The concentration of reducing sugar in the urine is commonly used in the management of diabetes in children. Supplemental doses of regular insulin are administered in response to the concentration of urine sugar according to a protocol termed the "sliding scale." This practice assumes that the concentration of sugar in urine is a good indicator of the plasma glucose concentration. This assumption was tested by comparing urine sugar concentrations in first and second voided urines with the plasma glucose concentrations in 220 children with diabetes. The correlation was good (r = .92) for both the first and second voided urine specimens. Thus, urine sugar concentrations in general define the level of plasma glucose. The large standard deviation of the plasma glucose at each concentration of urine sugar, however, limits the usefulness of urine sugar as an accurate reflection of the coincident plasma glucose concentration. The urine sugar concentration, although useful for the general management of diabetes, provides significant risk when used to guide frequent adjustments in insulin administration. Therefore, the "sliding scale" should not be used in the treatment of children with diabetes.