BIRDS OF ETHIOPIA AND KENYA COLONY 



393 



the differences between the adults, can be picked out and put into 

 their proper place." It would have been more to the point to say 

 that they could be picked out according to the preconceived notion 

 of the student and put where they best fitted that notion. I have 

 studied these birds without any preconceived ideas, and find that 

 after examinin<j them carefully (some 63 specimens in all) I have 

 no ideas. i)reconceived or otherwise, that enable me to do anything 

 definite with regard to subadult birds. The best I can do is to guess 

 and it happens that my guess is the same as that of Grant. I as- 

 sume, for instance, that the immature male listed above is granti 

 rather than P. purjmreus marwitzi because it has no green at all on 

 the nape, mantle, or breast, and that it is not a race of somaliensis^ 

 merely because the latter species is not known to occur as far south 

 as the Thika River, and because this specimen has the throat deep 

 purple, while two correspondingly young individuals from southern 

 Ethiopia, collected together with adults of somaliensls lack this 

 purple color. 



Zedlitz ^ considers somaliensis and damat'ensis conspecific, the 

 former being a race of the latter. This certainly seems improbable as 

 then granti and somaliciisls would be the same bird, but they are 

 not. Adults from British Somaliland, which, by locality, can not 

 possibly be anything but typical somaliensls, have black bills, while 

 adults of granti have red bills and are more and brighter greenish 

 on the breast, more violaceous above. Van Someren * writes that the 

 majority of adults of gi-anti have black and red bills, but that very 

 old individuals have bright red ones. A series of 14 granti examined 

 fails to corroborate this. All iuWy adult birds seen by me have red 

 bills. 



The size variations of adults may be judged from the following 

 table : 



^ Journ. f. Ornith., 1910, p. 778 ; 1915, p. X\ 

 * Nov. Zool., vol. 29, p. 82. 



