700 



THE AMERICAN NATURALIST [Vol. XL V 



as give one a low or a high hand at cards. It is then 

 allowable to consider the biological interpretation of the 

 differences. 



3. Two tests for differentiation were applied: {a) the 

 mean intra-line variability was compared with the popu- 

 lation variability, and (b) the significance of the devia- 

 tion of individual line means from the population mean 

 was tested by a coefficient of individual prepotency re- 

 cently suggested. Both of these tests indicate sensible 

 and statistically significant differences between the lines. 

 These differences may be said to be distributed according 

 to "Quetelet's Law" as the term is loosely used by 

 biologists. 



4. This fact per se furnishes no evidence at all for the 

 genotypic nature of the differences in Eoemer's lines. 

 Indeed, throughout Eoemer's work there is no conclusive 

 evidence of any kind concerning any problem of heredity. 

 At least one (and possibly both) of his series of material 

 is from his own explicit statements in reality a pure line. 

 The difference observed within these lines and considered 

 by him and other pure linists to he of genotypic value and 

 a confirmation of Johannsen's results with beans are 

 probably merely the result of faulty experimental condi- 

 tions. If they are not, Eoemer's evidence goes squarely 

 against Johannsen's theory. 



July 19, 1911. 



