Range3 223.15 0.4622 -0.0308 2.0 48.7-292.3 248.15 0.8968 -0.0273 1.9 55.7-391.2 273.15 1.1763 -0.0241 0.4 25.0-393.8 298.15 1.4513 -0.0221 0.9 0.1-353.5 0.1-392.4 3%. 15 1.7746 -0.0206 0.9 Values obtained by expressing h (l@D/rn*.~-~) as a linear function of (P/MPa). Lowest experimental pressure ( M R ) for which D ir prrdicted within eiprrmrntal uncertainty by Eq I , and highest erprirnental pressure. *Mean deviation (%)of experimental poinb from values calculated from equatiom to the lines d bpst fit.temperatures are plotted in Figure 1. The variation of D with P for n-hexane is typical normal liquid behavior. The slopes of the curves decrease considerably with increasing pressure, particularly at elevated temperatures. Clearly the pressure dependence of D is so strong that if data were available up to only 200 MPa, for example, it would be difficult to reliably extrapolate the curves to much higher pressures. The data at each temperature could no doubt be fitted to functions such as polynomials in P for extrapolation purposes, but (depending on the order of the polynomial) a considerable number of data points may be required, so that polynomials have limited value as predictive functions. Fits of the n-hexane data to Eq. 1 are shown in Figure 2, which demonstrates that apart from the low pressure (<50 MPa) region at low temperatures, the plots are linear up to the highest pressures attained in the measurements. The best-fit parameters of Eq. 1 are listed in Table 1. The estimated accuracy of the experimental data is f2.5%, so that for each temperature the mean deviation is smaller than the uncertainty in individual data points. Since Eq. 1 has only two adjustable parameters, as few as two experimental diffusion coefficients at pressures above say 50 MPa are adequate to make good estimates of D at other pressures. The general applicability of Eq. 1 suggests that it can serve also as a useful criterion of normal liquid behavior. The relationship does not hold for tracer and self-diffusion coefficients of water.However, the variation of D with P for water is abnormal in that with increasing pressure D first increases, then subsequently decreases (Woolf, 1975; Woolf and Harris, 1980). Work in progress on tracer diffusion of methanol has revealed a similar anomaly at low temperature. At higher temperature the diffusion of methanol in water approaches normal liquid behavior.