THE ORIGIN OF VARIATION 27 



between individuals of markedly contrasted genetic constitu- 

 tion. In view of this, Lotsy's speculations as to evolution by 

 crossing appear unlikely to have a wide application in the 

 animal kingdom. There is a further difficulty in the way of 

 Lotsy's theory. If it has been something more than a minor 

 factor, we would have to admit that all the material of variation 

 was in existence in the earliest forms of life, and evolution has 

 consisted in the allocation to the forms which diverged from an 

 ancestral stock of various portions of this fund of material and 

 the recombination of parts of it to form new genetic groupings. 

 That a good deal of factorial recombination (with the appear- 

 ance of ' novelties ' due to this cause) has taken place we 

 do not doubt. But, if recombination is the only or even the 

 main source of variation, we have to imagine evolution as 

 merely the revelation of latent possibilities — a picture very 

 difficult to harmonise with the facts, for, looked at in the 

 broadest way, evolution undoubtedly leaves the impression of 

 the continuous emergence of new types of organisms. Thus, 

 while recombination has an obvious importance in trying 

 out all the permutations of the material lying to hand, we feel 

 the need of another process which will provide new material. 



4. Gene-mutations 



In spite of the vast amount of genetical research carried 

 out during the past thirty years our knowledge of the origin 

 of gene-mutations is still extremely slight. In the first place, 

 if a given variant is a mutation and not merely a recom- 

 bination, it should appear suddenly in an inbred stock. Thus 

 only in very quick-breeding forms can much information be 

 accumulated. 



In the second place a distinction must be made between 

 agencies which actually produce mutations and those which 

 accelerate mutation-rates. We may illustrate this distinction 

 by recalling the effect of temperature on growth in inverte- 

 brates. Here, while differentiation, within wide limits, proceeds 

 independently of temperature, the actual rate at which it 

 goes on is directly dependent. 



In actual practice there is no known treatment which 

 regularly produces a high proportion of any definite type of 

 mutant. Such agencies as X-rays induce variation in all 



