In a recent issue of thisJournal, Mohabbat and Simos (1978) provided estimates of the rate of discount using Kendrick's (1976) series of total private wealth. Since Kendrick also provides estimates of the division of total wealth between human, Wh, and nonhuman, W,, wealth, it should be of interest to investigate how the discount rates differ between the two components of wealth. In order to accomplish this, it is necessary to divide the total income into two parts-one corresponding to the human wealth, Yh, and the other to the nonhuman wealth, Yn. Kendrick (1976, p. 22) and Christensen and Jorgenson (1973) provide an outline of the method for such a division of the total income. Following them, the time series for Y'h and Y1, were estimated. The estimated series closely approximate the estimates for "labor compensation" and "property compensation" given by Kendrick and Christensen and Jorgenson. As suggested in Mohabbat and Simos, the time-series estimates of the rate of discount for Wh and Wn were computed using the varying parameter regression technique. These series are given in table 1. The obvious inference that can be drawn from the results given in table 1 is that the discount rate on Wh is consistently higher than the rate on Wt. This evidence is important by itself. In the absence of direct empirical estimates, as provided here, economists had to make assumptions about the behavior of the two rates (see Pesek and Saving 1967). Additionally, since the major portion of WI, is education, the results provide further evidence on what Welch has called "one of the most important phenomena of our time . . . that the rates of return to investments in schooling have failed to decline under the pressure of rapidly rising average educational levels" (Welch 1970, p. 54). The existence of a relatively higher rate on Wh confirms the conclusions about the behavior of the rate of return on education arrived at by Griliches (1970, 1977), Welch (1970), Becker (1975), and others.