The study compares the ability of a simple double contrast technique with our standard single contrast barium meal to diagnose peptic ulceration. Two hundred and six patients were randomly allocated to either examination. Endoscopy was used as the definitive diagnostic procedure. Deformity of the duodenal cap was more accurately detected by the double contrast technique (P less than 0.01). There was no significant difference in the detection rates for duodenal ulcer. False positive or false negative diagnoses of duodenal pathology were similar by both techniques. The incidence of gastric ulceration in the series was too low for statistical analysis.