742 THE AMERICAN NATURALIST [Vol. XLVII 



to be heavier than those which fail. Sometimes, the gen- 

 eral average is positive and sometimes it is negative in 

 sign; it is always insignificant in magnitude. Nor, to 



return to the question of the more qualitative classifica- 

 tion of the experiments, can any great weight be attached 

 to such inequalities in the number of positive and nega- 

 tive differences as we have secured. 



The mean values of the ratios of the differences, A-C, 

 to their probable errors have also been struck. The 21 



