The influence of gastric emptying, intestinal transit, renal and hepatic function, and variations in the timing of urine collections, on the cellobiose/mannitol test of intestinal permeability has been studied. None of these extraneous factors influences the cellobiose/mannitol recovery ratio, and there is only a modest effect of renal function on urinary mannitol recovery. Cellobiose is almost totally recovered in the urine within 10 h of intravenous injection, whilst mannitol is less completely recovered, perhaps due to hepatic metabolism. The simultaneous administration of two probe molecules in the cellobiose/mannitol test, and the use of a ratio to express results, thus achieves the desired object of minimising the effect of extraneous factors on the test result. The cellobiose/mannitol test is therefore not subject to the limitations of previous tests of intestinal permeability using orally administered probe molecules, and is widely applicable as a screening test for coeliac disease.