1. Medication, social and symptom histories were compared in patients with severe haemorrhage from a peptic ulcer (n = 71) and matched control subjects. Self-medication with proprietary agents was catalogued in addition to therapy prescribed by general medical practitioners. 2. Prior to the bleed, only 4% of ulcer patients had been free of symptoms normally associated with peptic ulceration, whereas 76% of the control group had been asymptomatic. 3. Gastro-irritant proprietary medicines were used regularly by 23% of ulcer patients compared with only 4% of controls. However, proprietary antacids were used chronically by 46% of ulcer patients compared with only 7% of controls. Bicarbonate was the antacid of choice used by 13% of ulcer patients. The odds ratio for the association between development of bleeding peptic ulcer and the use of indigestion remedies was 11.5% (95% CI 1.1, 121). 4. Fifty-one percent of ulcer patients were prescribed agents known to cause gastro-intestinal damage, whereas only 25% of the control group were prescribed similar agents. Only 7% of the control group were prescribed anti-ulcer therapy compared with 37% of those with bleeding ulcer. 5. A large proportion of patients with haemorrhage from a peptic ulcer had had symptoms sufficient to warrant recourse to self-medication with antacids without medical knowledge. Exacerbation of peptic ulcer by self-medication with proprietary products is likely to be of lesser significance.