INHERITANCE OF VARIATES IN THE ARMADILLO 169 



2. There is a small sexual difference in the matter of the number 

 of scutes, females being a little more closely correlated poly- 

 embryonically than males. .In the banded region, it will be re- 

 membered, the males were a little more closely correlated than 

 the females. Probably there is no real sexual difference in this 

 matter that would hold good for any collection of quadruplets. 



3. With regard to uniparental correlation, however, there is a 

 very pronounced difference between the two sexes (at least in the 

 40 sets dealt with here). The correlation between mothers and 

 males, especially if we take into account the probable error, is so 

 low as to amount to practically no correlation. This would seem 

 to mean that, in this collection at least, the males inherit much 

 more strongly from the fathers than from the mothers. In the 

 female sets, however, the correlation coefficient is not far from 0.5, 

 and in that respect about the same as that which obtains for 

 both sexes in the banded region. Probably the statistical treat- 

 ment in this case conceals some of the truth as will be brought 

 out in the next paragraph. 



2. FURTHER EVIDENCE OF ALTERNATIVE INHERITANCE OF GROUPS 

 OF MERISTIC VARIATES 



A glance down the 'totals' column of table 2, A, will show that 

 in practically all cases the numbers of scutes of the four quadru- 

 plets either fluctuate about that of the mother, or else are quite 

 unlike that of the mother and are presumably inherited from the 

 father. The following may be said to be maternal: C 21, C 23, 

 C 46 (mixed), C 76, C 91, K 64, K 90. The rest are to be con- 

 sidered as paternal. In this collection of 20 sets of male quad- 

 ruplets it happens that the proportion of sets like the mother to 

 those like the fathers is 6| to 13^ (counting the mixed set as 

 belonging equally to the two classes). This great predominance 

 of sets resembling the fathers is undoubtedly responsible for the 

 almost total lack of correlation between the male offspring and 

 the mothers. I do not believe that such a condition would be 

 found to hold generally, but that in a small collection of 20 sets 

 one is very apt to have a disproportion between sets inheriting 

 the scute numbers of the mother and those inheriting them from 



