We aimed to produce a unifying hypothesis to explain the different locations of peptic ulcer and gastritis observed in different populations. The pre-Helicobacter pylori literature on patterns of gastroduodenal disease was reviewed and compared with recent human and animal findings on H. pylori infection. Early observations revealed that duodenal and non-pre-pyloric ulcers tend to be mutually exclusive. In duodenal ulcer patients, gastritis is usually restricted to the antrum, while gastric ulcer patients experience more severe pangastritis. The manipulation of acid output by surgery or acid suppressive therapy alters the distribution of gastritis. Recent experimental evidence in humans and animals has shown that these changes parallel changes in the distribution and cellular responses to H. pylori infection. We propose that the most important factor in the ecology of the H. pylori-infected stomach is local acid production. Local acid production determines the location and severity of inflammation and the clinical outcome of this bacterial infection. Priority research areas should be the investigation of the in vivo behaviour of H. pylori in the acid and the non-acid producing areas of the stomach and the measurement of acid output in populations known to have different patterns of gastroduodenal disease.