692 THE AMEBIC AN NATURALIST [Vol. XLVII 



TABLE III 



Weight of Seeds More or Less Abnormal IN Germination 

































53 



NHD-M 



NHH-J . . . .......... 



_ 



\ 









\;. 





3 





_ 



_ 



89 



NHH-M 



— 





— 

















2 



148 



NHHH-M . ' . . ....... 





z 





I 



10 17 



3 6 

 Jil 



2 2 

 7 













NHDD-J 















5 



- 







54 













12 24 



5 







z 



z 



57 

 112 



NDH-E '. . 



— 



— 



1 



17 27 40 

 4 9 19 



26 



10 



4 



i 



z 



E 



51 



NDD-E 



z 



z 





3 10 9 



10 





1 









38 



NDDD-D 



NDDD-E 







2 







\ 





I 







30 



NDHH-D 









22 14 



11 













62 



NDHH-E 



FSS-I 













l 











41 



67 



FSS-L 

























FSH-C 



FSD-C 























126 



FSHH-C 























52 



FSDD-C 























47 













lec 



ti\ 



e 



mortality must 



rest upon a comparison of these constants. 



The method of making these tests demands a word of 

 explanation. In the previous study, the comparison was 

 necessarily drawn between the constants of the seeds 

 which actually produced fertile plants and those of the 

 general population from which they were drawn; the 

 constant for the general population was subtracted from 

 that of the sub-sample. The positive or the negative sign 

 of the difference showed whether mortality had tended to 

 raise or to lower mean or variability. 



In these greenhouse experiments, on the other hand, we 

 have the constants for samples (A) normally germi- 

 nating, (B) germinating abnormally and (C) failing to 

 germinate. (B) may possibly be regarded as interme- 

 diate between (A) and (C). 



If we take the difference between the constants 

 Survivors less failed 

 we shall have plus differences of the mean if selection has 

 tended to raise the general average by eliminating the 



