depth. High pore pressures require larger porosity changes to achieve undersaturation, but when undersaturation does occur, the velocity anomaly will be slightly larger.Some of the approximations we have made deserve a closer look. Crack aspect ratio has been shown to have a significant effect on wave velocities in rock (WALSH, 1969;NUR and SIMMONS, 1969;O'CONNELL and BUDIANSKY, 1974). While aspect ratio will not affect the occurrence of undersaturation, it will affect the magnitude of the velocity anomalies. We looked into this by reducing our constant aspect ratio porosity increase went from 0.67 at 1km depth to 0.44 at 10km, so the effect of a lower aspect ratio increases with depth (compare to Fig. 4A, line (2)). However, as HADLY (1976) points out, average aspect ratio should increase with depth in the crust as confining pressure closes flat cracks and leaves round pores relatively unchanged. Aspect ratios increasing with depth will have the net effect of increasing the anomalies at shallow depth in Fig. 4 and reducing the anomalies at greater depth.In our crustal models we have not included pore porosity which takes the form of large aspect ratio voids in rock. has estimated pore porosity to be on the order of 0.001 for many rocks and independent of depth. Our results would not be significantly altered if we had considered this pore porosity. At shallow depths the low aspect ratio crack porosity, of comparable magnitude (or greater), will have a much greater effect on velocity than the high aspect ratio pore porosity. It will slightly increase the dilatant strain required for undersaturation. At greater depths the pore porosity will have a greater relative effect on velocities because the net porosity increase is significant, but the high aspect ratios should keep the effect small. However, this additional porosity at depth will significantly increase the dilatant porosity increase required for undersaturation. Therefore, while the affects of pore porosity are somewhat variable with depth, it appears that the net effect will reinforce the results arrived at above, namely, that the potential for velocity anomalies decreases with depth.While our assumption of a hydrostatic pore pressure gradient seems reasonable, it conflicts somewhat with the conclusions of HANKS (1974, BOOKER (1975) and ANDERSON and WHITCOMB (1975), who suggest that substantially lower pore pressures may exist at depth. If this is so, then the pore fluid at depth will be closer to the liquid-vapor transition pressure than we have assumed, and undersaturation will be achieved with smaller porosity increases than we have calculated. The smaller porosity increases will create correspondingly smaller velocity anomalies and so we again arrive at the conclusion that the largest velocity anomalies will be found at shallow depths (assuming undersaturation occurs everywhere). It is, of course, possible to conceive of pore pressure and dilatant strain distributions that will cause undersaturation only at great depths. However, this would severely l...