Oesophageal function was assessed in 52 patients with angina pectoris whose coronary angiograms were completely normal and in 21 patients with angina pectoris who had significant coronary artery disease. During a standard oesophageal manometric study, abnormalities were found in 23 (44%) patients with normal coronary angiograms but in only 2 (10%) patients with coronary artery disease (p less than 0.01). Twenty-four (46%) patients with normal coronary angiograms were found to have gastro-oesophageal reflux disease during 24-hour oesophageal pH monitoring. Of the 52 patients with normal coronary angiograms, 19 (37%) had gastro-oesophageal reflux disease and abnormal oesophageal motility, 5 (10%) had gastro-oesophageal reflux disease alone, and 7 (13%) had oesophageal motility disorder alone. The use of provocation procedures, including intravenous edrophonium during oesophageal manometry and treadmill exercise testing during pH monitoring, enabled the oesophageal abnormality to be demonstrated simultaneously with chest pain in 25 of these 31 patients. Typical angina pectoris, coincident with abnormal oesophageal motility, was precipitated in a subgroup of patients who had been shown to have oesophageal manometric abnormalities and gastro-oesophageal reflux disease by the infusion of hydrochloric acid into the oesophagus; both the chest pain and manometric abnormality resolved following the oral administration of antacid.