1. Studies were carried out in vitro using an ultracentrifugation method to quantify bile acid binding to the different components of a Lundh test meal, and to determine what factors influence bile acid binding to one of the components (casein). We validated the ultracentrifugation method by showing good agreement with the equilibrium dialysis method. Studies were carried out in vivo on jejunal aspirate from 10 ileal resection patients in order to determine whether bile acid binding to casein could be demonstrated, and whether this influenced aqueous-phase bile acid and fatty acid concentrations. 2. In vitro , the Lundh test meal was found to adsorb bile acid. The protein content of the meal (casein) alone accounted for this binding, which was abolished by use of casein hydrolysate. The binding to casein was a saturable process. Both binding affinity and binding capacity were significantly greater for taurocholate at pH 4.5 than at pH 6.5, and for dihydroxylated than for trihydroxylated bile acid, suggesting that hydrophobic bonding was involved. 3. In vivo , jejunal samples aspirated at pH > 6 from 10 ileal resection patients showed 25% binding of bile acid to protein. On substitution of amino acids for casein, mean binding was reduced to 16% ( P < 0.05), residual binding being attributed to endogenous protein. This was associated with an increase in fatty acid solubilization from 28% to 60% ( P < 0.025). 4. These findings suggest that protein binding may have a critical effect on aqueous-phase bile acid and fatty acid concentrations in certain patients with a reduced total bile acid concentration as in ileectomy steatorrhoea, and in patients with defective lipolysis as in pancreatic insufficiency. This finding may provide an additional rationale for the use of elemental diets in severely ill patients with steatorrhoea.