110 REPORT OF THE CALIFORNIA EARTHQUAKE COMMISSION. 
An explanation has been sought by supposing that waves of various periods are present 
in the disturbance and that they are propagated at various rates, just as light waves of 
different wave-lengths travel at different speeds in transparent substances. Altho the 
slow periods of the regular waves change into the quicker periods of the principal part, 
this change does not seem to continue during the remainder of the disturbance; nor 
has a similar change been discovered during the first two phases. 
A NEW EXPLANATION. 
The passage of sound thru air suggests a better analogy. A strong sound, like the firing 
of a cannon or a clap of thunder, is not heard at a distance as a sharp noise, but is ac- 
companied by a rumbling that lasts for many seconds; this is due to reflections and 
refractions of the sound at the surfaces of many layers of air of varying temperature, etc. 
Now the material of the earth for a few kilometers from the surface consists of rocks of 
varying density and elasticity; and when an elastic wave crosses the bounding surface 
between two different materials, it is in general split up into four waves, reflected longi- 
tudinal and transverse waves, and refracted longitudinal and transverse waves. When 
the reflected waves, returning, meet a boundary between different kinds of rock, they are 
again reflected and send waves forward, which are, however, retarded behind the original 
wave. In this way, by repeated reflections and refractions, a large part of the energy of 
the original wave would be, as it were, stored up in the heterogeneous surface layer of 
the earth and be slowly given out, thus keeping up a continuous supply at the surface for 
a limited time. 
If the whole earth were sufficiently heterogeneous, we should not have, at distant sta- 
tions, the distinction between first and second preliminary tremors, for there would be 
thruout the whole course of the waves such frequent transformations from longitudinal 
to transverse waves and vice versa, that they would arrive at a distant station thoroly 
mixed, and the supply of energy there would be fairly continuous, without the sudden 
variation which actually marks the arrival of the second phase. But we believe that, 
with the exception of a surface layer a few kilometers thick, the earth is fairly homo- 
geneous, or, rather, without sudden changes in density or elasticity; and that an earth- 
quake will set up both longitudinal and transverse vibrations, which will travel at differ- 
ent speeds and become entirely separated from each other in the homogeneous interior. 
When the longitudinal waves reach the heterogeneous layer near the surface they will 
be broken up; at every refracting surface both longitudinal and transverse waves will 
be sent forward; as the former always travel the faster, they will arrive first at the 
earth’s surface ; but, in general, the transverse waves, set up at the last refracting sur- 
face, will not be far behind them. The proportion of longitudinal and transverse waves 
in the first preliminary tremors, at a given station, will probably depend upon special 
characteristics of the rock in the neighborhood and also on the direction from which the 
waves come; for transformations depend on the angle between the vibrations and the 
refracting surface. In regions of stratified rocks such surfaces are very numerous and 
are usually parallel with each other; their influence would vary in accordance with 
the direction in which the vibrations met them. It might thus be possible for the 
first preliminary tremors to consist almost wholly of longitudinal waves, or to consist 
of both kinds equally; but it does not seem possible that transverse waves could pre- 
dominate in them. ; 
Let us now turn our attention to the group of transverse waves traveling by them- 
selves in the homogeneous interior of the earth. They fall farther and farther behind the 
longitudinal waves; when they reach the heterogeneous outer layer they also suffer trans- 
formations, giving rise to both longitudinal and transverse waves, and these, by continual 
