THE SERRANOCHROMINE GENERA REVIEWED 



39 



his and my personal observations of their large size and 

 restricted number (3 or 4) as compared with the small and 

 numerous anal spots in other serranochromines (see p. 34), 

 especially Chetia flaviventris. (ii) Jubb's (op. cit) statement 

 that the scales are ctenoid, and its implication that such scales 

 are the predominant form, (iii) The presence of bicuspid 

 teeth in the outer oral tooth rows of some specimens more 

 than 100mm S.L., contrasting with the situation in Chetia 

 flaviventris where, on the information then available, few 

 bicuspids were thought to be present in fishes 71-85 mm 

 S.L., and only unicuspids occurred in specimens above that 

 length. 



Having now been able to examine colour-transparencies of 

 live C. brevis, it is clear that the anal spots are not true ocelli 

 (as defined on p. 34), but are large versions of the maculae 

 found in other serranochromines, including Chetia flaviven- 

 tris. The taxonomic significance to be attached to their large 

 size, and small number, cannot yet be assessed. However, a 

 wide size-range of anal maculae, their size negatively corre- 

 lated with their number, occurs in the monotypic serrano- 

 chromine genus Pharyngochromis , some individuals of which 

 have spots as large as those in Chetia brevis. 



The larger collection of C. brevis specimens now available, 

 together with a re-examination of the entire type series, 

 shows that Jubb's blanket description of the scales as ctenoid 

 is somewhat misleading. As in Chetia flaviventris , the scales 

 of C. brevis are cycloid above the lateral-line, and mostly 

 cycloid below that level. A variable number, usually small, of 

 ctenoid scales is present anteriorly on the flanks, such scales 

 being most numerous in fishes less than 80 mm S.L. In other 

 words, the same pattern as occurs in Chetia flaviventris is also 

 found in C. brevis. 



An examination of larger samples of both C. flaviventris 

 and C. brevis also revealed that bicuspid teeth can be present 

 in idividuals of C. flaviventris up to a standard length of 

 90 mm, and that true unicuspid teeth as well as very weakly 

 shouldered (almost unicuspid) teeth occur in C. brevis speci- 

 mens between 80 and 90 mm S.L. This situation severely 

 weakens my third reason (see above) for excluding C. brevis 

 from the genus Chetia. 



Furthermore, the presence of two vertical rows of postor- 

 bital scales in both Chetia brevis and C. flaviventris (see 

 p. 36), a feature previously overlooked by all workers, pro- 

 vides additional evidence for returning the species 'brevis' to 

 the genus in which it was originally, and I now acknowledge 

 correctly placed by Jubb (1968). 



Since 1979 another two species have been included in the 

 genus, namely the taxon Chetia mola described by Balon & 

 Stewart (1983) from the Zaire river system, and Hap- 

 lochromis welwitschi (Blgr.) from Angola (see Greenwood, 

 1979 & 1984). 



Chetia mola differs from all other congeneric species in 

 having a greatly enlarged lower pharyngeal bone with a 

 heavily molarized dentition (see fig. 12 in Balon & Stewart 

 1983), a type of pharyngeal mill more usually associated with 

 species of Sargochromis . However, unlike members of that 

 genus, C. mola has a double or sometimes triple row of 

 postorbital scales, and a lower number of abdominal verte- 

 brae. 



To the existing four species of Chetia, I would now add a 

 fifth, a species from the Cutato river, Angola, originally 

 described (Greenwood, 1984) as Serranochromis (Sar- 

 gochromis) gracilis. The reasons for this transfer are the 

 double row of postorbital scales in 'gracilis', the persistence 



of compressed, bi- and weakly bicuspid teeth in the outer 

 tooth row of both jaws in specimens over 100 mm S.L. and of 

 tricuspid teeth in the inner tooth rows in such invididuals (see 

 Greenwood, 1984: 227, and above), and the presence of only 

 15 abdominal vertebrae, a combination of derived and plesio- 

 morphic characters not occurring in any known Sargochromis 

 or Serranochromis species. 



On the basis of comparisons with the few specimens 

 available, Chetia gracilis (n=2) differs from the only other 

 Chetia species recorded from southwestern Africa, C. wel- 

 witschi (n=5), in having a narrower interorbital width 

 (18.2-18.6 cf 21.0-23.3% of head length in C. welwitschi), a 

 larger eye (25.0-25.6, cf 18.6-22.25% head; the effects of 

 allometry are unlikely to account for this difference since the 

 two samples overlap in the size range of individuals repre- 

 sented), and a shallower cheek (22.7-23.3 cf 32.0-33.3% 

 head). The lower pharyngeal bone in C. gracilis is somewhat 

 stouter, and its median teeth coarser than in C. welwitschi (cf 

 figs 19 & 20 in Greenwood, 1984). 



In its morphometric and meristic features, C. gracilis 

 closely resembles C. flaviventris and C. brevis, species respec- 

 tively from the Limpopo and Incomati river systems in South 

 Africa. From Chetia brevis, C. gracilis is distinguished by its 

 narrower interorbital width (18.2-18.6 cf 23.0-26.0% head 

 length), and from C. flaviventris, especially when specimens 

 of approximately the same size are compared, by its shal- 

 lower cheek (22.7-23.3 cf 25.5-33.6% head). It also seems 

 likely that, when the effects of allometric growth are taken 

 into account, the eye diameter is greater in C. gracilis than in 

 C. flaviventris. The teeth in the median row of the lower 

 pharyngeal bone of C. gracilis are coarser and stouter than 

 those of C. flaviventris, in that respect being more like the 

 teeth in C. brevis (cf fig. 19 in Greenwood, 1984, fig. 19 in 

 Greenwood, 1974, and fig. 4B in Jubb, 1968). 



Regrettably, apart from Chetia flaviventris (see du Plessis 

 & Groenewald, 1953; Trewavas, 1961) and C. mola (see 

 Balon & Stewart, 1983; fig. 10b) little or nothing is known 

 about the live coloration of the other Chetia species. 



A revised generic diagnosis for Chetia is given on p. 43. 



Included species: 



Chetia flaviventris Trewavas, 1961. Limpopo river system. 



Chetia brevis Jubb, 1968. Incomati river system 



Chetia mola Balon & Stewart, 1983. Luongo river, Zaire 



system. 

 Chetia welwitschi (Boulenger), 1898. Cunene and Zaire river 



drainage systems, Angola (see Greenwood, 1979). 

 Chetia gracilis (Greenwood), 1979. Cutato river (Cubango 



drainage system), Angola. 



The genus Pharyngochromis Greenwood, 1979. 

 Type SPECIES: Pelmatochromis darlingi Boulenger, 1911. 



A detailed revision of this monotypic genus, together with 

 an annotated synonymy, has been published recently (Green- 

 wood, 1992). Three nominal species all classified in the genus 

 Haplochromis since Regan's revision of 1922 (or in one case, 

 Pharyngochromis) , viz. Pelmatochromis darlingi Boulenger, 

 1911; Chromis jallae Boulenger, 1896, and Pelmatochromis 

 multiocellatus Boulenger, 1913, are now treated as junior 

 synonyms of the single Pharyngochromis species, P. acuticeps 

 (Steindachner) 1866. 



There are some indications that P. acuticeps could be 



