There are two ways in which the significance of hormones may be demonstrated, namely, by studying the effects of diminishing or of increasing quantities of active hormone below or above the normal hormone level. Loeb and his associates, by ovariectomizing mice at different periods of life, used the first of these methods to test the relationship between the action of ovarian hormones and mammary cancer (1–4). These experiments led to the conclusion that the frequency of cancer and the age at which it appears in a given family or strain are directly proportional to the quantity of hormone, or hormones, which have had a chance to act on the mammary gland. These observations were confirmed and extended by Cori (5). An hereditary factor, also, was found by Loeb to interact with ovarian hormones in producing mammary cancer. This combined action of internal secretion and hereditary factors we have discussed in a series of previous papers (1–4). Furthermore, it was established in these early experiments that, after a certain quantity of the hormone had acted, the cancerous transformation would occur subsequently, even though the action of the hormone had ceased some time before the first appearance of cancer (2). It seemed, therefore, very probable that the manifestation of cancer was preceded by a preparatory period, during which changes took place in the mammary gland tissue which might not be visible, but which caused the reactions of the tissues from then on to be such that ordinary functional and metabolic activities were able to bring about cancerous transformation. It was shown, also, that the incidence of cancer is greater in breeding than in non-breeding mice (6), although the degree of this difference is more marked in some strains than in others. In every strain, however, removal of the action of the ovarian hormones early in life had a greater effect, and usually a much greater effect, on the reduction of the cancer rate than the prevention of breeding. Furthermore the tumors appeared later in life in non-breeding and castrated mice than in breeding mice. But this difference in tumor age was not actually as great as it appeared, inasmuch as it was found that prevention of breeding, as well as ovariectomy, tended to prolong the life of mice quite noticeably. Because these mice lived longer their chances of becoming affected by cancer in a higher age group was therefore greater. Since non-breeding and ovariectomized mice lived to about the same age, it seemed probable that castration affected the duration of life indirectly by the prevention of breeding.