Helicobacter pylcri (H.pylori) was studied in duodenal ulcer (DU) and gastric cancer patients. The relationship between dietary linoleic acid, antral H.pylori infection and DU was studied. A detailed dietary history was obtained from DU patients and matched non-ulcer dyspepsia controls. The fatty acid content of abdominal adipose tissue biopsies was measured in both groups. Levels of dietary and adipose tissue fatty acids were related in controls and non-smokers. The adipose fatty acids did not reflect dietary intake in DU patients or controls who smoked. DU is associated with reduced adipose tissue linoleic acid content and this is accentuated by smoking. In an in vitro test, linoleic acid was not shown to be metabolised to more saturated fatty acids by H.pylori. Eradication of H.pylori from the stomach prolongs remission in DU disease but gastric re-infection is common and leads to relapse. Tooth plaque was cultured from DU patients and from matched controls. H.pylori was cultured from the antrum in 89% and from tooth plaque in 19% of duodenal ulcer patients. The identity of H.pylori was confirmed by biochemical testing and DNA analysis. Some strains were examined by electron microscopy. Ribotyping showed that the organisms present in tooth plaque and the antrum were identical but each individual carries a separate strain. Gastric re-infection by H.pylori could originate from dental plaque. A general population of hospital inpatients with and without teeth were studied by bacteriological culturing of tooth plaque, by urease testing of tooth plaque and by measuring plasma antibody to H.pylori. No significant difference in antibody levels was found in either group. In an in vitro experiment, bile acids inhibited or prevented growth of H.pylori supporting the view that biliary post-reflux gastritis and type B H.pylori gastritis are of different aetiology. Gastrectomy specimens from 83 patients who underwent gastrectomy for primary gastric cancer and from 34 controls who underwent gastric surgery for peptic ulcer were reviewed histologically for H.pylori. No associations were found between H.pylori and gastric intestinal metaplasia, tumour extent or stage.