Published data on the renal clearance of creatinine, p-aminohippuric acid (PAH) and kanamycin in relation to glomerular filtration rate (GFR) in patients with various renal diseases were analysed by a physiological model of renal clearance. Fitting of the data by the general linear equation representing the model proposed by Levy [10] resulted in insignificant intercepts with the ordinate, indicating the unsuitability of the model for the detection of tubular secretory activity. Use of this model also did not lead to significant improvement in goodness of fit compared to simple proportionality of renal clearance and GFR. On the other hand, parameter estimates of the physiological model obtained from the data by nonlinear regression analysis revealed statistically significant tubular secretion both of PAH and creatinine. The much lower tubular secretory activity estimated from the kanamycin data did not reach statistical significance. For compounds exhibiting statistically significant tubular secretion, use of the physiologically based relationship between renal clearance and GFR significantly improved the goodness of fit to the data as compared to simple proportionality of both variables. It is concluded that analysis of the relationship between renal clearance of drugs and GFR using the physiological model of renal clearance can contribute to our knowledge of drug handling by the kidney, and may facilitate drug classification according to total extraction by this organ.