Recognition that vinyl chloride could be hepatotoxic led to a survey of workers to determine whether changes had been induced by past exposure, and to evaluate standard liver function tests as monitors of early liver abnormalities. Standard liver function tests were found to be unsuitable for the detection of such abnormalities in the population at risk. Of 487 workers examined, 102 (20-9%) had abnormalities on initial testing but only two were finally shown to have portal hypertension; in both cases, thrombocytopaenia provided the first diagnostic evidence since liver function tests were normal. Furthermore, 40 (35-7%) of 112 control subjects had initial test abnormalities. A sample of 19 workers with various exposures to vinyl chloride monomer were examined blind by greyscale ultrasonography. Five with minimal or no exposure were confirmed as normal but 12 of the remainder had abnormalities. These consisted of an enlarged portal vein (seven instances), splenomegaly (eight), and changes in hepatic texture (seven). Five of these 12 cases had previously been considered normal. It was concluded that greyscale ultrasonography had many advantages over standard methods for screening workers exposed to hepatotoxic chemicals, and should be the subject of a large scale evaluation.