The Mechanism of Adaptation 347 



considered to be one of the original properties of life, and 

 our inability to explain its origin is not different from our 

 inability to explain the origin of metabolism, reproduc- 

 tion, irritability, or of life itself. Thus the simple principle 

 of overproduction and the elimination of the injurious or 

 unfit, whether individuals, cells, or physiological responses, 

 would offer a possible mechanical explanation of both racial 

 and individual adaptations and of the almost universal oc- 

 currence of fitness in the living world. 



There are, however, certain difficulties which such an 

 explanation encounters, and there is always left an unex- 

 plained remainder which for the present at least is inexpli- 

 cable on purely mechanical principles. One of the most 

 serious of these difficulties is that the rate of adaptation 

 does not appear to be proportional to the rate of overpro- 

 duction and elimination, as it should be if these are the 

 only causes of adaptation. The rate of adaptation can be 

 measured by the rate of divergent evolution, that is, adap- 

 tations in different directions, or by the relative complexity 

 and perfection of corresponding adaptations in two groups 

 of approximately equal age. Measured in either of these 

 ways, we find that the rate and degree of differentiation 

 and adaptation are not always proportional to the amount 

 of overproduction and elimination. The rate of repro- 

 duction and of elimination is lowest in some genera and 

 classes in which adaptations are most varied and perfect. 

 For example, compare the rate of reproduction and elimi- 

 nation in lower animals and plants with that in higher ones. 

 If the variety or complexity of adaptations is dependent 

 entirely upon these two factors, why do not bacteria and 

 protozoa have the most numerous and complex adaptations 

 of all organisms; or why do not mice greatly surpass ele- 

 phants in these respects; or grasses, sequoias; or why is not 



