An investigator blind trial was performed comparing bismuth salicylate, erythromycin ethylsuccinate, and placebo in the treatment of Campylobacter pyloridis associated gastritis in patients without peptic ulceration. Fifty patients fulfilled the study criteria. There was a strong correlation between the presence of C pyloridis and histologically confirmed gastritis. Clearance of organisms led to improvement of the gastritis. C pyloridis was cleared from 15 patients; of these, 13 had gastritis initially, which resolved in 12. Conversely, gastritis resolved in only four of 32 patients not cleared of organisms (p less than 0.0001). There was significantly greater improvement in endoscopic appearances in the patients cleared of C pyloridis compared with those whose infection persisted (p less than 0.001). In the three treatment groups organisms were cleared from 14 of 18 patients receiving the locally active bismuth salicylate, only one of 15 patients receiving erythromycin ethylsuccinate, and none of 17 patients taking placebo. These findings suggest that the ideal antimicrobial for the successful eradication of C pyloridis associated gastritis should be locally active, stable at low pH, and should penetrate gastric mucus. The resolution of gastritis and improvement in endoscopic appearances associated with clearance of C pyloridis support the view that these organisms may play a part in this condition.