O'Donoghue, however, concludes from a study of the Marsupial _Dasyurus_ that the stimulus which upon the milk glands proceeds from the corpora lutea in the ovary. In this animal changes in the pouch occur in pregnancy, which are doubtless also due to hormone stimulation, but which we will not consider here. The most important evidence in O'Donoghue's paper [Footnote: _Quart. Journ. Mic. Sci_., lvii., 1911-12.] is that development of the milk glands takes place after ovulation not succeeded by pregnancy; that is to say, when corpora lutea are formed but no fertilised ova or foetus are present in the uterus. In one case eighteen days after heat, the milk gland was in a condition resembling that found in the stages twenty-four and thirty-six hours after parturition. In another specimen, twenty-one days after heat, the milk glands were still more advanced, with distended alveoli and enlarged ducts. The alveoli contained a secretion which was almost certainly milk, O'Donoghue states that the entire series of growth changes in these animals up to twenty-one days after heat in identical with that which occurs in normally pregnant animals. O'Donoghue's conclusion is in agreement with that of Basch,[Footnote: _Monatesschr. f. Kinderh. V._, No. ix., Dec. 1909.] who states that implantation of the, ovaries from a pregnant bitch under the skin of the back of a one-year-old bitch that was not pregnant was followed by proliferation of the mammary glands of the latter. After six weeks the glands were considerably enlarged, and after eight weeks they were caused to secrete milk by the injection of extract of the placenta. It has to be remembered, however, that the milk glands undergo considerable growth, especially in the human species, at puberty and at every menstruation, or at oestrus in animals, which correspond to menstruation. In these cases there is no question of any influence of the foetus, and experiment has shown that if the ovaries are removed before puberty, the milk glands nor the uterus undergo the normal development and menstruation does not occur. According to Marshall to Jolly [Footnote: _Quart. Journ. Exp. Phys._, i. and ii., 1906.] the symptoms of oestrus in castrated bitches were found to result from the implantation of ovaries from other individuals in the condition of oestrus.