EAST AFRICAN CHAMELEONS. 323 



C. quilensis (=parvilobus, Blgr.). — Occipital lobes larger than 

 C. gracilis. Male with tarsal spur. 



G. dilepis. — Occipital lobes larger than C. quilensis, and in 

 contact with each other in the middle. Male with tarsal spur. 



It is stated by Werner that he has examined a male G. roperi 

 which possesses the tarsal spurs ; he therefore, with Boettger, 

 regards it as only a variety of G. dilepis. I have, however, 

 examined the five specimens of males in the Museum collection, 

 but could not find one in which the spur is developed. But for 

 this character it would be pardonable to confuse C. roperi with 

 either C. dilepis or C. quilensis, as the configuration of the lobes 

 of G. roperi varies considerably, although always conforming to 

 the characteristic of being separated on the median line. 



This variation of the occipital lobes is also prevalent in the 

 species G. quilensis, as a female specimen labelled as from the 

 Niger has the lobes much as in C. roperi, a species believed to be 

 only found in East Africa. 



Boettger also considered that the difference in size of the 

 lobes of C. dilepis and G. quilensis only entitles the latter to rank 

 as a variety of C. dilepis. I find, however, that the character is 

 well marked and constant, there being no connecting variations, 

 and I see no reason therefore for not regarding them as distinct 

 species. 



The type (C. isabellinus), which Giinther separated by reason 

 of " the large scutes of the occipital flaps and of the occiput," I 

 am forced to the conclusion is only an extreme variety of G. di- 

 lepis, on comparing it with the Museum series of the latter (which 

 has been considerably augmented since the describing of G. 

 isabellinus). The large scutes of the occipital lobes, to which 

 Giinther attaches so much importance, are in three longitudinal 

 rows, and from that number I find among the specimens of 

 C. dilepis a complete gradation to the typical form of four, five, 

 and six more or less well-defined rows. The general scaling also is 

 variable, being on some specimens flat and on others tubercular. 



C. Icevigatus, which has been confounded with the West 

 African G. senegalensis, is closely related to the species of the 

 first group, but entirely lacks the occipital lobes. 



The second group of three species (G. bitceniatus, C. ellioti, 

 G. hoehneli) is interesting, as it introduces the question of gular 



2c2 



