Using a three-cell study design over two years, we assessed the efficacy of rice ORS (R-ORS) for home management of dysentery in rural Bangladesh involving 1,911 children of less than five years of age who were matched for socioeconomic status, morbidity, nutrition and other factors. Dysentery was defined as passing blood and/or mucus in the stool, and the total numbers of episodes studies were: 1472, 1335, and 1557 in the R-ORS, glucose ORS (G-ORS), and comparison areas respectively. Packaged R-ORS and G-ORS were supplied free of cost. The comparison area continued the usual treatment of dysentery as practised in the community. During the two-year period, weekly observations were made in all areas on the incidence, management, and outcome of all episodes of dysentery. We demonstrated that the use of R-ORS was associated with reduction in duration and a lower incidence of prolonged diarrhoea compared to treatment with G-ORS or drugs. The mean durations of dysenteric diarrhoea were 7.6 +/- 4.9, 10.1 +/- 6.3, and 14.0 +/- 8.0 days in the R-ORS, G-ORS, and comparison areas respectively. Similarly, the cumulative recovery rates on day 7 of the illness were 61%, 42% and 21% in R-ORS, G-ORS, and comparison areas respectively. Based on these observations, we conclude that R-ORS may be an effective treatment of dysentery and deserves further investigation.