340 



BOTANICAL GAZETTE 



[NOVEMBER 



The rubrinervis strain for these experiments had arisen as a 



mutant from O. Lamarckiana in 1895, and its second generation was 



cultivated in 1 905 . No dwarfs were produced in the first generation 



after the crosses, and in the second only from single individuals, the 



remainder giving either no dwarfs at all or only about i per cent, 



by ordinary mutation. 



TABLE I 



Exceptional production of dwarfs by single plants of Oenothera subrobusta 



If we compare these figures with the results of the crosses 

 between 0. rubrinervis and 0. nanella itself, as described in my 

 Gruppenweise Artbildung (p. 215), we find a complete analogy, 

 since these crosses give no dwarfs in the first generation, and in the 

 second about 10-14 per cent from the self-fertilized specimens of 

 0. subrobusta. It is evident, therefore, that the exceptionally high 

 yield of dwarfs in these crosses of O. Lamarckiana and O. rubrinervis 

 must be the product of latent mutations which occurred in some of 

 the sexual cells of one of the parents. And since 0. Lamarckiana 

 is known to produce ordinarily 1-2 per cent dwarfs, while O. 

 rubrinervis does not show signs of such a mutabihty, we may con- 

 fidently assume that our figures indicate latent mutations of sexual 

 cells of O. Lamarckiana. 



Bartlett* recently described a similar instance of an unexpect- 

 edly high mutabihty, and proposed for it the same explanation, on 

 the assumption of a latent mutation of a sexual cell in a previous 

 generation. This case is of the greatest interest since it relates to 

 a pure species and not to the discovery of mutated gametes by 

 means of crosses as in the experiments just described. The mutat- 

 ing species was O. Reynoldsii Bartlett, one of the forms of the old 

 0. biennis. It produced in 1913 three marked types, one repeating 

 the parental form, and the two others being dwarfs and called 



' Bartlett, H. H., Mutation en masse: Amer. Nat. 1915. 



