new South-African Sylviidse. 205 



mination by Messrs. Layard, Gurney, Verreaux, Sharpe, and 

 others, I have been somewhat appalled at the very unsatisfactory 

 state of our knowledge of the group Saxicolina. It is neces- 

 sary to observe that in no class of birds is a large series of spe- 

 cimens so absolutely indispensable for the accurate discrimina- 

 tion of species. Many of the characteristics which have been 

 relied upon by closet naturalists as specific distinctions will be 

 found to be merely variations of sex or age — and this in cha- 

 racteristics which, in allied genera, are infallibly distinctive. 

 For instance, in some dark-coloured species, the coloration of 

 the head may be black, grey, or white — and this, so far as we 

 can discover, simply from age, not sexual, and occurring in 

 breeding birds in the same locality. This fact was brought 

 prominently to my notice in the case of the North-African 

 groups, both in the Sahara and in Palestine. I find it holds 

 equally in the analogous species both from Scinde and from 

 South Africa. I may mention, as cases in point, the variations 

 in Saxicola eurymelana and S. monticola. In the same way a 

 comparison of a large series will show us that the proportion of 

 white and black on the rectrices is very variable in individuals 

 of the same species in some of the desert groups. 



My examination of the series sent by Mr. Layard leads me at 

 once to reject the specific value of Saxicola castor, Hartl. (P. Z. S. 

 1865, p. 747), which appears to me to be only one of the varia- 

 tions of S. cinerea, Vieillot, a species that has as many different 

 phases of plumage as S. monticola, the young birds being rusty- 

 brown, then blackish-brown, and finally assuming the uniformly 

 cinereous plumage. 



There are, however, in all the Chats some invariable points of 

 distinction ; among these I attach the chief value to the colora- 

 tion of the rump, and to its extent, which appears to be invari- 

 able at all ages, presenting no sexual variations in the subgenus 

 DromolxEa, generally differing in the sexes of the desert group of 

 the Saxicolina. 



Thus, in Saxicola monticola, we have at all ages the white 

 epaulettes in the male, and in both sexes the narrow white rump ; 

 but, according to age, we find specimens otherwise wholly cine- 

 reous excepting their rcmiges and rectrices, others black, with an 



