ON SEISMOLOGICAL INVESTIGATIONS, 97 
given interval, but not by any means exactly. The period found can 
then only be regarded as an average period, and the deviations in the 
table are intelligible. But it is noteworthy that Mr. Chandler detected 
a fifteen-month period in the Latitude Variation,” though the coefficient 
is extremely small (only 0”.03): and some preliminary calculations 
seem to show that the level error of the Greenwich transit circle also 
has an inequality: of this period. These matters are being further 
investigated. ; 
X. Intervals in Days from the Commencement of one Group to the 
Commencement of another. 
1899 . . . 47 21 24 21 28 32 25 33 39 41 33 
1900 . . . 658 13 27 69 79 40 33 39 
1k) . 20 38 18 13 19 51 73 32 31 31 28 20 
1902. . . 17 28 16 20 21 11 54 24 28 19 30 60 28 
1903. . . 22 11 18 23 26 37 15 20 22 30 19 24 46 43 22 
1904. . . 13 58 15 15 31 55 29 16 16 15 16 14 17 #11 16 
11 17 
1905 . . . 32 29 19 10 41 15 22 30 24 33 20 48 20 14 6 
7 
1906. . . 37 33 18 18 10 75 21 10 26 29 15 12 14 12 43 
1907 . . . 15 33 51 18 20 22 30 35 7 28 20 10 41 23 18 
1908 . . . 19 16 32 10 21 7 22 11°93 #40 22 20 28 12 
eee es LO. Ode Sak EGP AGS 1S, TOT HW LOWASo NSE. 1.) Tie 242 
A table of the above intervals shows that those of from fifteen to 
twenty-two days recur no less than fifty-five times, whilst those between 
eighteen and thirty-three days are repeated twenty-seven times. Inter- 
vals of seven, ten, eleven and twenty-four days are each repeated five 
times ; the remaining intervals only occur or recur once, with the excep- 
tion of the intervals thirty-seven and forty-one days, which recur twice. 
The inference is that for the world as a whole seismic strain usually 
finds relief every fifteen or thirty days. In other words there is fre- 
quently a rough regularity in the recurrence of megaseismic groups. 
X1.—Intervals and Days between Successive Megaseisms in 
Particular Districts. 
In the eleven years 1899 to 1909, off the East Coast of Japan to the 
North of Tokio, I find that thirty-two megaseisms were recorded. 
Twelve of these were separated by intervals lying between fifty-seven 
and ninety-five days. 
To the South of the Philippines round the Celebes and to the West 
of N.-W. New Guinea during the same period forty-two large earth- 
quakes originated ; fifteen of these are separated by intervals which lie 
between fifty-six and ninety-two days. 
XII.—Geographical Distribution of Megaseisms and Thermometric 
Gradients. 
For the five years 1899-1903 we have a list of 818 megaseisms, the 
origins of which are known; of these, sixty-one originated on continental 
areas, and 252 originated along the lines of troughs or ‘ deeps,’ beneath 
? Astr. Journel, No. 523, p. 152. 
1912 H 
