One thousand one hundred and twenty-eight patients treated for perforated gastroduodenal ulcer during the years 1935-1985 were studied at the Haukeland University Hospital. The majority of patients (97.7%) were treated surgically. The data was analyzed by contingency tables and chi square testing, and a stepwise logistic regression analysis was performed in order to reveal interactions between variables and to elucidate time trends in lethality rates. The total postperforation lethality was 7.4%, the postsurgical death rate was 6.6%, and the death rate among conservatively treated patients was 42.3%. Lethality was significantly influenced by year of hospital admission and increased markedly with the age of the patients. For all age groups, the lethality decreased markedly with time. Treatment delay was associated with a moderate but significant increase in lethality. In patients with gastric ulcer the lethality was 3.6 times higher than in those with duodenal ulcer. The death rate was similar in the duodenal and pyloric ulcer groups. Death rate decreased with time in both stomach ulcer, duodenal, and pyloric ulcer patients. There was no sex difference and no difference between patients treated with simple suture or gastric resection.