Intestinal permeability was investigated in adult patients with atopic eczema by in vivo and in vitro techniques. Patients with symptoms of 'immediate' food allergy were specifically excluded. A 51Cr-labelled ethylenediaminetetraacetate absorption test was carried out in eighteen patients. Their mean (+/- s.d.) 24-hour urine excretion following oral administration of the test substance (2.1 +/- 0.9%) did not differ significantly from that of thirty-four normal controls (1.9 +/- 0.5%). Small bowel permeability was estimated directly in jejunal mucosal samples in ten patients with three permeability probes of differing molecular weight. Mucosal permeability did not differ significantly from that of fifteen control patients for any of the test substances. Two patients had abnormal results by both tests and in one this was due to coeliac disease. These results suggest that altered intestinal permeability is not important in the pathogenesis of eczema. Patients demonstrating increased intestinal permeability should undergo jejunal biopsy to exclude significant small bowel disease.