30 Mr. 0. Thomas on 



2. Petrodromus (Mesoctenus) rovumce, Thos. 



<?. 206, 207, 208, 214, 220; ? . 209, 210,211, 212, 215, 

 216, 219, 221, 222, 223, 224, 225, 226. 



This fine series of a species hitherto very insufficiently 

 represented is of particular importance, as some doubt 

 appeared possible in regard to the relation of the thickened 

 tail-bristles to the age of the individual — a point on which its 

 distinction as representing a special subgenus mainly depended. 

 For it might have proved that in old age the bristles of this 

 species became as knobby as they are in the subgenus Cerco- 

 ctemis. Now, however, I am able to record that not in the 

 oldest specimens do the bristles become like those of P. sultan, 

 the type of Cercoctenvs, while, on the other hand, every 

 individual that has its permanent teeth in place has some 

 thickened bristles present, none occurring in true restricted 

 Petrodromus. Certainly the bristles do increase in number 

 and knobbiness with age, but they never equal those of 

 Cerccctenus. The bristles of the males also seem to become 

 in old age more knobby than those of the females. 



The palatal vacuities are in most cases of considerable size, 

 but in seme individuals are almost completely absent, so that 

 there is no absolute constancy in the character, although it 

 has undoubtedly a certain average value. 



It was largely on this character that I separated mossam- 

 bicus of Cabaceira from rovumce of the Rovuma River; but 

 it now appears that the character is not to be trusted when 

 only individual specimens are available. 



As to the other character of rnossambicus (the slaty grey on 

 the belly-hairs) there is in this series a most surprising and 

 abnormal range of variation — from none at all to cases where 

 each hair is broadly slaty at base. In consequence, I think 

 that the name nioasambicus should be withdrawn and all of 

 these southern forms of JiJesoctenus should be referred to 

 rovumce. It is, of course, still possible that a subspecific 

 difference in colour may prove to exist when good Rovuma 

 skins are available, but for the present the name rovumce 

 should be used for all. 



3. Mangos rnossambicus, Matsch. 



S . 204, 235 ; ? . 205. 



Practically topotypes, the name having been based on a 

 specimen obtained by Peters at Cabaceira. 



