Human pregnancy is associated with substantial increments in glomerular filtration rate (GFR) and renal plasma flow (RPF). We have previously demonstrated that permselectivity to neutral dextrans is altered in pregnancy, theoretical analysis of the dextran sieving curves suggesting that elevated GFR is due to increased RPF and decreased glomerular oncotic pressure (pi(GC)) with no evidence of increased transglomerular hydrostatic pressure difference (DeltaP). These conclusions have been challenged, with claims that the rise in GFR is primarily a result of a decrement in pi(GC). With refined laboratory and infusion protocols, we have reexplored the determinants of ultrafiltration in a serial study of 11 healthy women in late pregnancy (LP) and 4 mo postpartum (PP), both in the baseline state and after increasing GFR and RPF by infusion of amino acids. Results were analyzed using two computer modeling programs. Increased GFR in LP (38%, P < 0.05) was due to a combination of elevated RPF (22%) and a decrement in pi(GC) and associated with an increased ultrafiltration coefficient, without evidence of increased DeltaP, and additional amino acid-provoked GFR increments (P < 0.05) produced similar findings. In addition, refined methodology permitted collection of sufficient data on excreted large-radii dextrans (>60 A) to better define the nondiscriminatory "shunt" pathway (omega(0)) and the standard deviation of pore size (S) about the mean radius of the distribution. Thus it was possible to demonstrate that the physiological increase in total protein excretion in LP is associated with a prominent shunt and an upward shift in breadth of distribution of pore sizes. This ability to quantify omega(0) and S will now permit better evaluation of the pathophysiological changes in the glomerulus associated with pregnancy in women with renal disease and in gravidas developing preeclampsia.