forms under con- sideration. (2) Continuous variation of the feature throughout the species group being examined (instead of the variation being separable into two or more disjunct character states). This applies specially to characters that involve very simple differences, for instance the relative size or proportions of a morphological feature. These may often be the result of quite small variations in growth pattern and be under polygenic control so that they may be likely to change fairly swiftly in response to selection pressures because there is considerable stored variability in the genotype and fresh mutation is not initially necessary. (3) The suspicion of lability is increased if, especially in the case of a large and varied genus like Lacerta, such contrasted characters show a close correlation with environmental para- meters. Such correlation may well indicate that a selective pressure connected with the ecological parameter is responsible for the distribution. If the selective pressure and its mode of action in functional terms can be identified, then the case for assuming lability is very strong. Some examples of this sort of character assessment in Lacerta are given below. Ecologically correlated characters Sympatric species of lizards avoid competing ecologically by utilizing different food resources. But, as the majority of small lacertids are fairly general carnivores and eat whatever small, palatable animals they can overpower, this ecological separation does not usually depend on actively selecting prey species. Where more than one lizard species coexist in precisely the same habitat, separation is usually achieved either by hunting at different periods or by selecting different sizes of prey, in which case the coexisting species often differ in size too. But the commonest and most obvious means of dividing resources depends on each species being confined to and exploiting particular areas of the general environ- ment. Spatial separation is common among sympatric species of Lacerta and their allies, the division depending on such features as humidity, degree of shade and the physical nature of the lizard’s substrate. For instance in some upland areas of southwestern Yugoslavia (at about 1000 m), seven species of Lacerta may occur within a few hundred metres of each other. They divide the environment as follows. Species Size Typical habitat L. (A.) oxycephala Small Scansorial on sunny rock outcrops, pavements, boulder-screes, etc. L. (A.) mosorensis Small Scansorial on moister and more shady rock areas than oxycephala L. (P.) muralis Small Less scansorial than oxycephala and mosorensis but climbs frequently on the base of rock outcrops, on steep earth banks and on vegetated field walls and screes 318 E. N. ARNOLD Species Size Typical habitat L. (P.) melisellensis Small Largely ground-dwelling in dry places, especially on broken terrain often with some vege- tation. Rarely climbs on rocks, etc. but may occasionally climb small bushes L. (L.) viridis Large In and around bushes especially brambles L. (L.) agilis Medium Ground-dwelling in dry pastures often with small bushes L. (Z.) vivipara Small Ground-dwelling in moist well-vegetated places often near water It is with such differences in spatial niche that many of the characters used in defining Algyroides, Psammodromus and the subgenera of Lacerta can be correlated. Furthermore, they can often be interpreted as functional adaptations to survival in these niches. Ventral scaling (Fig. 9) Variation in ventral scaling provides an example of this type of correlation. There are two extreme conditions found in Lacerta which are illustrated in Fig. 9 : (1) belly scales shaped like inclined parallelograms with the posterior and lateral edges strongly overlapping ; (2) belly scales almost rectangular with little or no overlap. Many intermediates between these two conditions also occur. The most complete development of the first condition is found in the large lizards which make up the subgenus Lacerta s. str., and in L. princeps and L. lepida. It exists in a less extreme form in the following smaller species: L. (Z.) vivipara, L. prati- cola, L. derjugini, L. parva, L. fraasti and some species of the subgenus Podarcis also tend towards this condition (e.g. L. taurica). All these forms are largely ground dwelling in at least partly vegetated areas. If they climb, it is usually in scrub and bushes rather than on open surfaces. An even greater degree of ventral scale imbrication is found in Psammodromus algirus which often occurs in and around dense spiny vegetation. The second condition is best developed in the subgenus Archaeolacerta, L. (S.) perspicillata, L. (Ap.) cappadocica and in some Podarcis (populations of L. hispanica and L. muralis). In contrast to species having the first condition, these are largely scansorial and most of them climb principally on rock faces, boulders or walls but rarely in bushes. The two conditions can be ‘ explained’ in functional terms. A ground-living Lacerta carries its viscera close to the substrate from which they are separated only by a thin body-wall. When travelling at speed, there is a considerable danger that this wall will be pierced by sharp twigs, dry grass stems or thorns pointing backwards against the direction of motion. Similar risks are faced by lizards in dense bushes. The ventral plates provide protection from such damage, having a hard, mainly f-keratin epidermis (Maderson 1964) and a tough dermal layer. However, these plates are quite rigid and, to permit the sinusoidal movements occurring in locomotion, there are transverse ‘hinge’ regions housing flexible areas that allow expansion or contraction. Flexible (but non-expanding) hinges RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 319 a 3 4 Fic. 9. Ventral scaling. (1) Strongly imbricate, most scales shaped roughly inclined like parallelograms. (2) Little imbrication, individual scales almost rectangular. (3) and (4) see text. run in longitudinal series between the plates and facilitate the movements con- nected with respiration, lateral and dorsal compression of the body, etc. To provide the necessary flexibility these hinge regions have a thinner, less rigid epidermis largely made up of «-keratin (Maderson 1964) and a thinner dermal layer ; they consequently remain likely sites of penetration. (This can easily be confirmed on preserved material with a dissecting needle.) However, the trans- verse hinge regions are protected by the posterior overlap of each scale and are only exposed if the body is sharply flexed to one side. The roughly longitudinal hinge regions are also protected by imbrication but the protection is increased by the lateral scale margins being oblique. If these margins were parallel to the mid- line of the belly it would be possible for even slightly obliquely directed projections to penetrate under the overlap as in Fig. 9(3). The risk of damage is greatest with projections having axes approaching although not lying on a line parallel to the mid-line (Fig. 9(3a, 4a)), for at more oblique angles (Fig. 9(4b)) it is more likely that the projections will be deflected rather than penetrating. Therefore, by having the scale margins oblique, the risk of a projection entering a hinge is limited to the 320 E. N. ARNOLD potentially less damaging ones. All other things being equal, risk of damage by rigid projections increases with increase in size (since the mass and momentum of a lizard tends to increase with the cube of its linear dimensions, including skin thickness). It is therefore not unexpected that the system of overlaps should be better developed in the larger species. Lizards habitually climbing on rock-faces are much less likely to be damaged by sharp projections and so can usually ‘afford’ to do without these elaborate overlaps and their ventral scales themselves are not so resistant to penetration. (Rock-dwelling species of Lacerta if accidentally driven into thorn bushes during collection will sometimes get transfixed.) Collar (Fig. 10) When present, the collar provides an area of extension allowing movement of the gular skin during head raising. Thus when the head is thrown upwards, the collar fold is pulled out and the soft skin protected by the collar plates is extended (see Fig. 10(A), 10(B)). The various modifications of the collar that occur in Lacerta and related genera can, like belly scales, be interpreted as adaptations to particular structural environments. Collars with small, even-edged, only slightly imbricate plates that expose the vulnerable soft skin widely are, as might be expected, con- fined to habitats with few projections. In more spiky biotopes, there is a tendency for the collar plates to extend backwards to form a denticulate frill giving better protection to the soft skin (e.g. in Lacerta s. str.). There is also a trend in such habitats for the sites of expansion not to be confined to a single vulnerable collar area but to be dispersed between a number of transverse rows of gular scales, which in the species concerned are strongly imbricate and thus cover the areas of expan- sion even when the skin is extended. In the most extreme cases, the collar dis- appears altogether (e.g. Psammodromus algirus), see Fig. 10(C), t0(D). Dorsal scaling Dorsal scales show trends in relation to habitat structure that are similar to those found in ventral scaling. Thus species in open environments tend to have small, rather convex scales with little overlap while forms living in biotopes with many projections tend to have larger, more strongly built scales with greater imbrication so that the more vulnerable interstitial skin is better protected. This process reaches its greatest development in Psammodromus. Characters distinguishing the subgenus Archaeolacerta Many of the features distinguishing the members of the subgenus Archaeolacerta (and Scelarcis and Apathya as well) from the rest of Lacerta can be interpreted as functional adaptations to the kind of spatial niche that they inhabit. As stated on p. 318, all the species assigned to Avchaeolacerta, L. (S.) perspicillata and L. (Ap.) cappadocica are essentially rock-dwelling. (Evidence: L. graeca, horvathi, monticola, mosorensis, oxycephala — personal observation; L. bedriagae — Lambert 1967 ; L. danfordi — Wettstein 1967, M. R. K. Lambert, personal communication ; RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 321 EEE Fic. 10. Different ways of providing elasticity in the gular skin. A and B: restricted area of extensible skin protected by a collar (e.g. in rock-dwelling species of Lacerta) ; A-—skin at rest; B-—skin stretched. C and D : no collar, extensible skin distributed between a number of overlapping transverse gular scale rows ; C —skin at rest ; D —skin stretched. c = collar. e = regions of extensible skin, 24 322 E. N. ARNOLD L. saxicola and other Caucasian forms—Darevsky 1966; L. perspicillata —- Dou- mergue 1901; L. cappadocica— Reed & Marx 1959, Bird 1936, I. Nader, per- sonal communication.) A more detailed account of adaptations to rock-living in Lacerta will be published elsewhere ; the following remarks are abstracted from this. Typical Avchaeolacerta characters are printed in italics. As with many lizards living in open environments, saxicolous lacertids appear to be subjected to quite severe predator pressure, both from snakes and from birds (shrikes, hawks, etc.). The principal means of avoiding such predators is for the lizards to retreat deep into the narrow frost or sun-induced crevices characteristic of many rock formations and to attach themselves firmly with their claws and by flexing their bodies against the crevice surfaces. Birds cannot follow lizards into such crevices and snakes have difficulty in obtaining enough purchase to extract them. Many Archaeolacerta features are connected with this behaviour. The depressed body and low skull make entrance into very narrow crevices possible. The weak ossification of the skull with its thin flat roof, poorly developed osteoderms and large nasal openings gives it a measure of flexibility so that the head can be distorted and wedged into quite irregular fissures. The flat, non-imbricate dorsal scales, the narrow collar and non-overlapping ventral scutes allow these lizards to move as easily backwards as forwards in restricted spaces. This is advantageous since backing out of refuges is often necessary. Depression of the head has secondary effects, one of which involves the eyes. An animal that depends on sight to detect prey and predators cannot ‘ afford ’ to reduce the absolute size of its eyes, consequently in lizards with depressed heads these project well above the level of the skull roof. But when such a lizard enters a crevice the eyes must be accommodated within the reduced vertical dimensions of the head. This is accomplished by some of the orbital contents passing down- wards through the suborbital foramen which is enlarged, so that they project into the buccal cavity. These movements require a much greater flexibility of the supra- orbital region and this is provided by the limited ossification of the supraocular plates. The characteristic features of rock-living Lacerta species often allow these lizards to avoid predators but paradoxically make their plight worse if they cannot reach a suitable crevice. The fragile skull is easily smashed and the absence of scale overlaps and mechanically strong, raised scales means that the skin is more liable to damage. Furthermore, the lowering of the skull results in the efficiency and size of the jaw muscles being reduced so that specialized rock lizards are less able to actively defend themselves than other species of similar size. There is conse- quently great advantage in directing attack away from the body towards the more expendable tail. There are often behavioural mechanisms to do this, for instance some rock lizards switch the tail vigorously from side to side when attacked. Also while the body is usually more or less procryptically marked, the tail in many young vock lizards is blue or green and conspicuous at close quarters. In the most special- ized rock lacertas, e.g. L. oxycephala, L. perspicillata, L. cappadocica, L. cyanura, this colouring may be retained by adults too. Furthermore, the tails of rock lizards appear to be more easily autotomized than those of other forms (well over go per cent of adults in some populations of L. oxycephala have regenerated tails). RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 323 Evolution of low autotomy thresholds is probably facilitated by the tail being of less locomotory importance in scansorial lizards than in ground dwelling ones. Y. Werner (1968) has demonstrated a similar correlation between habitat and rate of autotomy in geckoes. Many of the characters typical of rock-dwelling lacertas also occur in saxicolous species of other families, e.g. Afroedura, Ptyodactylus and Quedenfeldtia (Gek- konidae) ; Oplurus, Sauromalus, Petrosaurus (Iguanidae) ; some Xantusia (Xan- tusiidae) ; Mabuya sulcata, M. laevis (Scincidae) ; Platysaurus (Cordylidae). The correlation between the development of this syndrome of characters and rock-dwelling is often quite precise. Thus L. monticola cyreni appears to be less strictly restricted to rocks and crevices than the three Yugoslav species of Archaeo- lacerta (L. horvatht, L. mosorensis and L. oxycephala) and has the characteristic saxicole features less well developed than in these species. Similarly although most species of Podarcis climb to some extent, only some populations of L. hispanica seem to be almost exclusively rock-dwelling. These are the only members of the subgenus to show very distinct development of features supposedly characteristic of Archaeolacerta, etc. Discussion Thus the distribution of many of the characters traditionally used in the sys- tematics of Lacerta and allied genera can at least be tentatively related to ecological parameters and often to particular strong selective pressures. This suggests that they are rather labile, a conclusion supported by many of them being continuously variable throughout the group, in some cases by their showing some intraspecific variability and by the fact that differences between contrasted characters are often the result of quite small changes in growth pattern. For example, the supra- ocular pattern shows two characteristic adult states, viz. ‘ supraocular bones com- plete ’’ and ‘supraocular bones fenestrated’ which both develop in ontogeny from an unossified juvenile condition. In most species the osteoderms develop fast, spreading outwards from the border of the frontal until the whole supraocular area is covered, but in Archaeolacerta and other rock species the process is much slower so that by maturity there is still an unossified (i.e. fenestrated) area. That this state is merely the result of a comparative retardation in growth rate is indicated by the occurrence of complete osteoderms in old specimens of some Archaeolacerta species (Klemmer 1957). The apparent lability of the characters discussed above suggests that they should be given relatively low weight in assessing relationships although it could of course be argued that the other characters available for classifying Lacerta and its allies are just as likely to be labile and that this is not recognized because the selective factors responsible for their lability are unknown. This is true, but provided the possibility of high lability has been carefully considered for all features, it seems better in cases where two sets of characters suggest different classifications, to base systematic decisions on the set not known to be potentially or actually very labile. Of course, low lability does not imply that characters are non-adaptive. There are 324 E. N. ARNOLD many situations where a character may be fairly stable, yet of great functional significance. In some cases, the contrasted characters, or systems to which they contribute, may represent different solutions to a particular problem, which are superior, in most circumstances, to intermediate conditions. They may be integral parts of complex functional systems that are unlikely to be interconvertible without dis- solution of the system concerned. Thus rock lizards have different means of pro- tecting the eyes when they enter crevices. One of these, where part of the orbital contents passes through a large suborbital foramen has already been described (p. 322). Cordylids (Platysaurus) use another: the suborbital foramen being small, part of the orbital contents is passed into the interpterygoid vacuity. Pre- sumably once a rock lizard is committed to one of these methods it could not change to the other without a period of greatly reduced efficiency. Contrasted characters may represent solutions to a particular problem unconnected with the specific requirements of different niche-types. For example, Cryptodire and Pleurodire chelonians both protect the head by withdrawing it into the shell but they employ quite different means of folding the neck. Again it seems likely that the two methods are not interconvertible and the anatomical modifications connected with these methods thus make very good characters for defining the two groups. The various types of adhesive pad found on the toes of geckoes are a similar case (A. P. Russell, personal communication). While it is not yet possible to explain the stability of the lacertid characters discussed here in this way, analogous reasons for low variability seem possible. Among the features of Lacerta and its allies that seem to be generally quite stable intraspecifically and for which there are no obvious external grounds for regarding as particularly labile are the following (although in the present context, stable and labile are purely relative terms since, among the species under con- sideration, most characters are subject to occasional exception): postorbital— postfrontal fusion, sexually correlated variation in vertebral number, shape of the sternal foramen, caudal vertebral pattern, position of the lateral border of the parietal shield and various aspects of hemipenial structure. There are additional reasons for considering the latter to often be good indicators of relationship and these are discussed below. Hemipenial characters As already stated, genital characters have often been found pragmatically to be sensitive indicators of relationship (although like any other character source they are not infallible). In the lacertids, the differences between hemipenial types do not appear to be related to differing methods of copulation, or often to be isolating mechanisms. Nor is it possible at present to relate them to particular environ- mental selection pressures, either acting directly on the hemipenis or indirectly by producing modification of some other part of the animal which in turn causes selection to alter the form of the hemipenis itself. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 325 Mayr (1970) suggests that genital differences might often be pleiotropic by- products of changes in the genotype induced by selection acting on other parts of the animal. This may well be so but does not entirely explain why genitals are often such particularly good indicators of relationship. The reason may be as follows. With most organ systems, any pleiotropically induced change is likely to result in a lowering of functional efficiency. However, provided this reduction in efficiency does not outweigh the original selective advantage of the genetic change concerned, the pleiotropic alteration will initially become fixed. But there will then be normalizing selection modifying the genotype further so that the pleiotropic changes will tend to be suppressed and the organ will return towards its original state and level of efficiency ; see Fig. 11(a). One would expect rather similar events to take place when pleiotropic changes occur in the genitals but the process will not be exactly the same, since the efficiency of these organs cannot be considered in isolation but only in relation to their co- ordinated functioning with the genitals of the opposite sex which are also under selective control. (Presumably copulatory efficiency depends largely on a good physical match of the genitals rather than their absolute size and shape.) Consequently, if there is a pleiotropic change in the male organs which reduces copulatory efficiency in some way, there will not only be normalizing selection acting on the male genotype to bring the male organ back towards its original con- dition, but also, simultaneously, selection acting on the female genotype to produce changes in the genitals compensating for, and adapting to, the pleiotropic alter- ations that have already taken place in the male system. This means that the two sets of genitals may return to their previous efficiency without reverting to their original morphological state; see Fig. 11(b). Presumably the converse situation also exists where pleiotropic changes in the female system produce com- pensatory alteration in the male. The genitals are thus likely to ‘ store’ pleiotropic changes which in other organ systems would probably be eradicated. Furthermore, the genitals of each sex will tend to be altered not only by direct pleiotropic effects on themselves but also by those causing changes in the genitals of the opposite sex. It would consequently be expected that differences in hemipenial structure would often bear some relation to the overall genetic differences between the species compared. Genitals, at least in lizards, are often good indicators of relationship partly because they alter in the way described above but also because their rate of change is not too fast, so that closely related forms tend to retain quite similar structural pat- terns. One possible reason for this is that the original extent of pleiotropic changes is reduced by compensating selection in the opposite sex. Another possible explanation of hemipenial variability is that the female genital system in lizards is much more flexible and much less precisely structured than the hemipenis and could perhaps therefore often accommodate itself immediately, without genetic change, to slight pleiotropic alterations in the hemipenis. This would mean that such minor pleiotropic changes in the male organ would not be subject to normalizing selection or the rather more complex process described above. But if this was generally so, one would expect greater intraspecific variation than 326 E. N. ARNOLD Normalising selection suppressing pleiotropic effects Return to + original condition and efficiency a) Degree of change from original condition of organ a ie) Pleiotropic change reducing efficiency Original condition of organ Time ——————_> Pleiotropic change in male : Z Normalising selection reducing copulatory efficiency suppressing pleiotropic effects b) Degree of change from original condition of organ 2 Restoration of full efficiency but with changed structure Compensating selection acting on female system Original condition of genitals Tim Fic. 11. For explanation see text. has so far been encountered in lizard hemipenes. However, this factor may have some effect. Although there are reasons for believing that hemipenial characters may fre- quently be particularly good indicators of relationship, they are not infallible pointers and must be treated with as much caution as any other type of character. For it is quite probable that they are altered by direct environmental selective RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 327 pressure on occasion. Also, as the overall form of the hemipenis is relatively simple, one must sometimes expect parallel development of similar structure in forms which are not particularly closely related, especially at the lower taxonomic levels. Béhme (1971) has placed considerable emphasis on hemipenial micro-ornamen- tation as an indicator of relationship (see p. 299). However, when all available characters including other hemipenial features are taken into consideration, it appears that although hemipenial micro-ornamentation is usually similar in closely related species there are exceptions and the character consequently does not always indicate natural groupings. RECLASSIFICATION OF LACERTA, ALGYROIDES AND PSAMMODROMUS If the species of Lacerta, Algyroides and Psammodromus are arranged on the basis of those features listed on p. 324 as seeming relatively stable, it is found that there are a number of clearly defined species-groups, the members of each having more or less the same combination of these characters. Furthermore, some other features of less certain relative stability correlate quite well with this arrangement thus providing additional evidence that the groups are natural ones, e.g. number of postnasal scales, degree of sexual dimorphism, scale size, etc. The principal characters of the species-groups are listed in Table IT. Algyroides (excluding the African species) and Psammodvomus appear to be natural assemblages while Palaearctic Lacerta can be divided into four main groups. The principal features of these taxa are given in Table II, p. 328. The Lacerta groupings are as follows. 1. Equivalent to the subgenus Lacerta s. str. including L. princeps and L. lepida. 2. Including the subgenera Zootoca, Archaeolacerta, Scelarcis and Apathya plus all species not included in other groupings. 3. Members of the subgenus Podarcis excepting L. dugesii which is placed in group 2 ; and L. peloponnesiaca. 4. Equivalent to the subgenus Gallotia. As defined here, groups 1, 3 and 4 are relatively homogeneous while group 2 is much more varied and is in fact a polythetic assemblage. Groups 3 and 4 are quite distinct, both from each other and from groups 1 and 2 which in contrast are generally quite similar. In the interests of nomenclatorial stability it would be desirable, if possible, to retain Lacerta more or less as it is presently understood and recognize the groupings within it at an intrageneric level. However, this cannot easily be done for two of the groups have greater affinities with other genera than they do with the rest of Lacerta. Thus group 4 seems to be more nearly related to Psammodromus than to the other Lacerta groupings. Similarly there is a much closer relationship between group 2 and Algyroides than between groups 2 and 4 or groups 3 and 4. Lacerta, as it stands, is therefore a rather artificial assemblage and it seems best to divide it to produce more homogeneous entities. The most con- venient way of doing this is to leave the very similar groups 1 and 2 in Lacerta and raise the other two groupings to the rank of full genera, group 3 becoming Podarcis ARNOLD E. N. 328 (uondeaoxe jenied) har sapor1eqn} padeys -UMOIO JO UOI}e}UIUTeUIO-OINIPY - sdy oqoy uo seqtded [estos ‘SuoT - ase] torus aqoy jo sdry suo - saqoy] [etuediumoey yo uorsod [eody saanqwaf poruadrma zy q uta}yed + Io y uso}ed aviqezI0A [epned podeys-zire9y + yOU ‘TRAO o][AUe}UOF [eUIE}S + saroads 4soutr Io [[e JO sTenprarputr AWIOS UL Po}eUISIeWA VPIIARIO + yuasaid ‘ou [eiqez10A [eroesaid UI UOILLILA Pd}e[AIIO9-xaS sSuno< ut seo] 32 oyeIedas =/+ = =. =| 57 =//SF —/+ + = = = (suondaoxe) —/+ — + = = o}eIPAUIIE}yUL IO (suondaoxe) —/+ + + - - (suor;daoxa) - - - 7 a (saseo ysour ut) (suortydaoxe yenzed) ~ + - + ~ = = ar + at - - + 7 a (uordaoxe) (uondaoxe) - - + ~ + ae ei" (Sie Su) gz =u L=u Il #ed I Hed SNULOAPOUUDS | yoy SIIADPOT vjAIIDT DIAIIDT & dnory € dnorg z dno15 1 dnory SNULOAPOUUMDS J DIAIIDT + sauoq [e}U013s0d pue [e}IGQ10}s0q 430103380 F=4u sotoads Jo Joquinny sapioAna) uoryeoyissepa pasodorg saproans) SNUOAPOMUMDST PU ‘YjAIIVT UI papupout AT[euIsII0 exe pasodord anoj ‘sapro«ts) 7 JO somnyeoy jediourrq : [] ATAVL 329 RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS = Et: = (uo1daoxe) (suorydaoxa) + + + (suordaoxa may) = + ap or ‘9 oz-o1 (g) (8) 9 (eam 30) — + fh (soroads auo ur adr] Ayres) + — — + —/+ + (uordaoxa) (uo1rdaoxa) + (suor,daoxa) OI-9 + (uondaoxa) + — (jU9A 0} ynoUS WIUI OOI I9AO Ud}JO) A81R] OzIs y[Npy sayTeur = + SuIpee1q Ur pernojoo APY S11q ATED (suorydaoxa) (uo1ydaoxa) uio}}ed ar - [esiop ur wsrydsoump tenxes oI-9 9 SMOI OTROS [eI}UDA [eUIPNzsuoT Jo 1oquinyy + + yuasaid reyjop (sarsads auo ur a81ry Aparey) (suor,daoxa) - + ase] ATAAT}LIAI sareos [esIoq A|ioli1e}ue aU [e}1qG10}sod jo - - espe Ja}No sayoeor prarys [ejorreg (suotdaoxa) + + sjeseujsod om} ATTensq SaAnqwaf [DUAaIXTT 330 E. N. ARNOLD and group 4 Gallotia. This would produce genera with roughly the same degree of difference that is found in other parts of the Lacertidae. Such a course will result in new name combinations for the sixteen members of groups 3 and 4. This is unfortunate but probably will not cause much confusion as the new generic names have already been widely used as subgenera. An alternative course would have been to include both European Algyroides and Psammodromus in Lacerta, but this would also necessitate some name changes and would result in a large and very varied genus difficult to diagnose adequately. LACERTA Linnaeus Linnaeus, 1758, Syst. Nat., ed. 10, vol. 1: 200. TYPE SPECIES: Lacerta agilis Linnaeus. DraGNosis. Small to very large lacertids (from 35 mm to over 210 mm snout to vent). Skull shape variable, parietal foramen present, frontal bones paired throughout life, postorbital and postfrontal bones separate in young of most species but sometimes fuse during life; clavicle strongly expanded medially, foramen either emarginated or not (the two conditions often existing within the same population) ; interclavicle cruciform, the lateral arms not strongly directed for- wards ; sternal fontanelle almost invariably present, nearly always roughly oval ; sexual variation in presacral vertebral number present; all presacral vertebrae with ribs, except the first three cervicals; free ribs divided into two series — an anterior one of long ribs and a posterior one of short ribs ; caudal vertebrae typically of A or B pattern (rarely C). Hemipenes symmetrically bilobed. Typically armature absent and lobes not complexly folded (there are exceptions) ; lobes usually plicate, micro-ornamentation variable ; apical regions of the lobes usually short (exceptions) with small sulcal lips (exceptions) ; no large conical papillae at lobe tips. Head shields normal; nostril usually in contact or close to the first upper labial scute, bordered posteriorly by one, two or rarely three postnasals ; lower eyelid usually scaly although a small transparent window may be present ; anteriorly parietals typically do not extend to outer margin of postorbital bone, first supratemporal scale often large ; masseteric shield often present. Dorsal body-scales small or moderate (smaller than proximal caudals) ; collar well marked ; ventral scales smooth, truncate, strongly overlapping or not, in six to ten longitudinal rows. Toes cylindrical or compressed, usually tubercular beneath (occasionally strongly keeled) ; femoral pores present. Tail long, un- modified. Sexual dimorphism in dorsal pattern absent in most although not all species. Belly often, but not always, brightly coloured in at least the breeding males. SPECIES REFERRED. agilis, andreanszkyi, armeniaca, bedriagae, brandtiw, cap- padocica, caucasica, chlorogaster, cyanura, dahli, danfordi, derjugini, dugesit, fraasii, graeca, horvathi, jayakari, laevis, lepida, monticola, mosorensis, oxycephala, parva, perspicillata, praticola, princeps, vostombekovi, rudis, saxicola, schreibert, strigata, trilineata, unisexualis, viridis, vivipara. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 331 Lacerta as defined here is an extremely varied genus within which only one species group can be clearly separated ; this is Lacerta part I which is more or less equivalent to Lacerta s. str. and will be dealt with separately ; the remainder of the genus forms Lacerta part II. I do not propose to use formal subgenera for these two groupings, especially as the type species of Zootoca (available for Lacerta part II) is rather anomalous. LACERTA part I (= Lacerta s. str.) SPECIES REFERRED. agilis, lepida, princeps, schreiberi, strigata, trilineata, viridis. DISTINGUISHING FEATURES. A closely related group of species distinguished from the rest of Lacerta by possession of the following combination of characters. Body-size medium to very large (adults 70 mm to over 210 mm snout to vent) ; skull undepressed, very robust, often with ossified temporal skin. Usually two postnasal scales, dorsals strongly raised, often keeled; collar strongly serrated, ventrals with very marked imbrication, in six to ten longitudinal rows; sexual dichromatism in dorsal pattern quite frequent ; in young, dorsum often ocellated or with pale, narrow stripes, dorsal ground colour of adults often bright green, belly white in young, yellowish in adults, never brightly coloured. SKELETAL FEATURES. Skull robust; undepressed with a thick osteodermal layer. Supraocular lamellae complete in non-juvenile specimens, temporal ossifi- cation often extensive (not agilis or some populations of trilineata). Postorbital and postfrontal bones usually separate in hatchlings (not schreiberi) but some- times fused in adults (always in lepida) ; pterygoid teeth nearly always present. Sex-correlated variation in the number of presacral vertebrae present ; com- monest numbers 27 in males and 28 in females (often 28 and 2g respectively in agilis). Clavicle expanded medially, exists in both intact and emarginated con- ditions in all species. Interclavicle cruciform, lateral arms not obviously directed forwards or backwards. Sternal fontanelle oval, occasionally imperforate. Usually one, two or even three pairs of inscriptional ribs. Typically five or six non- autotomic tail vertebrae (rarely seven, four in some agilis). Caudal pattern most frequently B, less often A (some agilis, schreiberi and viridis). HEMIPENIS. Symmetrically bilobed without an obvious armature ; lobes not folded in retracted organ, plicate. Apical sections of lobes subequal to or shorter than basal sections; lobe tips without prominent papillae. Lips of lobe sulci small to moderate. Micro-ornamentation variable ; L. lepida has simple recurved spines but in most species the lobe apices have short tubercles with a fine denticu- lation at their tips, while the lobe flanks are covered by long, fairly straight projec- tions that sometimes end in a point and sometimes in an irregular series of spinules. In L. princeps these flank spines tend to be recurved. EXTERNAL FEATURES. Medium to large lizards (adults from about 70 mm to 210mm snout to vent). Nostril in contact with first upper labial and usually bordered posteriorly by two postnasals (one in many agilis). Eyelid scaly. 332 E. N. ARNOLD Parietal shield not clearly emarginated, but its outer margin generally running medial to the postorbital—postfrontal suture if present. In most cases, two large, deep supratemporals (posterior one sometimes broken into two or three parts). Masseteric shield frequent. Dorsal scales small to moderately large, 32 to 98 across the mid-body, raised and usually strongly keeled (keeling weak or absent in lepida). Collar well developed, strongly serrated. Ventrals smooth, truncate, strongly imbricating, in six to ten longitudinal rows, their borders forming laterally inclined parallelograms. Preanal scale large in most cases (somewhat reduced in L. lepida and to a lesser extent in L. schreibert and some L. trilineata). Toes tubercular beneath, not strongly compressed. Femoral pores usually present, each series extending to knee (except in some L. trilineata), 11 to 22 pores in each series. Tail long, unmodified. CoLoRATION. Sexual dichromatism occurs in a number of populations and is particularly well developed in L. schreiberi and many L. agilis. In juveniles the dorsum often has up to seven light longitudinal stripes although these may be absent (especially some lepzda, trilineata and viridis which are uniform) or replaced by ocelli (many agilis ; lepida and schreiberi). In adults the stripes, if originally present, are frequently lost and the dorsum becomes bright green (not in princeps, or some agilis, lepida, viridis, schreibert and strigata). Dorsum may sometimes be heavily marked with black (especially some agilis and female schreibert) but lizards of this group lack the dark lateral bands often found in Lacerta part II, Podarcis, Gallotia, Algyroides and Psammodromus. The throat is often brightly coloured in breeding males (not L. /epida and some populations of agilis, strigata, trilineata and viridis), but this does not apply to the belly which never contrasts strongly with the dorsum and which also always lacks blue spots on the outer ventral scales. RELATIONSHIPS. The seven species that make up Lacerta part I form a tight, closely related assemblage, the members of which have a detailed morphological similarity to each other. However Peters (1961, 1962a) has suggested that L. princeps and L. lepida have no near relationship to the other species and indeed this author placed L. lepida in Gallotia. Among the grounds given for excluding princeps from Lacerta part I are (i) its large dorsal scales, (ii) lack of green pig- ment and (iii) the presence of ‘ real’ (i.e. black-bordered) blue lateral ocelli. Eiselt - (1968) regards these differences as unimportant and certainly they seem rather trivial compared with the overall resemblance of princeps to other Lacerta part I lizards. Although the dorsal scales are large they are approached in size by those of some populations of L. strigata, a geographical neighbour of princeps ; also the differences in scale size within Lacerta part I (as defined here) are not particularly large, being considerably less than those found in Gallotia. Absence of green dorsal coloration is not confined to princeps, it occurs as well in some members of other species (see above). Similarly the presence of black-bordered, lateral ocelli is not really distinctive, indeed they do not occur in all princeps being typical only of the nominate subspecies ; L. p. kurdistanica Suchov 1936 has simple, unbordered blue spots on the flanks similar to those occurring in some L. ¢rilineata. The blue ocelli of adult L. p. princeps appear to develop from paler ocelli in juveniles (see RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 333 photographs in Eiselt 1969) which are very similar to the ocelli found in juveniles of L. schreibert. The case for transferring L. lepida to Gallotia must also be rejected. Peters (1961) again cites the presence of blue ocelli as grounds for exclusion from Lacerta s. str. (= Lacerta part I), but as in L. princeps, those of L. lepida develop from white ocelli which are also found in juveniles of some other Lacerta part I species. Furthermore, some L. lepida have no lateral ocelli at all. Admittedly members of Gallotia have bright (blue or yellow) flank spots but these are not clearly dark-edged. Bright flank-markings seem often to be sign stimuli and are likely to develop in any forms that use a lateral display in court- ship or in aggressive activities; on their own therefore they do not necessarily indicate relationship. Other resemblances between L. Jepida and the large members of Gallotia pointed out by Peters are: small anal plate surrounded by many small scutes ; high counts for dorsal scales across mid-body, longitudinal rows of ventrals, gular scales between chin-symphysis and collar, caudal scales in a basal whorl and femoral pores. All these characteristics indicate a general reduction in relative scale-size. It has already been pointed out (p. 313) that this is often correlated with large body-size and dry habitats, conditions that apply both to L. lepida and to Gallotia so again this resemblance may be convergent. More important L. lepida agrees with the majority of species in Lacerta part I in a long series of charac- ters not found in Gallotia, viz. postorbital and postfrontal bones separate in juveniles, sexual variation in presacral vertebral number (usually 27 in males and 28 in females as against 26 (rarely 25) in both sexes in Gallotia), clavicle sometimes emarginate, AB type caudal vertebrae, hemipenis without large apical papillae, two postnasals present, green dorsal colouring and no bright belly pigment. It seems certain therefore that the affinities of L. lepida do lie with Lacerta part I. Within this grouping, L. lepida, particularly the northwest African L. 1. pater, has a considerable resemblance to L. trilineata and especially L. schreiberi. L. 1. pater is similar to the latter in having eight rows of ventral shields, a small pre- anal, spotted juvenile livery and early fusion of the postorbital and postfrontal bones (before hatching in schreiberi, soon afterwards in lepida pater). The head shape and male colouring are also very close. Contrary to the opinions of Boulenger (1916) Lacerta part I is not closely related to Gallotia nor does it have any clear relationship with Podarcis. Its members are, however, very similar in skeletal and hemipenial features to some species of Lacerta part II. DISTRIBUTION (see Fig. 12). Extensive and continuous, covering much of the total range of the genus Lacerta (except some parts inhabited only by L. vivipara, some islands etc.). N.W. Arrica: Morocco, N. Algeria, N.W. Tunisia (perhaps Rio de Oro — Valverde 1957). MainLAND Europe: Southern peninsulas north to England, Denmark, Sweden, south Finland and about latitude 60° in Russia. MEDITERRANEAN IsLANDS: Sicily, Elba, islands on eastern Adriatic coast, Ionian Islands, Corfu, Crete, Cyclades, Euboa, Rhodes, Lesbos and several other Aegean islands. WESTERN ASIA: Central Asia to Turkey, east Mediterranean coast south to Israel, N. Iraq, N. and western Iran south to Shiraz and Neyriz. 334 E. N. ARNOLD BioLtocy. Lacerta part I species are usually among the largest lizards occurring in their ranges. They tend to eat larger prey than other sympatric lacertid species and thus avoid competition with them. Most forms are typically associated with areas having dense bushy vegetation (or occasionally with rough grass-land). L. agilis is mainly ground-dwelling but the other species may climb in bushes or even trees ; they do not climb on rock surfaces to any great extent. LACERTA part II SPECIES REFERRED. Lacerta part II contains members of Lacerta that do not possess the combination of characters distinguishing Lacerta part I. DISTINGUISHING FEATURES. Most species are under g0 mm snout to vent when adult and have brightly coloured bellies in at least the breeding males. L. jayakart, however, is much larger (up to 165 mm) and appears to lack a bright belly at all times ; in spite of this its affinities appear to lie with other members of Lacerta part II. SKELETAL FEATURES. Skull variable, depressed and delicate in some species, very robust in others, many forms intermediate between these two extremes. Supraocular lamellae often fenestrated and external nares rather large in adults of about half the species ; temporal ossification usually absent. Postorbital and postfrontal bones usually unfused (fused in L. vivipara) ; pterygoid teeth usually absent (not L. brandtii, L. dugesit, L. jayakari and L. laevis). Presacral vertebral number shows sex-correlated variation: usually between 25 and 28 in males, 26 and 29 in females. Clavicle expanded medially and existing in both intact and emarginated conditions in most forms (perhaps only the intact condition present in L. parva and L. fraasi). Interclavicle cruciform with lateral arms usually not strongly directed forwards or backwards (angled distinctly backwards in L. dugesii and L. perspicillata). Sternal fontanelle usually oval. (It may approach the cordi- form condition in L. laevis, L. andreanszkyi and L. danfordi.) Inscriptional ribs often absent or only one pair present. Non-autotomic caudal vertebrae most usually number 5 or 6, less commonly 4 and rarely 7. Tail vertebrae frequently B pattern or, less often, A: L. d. danfordi and L. d. anatolica but not L. d. kulzeri are peculiar in having C-type vertebrae while L. brandti, L. fraasii and L. parva have caudal sequences more or less intermediate between the B and C types. HEMIPENIS. Symmetrically bilobed; in most cases no obvious armature and the lobes not folded in repose ; armatures and complexly folded lobes do occur in L. vivipara, L. cyanura and L. jayakari and a more limited armature development may occur in L. cappadocica and L. graeca. Outer surfaces of lobes usually regu- larly plicate but not in L. parva or L. fraasii where the walls of the retracted organ are arranged in irregular folds and have longitudinal flaps. Lobes without large apical papillae ; their sulcal lips usually small or very small (larger in andreanszkyi, danfordi, perspicillata and laevis). Micro-ornamentation variable: most forms have crown-shaped tubercles but simple recurved spines occur in andreanszkyi, bedviagae, cappadocica, graeca, jayakari RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 335 and perspicillata. L. dugesi has spines that at first sight appear simple but when examined with a scanning electron microscope can be seen to have a number of spinules at their tips so that they are intermediate between the two types of ornamentation mentioned so far. L. brandtii, L. fraasii and L. parva all have irregular tubercles that seem to end in spinules. They often appear bifurcate but this is uncertain in the case of brandtii. (Micro-ornamentation data from Boéhme 1971, and personal observations.) ' The above comments on hemipenial structure do not cover the apparently all- female parthenogenetic forms allied to L. saxicola, viz. armeniaca, dahli, rostom- bekovt and unisexualis. EXTERNAL FEATURES. Small to large lacertids: adults from about 35 mm (andreanszkyt) to 165mm (jayakart) snout to vent, but most species between 50mm and gomm. Nostril usually in contact with first upper labial (often narrowly separated in some species, e.g. andreanszkyt, brandtii, cappadocica, fraasit, monticola, parva, perspicillata, vivipara), bordered posteriorly by one or two super- posed postnasals (three in some L. cappadocica), usually one in about half the species, one or two in bedriagae, fraasii and mosorensis, two in the remainder, viz. brandtit, cappadocica, cyanura, danfordi, dugesii, graeca, jayakari, laevis, oxy- cephala, parva and perspicillata. Eyelid fairly opaque and scaly in most forms ; a window of transparent, black-edged scales in L. cappadocica, a single-scaled window in L. perspicillata. Anterior part of lateral border of the parietal shield often emarginated and running medial to the edge of the postorbital bone in the majority of species but along the postorbital margin in andreanszkyi, dugesii, perspicillata, vivipara and some bedriagae. Supratemporals well developed in most cases, the first usually being longer and deeper than the rest ; sometimes the supratemporals not easily distinguishable from temporals, e.g. particularly dugesti, perspicillata. In cappadocica, cyanura and jayakari all the supratemporals rest at least partly on the bony parietal table; in jayakari the posterior supra- temporals are diagonally enlarged so that the anterior of each lies above the posterior portion of the one in front. Masseteric shield usually distinguishable in most species but not in cappadocica, cyanura, dugesii, graeca, jayakari, perspicillata and many andreanszkyi, bedriagae, danfordi and oxycephala. Dorsals small, 25 (some vivipara) to gi (some jayakari) in a transverse series at mid-body, in most species smooth or faintly keeled but more strongly so in chlorogaster, laevis, parva, praticola, rudis and vivipara. Collar well developed, serrated or smooth. Ventrals smooth, truncate, the degree of imbrication varying with habitat (least in rock-dwellers, greatest in ground forms); usually in six longitudinal rows (exceptions: brandtii eight to ten, jayakari and parva eight, perspicillata ten ; some individuals of cappadocica, danfordi, fraasii and laevis may also have eight rows). Preanal shield usually large, occasionally rather reduced with two or more semicircles of small scales anterior to it, e.g. cappadocica, danfordi, fraasti, jaya- kart, parva. Toes varying in degree of compression, usually tubercular beneath but unicarinate in L. cappadocica, strongly bicarinate in L. cyanura and weakly so in some L. parva. Femoral pores present, extending to, or nearly to knee; 7 (some vivipara) to 31 (some bedriagae) in each series. Tail long, unmodified. 336 E. N. ARNOLD CoLoRATION. Sexual dimorphism in dorsal patterning of adults absent in most forms (few exceptions, e.g. L. dugesti and to a lesser extent some populations of L. monticola and in L. graeca). Pattern very variable but most species lack the lateral pairs of narrow, well-defined pale stripes found in Podarcis and Gallotia (they are often present in andreanszkyi, parva and vivipara). Pattern usually consists of dark longitudinal stripes or series of markings which sometimes coalesce to form a reticulation (e.g. bedviagae, jayakari, oxycephala). In cappadocica, dan- fordi and perspicillata there seem to be distinct colour-morphs, the last two species including individuals with almost no markings, perhaps equivalent to the ‘ con- color’ morph of Podarcis. Venter usually brightly pigmented, particularly in breeding males although both sexes are often involved. L. jayakari appears to be exceptional in lacking any bright ventral coloration. The outer ventrals often have blue spots and these may extend onto the flanks in some cases. RELATIONSHIPS. Lacerta part II unlike the other groupings discussed in this paper contains a very varied assemblage of species (many of their principal differ- ences are summarized in Table III). It includes all the species originally assigned to the subgenera Zootoca, Archaeolacerta, Apathya and Scelarcis plus many of the species of uncertain position listed on pp. 298-299 and Lacerta dugesii which is usually classified with Podarcis (group 3). Many of the characters on which these subgeneric divisions were based appear to be ecologically labile and are thus relatively unimportant in assessing relationships (see pp. 315-324) ; once they are excluded from consideration, Lacerta part II in spite of its variability cannot be divided into discrete groups although it is often possible to suggest which forms are particularly closely related. However, for convenience, Lacerta part II will be discussed under a number of separate headings, even although the divisions between them are to a large extent artificial. (a) The more typical members of Lacerta part II. SPECIES DISCUSSED. armeniaca, bedriagae, caucasica, chlorogaster, dahli, dan- fordi, derjugini, graeca, horvathi, laevis, monticola, mosorensis, oxyvcephala, praticola, rostombekovi, rudis, saxicola, unisexualis. The above species occur in a broad, but disjunct band from northern Spain and Portugal across the northern Mediterranean region to the Caucasus and south Caspian coastal area. The majority of these forms have all the features listed as characteristic of Lacerta part II (although this does not necessarily imply that they are primitive). This applies especially to the more northern ones ; some of the southern species are rather different, frequently having two postnasals (dan- fordi, graeca, laevis, oxycephala and some bedriagae and mosorensis) and the mas- seteric shield absent or often reduced (bedviagae, danfordi, graeca and oxycephala). These southern species are often better differentiated than the more northern ones and some possess peculiar features. For instance, some populations of L. danfordi have C-type caudal vertebrae and very large sulcal lips on the hemipenis ; L. laevis and some L. danfordi often have a more or less cordiform sternal fontanelle and L. graeca has a distinct armature development in the hemipenis and simple spines (as opposed to tubercles with a ring of spinules at their tips) on the hemipenial The easily appreciated. presently available. Clear armature present Postorbital and postfrontal bones No microchromosomes in karyoty| Parietal shield extends to outer e postorbital bone . : Lateral arms of interclavicle clear posteriorly . . Masseteric shield absent or reducd Two postnasal scales present Hemipenial micro-ornamentation spines (+) or irregular often bifu tubercles (x ) Eight or more rows of ventral sca Prominent blue spots on flanks Caudal vertebrae not of A or B t Hemipenial lobes with longitudin unplicate : Chromosome number greatly rea Sternal fontanelle approaches hea Very large body-size Supratemporal scales diagonally e Supratemporal scales all rest on s} Belly always pale, without bright even in breeding males . Toes carinate beneath . Lower eyelid with ‘window’ Strongly expanded scales under fa 4agsvs040]Y9 vjooywad wuisntlsap DIISVINDI SUPNA * dnois-vjoo1xps vavdinrn Taste III: Distribution of characters present in a minority of Lacerta Part II species The characters are arranged not according to anatomical proximity but so that the common features of each loose species-group can be easil Y appreciated. t Parentheses indicate that the character is only sometimes present in limi i tai or present in limited form, a question-mark th: i at no certain data are F ; : Presently available, > o a > & = : z Sy S s y 8 = s 8 S a bc] > ° > be 9 ge ee ee Ee eS eS ee © Ne eee 4 2 § & s & 32 a ees Se § £ See oe y & 2 Foes 5 ; =: 2 2 = = ¢ & a 8 Oe EE ee 8 oe ee ee e = § 8. os Sth Sua: ho Ce 4S ee = : SF Ss 8 S a g ne og 2 r 3 S : postorbital bone . c. , ee: Lateral arms of interclavicle clearly directed posteriorly . - : ‘ 5 5 # Masseteric shield absent or reduced : Lo = = (+) de Two postnasal scales present : i Hemipenial micro-ornamentation of simple spines (+) or irregular often bifurcate tubercles (x) : 5 é 5 , x x Eight or more rows of ventral scales 3 ~ (8) 8 Prominent blue spots on flanks a - + + Caudal vertebrae not of A or B type . et a a = ay = is = es = a : = _ = = : + - + at ? 3 on Hemipenial lobes with longitudinal flaps ; unplicate . i 3 6 a Chromosome number greatly reduced 68 Sternal fontanelle approaches heart-shape _o- ae Very large body-size_ . ‘ 0 : a = = = = = ae Supratemporal scales diagonally enlarged 2 = _ - = = = a8 Supratemporal scales all rest on skull table . — = - - - - + + + = - = = = = = = = = = = = + apn ae ER Ge! TE EO CR” EP ee ES F Belly always pale, without bright colour, even in breeding males . : i é = - - = = = = Toes carinate beneath . 4 ni 2 . o> (+) - = = = fe Lower eyelid with ‘window’ . . . _ = - - - + - = = Strongly expanded scales under forearm Lo - = = = = ab + es ls Nl = le -- RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 337 epithelium ; similar spines occur in L. bedviagae (Bdhme 1971). Many of the northern species, on the other hand, are very similar to each other, indeed some of the key characters allegedly distinguishing allopatric forms do not always do this. This is true for L. horvathii and L. mosorensis. It seems probable that the geographical isolation of some of these northern species may be a relatively recent occurrence. (b) Lacerta vivipara. This species has an immense and largely continuous distribution mainly to the north of the area occupied by the species discussed in the last paragraph. It is probably related to the more northern of these forms but it differs from them all in a number of features: the hemipenis has a well-differentiated armature and folded lobes, the postorbital and postfrontal bones are fused even in newly deposited young, the edge of the parietal shield reaches the outer margin of the bony parietal table, one pair of microchromosomes is lacking from the usual Lacerta karyotype (which is present in the 12 typical members of Lacerta part II that have been investigated to date; see p. 310) and unlike the rest of Lacerta, L. vivipara is Ovoviviparous over most of its range. L. vivipara is usually associated with L. derjugini and L. praticola in the sub- genus Zootoca but there are no good grounds for doing this. The resemblance between the three species (body-shape, rather large dorsal scales, serrated collar, some overlapping of ventrals, often complete supraocular lamellae) is connected with their similar habitats — all are ground-dwelling lizards occurring most usually in rather moist herbaceous places (Lantz & Cyrén 1947). Neither L. derjugini nor L. praticola has any of the vivipara characters listed above, instead they are very similar to the other typical members of Lacerta part II in their general features and it is almost certain that this is where their close affinities lie, indeed L. der- jugini frequently hybridizes in the wild with one of the typical species, L. saxicola _ (Darevsky 1966). (c) Southeastern species. SPECIES DISCUSSED. cappadocica, cyanura, jayakart. These three well-defined species have some resemblance to each other. Like _ the more southern typical members of Lacerta part II they have two post-nasal scales (three in some L. cappadocica) and no masseteric shield. In addition all their supratemporals rest on the parietal table and they have at least some indi- cation of an enlarged series of scales beneath the forearm. L. cappadocica and L. jayakari both have the hemipenial epithelium micro-ornamented with simple recurved spines while that of L. cyanura has crown-shaped tubercles. L. cappadocica has a number of rather peculiar features: its hemipenis may have a fairly weak but definite armature and the lobes of the hemipenis may be somewhat folded, the parietal shield has a concave lateral margin, the eyelid a transparent window of black-edged scales and the digits a single row of keels beneath. Although quite distinct from it, this species probably has affinities with its geographical neighbour, L. danfordi, which it resembles in some features of 25 338 E. N. ARNOLD head-scaling, in body proportions in similar polymorphic colour patterns and in their both often having a series of feebly enlarged scales beneath the forearm. L. jayakari resembles L. cappadocica in the features common to the three south- eastern species and also in possessing a hemipenial armature and folded hemipenial lobes. However, the armature and lobe-folding are much better developed than in L. cappadocica and the degree of enlargement in the scales under the forearm is much greater. JL. jayakari also differs from L. cappadocica and from the greater part of Lacerta part II in its large size, lack of belly pigment (apparently even in breeding males), diagonally elongated posterior supratemporals and its strongly keeled, straight caudal scales. In the first two characters it resembles Lacerta part I but does not possess the other features characteristic of the members of that group. L. cyanura, which is sympatric with L. jayakari in Oman, seems to have close affinities with this species in spite of its much smaller size and possession of bright belly colouring (blue) in the male (Arnold 1972). The two species are alike in most skeletal features including the number of presacral vertebrae in females (26), most aspects of hemipenial structure, the very enlarged plates under the forearm and the strongly keeled caudal scales. The pattern of the one known juvenile has some resemblance to that found in the wolteri Bird 1936 form of L. cappadocica. Sub- digital keeling is also found in both L. cyanura and L. cappadocica but the pattern differs, L. cyanura having two rows of keels beneath each digit and L. cappadocica one. In spite of their considerable differences it is likely that the two Oman species are closer to L. cappadocica than to any other member of Lacerta. (d) Southwestern species. SPECIES DISCUSSED. andreanszkyi, dugesii, perspicillata. One of these species, L. dugesii, has usually been classified with the members of Podarcis but it lacks many of the features that occur in all members of that genus and to be much closer to members of Lacerta part II especially the other southwestern species. Among the features it possesses that do not occur in Podarcis are the following. (1) Interclavicle with lateral arms directed posteriorly. (2) Oval sternal fontanelle. (3) A and B pattern caudal vertebrae. (4) Hemipenis with short apical sections to the lobes and small sulcal lips. (5) Almost always two superposed postnasals. (6) No masseteric shields or well-defined supra- temporals. (7) Often a dorsal pattern including rather broad pale supraciliary stripes that become broader and fainter posteriorly (similar to that found in some members of Lacerta part II, e.g. L. caucasica) ; other specimens have a reticulated or striated pattern rather like that of some L. perspicillata. (8) Hemipenial orna- mentation consisting of recurved spines with minute spinules on their tips. The three southwestern species are alike in having the parietal shield reaching the edge of the postorbital bone, in usually lacking the masseteric shield (a masseteric sometimes present in andveanszkyi) and in having a micro-ornamen- tation of more or less simple spines on the hemipenial plicae (distinctive pattern of L. dugesii is mentioned above). Each species is well differentiated from the RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 339 others and has its own peculiarities but, as with the southeastern species, each pair of the trio has features in common that, with the characters listed above, suggest that they are perhaps quite nearly related. Thus L. perspicillata and L. dugesti both have a peculiar interclavicle structure, two postnasals and similarity in dorsal patterns; L. andreanszkyi and L. dugesii have six rows of ventrals and L. andreanszky1 and L. perspicillata share very large sulcal lips on the hemipenial lobes. L. perspicillata has customarily been separated from the rest of Lacerta in the subgenus Scelarcis because it has ten longitudinal rows of ventrals and a brille in the lower eyelid. Ten rows of ventrals occur elsewhere in Lacerta part II (in some L. brandti:) and, as explained on p. 311, windows in the lower eyelid are a relatively frequent development in several lines of lacertids ; indeed L. dugesii has a number of enlarged and translucent scales in this region. Therefore it does not seem necessary to separate L. perspicillata from the rest of Lacerta part II on these grounds. It is uncertain how the three southwestern species relate to the rest of Lacerta part II. Their nearest geographical neighbour is L. monticola of the northern and central Iberian Peninsula but this species, which appears to have its closest affinities with the more typical members of Lacerta part II (see p. 336), does not resemble the southwestern species very closely. The latter have more in common with L. danfordi and neighbouring species in the eastern Mediterranean region. Like them they may have two postnasals, no masseteric shields and large hemipenial lips. There is also some resemblance in habitus and in dorsal patterns, especially those of L. danfordi to L. dugesii and L. perspicillata. Such a relationship would not be entirely unexpected as there are other cases of zoogeographical links between the eastern Mediterranean region and Northwest Africa. (The nearest relations of the following Nortl:west African species, Acanthodactylus erythrurus (Schinz) and A. savignyi (Audouin), Ophisaurus koellikeri (Ginther) and Vipera lebetina mauri- tanica (Gray) seem to be respectively Acanthodactylus tristrami (Giinther), Ophi- saurus apus (Pallas) and Vipera 1. lebetina (Linnaeus) which are all essentially east Mediterranean forms. The southwestern species may also have some affinities with Podarcis, see p. 355.) (e) Lacerta parva and its relations. SPECIES DISCUSSED. brandtit, fraasit, parva. As Peters (1962) has pointed out, two of these species, namely L. parva and L. fraasii, have a considerable external resemblance to each other and this author thought that they were closely related but set somewhat apart from the rest of Lacerta. The close affinity of the two species is confirmed by their internal mor- phology. Both species have similar skeletons with most commonly 28 presacral vertebrae in males and 29 in females, unemarginated clavicles in all the specimens examined and tail vertebrae that are peculiar in being intermediate between the B and C patterns. In both, the hemipenes are unlike those of any other Lacerta in lacking regular plicae and having longitudinal flaps on the lobes instead. The hemipenial micro-ornamentation consists of irregular, sometimes bifurcate tubercles. L. parva is also singular in having a very reduced chromosomal formula (21 = 24) 340 E. N. ARNOLD (Gorman 1969) and is the only species of the genus known to have this degree of reduction. The karyotype of L. fraasii has not been investigated to date. L. parva and L. fraasii would occupy an isolated place in the genus if it were not for L. brandtii which is morphologically intermediate between these forms and the more typical members of Lacerta part II. L. brandtii shares the following characters with L. parva and L. fraasii: tail vertebrae intermediate between the B and C patterns, two postnasals present, nostril often narrowly separated from the first upper labial, often eight longitudinal rows of ventrals, usually two well- separated series of dark markings on the perivertebral area, well-marked blue ocelli on the flanks and perhaps similar hemipenial micro-ornamentation. It differs from them in having a lower presacral vertebral count (26 in males, 27 in females), usually emarginated clavicles, and a hemipenis with clearly defined plicae on the lobes and no longitudinal flaps. In all these features it agrees with at least some of the less aberrant members of Lacerta part II. DISTRIBUTION (see Figs. 12 and 13). Members of Lacerta part II are scattered over almost the whole range of Lacerta, but with the exception of L. vivipara, which is found over an enormous area, the species have small or disjunct ranges. This is almost certainly a relict distribution and indicates that the group has under- gone considerable reduction in the area of its total range. In the more southern regions many of the allopatric species are strongly differentiated, probably indi- cating that their separation is of long standing but, in the north, the species are more similar to each other and here the reduction in range presumably was rather more recent. It is uncertain what caused the shrinkage in distribution ; most species are now confined to relatively moist or highland habitats (exceptions : some L. danfordi — Mediterranean islands ; L. dugesii — oceanic islands ; L. brandtit, L. fraasii and L. parva—steppe-type habitats) ; possibly the post-glacial tem- perature-increase may be important. It is difficult to assess how relevant the spread of Podarcis has been in this process ; these lizards seem to be better adapted to drier and warmer environments than the members of Lacerta part II. Where the two groups occur together, the latter seems best able to compete at high alti- tudes and, where the two groups are found living side by side, the Lacerta part II species usually occupy scansorial or moist terrestrial niches. Whether there has been active elimination of Lacerta part II populations by direct competition is uncertain. Certainly the expansion of Podarcis can only have been one of several factors in the reduction of the range of Lacerta part II, as the latter has undergone an apparent contraction in distribution in many areas where Podarcis does not occur. BroLtocy. The members of Lacerta part II are mainly small lizards that occupy a wide variety of structural niches but the great majority are adapted to living on and around rock faces, the principal exceptions being as follows. L. chlorogaster spends considerable amounts of time on tree-boles (Lantz & Cyrén 1947, Terentiev & Chernov 1965, Droedov 1967) and this may originally have been the main habitat of L. laevis (e.g. Béhme 1971, Zinner 1967). This would ‘explain’ why these species lack certain of the characters of their close relatives originally placed in the subgenus Archaeolacerta, viz. smooth, flattened dorsal scales and strongly RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 341 depressed skull with incomplete ossification of the supraocular lamellae, these features being adaptations to utilizing rock crevices as hiding-places (see p. 320). L. brandtii, L. fraasii and L. parva all occur on the ground in relatively dry and open habitats (Lantz & Cyrén 1939, Wettstein 1928). L. derjugini, L. praticola and L. vivipara also tend to be ground-dwelling but in damp places (Lantz & Cyrén 1947). L. andreanszkyi does not climb high on rock faces but is at least partly saxicolous, often occurring on stones and coarse gravel in or near stream- beds (K. Klemmer, personal communication ; S. D. Busack, personal communi- cation). L. dugesii is often scansorial but occupies a wide range of habitats. There is some evidence that at least the more northern species may operate at rather lower preferred temperatures than sympatric species of Podarcis (personal obser- vations on L. horvathi, L. mosorensis, L. oxycephala, L. monticola and L. vivipara and on P. melisellensis, P. muralis and P. sicula). ALGYROIDES Bibron & Bory Bibron & Bory, 1833, in Bory, Expéd. sci. Morée, 3, 1: 67. TYPE SPECIES. Algyroides moreoticus Bibron & Bory. Diacnosis. Small lacertids (adults not usually over 70mm snout to vent). Skull fairly robust or delicate and depressed, parietal foramen present, frontal bones paired throughout life, postorbital and postfrontal bones unfused ; clavicle strongly expanded medially and exists in both emarginated and unemarginated conditions in all species; interclavicle cruciform, the lateral arms not strongly directed forwards or backwards ; sternal fontanelle usually oval; sexual variation in presacral vertebral count present ; all presacral vertebrae, except the first three cervicals, with ribs; free ribs divided into two series, an anterior one of long ribs and a posterior series of short ones ; caudal vertebrae of the A or B pattern. Hemipenis symmetrically bilobed with no obvious armature, the lobes unfolded in repose and regularly plicate with variable micro-ornamentation ; apical region of lobes short with very small sulcal lips ; no large conical papillae at lobe tips. Head shields normal, nostril usually in contact with first labial scute, normally two superposed postnasal shields (occasionally one) ; lower eyelid scaly; an- teriorly the parietal scute does not extend to the outer margin of the postorbital bone ; first supratemporal shield large ; masseteric shield frequently present. Dorsal body-scales large (larger than caudals in at least the mid-dorsal region), collar present and well defined ; ventral scales smooth, truncate, not overlapping very strongly, in six longitudinal rows. Toes cylindrical or compressed, tubercular beneath, femoral pores present. Tail unmodified. Sexual dimorphism absent in most species. Belly brightly coloured in breeding males. SPECIES REFERRED. fitzingeri, marchi, moreoticus, nigropunctatus. SKELETAL FEATURES. Skull robust in A. moreoticus and A. nigropunctatus, more delicate and depressed in A. fitzingeri and A. marchi which also have the osteodermal layer reduced and fenestrated supraocular lamellae. No ossification 342 E. N. ARNOLD in the temporal area. Postorbital and postfrontal bones always separate, pterygoid teeth present in A. moreoticus and A. nigropunctatus. Presacral vertebrae usually 25-26 in males, 26-27 in females (total range 24-28). Clavicle exists in both intact and emarginate conditions in all species. Interclavicle cruciform, the lateral arms not strongly directed forwards or backwards. Sternal fontanelle oval (may ap- proach cordiform condition in some A. moreoticus). Inscriptional ribs usually nil, occasionally one pair present. Non-autotomic caudal vertebrae four or five. Caudal pattern usually B, occasionally A. HeEmIPENIs. As diagnosis. Micro-ornamentation consists of recurved spines in A. moreoticus and A. nigropunctatus and crown-shaped tubercles in A. /fitzingert. EXTERNAL FEATURES. Small lizards (adults from 30 to 70mm snout to vent). Nostril in contact with first upper labial ; normally two postnasals present (excep- tions rather frequent in A. moreoticus, rare in the other species). Eyelid scaly. Parietal border emarginated running medial to the outer margin of the postorbital bone (this condition least developed in A. marcht). Supratemporals well developed, the first larger and deeper than the rest. Masseteric shield usually well defined. Dorsal scales large (larger than the caudals), imbricate, strongly keeled (less so in A. marchi), either pointed (A. fitzingeri, A. moreoticus) or truncate (A. march, A. nigropunctatus) ; number of dorsals in a transverse series at mid-body 15 to 28. Collar well developed, fairly smooth or serrated. Ventrals truncate, smooth, with- out very strong overlap, in six longitudinal rows. Preanal shield large, bordered by one or (more rarely) two semicircles of small plates. Toes cylindrical or com- pressed, tubercular beneath. Femoral pores present, each series extending to the knee and containing 11 to 18 pores. Tail unmodified. COLORATION. Sexual dimorphism in dorsal pattern absent except in A. moreo- ticus. Most individuals basically brown or bronze-brown above, the flanks often darker (marchi, nigropunctatus, some fitzingeri and male moreoticus; the latter also having light spotting on the flanks). Dorsum may have irregular dark spots (nigropunctatus, male moreoticus), or a vertebral stripe (marchi, fitzingert), or be uniform (some fitzingeri and nigropunctatus, female moreoticus). Male moreoticus have light, narrow, dorsolateral stripes. Venter brightly coloured in males, the colour extending onto the flanks in A. nigropunctatus. Throat and belly may contrast (blue: orange in nigropunctatus, blue or white: yellow in marcht) ; blue spots may be present on the outer ventrals. RELATIONSHIPS. The four species of Algyroides are generally quite similar and they seem to constitute a natural group. They resemble each other quite closely in external morphology, osteology, and in having very small sulcal lips on the hemipenial lobes. Externally A. nigropunctatus seems most similar to A. marchi and A. fitzingeri to A. moreoticus. The first two species have small lateral scales and truncate dorsals while the last two have no differentiation between dorsals and laterals, all scales being both large and pointed. It is uncertain that these superficial resemblances indicate true relationships; the pattern of hemipenial micro-ornamentation does not support such an arrangement since A. moreoticus has simple recurved spines on the lobe plicae which are like those of its geographical RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 343 neighbour A. nigropunctatus and unlike the crown-shaped tubercles occurring in A. Jitzingeri (the pattern of micro-ornamentation in A. marchi is not yet known). DISTRIBUTION (see Fig. 15). The four species each have relatively small ranges in southern Europe and are completely or almost allopatric (moreoticus and nigro- punctatus coexist in the Ionian Islands). Algyvoides occurs in southeastern Spain (perhaps also central Spain, see, e.g., Buchholz 1965), Corsica, Sardinia, the eastern Adriatic seaboard as far south as Epirus, the Ionian Islands and southern Greece. BioLocy. There is some evidence that these small lizards were perhaps originally largely associated with semi-shaded woodland habitats (marchi —-Klemmer 1960, personal observation; moreoticus—Clarke 1970, personal observation; mnigro- punctatus — personal observation). Populations are now found in deforested areas but some Algyroides occur quite frequently in and around fallen timber and others live among and climb in bushes. PODARCIS Wagler Wagler, 1830, Syst Amph. p. 154. TYPE SPECIES. Seps muralis Laurenti. DracGnosis. Small, occasionally medium-sized lacertids (adults not usually over g0 mm snout to vent). Skull usually fairly robust but may be relatively depressed with a fairly thin osteodermal layer ; parietal foramen present, frontal bones paired throughout life, postorbital and postfrontal bones unfused; clavicle strongly expanded medially, existing in both emarginated and unemarginated conditions (often within the same population) ; interclavicle cruciform, the lateral arms not strongly directed forwards or backwards ; sternal fontanelle usually heart-shaped ; sexual variation in presacral vertebral count present; all presacral vertebrae, except the first three cervicals, with ribs ; free ribs in two series, long anterior ones differentiated abruptly from short posterior ones ; caudal vertebrae of the C pattern. Hemipenis symmetrically bilobed with no obvious armature, the lobes unfolded in repose and regularly plicate with a micro-ornamentation consisting of simple, often recurved spines; apical regions of the lobes relatively long (usually longer than basal portions of the lobes) with extremely large sulcal lips; no large conical papillae at lobe tips. Head shields normal, nostril usually in contact with first upper labial scute and usually bordered posteriorly by a single postnasal shield (occasionally two), eyelid scaly ; anteriorly the outer edge of the parietal shield runs along the lateral margin of the postorbital bone; supratemporals present, rather narrow; a masseteric shield frequently present. Dorsal scales small (less than half the length of the proximal caudals) ; collar well developed ; ventral plates smooth, truncate, not strongly overlapping, in six (rarely eight) longitudinal rows. Toes somewhat compressed, tubercular beneath ; femoral pores present. Tail unmodified. Sexual dimorphism in dorsal pattern usual in most populations. Belly usually brightly coloured, at least in breeding males. 344 E. N. ARNOLD SPECIES REFERRED. erhardit, filfolensis, hispanica,* lilfordi, melisellensis, milen- sis, muralis, peloponnesiaca, pityusensis, sicula, taurica, tiliguerta, wagleriana. SKELETAL FEATURES. Skull usually fairly robust, sometimes quite depressed with a reduced osteodermal layer, particularly in P. hispanica. Supraocular lamellae complete in most adults (some exceptions especially in P. hispanica), temporal region usually unossified (fairly extensive ossification in some P. pelopon- nestaca and traces in individuals of some other forms, e.g. P. taurica and P. melisellensis). Postorbital and postfrontal bones always separate in hatchlings, occasionally fused in adult males, pterygoid teeth present in at least some individuals of most species (not P. lilfordi or P. wagleriana fide Klemmer 1957). Presacral vertebral number shows some correlation with sex, 27 commonest number for males (26 in P. tiliguerta, P. filfolensis, P. muralis and P. pityusensis), 28 in females (27 in P. filfolensis, P. lilfordi, P. pityusensis and P. tiliguerta) ; males occasion- ally have 25 presacral vertebrae, females 29. Clavicle expanded medially, most frequently emarginated but some specimens of the majority of species have the bone intact (not in L. filfolensis and L. muralis). Interclavicle cruciform, the laterally directed arms not strongly angled forward but may occasionally be angled very slightly backwards, especially in L. wagleriana. Sternum of most individuals with a cordiform fontanelle having a very distinct, posteriorly directed process arising from its anterior border. One inscriptional rib-pair may be present. Non- autotomic caudal vertebrae usually five or six, sometimes four. Caudal pattern always of the C type. HEMIPENIS. Symmetrically bilobed without an obvious armature; lobes not folded in the retracted organ, plicate. Apical sections of the lobes usually longer than basal parts, sometimes subequal; lobe tips without prominent papillae. Lips on lobe sulci very large, the outer one varying in its mode of attachment to the lobe wall as follows. P. hispanica, P. muralis: the outer lip is free laterally for its whole length and can be easily reflected to expose its underside which shows traces of the plication that covers the rest of the lobe wall. All other species: here there is an obvious longitudinal fold in the lobe wall directly dorsal (in the retracted organ) to the lip which is fused to the fold apically but is free basally. Hemipenial micro-ornamentation in Podarcis consists of simple recurved spines (Bohme 1971 ; personal observations). EXTERNAL FEATURES. Small to medium-sized lizards (adults usually from 50mm to 80mm snout to vent, occasionally over go mm). Nostril in contact with first upper labial, bordered posteriorly usually by a single postnasal (occasional exceptions, especially unilateral ones). Eyelid scaly. Anteriorly the lateral border of the parietal is unemarginated, usually following or close to the outer edge of the postorbital bone. Supratemporals often distinguishable, rather narrow, * P. hispanica, as presently understood, may not be a single species (Klemmer, personal communica- tion). RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 345 the first often the longest ; in some forms supratemporals are small and scarcely separable from the similarly sized temporals (e.g. hispanica, lilfordi, pityusensis and some ¢zliguerta). Masseteric present in most individuals but frequently small or absent in P. hispanica. Dorsals small, 42 to 90 in a transverse row across the mid-body, usually granular with a raised transverse section (fairly flat in some P. hispanica), sometimes smooth or faintly keeled (erhardii, filfolensis, hispanica, lilfordi, milensis and pityusensis), sometimes quite strongly keeled (melisellensis, sicula, taurica and wagleriana). Collar well developed, usually fairly smooth-edged but sometimes distinctly serrated (P. taurica). Ventrals smooth, truncate, degree of overlap variable but never very strong, arranged in six (rarely eight) longitudinal rows. Preanal shield with one or, more rarely, two semicircles of small scales anterior to it. Toes vary in degree of compression, tubercular beneath. Femoral pores extend to knee, 13 to 29 in each series. Tail long, unmodified. COLORATION. Sexual dimorphism in the dorsal pattern is usual in adults. Some specimens have continuous longitudinal stripes on the body: there may be a pair of light stripes on each flank (starting from the supraciliary and supralabial regions) ; these may be separated and bordered by dark pigment and there is often a dark vertebral stripe as well. The pattern may vary, for instance particular elements may be absent or divided. In other specimens, the dark stripes are broken up into discrete, although often very irregular transverse bars which in extreme cases may coalesce to form a reticulation. In most forms, the females tend towards the striped condition while the males, by comparison, have the pattern more broken up. This does not apply to certain populations where both sexes have reticulated markings (e.g. P. f. filfolensis, P. muralis insulanica and P. m. nigriventris). Ground colour is very variable and may be brown, grey or green. In some species there may be a distinct polymorphism in dorsal pattern, some specimens having the normal dark markings while others lack them completely (usually referred to as “concolor-mutants’). This morph occurs in P. erhardii, P. filfolensis, P. hispanica, P. melisellensis, P. sicula, P. taurica and P. wagleriana. Insular melanic popu- lations are frequent. Ventral surface usually brightly coloured in breeding males, but not in most mainland populations of P. sicula, some populations of P. pelopon- nestaca (fide Buchholz 1960) or in many P. milensis ; other individuals of this last species have only the second rows of ventrals from the midline brightly coloured. Throat and belly may contrast ; the outer ventrals often bear blue spots. RELATIONSHIPS. The thirteen species of Podarcis form a very homogeneous assemblage. Lacerta dugesii which used to be classified with these forms in the sub- genus Podarcis in fact differs from them in many features and is more fully discussed on p. 338. Klemmer (1957) considered Podarcis peloponnesiaca to be incertae sedis because he felt that its often very robust skull and usually quite extensively ossified temporal region separated it from the other species here placed in Podarcis. How- ever, P. peloponnesiaca possesses all the features which in combination distinguish Podarcis from other lacertid genera (see Diagnosis). The characters that Klemmer thought separate this species from Podarcis involve only relatively slight changes in degree of ossification: thus the heavy skull of P. peloponnesiaca is approached 346 E. N. ARNOLD by that of other Podarcis species such as P. taurica, especially its southern populations, and other members of this genus may also have some (admittedly minor) temporal ossification, e.g. individuals of P. melisellensis and P. taurica. Because the species of Podarcis are all morphologically very similar, it is difficult to decide which forms are most closely related to each other. Most attempts to subdivide the group (e.g. Kopfstein and Wettstein 1921, Boulenger 1920, Klemmer 1957) have been based on external morphology, especially skull and body shape and colour pattern. Unfortunately the morphological variables are largely the sort of characters that often correlate fairly closely with the kind of niche occupied and on their own are uncertain indicators of relationship. Colour and pattern are also of rather limited use since it is often highly variable within species. The most recent subdivision of Podarcis was made by Klemmer (1957) who separated it into three assemblages. I. muralis-group : muralis, filfolensis, milensis, tiliguerta and wagleriana. 2. bocagei(= hispanica)-group : bocagei (= hispanica), lilfordi, pityusensis and dugesi.* 3. sicula-group : sicula, erhardit, melisellensis and taurica. Klemmer’s classification produces zoogeographically coherent groups: one in the West (ispanica-group), another in the East (s¢cwla-group) and the third (muralis- group) consisting of a single, widespread continental species, P. muralis, and four Mediterranean island forms, all occurring south of the main range of P. muralis, which are presumed to have originally been connected with it either directly as in the case of P. milensis and P. tiliguerta or via another member of the group: thus Klemmer believed that P. wagleriana and P. filfolensis are closest to P. tiliguerta. However, many of the characters on which Klemmer’s arrangement is based are rather equivocal and a wider range of evidence will be necessary to produce a con- vincing theory of Podarcis relationships. Chemotaxonomic studies may provide useful information (G. C. Ross and Arnold, in progress). But until this and other lines of evidence are adequately surveyed, it seems best to leave the question of Podarcis relationships open. Comment is limited here to two minor points. 1. Gross hemipenial structure often proves to be a good indicator of lacertid relationships. P. hispanica and P. muralis are alike in having the outer sulcal lips of the hemipenis free and differ from all other Podarcis in this respect. This may indicate that they are fairly closely related to each other. Such free lips are found in some of the members of Lacerta part II that may be related to the stock from which Podarcis was derived (see p. 355); so it is possible that /ispanica and muralis are the most primitive members of Podarcis in this respect. 2. P. erhardii and P. peloponnesiaca seem to be closely related. Although quite distinct on the Greek mainland, where they have a small area of sympatry, the two species are to some extent connected by island populations of P. erhardii that approach P. peloponnesiaca in build and colouring. P. peloponnesiaca and most populations assigned to P. evhardii share a similar hemipenis shape in which the outer sulcal lips are broader than in other species of Podarcts. * As stated above, L. dugesii is here removed from Podarcis; Klemmer did not include P. pelopon- nesiaca in any of his assemblages because he was uncertain whether it was related to the rest of Podarcis. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 347 The neighbouring P. ¢awrica might also be close to these two species, especially as some island populations usually identified as P. erhardii approach P. taurica in form and colouring. Thus the Podarcis population on Skyros (N. Sporades) was originally named as a subspecies of P. taurica (P. taurica gaigeae Werner 1930) but is now usually referred to P. erhardii. Its hemipenial lips, however, are rather narrow and tend towards the condition found in taurica. DISTRIBUTION (see Fig. 14). Mainly continuous including both highland and lowland areas but limited to the western part of the total region occupied by the West Palaearctic endemic genera dealt with in this paper. N.W. Arrica (P. hispanica only): Tunisia, N. Algeria, Morocco. MAINLAND Europe: All southern peninsulas and northwards to S. Netherlands, Rhine Valley, Bavaria, Czecho- slovakia, Hungary and Roumania; also Northwest Black Sea coast to Crimea (P. taurica) and N.W. Asia Minor (P. muralis). MEDITERRANEAN ISLANDS: Most islands with the exception of those lying east of a line running just west of Samothraki, Limnos, Agios Evstratios, Psara, Ikaria, Kos and Kasos in the Aegean Sea. Brotocy. These small lizards occupy a relatively broad spectrum of spatial niches. Most are at least partly scansorial and some spend much of their time on semi-vertical surfaces (e.g. P. muralis and P. hispanica). Others (like P. taurica) climb relatively little. PSAMMODROMUS Fitzinger Fitzinger, 1826, Neue Classif. Rept. : 22. TYPE SPECIES : Psammodromus hispanicus Fitzinger. Diacnosis. Small lacertids (adults usually under 80 mm snout to vent). Skull moderately built or robust, not depressed, a parietal foramen present, frontal bones paired throughout life; postorbital and postfrontal bones fused even in hatchlings ; clavicle strongly expanded medially with a large always unemarginated foramen ; interclavicle cruciform, the lateral arms not directed strongly forwards or backwards; sternal fontanelle oval; sexual variation in presacral vertebral count present ; all presacral vertebrae, except first three cervicals, with ribs ; free ribs in two series, long anterior ones differentiated abruptly from short posterior ones ; caudal vertebrae of the C pattern. Hemipenis symmetrically bilobed with no obvious armature; lobes not folded in repose, regularly plicate with variable micro-ornamentation ; apical region of lobes short or quite long with small or moderately sized sulcal lips; fairly large conical papillae sometimes present on lobe tips. Head shields normal, nostril usually in contact with first upper labial and bor- dered posteriorly by a single postnasal scale; lower eyelid scaly; anteriorly parietal scale borders outer margin of fused postorbital-postfrontal bone ; supra- temporal scales well developed ; masseteric shield not usually discernible. Dorsal body-scales large, pointed, strongly keeled and overlapping ; collar absent or very weakly developed ; ventral scales smooth, truncate and overlapping, often 348 E. N. ARNOLD strongly so, in six or ten longitudinal rows. Toes cylindrical or slightly com- pressed with smooth or keeled lamellae beneath; femoral pores present. Tail long, unmodified. Sexual dimorphism in dorsal pattern absent. Belly usually brightly coloured in breeding males. SPECIES REFERRED. algirus, blanci, hispanicus, microdactylus. SKELETAL FEATURES. Skull undepressed, robust with a thick osteodermal layer in P. algirus, less so in the other species. Supraocular lamellae complete in adults, temporal region ossified in P. algivus. Postorbital and postfrontal bones always fused, pterygoid teeth present only in P. algirus. Presacral vertebrae usually 26 or 27 in males, 27 or 28 in females (total range 26 to 29). Clavicle expanded medially, always unemarginated. Interclavicle cruciform, the lateral arms not obviously directed forwards. Sternal fontanelle oval. One pair of inscriptional ribs often present (but not usual in P. hispanicus). Non-autotomic caudal verte- brae four or five. Caudal pattern always C. HEMIPENIS. Symmetrically bilobed without an obvious armature; lobes not folded in retracted organ, plicate. In P. blanci, P. hispanicus and P. microdactylus the apical parts of the lobes are short with quite large conical papillae at the tips and relatively small sulcal lips. P. algirus differs in having long apical sections to the lobes (twice as long as basal section) and the lobes themselves are slender with large lips that are free and expanded distally, and no conical apical papillae. Micro-ornamentation consists of simple recurved spines in P. algivus but in the other species the spines are irregular (and often bifurcated, Bohme 1971). EXTERNAL FEATURES. Small lizards (adults usually 35 mm to 80 mm snout to vent). Nostril in contact with first upper labial, bordered posteriorly by a single postnasal. Eyelid scaly. Border of parietal shield unemarginated, usually running along outer margin of fused postorbital and postfrontal bones. Supratemporals well developed, the first often deepest in most species (last frequently deepest in P. algivus). Temporal scaling coarse, masseteric shield not clearly defined. Dorsal scales large (at least half the length of the proximal caudals), imbricate, strongly keeled and pointed. Number of dorsals in a transverse series 21 to 32. Collar very weak (hispanicus, some blanci and some microdactylus) or absent. Ventral plates smooth, truncate, strongly or extremely strongly imbricate (P. algivus), in six or ten* (P. algirus) longitudinal rows. Anal shield large, bordered by one or two rows of smaller plates. Toes not strongly compressed, tubercular beneath or with two series of keels (keeling strong in P. hispanicus, obtuse in some P. mucro- dactylus and some P. algirus). Series of femoral pores extend to knee and consist of 10 to 21 poresin each. Tail unmodified, very long in P. algivus, as much as three times the snout—vent length. COLORATION. Sexual dimorphism in dorsal pattern absent. Most individuals have light, narrow supraciliary and supralabial stripes extending along the body * In lacertids with large dorsal scales, it is often difficult to decide where these end and the equally large ventrals begin. But in forms with small dorsal scaling easily distinguished from the ventrals, the distribution of the latter corresponds closely with the rectus abdominis superficialis muscle. If this muscle is used as a criterion in defining the ventrals of Psammodromus species, it appears that P. algivus has ten longitudinal rows of ventrals rather than six as often stated (e.g. Boulenger 1921). RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 349 (these may be absent, especially in P. blanci and P. microdactylus). Other elements of the dorsal pattern rather variable. Venter brightly coloured in breeding males of the small species. P. algiyus has a bright throat but a pale belly in breeding males. RELATIONSHIPS. It is usually assumed that the three small species of Psammo- dromus, viz. blanci, hispanicus and microdactylus, are more closely related to each other than they are to P. algirus. This appears to be true; P. algirus differs from them in several features including hemipenial structure, peculiar almost skink-like ventral scaling, extremely long tail and heavily armoured head. However, this species seems to be more closely related to the remainder of Psammodrvomus than it does to any other group of lacertids. The small species, P. blanci and P. micro- dactylus, appear to be very closely related and may even be conspecific. DIsTRIBUTION (see Fig. 15). N.W. Arrica: Tunis, Algeria, Morocco. Europe: Spain, Portugal and southern France. Brotocy. The three smaller species all typically occur in open areas often with low, dense vegetation in which they hunt and hide. P. algirus is often found in scrub regions, especially in the vicinity of bushes and shrubs. When disturbed, it retreats into the base of these and may sometimes climb in them. GALLOTIA Boulenger Boulenger, 1916, Ty. Zool. Soc. 21: 3. TYPE SPECIES: Lacerta galloti Duméril & Bibron. Diacnosis. Medium to large lacertids (adults of extant species up to 210 mm snout to vent). Skull robust, a parietal foramen present (not in the extinct Lacerta (= Gallotia?) maxima according to Bravo 1953); frontal bones paired throughout life, postorbital and postfrontal bones fused, even in hatchlings ; clavi- cle strongly expanded medially, foramen always unemarginated ; interclavicle cruciform, the lateral arms not strongly directed forwards or backwards ; sternal fontanelle oval; no sex correlated variation in presacral vertebral count; all presacral vertebrae, except first three cervicals, with ribs; free ribs in two series, long anterior ones differentiated from the short posterior ribs; caudal vertebrae of the C pattern. Hemipenis symmetrically bilobed with no obvious armature; lobes not folded in repose, regularly plicate with micro-ornamentation of simple recurved spines. Apical regions of lobes usually fairly long, sulcal lips small, large conical papillae present on lobe tips. Head shields normal, nostril in contact with first upper labial scute and bor- dered posteriorly by a single postnasal scale, eyelid scaly ; anteriorly parietal scute may or may not reach the outer margin of the fused postorbital and postfrontal bones ; supratemporals present, the anterior ones sometimes narrower than the rest ; a masseteric shield may be present. Dorsal body-scales small or moderate sized (largest may be half the length of the proximal caudals) ; collar present and well defined; ventral plates smooth, 350 E. N. ARNOLD truncate, not strongly overlapping, in (8) 10 to 20 longitudinal rows. Toes not strongly compressed, tubercular beneath; femoral pores present. Tail long, unmodified. Sexual dimorphism in dorsal pattern usual. Belly brightly coloured at least in breeding males. SPECIES REFERRED. atlantica, galloti, simonyt. Perhaps also the extinct Lacerta goliath Mertens 1942 and Lacerta maxima Bravo 1953- SKELETAL FEATURES. Skull robust, undepressed with thick osteodermal layer. Supraocular lamellae complete in mature specimens, temporal ossification present in G. atlantica and G. simonyi simonyi, much reduced or absent in G. galloti and G. simonyi stehlint. Postorbital and postfrontal bones fused throughout life, pterygoid teeth present. Almost unique in the family Lacertidae in having ne sex-correlated variation in the number of presacral vertebrae there being 26 in nearly all the specimens examined (occasionally 25). Clavicle expanded medially, always unemarginated. Interclavicle cruciform, the lateral arms not directed strongly forwards, although they may be swept slightly backwards in some G. galloti. Sternal fontanelle oval. Most frequently one well- developed pair of inscriptional ribs in G. galloti and G. simonyi but usually none in G. atlantica. In most cases five non-autotomic caudal vertebrae (sometimes four in G. galloti and G. atlantica and sometimes six in G. stmony1). Caudal pattern always C. HEMIPENIS. Symmetrically bilobed without an obvious armature ; lobes not folded in retracted organ, plicate. Apical sections of lobes longer than basal parts or approximately equal to them ; lobe tips covered by prominent papillae which are conical, rather flattened and about equal in length to the widths of two to four plicae. Lips on lobe sulci relatively small, their free edges papillate. Micro- ornamentation consists of simple recurved spines. EXTERNAL FEATURES. Medium to large lacertids (adults from about 60 mm to 210mm snout to vent). Nostril in contact with first upper labial, bordered posteriorly by a single postnasal. Eyelid scaly. Border of parietal shield running along the outer edge of fused postorbital and postfrontal bones in G. atlantica and G. galloti, medial to this in G. simonyi. Supratemporals present, the first deep in G. simonyi, shallow in the other two species. A masseteric shield present in many G. galloti and G. simonyi, absent in G. atlantica. Dorsal scales very small (G. galloti and G. simonyt) or moderately large (G. atlantica), 45 to 110 in a trans- verse row across the mid-body, distinctly keeled, very strongly so in G. atlantica. Collar strong ; serrated or smooth. Ventrals smooth, truncate, not very strongly imbricate, arranged in 10 (rarely 8) longitudinal rows in G. atlantica, 10 to 14 in G. galloti and 16 to 20 in G. simonyi. Preanal shield rather small, the scales in this region rather fragmented. Toes not strongly compressed, tubercular beneath. Femoral pores extend to knee, 17 to 33 in each series. Tail unmodified. CoLoRATION. Sexual dimorphism in dorsal colouring of adults present in most populations. Young typically with two pairs of light, well defined flank stripes, RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 351 one supraciliary, the other supralabial. These are usually separated and bordered by black pigment which may be either continuous or broken up. There may also be a light vertebral stripe and the pattern may be overlaid by irregular transverse rows of small white ocelli. The throat often bears a series of dark chevrons or is entirely dark. Females (and a few males of some forms) tend to retain the juvenile pattern of stripes although it may be fainter. In males (and some females, especially of G. simonyi) the dorsal coloration becomes more uniform (and in G. galloti and G. stmonyi darker as well). Males (and females of G. atlantica) have prominent blue (G. atlantica, G. galloti) or yellow (G. simonyz) ocelli on the flanks and the outer ventrals are also often blue. In G. g. galloti the transverse rows of white ocelli on the back are also replaced by blue. The belly is also often brightly pigmented in adults although this colour is often largely obscured by melanin. RELATIONSHIPS. In spite of considerable external differences, the three extant species of Gallotia appear to be relatively closely related to each other, albeit perhaps not as closely as the members of Podarcis or of Lacerta part I. It is generally assumed that the fossil forms discovered in Pleistocene deposits on the Canary Islands are also members of Gallotia. These have been found on some of the large islands in the western group (La Palma, Tenerife, perhaps Gomera) and have been named as Lacerta (Gallotia) goliath Mertens 1942 and Lacerta (Gallotia) maxima Bravo 1953. As indicated by their names, these fossil forms were larger than any extant species of lacertid, L. maxima having a skull length of over 120 mm and an estimated total length of 1200 mm (Bravo 1953). Insufficient data are available from the published descriptions to allow these giant forms to be firmly allocated to Gallotia but their skull shape and head scale impressions in the osteodermal layer resemble those of G. stmony1. If this indicates a real relationship, then this apparent species-group (i.e. goliath, maxima and simonyi) has undergone a con- siderable reduction in range. It does not now occur on La Palma, Tenerife or Gomera and has recently become extinct on Hierro and the nearby Roques Zalmor (Klemmer, personal communication) where it was represented by G. simonyi simonyt. The only known population remaining is on Gran Canaria (G. simonyi stehlint). On all the islands where the lizards of the putative simonyi-group have disappeared, G. galloti now exists. Possibly competitive exclusion by this species has been important in bringing about the extermination of the s?monyi-group on some islands but it is uncertain if this was the case on Hierro. Here G. stmony1 was represented by a very large-bodied population (adults up to 210 mm snout to vent) while the resident race of G. galloti (G. galloti caesaris (Lehrs)) is much smaller (less than 100 mm snout to vent) which suggests that the two forms were unlikely to compete directly, unless of course there was competition between the young simonyi and the adult galloti. The three species definitely assigned to Gallotia have no very close relationship to the lizards assigned to Lacerta part I. The affinities of Lacerta lepbida which was previously placed in Gallotia by Peters are discussed on p. 333. DISTRIBUTION (see Fig. 15). Restricted to the Canary Islands. fa) =) fo) A fe < 352 dnois sawads vadwd “7 ay} Jo uoryNqiystp ayeurxoidde :sevare pasopuy ‘"y Wed vjsaan7 :Suyddys ‘py yaed vysaov7 yo dnois-vaswd -T oy} pue J pred vysaov7T Jo uoNNqiysIq jo uoljnqiuysip 9yeunxoidde 1ypue ‘CI “SI 353 RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS (dno1s-satsads vaswd 7 ay} 1Oy ZI “BY pue vavdiara “TJ 10p 1 “Bt aas) “zavyndvl 7 pue vanuvio ‘T :O ‘v2090pvddvo JT :N “Aagsvso41ojy9 “JT IN ‘SYvnxasiun “JT pue sipna “J ‘1a0yaquiojsos “J ‘wjooywad “J yO suoyetndod ussjsea ‘2weSnluap -T ‘yyvop “TJ ‘DIISVINDI “T ‘VIDIUGUAD “T : SUSBONLD 9Y} JO UOISAI [eIOUIS dy} 0} PpouyUoS ore SUIMOTIOJ OY} JNq voIe B[OYM IOAO SIN990 YjoILXDS “7 :soreds yo dno1rs wjosxys “J: ‘stoavy “JT zy “ypsofuwp “JT :[ -woaw4S “J : 1 “wjooyvad “Tz yo suorze[ndod u1ajsom :H ‘vyoydaakxo “TT pue srsuasosom “J ix) “tygwasoy “JT : 4% ‘avsviupsq] :q ‘“Wyzsuvaes—pun “J :q ‘vwyprordsaad STS) ‘vjosyuom "JT :q ~usadnpyT :w ‘dnors-soiads vaso “7 ay} pue vavdrata “J Surpnpoxe [J yAed wysaav7T Jo uonqiuysiq “€1 ‘DW ED 26 E. N. ARNOLD Fic. 14. Distribution of Podarcis. Broken line indicates eastern limit of the genus. i) n———— / GALLOTIA\ Se ee Fic. 15. Distribution of Algyroides, Psammodromus and Gallotia. Black: range of Algyroides (A = A. marchi, B = A. fitzingeri, C = A. nigropunctatus, D = A. moreoticus). Stipple : range of Psammodromus. Broken line: range of Gallotia. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 355 BioLtocy. The lizards of the genus Gallotia are among the largest terrestrial vertebrates native to the Canary Islands. The eastern G. atlantica is partly scan- sorial and feeds mainly on small invertebrates, while the western G. galloti and G. simonyt climb less and eat a substantial amount of vegetable food (see, e.g., Krefft 1950). These vegetarian habits are sometimes attributed to the lack of small animal prey on the islands (e.g. Peters 1961) but it may be more likely that these lizards have partly occupied plant-eating niches because, in the depauperate fauna of the Canaries, these are not completely filled by more efficient herbivores. The species of Gallotia tend to produce small clutches of eggs (usually 1 to 3 - Krefft 1950) for their size. This compares with broods of 8 to 18 in species of Lacerta part I of equal body size (Peters 1961). RELATIONSHIPS OF ALGYROIDES, LACERTA, PODARCIS, PSAMMODROMUS AND GALLOTIA The fossil record of the Lacertidae is far too fragmentary to give any useful information about relationships within the family, so all speculation about this must be based very tentatively on comparisons of extant forms. If the groups of lizards discussed in this paper are compared with the rest of the Lacertidae, it is apparent that Lacerta part II has more features that are widespread in the Lacer- tidae as a whole than the other groups (see Table IV). Lacerta part II is a very varied assemblage with many, often well-differentiated species and has a complex relict distribution. These features suggest that it is long established. It could be argued from this and its greater resemblance to the rest of the family that Lacerta part II might be closest to the stock which gave rise to the endemic West Palae- arctic group of genera. Such a conclusion is of course highly provisional. Lacerta part I is morphologically very like Lacerta part II and is separated partly because its members have a close and detailed resemblance to each other; the actual common features that divide the species of part I from part II are relatively slight (large body-size, dorsal pattern, no bright belly colour, distinctive type of ecological niche) and it seems probable that they are a relatively recent offshoot of Lacerta part II although it is not possible to suggest at present what part of this assemblage gave rise to them. Comparatively recent origin, or at least recent expansion, is also suggested by the close similarity of the species in Lacerta part I and their continuous distribution. Similarly Algyroides is also close to Lacerta part II, being differentiated only by large dorsal scales and to some extent by the very small size of the sulcal lips on the hemipenial lobes. This group appears to have most features in common with the more typical members of Lacerta part II. In the case of the genus Podarcis, homogeneity and continuous (and relatively small) range indicate recent origin or expansion. Most of the characters that in combination distinguish its members occur sporadically in Lacerta part II, more particularly in L. danfordi and its closer relations, viz. L. laevis, L. graeca and the southwestern species — L. andreanszkyi, L. dugesii and L. perspicillata. Among these characters are large apical sections to the hemipenial lobes (danfordi) and E. N. ARNOLD 356 —/+ DYO]]VL) SNWUMOAPOUMDS J (suoydaoxe may) (uondaoxa) - + - + (suoydaoxa) (suodaoxa) - + s[enptarput - aul0s + + + - SIIADPOT sapioaaa) (suoydaoxa) (suondaoxe) quasqe u1o}yed [esiop ut wistydiounp penxag 91qe} jeqotred Auoq JO a8pa yoRer yOu seop Prerys [eqzorsred jo ursieyy sjeseujsod pasodiadns om 7, ainzeusre youljsip qytM stuedruey uia}ed ye1qez190A Tepnes odAq-y Jaquinu [e1qa}I0A [Te1oes -aid ul uoTereA [enxas ssurlyyo}ey ur pasnyun sauoq [ejuo1s3S0d pur [e}1q10};sog + - SUIIO} }SOJ] (suorydaoxa) + ae saroads jo %06 07 %og (suoiy;daoxa) —/+ + saroads jo %og OUIOS - satoads Jo %0g 102AQ S[enprAipul = s[eNprAtpur aul0S 90S satoads yo %od 19AG “dds snjAjonpoyjuno a7 + Moy & 3da0xe ‘TLV (uoydeoxe) (uomdeoxe) + + satoads Jo %o9 19AQ II Hed J ued eigues IIUIEpua DIAIIVT DIABIDT o1ZoIeSRTe SAA SuIpnjoxs sepiyi0oeT Ul 99Uaprout ayeurxoiddy ainzea,y depiz1a0eT Ay} JO 4Ser yy ur peaidsaprM sainjzesy Jo uoINGIySIq : A] AIAVL RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 357 extremely well-developed sulcal lips (danfordi, andveanszkyi), hemipenial micro- ornamentation of simple spines (graeca and the southwestern species group), C-type caudal vertebrae (some danfordz), sternal fontanelle approaching cordiform shape (andreanszkyt, laevis, danfordi), parietal shield margin reaching outer edge of postorbital bone (southwestern species group), well-developed masseteric shield in some cases (andreanszkyt, laevis), Podarcis-type dorsal pattern (andreanszkyi). In spite of the above individual peculiarities, all these species are closer to other members of Lacerta part II than they are to Podarcis. Of the species cited above, L. andreanszkyi appears to have more in common with Podarcis than the others, and it may be closest to the stock that gave rise to the genus. This tentative hypothesis receives some support from the fact that one species of Podarcis, P. hispanica, can be quite similar in appearance to L. andreanszkyt, especially in the Atlas of Morocco where both of these forms occur. As stated (p. 346), there are reasons for believing that P. hispanica may be one of the more primitive members of Podarcis. Boulenger (1921) thought that Psammodromus might be related to Lacerta parva. Certainly, there is an external resemblance between the smaller species of Psammo- dromus and this form that extends to habitus, scale-shape and coloration, although this could be convergent as these lizards occupy similar habitats. L. parva resembles the small Psammodromus species in having a rather similar, peculiar pattern of hemipenial micro-ornamentation (Bohme 1971); also its caudal ver- tebrae approach the C type. However, there are a number of important differences between L. parva and Psammodromus ; the latter lacks the following features that occur in L. parva, viz. postorbitals and postfrontals unfused, double postnasals, reduced chromosomal formula and lack of plicae on hemipenial lobes. It appears unlikely therefore that there is a very close connexion between Psammodromus and L. parva although the former may well be derived from Lacerta part II. Gallotia is usually said to be closely related to Lacerta part I (e.g. Boulenger 1920, Peters 1961) but the two groups show many points of difference and the closest relatives of Gallotia appear to be the small species of Psammodromus. Although the more extreme forms of these two groups are superficially very different, they have many similarities especially in skeleton and hemipenis. The smallest species of Gallotia, G. atlantica, has large dorsal scales approaching those of Psammodromus in relative size. SUMMARY OF PROPOSED SYSTEMATIC CHANGES Algyroides Bibron & Bory This genus is limited to the four European species, viz. fitzingert (Wiegmann 1834), marchi Valverde 1958, moreoticus Bibron & Bory 1833 and migropunctatus (Duméril & Bibron 1839). The three African species originally placed in Algyrozdes are now placed in Adolfus Sternfeld 1912; these are africanus Boulenger 1906, allent Barbour 1914 and vauereselli (Tornier 1902). 358 E. N. ARNOLD Lacerta Linnaeus 1758 The central and southern African species, viz. australis Hewitt 1926, echinata Cope 1862, jacksont Boulenger 1899 and rupicola Fitzsimons 1933, are not con- sidered to be congeneric with the Palaearctic species of Lacerta. Their affinities lie with other Ethiopian lacertids and will be discussed more fully elsewhere. L. echinata and L. jacksoni are related to the Central African species placed in Adolfus, Bedriagaia and Gastropholis ; L. rupicola is perhaps related to Tropidosaura ; and the precise affinities of L. australis are unknown, as the type cannot be found. Palaearctic Lacerta, as usually understood (and including L. cappadocica Werner 1902) is divided into three main groups two of which are here raised to the status of independent genera. Species are allocated as follows. Podarcis Wagler 1830, including erhardii Bedriaga 1882, filfolensis Bedriaga 1876, hispanica Steindachner 1870, lilfordi (Gunther 1874), melisellensis Braun 1877, milensis Bedriaga 1882, muralis (Laurenti 1768), peloponnesiaca Bibron & Bory 1833, pityusensis Bosca 1883, sicula Rafinesque 1810, taurica Pallas 1814, tiliguerta Gmelin 1789, wagleriana Gistel 1868. Gallotia Boulenger 1916, including atlantica Peters & Doria 1882, galloti Duméril & Bibron 1839, stmonyt Steindachner 1889. The remaining species are left in Lacerta but this is divided into two parts: Lacerta part I and Lacerta part II. Lacerta part I contains the following species : agilis Linnaeus 1758, lepida Daudin 1802, princeps Blanford 1874, schreiberi Bedriaga 1878, strigata Eichwald 1831, trilineata Bedriaga 1886, viridis (Laurenti 1768). Lacerta part II contains all other species presently assigned to Lacerta. Psammodromus Fitzinger 1826 This genus is not changed and contains the four species algivus (Linnaeus) 1758, blanct (Lataste 1880), hispanicus Fitzinger 1826, microdactylus (Boettger 1881). ACKNOWLEDGEMENTS I should like to acknowledge the very considerable help given to me during the course of this study. The following either lent material or provided helpful comment: Dr W. Bohme (Zoologisches Forschungsinstitut, Bonn), Dr J. Eiselt (Naturhistorischesmuseum, Vienna), Dr U. F. Gruber (Zoologische Sammlung des Bayerischen Staates, Munich), Dr M. Hoogmoed (Rijksmuseum van Natuurlijke Historie, Leiden), Dr K. Klemmer (Senckenbergische Naturforschende Gesellschaft, Frankfurt a.M.), Dr A. E. Leviton (California Academy of Sciences, San Francisco), Dr G. Underwood (London) and Mr D. Western. Miss T. I. Molleson instructed me in the use of low-voltage X-ray equipment and Miss A. G. C. Grandison and Dr G. Underwood read the manuscript and made useful criticisms. Many of the observations on lacertid ecology and morphology were made when I was the recipient of a post-graduate research grant from the Scientific Research Council. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 359 REFERENCES ARNOLD, E. N. 1972. Lizards with northern affinities from the mountains of Oman. Zool. Meded. Leiden, 47 : 111-128. (in press). Structure and variation of the hemipenis in lacertid lizards (Reptilia: Lacertidae). Bargour, T. 1914. Some new reptiles. Pyoc. New Engl. zool. Club, 4: 95-08. BEDRIAGA, J. VON 1876. Die Faraglione Eidechse, p. 19. 1878. Herpetologische Studien. Avch. Naturgesch. 44(1) : 259-320. 1882. Die Amphibien und Reptilien Griechenlands (part). Bull. Soc. Nat. Moscou, 56(3) : 43—-103.! 1886. Beitrage zur Kenntnis der Lacertidenfamilie. Abh. senckenb. naturforsch. Ges. 14 : 17-443. Bisron, G. & Bory, J. B.G.M. 1833. Reptiles et poissons, in Bory, J. B. G. M., Expédition scientifique de Morée, 3. Paris. Birp, C.G. 1936. The distribution of reptiles and amphibians in Asiatic Turkey with notes on a collection from the Vilayets of Adana, Gaziantep and Malatya. Ann. Mag. nat. Hist. (10) 18 : 257-281. BLanForpD, W.T. 1874. Descriptions of new Reptilia and Amphibia from Persia and Baluchi- stan. Ann. Mag. nat. Hist. (4) 14: 31-34. BoeEttGER, O. 1881. Diagnoses reptilium novorum Maroccanorum. Zool. Anz. 4: 570-572. B6uME, W. 1971. Uber das Stachelepithel am Hemipenis lacertider Eidechsen und seine systematische Bedeutung. Z. Zool. Syst. EvolForsch. 9(3) : 187-223. BoscA, E. 1883. Exploracién herpetolégica de la Isla de Ibiza. An. Soc. esp. Hist. nat. 12 : 241-250. BouLencer, G. A. 1887. Catalogue of Lizavds in the British Museum (Natural History), vol. 3. London. xii+575 pp. 1899. Descriptions of two new lizards from the interior of British East Africa. Pvoc. zool. Soc. Lond. 1899 : 96-08. 1905. A contribution to our knowledge of the varieties of Wall Lizard (Lacerta muralis) in Western Europe and North Africa. Tvans. zool. Soc. Lond. 17(4) : 351-437. 1906. Additions to the herpetology of British East Africa. Proc. zool. Soc. Lond. 1906 : 57°. — 1907. Remarks on Prof. L. von Méhely’s paper Ann. Mag. nat. Hist. (7) 20: 39-46. 1908. Description of a new species of Laceyta from Persia. Proc. zool. Soc. Lond. 1908 : 934- 1910. Remarks on Prof. L. von Méhely’s recent contribution to the knowledge of the lizards allied to Lacerta muralis. Ann. Mag. nat. Hist. (8) 5 : 247-256. 1913. Second contribution to our knowledge of the varieties of Wall Lizard (Lacerta muralis). Trans. zool. Soc. Lond. 20(3) : 135-216. 1916. On the lizards allied to Lacerta agilis and Lacerta parva. Trans. zool. Soc. Lond. 21(1) : 1-90. 1920. Monograph of the Lacertidae, vol. 1. London. x+352 pp. 1921. Monograph of the Lacertidae, vol. 11. London. viii+451 pp. Braun, M. 1877. Lacerta lilfordi und Lacerta muralis zugleich ein Beitrag zur Reptilienfauna der kleinen Inseln des Mittelmeers. Ayb. zool.-zootom. Inst. Wurzburg, 4: 1-64. Bravo, T. 1953. Lacerta maxima n. sp. de la fauna continental extinguida en el Pleistoceno de las Islas Canarias. Estudios geol. Inst. Invest. geol. Lucas Mallada, 9 : 7-34. BRELIH, S. 1961. Seven new races of the species Lacerta (Podarcis) sicula Rafinesque from the Rovinj-Porec region. Biol. Vest. 9: 71-91. 1963. Ein Beitrag auf Kenntnis der Eidechsen aus dem Quarnero-Gebeit. Bzol. Vest. 11: 107-113. Bucuuorz, K. F. 1960. Zur Kenntnis von Lacerta peloponnesiaca (Reptilia: Lacertidae). Bonn. zool. Beitr. 11 : 87-107. Zur Losung der ‘ Muralis-Frage ’ ’’ 360 E. N. ARNOLD CAMERANO, L. 1885. Monografia dei sauri italiani. Zool. Anz. 8: 417-419. CrarK, R. J. & CrarK, E. D. 1970. Notes on four lizard species from the Peloponnese, Greece: Algyvoides morveoticus Bibron & Bory, Anguis fragilis peloponnesiacus Stepanek, Ophiomorus punctatissimus (Bibron & Bory) and Ophisaurus apodus (Pallas). Br. J. Herpet. 4: 135-137. Cope, E. D. 1862. On Laceyta echinata and Tiliqua duva. Proc. Acad. nat. Sct. Philad. 1862 : 189-1091. DareEvsky,I.S. 1957. Systematic and ecological aspects of the dispersal of the lizard Lacerta saxicola Eversmann in Armenia. Zool. Sb. Evevan, 10: 27-57. (In Russian.) 1966. Natural parthenogenesis in a polymorphic group of Caucasian rock lizards related to Lacerta saxicola Eversmann. J. Ohio herpet. Soc. 5(4) : 115-152. —— 1967. The Rock Lizards of the Caucasus: Systematics, Ecology and Phylogeny of a Poly- morphic Group of Caucasian Lizards of the Subgenus Archaeolacerta. Leningrad. 214 pp. (In Russian.) Daupin, F.M. 1802. Histoive Naturelle des Reptiles, vol. 3. Paris. 452 pp. DOUMERGUE, F. 1901. Essai sur la Faune Erpetologique de l’Ovanie. Oran. 404 pp. Dow inc, H. G. & SavaceE, J. M. 1960. A guide to the snake hemipenis: a survey of basic structure and systematic characteristics. Zoologica, N.Y. 45: 1-28. DroEpov, N.N. 1964. Biology of the green-bellied lizard (Lacerta chlorogaster), in TERENT’EV, P. V., Voprosy Gerpetologit. Leningrad. 84 pp. (In Russian.) DumErIL, A. M. C. & Brpron, G. 1839. Erpetologie Générale ou Histoive Naturelle Complete des Reptiles. Paris. viii+854 pp. EICHWALD, L. E. von. 1831. Zool. spec. Ross. Polon. 3: 189. Eimer, T. 1881. Untersuchungen iiber das Variiren der Mauereidechse, ein Beitrag zur Theorie von der Entwicklung aus constitutionellen Ursachen, sowie zum Darwinismus. Arch. Naturgesch. 47(1) : 239-517. EIsELT, J. 1968. Ein Beitrag zur Taxonomie der Zagros Eidechse, Lacerta princeps Blanf. Annin naturh. Mus. Wien, 72 : 409-434. —— 1969. Zweiter Beitrag zur Taxonomie der Zagros-Eidechse, Lacerta princeps Blanford. Annin naturh. Mus. Wien, 73 : 209-220. EVERSMANN, E. 1834. Lacertae imperii Rossici variis in itineribus meis observatae. Mém. Soc, nat. Moscou, (2) 3 : 337-369. ETHERIDGE, R. 1967. Lizard caudal vertebrae. Copeia, 1967 : 699-721. Firiprr, F. bE. 1863. Nuove o poco note specie di animali vertebrati raccolte in un viaggio in Persia nell’estate dell’anno 1862. Aycho Zool. Anat. Fisiol. 2 : 377-394. Firzincer, L. J. F. J. 1826. Neue Classification der Reptilien nach thren natiivlichen Ver- wandtschaften. Vienna. vili+66 pp. 1843. Systema Reptilium: fasciculus primus, Amblyglossae. Vienna. 106 pp. Fitzsimons, V. F. 1933. Descriptions of five new lizards from the Transvaal and Southern Rhodesia. Ann. Transv. Mus. 15(2) : 2, 3-280. GIsTEL, J. 1868. Die Lurche Europa’s, in Blicke in das Leben dey Natur. Leipzig. GMELIN, J. F. 1789. In Linnaeus, C., Systema naturae, ed. 13, vol. I: 100. GorMAN, G. 1969. New chromosome data for 12 species of lacertid lizards. Journal of Herpetology, 3: 49-54. —— Jovanovic, V. & McCottum, F. 1971. Conservative karyotypes among lizards of the genus Lacerta from the Adriatic Islands. Genetika Beograd, 2 (2) : 149-154. Gray, J. E. 1838. Catalogue of the slender-tongued saurians, with descriptions of many new genera and species. Ann. Mag. nat. Hist. 1: 274-283. GUNTHER, A. 1874. Description of a new European species of Zootoca. Ann. Mag. nat. Hist. (4) 14: 158-159. —— 1876. Description of a new species of lizard from Asia Minor. Proc. zool. Soc. Lond. 1876 : 818. Hewitt, J. 1926. Some new or little-known reptiles and batrachians from South Africa. Ann. S. Afr. Mus. 20(6) : 473-490. RELATIONSHIPS OF LACERTA, ALGYROIDES AND PSAMMODROMUS 3061 Jacquin, J. F. 1787. Lacerta vivipava, observatio Jos. Francisci de Jacquin. Nova Acta helv. 1: 33. KLEMMER, K. 1957. Untersuchungen zur Osteologie und Taxionomie der europdischen Mauereidechsen. FULL GROWN : ; i : BODY I- BECOMING apuLT+———___51-------- WING P. TRICARUNCULATA Ist 2ND 3RD 4TH CROWN GREEN +8ECOMING SOOTY—_—> / BECOMING WHITE> : ‘ THROAT +BECOMING BARE——> : BARE : : \ BODY JUVENILE {BECOMING ADULT ———;—————>! : : 1 : : WING . JUV — ADULT . . M P. NUDICOLLIS Fic. 5. Generalized time-table of acquisition ot adult plumage by males of Procnias tricarunculata and P. nudicollis. Body plumage: the broken line represents the period within which fully adult plumage is acquired (individually variable). Wing: the lines labelled M indicate periods of moult of the flight-feathers. they are a little under a year old. At the first complete moult the wing and tail are replaced again by juvenile feathers ; the head and body feathers are replaced, vari- ably and perhaps more gradually than the wing and tail, by feathers of more adult type. In some birds, the last wing and tail feathers to be replaced also show a EVOLUTION OF THE BELLBIRDS (PROCNIAS) 381 tendency towards adult colouration. When the bird is a little under two years old another complete moult leads to the assumption of wing and tail feathers of adult or nearly adult type, and the head and body also become nearly adult in colour. In some individuals completely adult body plumage is acquired without a further wing moult, presumably by gradual replacement ; in others, the fully adult plumage is not acquired until after the third complete moult. The replacement of odd wing and tail feathers between these complete moults may result in irregularities, with a few feathers of more adult type than the rest. In P. tricarunculata, the wattles begin to grow when the male is between six months and a year old (assuming hatching in June), and they evidently grow fast, those at the gape much faster than the central one, since by the end of the first year they may be up to 20mm long. They do not then grow much more during the following year. In P. nudicollis the male’s throat becomes bare in the course of the first year. It is difficult to investigate the body moults which do not involve a colour change, but there is probably a complete body moult in the first few months of life. In P. nudicollis the crown becomes sooty-black in this period, to be replaced by white in the first months of the second year. In P. averano and P. alba the less well-synchronized annual cycle (which may also differ in different parts of the species’ ranges, according to the regime of wet and dry seasons) and more frequent irregularities of moult sequence make a reconstruction of plumage changes more difficult ; but the observed changes in plumage in a male P. averano in Trinidad, of known age (Snow 1970), provide a useful time-scale for part of the juvenile period. It seems that the plumage sequence and timing are very much the same as in P. tricarunculata and P. nudicollis. There is a succession of full moults, apparently starting when the bird is nearly a year old. After the first, the wing and tail are still of juvenile type, after the second they are sub-adult, and after the third fully adult in colour. In P. averano an early moult, probably of the whole head and body, leads to the assumption of a sooty crown, and this is replaced by the brown crown of the adult during the second year. After the first full moult the back also becomes sooty in colour rather than green, and later often grey-brown, before the very pale grey adult colour is acquired. The change is variable and probably depends on the relative timing of hormonal changes and of moults. In P. averano the pale juvenile throat feathers are replaced by sooty feathers at about the same time as the sooty crown is acquired. The wattles probably develop at the age of about a year ; in the Trinidad juvenile they were just visible, in the field, at the age of 16 months. In the only specimen of P. alba examined at the appropriate stage, the wattle was present but short when the bird was half-way through its first wing moult, presumably at an age of just over a year. THE WATTLES The age at which the wattles first appear has just been mentioned. In all three of the species with wattles, they begin to grow in the young male while the plumage is still wholly juvenile, and are a good size, though probably not full-grown, by the time that the adult plumage is complete. The wattles are bare, shiny and blackish 382 D. W. SNOW in P. tricarunculata and P. averano ; in P. alba the single wattle is black and is sparsely covered with small white star-like feathers. The bare skin of the throat of P. nudicollis is greenish ; it is not completely bare, but is sparsely covered with black bristles. Similar bristles grow between the wattles on the throat of P. averano. The histological structure of the wattles has not been investigated. Schomburgk (1848) asserted that the wattle of P. alba is hollow and muscular, and in direct com- munication with the palate, and that it can be inflated at will. Waterton (1824) had earlier said much the same thing. As a result the older illustrations of P. alba usually show it with the wattle standing straight up like a unicorn’s horn. From close observation of a captive bird, Quelch (1892) refuted Schomburgk’s assertion, describing the wattle as made up of fine elastic tissue. Little can be added so far as the histology goes, except that the wattle must also contain muscle fibres, as Salvin (1865) suggested for P. tricarunculata. The first good account of the wattles in life was given by Quelch. Writing of a male P. alba in captivity in British Guiana (Guyana), he wrote: ‘When the bird is about to utter its characteristic notes, this appendage slowly becomes greatly elon- gated — to as much as five inches, I have observed at times. At the conclusion of the note, the organ may remain extended till the next note, or it may be partially retracted ; but when a long interval takes place, the structure is always allowed to shrink up to about half an inch or an inch in length, at will ; and it then hangs against the beak.’ Similarly Snow (1961) observed that when a male P. alba, observed in the wild, had finished a bout of calling the wattle contracted to about a third of its former length before the bird flew off to feed. Crandall (1948) also reported that in a captive P. tricarunculata the wattles enlarged to about three times their usual dimensions when the bird was calling. The mass of wattles on the throat of P. averano does not alter so noticeably when it is calling ; but they are made more conspicuous by the erection of the feathers of the upper breast, which push the bird’s black ‘beard’ forwards (Snow 1970). It is probable, however, that in this species too the wattles can be extended by muscular action. Thus W. G. Conway (in litt.) noted that in a captive P. averano in the Bronx Zoo, New York, the wattles appeared shrivelled up when it stopped calling and began to moult. It seems clear that the display of the wattles, in all three species, is closely con- nected with calling, and there is no doubt that the main function of all the male’s displays is to attract females to the calling perch, on which mating takes place (Snow 1970). Presumably the wattles, in combination with the striking plumage and associated behaviour, play an important part when the female is at close quarters ; certainly in P. averano the male’s performance appears to rivet the female’s attention. Crandall (loc. czt.) reported that a female P. trvicarunculata in captivity appeared to be attracted by the dangling wattles of a male and at least once reached forwards as if to peck at them. This is almost certainly not normal behaviour, since in P. averano, and probably also in P. alba (Snow, in press), the male and female stay two or three feet apart during courtship, until the male leaps, with an explosive call, onto the female. Nevertheless, the wattles may perhaps be regarded as structures analogous to the dangling silky fringes of the modified secondary feathers of the Cock EVOLUTION OF THE BELLBIRDS (PROCNIAS) 383 of the Rock Rupicola rupicola, which attracted a courting female to lean forward and nibble at them (Snow 1971). The evolution of extensible wattles as a result of sexual selection has brought with it a concomitant disadvantage, though probably not a serious one, for the males of the two species whose wattles grow from the base of the beak. Crandall mentioned that the long wattles of a captive P. tricarunculata were a positive nuisance to the bird when it wanted to feed ; and Snow (1961) noted that a male P. alba, when calling, had to ensure that its wattle was hanging down on the right side before it could begin one of its calls, which involved swinging the head to the left. VOICE Male bellbirds utter what are perhaps the loudest of all bird calls; the females and nestlings, at least of P. averano, are almost or completely silent (Snow 1970). The main characteristics of the calls of the four species are described and illustrated in the following sections, which are followed by a brief discussion of functional, mechanical and evolutionary aspects of the calls. Procnias averano Males of the Venezuelan population have three distinct calls: a single, very loud bock (Plate 1(A)) ; a repeated, less loud tock, tock, tock .. . (Plate 1(B)) ; anda more musical, double kay-kong, kay-kong, . . ., repeated more slowly than the fock, tock, tock (Plate 1(C)). The single bock is remarkable for the rather even distribution of energy over a wide frequency range, with the emphasis on the second harmonic, and for the very sudden beginning of the note. These features combine to give the note its sudden, unmusical hammer-like quality. The repeated focks are very similar in form to the bock, but only two instead of six harmonics above the fundamental are apparent in the sonogram. The double kay-kong is of more complex form: the main part of the first syllable is very similar in form to the tock but more drawn out, while the second syllable has almost all its energy concentrated at the fundamental frequency, thus producing a more musical and lower-pitched note. In the first syllable, the harmonics begin a little after the fundamental frequency is sounded. Both syllables begin with a preliminary pulse of energy whose upper frequency is intermediate between the frequency of the second and third harmonics. Exactly the same preliminary pulse is apparent at the beginning of the tock, where it is less distinct in time from the main part of the note, while in the bock it is, as it were, attached to the main part of the note, so that there is a sudden descent in pitch which is most marked in the higher harmonics. The bock of the Trinidad bird (Plate 1(D)) is even more remarkable for the even distribution of energy over a wide frequency range, and for its short duration, which gives it an even duller, more hammer-like quality than the bock of the Venezuelan bird. The repeated tocks (Plate 1(E)) are simply less loud versions of the bock. The refinements of the beginning of the call, noted for the Venezuelan bird, are not present. 29 384 D. W. SNOW Males of the Trinidad population now use only these two calls, the bock and the repeated tock, but, as Snow (1970) has pointed out, there is good evidence that a third call was being used by the Trinidad population in 1893, when Brewster and Chapman visited the island (Brewster & Chapman 1895). This third call was disyllabic and may have been similar to the kay-kong of the Venezuelan population (but see footnote to p. 388). The implications of this loss of part of the repertoire are discussed in a later section. Snethlage (1928) has given the only account of the calls of the population of P. averano in northeastern Brazil. He wrote: ‘Zwar verfiigt er nur tiber zwei Tone : einen schweren und einen leichteren Schlag, die den Hammerschlagen auf Eisen gleichen. Ein schwerer Schlag, kleine Pause, dann regelmassige leichte, das ist der Gesang, der zur Zeit der ersten Regen (im August-September) weithin zu vernehmen ist.’ (They use only two notes: a loud and a softer note, which sound like the blows of a hammer on iron. A loud blow, a short pause, and then regularly repeated and less loud notes — that is the song, which is to be heard from afar at the beginning of the rains (August-September).) Before uttering the explosive bock the bird opens its beak very wide, and the call is given when the beak is fully open (Snow 1970). Brewster & Chapman described the calling bird as throwing its head forward and downwards witha violent, convulsive jerk, but this is certainly not normal behaviour. When the repeated ¢ocks are being uttered the beak remains open, the lower mandible merely moving down and then upward again with each note. Procnias nudicollis Two calls are uttered : a single bock (Plate 2(A)) and a repeated call (Plate 2(B)). The bock is extremely similar to that of P. averano; the repeated call is probably homologous with the repeated tock of P. averano, but it is of most peculiar structure : a less loud but otherwise apparently identical version of the bock is followed, about 0-08 second later, by a very loud and pure high-pitched note, exactly an octave above the fundamental frequency of the first note. The two very different and partly overlapping notes give the impression of a hammer striking an anvil a glancing blow so that it rings. Greenewalt (1968) presented an oscillogram (time-amplitude diagram) of this call, together with a time-frequency plot, giving it as an example of a call in which two independent sound sources must be involved. This aspect of the call is discussed more fully in a later section (p. 387). When preparing to utter the explosive bock, the bird is described as opening its mouth very wide, perhaps inhaling ; it then bows forward as the bock is uttered. The continuing, repeated notes are accompanied by coordinated nodding or slight bowing (manuscript notes by R. Ward, who made the recordings used for the sonograms). Procnias alba Of the four bellbirds, this species has by far the most musical voice, and is the one which gives the genus its popular name. There are two calls: a double kong-kay EVOLUTION OF THE BELLBIRDS (PROCNIAS) 385 (Plate 2(C)) and a long drawn-out do-1-1-1-ing (Plate 2(D)). This second note, attenuated to a silvery chime as it filters down through the forest trees, is among the most beautiful sounds of the Guiana forests. The two notes of the kong-kay are similar in structure, but the second is a little higher pitched than the first (fundamental frequencies of about 1-8 and 1-7 kHz respectively). Though it is difficult to be certain of this, it seems from detailed inspection of the sonograms that the two overlap in time, and thus are presumably produced by two independent sound sources. The utterance of the double kong-kay is sometimes accompanied by a characteristic movement (Snow 1961). The bird turns sharply to the right to utter the first syllable ; then with beak still wide open it rotates the body rapidly through an arc of about 100 degrees, to make the second note facing to the left. The wattle nor- mally hangs down to the right of the beak, so that it flies out horizontally as the bird swings to the left. If it happens to be hanging down on the left side, the bird manceuvres it to the right side before calling. The drawn-out do-i-1-i1-ing is a pure note of very even pitch, with a single harmonic. It is usually more or less clearly broken into two notes (Plate 2(D)), the second syllable being shorter and dying away with a slight tremolo. Because it tends to be disyllabic, it has been written as do-vong (Quelch 1892, Snow 1961). Quelch’s detailed description of the calling of a captive male of P. alba in Guyana is worth quoting in full, since the way in which the air is inhaled and the movements associated with the calls are relevant to the problem of the two sound sources, discussed in a later section (p. 387). ‘When the appendage is fully elongated [see earlier, p. 382, for distension of the wattle], the bird suddenly inflates its lungs, right and left, by inhaling — almost by a swallowing action — two great draughts of air; but the method by which this is done depends upon which of its two charac- teristic notes it intends to utter. When the notes “Kong-Kay” are uttered, the action of inflation has been performed by two distinct inhalations of air, one with the head turned to the right and the other immediately after to the left. At the moment of utterance of the notes, the head is turned to the right for the “‘ Kong’’, and then suddenly —so suddenly that it almost startles the observer —the head is swung round to the left for the ‘‘ Kay”’, which is issued with a strikingly loud, piercing and metallic ring or clang — so loud and shrill indeed that, if the observer is close by, the ears are actually deafened for the moment by the sharpness of the sound. ‘When, however, the sweet, musical, and deeply-toned bell-like notes ‘‘ Do-rong”’ are about to be uttered, the bird is observed simply to hold its head forward, and to make two distinct gulps of air; and then, holding its beak upwards and slightly extending its neck, the notes are rolled out, as it were, with full voice and roundness and resonance.’ Procnias tricarunculata Of the four bellbirds, this species has the least highly developed repertoire. Skutch (1969) has given the fullest account of its voice. There are apparently only two dis- tinct notes, one ‘loud and strong but dull and throaty’ (buck), the other much sharper and higher in pitch (wheat). The sonograms show less difference between these two 386 D. W. SNOW notes than might be expected from Skutch’s description (Plate 2(E)), but they are in fact quite distinct. The buck has much of its energy centred, rather diffusely, round frequencies of about 1-75 and 3-5 kHz, with some energy at very low frequen- cies. This frequency distribution presumably accounts for the dullness of the tone, which Skutch likens to a wooden clapper, while the roughness of the sound probably comes from the irregular pattern of frequencies at the beginning of the note. The wheat, on the other hand, has its main frequency band about 0-3 kHz higher than that of the buck, has no very low frequencies and shows an orderly series of diminishing harmonics.* It is clear from Skutch’s account that, although the two notes are commonly uttered in succession, to form the phrase buck wheat, they are also given in every possible combination. The significance of this lack of organization in calling behaviour is discussed in a later section (p. 388). FUNCTION OF THE CALLS The epigamic function of these very loud calls has already been briefly mentioned (p. 369). There is no doubt that they are advertising calls, and one of the means by which males attract females to their display perches. Why such a very loud call should have evolved in this genus is not clear ; but it should be noted that many forest birds, in which the males display at traditional sites, have loud advertising calls. In the Cotingidae very loud calls, though of a different kind from those of the bellbirds, are uttered by males of other large species, in the genera Tijuca, Perisso- cephalus, Pyroderus, Cephalopterus and Lipaugus. Males of other genera are less noisy, but these advertise themselves from a distance by making themselves visually conspicuous. The bellbirds are peculiar in having two extremely effective methods of making themselves conspicuous from a distance. A mainly white bird calling very loudly from an exposed tree-top — a usual calling perch - could hardly advertise itself more effectively. At least in Procnias averano, the later stages of courtship display are carried out beneath the canopy of the forest, on comparatively low perches, and are silent except for the loud bock which accompanies the mating leap. One of the ways in which vertebrates can locate sounds is by comparison of the time of arrival of the sound at the two ears. This is possible only if the sound begins or ceases abruptly (Marler 1955). The extraordinarily sharp beginnings of the bell- birds’ calls should therefore make them especially effective in this respect, and in fact a bellbird calling at a distance can be accurately located by the human ear, as long as one is in the open. It is, however, rather difficult to locate a calling bellbird if one is under the forest canopy, probably because the sound is reflected with varying degrees of strength from the trees all round. MECHANISM The vocal mechanism which enables bellbirds to produce explosively loud calls of varied structure and complexity has not been fully elucidated. Greenewalt * In an earlier account Ridgway (1905) described the call as ‘like a heavy stroke with a hard mallet on a hollow log of hard wood followed immediately by a wonderfully loud, clear and prolonged whistle’. This is endorsed by Slud (1964) ; but the onlyavailable recording matches Skutch’s description closely, since the wheat is not prolonged. Perhaps there is local or even individual variation in the calls of this species. EVOLUTION OF THE BELLBIRDS (PROCNIAS) 387 (1968) has discussed the physical basis of sound production by the avian syrinx and gives the reasons for believing that the sounds produced depend on the vibration of syringeal membranes alone, and that little or no importance is attributable to the resonance of the trachea, or to modulation of the sounds in the larynx. Greenewalt concluded from the anatomical evidence, and from early experimental evidence, that the only membranes actively involved in sound production were the two internal tympaniform membranes. Thorpe & White (1969), however, have pointed out that the external tympaniform membranes are also involved, at least in some species. It has been known for a very long time (Herissant 1753) that the production of sound by the syrinx depends on air pressure in the interclavicular air sacs which surround the syrinx, and Rtippell (1933) confirmed this with experiments on an isolated syrinx. Greenewalt took the theoretical analysis much further, and showed that ‘high amplitude, for phrases rich in harmonic content, must be associated with increased pressure in the clavicular sac’. Since the calls of bellbirds are both rich in harmonics and extremely loud, it seems certain that they must build up an unusually high air pressure in their interclavicular air sacs. Discussing the relationship between pulse length and harmonic content, Greenewalt further pointed out that ‘the more nearly one approaches a simple pulse the more terms one will find in the harmonic spectrum. The extreme case (a sharp pulse of very short duration) will produce a harmonic spectrum containing an infinite number of terms at constant amplitudes.’ The single very loud calls of bellbirds perhaps illustrate this principle more clearly than any other bird calls. Both P. alba and P. nudicollis utter calls in which two separate sound sources appear to be involved. The existence of two separate sound sources in birds’ calls has been dealt with at length by Greenewalt, who considered that the right and left internal tympaniform membranes are responsible, acting independently under the influences of their separate musculature and innervation. This interpretation has been questioned by Thorpe & White (loc. cit.). The description of the calls of the four bellbird species, given above, can add nothing to this highly technical contro- versy. The behaviour associated with the double call of P. alba does, however, suggest the possibility that the production of sound by two vibrating membranes may be linked with the exhalation of air from the two bronchi successively. In the literature on the working of the avian syrinx there appears not to have been any serious suggestion of independent operation of the two lungs and bronchi ; but this would seem to be mechanically more efficient than their obligatory simultaneous operation in the case of sound produced independently by the two internal tympani- form membranes, since the flow of air from the right bronchus, for instance, would be largely wasted if the left membrane alone were being activated. Quelch did not suggest it, but his account of the calling of P. alba, quoted above, suggests the pos- sibility that the turning of the head to right and left might be linked with the asymmetrical functioning of the bronchi and lungs. EVOLUTION OF THE CALLS The characteristics of the calls which suggest the phylogenetic division of Procnias into two pairs of species have been briefly mentioned on pp. 371-372. P. averano 388 D. W. SNOW and P. nudicollis both have a single very loud monosyllabic call, the bock, of very similar structure in each. Both also utter a less loud, regularly repeated call, in structure very similar to the bock, except that in P. nudicollis there is a second, high- pitched element in the call, which is missing in P. averano. Both P. alba and P. tricarunculata utter a disyllabic call, of which the second note is higher-pitched than the first, but there the resemblance ends ; the two parts of the call of P. tricarunculata are not so closely coordinated as they are in P. alba, and the quality of the calls differs in the two species. Snow (1970) described how a male P. avervano in Trinidad, colour-ringed as a nestling, had still not perfected its call when it was over two years old. It seems probable that a slow learning process is involved in the achievement of the fully adult call by young birds, and it may well be that the presence of calling adults is necessary. If this is so, some elements of the full vocal repertoire might well be lost in a very small, isolated population if for some reasons the numbers sank very low and included few or no old males. Snow (loc. cit.) suggested that this may account for the loss, in the Trinidad population of P. averano, of an element present in the vocal repertoire of the Venezuelan birds.* In the case of long-distance dispersal, leading to the establishment of a new isolated population, this is even more likely to be accompanied by a reduction of the vocal repertoire, since dispersal most commonly involves young birds. This could account for the lack of differentiation in the calls of P. tricarunculata, which on quite different grounds seems likely to have arisen as a result of long-distance dispersal from some area far to the east, perhaps the Guiana highlands (Figure 4). The throaty quality of the buck, and the varied sequence of buck and wheat described by Skutch, are rather reminiscent of the harsh and irregular calls of young males of P. averano. Subsequent observations of P. alba in Guyana (Snow, in press) have shown that the calls of young males are very similar to those of P. averano. It may well be that in all four species the calls pass through very similar developmental stages before they are crystallized under the influence of the calls of fully adult birds. SUMMARY The genus Procnias is one of the many isolated and peculiar genera in the neo- tropical family Cotingidae. It represents the culmination of the tendency towards specialization for fruit-eating characteristic of the family, associated with marked sexual dimorphism and the emancipation of the male from the nest. * Brewster & Chapman’s description (1895) of the voice of P. averano in Trinidad raises some interesting questions which will never be fully answered. In addition to the calls that are easily identifiable as the same as those given by Trinidad birds today, described by Snow (1970), these two authors describe a double ¢wi-ting, which certainly suggests identity with the disyllabic call of the Venezuelan population (p. 383). There were, however, some clear differences in the Trinidad call, if Brewster & Chapman's description is to be relied on. ‘Each ¢ui is followed closely by a metallic ting which sounds exactly like an echo and appears to be of about the same duration and nearly as loud as the note it supplements. The tui notes are given so quickly that at first it did not seem possible for the bird to produce another note between them, and it was only after repeated observations we became convinced that the ting was an integral part of the ¢ui call.’ This is not the impression given by the distinctly disyllabic and well- spaced kay-kong of the Venezuelan bird (Plate 1(C)); indeed it seems much more like the repeated call of P. nudicollis (Plate 2(B)). One can only conclude that the vocal repertoire of isolated populations of bellbirds is labile, and that new elements may be gained as well as lost. EVOLUTION OF THE BELLBIRDS (PROCNIAS) 389 The four species, which are almost entirely allopatric, fall into two closely related pairs, P. averano and P. nudicollis, and P. alba and P. tricarunculata. Their present distribution suggests that the ancestral stock inhabited northeastern South America, that it first split into an eastern and a western form, and that each of these subse- quently split into two species. It is probable that the bellbirds’ general body size, and especially the size of the gape, are related to the sizes of fruits of trees in'the family Lauraceae, which provide a great part of their food. The acquisition of fully adult plumage by the males takes about three years. In P. averano at least, the adult male’s calls are not perfected until the third year of life. The mechanism by which the extremely loud calls are uttered is not fully understood, but there is evidence that two independent sound sources are involved. There is some evidence that learning is important in the acquisition of the full vocal repertoire. The isolated Trinidad population of P. averano lacks one of the calls of the Venezuelan population. The lack of clear differentiation in the calls of P. tricarunculata may have resulted from a break in learning, consequent on the pre- sumed long-range colonization of Central America by the ancestors of this species. ACKNOWLEDGEMENTS I am most grateful to the following museum curators for allowing me to examine the specimens of bellbirds in their care : Dr D. Amadon (American Museum of Natural History, New York), Dr H. F. de A. Camargo (Museu de Zoologia, Sao Paulo), Dr J. Dorst (Muséum National d’Histoire Naturelle, Paris), Dr F. Gill (Academy of Natural Sciences, Philadelphia), Dr F. C. Novaes (Museu Goeldi, Belém), Dr K. C. Parkes (Carnegie Museum, Pittsburgh), Mr W. H. Phelps (Coleccién Phelps, Caracas), Dr C. G. Sibley (Peabody Museum of Natural History, New Haven), Dr H. Sick (Museu Nacional, Rio de Janeiro), Dr G. E. Watson (National Museum of Natural History, Washington). Visits to the museums above mentioned, and field work on bellbirds in Brazil and Guyana, were supported by grants from the Royal Society and the Frank M. Chapman Memorial Fund of the American Museum of Natural History. Mr Paul Sewartz, Mr Jan Lindblad and the Cornell University Library of Natural Sounds kindly made available the tape recordings from which the sonograms on Plate x (A-C) and Plate 2 were made. I am much indebted to Mrs J. Hall-Craggs of the Sub-department of Animal Behaviour, Madingley, Cambridge, for making the sonograms. Dr H. F. de A. Camargo, Dr W. G. Conway, Dr R. Meyer de Schauen- see, Dr L. L. Short, Mrs E. H. Stickney and Dr C. Vaurie kindly supplied information on specimens. REFERENCES ALLEN, C. K. 1948. Flora of Panama. Lauraceae. Ann. Mo. bot. Gdn, 35 (1) : 1-68. Ames, P.L. 1971. The morphology of the syrinx in passerine birds. Bull. Peabody Mus. nat. Hist., 37. 194 pp. ‘ BeEarp, J.S. 1946. The natural vegetation of Trinidad. Oxf. For. Mem., 20. BREwsTER, W. & CHAPMAN, F. M. 1895. Notes on birds observed in Trinidad. Auk, 12: 201- 2it. 390 D. W. SNOW BurMEISTER, H. 1856. Systematische Uebersicht dey Thiere Brasiliens, Vol. 2. Georg Reimer, Berlin. x + 466 pp. Camarco, E. A. DE 1957. Resultados ornitolégicos de uma excursao ao Estado do Maranhao. Papéis Dep. Zool. S. Paulo, 13 : 75-84. CRANDALL, L.S. 1948. Notes on the display of the Three-wattled Bellbird (Pyocnias tricarun- culata). Zoologica, N.Y., 33: 113-114. Euter,C. 1867. Beitrage zur Naturgeschichte der Végel Brasiliens. J. Orn. Lpz., 15: 177-108. Goopwin, D. 1970. Pigeons and Doves of the World. British Museum (Natural History), London. 446 pp. GREENEWALT, C. H. 1968. Bird Song : Acoustics and Physiology. Smithsonian Institution, Washington. 194 pp. HAvERSCHMIDT, F. 1955. List of the birds of Surinam. Ujitg. natuurw. StudKving Suriname, No. 13, 153 pp- 1968. Birds of Surinam. Oliver & Boyd, Edinburgh and London. xxix + 445 pp. Hetimayr, C.E. 1929. Catalogue of birds of the Americas. Part VI. Publs Field Mus. nat. Hist., Zool. Ser., 13 : 1-258. Kopp, L. E. 1966. A taxonomic revision of the genus Persea in the Western Hemisphere (Persea-Lauraceae). Mem. N.Y. bot. Gdn, 14 (1). 117 pp. Laupmann, A. 1940. Die Vogel von Paraguay, Band 2. Strecker, Stuttgart. 228 pp. LeMEE, A. 1952-56. Flove de Guyane francaise. Lechevalier, Paris. Marter, P. 1955. Characteristics of some animal calls. Nature, Lond. 176 : 6-7. MEYER DE SCHAUENSEE, R. 1950. The birds of the Republic of Colombia. Caldasia, 5 : 751- 775 (Cotingidae). Monroe, B.L. 1968. A distributional survey of the birds of Honduras. Ovn. Monogyr., No. 7. American Ornithologists’ Union. 458 pp. Merz, C. 1889. Lauraceae Americanae. Jb. bot. Gart. Berl., 5: 1-566. Otson, S. L. 1971. Taxonomic comments of the Eurylaimidae. Ibis, 113 : 507-516. PELZELN, A. voN 1868. Zur Ornithologie Braziliens. A. Pichler’s Witwe & Sohn, Vienna. 462 + lix pp. Puerps, W. H. & PuHetps, W. H. Jr. 1962. Cuarentinueve aves nuevas para la avifauna brasilefia de Cerro Uei-tepui (Cerro del Sol). Bol. Soc. venez. Cienc. Nat., 23 : 32-39. 1963. Lista de las aves de Venezuela con su distribucién, Tom. 1, Pt. 2. Bol. Soc. venez. Cienc. nat., 24: 5-479. Pinto, O. M. DE O. 1944. Catalogo das aves do Brasil, Pt. 2. Sao Paulo, Brazil. 700 pp. 1954. Resultados ornitolégicos de duas viagens cientificas ao Estado de Alagoas. Papéis Dep. Zool. S. Paulo, 12: 1-97. Quetcn, J. J. 1892. Position of the caruncle and method of song of the Bell-Bird. Timehri, N.S., 6: 164-166. Rimpeway, R. 1905. A winter with the birds in Costa Rica. Condor, 7 : 151-160. RUPPELL, W. 1933. Physiologie und Akustik der Vogelstimme. J. Orn. Lpz., 81: 433-542. Satvin, O. 1865. Notes on the Costa-Rican bell-bird. Jbis, (2) 1: 90-95. ScHOMBURGK, R. 1847. Reisenin Britisch-Guiana, Vol.1. J.J. Weber, Leipzig. x + 469 pp. Sxutcu, A. F. 1969. Life histories of Central American birds III. Pacific Coast Avifauna, No. 35. Berkeley, California. 580 pp. Stub, P. 1964. The birds of Costa Rica. Bull. Am. Mus. nat. Hist., 128 : 1-430. SNETHLAGE, H. 1928. Meine Reise durch Nordostbrasilien. II. Biologische Beobachtungen. J. Orn. Lpz., 76 : 503-581. Snow, B. K. 1961. Notes on the behaviour of three Cotingidae. Auk, 78 : 150-161. 1970. A field study of the Bearded Bellbird in Trinidad. Jbis, 112 : 299-329. Snow, D. W. 1965. A possible selective factor in the evolution of fruiting seasons in tropical forest. Oikos, 15: 274-281. 1971. Evolutionary aspects of fruit-eating by birds. Ibis, 113 : 194-202. Snow, D. W. & Snow, B. K. 1964. Breeding seasons and annual cycles of Trinidad land-birds. Zoologica, N.Y., 49 : 1-35. EVOLUTION OF THE BELLBIRDS (PROCNIAS) SNYDER, D.E. 1966. The Birds of Guyana. Peabody Museum, Salem, Mass. Tuorpe, W. H. & WuitE, R. E. C. 111 : 629-631. 391 308 pp. 1969. Special review — the mechanics of bird song. Ibis, Waterton, C. 1824. Wanderings in South America. London. Dr D. W. Snow Sub-Department of Ornithology _ British Museum (NaTuRAL History) TRING : HERTFORDSHIRE 30 PLATE 1 Sonograms of calls of Pyocnias averano. HOOWP> Venezuela. The bock. Venezuela. The repeated fock. Venezuela. The repeated kay-kong. Trinidad. The bock. Trinidad. The repeated tock. I PLATE Bull. Br. Mus. nat. Hist. (Zool.) 29, 5 wilt uit CC TENT TLL nt HH HE: {\ SONOD3S NI 3WiL WHbH un f | nue i) OD wee MAYA woe i LT it ia {iy PLATE 2 Sonograms of calls of Procnias nudicollis, P. alba and P. tricarvunculata. BOOP tu ty tot . nudicollis. The bock. . nudicollis. The repeated call. . alba. The kong-kay. . alba. The do-i-i-1-ing. . tricarunculata. The buck wheat. Bull. By. Mus. nat. Hist. (Zool.) 29, 5 (0) ] TIME IN SECONDS PLATE A LIST OF SUPPLEMENTS TO THE ZOOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) Kay, E. ALison. Marine Molluscs in the Cuming Collection British Museum (Natural History) described by William Harper Pease. Pp. 96; 14 Plates. 1965. (Out of Print.) £3.75. WHITEHEAD, P. J. P. The Clupeoid Fishes described by Lacepéde, Cuvier and Valenciennes. Pp. 180; 11 Plates, 15 Text-figures. 1967. £4. TAYLor, J. D., KENNEDY, W. J. & Hatt, A. The Shell Structure and Mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Pp. 125; 29 Plates 77 Text-figures. 1969. £4.50. Haynes, J. R. Cardigan Bay Recent Foraminifera (Cruises of the R.V. Antur) 1962-1964. Pp. 245; 33 Plates, 47 Text-figures. 1973. £10.80. Printed in Great Britain by Fohn Wright and Sons Ltd. at The Stonebridge Press, Bristol BS4 s5NU ae . S64 RN eS AD Sing phar INDEX TO VOLUME 25 The page numbers of the principal references and the new taxonomic names are printed in bold type. Acanthamoeba 129, 132 Acanthodactylus 302, 310, 339 acuticeps, Haplochromis. 4 3 eaest acuticeps, Tilapia . ” : : eS Adeona 245-247, 249-250, 2 256-259, PP I Adeonella 245, 247, 253-255, 250, 258-259, Pl. 3 Adeonellidae . 243-263, 3 Pls. Adeonellopsis 246, 250-253, 256-258, Pls. 2-3 Adeonidae 243-263, 3 Pls. Adolfus ci 357-358, 364 adolphifrederici, Haplochromis : . 234 aeneocolor, Haplochromis 147, 149, 150-158, 164, 199, 213, 224, 234, 239, 242 aerogenes, Klebsiella 0 Cee 2ei2 7, afra, Crenopharynx 2 : 3 5 Mae africanus, Adolfus . 357, 304 africanus, Algyroides 204, 357, 303-304 Afroedura. 7 A 5 323 Agama . 6 - 314 agilis, Teicerta 292— -293, ose 206, 301-302, 310-311, 316, 330-332, 334, 358, 364-365 alata, Sabatieria é 87, 90, 100-102 alba, Nitzschia : eEtsO) alba, Procnias 370-379, “381- 385, sire 389, Pl. 2 alcalicus, Sarotherodon . . 20-21, 23 alcalicus grahami, Sarotherodon c Be Alectroenas . é ‘ 2 SYhi algirus, Pe tmodzarius) 299, 302, 308, 310, 312, 314-315, 318, 320, eS Ee 364, 366 Algyroides 289-366 alisonae, Synonchus 87, 90, 106-108 alleni, Adolfus 357, 304 alleni, Algyroides 294, 357, 303 Alloporodontius” . 5 2 69, 81 alluaudi, Wem eccochroms . - 174, 214, 224-225, 228, 231, 233, 236, 239, 241 alluaudi occidentalis, Astatoreochromis . 224 Alokobius. < 3 é 63-64 altianalis, Barbus . 5 5 é 233 alveolata, Euglypha 121-123 alveolata minor, Euglypha_~. ¢ Eeb23 Amoeba : - A 6 ksi amphimelas, Sarotherodon 20, 23 Anachalypsicrinus 269-274, 287, Pl. 1 Anarthropora r 260 anatolica, Lacerta Gantord! 306, 334 Andebius 60, 62, 79 andersonii, Sarotherodon a ° 202i andersonii, Tilapia 0 5 27, 32 andreanszkyi, Lacerta 295- 296, 301, 304, 308, 312, 330, 334-336, Tab. III (Opp. P- 336), 338-339, 341, 353 355» 357, 364-305 andreanszkyi, Zootoca 295, 308, 312 angolensis, Loruwiala_. 7 0 28 angolensis, Sarotherodon 19, 21-33, Jel ot angolensis, Tilapia . : me. he} angustifrons, Haplochromis 177-183, 187-188, 213, 239, 242 annectidens, Haplochromis . g neti72 Ansorge, W. S. : 5 , : = 2 antarcticus, Ptilocrinus . : 5 = 269) anterides, Sphaerolaimus 5 : 95 Apathya 292, 294-295, 297, 301, 306- 307, 310-311, 313-314, ore 320, 327, 336 apus, Ophisaurus . - 339 arabium, Meee nantnionn 87, go, 114-116 araichensis, Lithobius 60, 62-63, 79 Arcella . - 132-133 Archaeolacerta 294- 296, 299, 301, 304, 306-307, 308, 312, 314-315, 316-317, 318, 320, 322-323, 327, 336, 340 Archilithobius 65-67, 70-71, 73, 78-79 Arenobius : 53-55, 57, 59 armata, Euglypha. 0 5 22 armeniaca, ee ae a 6 e295 armeniaca, Lacerta , 295, 310, 316, 330, ce Be 353, 304-365 Arnold, E. N. : . 289-366 asperatus, Bothropolys . Q c 75,79 astatodon, Haplochromis Astatoreochromis . 172, 234 174, 214, 224-225, 228, 231, 233, 236, 239, 241 atlantica, Gallotia . 295, 301, 312-313, 314, 350-351, 355» 357-358, 364, 366 atlantica, Lacerta . 295, 301-302, 312-313, 314 aureus, Sarotherodon : , 21, 23 australis, Lacerta . 5 292-293, 358 Australobius . 66-67, 68- 70, 71, 73, 75, 80-81 averano, Procnias . 369-379, 381-384, 386-389, Pl. 1 aztecus, Neolithobius. s é - 48 Barbus 233-234 Barilius 5 rs e234 batesii, Chromidotilapia,. j 6 16 Bathycrinidae é 286-287 beadlei, Haplochromis . 5 c . 199 bedriagae, Archaeolacerta 295, 304 394 INDEX bedriagae, Lacerta . . 295-296, 299, 304, 320, 330, 334-336, Tab. III (opp. p. 336), 337, 353, 364-365 Bedriagaia 292, 358, 364 bellbirds 367-391, 2 Pls. bermudensis, Topics 5 60, 62-63, 78-79 bethellianus, Hyocrinus . 269, 271, 274 bicolor, Macropleurodus . : . 228 bimaculata, Cichlasoma . : ‘ 4 23 bimaculatus, Hemichromis 3, 8-9, 16 bimunita, Laminopora 255-256 birmanicus, Archilithobius : : . 66 birmanicus, Australobius 5 ¢ 66 birmanicus, Lithobius 66— 68, 71, 73, 78 birmanicus chandellensis, Archilithobius 66, 68, 78 blanci, Psammodromus . 299, 309, 348-349, 358, 364, 366 bloyeti, Haplochromis 34, 149, 235 Bee ee Podarcis zs 5 - 346 Bodo . 5 3 5 EP enS2 Bothropolys . : 75,79 Bracebridgia . - 246— Daa apa ae 256, Pl. 2 brachiata, Euglypha < 7 S122 brandtii, Lacerta 208, OR. ao, 306, 330, 334-335, Tab. III (opp. p. 33), 339-341, 364-365 brucei, Ptilocrinus “ = 1269 Bryozoa . : ee 263, 3 Pls. buettikoferi, Pelmatochromis : 3, 7, 10-19 Burmobius - 3 68-70, 81 burtoni, Haplochromis c 5 2 4 busallae, Archilithobius 5 0 - 65 busallae, Lithobius : 2 65-66 busumana, Tilapia . cabrae, Tilapia : “ : = 31-32, 33 Cabrita a 5 5 ~ Sir caesaris, Gallotia galloti . 3 : - - 351 Calamocrininae . : 0 . 268, 287 Calamocrinus : 5 : - 267-268 callao, Andebius 2 6 60, 62, 79 callipterus, Haplochromis = : j 36 Callochromis 3 5 2 ‘ 17 Calyptomena : ° 4 : . 369 Camacolaimidae . z 2 A 92-93 (?)camerunensis, Tilapia . 3 : 2 campbelli, Sphaerolaimus a 95 cappadocica, Apathya 295, 301, 306-307; 310-311, 313-314, 318, 320 cappadocica, Lacerta 292, 294—295, 301, 306-307, 310-311, 313-314, 318 320, 322, 330, 334-336, Tab. III (Opp. P- 336), 337-338, 353, 364-365 cappadocica wolteri, Lacerta . : - 338 caroli, Sarotherodon c 2 23 carterae, Cricosphaera . a > Eee 20) ceascudum, Didelta. 3 é 3 a @2 castaneus, Lithobius 57, 59, 65, 79 castellanii, Acanthamoeba 7 Pee oe) caucasica, Archaeolacerta 3 205 caucasica, Lacerta . -295, 310, 316, 330, 336, Tab. III (opp. p. 336), 338, 353 364-305 caudatus, Bodo. ‘ 3 : 4 tee Cephalopterus : . _ . 386 ceylanicus, estalobicse c F 73 ceylanicus, Monotarsobius Ghat =. 79 ceylanicus, Tamulinus . 7° 993 chandellensis, Archilithobius birmanieas 66, 68, 78 Cheilostomata 243-263, 3 Pls. Chilebius . : 60, 62, 79 Chilopoda_. : 5 A 39-83 chlorogaster, Lacerta 290, 306, 310, 330, 335-336, Tab. III (opp. p. 336), 349, 353, 364-365 Chromidotilapia 16-18, 22 Chromis , 27, 33-34 Chromys : 5 9 7 4 35 Chrysochromulina . : 2 : 2Zg chuni, Democrinus . 267, 282, 284, 286-287 chuni, Rhizocrinus. < : 5 > 286 Cichlasoma_ . . : 0 23 Cichlidae 2 + 2 -37, ¥ PI. ; 139-242 ciliata, Euglypha . PEe2g cinereus, Haplochromis . 149, 158, 187 Clark, Ailsa, M. 265-288, 2 Pls. cobbi, Didelta 87, 89-92 Coccolithus . . 3 f : + 529) Comesomatidae B 100-104 congicus, Pelmatochromis , = 1-26 congicus, Pterochromis : 10, 11-15, 18 contorta, Laminopora_. : ; . 256 Cook, Patricia L. 243-263, 3 Pls. Coptodon < 15, 20, 22 coquimbo, Chilebius - 60, 62, 79 coralliformis, Adeonella . . 254-255, Pl. 3 corbali, Pelmatochromis . 0 4 : 15 Cotinga ‘ a : 5 - »360 Cotingidae aie 391, 2 Pls. crassilabris, Haplochromis 2 199, 234 crassipes, Lithobius . : Be Ad) crassipes, Monotarsobius 2 ce WS crassipes Holstii, Monotarsobius ° “ 715 crassus, Crenopharynx . o > DEL crenellatus, Sphaerolaimus 87, 90, 93-95 Crenopharynx 87, 90, 109-111 Cricosphaera x > 7 29 Crinoidea 5 265 288, 2 Pls. Crithidia 5 : : z . 129 Ctenochromis 5 34 208, 301, 306-307, 313- 314 322, 330, 334-335, Tab. III (opp. p- 336), 337-338, 353, 364-365 Cyatholaimidae. 5 é 1 tae cyreni, Lacerta sortie 5 323 cyanura, Lacerta IN dageti, Tilapia : 5 - 20-21, dahli, Archaeolacerta. é a 295 dahli, Lacerta 295, 310, 316, 330, 335-336, 353 danfordi, Archaeolacerta 295, 306, 308 danfordi, Lacerta . 295, 304, 306, 308, 320, 330, 334-336, Tab. III (opp. p. 336), 339-340, 353, 355, 357, 304-365 danfordi, Lacerta danfordi 306, 334 danfordi anatolica, Lacerta 306, 334 danfordi danfordi, Lacerta 306, 334 danfordi kulzeri, Lacerta 306, 334 Dayellus dayi, Dayellus 87, 90, TOBE 10g, 110 ae 90, 108-109, 110 decodontus, Lithobius 44-45, 78 decodontus, Sotimpius . : é 4d degeni, Platytaeniodus . ‘ he Ei Democrinus . 2 267, 282, 284, 286-287 derjugini, Lacerta . 299, 301, 306, 310, 315-316, 318, 330, 336 Tab. III (opp. p. 336), 337, 341, 353, 364-305 desfontainesi, Haplochromis . .- 236 Desmodora 87, 90, 98-100 Desmodoridae - 90, 98-100 Didelta 87, 89-92 Difflugia 131-132 Difflugiella o 131-132 dikume, Konia 5 z . 23 dimorphica, Platycomopsis isn 104, 106 Dimorphocella 3 c > 21260 diomedae, Gitmocrnuse 267-268 diplechma, Mesacanthion 87, 113-114 discolor, Tilapia 20, 23-24 dispersus, Pseudocrenilabrus philander 33-34 distoma, Adeonella : : e253 distoma imperforata, Adeonella : 7 253 dolichostylum, Procamacolaimus — . 87, 93 doriae, Archilithobius . ; : 65, 78 doriae, Lithobius . 5 : - 65-66, 78 doriae, Lithobius pilicornis . 5 65, 78 doriae uccianensis, Lithobius . t 65, 78 dugesii, Lacerta 295-296, 301-302, 304, 306, 308, 311, 316, 327, 330, 334-336, Tab. III (opp. p. 336), 338-340, 341, 346, 353, 355 364-365 dugesii, Podarcis 295, 301, 304, 306 Eason, E. H. . 2 F 5 39-83 easoni, Lithobius . ; 4 65 echinata, Lacerta . Been 358, 364 eduardiana, Schubotzia . F 215-216, 221 eduardianus, Haplochromis 215-221, 239, 241 eduardianus, Schubotzia. 23m eduardii, Haplochromis . 153, 199-200, 238 eisentrauti, Konia . a : ees elegans, Difflugia . : : nse DEX 395 elegans, Haplochromis 145-150, 153-154, 156-158, 164-167, 169, 178, 183, 186, 198-199, 213, 219, EIGER ae empodisma, Haplochromis é 183 Enantiosula . 4 : : é ~ 260 Engraulicypris : 9 : 228 engystoma, Haplochromis 153, 162 Enoplidae : I1I-116 Eremias c O i 310-311 Eremobius . . . , 5 A 59 erhardii, Lacerta , 295, 316 erhardii, Podarcis 295, 344-347, Bea 364, 366 erythrocephalus, Haplochromis 158-159, 229 erythrurus, Acanthodactylus . f 339) Escharoides . 5 247, 256-257, 260 ethochaetus, iethebine 7 . 60, 79 Euglypha é 119-137, Pls. 1-7 Euporodontus 4 2 4 7) 50) excrescens, Walesobius 60, 62, 79 Exechonella . és 5 . : - 256 Exochella. : 5 9 : . 260 exsul, Pelmatochromis . 3 a 3, 8-9 fasciatus, Hemichromis . 5 : ets fasciculata, Crithidia r 3 - 129 feae, Australobius . A A g 3 68 feae, Burmobius_ . a 5 F 4 68 feae, Lithobius é a ¢ - 68-70, 78 Fenestrulina . a 5 9 - 254; 260 filfolensis, Lacerta . 295, 304, 358 filfolensis, Podarcis 295, 304, 344-346, 364, 366 filfolensis, Podarcis filfolensis . 5 - 345 filfolensis filfolensis, Podarcis . : =) 345) filicaudatus, Synonchus . rs - 108 fitzingeri, Algyroides 299, 301, 341-343, 354) 357, 364-365 flavomarginata, Tilapia . c : 2 foliacea, Adeona . 3 5 @ . 250 forficatus, Lithobius 3 5 59 fraasii, Lacerta 298, 301, 303- 304, 306, 309, 318, 330, 334-335, Tab. III (opp. p. 336), oe 364-365 franchettii, Sarotherodon 0 20-21 fringilla, Crenopharynx 87, "90, 109-111 furcillata, Sabatieria : : 87, 102 fuscus, Haplochromis_. 3 : 0 (Lys fusiformis, Haplochromis 203-204 gaigeae, Podarcis taurica 0 4 S47. galilaeus, Sarotherodon . : F 23-24 galloti, Gallotia « 295, 312-313, 314, 350-351, 355, 358, 364, 366 galloti, Gallotia galloti . Pees sw galloti, Lacerta 295, "302, 312-3 14, 349 galloti caesaris, Gallotia . 5 Fi ae g5r galloti galloti, Gallotia . ‘5 A E355 306 INDEX Gallotia 292, 295-296, 298-299, 301-303, 306, 308-309, 311-314, 327-329, 339, 332-333, we 349-358, 363-364, 366 Gastropholis - 292, 358, 364 Gephyrocrinus 5 269, 274 Gigantopora . - . . 258 girigan, Sarotherodon pangani. : 21 gisleni, Zeuctocrinus . 277- 281, 287, Pl. 2 gladiator, Siphonolaimus 5 : 87, 98 godmani, Lithobius 46-47, 53, 78 godmani, Vulcanbius_. a vhs goliath, Gallotia . i A A - 351 goliath, Lacerta 350-351 gowersi, Haplochromis . : a - 234 gracilis, Sphaerolaimus . = 4 5 graeca, Archaeolacerta graeca, Lacerta 295, 299, 307 "205- 296, 304, 307, 316, 320, 330, 334-336, Tab. III (Opp. P- 336), 353, 355, 357) 364-305 grahami, Sarotherodon alcalicus “ 5 2 graueri, Haplochromis . 2 ; 234 Greenwood, P. H. . Fi “ 139-242 grimaldii, Gephyrocrinus 269, 274 guatemalae, Lithobius . 5 5 5 O Guerrobius . . b & 53,55 guiarti, Haplochromis 155, 178, 202-203, 207, 231-233 guinasana, Tilapia TH 2S eS 2 guineensis, Tilapia . 20, 23, 32 Gymnophthalmus . c “ if Be egrx Haliclona 5 z : : ' on tEgO Haliphysema ° 4 Oe tis }s} Haplochromis 18, 27, 33- <6) 139-242 Haplochromis group 7 L422, 22s4r Hartmannella : i ¢ 0 8G} haugi, Tilapia c 31 hawaiiensis, Nanmachoeraus Ds, 281, 28%, 285 Hedley, R. H. 119-137, Pls. 1-7 Hemichromis 3, 8-9, 14, 16, 31 Hemihaplochromis 34, 225-227, 228, 231, 233, 241 Hippomenella . : : 260 Hippopleurifera . é 5 ‘ - 260 Hippopodinella —. . 5 : =) 265 Hippoporidra : 261 hispanica, Lacerta hispanica, Podarcis | 295-296; 301, 318, 323 295, 301, 344-347, ; 357-358, 364, 366 hispanica-group, Podarcis 4 : - 346 hispanicus, Psammodromus 299, 309-310, 314, 347- ase 364, 366 Holaspis 2 5 BEE holstii, Lithobius_ . 5 : 0 75-78 Holstii, Lithobius . 3 : 0 75 Holstii, Monotarsobius crassipes : 5 Hoplotilapia . 216, 221 Hopperia 87, 90, 102-104 horvathi, Archaeolacerta d . - 205 horvathi, Lacerta . 295, 320, 323, 339, 336, Tab. III (opp. p. 336), 337, 341, 353, 304-305 humberti, Arenobius - 5 2 ci 55 humberti, Guerrobius . ci - 2 55 humberti, Kunobius F 55 humberti, Lithobius 48, 54, B25 —57, 78 hunteri, Sarotherodon . “ 23 Hydrocynus ; a 5 e237, Hyocrinidae . AS fears 287 Hyocrininae . : 268-269, 287 Hyocrinus 267-268, 269, 271, 274 imperforata, Adeonella distoma 3 5 253) implicatus, Platycomopsis 87, 90, 104-106 incrassatus, Monachocrinus 281-282 insulanica, Podarcis muralis_ . . - 345 Inversiula. : . : , 2 200 irvinei, Leptotilapia 0 : 6 16 ishmaeli, Haplochromis . 172-174) 176, 231-232, 233 jacksoni, Lacerta 292-293, 358, 364 japanicus, Tamulinus. 2 OTS: javanicus, Archilithobius 707%; 78-79 javanicus, Australobius . 2 - 70-71 javanicus, Lithobius 70, rhe, 75, 78 javanicus, Monotarsobius “ : - 70-71 javanicus, Tamulinus. - 70-71 jayakari, Lacerta . 208, seat 306-307, 313, 315, 330, 334-336, Tab. III (opp. Pp. 336), 337-338, 353 364-365 jentinki, Pelmatochromis 5 q : 17. joka, Tilapia é - : g eg kingsleyae, et aes ae 2 ee: Klebsiella. . : - 7 £20 koellikeri, Ophisaurus . a 3 - 339 Konia . 5 6 - . = 2g kottae, Tilapia . 3 c “ 23 kulzeri, Lacerta danfordi 306, 334 Kunobius”. - 53-55 kurdistanica, Taberta princeps 6 « 3932 labiatus, Haplochromis . eae 228, 234 Labrobius”s . ; a 50, 52 laccatus, Teehebine! 0 eC 3 = ) 865 Lacerta 0 c E 289-366 Lacertidae . c “ . 289-366 lacustris, Spongilla c - 130 laevis, Lacerta . 298-299, hoa 306, 308, 310, 334-336, Tab. IIT (opp. p- 336), 349, 353, 355 357, 364-365 laevis, Mabuya é - = 1308 Laminopora . "247, 255- 256, 258 INDEX 397 lapidicola, Lithobius > é : = 163 Lates . : : 0 e230) lebetina, Vipera lebetina, : 4 - 339 lebetina lebetina, Vipera c z S89 lebetina mauritanica, Vipera . . 339 298-299, 301— Bon 306, 310-311, 313-314, 315, 318, 327, 330-333, 351, 358, 364-305 lepida, Lacerta lepida pater, Lacerta. : e333 lepidura, Sarotherodon . : . gO) Leptosomatidae ; 104-108 Leptotilapia . : a < 5 : 16 lilfordi, Lacerta lilfordi, Podarcis limax, Haplochromis 2957310) 205, 344- 346, 358, 364, 366 . 151, 154, 158, 167-172, 198, 219, 227-228, 239, 242 Linhomoeidae - a é E 89-92 linnellii, Sarotherodon . F 5 é 23 Lipaugus 0 9 P d 4 3c6 Lithobiomorpha_ . : ® . 39-83 Lithobius : : 5 5 39-83 lividus, Haplochromis c 169, 171-172 lizards - : o 289-366 lohbergeri, Garathersdort ; : 4 23 longipinnis, Paratilapia . 4 a : 6 longirostris, Haplochromis : : e207, Loruwiala. , F a F 28-30 louka, Tilapia 4 : 5 6 : 16 lucullae, Haplochromis . é £ é 31 lucullae, Tilapia. : 5 3 32 Mabuya : 2 = 314, 323 maclareni, Pungu . : 5 a 23 macrocephalus, Serranochromis eng macroceros, Lithobius 42-45, 78 macroceros, Sotimpius . P 5 eryac macrochir, Sarotherodon F 6 29-30 Macropleurodus_. 5 2 eeez28: macrops, Haplochromis . 166, 231, 232 macropsoides, Haplochromis 158, 162-167, 169, 183, 219, 232, 239, 242 maculatum, Didelta a : 3 87, 92 malacophagus, Haplochromis . 176-177 malayicus, Alloporodontius’ . - Sr malayicus, Lithobius’. c =) 8a Manacus P 5 2 5 esi. manegitus, Terthobius : 5 55 marchi, Algyroides 299, 301, 341-343, 354, 357, 364-365 mariae, Stomatepia 5 3 16, 23 mariae, Tilapia 16, 20-21, 23 marine nematodes 4 85-117 masira, Desmodora 87, 90, 98-100 massiliensis, Hopperia 87, 104 mauritanica, Vipera lebetina . Q - 339 maxillaris, Haplochromis 194, 196 maxima, Gallotia? 3 " ss - 349 maxima, Lacerta 349-351 Mayorella. ‘ 133 meandrina, Adeonellopsis 250-251, 258, IL 3 mediterraneus, Lithobius 2 : 65, 78 Melanoides_ . 2 : 5 LTS melanopleura, Tilapia 3 ; . : 32 ‘melanopleura, Tilapia’ . a : D 24 melanotheron, Sarotherodon . 2 23-24 melisellensis, Lacerta melisellensis, Podarcis 295, 302, 310, 316 295, 341, 344-340 358, 364, 366 mentatus, Haplochromis 5 E - 204 mento, Haplochromis 207, 209 Mesacanthion 87, 90, 111-116 Mesalina fi ; . ° Ae es ite) Metrarabdotos 256-257, 260 microdactylus, pesrmodromes 299, 309, 314, Boer 349, 358, 364, 366 Microporella . é im 259 milensis, Lacerta 295, 302, 315 milensis, Podarcis 295, 302, 315, 344-340, 2 364, 366 minor, Euglypha alveolata - = 23 moffattii, Chromys : 4 5 a 35 Monachocrinus 5 ; 3 : 281-282 mongo, Stomatepia 2 : 22-23 Monotarsobius 702 71, 73, 75-78, 79 monticola, Archaeolacerta Q ZO 5 monticola, Lacerta 295, 320, 330, 335-336, Tab. III (opp. p. 336), 339, 341, 353, 364-365 monticola cyreni, Lacerta c : > 3238 moorii, Barilius. = - 234 moreoticus, Algyroides Dep, 299, 302, 304, 315, 341-343, 354, 357, 364-305 mosorensis, Archaeolacerta. 0 295, 317 mosorensis, Lacerta 295, 317, 320, 323, 330, 335-336, Tab. III (opp. Pp. 336), 337, 341, 353, 364-365 multicolor, Chromis : 5 2 multicolor, Haplochromis - 7 18, 34 multicolor, Hemihaplochromis 226-227, 228, 231, 233, 241 multicolor, Pseudocrenilabrus . d - 34-36 multifasciatus, Sarotherodon . < c 2 multigranosus, Sotimpius 5 2, 44, 78 multiocellatus, Haplochromis . 5 a ayy muralis, Lacerta 295, 298, 304, 310, 316-317, 318 muralis, Podarcis 295, 304, 317, 341, 344- Se ae 366 muralis, Seps . 343 muralis insulanica, Podarcis : 5 - 345 muralis nigriventris, Podarcis . : - 345 muralis-group, Podarcis ‘ 346 muscatensis, Hopperia. Brn 90, 102-104 Myaka . 4 : 5 0 ” 23 myaka, Myaka . : ‘ 23 mylodon, Haplochromis 172-1 77, 187, 227-228, 231-232, 233, 239, 241 natalensis, Pseudocrenilabrus . =) 127; 33-30 Naumachocrinus 267, 275, 281, 284-285 308 INDEX nefertiti, Anachalypsicrinus 269-274, 287, Pl. Neolithobius : a 2 : A 48 ngomoensis, Tilapia é ; c 31 nigricans, Haplochromis . 169, 171, 215 nigripinnis, Haplochromis 155-159, 161-166, 199-200, 229, 232, 239, 242 nigriventris, Podarcis muralis . . - 345 nigrofasciata, Paratilapia c 4 - 5-6 nigrofasciatus, Pelmatochromis 3-4, 6-15, 17-19 nigropunctatus, Algyroides 299, 302, 341-343, 354, 357, 363-365 nilotica, Tilapia c 0 18 niloticus, Sarotherodon . 5 : 21, 23 nitidissima, Alectroenas . , A A eye) Nitzschia 4 c ESO nubilus, Haplochromis on ee 150, 154, 167, 178, 221-224, 228, 231, 233, 236, 239, 242 nuchisquamulatus, Haplochromis 171-172 nudicollis, Procnias 369-382, 384, 387-389, Pl. 2 nudus, Phrynocrinus 274-275, 276-277, 280, 282 Nyasalapia . : . 2 5 . 30 obesus, Haplochromis — . 0 194, 215 obliquidens, Haplochromis ise 169, 171-172 obtortus, Phrynocrinus . 5 e270) occidentalis, Astatoreochromis alluaudi . 224 ocellifer, Pelmatochromis 2 3-4, 5-15, 19 Ocotea 5 5 ; 7S oedipes, LING NS - 0 9 46-47 Ogden, C. G. 119- Pay Gaelss Oncholaimidae 5 5 5 C 90 Onchoporella : : 2 254, 2 259-260 Ophisaurus . 6 : 5 : F we Ophisops a 310-311 Oplurus é . 3) 1323 oregosoma, Haplochromis 159-162, 166, 239, 242 oxycephala, Archaeolacerta 295, 304, 316-317 oxycephala, Lacerta 295, 304, 310, 320, 322-323, 330, 335-336, Tab. III (Opp: P- 336), 341, 353, 364-365 palestinensis, Mayorella . ’ é 183 pallidus, Haplochromis . : 149, 154 palnis, Lithobius . 5 - 7 77-81 pangani, Sarotherodon . 5 ‘ eee pangani girigan, Sarotherodon : 21 papillosus, Procamacolaimus 87, 90, 92-93 pappenheimi, Haplochromis 199-204, 228, 239, 241 pappenheimi, Tilapia. “ : « I99 Paramesacanthion . 7 A . 90 Paranticoma S ‘ : : ; 90 Paratilapia_ . - 5-6 parva, Lacerta “208, BGR) B06} 309-310, 314, 318, 330, 334-336, Tab. III (opp. p. 336), 339-341, 357, 364-365 parva—group, Lacerta 352, 353 parvidens, Haplochromis 5 “ . 194 pater, Lacerta lepida 4 5 SESSG paucidens, Haplochromis : : 4. 234 pectoralis, Haplochromis c ose pelagicus, Coccolithus . > - Brie) pellegrini, Barbus . 5 z34 Pelmatochromis . 3 é 5 1-26 ‘Pelmatochromis’ . 5 : : : 14 Pelmatolapia z 5 Dee ee) peloponnesiaca, Lacerta 298- 299, 302, 304, 306, 308, 312, 315, 327 298, 315, 344-346 358, 364, 366 F b 16 peloponnesiaca, Podarcis Pelvicachromis ; : percivali, Sarotherodon . ‘ 2 21, 23 peregrinus, Lithobius 59-60, 63, 79 Perissocephalus —. 3 : 5 . 386 Persea . % 5 2 «, 378 persicus, Alokobius : 5 63 persicus, Lithobius 63-65, 78 perspicillata, Lacerta 295, 301, 304, 306, 308, 311-312, 314, 318, 320, 322, 330, 334-336, Tab. III (opp. p. 336), 338-339, 353, 355, 364-365 perspicillata, Scelarcis 295, 301, 304, 3006, 308, 311-312, 314, 318, 320 petronius, Haplochromis 209-215, 219, 228, 235-236, 241 Phanodermatidae . 108-111 pharyngalis, Haplochromis 176-177, 213-215, 228, 235-236 pharyngomylus, Haplochromis 173-174, 176, 232, 233 philander, Chromis 27, 33-34 philander, Haplochromis 3 33-34 philander, Pseudocrenilabrus . A 32-36 philander dispersus, Pseudocrenilabrus . 33 Phrynocrinidae 274-281, 287 Phrynocrinus 274-277, 280-281, 282 Physarum. : BH Hgy piceus verhoeffi, Tecnones “ c ae 65 pilicornis, Lithobius F , c . 78 pilicornis doriae, Lithobius’. : 65, 78 pindu, Stomatepia 7 ( 5 22-23 pinnatus, Ptilocrinus . 0 « 269 Pipra . 3 : 3 9 : 377 Pisces . 6 5 5 - I 30-242 pityusensis, Lacerta pityusensis, Podarcis 292, 295, 316 295, 344-346, 358 364, 366 placidus, Sarotherodon . - oP 2u INDEX 399 placodus, Haplochromis . e . 176-177 platensis, Lithobius : : . 60, 62, 79 Platycomopsis 0 87, 90, 104-106 Platyplacopus é ; 5 . eso Platysaurus . 311, 314, 323-324 Platytaeniodus S : 5 8 2 é e2ZO Pocock, R. I. D : : 39-83 Podarcis 292-296, 297, Aseria 301-302, 304, 306, 308-309, 311-312, 314-316, 318, 323, 327-329, 332-333, 336, 338-339, 340-341, 343-347, 354-358 363-364, 366 polyarthra, Porphyrocrinus 282-285, 287 Polypterus . : : : ezs6) polystomella, Adeonella i ‘ 5 = 255) pontifer, Thalassocrinus 269, 271, 273, 274-275 pontifex, Arenobius ; ° rSS pontifex, Guerrobius 3 a : : 53 pontifex, Kunobius ; 2 53 pontifex, Lithobius 53s55y 57,78 Porina . A 5 a Ob Porphyrocrinidae c 3 281- 285, 287 Porphyrocrinus 267, 275-— ee ae 287 Posterula a 260 prasinus, Gastropholis : 4 : - 364 praticola, Lacerta . 295, 302, 310, 318, 330, 335-336, Tab. III (opp. p. 336), 337, 341, 353, 364-365 praticola, Zootoca . 295, 302 princeps, Lacerta . " 298- 299, 301- 302, 306, 311, 314-315, 318, 327, 330-333, 358, 364-365 princeps, Lacerta princeps 9 5 2332 princeps kurdistanica, Lacerta 5 eegg2 princeps princeps, Lacerta P e332 Procamacolaimus . , S 87, 90, 92-93 Procnias : - 367-391, 2 Pls. profundus, Siphonolaimus 87, 90, 96-98 ‘prognathus’ group, Haplochromis_ . 209, 234 proteus, Amoeba . s 5 Sue Protozoa é I 19-137; 7 Pils. provocator, Eremobius 4 5 59 provocator, Lithobius 59- 60, & 79 Prymnesium 5 2 c 5 5 AS) Psammodromus.. a C : 289-366 Pseudocrenilabrus . 27, 32-30 Pseudotropheus”~ . * - . 14 Pterochromis a 10, 11-15, 18 Ptilocrinus . - 3 5 269;,.274. Ptyodactylus 2 8 3 3 323 Pungu . 3 5 ; 23 pyriformis, Benccbrdeia 5 a ZA Pyroderus . 5 5 é é - 386 Quedenfeldtia : A S23 quinquetaeniata, Mabuya ‘ 0 or ahi regularis, Adeonella ° 5 0 - 254 remotorostrata, Smittina is a 25S rendalli, Tilapia b 15, 20-24, 32 Reptadeonella 3 3 246-251, 250-257 Reptilia ‘ 5 : : 289-366 retrodens, Hoplotilapia 5 d : 29 2225: Rhizocrinus . - : p : . 286 rhysodes, Hartmannella 5 ei : = 133 riponianus, Haplochromis 185, 188, 192 rodiaei, Ocotea 5 ‘ : : = 378 rosea, Haliclona’. 5 eErSO rostombekovi, Archaeolacerta 2 295 rostombekovi, Lacerta 295, 310, 330, 335-339, 353 rotunda, Euglypha 119-137, 7 Pls. rudis, Archaeolacerta. : - 295 rudis, Lacerta 295, 310, Ais 339, 335-336, Tab. III (opp. p. ooo) 353, 364-365 rukwaensis, Sarotherodon 2 , 30 Rupicola - 5 5 : 5 SSeS rupicola, Lacerta 6 6 292-293, 358 rupicola, Rupicola : c : mses tuweti, Tilapia 4, 14-15, 19 Sabatieria 2 87, 90, 100-102 salvini, Lithobius . 48-50, 53, 2 79 salvini, Vulcanbius : a 48, 79 Sarotherodon I- 26; 27-33, 35> PL I Sauromalus . 5 3 323 sauvagei, Haplochromis . a a 228 savignyi, Acanthodactylus : EeeSSO saxicola, Archaeolacerta : = 295 saxicola, Haplochromis . 0 ‘ 188, 192 saxicola, Lacerta . 292, 295, 310, 316, 322, 3)3f2h 335- 336, 337, 364-365 saxicola-group, Lacerta Tab. III (opp. p. 336), 353 scabrior, Australobius— . 5 5 67, 80 Scelarcis 293, 295, 297, 301, 304, 306, 308, 311-312, 314, ee 320, 327, 336, 339 Schizoporella ‘ : = 254 Schizostomella : 2 ; . 260 schoutedeni, Haplochromis 5 5 214234 schreiberi, Lacerta 295, 302, 316, 330-333, 358, 364-365 schubotzi, Haplochromis 177-178, 183-188, 190, 192, 202, 228, 234, 239, 242 Schubotzia_ . 215-216, 221, 231 schubotziellus, Haplochromis 180, 187, 188-192, 202, 213, 234, 239, 242 schwebischi, Sarotherodon : E 27, 30 schwetzi, Haplochromis . : é wee sculpturatus, Archilithobius . 73 sculpturatus, Lithobius . 41, 73-75, 77,79 sculpturatus, Tamulinus 5 3 73 scutata, Didelta . n 5 a : 2 scutellata, Didelta 4 : - 92 Seps. : 3 - Ro 825} Serranochromis 5 ; 6 5 2 2 serranus, Haplochromis . 207, 209, 232 400 INDEX ‘serranus’ gvoup, Haplochromis c - 209 serridens, Haplochromis . 170-171, 234 shimensis, Australobius . 3 : 75 shimensis, Lithobius . F : 75,79 shiranus, Sarotherodon . : 22-23 sicula, Lacerta sicula, Podarcis io 310, 315-316 a 315, 341, 344-346, 358, 364, 366 sicula-group, Podarcis . F cl - 340 Simochromis - . ¢ : 14 simonis, etramiella - : ; 2 simonyi, Gallotia 295, 302, 308, 314, 350-351, 355, 358, 364, 366 simonyi, Gallotia simonyi 6 350-351 simonyi, Lacerta 295, 302, 308, 311, 314 simonyi simonyi, Gallotia 5 : 350-351 simonyi stehlini, Gallotia 350-351 simonyi-group, Gallotia . . » - 351 simpsoni, Haplochromis . 5 . ; uss Siphonolaimidae . : 5 6 96-08 Siphonolaimus 87, 90, 96-98 Smittina : ; c F -) 258 Snow, D. W. 367- 391, 2 Pls. Sotimpius 2, 44-45, 50, 53, 78-79 southerni, Meededtatnion 87, 90, 11-114 Sowubius 5 3 0 57, 59 sparrmanii, Tilapia ot 19, 22-23, 32, 35 Sphaerolaimidae . ° - 93-95 Sphaerolaimus 6 eb 90, 93-95 spilurus, Sarotherodon . . ae Ezs Spongilla E 7 =L30 Squamipinnis, Haplochromis 204-209, 228, 230, 234, 239, 241 squamipinnis, Sarotherodon . . : 30 squamulatus, Haplochromis . . e232 stehlini, Gallotia simonyi 350-351 steinbachi, Sarotherodon a ; 2023 Steineria ; 5 F 6 é : 90 stolli, Arenobius . ; c Q ; 57 stolli, Lithobius ee 79 stolli, Sowubius. 4 : 5 57 Stomatepia r6h 22-23 strigata, Lacerta (265; 310, 316, 330-332, 358, 364-365 strigosa, Euglypha : : 2 e283 Stylopoma . F ‘ - 259 subsulcata, Bracebrideta 252, 253, Pl. 2 subsulcata, Porina ¢ : I 252 sulcata, Mabuya a sia 23 sydneyensis, Lithobius a 78-79 sydneyensis, Walesobius 5 ‘ 60 Synodontis . A a . c zs) Synonchus 87, 90, 106-108 takakuwai, Monotarsobius : - 75-78 Takydromus . 3 : 2 07,3 K0) Tamulinus 70-71, 73, 75 taurica, Lacerta taurica, Podarcis 295, 298, 302, 310, 318 295, 302, 318, 344-347, 358, 364, 366 taurica gaigeae, Podarcis c 5 =~) 347 taurinus, Haplochromis . 192-196, 227-228, 239, 241 Testacea 119-137, 7 Pls. Thalassocrinus ci go 271, 273-275 Theristus A F é c : tele) Thetia . c : - - 293 tholloni, Tilapia. 2 13, 18, 2 thomasi, apeiedatochicounet 5 : 2 14 thysi, Haplochromis 2 5 : , 31 Tijuca 5 : 5 . 386 Tilapia 205273 28, 302 33, 35, 199 tiliguerta, Lae, - 295, 316 tiliguerta, Podarcis 295, 344- 346) 358, 364, 366 Trewavas, E. ¢ 1-37, 1 Pl. Trewavasia . é Fi . , 15 tricarunculata, Proenias 370-372, 374-383, Re see, 388-389, Pl. 2 Trichometra . 5 5 neez75 Trigonopora . z 4 - 260 trilineata, Lacerta 295; 30%, 310, 316, 330-333, 358, 364-365 Triporula ‘ 2 . ‘ ‘ - 260 Tristramella . . : - 22-23, 24 tristrami, Acanthodactylus q 2330) Tropheus 5 c ae: tropidopholis, Bedriagaia : 7 - 364 Tropidosaura . c 3 A 292, 358 tuberculata, Melanoides . : 5 SLi Tylochromis . “ F : . é 17 uccianensis, Lithobius doriae . - 65, 78 Umbonula 256-257, 260 unisexualis, echaeolacerta 3 + 295 unisexualis, Lacerta som 295, 310, 316, 330, 335-336, 353 upembae, Sarotherodon . c F 2 1 30 urolepis, Sarotherodon . . 5 : 21 Uromastyx . 5 6 ¢ 5 - 314 variabilis, Sarotherodon . ‘ c ee sto vauereselli, Adolfus 357, 364 vauereselli, Algyroides 294, 357, 363 velifer, Haplochromis 149, 166-167 venator, Haplochromis . . : + Ze verhoeffi, Lithobius piceus - = 65 verrucosus, Porphyrocrinus vicarius, Haplochromis . 275, 2812 282, 285 153, 162, 171, 178, 232, 238 207, 209, 232 248-240, 250-251 victorianus, Haplochromis violacea, Reptadeonella . Vipera . . - 339 viridis, ikavoara "205; aro; 316, 330-332, 358, 364-365 Viscosia ‘ : 5 : : 90 vittatus, Gastropholis 2 2 A - 364 vittatus, Haplochromis . 5 3 5 By. vivipara, Lacerta INDEX 401 292-293, 295, 298, 302, 307, 310, 312-313, 318, 330, 333-336, Tab. III (opp. p. 336), vivipara, Zootoca . Vulcanbius vulcani, Labrobius. vulcani, Lithobius . vulcani, Vulcanbius vulgaris, Arcella wagleriana, Lacerta wagleriana, Podarcis Walesobius Warwick, R. M. Watersipora . weberi, Recraitehobius 337, 340-341, 353, 364-365 295, 302, 312-313, 318 40-48, 50, 52 : 5°, 79 50-53, 79 5°, 79 132-133 295, 344 295, 316, 344-346, 358, 364, 366 60, 62, 79 85-117 249, 254, 259 : 70-71 MAR, HisT. 2 small "> BE ARARSTEE Noa, pe : weberi, Australobius : : : 70 weberi, Lithobius . é 70-71, 75,79 wingatii, Haplochromis . 214-215, 235-236 wittei, Haplochromis” . 5 f 6234 wolteri, Lacerta cappadocica . és 338 Xantusia 0 a 5 = 323 ‘xenognathus, Haplochromis 5 215, 221 Xipholena . ° : 5 a - 369 yarraensis, Adeonellopsis 251-252, 258 Zeuctocrinus zilchi, Sotimpius zillii, Tilapia Zootoca 276-281, 287, Pl. 2 50, 52, 79 15, 20-23 293, 295— 296, 299, 302, 306, 308, 312-313, 315-316, 318, 327, 331, 336 a aoe bin Ke Ry i fe BREUER