74 



WILLIS : OEOGRAPHIOAL DI8TRIBirTIO> 



the whole order, and we arrive at a surprisingly simple and 

 straightforward result. Hitherto we have tended to look upon 

 the geographical distribution of an order as a fact to be accepted 

 as a fact, but as little else, and not a fact that can be further 

 explained. So long as we accept the theory of infinitesimal 

 variations, this must necessarily be so to a large extent, for 

 if we suppose the origin of any three species, say, A, B, C, to 

 be represented by such a diagram as— 



Or 



we are necessarily in the dark as to what x and y were hke, 

 and can for a long time have no special hope of finding out. 

 But if we assume that the origin was — 



Or 



we simplify the problem enormously, and this simpUfication 

 alone makes a very great claim for the new theory, if it cannot 

 be shown to break down in any way in explaining geographical 

 distribution. 



The figures which we have just given explain the essential 

 difference between the two theories. On the theory of infini- 

 tesimal variations it is necessary to assume a vast amount 

 of destruction, while on that of mutation there is no need to 

 assume this, unless the mutations are actually harmful in the 

 struggle for existence. There is consequently no reason, so far 

 as we can tell, why the ivhole tree of a natural order sJwuld not 

 acttwUy exist on the earth at this present moment. If we apply 



