FACTORS IN THE COLORATION OF ANIMALS 195 



had opportunities of leaving descendants, the original characters 

 will somehow sooner or later reappear. 



Then, from the point of view of natural selection, there is no 

 good reason why there should be such a sharp contrast of pigments 

 on the flank of the black Buck, and many other animals ; but, from 

 the point of view of a hereditary impression^ there seems very good 

 reason why the pigments which picture the separation of the 

 formerly armoured and unarmoured surfaces should be contrasted. 

 It seems to represent the line of division which we now see be- 

 tween the back and abdomen of the Armadillo (Fig. 71), and of 

 some Pangolins. The white or differently coloured abdomen would 

 then mean that the abdominal surface lost its armour, and may be 

 also its spotting, long before its ancestral forms lost their carapace. 



This theory of the origin of contrasted colours is based on the 

 early disappearance of armour from certain parts of the surface 

 while it continued to a later period on certain others ; and the 

 contrast of pigments, when finally the armour wholly disappeared, 

 represents the two stages. It does not attempt to account for 

 every bit of colour, but only for those colorations which appear 

 to be formed on a. plan. 



To put this idea more clearly before the reader The repetition 

 through ages of the nerve-centre action which resulted in the forma- 

 tion of dermal bone-rosettes, left, as it were, a memory in the nerve- 

 centre. This memory, when the dermal plates could no longer be 

 produced, externated itself by producing only /zV/^r^-rosettes, as 

 we see them in the Jaguar. When even the memory of separate 

 rosettes had been effaced, it still continued to externate itself by 

 producing a contrasted pigmentation of ventral and dorsal surfaces ; 

 that is, of the sites of the earlier and the later loss of armour. 



Among existing animals we have survivals of all the stages 

 that mammals have gone through in their evolution. We have 



