THE BOLTI. 347 



Having previously received, after writing the description of 

 Tilapia nilotica for the ' Catalogue of African Fresh-water 

 Fishes,' two young specimens from another affluent of the same 

 river, the Simba River, it occurred to me to examine them 

 ■carefully, as I should have done before, and I found four to be 

 the number of spines in both. In their physiognomy, in their 

 coloration and markings, and in all structui^al particulars, these 

 fishes are indistinguishable from Tilapia nilotica ; and although, 

 in view of the constancy of the increased number of anal spines, 

 the Athi River specimens may be recognised as a new local form, 

 under the name of var. athiensis, I should not think of proposing 

 for them a new species. 



A further remarkable fact is the presence of four anal spines 

 in another Tilapia very closely related to, though sufficiently 

 distinct from, 2\ nilotica, viz. T. (^Oreochromis) nigra Gthr., also 

 from the Athi basin. Why in the Tilajna from this river-system 

 a.n increase should have taken place in the number of anal spines 

 is difficult of explanation, unless it be that an abnormal trans- 

 formation of a soft ray into a spine, as happens elsewhere, should 

 have been a peculiarity of the first settlers in that basin of the 

 widely distributed T. nilotica, and, becoming fixed, been passed 

 on to T. nigra, which may well be regarded as derived from that 

 species. Whatever this explanation be worth, the fact is clear 

 that, unless our classification of the Oichlid^e be made still more 

 artificial than it unfortunately is at present, the number of anal 

 spines must not be used, as a single character, for the division 

 into genera, and it affords the best justification that could be 

 wished for the course I have followed in the past. 



