240 THE VARIATION OF ANIMALS IN NATURE 



surface than on the under. (2) It is difficult to explain why 

 the predators which are characteristically nocturnal should 

 be effacingly coloured, and why the subterranean form (like 

 pocket gophers) should be ' desert-coloured.' (3) The habits 

 of certain desert birds seem to frustrate the advantage of their 

 coloration, as they come out to feed at sundown when their 

 shadows render them quite conspicuous. (4) The theory of 

 protective coloration cannot apply to animals (p. 168) which 

 appear to be without enemies. Buxton (pp. 168-70) con- 

 cludes by avowing the belief that the origin of desert colora- 

 tion ' will be eventually found by studying the effects of 

 physical conditions upon the animal life,' though he admits 

 that no factors hitherto studied (heat, etc.) can be responsible. 

 Buxton has put in a very forcible manner difficulties 

 voiced by other naturalists (e.g. Grinnell, Sumner). It is 

 true that other observers have emphasised special features in 

 desert coloration that seem to lend support to the ' protective ' 

 theory. Thus Cheesman (I.e. p. 316) points out that ' pro- 

 tectively coloured ' forms are found among the ground-breeding 

 birds and not among those which nest in holes (bee-eaters, 

 rollers, etc.). 



Colour of the lizard Anolis. 



Doflein (1908, p. 245) describes three species of Anolis that 

 are of very different colour living on the island of Martinique. 

 They live together, but if disturbed they dash off each 

 to different-coloured vegetation against which they are 

 invisible. He observed similar behaviour in two species of 

 grasshoppers (I.e. p. 246), and claims that there is a definitely 

 established type of flight instinct which leads such animals to 

 seek appropriately coloured backgrounds. These cases are 

 not worked out in any detail, and there is no statistical treat- 

 ment nor any intensive study of the behaviour. It is not 

 stated how far the natural enemies are deceived. 



Coral fishes. 



Reighard (1908) made an extended series of observations 

 and experiments designed to elucidate the significance of the 

 bright colour and striking patterns of twenty- two species of coral- 

 reef fishes in relation to the attacks of one of their habitual 

 enemies, the Grey Snapper (Lutianus griseus) . It was found that 



