LIMITS OF NA TURAL SELECTION. 9 



indefinite variation, that is, a variation without dela- 

 tion to the needs of the organism or the demands 

 of the environment ; and consequently they believe 

 in a natural selection which takes note of the 

 minutest deviation of growth. To test this idea, 

 let us consider for a moment the millions of genera- 

 tions which compose the line of ancestry of the 

 mammals, and the inconceivable number of millions 

 of variations that have taken place along the line. 

 If the struggle for existence be so severe, and natural 

 selection so relentlessly active and acute as it is 

 claimed to be, then all these variations must have 

 been useful, or the mammal's ancestors would have 

 died — perhaps as the Devonian fishes or Silurian 

 invertebrates. Each one of this inconceivable num- 

 ber of variations must have occurred, not only at the 

 right period of geological time, but also at the right 

 moment in the life of the individual in whom it 

 first appeared : some of them occurred in embryonic 

 stages, some in adult stages ; many of them were 

 complex and correlated, that is, advantageous only 

 when accompanied by certain other variations. In 

 each case there could have been only one chance 

 for a useful variation against countless chances for 

 detrimental variations : for an organism is a delicate 

 piece of machinery, and while there is, perhaps, one 

 point where a slight change will make it more per- 



