37O HARRY BEAL TORREY AND BENJAMIN HORNING. 



We are aware that such variability is a source of difficulty in the 

 interpretation of results. It is greatest in the first few weeks 

 after hatching, when it is attributable partly to irregularities in 

 nutrition and to differences of vigor referable to other causes. 

 These tend to disappear with age. But the constitutional varia- 

 bility of the Rhode Island Red breed, which we have learned to 

 recognize in the course of our work, remains as a source of error. 

 All of our experiments have been checked in other breeds, notably 

 in Leghorns (Torrey and Horning, 1925) and Campines, where 



FIG. 5. R. I. R. female (No. 13, Table I.) twelve weeks old; thyroid fed for eight 



weeks. 



the forcing effect of thyroid feeding on the development of the 

 plumage is clear and definite. In the latter breeds, however, the 

 first rectrices normally develop simultaneously in both sexes. 



The variability of Rhode Island Reds led us formerly into a 

 conclusion that proved to be erroneous. In the first series of 

 experiments reported, our five thyroid-fed capons failed to 

 develop rectrices like unaltered thyroid-fed males. Later it was 

 discovered that this result was exceptional. The gonad had, in 

 the male, no bearing on the result. 



