As no adequate comparison of these widely used drugs has been made, we have performed a double-blind cross-over trial in 30 individuals with chronic diarrhea. Each underwent three randomized treatment periods of 4 wk duration. Patients were instructed to increase the daily dose gradually until control was achieved or side effects became intolerable. Stool frequency, consistency, urgency, and incontinence were then compared when a stable dose was reached. Though 2.3 capsules (4.6 mg) of loperamide, 2.3 capsules (103.5 mg) of codeine and 2.5 capsulses (12.5 mg) of diphenoxylate all reduced stool frequency to the same extent, diphenoxylate was significantly less effective in producing a solid stool. Before treatment 95% of patients experienced urgency, sometimes associated with fecal incontinence, often as their major diability. Loperamide and codeine were more effective in relieving this than was diphenoxylate. Side effects, particularly central nervous effects, were greatest with diphenoxylate and least with loperamide. Approximately equal numbers discontinued each preparation; poor control and central-nervous-system side effects were the usual reasons for stopping diphenoxylate and codeine, and abdominal pain and constipation for stopping loperamide. We conclude that both loperamide and codeine phosphate are superior to diphenoxylate in the symptomatic treatment of chronic diarrhea.