INHERITANCE OF VARIATES IN THE ARMADILLO 179 



rion with much more advanced fellow fetuses, but there is no un- 

 usual difference as to scutes. Such dimensional differences within 

 sets are no doubt due to accidental interferences with the food 

 supply of a more or less radical character. The fact that these 

 very pronounced inequalities in growth conditions fail to disturb 

 the inheritance ratios proves in the most unequivocal way, I be- 

 lieve, that the resemblances and differences that we have studied 

 in connection with the scutes and their aggregates are purely 

 matters of heredity and as such are ideal for our purposes. 



4. Unless the writer is totally, misled by the data, which are 

 given in extenso so that the reader may have the opportunity of 

 making his own judgment, the mode of inheritance of larger and 

 smaller groups or aggregates of scutes is primarily alternative with 

 only a minor degree of blending. It is highly probable that blend- 

 ing would disappear entirely if smaller aggregates than those 

 treated were to be dealt with, for the apparent blending in the 

 larger armor regions may be merely the result of an averaging of 

 the scute numbers belonging to the mosaic of minor inheritance 

 regions of which some are maternal and others paternal in charac- 

 ter. There is doubtless a considerable degree of incompleteness in 

 dominance, just as there is in practically all cases of alternative 

 inheritance, which would account for the smaller fluctuations of 

 the four quadruplets about the maternal scute number or, pre- 

 sumably, the paternal. It is interesting, however, to note that 

 in the smaller aggregates, such as the tail rings this incompleteness 

 in dominance disappears to a large extent and we find in the 

 offspring a very large number of rings with exactly the same num- 

 ber of scutes as occur in the corresponding rings of the mother. 

 Presumably this is also true for the father. Incomplete dominance 

 is the rule even in the case of the simplest characters and hence 

 we must expect the same or a greater degree of incompleteness in 

 the case of such complex characters as those which form the basis 

 of the present study. And we would expect also to find that the 

 more precise are the methods of comparison between parent and 

 offspring the less complete would be the dominance. No such pre- 

 cise and searching comparison of genetically related individuals 



