80 THE GENESIS OF SPECIES. [CHAP. 



Selection " only it is so improbable as to be practically im- 

 possible for two exactly-similar structures to have ever 

 been independently developed. It is so because the num- 

 ber of possible variations is indefinitely great, and it is 

 therefore an indefinitely great number to one against a 

 similar series of variations occurring and being similarly 

 preserved in any two independent instances. 



The difficulty here asserted applies, however, only to 

 pure Darwinism, which makes use only of indirect modifi- 

 cations through the survival of the fittest. 



Other theories (for example, that of Mr. Herbert Spen- 

 cer) admit the direct action of conditions upon animals and 

 plants in ways not yet fully understood there being con- 

 ceived to be at the same time a certain peculiar but limited 

 power of response and adaptation in each animal and plant 

 so acted on. Such theories have not to contend against 

 the difficulty proposed, and it is here urged that even very 

 complex extremely similar structures have again and again 

 been developed quite independently one of the other, and 

 this because the process has taken place not by merely 

 haphazard, indefinite variations in all directions, but by the 

 concurrence of some other and internal natural law or laws 

 cooperating with external influences and with "Natural 

 Selection " in the evolution of organic forms. 



It must never be forgotten that to admit any such con- 

 stant operation of any such unknown natural cause is to 

 deny the purely Darwinian theory, which relies upon the 

 survival of the fittest by means of minute fortuitous indefi- 

 nite variations. 



Among many other obligations which the author has 

 to acknowledge to Prof. Huxley are, the pointing out of 

 this very difficulty, and the calling his attention to the 

 striking resemblance between certain teeth of the dog and 

 of the thylacine as one instance, and certain ornithic pe- 

 culiarities of pterodactyls as another. 



