INHERITANCE OF VARIATES IN THE ARMADILLO 153 



work has been primarily statistical in character, but the vahie 

 of individual cases h^s not been lost sight of. 



A determination, for the aggregate of scutes in the banded re- 

 gion, of the coefficient of fraternal correlation is made separately 

 for males and for females. Following this the degree of correla- 

 tion existing between mothers and male, and that between mothers 

 and female offspring are determined separately and compared to 

 discover whether there is any sex-limitation in the inheritance of 

 these aggregates. A study is then made of the correlations exist- 

 ing in smaller aggregates, the individual bands, and these are 

 compared with those detemined for larger aggregates. After the 

 banded region is thus disposed of the caudal armor is dealt with 

 in a similar fashion and a comparison of the two regions is 

 attempted. Finally, by way of further comparison, the cephalic, 

 scapular and pelvic shields are briefly dealt with. 



THE INHERITANCE OF AGGREGATES OF SCUTES IN THE NINE 



BANDS OF ARMOR 



This region was dealt with rather extensively in a former paper 

 (Newman and Patterson '11), but at that time only twenty sets 

 of fetuses were available, a number considered by biometricians 

 as inadequate for statistical treatment. Moreover, no attention 

 was paid to the conditions of the mothers, although the scute 

 counts of a few of the latter were given in table 6, pages 874-875. 

 This table will also serve to indicate the distribution of scutes in 

 the separate bands of mothers and offspring and will obviate the 

 necessity of publishing so large a mass of figures as would be in- 

 volved in a table of the type of that just referred to if 115 sets 

 were included. It will be sufficient for present purposes to give in 

 tabular form the totals for the nine bands of mothers and offspring 

 (table 1, A and B). It seems necessary to publish a considerable 

 amount of the data upon which the subsequent calculations are 

 based in order that it may be possible for anyone interested in so 

 doing to check any errors of calculation, which may have slipped 

 in in spite of a painstaking effort to avoid them. If errors exist 

 they are unquestionably of a minor sort and not of sufficient 

 import to invalidate the significance of the results. 



