368 THE EVIDENCE OF EVOLUTION. 



proposition which everybody knows to be impossible. Many highly 

 improved races of forage plants and agricultural crops have been 

 produced by intelligent breeders simply on the ground of these 

 always available dissimilarities. They can be selected and accumu- 

 lated, augmented and heaped up, until the new race is distinctly 

 preferable to the original strain. 



In ordinary agricultural breeding, however, it is very difficult to 

 distinguish sharply between these two principles. Moreover, for 

 practical purposes, this distinction has no definite use. The practise 

 of selection is nearly the same in both cases, and, besides hybridizing, 

 with which we are not now concerned, selection is as yet practically 

 the only means for the breeder to improve his races. Hence it came 

 that at Darwin's time there was no clear distinction between the 

 two types of variations, at least not to such an extent that a theory 

 of the origin of species could confidently rely upon it. 



Quetelet's celebrated law of variability was published only some 

 years after the appearance of Darwin's 'Origin of Species.' Varia- 

 bility seemed until then to be free from laws, and nearly everything 

 could be ascribed to it or explained by it. But the renowned Belgian 

 scientist showed that it obeys laws exactly in the same way as the 

 remainder of the phenomena of nature. The law which rules it is 

 the law of probability, and according to this law the occurrence of 

 variations, their frequency and their degree of deviation can be 

 calculated and predicted with the same certainty as the chance of 

 death, of murders, of fires and of all those broad phenomena with 

 which the science of sociology and the practise of insurance are con- 

 cerned. 



The calculations of probable variations based on this most im- 

 portant law did not, however, respond to the demands of evolution. 

 Specific characters are usually sharply defined against one another. 

 They are new and separate units more often than different degrees 

 of the same qualities. Only with such, however, Quetelet's law is 

 concerned. It explains the degrees, but not the origin, of new pecu- 

 liarities. Moreover, the degrees of deviation are subject to reversion 

 to mediocrity, always more or less returning in the progeny to the 

 previous state. Species, on the contrary, are usually constant and 

 do not commonly or readily revert into one another. It is assumed 

 that from time to time specific reversions occur, but they are too 

 rare to be comparable with the phenomena which are ruled by the 

 law of probability. 



A thorough study of Quetelet's law would no doubt at once have 



