38 



PH. GREENWOOD 



have examined from the Okavango swamp and river system 

 in Botswana seems to indicate that such variation may occur. 

 A soft margin, for example, was absent in specimens of 

 Serranochromis thumbergi and S. longimanus, but is present 

 in S. macrocephalus . From the same samples some individual 

 variability was noted both in the extent of the granular area 

 on the caudal field of flank scales below the upper lateral- 

 line, and the presence or absence of a granular area in at least 

 some scales above that line. Apparent interspecific variation 

 is also seen in the extent and nature of cheek, opercular and 

 interopercular squamation patterns. The value of these fea- 

 tures as a basis for intrageneric classification has, however, 

 yet to be tested on a wider geographical and taxonomic basis. 



A feature of Serranochromis, Sargochromis and Chetia 

 noted in an earlier paper (Greenwood, 1979), where it was 

 ranked as an apomorphy, is a reduction in the maximum 

 number of inner tooth rows in both jaws to a single or 

 irregularly double series. The new species, 5. altus, described 

 by Winemiller & Kelso-Winemiller (1991), is exceptional in 

 having as many as six rows anteriorly, reducing to a single or 

 double row posteriorly (the figures adjusted from Winemiller 

 & Kelso-Winemiller who include the outer tooth row in their 

 counts). These authors also note that the number of inner 

 rows increases with growth ". . .in all species of Serrano- 

 chromis that we have examined. . .". Their list of study 

 material indicates, however, that only one other species, S. 

 angusticeps, was studied, and that the maximum number of 

 inner or outer rows in that species is three. Since museum 

 specimens are often not fully representative of a species' full 

 size-range, the possibility of growth related changes in the 

 number of tooth rows should perhaps be treated, for the 

 moment, as an open question. In that connection, however, it 

 should be noted that in Sargochromis, one species, Sargo. 

 thysi, has four inner rows in both jaws at a size when only one 

 or two rows are present in other congeneric species. 



Another dental feature in Serranochromis, one also shared 

 with Sargochromis, (and probably Chetia), is for the two 

 median teeth in the outer row of the inner tooth series to be 

 enlarged and displaced anteriorly (Greenwood, 1984: 216; 

 also see figs 3 & 8 in Trewavas, 1964). Amongst the sample of 

 Serranochromis specimens examined, which included all spe- 

 cies of the genus except S. altus. (see p. 35, and Greenwood, 

 1979 & 1984 for additional material), such tooth displacement 

 occurs in most species but not in all individuals of a species. 

 Its frequency of occurence is lower in Sargochromis than in 

 Serranochromis, and the condition is known from only one 

 Chetia species (C. gracilis; see below and p. 39). Because of 

 this inconstancy in its expression I would now consider the 

 character more in the nature of a trend, albeit a derived one, 

 rather than the trenchant synapomorphy recognised earlier 

 (see Greenwood, 1984; 216). 



A revised generic diagnosis for Serranochromis is given on 

 p. 42, and the possible relationships of the genus are dis- 

 cussed on p. 40. To the list of included species published in 

 Greenwood, 1979 (pp. 302-303) must be added 5. altus 

 Winemiller & Kelso-Winemiller (1991) from the Zambezi 

 system. 



Little need be added to previous accounts of the genus 

 Sargochromis, type species Paratilapia codringtoni Blgr., 

 1908 (see Greenwood, 1979; 1984, in both papers the taxon 

 treated as a subgenus) except to note that the species 

 described as Serranochromis (Sargochromis) gracilis Green- 

 wood (1984) from the Cunene river, is now considered to be a 

 member of the genus Chetia (see p. 39). The transfer of this 



species reduces to two the number of Sargochromis species 

 having, relative to other congeneric taxa, slender lower 

 pharyngeal bones with few and only partially molariform 

 teeth, viz. Sargochromis greenwoodi (Bell-Cross) and Sarg. 

 coulteri (Bell-Cross). The latter species, however, appears to 

 show considerable variability in these features, with some 

 individuals having noticeably coarser lower pharyngeal bones 

 than do others, a feature invariably correlated with an 

 increased number of molariform teeth (Greenwood, 1984: 

 217-221). As was discussed in that paper, the species level 

 taxonomy of Sargochromis is far from satisfactory, so this 

 seemingly intraspecific variability should only be accepted 

 with some reservation (see also Greenwood, 1965 and 

 Hoogerhoud, 1986). 



In lacking a pronounced ventral expansion of its sphenotic 

 postorbital process(see p. 37), the neurocranium in Sar- 

 gochromis is immediately distinguishable from that of Serra- 

 nochromis (see Fig 3; also comments in Greenwood, 1979: 

 303, and figs 13 & 16). 



A list of Sargochromis species is given in Greenwood, 

 (1979, pp. 304-305), and a revised generic diagnosis on p. 42 

 below. 



The genus Chetia Trewavas, 1961 



Type species. Chetia fla viventris Trewavas, 1961. 



It has not proved possible to formulate a trenchant generic 

 definition for Chetia since the taxon has yet to yield a single 

 diagnostic autapomorphy. Its definition, therefore, is based 

 on, as it were, negative features, namely those which exclude 

 the five member species from inclusion in any of the three 

 other serranochromine genera, especially Serranochromis 

 with which Chetia has the greatest superficial and some 

 detailed similarity. Thus, although Chetia shares with Serra- 

 nochromis the apomorphy of two postorbital scale rows (see 

 p. 36), a high number of caudal vertebrae (15-17, modes 16 

 or 17) and an increased number of circumpeduncular scales 

 (18 or 20, rarely 16), it does not share with Serranochromis 

 apomorphies of an increased number of abdominal verte- 

 brae, (and as a consequence, an increase in the total number 

 of vertebrae), nor the increased lateral-line scale count of 

 that genus (35-41, a character probably correlated with the 

 increased number of vertebrae). Neither do the two genera 

 share the autapomorphic dental character of Serranochromis , 

 namely the development of a totally unicupsid jaw dentition 

 at a very small size; i.e. , at some length, yet to be determined, 

 less than 29 mm S.L.; see Greenwood, (1979: 300). 



On the basis of that analysis, the simplest diagnosis for 

 Chetia is: a Serranochromis lacking the apomorphic features 

 of that genus (see also Greenwood, 1992: 49-50 and p. 43 

 below). 



At the time of my earlier generic review (Greenwood, 

 1979), Chetia comprised two species, namely the type, Chetia 

 flaviventris , and Chetia brevis, Jubb, 1968; the species coming 

 from the Limpopo and Incomati river systems of South Africa 

 respectively. 



Based on Jubb's original description together with an 

 examination of the holotype and two other Chetia brevis 

 specimens, I excluded, in my 1979 paper, the species from 

 Chetia and placed it, as incertae sedis, in the genus Astatotila- 

 pia (Greenwood, 1979: 284 & 307). That decision was based 

 on three supposed features of C. brevis. (i) Jubb's (1968) 

 description of the anal fin markings as ocelli, together with 



