62 



A. FRANKLIN SHULL 



compute the standard deviation of laying, since in the four exper- 

 iments to be compared the number of days on which the eggs 

 were obtained does not vary so greatly. 



To test whether the viability of the eggs could be altered by 

 selection, two parthenogenetic lines were started from fertilized 

 eggs of the lot (C X D) x (C X D). Of the 54 females laying 

 these eggs, twelve were kept isolated and the hatching of their 

 eggs observed separately. There was considerable variation in 

 the proportion of the various famihes that hatched. The least 

 viable family hatched 4 out of 18 eggs, or 22.2 per cent of the 

 family. From the first of the 4 that hatched a parthenogenetic 

 hne was reared, as in table 10. The most viable family hatched 

 11 out of 15, or 73.3 per cent. From the first of the 11 that 

 hatched a parthenogenetic line was reared, as in table 11. The 



TABLE 9 

 Showing number of eggs that hatched, out of 904 in lot {C X D) X {C X D), figure 1 



Percentage of viable eggs. 



47.0 



Standard deviation: o-h = 2.30 



T.\BLE 10 



Showing details of parthenogenetic line derived from low viability family in 

 (C X D) X (C X D), figure 1 



Percentage of cf 9 



37.0 



