780 



PROCEEDINGS OP THE AMERICAN ACADEMY. 



iV30 (Figures 6 and 7) and Carpenter's "normal" individuals (Figure 

 5, N). This is true irrespective of the inbredness of the different cul- 

 tures. This conclusion is supported by a comparison of the A series in 

 the sixth inbred generation (A 6, B 6, and G 6) and in the sixty-first 

 generation (A 61). Conditions were optimum in both cases; the aver- 

 age number of spines and the variability are similar. Contrary to the 

 common opinion, the variability would seem to have been increased 

 rather than diminished by the inbreeding, since the coefficient of varia- 



TABLE XXIII. 



Correlation in Number of Spines between Sex-combs of Right and 

 Left Legs, Series X, Generation 1. 



Left leg (x). 



H 



A x = 10.74 

 A y = 10.81 



<rx = 1.094 

 ay = 1.186 



r = 0.301 -L_ .06133 



tion is greater for the sixty-first than for the sixth inbred generation ; 

 but this difference may be due to the small numbers of individuals exam- 

 ined, particularly in the cultures of the sixth generation, each of which 

 furnished only forty males. 



The measurement, tibial length, is probably a good index of relative 

 size of the individuals in the cultures measured. If so, we may infer 

 that the individuals of X 1 are larger than those of the other three cul- 

 tures measured, which fact would indicate that inbreeding tends to 



