On the basis of previous experimental studies we postulated that individuals who were phenotypically good hydroxylators but poor sulphoxidisers would be susceptible to chlorpromazine jaundice. Sulphoxidation capacity was assessed in 12 subjects with a history of chlorpromazine jaundice, using S-carboxymethyl-L-cysteine as an in vivo probe. Following an oral dose of 750 mg, unchanged compound and sulphoxide metabolites were measured in urine. All 12 subjects (100%) were shown to be poor sulphoxidisers compared to 22% of normal controls (P less than 0.001) and 23.8% of liver disease controls (P less than 0.001). No subjects with a history of chlorpromazine jaundice had an impaired hydroxylation capacity as assessed by recovery of 4-hydroxydebrisoquine in urine following oral debrisoquine. The results support the hypothesis and demonstrate an inherent metabolic basis of susceptibility to chlorpromazine jaundice.