We tested the accuracy of the micropuncture technique by performing total collections of tubular fluid followed by immediate recollections from the same site. We studied thirty-one Wistar rats under different conditions of hydration and during maintenance infusions. To assure constancy of body fluid volumes the experiments were performed during continuous reinfusion of urine. Overall we performed 190 paired collections and immediate recollections from 147 proximal and 43 distal tubules. The mean values measured during the first collection were not different from the paired means obtained by recollection. Mean values (+/- S.E.M.) for first collection and recollection were: single nephron glomerular filtration rate (SNGFR), 34.7 +/- 1.3 vs. 34.2 +/- 1.3 nl min-1, P > 0.42; percentage reabsorption, 74 +/- 1 vs. 73 +/- 1, P > 0.53; absolute reabsorption, 26.4 +/- 1.2 vs. 25.9 +/- 1.2 nl min-1, P > 0.47; collection rate, 8.4 +/- 0.6 vs. 8.3 +/- 0.4 nl min-1, P > 0.92. When subdivided according to different physiological conditions and sampling sites, the results for the paired collection-recollection means were still not significantly different. The regression between collected and recollected SNGFRs had a slope of 0.96, r = 0.85, P < 0.0001. In order to exclude the possibility that urine reinfusion per se could obscure putative differences in collection-recollection pairs, the effect of urine reinfusion was separately evaluated. We measured the differences obtained in collection-recollection pairs before and during urine reinfusion in twenty-eight tubules from eleven rats, a sample size that allows the detection of 10% difference as significant. While SNGFR did not change, percentage reabsorption fell significantly from 76 +/- 3% before infusion to 60 +/- 3% during reinfusion, P < 0.003. We conclude that the recollection technique yields a reproducible estimate of SNGFR and tubular reabsorption, independent of the sampling site and of the technique used for fluid maintenance. Thus, it can be used to study the effect of different manoeuvres on the rate of proximal tubular transport. Urine infusion per se depresses proximal transport compared with that measured during maintenance with equivalent amounts of isotonic saline.