There are a number of lines of evidence suggesting that transforming growth factor beta (TGF beta) has an important role in the control of intestinal growth and differentiation. In vivo localization studies show that TGF beta expression occurs predominantly in the differentiated non proliferating cells of the intestinal epithelium. The use of an antisense expression vector for TGF beta resulted in an increased tumorigenicity in an antisense-transfected cancer cell line. In vitro proliferation studies showed colorectal premalignant adenoma cells to be more sensitive to the growth inhibitory effects of TGF beta than colorectal cancer cells. Furthermore the conversion of an adenoma to a carcinoma was accompanied by a reduced response to the inhibitory effects of TGF beta. The acquisition of partial or complete resistance to the inhibitory effects of TGF beta may be an important late event in colorectal carcinogenesis. Of further interest is the possibility that clonal selection could occur even more rapidly in colorectal tumour cells which not only had lost response to TGF beta inhibition but produced TGF beta and were growth stimulated by it. This could have the advantage of not only inhibiting the growth of surrounding less malignantly advanced cells but of also escaping from their potential growth suppressive influence. Carcinogenesis is not, however, simply losing response to negative regulators of growth; the fully malignant cell has to acquire new characteristics of invasiveness and metastatic potential. Growth factors including TGF beta may have a role in the complex cascade of events leading to the activation of proteolytic enzymes which are involved in progression to an invasive phenotype. Cell proliferation in the large bowel, as well as being under the control of endogenous growth factors, is also under the influence of dietary components in the lumen such as the naturally occurring fatty acid sodium butyrate. Sodium butyrate at physiological concentrations induces apoptosis (programmed cell death) in colonic tumour cell lines. Since sodium butyrate occurs naturally in the colorectum, being produced by bacterial fermentation of dietary fibre, it may be involved in the control of cell death in human colorectal epithelium. This could, in part, explain the apparent protective effects of dietary fibre. Clonal evolution and tumour progression in colorectal carcinogenesis could therefore involve loss of response to endogenous growth factors such as TGF beta and an escape from the induction of programmed cell death by dietary factors.