Studies were carried out on multiple urine samples from eight patients with recurrent idiopathic calcium oxalate stone formation and eight normal persons to define an index of the risk of forming calcium oxalate stones. Under the same conditions of dietary and fluid intake the urine samples of the patients with stone formation were more supersaturated with calcium oxalate (P less than 0.001) and had lower concentrations of protective inhibitors of crystallization (P less than 0.001) than those of the controls. However, the best separation between the groups was defined by a discriminant line relating inhibitory activity and urine saturation. A measure of the risk of forming large crystals, the saturation-inhibition index, was defined as the distance of each urine from the discriminant line. The patients with stone formation had a significantly higher mean saturation-inhibition index than the controls (P less than 0.001). Both the percentage of large calcium oxalate crystals excreted (P less than 0.001) and the stone episode rate (P less than 0.005) were significantly correlated with the saturation-inhibition index.