758 



THE AMEBIC AN NATURALIST [Vol. XLVIT 



controvertible evidence of a change in mean weight as a 

 result of selective mortality. But individually considered, 

 more differences in mean weight are from two to four or 

 more times their probable errors than can possibly be 

 attributed to experimental or sampling errors. Some of 

 these differences are positive, others are negative. There 

 seems in view of these facts, no escape from the conclu- 

 sion that there is a real biological relationship between 

 weight and viability of such a nature that in some experi- 

 ments the heavier and in other experiments the lighter 

 seeds are most heavily drawn upon in the mortality. 

 This seems clear from the greenhouse experiments in 

 whatever way the differences are taken. There are in- 

 dications of the same condition in field cultures, although 

 here the criterion of statistical trustworthiness is, because 

 of the dual errors of sampling, less dependable. 



There is strong evidence for varietal differences with 

 respect to mortality. In some strains the heavier, in 

 others the lighter, seeds seem less capable of develop- 

 ment. The reason for these differences may be sought 

 in the inherent characters of the stocks used or in the 

 environments to which they have been subjected. This 

 question is, however, so complicated that larger and more 

 diverse series of data must be gotten together for its 

 final solution. 



4. Consider now the variabilities. There can be no 

 question whatever concerning the reality of the reduction 

 in variability, either absolute or relative, as a result of 

 differential mortality. The following conditions seem to 

 prevail for the individual comparisons which may be 

 made. 



There is probably a reduction in absolute variability, 

 and there is certainly a reduction in relative variability, 

 in passing from seeds which fail to germinate to those 

 which produce abnormal seedlings. 



There is probably also a reduction in standard devia- 

 tion in weight in passing from seeds which give abnormal 

 seedlings to those which germinate normally. This re- 



