ON SEISMOLOGICAL INVESTIGATIONS. 



61 



Table XVI. — contimied. 



The observed S (i.e., the interval by which it follows 8^ 44™ 49') is 

 given in the third column, and it is compared with the adopted (Shide) 

 tables in the next column — Ci; except that in the latter part of the 

 table this comparison has been omitted when it obviously fails. The 

 corrections, taken for A>45° from the last two Eeports, and for 

 A< 45° by use of the factor 1"80 on the corrections for P in Table V., 

 are given in the column ' Corr.,' and applied in the column O — Ca. 

 When A > 90° a large number of records will not fit S at all, but at 

 first agree with the phenomenon Y or polychord suggested in the last 

 Eeport. A comparison with the times suggested for Y is therefore 

 given in the last column O-Y. We now take in order certain matters 

 brought out by this table. 



(a) There are three records near the epicentre for which no explana- 

 tion has as yet suggested itself, viz. Manila, Taihoku, and Tokyo. 

 They may, of course, be mistakes, but there is a systematic character 

 about them which seems opposed to the idea of mistakes. The average 

 velocities are 19^"5, 18^'9, and 17*'9 per degree of A, intermediate 

 between those of P and S, and it may ultimately be found possible to 

 assign some combination of P with S which shall explain the records; 

 but up to the present no success has been attained in this direction. 



(j8) With these three exceptions all the records for stations up to 

 A =95° are brought into fair accord by the suggested corrections. 

 Particularly noteworthy are the records for Zi-ke-wei, Osaka, and 

 Tsingtau near A ==30°, where the correction is near one of its maxima 



K 2 



