754 



PROCEEDINGS OP THE AMERICAN ACADEMY. 



lines is shown separately in Figure 4, A and A", M and 3f, N and N 7 . 

 The two lines of each series pass through similar cycles, though not quite 

 synchronously ; yet the two lines of each series were reared synchronously. 

 Thus, A reaches a maximum two generations (six or eight weeks) later 

 than A" ; il/ culminates a generation later than M' ; and N' culminates 

 four generations later than N. The last case is particularly interesting 

 because in this series alone of the three the attempt was made to select 



Average 

 Ho. Young 



Generation 50 52 54 50 5S 



21 23 25 27 



.19 21 



Figure 4. Productiveness of the two lines maintained in 1904-1905 in each of the 

 three races, A, M, and N. 



pairs from the most productive and the least productive broods respec- 

 tively in the two lines. To carry out the selection proved in practice a 

 difficult matter, because one cannot tell in the early stages of the cultures 

 which brood will prove to be largest, so that he must either breed from 

 all the cultures or else use the imagoes which emerge late. To do the 

 former is very laborious, and the latter is undesirable because the late 

 hatched individuals may be under normal size owing to deficient or too 

 sour food. 



