The objective of this magnetic resonance imaging (MRI) study was to (1) test the validity of the ellipsoid formula for estimating kidney volume using ex vivo and in vivo models and (2) establish a normal range of values for kidney length and volume in patients with no known history of renal disease. The volumes of five excised porcine kidneys were measured by (1) disc-summation method, (2) ellipsoid formula, and (3) water displacement method. In a retrospective, consecutive group of clinically referred patients (n ؍ 150; 300 kidneys), individual kidney volume and length were calculated by the disc-summation method and by multiplanar reformation of MRI data, respectively. For comparison, kidney volumes also were calculated using the ellipsoid formula in all patients. Renal volume that was obtained by MRI using the disc-summation method was within 5% of the volume that was determined by the water displacement method, independent of the spatial resolution of the MRI technique used. Data from both the in vivo and the ex vivo models revealed that the ellipsoid formula that commonly is used in ultrasonography underestimates renal volume by 17 to 29% compared with the disc-summation method (P < 0.05). As measured by MRI (mean ؎ SD), kidney lengths were 12.4 ؎ 0.9 cm for men and 11.6 ؎ 1.1 cm for women, and kidney volumes were 202 ؎ 36 ml for men and 154 ؎ 33 ml for women. The results from the ex vivo MRI study show that the kidney volume that was obtained using the disc-summation method is within 5% of the true kidney volume as measured by the water displacement method. The ellipsoid formula consistently and significantly underestimates the true kidney volume. The length and the volume of kidneys that are obtained by MRI in patients with no known history of intrinsic renal disease are greater than the commonly quoted reference values that are obtained by ultrasonography.