No. 564] INHERITANCE OF LEFT-IIANDEDNESS 



747 



This anaylsis of means by varieties is most suggestive. 

 Leaving out of account Flageolet Wax for which there is 

 only a single experiment, it appears that in Navy and 

 White Flageolet there is a distinct increase in mean 

 weight of survivors, that in Ne Plus Ultra and Burpee's 

 Stringless there is no marked change in mean weight, 

 while in Golden Wax there is a pronounced tendency for 

 the survivors to be lighter than the general population. 17 



It is clear that such a condition as this would give, 

 with a proper combination of strains, precisely the gen- 

 eral result that we have found for the means : that is, an 

 average of no change by selective elimination but signifi- 

 cantly positive differences in some experiments and sig- 

 nificantly negative differences in others. Here is a 

 problem requiring further analysis — which, however, can 

 be profitably undertaken only when larger bodies of 

 experimental evidence are at hand. 



Comparison of Absolute Variabilities 

 For the standard deviations for seeds germinating 

 normally and seeds failing to germinate, A-C, in the 

 whole material the results are: 



experiments are not sufficient to be of material importance. Much greater 

 weight is probably to be attached to certain experimental and biological 



measured in terms of capacity to produce mature fertile plants, while in the 



(normal or abnormal) germination only. 



it Possibly these results are due merely to the unavoidable errors of 

 experiment and of sampling. Only far wider series of data can settle this 

 point; until then no stress whatever is to be laid upon it. But a priori 

 there is nothing unreasonable or improbable in such results. These varieties 



prising in the indication that in one variety the death rate is more concen- 



