268 RAYMOND PEARL 



From table 10 we derive the following series of differences and 

 their probable errors in respect to mean hatching weight of the 

 chicks in the different series : 



Male chicks 



(Treated & X untreated 9 9) - (controls) = +0.67±0.47 gram. 

 (Treated cf X treated 9 9) - (controls) = +2.73±0.58 gram. (X 4.7) 

 (Treated cf X treated 9 9) - (treated d' X untreated 9 9) = +2.06±0.54 



(X 3.8) 

 Female chicks 



(Treated cf X untreated 9 9) - (controls) = +0.31 ±0.35 gram. 

 (Treated cf X treated 9 9) - (controls) = +2.44±0.38 gram. (X 6.4) 

 (Treated cT X treated 9 9) - (treated cf X untreated 9 9) = +2.13±0.39 



(X 5.5) 

 Sex differences 



Treated cf X untreated 9 9 series: 9 — & = 0.13 ±0.39 gram. 

 Treated cf X treated 9 9 series: 9 - cf = 0.20±0.54 gram. 

 Control series: 9 — c?' = 0.49±0.44 gram. 



From these figures the following conclusions appear to be 

 warranted : 



1. In the present series of experiments there is no significant 

 difference in mean hatching weight between the offspring of 

 treated males and the offspring of normal untreated control 

 males when both are mated to normal untreated females. The 

 slight differences which do appear are of the same order of 

 magnitude as their probable errors. 



2. Both the male and the female offspring of matings in which 

 both parents were treated have a larger mean hatching weight 

 (i.e., are heavier when hatched) than the offspring of either 

 completely normal control matings, or of matings in which the 

 father only is treated. The differences here are relatively large 

 and are statistically significant in comparison with their probable 

 errors. The figures in parentheses following the difference lines 

 give the ratio Diff./P. E. diff. These figures range from 3.8 to 

 6.4. From the table given by Pearl and Miner (24, p. 88) it 

 appears that the odds against the fortuitous occurrence of devia- 

 tions as great as or greater than these range from about 95 to 

 1 to about 20.000 to 1. They may fairly be considered real and 

 significant differences. 



