In the past century, the homogeneous nucleation of light water (H2O) has repeatedly been studied using various experimental techniques. Generally, the onset of nucleation was recorded, while less frequently, the actual nucleation rates were determined. In contrast, the nucleation of heavy water (D2O) has been examined only in a single instance with no nucleation rates measured. Here, we report the first nucleation rate study of D2O along with nucleation rate measurements for H2O, which we repeated for comparison under identical conditions. We find that the nucleation rates for H2O and D2O differ by a factor of 2500, if compared at the same respective vapor pressure p(nu) and temperature T, whereas the comparison at the same supersaturation S shows an agreement within experimental scatter. Also, the numbers of molecules in the critical clusters, which are determined from the slopes of the In J versus In S curves, are nearly the same for both isotopic waters. A satisfactory agreement with previous nucleation rate measurements of H2O made by Viisanen et al. (Viisanen, Y.; Strey, R.; Reiss, H. J. Chem. Phys. 1993, 99, 4680; 2000, 112, 8205) is observed, if the onset supersaturations So at nucleation rates of J(0) = 10(7) cm(-1) s(-1) are compared. Using the most recent expressions for temperature-dependent vapor pressures, we calculated surface tensions and densities predictions by the classical Becker-Doring nucleation theory. Around T = 240 K, the predictions quantitatively agree with the experimental data. However, as in the case of other systems (e.g., alcohols and alkanes), classical theory shows a stronger temperature dependence than experimentally observed. A temperature-dependent correction of the classical theory is developed which permits analytical calculation of nucleation rates as function of supersaturation and temperature over extended ranges.