38 STUAHT : THB, SPJIMANGAL EARTHQUAKE OF 8TH JULY 1918. 
little doubt that it was considerably affected by the earthquake. 
The following is the rate of error (seconds gained per day) of that 
clock, from observations of well-placed stars for the period around 
8th July, 1918:— 
Seconds. 
June — 
5th +0-J50 
7th +0-6-1 
11th +0-54 
20th +0-G2 
2.5th +0-42 
J lily — 
.5th +048 
9th -002 
10th +0 05) 
15tli +0-3C 
10th +0-.51 
To explain the above, the following is given for illustration. The 
error of the clock was obtained by star observations on June 5th 
and again on June 7th. The difference between these errors 
divided by the interval in days between the times when the obser- 
vations were taken gives 0 64 seconds as the daily rate of error 
during that interval, and it is put down against the 7th June. On 
studying the above list the big jump in rate against the 9th July 
is at once obvious, as is the gradual recovery to the old rate during 
the week following. The change in rate revealed by the obser- 
vations of the 9th was so great, that observations were taken again 
on the following night. The chronometers which are systematically 
compared with the clock also showed that something happened to 
the clock between 10-30 a.m. and 7 p.m. on 8th July. 
The above is not merely of general interest. In view of the 
distance of Madras from the earthquake centre at Srimangal and 
the time that the irregularity of beat of pendulum occurred it seems 
that the disturbance to the Madras clock had nothing whatever to 
do with the long waves from the Srimangal centrum, but that the 
disturbance was caused by earthquake waves emanating from some 
other centrum such as that of the sympathetic shock which was dis- 
cussed in the prehminary report as having occurred near Madras 
and which will be discussed more fully hi a later portion of this 
Memoir. 
