General Characters of the Disturbance. 33
never realized in actual earthquakes. In the first place, the
disturbance never originates at a mere point, but always along
an axis of greater or less extent, such as a fissure or fault.
The vibrations thus set up, in travelling to the surface,
undergo so many changes of direction by reflection and re-
fraction in passing through the very heterogeneous materials
of which the earth's crust is composed, that the motion of a
particle at the surface is, in all cases where the movement has
been traced by seismographs, of an extraordinary degree of
complexity, as may be seen by reference to any of the curves
published by the Japanese observers. Thus Prof. Ewing,
who, with Prof. Milne, has had the opportunity of recording
large numbers of earthquakes in the Plain of Yedo, Japan,
summarises his results as follows :—
"Automatic records given by seismographs confirm what
has been said, on theoretical grounds, as to the complexity
which earthquake motions may be expected to present. They
show that, as observed at a station on the surface of the
earth, an earthquake consists of a very large number
of successive vibrations—in some cases as many as three
hundred have been distinctly registered. These are irregular
both in period and amplitude, and the amplitude does not
exceed a few millimetres, even when the earthquake is of
sufficient severity to throw down chimneys and crack walls;
while in many instances the greatest motion is no more than
a fraction of a millimetre. The periods of the principal
motions are usually from half a second to a second, but, as
has been already said, the early part of the disturbance often
contains vibrations of much greater frequency. The earth-
quake generally begins and always ends very gradually, and
it is a noteworthy fact that there is in general no one motion
standing out from the rest as greatly larger than those which
precede and follow it. The direction of motion varies
irregularly during the disturbance—so much so that in a pro-
tracted shock the horizontal movements at a single station
occur in all possible azimuths. The duration, that is to say
the time during which the shaking lasts at any one point, is
rarely less than one minute, often two or three, and in one
case in the writer's experience was as much as twelve
minutes."17
17 Loc. cit., p. 13. Almost identical results are stated by Milne and
Gray in their paper on " Earthquake Observations and Experiments in
Japan," Phil. Mag., Nov., 1881, pp. 356—377.
D