Résumé : The present study investigates the possibility that insulin dependence, a property exhibited in organ culture by a majority of the 7,12-dimethylbenz(a)anthracene-induced rat mammary carcinomas, might also be operating in vivo. It was found that induction of alloxan diabetes 3 or 4 weeks after 7,12-dime thylbenz(a)anthracene administration completely prevented mammary tumor formation. Moreover, induction of alloxan diabetes in tumor-bearing rats produced the rapid regression of 90% of the tumors, with a time course similar to that observed after oophorectomy or hypophysectomy. In contrast to what occurs after oophorectomy but in accordance with what is observed after hypophysectomy, administration of estradiol benzoate failed to prevent the tumor regression produced by alloxan diabetes. Because induction of alloxan diabetes involves a considerable loss of body weight, the effect on the mammary tumors of a severe food restriction, leading to an even greater loss of body weight, was also studied. It was found that food restriction resulted in a rapid regression partially counteracted by estradiol benzoate. Tumors regressing as a result of alloxan diabetes, those progressing (about 10%) in spite of diabetes, and those that progressed under the stimulating effect of estradiol benzoate in rats subjected to food restriction were investigated with respect to their insulin dependence in organ culture. It was observed that the tumors regressing in alloxan-diabetic rats were, by and large, markedly insulin dependent in vitro, whereas those growing despite diabetes were little-or noninsulin dependent. Eventually, most tumors that were stimulated to grow by estradiol benzoate in food-restricted rats proved very insulin dependent in vitro. These observations suggest that the tumors that are insulin dependent in organ culture similarly have a stringent requirement for insulin to grow in vivo and that they regress after induction of alloxan diabetes as a consequence of insulin deprivation. This would explain why only tumors that are insulin independent in vitro are able to grow despite the insulin deprivation of alloxan diabetes. It is conceivable that food restriction results in tumor regression also as a consequence of a decreased rate of insulin secretion. However, this decrease would be smaller than in the diabetic state, and the available insulin would be present in amounts large enough to allow insulin-dependent tumors to grow, provided other limiting factors such as estrogens or estrogen-stimulated hormones are restored. © 1972, American Association for Cancer Research. All rights reserved.