Differentiation of the metanephrogenic mesenchyme is triggered by an inductive tissue interaction between an inducer tissue and the mesenchyme. It is generally believed that the epithelial ureter bud acts as an inducer during in vivo development. In response to the inductive stimulus most of the mesenchymal cells convert into epithelial cells, while a small fraction differentiates into stromal cells. In vitro, differentiation of isolated mesenchyme to epithelium can be induced by a variety of embryonic tissues, but nothing is known about the molecular nature of the inducing stimulus. In recent years, large numbers of polypeptide growth factors have been described, which in addition to proliferative effects were shown to exert effects on a variety of biological phenomena such as chemotaxis, inflammation, tissue repair, or induction of embryonic development. We therefore analyzed whether growth factors in the absence of inducer tissue can induce isolated kidney mesenchyme to differentiate into epithelium or interstitium. As expected, both growth and differentiation into epithelium were stimulated by an inducer tissue, the spinal cord. We found that none of the various growth factors tested (including epidermal growth factor, transforming growth factors alpha and beta, insulin-like growth factors I and II, fibroblast growth factor, platelet-derived growth factor, and retinoic acid) could mimick the effect of an inducer tissue, although we tested the factors over a wide concentration range. One of the tested factors, epidermal growth factor (EGF) stimulated the mesenchymal cells to become stromal cells, although it could not stimulate development into epithelium. EGF could stimulate stromal development both when the mesenchyme was cultured in isolation and when the mesenchyme was stimulated by an inducer tissue to become epithelium. The expansion of the stromal compartment in response to EGF treatment occurred at the expense of the epithelial cells, but EGF could not completely suppress the formation of epithelium. These data suggest the presence of EGF receptors in the developing kidney, but since application of soluble EGF leads to abnormal development, soluble EGF cannot be the natural ligand. We suggest that locally produced mitogens with an EGF-like structure may regulate the relative amounts of stroma (interstitium) and epithelium in the developing kidney.