12 FRIEDA COBB 



character flatness vs. revoluteness is one which is evident in even very 

 young seedlings, many sowings were made merely for the sake of recording 

 the nature of the seedlings, and the seedlings were then discarded. In 

 order to give all possible space and attention to the cultures intended for 

 the summer garden at the time which is most favorable for planting, the 

 sowings for seedling counts were made early and got out of the way. This 

 meant that the seeds did not have as long a rest period as they apparently 

 need before germination. Had they been planted a month or two later 

 (in February and March instead of December and January) the percent- 

 ages of germination would have been very much higher. 



In problems connected with comparative fertility and sterility it is 

 of course quite necessary to know that all viable seeds are forced to ger- 

 minate. The method worked out by DE VRIES (1915 a) and applied by 

 DAVIS (1915) for forcing germination to completion, is especially suited 

 to such problems. It will also throw light on the types that may be lost 

 through selective mortality and selective germination rate when less 

 thorough methods are used. This information is most valuable. But 

 that the low germination percentage has no significant effect on the results 

 of the experiments recorded here is shown by table 30. Here the ratios 

 of flat-leaved to revolute-leaved plants are assembled from all of those 

 cultures in which the expected ratio is 3:1, and arranged according to the 

 germination percentage. It will be seen that three-fourths of the cultures 

 have a germination percentage under 25 percent. But it may also be 

 seen that the average of the ratios of cultures in which the germination is 

 from 50 to 81 percent is no nearer the expected 3:1 ratio than are the 

 averages from cultures with poorer germination. Even when less than 5 

 percent of the seeds germinated, the average is as close to the theoretical 

 ratio as is that of any one of the five cultures with a germination over 50 

 percent, or as the average of these five cultures. 



That selective germination, in connection with the types with which this 

 problem is concerned, occurs to any significant extent seems impossible. 

 The only evidence of such selection is a slight excess of revolute-leaved 

 individuals, especially noticeable when the germination is poorer. There 

 are four tables which show in the total an excess of flat-leaved plants, ten 

 which show an excess of revolute-leaved plants. In work with this species 

 it has been noticed that when the percentage of germination is very low 

 the percentage of mutations is very high (BARTLETT 1915 a). General 

 observations lead to the conclusion that poor germination tends to bring 

 the unusual types into prominence rather than to conceal them. 



