The clinical effectiveness of metformin was compared with that of chlorpropamide in closely similar groups of 216 non-obese patients recently diagnosed as cases of maturity-onset diabetes that could not be controlled by diet. The incidences of primary and secondary drug failures in each group and the numbers of patients satisfactorily maintained on each of the hypoglycaemic agents throughout the first year proved remarkably similar. In 61 of the successfully treated patients who were studied by crossover to the other drug and observed for a further year the mean blood glucose concentrations at the end of the year were roughly comparable, but the mean weight response was a small loss of 1.5 +/- 3.8 kg with metformin but a gain of 4.6 +/- 3.9 kg with chlorpropamide. Thus for non-obese, maturity-onset diabetics whose disease cannot be controlled by diet and who require oral treatment sulphonylureas and biguanides are equally effective, the choice depending on whether the patient is underweight and the severity of symptoms.