142 Cockerdl — The Scales of the African Cyprinid Fishes. 



Hner, th(_' transition :il)rnpt; al)out nine delicate apical radii, and 

 a few irregular, poorly developed l)af^al ones. Tlie nuclear area is 

 a little basad of the middle, and may he hroadly free from circuli, 

 or circulate nearly to tlie middle. The lish is small and silvery; 

 the scales are large in proportion, though only about 1% mm. across. 

 There are no distinct laterobasal angles. Both the fish and the 

 scales are rather suggestive of the American Xotropis, though nuuiy 

 differences are apparent on close comparison. 



The Indian Chela arrjentea Day (Wynad, F. Day) has larger 

 scales, with evident laterobasal angles. The apical radii are well 

 defined, al)out 14, l)ut several imperfect centrad ; the basal radii are 

 two or three, but very strong. All this shows evident reseml^lance 

 to the type of radiation found in Alestes, and in the ]ilcurol;vnia 

 group of Barbus; that is to say, with few strong radii arising from 

 the nucleus. In Chela ur<ientea tiiis is superimposed, as it were, 

 on a system of more ordinary apical radii, and it is only the latter 

 that occur in Chehi'thiops. 



A further study of the scales of Chela shows that C. argentea is 

 quite exceptional, and I find one species which has scales of essen- 

 tially the same type as those of Clwhcthiops, except that there are 

 fairly well developed laterobasal angles, no basal radii, and the 

 apical radii are greatly reduced, with not more than two or three 

 complete. This species is C. (Paralanbuca) typiis (type of Para- 

 laiihuca lateralis Sauvage) from the Me Kong (Paris Museum). 

 C. anomalurua from Sarawak has a scale of the same general char- 

 acter as C. typus, but it is excessively short and liroad (long, (i, 

 lat. 10 mm.), with no laterobasal angles, and quite munerous l)ut 

 very weak apical radii. On the other hand C. chipeoidcK (Alnrree, 

 F. Day) has .small scales which are much longer than broad, with 

 six or seven apical radii and no basal. C. hacaila (Orissa, F. Day) 

 has scales of the type of C. clupcohli's, ]m{ with at least twice as 

 many apical radii. 



The scales of C. macrochir ( Menam River, Royal Siamese Museum ) 

 are transversely oval, scarcely radiate, mucli like Chehvthiui)s hut 

 with the coarse lateral circuli extending over into the a]iical region. 

 This species has the dorsal tin very far back, and the scales were 

 taken from the middle of tln'side. Chela sardiurlla (Sittang liiver, 

 E. W. Gates) has broad weak scales with al)out 12 apical radii; 

 there is a dark spot at the fork of the camlal lin. 



Thus Chela, on the scales, seems to fall into at least three distinct 

 groups: (1.) C. argentea, (2.) C. clupeoides and hacaila, (:'..) C. 

 typus, macrucliir and sardinella, with asul)group f<jr C. anomalurus. 

 Chehethiops is evidently to be compared with the third group. 



The ChcUvllviops scale has (|uite a strong suiterlicial resend)lanci' 

 to that of the Characinid Cithariims cuiiyicus, although in the 

 latter the lateral cinaili are more numerous, and less abruptly <lif- 

 ferentiated from the apical. 



