The peritoneal transport of five serum proteins and intravenously-administered neutral dextran was studied in 13 CAPD patients. In all patients a study was done three hours after the administration of dextran. In nine the study was repeated after 14 hours, and in six also after 38 hours. Using gel permeation chromatography of serum and dialysate it was possible to calculate fractional (relative to inulin) clearances for dextran fractions with Einstein-Stokes radii ranging from 30 to 90 A. In contrast to glomerular transport, we found that peritoneal clearances of serum proteins and corresponding dextran fractions were of a similar magnitude up to a radius of 90 A, especially in the follow-up studies. No evidence was found for a charge-selective barrier function of the peritoneum. It appeared likely that the peritoneal interstitium is probably a third compartment between blood and dialysate in which equilibrium with blood and dialysate is present for serum proteins, but initially not for dextran. For dextran this equilibrium is probably approached 14 to 38 hours after administration. The linear relationship between the reciprocal of fractional clearances and Einstein-Stokes radii indicates that restricted diffusion is the main transport mechanism of macromolecules across the peritoneal membrane. The slopes of the linear regression between the reciprocal of fractional clearances and radii ranged from 0.32 to 6.95 (slope for free diffusion 0.07) and provided an index of the permeability characteristics of the peritoneal membrane.