8 REPORT OF THE CALIFORNIA EARTHQUAKE COMMISSION. 
apparently too large. The 15 seconds which Professor Townley allowed for the interval 
between the arrival of the shock at Ukiah and the moment that he read his watch seems 
to me rather long; if we reduce this interval to 12 seconds and take the time of arrival 
at Ukiah at 55 12™ 20°, the observations become more accordant; we find that the best 
position for the origin has the same geographical position mentioned above,: but the 
centrum is near the surface, and the time of the shock becomes 5" 11™ 59°; the errors 
of observations are: San Francisco, + 1.1; Berkeley, — 2.8; Mount Hamilton, + 0.9; 
Ukiah, + 0.7. The sum of squares of the error is 10.4 seconds. But the position of 
the centrum can not be lookt upon as being determined very accurately; even if we put 
its depth at 30 km. we find the sum of the squares of the errors only 10.8 seconds; and 
the individual errors are: San Francisco, + 1.9; Berkeley, — 2.6; Mount Hamilton, 
+0.6; Ukiah, + 0.2. This is a better group of errors, as that of Mount Hamilton is 
very small. The time is 5° 11™ 57.7%. 
It will be noticed that the groups of errors seem slightly to favor the idea of a simul- 
taneous slip along the fault-line in preference to the slip beginning over a small area and 
then gradually spreading along the line. But let us notice what this really involves. In 
the first place it requires a velocity of propagation of only 1.8 km./sec., a value less than 
a quarter as great as the most probable value of this velocity.t. It may be urged that 
these times refer to the arrival of the large surface waves, whose velocity has been deter- 
mined as about 3.3 km./sec.; but with this value of the velocity we find much larger 
errors, the sum of the squares amounting to 36.9; and therefore a consideration of the 
errors alone renders this supposition less probable than either of the other two; and, 
moreover, the preliminary tremors and large waves are not separated at such short dis- 
tances from the origin as San Francisco and Berkeley. 
Secondly, it is clear from the surveys of Messrs. Hayford and Baldwin (vol. 1, pp. 114- 
145) and from the discussion of them (pp. 16-28) that the rupture along the fault-line was 
the result of gradually-increasing forces which finally became greater than the strength 
of the rock; before rupture the rock yielded elastically to the forces and it seems abso- 
lutely impossible that its ultimate strength, varying locally, should have been reached 
simultaneously over the whole area of the fault-plane, whose length was 435 km., or in- 
deed over any large area. It would require a nice adjustment of the forces concerned, 
which the nature of the forces in no way leads us to expect. It is only in the case of 
absolute rigidity, which is far from the true nature of rock, that we can conceive of a 
simultaneous movement along the whole fault ; and then we should be at a loss to account 
for the dying out of the fault at its ends. Moreover, our general experience is entirely 
against simultaneous yielding; when structures, such as bridges, break, they give way 
first at a particular point; when an ice-jam in a river yields, one part yields before the 
rest; and, indeed, many such examples might be cited. We are therefore constrained 
to believe that the rupture on the fault-plane began over a small area and rapidly spread 
to other parts of the fault. 
We may then consider the position of the origin as determined within, perhaps, 30 km. 
along the fault-line and within 20 km. in depth; and the time, within 3 seconds; and we 
may write for 
t, = 13" 11™ 58° + 3 seconds G. M. T., 
== fe) , , 
The beginning of the shock i ia ae: en aie , 
z, =10 km. + 20 km. or — 10 km., 
where ¢, is the time of the occurrence of the shock; A, the longitude, and ¢, the latitude, 


‘If we had taken the Ukiah time as 5" 12™ 205, the hypothesis of simultaneous slip would have 
required a velocity of 1.6 km./sec., and the sum of the squares of the errors would have been 17.4. 
