The need exists for a fast and accurate method of estimating glomerular filtration rate (GFR). A number of nomograms and formulae which estimate GFR on the basis of a plasma creatinine level have appeared in the literature. We examined some of these nomograms and formulae and showed that when the formula published by Cockcroft and Gault was compared with the GFR as estimated by 51Cr EDTA clearance, the coefficient of correlation was better than 0.94. We also compared creatinine clearance with the Cockcroft and Gault formula, and the Siersbaek-Nielsen nomogram. In each case, the correlation coefficient was less than that for the Cockcroft and Gault-51Cr EDTA comparison. Because of the inherent difficulties in making accurate 24 h urine collections, and problems associated with 51Cr EDTA studies, it is suggested that the formula prediction of GFR is more reliable than a GFR estimate using a 24 h collection.