INHERITANCE OF VARIATES IN THE ARMADILLO 155 



For the sake of brevity the results of an intricate series of arith- 

 metical computations are set down in a few short paragraphs as 

 follows : 



a. Polyemhryonic correlation for 56 sets of male quadruplets 

 Mean = 558.62 ± 0.65 scutes 



(TX^ = 236.18 

 av~ = 33.36 

 Substituting for the general formula, we have: 



r=(l -I j33^\ Q g294 ± 0.0057 

 \ 2 236.18/ 



b. Polyemhryonic correlation for 59 sets of female quadruplets 

 Mean = 559.05 ± 0.57 scutes 



(7.x2 = 169.95 

 av"" = 33.05 

 Substituting for the general formula, we have: 



r = 1 - f- l?:2i ) = 0.9129 ± 0.0059 

 \2 169.95/ 



c. Polyemhryonic correlation for the 20 sets of quadruplets used in 

 the former paper 



Mean = 558.48 ± 2.51 scutes 

 <rx^ = 257.62 

 av'- = 33.5 

 Substituting, we get: 



r = ( 1 - 1 J^')= 0.9348 ± 0.0111 

 \ 2 257.62/ 



A survey of these results brings out the following noteworthy 

 points : 



1. The coefficients of correlation for the new collection of quad- 

 ruplet sets are sensibly equivalent to those determined for the 

 original 20 sets, indicating that 20 sets consisting of 80 individ- 

 uals are enough for statistical purposes, the statement of experts 

 to the contrary. 



