The precision of renal mean transit time (MTT) measurements by 99Tcm-DTPA scintigraphy was studied by a Monte Carlo procedure. Data were obtained from twenty mild essential hypertensives with normal renal function. An ensemble of equivalent data sequences was constructed for each renal time-activity curve (TAC), assuming the dominant noise in the data to be Poisson in character. TACS were deconvolved by the matrix method and MTT was inferred from retention functions by a procedure entailing a nonlinear least-squares minimization. For every ensemble of data sequences a distribution of MTTs was generated. The population-averaged coefficients of variation for these distributions were 1.6% and 2.6%, for whole-kidney and cortical data, respectively. We conclude that the precision of MTT determination from clinical data is comparable to that from simulated data. MTT measurements may be useful for detecting functional changes in individual kidneys following an intervention such as angiotensin-converting-enzyme inhibition.