A decline in hospital admissions, surgery, and death rates resulting from peptic ulcer disease has been observed in the United States. It is as yet unsettled whether these changes represent a real fall in ulcer incidence or are only due to changes in the practices of treatment and coding of gastric and duodenal ulcer. Based on monthly surveys among random samples of office-based physicians from different specialities and regions in the United States, the National Disease and Therapeutic Index (NDTI) publishes annual information on how often physicians were visited for different kinds of diseases. The physician visits for gastric and duodenal ulcer between 1958 and 1984 were used to study the time trends of gastric and duodenal ulcer disease. The visits for duodenal ulcer markedly declined during the past 20 years. The decline in duodenal ulcer involved first and subsequent visits and both sexes similarly. The rate of first visits for gastric ulcer declined in men but remained unchanged in women. From 1958 to 1984, the rate of all visits for gastric ulcer decreased in men and increased slightly in women. The ratio of male over female ulcer disease decreased in both ulcer types. For duodenal ulcer, this change resulted predominantly from a steeper decline in men than in women rather than from an increase of female ulcer disease. For gastric ulcer, the relative rise of female ulcer disease was most marked in the old age groups. The similarity between the temporal trends of visits for gastric and duodenal ulcer and other statistics of ulcer morbidity suggest that these trends reflect real changes in the occurrence of the two diseases.