Marcu 31, 1923] 
NATURE 
433 

of the line of the Allah Bund, but that along that 
line the ground on the north was upraised by some 
_ twenty feet and on the south depressed by some 
fifteen at the time of the earthquake. 
- The next earthquake to be considered is one which 
has been investigated with great care and in great 
detail; it is the Californian earthquake of April 18, 
1906. In this there was a visible fracture, following 
what is known as the San Andreas fault, along which 
the shock attained its maximum intensity. This 
fracture was crossed at various points by roads and 
fences, and after the earthquake it was noticed that, 
where these crossed the line of fracture, they were no 
Jonger continuous, but the ends were shifted laterally 
distances which varied at different places, but 
frequently amounted to twenty feet. This was not 
all, for the displacement was of a very curious nature, 
revealed by surveys of the displaced fences, and by a 
epetition of the original trigonometrical survey of the 
esion. These showed that for a distance of some 
miles back from the fault line the stations had been 
displaced, those nearest to the line by the greatest 
amount, which lessened as the distance increased and, 
t about five miles or so to the east, became very small. 
foreover, it was found that the movement on the 
eastern side of the fault had been to the southwards, 
and on the western towards the north. 
Here we have a result very like that in Cutch; in 
both cases there was a well-defined line along which 
permanent change of position of the ground took place, 
imultaneously, in opposite directions on either side 
of the line of separation, and in both cases the displace- 
ment decreased in amount with increasing distance, 
till it ceased to be measurable at a distance of about 
-a-dozen miles. The only difference was that in 
alifornia the displacements were horizontal, with 
e or no change of level, while in Cutch they were 
vertical, whether accompanied or not by horizontal 
shifting cannot be known. 
Paradoxical though it may seem, this movement on 
pposite sides of the fracture in opposite directions is 
quite in accord with known physical principles. If 
any block of material is compressed or stretched in 
ie way, while free to expand or contract in a trans- 
se direction, or if it is twisted by two opposite sides 
g forced in opposite directions, a complicated 
stem of strain is set up, and if the strain is more 
an the material will bear, disruption will take place, 
On opposite sides of which the material will move in 
opposite directions. 
Models to illustrate this principle have been con- 
tructed by others and myself, and from these con- 
iderations there has arisen what is known as the 
astic rebound theory of earthquake origin, and as 













































rowth of strain and a sudden release by fracture. 
The former, however, is by no means necessary, and 
same result, as regards displacements of the ground, 
vould be attained if the strain was rapidly, even 
suddenly, produced. There are, in fact, reasons for 
pposing that the growth of strain is not slow but 
apid, yet the fracture and elastic rebound theory 
might be accepted as sufficient, if earthquakes could 
always be attributed to a single fracture, or to a close- 
t group of fractures ; but in the case of great earth- 
NO. 2787, VOL. 111] 
enerally expressed this takes the form of a very slow, 
quakes, and sometimes of minor earthquakes also, the 
interpretation is put out of court by a study of the 
distribution of the intensity of the disturbance. 
In illustration I may take first the great Indian 
earthquake of June 12, 1897, in which the central 
region of greatest intensity covered an area of about 
140 miles long by 4o miles broad, over which there was 
a complicated series of faults, fractures, and distortion, 
which was certainly widely different from the com- 
paratively simple origin generally assumed for earth- 
quakes. This seemed at the time sufficient to account 
for all the facts, though there were some recorded as 
difficult to explain, and later examination seems to 
have established the conclusion that the origin of the 
earthquake cannot be limited even to this extensive 
area. In this earthquake only two of the isoseists 
could be plotted in detail, those of eight degrees and 
of two, or the extreme limit at which the shock could 
be felt; both exhibit considerable irregularities of 
outline, the most conspicuous of which is a pronounced 
projection to the westwards, and on the continuation 
of this line is a detached area, where the shock was 
again felt, after a gap where it was not felt. Col. 
Harboe has suggested, from a study of the recorded 
times, that there was an extension of the origin along 
this line, and though his plotting of the origin cannot 
be accepted in detail, I am convinced that, in the main, 
his conclusions are correct, for they very materially 
help to explain some peculiarities of the recorded 
observations, which remained inexplicable on the older 
and more generally used interpretation. 
From this it appears that in earthquakes covering 
a large area we are not dealing with a simple disturb- 
ance, starting from an origin of restricted dimensions 
and propagated outwards, but with one of complex 
origin ; and that in the outer regions of the seismic 
area the disturbance may be compounded of wave 
motion propagated from a more or less distant origin, 
where the initial severity was great, and of that coming 
from a nearer origin, of a lesser degree of severity, so 
that, instead of a fracture of at most a few tens of miles 
in length, we have to deal with a cobweb-like system 
of fractures, or something analogous, which may run 
to hundreds of miles. 
The general drift of the argument I wish to set forth 
is probably best illustrated by the California earth- 
quake of 1906. In this the greatest degree of violence 
was found along the line of the San Andreas fault, but 
the plotting of the isoseists shows that there was not 
only an independent centre in the San Joaquin valley, 
some forty miles to the eastwards, but also several inde- 
pendent centres of great intensity at lesser distances from 
the San Andreas fault. Moreover, the displacements 
recorded by trigonometrical survey make it probable 
that other similar independent centres would be found 
to the west, if the waters of the ocean had not made 
observation impossible. The records, therefore, indi- 
cate a set of separate centres of disturbance, scattered 
over a region of about three hundred miles in length 
by very possibly one hundred miles in width, and these 
separate centres, though independent as regards the 
surface shock, were all evidently connected with some 
common cause. Had they been the result of breakage 
under a slowly growing strain it is difficult to under- 
stand how so complicated, scattered, and extensive a 
N2 
