Glomerular filtration rate and urine flow were measured in 40 babies of 26 to 40 weeks' gestation without the need for accurately timed collections of urine by infusing inulin continuously for greater than or equal to 24 hours. Estimates of GFR and urine flow were similar to those obtained when timed specimens of urine were also collected, the coefficient of variation being only 6.1% (22 studies). Although the plasma inulin concentration 80 minutes after a bolus and continuous infusion changed only slowly, it still closely reflected the size of the bolus dose (eight studies, P less than 0.001) and did not correlate with GFR (P = 0.48) as previously claimed. The GFR measured over two hours with a single injection of inulin correlated poorly with values recorded after a continuous 24-hour infusion of inulin (n = 5, P greater than 0.1). The agreement was closer (P less than 0.1) and the values lower (P less than 0.02) when the single-injection studies were extended to five hours. The half-life of inulin in the plasma of preterm babies was 4.3 +/- 0.8 hours (14 estimations in nine babies). This is longer than in adults because babies have a low GFR and a relatively large extracellular fluid compartment. The duration of some renal function test protocols needs to be extended because the long half-life leads to slower changes in plasma concentrations of test substances cleared by the kidney, and delayed equilibration.