36 



P.H. GREENWOOD 



SERRANOCHROMINE TAXONOMY 



Introduction 



In their revision of certain haplochromine genera from Lake 

 Malawi, Eccles & Trewavas (1991: 21) divide the non- 

 tilapiine taxa of that lake into three groups, one of which they 

 call the Pharyngochromis - Chetia - Serranochromis group. 

 That action I consider to be premature, both because there is 

 little concrete information available about the phyletic inter- 

 and intrarelationships of the Malawi species, and because, as 

 Eccles & Trewavas point out, there are differences between 

 the squamation of those species and that of the serrano- 

 chromines as construed in this paper (see p. 40). Granted, I 

 have suggested (Greenwood, 1979: 314) that Serranochromis- 

 and Chetia-like taxa could have been involved in the origin of 

 the Malawian cichlid flocks, but that idea was not put forward 

 on the basis of characters constituting a testable hypothesis. 

 Rather, it was intended, because of the superficial resem- 

 blance between the two groups, to promote an awareness that 

 possible intergroup synapormophies should be looked for in 

 future research. 



As recognised in this paper, the serranochromines are an 

 assemblage of mainly fluviatile taxa whose geographical 

 range encompasses the Zambezi, Save-Runde, Limpopo, 

 Cunene, Quanza, Okavango and Zaire river systems, with 

 one species (Serranochromis robustus robustus [Giinther]), 

 occurring in Lake Malawi (see Balon & Stewart, 1983; 

 Bell-Cross, 1975; Eccles & Trewavas, 1989; Greenwood, 

 1984, 1992; Jubb, 1967, 1968; Ladiges, 1964; Poll, 1967; 

 Skelton, 1993, Trewavas, 1961, 1964). 



Since the last published inventory of serranochromine 

 species (Greenwood, 1979), revisional studies (Greenwood, 

 1984, 1992) and the description of new species (Balon & 

 Stewart, 1983, Winemiller & Kelso-Winemiller, 1991, and 

 Greenwood, 1984 [see p. 38 below]) have both increased the 

 number of included taxa and extended the geographical range 

 of the group. 



Morphologically, the two distinguishing features of the 

 serranochromines are the following, (i) The presence, often 

 in both sexes, of non-ocellate maculae (see p. 34) on the anal 

 fin. Generally these spots are very numerous with, in certain 

 species, as many as 30-40 covering almost the entire fin. 

 However, their number and size show considerable intraspe- 

 cific (and intergeneric) variability, with as few as three or four 

 large spots occurring in some individuals of a species where 

 the maximum number is 18-20 (Greenwood, 1992). When 

 many spots are present their arrangement may give the 

 impression of an irregular distribution on the fin, but (as was 

 noted by Oliver [1984], pace Greenwood, 1979: 315) there is 

 a basically linear regularity in their arrangement, (ii) All 

 scales above the lateral-line series are cycloid, as are the 

 majority of scales below that level. Some weakly ctenoid 

 scales may occur anteriorly on the flanks, especially in small 

 specimens, the ctenii being confined to a narrow arc situated 

 near the centre of the scale's free margin. 



A third but less trenchant feature of the serranochromines, 

 as compared with other fluviatile non-tilapiine and non- 

 pelmatochromine taxa (both sensu Greenwood, 1987: 

 194-199) is a tendency for there to be a higher modal number 

 of abdominal vertebrae (modes 16 or 17 in serranochromines, 

 [but 14 in one taxon] cf 12 or 13 in the other taxa); however, 



the ranges for total vertebral counts in the two groups 

 overlap. 



Where information is available on breeding habits, the 

 serranochromine species are known to be female mouth- 

 brooders, and in all taxa the neurocranial apophysis for the 

 upper pharyngeal bones is formed from the parasphenoid and 

 the basioccipital bones. In those respects, the group would 

 conform with Eccles & Trewavas' (1989) definition of the 

 tribe Haplochromini. 



Recently, Lippitsch (1991) drew attention to the possible 

 value of scale morphology and patterns in resolving certain 

 problems of intergeneric relationships within the serrano- 

 chromines. To determine if the features described by Lip- 

 pitsch could also provide another 'group' characteristic, I 

 examined, at a magnification of 50x, the superfical morphol- 

 ogy of flank and other scales in several species belonging to 

 both subgenera of Serranochromis (sensu Greenwood 1979; 

 but see p. 37), all species of Chetia, and in different popula- 

 tions of the single Pharyngochromis species. In general I 

 found a fairly high level of intrageneric variability, as well as 

 some individual variability in the features noted by Lippitsch 

 (1991: 99-100; figs D & E) i.e. ornamentation of the caudal 

 field, and the presence of a soft caudal rim to the scale. 

 Further studies will be necessary before the value of these 

 features can be established, both at the group and lower 

 levels of serranochromine taxonomy. However, another 

 squamation feature noted and discussed by Lippitsch, namely 

 the number of scale rows between the posterior orbital 

 margin and the preoperculum, has proved of considerable 

 value in reviewing generic level taxonomy within this group. 



The generic or subgeneric status of Serranochromis 

 Regan, 1920, and Sargochromis Regan, 1920, 

 reconsidered. 



In an attempt to revise the Haplochromis generic concept on 

 phyletic lines, several so-called Haplochromis species were 

 assigned, as a subgenus, to the genus Serranochromis (Green- 

 wood, 1979). Since one of these species is the type of Regan's 

 genus Sargochromis (S. codringtoni [Big, 1908]) that name 

 was resurrected for the new subgenus. 



The principal argument for this taxonomic rearrangement 

 was that the taxa included in the new concept of Serrano- 

 chromis all share what appeared to be three derived features, 

 viz: an increased number of abdominal vertebrae, higher 

 gill-raker counts, and an increased number of branched rays 

 in the dorsal fin (Greenwood, 1979: 299). 



In the light of new data, especially those stemming from an 

 increased knowledge of the genus Chetia (see p. 38) and, 

 particularly, Lippitsch's (1991) observations on the different 

 type of postorbital squamation patterns present in the two 

 presumed subgenera, I would now revise my earlier action 

 and recognise both Serranochromis and Sargochromis as 

 distinct lineages, and thus accord each generic status. The 

 reasons for that action are as follows. 



Firstly, as Lippitsch (1991) noted, in Serranochromis but 

 not in Sargochromis there are at least two, and often more, 

 vertical rows of scales between the posterior orbital margin 

 and the upper part of the preoperculum's ascending arm. In 

 Sargochromis only a single row is present. A double row, 

 however, also occurs in Chetia (personal observations, see 

 below). Further investigation of both Serranchromis and 

 Chetia reveals that the double and multiple scale row condi- 

 tion is correlated with underlying osteological and myological 



