64 



A. FRANKLIN SHULL 



Attention may now be called to the' fact that although the 

 eggs oi D X C showed lower viabilit}' and greater variability in 

 the duration of the egg stage than did those of C X -D, yet when 

 the parthenogenetic lines from each were inbred, the eggs in 

 lot {D X C) X {D X C) were a little more viable and a little 

 less variable in their hatching time than were those of {C X D) 

 X (C X D). The differences are small, however, as compared 

 with the differences between D X C and C X D, and are prob- 

 ably insignificant. The inequality of the reciprocal crosses dis- 

 appears when the parthenogenetic lines derived from them are 

 compared. Further evidence that the reciprocal crosses C X D 

 and D X C are not after all unequal is found in the reciprocal 

 crosses between the reciprocal crosses, now to be described. 



(C X D) X {D X C). Females from the parthenogenetic line 

 C X D were paired with males from line D X C No^'ember 9 

 and 10, 1911. From 20 matings thus produced, 278 eggs were 

 obtained. Up to December 20, 174 of these eggs, or 62.5 per 

 cent, had hatched (table 13). The duration of the egg stage 

 is not greatly variable, hence the standard deviation is low 

 (<x = 2.42). 



{D X C) X (C X D). Females of the line D X C were mated 

 with males of the line C X D November 7 to 10, 1911. From 

 23 such matings were secured 357 eggs. Up to December 20, 



TABLE 13 

 Showing number of eggs that hatched, out of 278 in {C X D) X (D X C), figure 1 



Percentage of viable eggs. 



Standard deviation: a^ = 2.42 



62.5 



