210 



Gates. 



which appear also in these crosses, are not the result of the recombi- 

 nation or blending of characters in the hybrids, but are independent 

 phenomena having a different cause. Mendelians have been particu- 

 larly prone to claim that the mutation phenomena in Oenothera La- 

 marckiana are merely an example of complex Mendelian splitting with 

 various suggested complications. But this claim has almost invariably 

 been made without any direct evidence in its support. 



Notwithstanding the large amount of work which has been devoted 

 to the Oenotheras since the original investigations of de Vries, we have 

 been with.out an adequate theory or explanation of the process of 

 mutation. With the accumulation of cytological facts, however, it has 

 become increasingly evident that the behaviour of the chromosomes 

 must be included in any attempted explanation. It is now not too 

 much to say that the cytological investigations begun by myself in 

 1905 and since taken up also by a number of others, have furnished 

 a key to an explanation of the nature of the mutation behaviour in 

 these forms. The cytological evidence shows that germinal changes 

 may and do occur which are independent of all the laws of hybrid 

 combination and hybrid splitting. This generalization is of more 

 fundamental significance than might at first appear; for it shows 

 that mutation in Oenothera is a process sui generis, and that no 

 amount of hybrid combination and splitting, Mendelian or otherwise, 

 is sufficient to account for it. 



At the present moment we are not concerned with the value of 

 mutation as a general factor in evolution. This paper is concerned 

 only with the nature of the mutation process as observed in Oenothera, 

 in comparison with the hybridization results. 



The pre-mutation hypothesis of de Vries was a purely formal 

 attempt, in the absence of definite evidence, to picture the process by 

 which new forms might suddenly arise through germinal changes. 

 That hypothesis, together with the equally formal Mendelian sug- 

 gestions, is now superseded by the definite facts derived from com- 

 bined cytological and breeding investigations. Instead of vague and 

 formal hypotheses we can now deal with concrete facts. 



Any theory of the mutation process must explain the following 

 phenomena: (i) The sporadic and repeated appearance of the same 

 mutant types from a variety' of parents, including 0. Lamarckiana 

 and various of its mutants. (2) The sporadic appearance of individuals 

 in crosses between species and mutants, in which the mutant cha- 

 racters are combined with characters derived directly from the parents 



