The correlation between radiography, endoscopy, and histology in the diagnosis of reflux esophagitis, as well as the effect of glucagon on double contrast radiography was studied. The material consisted of 220 out-patients sent to the Oulu University Central Hospital for upper gastrointestinal endoscopy. 109 of these were shown to have reflux esophagitis at endoscopy, the other 111 formed a control group with normal esophageal mucosa. Radiologic examinations were performed after endoscopy on the same day by a radiologist, who knew neither the clinical history of the patients nor the findings at endoscopy. Using endoscopy as a reference, 56% (28/50) of the patients with grade E I reflux esophagitis (erythema, oedema) were diagnosed correctly by double contrast radiography. The corresponding figures concerning grade E II (erosions), grade E III (localized deformity, ulcer), and grade E IV (stricture) reflux esophagitis were 84% (41/49), 100% (4/4), and 100% (6/6). False positive findings were found in 4.5% (5/111). The sensitivity of double contrast radiography as compared to endoscopy in all grades was 73%, its specificity was 96%, and accuracy 84%. The corresponding figures, when only grades E II, E III, and E IV are considered, were 86%, 96%, and 92%. In double contrast radiography, signs sometimes visible in grade E I reflux esophagitis were thick mucosal folds and mucosal granularity. Reliable signs of grade E II reflux esophagitis were streaks and dots of barium against the mucosa either alone or together with thick mucosal folds and mucosal granularity. Specific signs of grade E III and E IV reflux esophagitis were--along with the above--localized deformities, ulcers, and strictures. A hiatus hernia or wide hiatus was detected radiologically in 2/3 of the reflux esophagitis patients, and in 1/3 of the controls. Histologic findings correlated poorly with both endoscopic and radiologic findings. Single contrast radiography was less sensitive than double contrast radiography in detection of superficial mucosal lesions. Glucagon had no advantagous effect on esophageal double contrast radiography. Its use, however, in connection with double contrast radiography of the stomach is unlikely to have any disadvantagous effect on the evaluation of the hiatus and gastroesophageal reflux.