I9I91 



DeVRIES— OENOTHERA RUBRINERVIS 



1 9 10 and 19 1 5, and in that of two specimens of the fourth genera- 

 tion of 19 1 5. I found 53-68 per cent of germs, with a mean of 60 

 per cent. Among the specimens of deserens, quoted in the same 

 table, 5 belonged to this race; their seeds contained 96, 99, 94, 83, 

 and 58 per cent of good germs. Thus we see that the empty grains, 

 which are a character of 0. rubrinervis, have disappeared almost 

 wholly in the new mutant. 



My second strain of O. rubrinervis was derived from another 

 mutant of 1895. It has not been used for any crosses except those 

 mentioned in this article, and which served as control for the experi- 

 ments in the main Hne. Part of the seeds of 1895 were sowed in 

 1907 and yielded a second generation from which a third has been 

 derived in 19 13 and a fourth in 19 14. I counted the deserens for 

 three parent plants as previously described and found the per- 

 centages as follows: 



Percentage of 0. deserens in cultures of 0. rubrinervis 



Seeds of 



Culture 



Number of 

 specimens 



Percentage of 

 deserens 



Mean 



Second generation in 1910. 

 Third generation in 19 13. 

 Third generation in 1913. 



1913 

 1914 (A) 



1914 (B) 



25 

 70 

 70 



19 



The results agree exactly with those deduced from the previous 

 table. The suspicion, however, that in the two last cases the per- 

 centage figures were found too low, on account of losses of speci- 

 mens of deserens at the time of planting out, induced me to repeat 

 these sowings in 1916 from preserved seeds, giving them all the 

 care which the previous cultures and the first ones of 19 16 had 

 shown to be necessary. Moreover, in 19 13 I had self-fertilized a 

 third plant, besides the two mentioned in the table, and also sowed 

 its seeds. In this way I got in 191 6 the following percentages for 

 the seeds of the three self-fertilized plants of 1913: 



Percentages of 0. deserens in cultures of 0. rubrinervis, strain B 



