Abstract
Three years' observations of Tonga-Fiji-Kermadec earthquakes at stations in New Zealand have been analysed by the statistical techniques applied to Herrin's “1968 Seismological Tables for P Phases”, to yield smoothed regional travel times for P and S. The travel-time curve for P is later than Herrin's until it merges with it at about 30°, implying lower velocities than those of Herrin's model at shallow levels in the upper mantle, and higher velocities at deeper levels. The travel-time curve for S crosses the Jeffreys-Bullen curve at 21° and 27° and has a greater curvature in this range. A correlation between the isostatic gravity anomaly and the mean travel-time residual at the stations is tentatively explained in terms of higher velocities in the deep seismic zone.

This publication has 5 references indexed in Scilit: