The incidence of several forms of liver disease associated with inflammatory bowel disease has been putatively ascribed to a toxic effect on the liver of portal vein bile acids abnormal in type or amount. To examine this possibility, total bile acid concentrations (sulphated and non-sulphated) were measured by gas-liquid chromatography in inferior mesentric vein serum of 19 patients undergoing colectomy for severe inflammatory bowel disease (IBD). Similar determinations were obtained on a control group of eight patients requiring colectomy for other non-inflammatory conditions. Mean values for mesenteric vein serum bile acid concentrations (μM/1) were 19·6±1·8 in controls and 16·3±2·0 in IBD. The mean sulphated bile acid fraction did not exceed 10% of total, although there was considerable variability (up to 40% of total). Lithocholic acid levels (entirely sulphated in all patients) were low. Although the IBD group showed a more than two-fold increase in mean lithocholate concentration (0·54±0·15 μM/1) over controls (0·21 ± μM/1), this difference was not statistically significant. No significant intra-group difference was noted in the non-sulphated and sulphated fractions for cholic, chenodeoxycholic, and deoxycholic acid species, respectively. No unidentified or unusual bile acids were observed. There was no correlation between bile acid measurements and liver histology. These findings fail to support the hypothesis that liver disease often found in association with severe inflammatory bowel disease represents a form of bile acid toxicity. The invariable finding of total sulphation of the lithocholic acid fraction even in the presence of severe mucosal disease was unexpected.