Abstract. Severance of the ureter beyond the renal papilla causes a fall in urinary osmolality, which suggests that exchange of water or solute between urine and renal parenchyma normally occurs in the intact renal pelvis. We examined water and solute flux in the renal pelvis with micropuncture and microcatheterization techniques. Four groups of antidiuretic rats were studied. Group I (n = 17) underwent micropuncture through the intact contracting ureter. Urine samples were obtained at the papillary tip, and in the pelvis beside the base of the extrarenal papilla. Urinary osmolality at the base, 880±97 mosmol/kg H20 (mean±SE), was less than that at the tip, 1,425±104i mosmol/kg H20 (P < 0.005). In group II (n = 24), samples were analyzed for inulin and osmolality. In 15 rats (group IIA), comparison was made between base and tip samples. In thei other nine animals (group IIB), comparisons were made among base, tip, and bladder samples and urea was also measured. In group II (A and B combined) urine-to-plasma (U/P) osmolality was lower at the base, 4.31±0.27, than at the tip, 6.08±0.23 (P < 0.001), and U/P inulin was lower at the base, 192±25, than at the tip, 306±16 (P < 0.001). In group IIB, the bladder urine had a lower U/P osmolality, 5.27±0.25, than the tip, 6.01±0.31 (P < 0.02). The U/P urea was 59±10.6 (base), 98±9.4 (tip) (base vs. tip, P < 0.05), and 81±6.5 (bladder, P < 0.005, compared with tip). In group III (n = 8), samples were A preliminary report of these findings has been published in abstract form: Leonard, S., and R. Jamison. 1983 obtained by microcatheter from the fornices, the deepest intrarenal extensions of the pelvis, and compared with samples at the tip. Urinary osmolality was lower in the fornix, 646±106 mosmol/kg H20, than at the tip, 1,296±99 mosmol/kg H20 (P < 0.001). Similarly, U/P inulin was lower in the fornix, 48±14, than at the tip, 128±12 (P < 0.001). The lower U/P inulin in the pelvic urine is the result of either the addition of fluid to the pelvis, or the backleak of inulin across the epithelium lining the pelvis. To verify that the pelvic epithelium was impermeable to inulin, in group IVA (n = 4) the left renal pelvis was superfused with a solution of chemical inulin. Cumulative absorption of inulin from the left kidney was 0.15±0.08% of that superfused. Using ['4C]inulin in group IVB (n = 3), similar results were obtained (0.05±0.02%).These findings indicate that in the renal pelvis, fluid is added to urine after it emerges from the collecting ducts. We suggest that reflux of hyperosmotic urine over the renal papilla creates a transepithelial gradient for the flux of water into the pelvis. A model that incorporates diffusive and convective forces for water and solute transport is proposed to account for these findings.