746 



THE AMERICAN NATURALIST [Vol. XL VII 



Somewhere between this minimum value and the one 

 given above by (A -f- B) - C probably lies the true meas- 

 ure of the change in mean weight as it would occur if the 

 plants were required (as they would be in nature) to 

 grow to reproductive maturity. 



I now turn to the individual varieties. This demands, 

 for results which shall be at all trustworthy, the com- 

 bination of both sets of experiments. 16 



The accompanying table gives the results for the rela- 

 tive differences in mean weight (differences expressed 

 as percentages of the general population constant). 



