Four hundred adults presenting with acute watery diarrhoea were entered into a randomised, placebo controlled, double blind clinical trial of berberine, tetracycline, and tetracycline and berberine to study the antisecretory and vibriostatic effects of berberine. Of 185 patients with cholera, those given tetracycline or tetracycline and berberine had considerably reduced volume and frequency of diarrhoeal stools, duration of diarrhoea, and volumes of required intravenous and oral rehydration fluid. Berberine did not produce an antisecretory effect. Analysis by factorial design equations, however, showed a reduction in diarrhoeal stools by one litre and a reduction in cyclic adenosine monophosphate concentrations in stools by 77% in the groups given berberine. Considerably fewer patients given tetracycline or tetracycline and berberine excreted vibrios in stools after 24 hours than those given berberine alone. Neither tetracycline nor berberine had any benefit over placebo in 215 patients with non-cholera diarrhoea.