The result of the mating of Nos. 2464 and 1636 in pen 736 was 25 chicks, of which 24 had tails and 1 (No. 5335) was without tail or oil-gland. This, unfortunately, died early, so it was impossible to breed it. In 1908, the hen No. 1636 having in the meantime died, I mated No. 2464♂ to 6 of his (tailed) daughters. He was not well and soon died, leaving no descendants by them, but 5 offspring by a female cousin, all tailed. Then one of his sons (tailed) was mated to its own sisters and produced 49 offspring, all tailed. Thus the strain seems to have died out. The whole history is important both because an apparently new mutation had taken place and because it was, in a degree, "hereditary." How, if at all, can this case and those of the bantams be brought under known laws of inheritance? First of all, it must be confessed that the provisional hypothesis, suggested in my earlier report, that rumplessness is in my strain recessive has not been supported by the newer facts. In the light of the principle of imperfect dominance to which the facts of the last two chapters have led us, everything receives a satisfactory explanation. The only conclusion that meets all the facts is this: _The inhibitor of tail development--the tailless factor--is dominant_; _its absence_--permitting a continuation of the normal development of the tail region--_is recessive_. The application of this hypothesis to the various matings may now be attempted. No. 117 is to be regarded as a heterozygote. The matings with tailed birds is of the order DR × R, and expectation in the typical case is 50 per cent DR (interrupted tail) and 50 per cent RR (non-interrupted). But, owing to the relatively weak potency of the interrupter derived from No. 117, growth of the tail is not interrupted in the heterozygous offspring. These offspring are, by hypothesis, so far as their gametes go, of two equally numerous sorts, DR and RR. Mated to No. 117, two sorts of families are to be expected, namely, the products of DR × RR (=50 per cent DR, 50 per cent RR) and the products of DR × DR (=25 per cent DD, 50 per cent DR, 25 per cent RR). The first lot of families might be expected to resemble the preceding generation in consisting entirely of tailed birds; the latter might be expected to show in the 25 per cent extracted DD's evidence of the presence of the undiluted interrupter. Actually in matings of the latter sort (table 27) 3 families show no trace of the tail-interrupter, but in 7 there is evidence of a disturbance, as shown by the small size of the uropygium and the bent back. In these families there are 13 cases of small uropygium to 53 of large, being about 20 per cent of the affected uropygium where 25 per cent was to be looked for--not a wide departure, considering the liability of not recognizing the reduced uropygium as such. This failure even of the extracted dominants completely to stop the development of the tail gives a measure of the weakness of the inhibitor in this case.