94 REPORTS ON THE STATE OF SCIENCE.—1912. 
days, there were 976 megaseisms, or on the average one every 4°1 days. 
Of these 117 at least originated on land, and the remaining 859 beneath 
oceans. The suboceanic activity was therefore seven times that on 
land. The average for the latter was one megaseism for every thirty- 
four days, but beneath the ocean one every 4°6 days. 
Frequency in Districts F,, F, and F, (East Indies).—These dis- 
tricts are taken collectively for five years, 1889 to 1893. I select these 
three overlapping areas because at the present time they are more 
active than any other. In the time considered, 1,826 days, sixty 
megaseisms originated in them. The average frequency was therefore 
one disturbance in thirty days. If we regard those disturbances which 
occurred within an interval of less than five days of each other as being 
parts of one effort, the number sixty is reduced to forty-eight, and the 
average frequency becomes one disturbance in thirty-eight days. The 
time intervals between these forty-eight efforts expressed in days were as 
follow :— 
6 8 8) UL0) St LL I pele 12: 133 — — 15 
US Se Ss Sees 6 19 YOO eo: pede 23 23 23 25 
30%e Sky (3) -3L ) Ol p32) 43323456 36 37 39 42 
47 --4j_ +54 57°60 62 65 93 124. 129 132 
One inference which may be drawn from these figures is that in 
these districts the time taken to bring about conditions of seismic insta- 
bility lies between eighteen and thirty-one days. 
Frequency in Districts EH, and EH, (Hast of Japan).—In the five 
years 1889 to 1893 forty-one megaseisms originated in these districts. 
The average frequency was therefore one disturbance in forty-four days. 
Four of these disturbances may however be regarded as parts of single - 
efforts. With this assumption, the average frequency becomes one 
disturbance in forty-nine days. 
The time intervals between these efforts expressed in days are as 
follow :— 
BEB PAI TO PRO OTE! VUSHOTET MLS AL ONe QE 4 26 Fee 
30 32 36 38 39 43 43 46 46 46 47 50 
57 60 64 87 91 102 105 109 138 160 
The time interval required to bring about seismic instability may 
lie between thirty-six and forty-seven days, but it is not so well marked 
as it is in the East Indies. 
VIII. Earthquake Periodicity. 
If we plot the megaseisms which have been recorded since 1889 on 
a sheet of paper ruled to show the days of the year (see Plate II.), it 
is seen that these disturbances have occurred in groups separated by 
periods of rest. The number of groups in a year have varied from eight 
to seventeen. Between November 23 and December 24, 1889, there was 
a period of quiescence extending over thirty-one days. If we start on 
December 14, which is in the middle of this period, we find that every 
successive 443 days we arrive at other periods of rest. These are met 
with on the following dates: February 28, 1901, May 17, 1902, August 
4, 1903, October 22, 1904, January 17, 1906, April 1, 1907, June 17, 
