504 CICHLIDJE. 



support Dr. Pellegrin's opinion, and although I do not feel quite satisfied as to the 

 course here followed, the enormous amount of variation with which I have met has 

 compelled me to give it my assent, provisionally at least. 



I have had large series from two points only in the vast habitat assigned to this 

 species, viz. from Algeria (the typical H 9 desfontainesi) and from Lake Victoria. The 

 specimens from the latter usually differ in having a longer pectoral fin and in attaining 

 a larger size, but many specimens, which are sexually mature, are no larger than the 

 Algerians, and the length of the pectoral is subject to considerable variation. Besides, 

 the two extreme types are connected by the form which has been described as 

 Hemichromis bloyeti. There is a striking resemblance in form, dentition, and 

 coloration between the small specimen from Gondokoro described by me as Para- 

 tilapia wingatii, and the Syrian fish described by Dr. Lortet as Chromis flavii-josephi. 



The above table of measurements, &c, will enable the reader to form some opinion 

 as to the range of variation according to size and localities, and I append further notes 

 on the variation of the fin-rays in large series of specimens. 



In forty specimens from Algeria and Tunisia, I find the dorsal fin-formula to be : 

 XIV 10 once, XIV 11 five times, XV 10 five times, XV 11 twenty-three times, XV 12 

 once, XVI 10 three times, XVI 11 twice ; the anal formula : III 9 nine times, III 10 

 twenty-nine times, III 11 once, and IV 9 once. In fifty specimens from Lake Victoria, 

 the dorsal formula is XIV 10 once, XV 8 once, XV 9 twenty-two times, XV 10 seven 

 times, XVI 8 twice, XVI 9 sixteen times, XVI 10 once ; the anal formula is III 8 

 four times, III 9 thirty-seven times, III 10 nine times. In five specimens from Lake 

 Manjara I count D. XIII 10 once, XIV 9 once, XIV 10 twice, XV 9 once ; A. Ill 8 

 three times, III 9 twice. 



The teeth on the pharyngeal bones are usually conical, but in some of the larger 

 specimens, from Lake Victoria, they have a tendency to assume the molariform 

 character exhibited by H. alluaudi, the passage between these two types of dentition 

 being as complete and as perplexing as that shown by the external characters of the 

 two species. 



As in so many of the Cichlidse, the female, which is considerably smaller than the 

 male, protects the eggs and young by sheltering them in her mouth and pharynx. This 

 observation has recently been made near Biskra by Lord Walsingham. He collected 

 a large number of specimens in. a tepid spring at Ain Oumash, and most of the females 

 had the mouth and pharynx full of eggs or embryos. The egg of this fish is not 

 perfectly round, the upper pole being somewhat pointed ; the vitelline sac, when 

 hanging from the embryo, is pyriform. 



Haplochromis moffati {Chromis moffati, Castelnau, C. philander, M. Weber), from 

 Natal and the Transvaal, is barely distinguishable from H. desfontainesi. Its dentition 

 varies to the same extent, and examples of it have likewise been described under two 

 genera. H. callipterus, Giinther, from Lake Nyassa, is another close ally. 



