Disparate reports on intrarenal blood flow distribution prompted a direct comparison between microspheres (Ms) and inert diffusible tracers (DT). The "tissue sampling technique" for estimating local flow with DT (Kety) was adapted for the dog kidney, using 125I-iodoantipyrine (1-Ap) and tritiated water (THO). Ms (15 micron) were injected 2-3 min prior to 10-15 s DT infusion made during continuous 1 s arterial blood sampling. Tracers were measured in 7 to 20 samples from each of the following zones: Outer, middle and inner cortex (C1, C2, C3), outer and inner halves of outer medulla (OM1, OM2), and inner medulla (IM). I-Ap and THO gave closely similar flow distribution, and average total renal blood flow (RBF) of 3.90 and 3.78 as compared to 3.94 ml/min . g with Ms. Flow in C2 (ml/min . g) was similar with all tracers, and in per cent thereof average local flows were: C1 102, C3 70, OM1 34, OM2 12, and IM 2 with DT versus 117, 53, 12, 3, and 0 with Ms. Zonal flow fractions of total RBF obtained with DT were: C1 0.41, C2 0.33, and C3+medulla 0.26 versus 0.51, 0.33 and 0.16 with Ms. Thus, a Ms surplus in C1 relative to DT flow, representing 10% of total RBF, matched a Ms deficit in C3+medulla. This disparity might result from: (1) Failure of Ms to enter deep afferent arterioles in proportion to blood flow, (2) diffusion of DT from deep portions of the interlobular arteries, and/or (3) postglomerular inward flow of blood and DT.