Out of five females on which the operation was performed only two lived more than a few days afterwards. One of these (a) was castrated in the spring of 1909 when a little less than a year old, the other (b) on August 13 when twelve weeks old. In October 1909 they showed no marked modifications. In July 1910 it was noticed that they had the male curled feathers in the tail, and (a) had breast feathers similar to those of the male in summer plumage, (b) was rather more strongly modified: she had a very narrow white neck-ring, and breast feathers distinctly of male type. The next moult began in September, and in November was well advanced. On the whole (a) had made little advance towards the male type, but (b) closely resembled the male in nuptial plumage. It had brilliant green feathers on the head, a white neck-ring, much claret colour on the breast, and some feathers indistinguishable from those of the male, and also the male sex feathers on the tail. Goodale concludes that the female owes her normal colour to the ovaries or something associated with them which suppresses the male characters and ensures the development of her own type. He considers it is quite as conceivable that selection should operate to pick out inconspicuously coloured females as that selection of brilliantly coloured males should bring about an addition to the female type. But as pointed out above, selection cannot explain the dimorphism in either case. It may be mentioned here that owing to the fact that the single (left) ovary in birds is very closely attached to the peritoneum immediately covering the great post-caval vein, it is generally impossible to remove the whole of the ovary without cutting or tearing the wall of the vein and so causing fatal hemorrhage. The above results observed by Goodale are therefore all the more remarkable, and it may be assumed that he removed at any rate nearly all the ovary.