PURE LINE INHERITANCE AND PARTHENOGENESIS. 33 



TABLE OF MEASUREMENTS AND RATIOS (Continued). 



gave me five young, only two of which came to maturity and 

 gave an average index of 1.80:1. F 6 3 gave birth to eight indi- 

 viduals, four of which came to maturity, and gave as a fraternal 

 mean an index of 1.88:1. Although the number of individuals 

 obtained are too small to draw any conclusions, if they were 

 not supported by a repetition of similar results, yet they are 

 very significant, as is shown by the plot in Fig. i. The curves 

 running from the two extreme individuals selected of the F& 

 generation, to the fraternal means of the offspring of these two 

 individuals actually cross! In other w T ords, offspring from an 

 individual with a high index had a lower fraternal mean than 

 the offspring from the individual (of the same fraternity as the 

 former) w r ith a low index. Among the F 7 individuals two were 

 selected; one, F ? 3 with an index of 1.89:1, and one, F 7 4, w r ith an 

 index of 1.79:1. From the former, F 7 3, eleven young w r ere 

 reared to maturity. They gave a fraternal mean of 1.93:1. 

 This is almost the same formula as was obtained for the mean 

 of the offspring of the latter individual, F 7 4, which was 1.95:1. 

 Here w r e get convergence almost to the same point in the offspring 

 of two variants of a fraternity. However, both of these fra- 

 ternities, that is, of the F 8 generation, did not show a reversion 

 tow r ard the mean of the original strain, but a deviation from it. 

 This is what occurred in the case of the F 5 generation. In the Fg 

 fraternity, F 8 i, formula 1.86:1, and F 8 I9, formula 1.67, were 



