Feline mammary carcinoma and canine mammary cancer were evaluated as models for future experimental therapy. Those tumor characteristics known to be of special prognostic significance in human mammary cancers were tested for their prognostic significance in the cat and were compared with those in the dog. The statistical analysis presented is based on a prospective follow-up study of 202 cats treated surgically by mastectomy and by block dissection. Thirty-five factors (general, anamnestic, clinical and histologic data, and data on therapy) were analyzed for relationships with survival, with local recurrence, and with each other. Of the 17 significant relationships found between survival and the direct factors, only 7 remained significant after correction. The factors that related to survival independently of each other were age, diameter of the primary tumor, presence of tumor-positive lymph nodes as judged by microscopic examination, number of mitotic figures, necrosis of the primary tumor, and histologic verification of completeness of surgical treatment. The value of statistical analysis for use in prospective studies of human mammary cancer is discussed.