NATURAL SELECTION 239 



live. The general result is rather obscure. He says (p. 97) : 

 ' For the most part the larks harmonise fairly closely with the 

 ground upon which they were shot, but the same subspecies 

 may be found on two or more soils widely differing in colour 

 and composition but still matching closely the plumage of the 

 Crested Lark' — e.g. G. cristata carthaginis was shot on pinkish 

 buff soils and drab grey soils ' ... on each surface the . . . 

 Lark was practically invisible to the eye.' He notes that 

 another form, G. theklae harterti, was not nearly so difficult to see, 

 contrasting with the dark soil on which it was shot. He goes 

 on to make the important observation (p. 98) that in winter 

 these birds move about and are often found on soils which they 

 do not resemble so closely, the same subspecies being found on 

 several differently coloured soils ; and the same statement is 

 made by Rothschild and Hartert (191 5). 



Much of Bannerman's evidence does show that some of 

 these races (see especially p. 98, on G. theklae hilgerti) resemble 

 very closely the soils on which they were shot. It is similarly 

 clear that the coloration tends to be of a generalised tint, so 

 that the owner is invisible on more than one soil (p. 97). It 

 seems, however, that the birds are sometimes found on soils on 

 which they are conspicuous, and that there is no very definite 

 preference for soils with which they harmonise. We do not 

 think it is possible to say more than this — that, as far as the 

 human observer is concerned, there is probably a definite 

 concealing value in the colours of these birds, but we do not 

 know how far the natural enemies are deceived. 



The colours of desert animals considered generally. 



This problem has been studied with great fullness by 

 Buxton (1923, chapter vii). He first of all insists on the 

 general resemblance of the desert fauna to its background and 

 (quoting Meinertzhagen and others) of particular desert races 

 and species to particular shades of sand, and he admits that 

 their colour does in fact tend to make such animals difficult to 

 see as long as they remain motionless. He next alludes to 

 certain exceptions {e.g. black forms — Tenebrionidae, chafers, 

 ravens, wheatears, chats). He then proceeds to pose a set 

 of difficult questions. (1) A predacious bird like the desert 

 Merlin, which, as it hunts on the wing, should be ' effacingly ' 

 coloured on the under-side, is coloured paler on the upper 



