FACTORS IN THE COLORATION OF ANIMALS 197 



animals by * natural selection ' would be an idle task. No amount 

 of natural selection, for instance, would account for the fact of 

 so many mammals having ringed tails and plain bodies (see 

 Appendix E. Nos. 4, 7, n,etc.); for we cannot for one moment 

 suppose that it was the tail alone that required protection ! In 

 such cases we can only fall back on the supposition that, for 

 some reason unknown, the tail has been slower in losing the 

 modified rosettes than the body. 



As happens with most theories, to pursue this one further 

 would entangle it in difficulties, as so many modifications in the 

 coloration of mammals have occurred since their ancestors wholly 

 lost their armour. It is only by an occasional survival here and 

 there that some sort of concatenation can be conjectured to link 

 the armoured stage of mammals with the existing stage, which 

 in most cases is totally different. 



The upshot of this discussion is that we come to the con- 

 clusion that the rosettes of Leopards are the inherited imprints 

 of armour-clad ancestors, and not the result of natural selection 

 by slow and useful modifications ; and that their strange colora- 

 tion has been preserved by natural selection because it may have 

 harmonised with the surroundings that these animals frequented. 



