1915] BARTLETT—MUTATION IN OENOTHERA III 



ratio of nummularia mutations to plants showed a variation roughly 

 commensurate with the difference in germinabiUty between the 

 Fi and F2 seeds. In other words, the mortaUty among the year- 

 old Fi seeds appears to have been largely confined to seeds of 

 typical 0. pratincola. The ratio of nummularia mutations to seeds 

 planted is seen from table XII to be reasonably constant for all 7 

 strains in both the Fi and F2 generations. The ratio of mutations 

 to total plants, however, varies between wide limits, and in every 

 case a low percentage of germination is associated with a high 

 frequency of mutation. The Fi progeny of Lexington A, for 

 example, included 4 individuals of mut. nummularia among 255 

 plants, a ratio of 1:64. These 255 plants, however, were obtained 

 by sowing 1,083 seeds, of which only a small proportion (23.5 

 per cent) germinated. There seems no escape from the conclusion 

 that the percentage of germinable seeds of mut. nummularia had 

 increased by virtue of the greater mortaUty among the seeds of 

 typical 0. pratincola. 



The evolutionary significance of differential mortaUty is too 

 obvious to require any lengthy discussion. Mut. nummularia has a 

 distinctly greater survival value than its parent when subjected 

 to conditions which delay germination. It has already been shown 

 that mut. nummularia has an enormously greater chance 

 to survive than typical 0. pratincola when subjected to certain 

 unfavorable soil conditions. These facts should be carefully 

 weighed by critics of the mutation theory who persist in assimiing, 

 as a matter of course, that mutations would have no chance to 

 survive in competition with the more numerous typical plants. 

 De Vries^° has already shown that the percentage of mutation 

 in a culture of 0. Lamarckiana from seeds 5 years old was 40 per cent 

 instead of the usual 6 per cent. In his comment on this remark- 

 able result he states that in general the seeds of the mutation 

 remain germinable longer than those of typical 0. Lamarckiana, 

 and suggests that it might be possible to make use of differential 

 mortality to increase the proportion of mutations in seeds, and 

 thereby to faciUtate the discovery of the mutations. The writer 

 unconsciously put this suggestion to a test at the time mutations 



2° De Viues, H., Die Mutationstheorie i : 186. 1901. 



