In order to determine: (1) whether there is a growth-regulating interaction between the mammary fat pad and mammary epithelium; (2) whether this interaction could be modified by dietary fats; and (3) whether these effects could be demonstrated in vitro, the following experiments were performed. Virgin Balb/c mice had the left inguinal mammary fat pad cleared of epithelium and were then maintained on one of four fully defined diets. These diets contained the following proportions of fat by weight: 5% or 10% mixed fats; 20% saturated fat plus cholesterol; or 20% polyunsaturated fat. To test for effects in vivo, animals received subcutaneous injections into the cleared fat pad of tumorigenic mammary cells (WAZ-2T(+SA) or WAZ-2T(-SA)) or preneoplastic mammary cells (CL-S1). Dietary fat had little effect on the latent period of tumour formation, but a low-fat diet increased the invasive/metastatic potential of both tumorigenic cell lines. A high-saturated-fat diet inhibited the growth of normal and preneoplastic epithelium in vivo. To test for effects in vitro, CL-S1 cells were co-cultured with explants of cleared mammary fat pad embedded within collagen gels. CL-S1 cells co-cultured with adipose explants obtained from mice fed on a diet containing 20% polyunsaturated fat showed a threefold increase in incorporation of [3H]thymidine into trichloroacetic acid-precipitable material. These results imply that dietary fats may affect the growth of mammary epithelium in two ways: the inhibition of growth caused by the high-saturated-fat diet may be due to systemic effects as it was not apparent in vitro; the increase in growth seen in vitro and caused by a high-polyunsaturated-fat diet is due to a direct interaction between the mammary fat pad and mammary epithelial cells. This interaction may be masked by systemic effects in vivo.