ON SEISMOLOGHCAL INVESTIGATIONS. 71 



Another example is afEorcled by the earthquake of June 25, 1914, 

 details of which have already been published in the Shide Bulletin for 

 June. Let us first take the following residuals : 



It is clear that if Eskdalemuir is even approximately correct, the 

 S for West Bromwich is sensibly in error. Now, the West Bromwich 

 trace shows, following the printed S at 19 31 45, another disturbance 

 about 45 s. later, say at 19 32 30, M^hich is probably the true S. The 

 printed A for West Bromwich is 100-3°, for which Y— ^S) would be only 

 —22 s. according to the figures above given. But the position of the 

 epicentre is subject to revision. The printed distance 100 3° would 

 give S— P=ll m. 45 s. (11 m. 22 s. corrected), whereas the West Bromwich 

 trace suggests the smaller value 11 m. 13 s., corresponding to A=(98-3°) 

 say : for which Y-(S)=(-35 s.). 



For this earthquake the S residuals for Florence, Aachen, Barcelona, 

 Dyce, Honolulu, &c., all accord with a mistake of Y for S. On noticing 

 the above discrepancy between Eskdalemuir and West Bromwich, inquiry 

 was made of Eskdalemuir, and on April 23 Mr. Richardson kindly replied 

 ' the time stated (for S) in our tabulation was the correct one for a certain 

 disturbance on N.-S. But on unravelHng the E.-W. trace it is seen to 

 contain a disturbance about a minute earlier. Photo, prints are being 

 sent to you herewith.' A rough tracing of the three records is given 

 in the illustration. It will be seen that Y is small on the N.-S. component ; 

 large on the E.-W. component, and quite noticeable (though S is absent 

 entirely) on the vertical component. These facts seem to accord with 

 a wave of P character, for the epicentre is nearly due east of Eskdalemuir, 

 the azimuth (from the north point) being 82-6 if we accept the epicentre 

 of the June Bulletin (4-5° S., 99° E.). A thrust from this direction would 

 be 7-7 times as much to the W. as to the S. Now measurements of the 

 traces give for the first three south movements in milhmetres +8-5, —12-0, 

 +9-0; and for the corresponding west movements, +57, — 72*5, +65. 

 Dividing the latter by 7-7 we get +7 4, —9-4, +84, which are all a little 

 smaller than the former. The ratio (57+72-5+65)/(8-5+12-0+9-0)=6-6 

 in fact, giving an azimuth for the epicentre of 81-5° instead of 82 6°. 

 The epicentre determined at Eskdalemuir (1° S., 102° E.) gives azimuth 

 78°, so that the correspondence is thus well within the limits of acci- 

 dental errors of various kinds. It would be interesting to check this 

 ratio from the S movements, which ought to be, and are, large in the 

 upper trace where Y is small, and small in the lower trace. But, un- 

 fortunately, they are masked in the lower trace by the end of the Y move- 

 ments ; at any rate this is a reasonable interpretation of the trace. In 

 fact, the epicentre is in a particularly favourable azimuth for separating 

 Y from S, and the sentence above quoted from Mr. Richardson's letter 

 takes on a new significance when this is realised. If the azimuth had 

 been nearer 45° or one of the other octants, Y and S might have been 

 mixed up in both traces, and the beginning of Y would probably have 

 been read as the beginning of S, The possible eSect of this confusion 



