The velocity of propagation of concussion pulses has been calculated for a mountain of quartz, in which it would be 3.6 km. (2.2 miles) per second, very nearly the same as the last-mentioned figure. We should expect this, since the firm crust of the earth consists essentially of solid silicates—_i.e._, compounds of quartz endowed with similar properties. Measured at small distances from the origin, the velocity of propagation of the wave appears smaller, and the first preliminary tremor is frequently not observed. The velocity may be diminished to 2 km. (1-1/4 miles) per second. The reason is that the pulse partly describes a curve in the more solid portions of the crust, and partly passes through looser strata, through which the wave travels at a much slower rate than in firm ground; for instance, at 1.2 km. through loose sandstones, at 1.4 km. through the water of the ocean, and at 0.3 km. through loose sand. We recognize that it should be possible to calculate the distance between the point of observation and the origin of the earthquake from the data relating to the arrivals of the first preliminary tremor and of the principal shock of maximum amplitude. The violent shock is sometimes repeated after a certain time, though with decreased intensity. It has often been observed that this secondary, less violent, shock seems to have travelled all round the earth _via_ the longest road between the origin and the point of observation, just like one portion of the aerial waves in the eruption of Krakatoa (compare page 27). The velocity of propagation of this secondary shock is the same as that of the principal shock. Milne has deduced from his observations that, when the line joining the origin of the earthquake and the point of observation does not at its lowest level descend deeper than 50 km. below the surface of the earth, the pulse will travel undivided through the solid crust of the earth. For this reason we estimate the thickness of the solid crust at 50 km. The value is in almost perfect agreement with the one which we had (on page 16) derived from the increase of temperature with greater depths. It should further be mentioned, perhaps, that the density of the earth in the vicinity has been determined from pendulum observation, and that this density seems to be rather variable down to the depths of 50 or 60 km., but to become more uniform at greater depths. These 50 or 60 km. (31 or 37 miles) would belong to the solid crust of the earth. The movement of earthquake shocks through the earth thus teaches us that the solid earth-crust cannot be very thick, and that the core of the earth is probably gaseous. The similar conclusions, to which these various considerations had led us, may therefore come very near the truth. A careful study of seismograms may, we hope, help us to learn more about the central portions of the earth, which at first sight appear to be absolutely inaccessible to scientific research.