ON THE EARTHQUAKE PHENOMENA OF JAPAN. 363: 
that in these diagrams the first sudden movement, which invariably has 
the appearance of a quarter-oscillation, ought apparently to be considered 
as a semi-oscillation. 
The second set of experiments consisted in determining the quantity 
to be calculated from an earthquake diagram which would give a measure 
of the overturning or shattering power of a disturbance. 
For this purpose a light strip of wood was caused by means of a strong 
spiral spring and a heavy weight to move horizontally back and forth 
with the period of the spring. On this strip small columns of wood were 
stood on end, and it was determined how far the spring had to be deflected 
and then suddenly released to cause overturning. 
Knowing the period of the spring and its amplitude, the — 
velocity v, the meantime acceleration - ,and the maximum acceleration . 
might be calculated. An acceleration, f, sufficient to cause overturning 
could also be calculated from the dimensions of the column. The result 
of various experiments showed that for small deflections f was nearly 
; NG 
equal to . , but for larger deflections the value for f lies between = 
2 
and». From this it seems that the overturning power of an earthquake 
can only be approximately determined from the dimensions of a body 
which has been overthrown. The quantity f is not the quantity v, or 
‘maximum velocity of an earth particle,’ so largely employed by Mallet 
and other earthquake investigators. Such a quantity, which may be 
determined for bodies of definite dimensions on the assumption that they 
are overturned by a sudden blow, so far as my experiments have gone, 
does not tell us anything about the nature of earthquake motion. 
An account of these and other experiments which I have from time to 
time been engaged upon will shortly be published as a part of Vol. VIII. 
in the ‘ Transactions of the Seismological Society.’ The more important 
results of all these experiments are as follows. 
In reading these conclusions it must be remembered that they only 
refer to experiments performed in certain kinds of ground. 
I. Effect of Ground on Vibration. 
1. Hills have but little effect in stopping vibrations. 
2. Excavations exert considerable influence in stopping vibrations. 
3. In soft damp ground it is easy to produce vibrations of large 
amplitude and considerable duration. 
4, In loose dry ground an explosion of dynamite yields a disturbance 
of large amplitude but of short duration. 
5. In soft rock it is difficult to produce a disturbance the amplitude 
of which is sufficiently great to be recorded on an ordinary seismograph. 
II. General Character of Motion. 
1. The pointer of a seismograph with a single index first moves in a 
normal direction, after which it is suddenly deflected, and the resulting 
diagram yields a figure partially dependent on the relative phases of the 
normal and transverse motion. These phases are in turn dependent upon 
the distance of the seismograph from the origin. 
