Human proinsulin connecting peptide (C-peptide) was measured by immunoassay in urine from 25 normal subjects, 18 patients with diabetes mellitus, and 34 patients with various degrees of renal insufficiency. Assay validation studies showed that pancreatic C-peptide was quantitatively recovered when added to urine. Fractionation of urine by gel filtration indicated that most endogenous C-peptide eluted in fractions that corresponded to the C-peptide standard. In 34 nondiabetic subjects with normal kidney function or various renal diseases, C-peptide clearance was independent of creatinine clearance over a range of 6 to 190 ml./min. Urine C-peptide clearance (5.1 +/- 0.6 ml./min.) is greater than that of insulin (1.1 +/- 0.2 ml./min.), and the total quantity of C-peptide excreted in the urine per day represents 5 per cent of pancreatic secretion, as against only 0.1 per cent of secreted insulin. Healthy subjects excreted 36 +/- 4 mug. C-peptide per 24 hours, while this value in juvenile-onset diabetics was only 1.1 +/- 0.5 mug. Adult-onset diabetics excreted 24 +/- 7 mug./24 hr., the range overlapping the excretory rates of both normal subjects and juvenile-onset diabetics. Two insulin-requiring adult-onset diabetics showed significant beta-cell reserve during the course of acute infections. These results suggest that urine C-peptide provides a useful means of assessing beta-cell secretory capacity over a period of time and is especially advantageous when frequent blood sampling is not feasible.