The hypothesis that mild hyperoxaluria is more important than hypercalciuria in the pathogenesis of urolithiasis is re-examined in the light of new evidence. Small increments in urinary oxalate in the normal to high-normal range are much more critical than similar rises in urinary calcium for increasing the relative supersaturation of urine with respect to calcium oxalate, the oxalate/calcium ratio in urine, the total volume of calcium oxalate crystals excreted, the proportion of abnormally large crystals and aggregates of calcium oxalate and the severity of the disorder as defined by the recurrence rate of stone-formation. Data from the Arabian Peninsula, where the prevalence of calcium-containing stones is considerably higher than in the West, have shown that this occurs in spite of the almost complete absence of hypercalciuria. On the other hand, there is a strong association between stone-formation and the occurrence of mild hyperoxaluria. The life-time expectancy of stone-formation in men from various countries is strongly correlated with the average daily excretion of oxalate in the urine of the normal men in these countries. This relationship extends to include patients with enteric and hereditary hyperoxaluria. There is no such relationship, however, between the life-time expectancy of stones and urinary calcium excretion in the same populations. Studies on the regulation of urinary oxalate indicate that it is largely controlled by the quantity of "free" dietary oxalate available for absorption in the lower intestine. This can be calculated from the intakes of calcium and oxalate and the urinary excretion of calcium.