204 



MASS MUTATIONS AND TWIN HYBRIDS OF 



the boxes almost daily, it had been ascertained that it was always 

 the pale ones which died, whereas the green seedlings grew without 

 trouble. Thus I was confident that the dead had been ochracea, as 

 wel as the surviving pale ones. The number of the decayed was 

 derived from the number of specimens at the time of planting out, 

 minus the number of survivals. The results of the count on April 

 25, 1915, are given in table II. 



Table II 



The seedlings of the remaining five parents of 1914 were not 

 counted in April, but at the time of flowering in August. Each of 

 the groups yielded a large number of ochracea and one or more 

 lorea, but the percentage for the first was now only 12. This figure 

 is evidently due to the losses mentioned, since even during the 

 summer usually many specimens of the type of ochracea are lost 

 on account of their weakness. In 1916 I got about the same per- 

 centage at the time of flowering, but did not estimate the losses 

 during the spring. 



For this reason I repeated the sowing in the spring of 1917 with 

 the preserved seeds of the same self-fertilized individual of 1915, 

 taking every possible care to avoid the presumed losses. I planted 

 out 70 seedlings, only one of which died; 20 were found in May to 

 be ochracea, giving a total percentage of 30. This figure, therefore, 

 should be substituted in table I for the 15 per cent given for the 

 fourth generation. 



Moreover, in 1917 I sowed the seeds of four other self-fertilized 

 individuals of 1915, taking the same precautions. The culture 

 embraced 224 seedlings, of which 8 per cent were pale and weak and 

 died soon after being transplanted, while 31 per cent were recog- 

 nized in May as ochracea. This gives a total of 39 per cent, which 

 corresponds to the figures found in the best of the previous trials. 



