Radioimmunoassay measurements of hormones in sera show that contrary to expectations, natural mammary tumorigenesis in mice is not always accompanied by high prolactinemia. Basal levels of circulating PRL follow no consistent pattern in relation to the incidence of mammary tumors. Circulating PRL after a provocative stimulus, however, presents a profile characteristic of the mammary tumor incidence: strains with high incidence of mammary tumors generally have have lower levels, whereas strains with low incidences exhibit higher levels. These differences in circulating PRL levels are apparently not caused by differences in the pituitary stores or pituitary release of PRL. The PRL release patterns appear not to be entirely genetic, since they can be modified by foster nursing, a procedure that may introduce MTV in a virus-free strain. The urine concentrations of PRL and GH and the tissue uptake data suggest that the C3H/St strain may metabolize PRL and GH differently from the C5BL/St strain. It remains to be determined whether the observed differences are significant factors in the development of spontaneous mammary tumors in mice.