10 REPORTS ON THE STATE OF SCIENCE.—1918. 
TaBLeE VI.—continued. 
Errors of P deduced from PR,. 
| 4 | From PR, |From (23) Old Tables ap New | Possible 8P,| Corn. 
gen | —io | —6 | 446 || 436 | 437 ag 
| 20 | 21 21 
AO ed).p = oy 466 457 458 se) 
| 21 | 22 21 
a1 god: 8 a 487 | 479 479 ig 
. | 19 20 20 | 
i ar 6 506 499 499 Sg 
| 18 | a | 20 
9488 ko 4 a4 524 | 5200 | 9 ae 
We see that the investigation of the 22nd Report (shown in the third 
column of Table VI.) fails for the smaller values of A in much the same 
way as is manifested in Table II. above ; viz. from A = 10° to A = 20° 
the errors are not sufficiently positive if we may accept the indications 
of PR,, while near 4 = 35° they are not sufficiently negative. The 
column deduced from PR, starts with a hill and falls into a valley; in 
the investigation (22) the hill is cut down and the valley filled up; and 
it is reasonable to attribute these modifications to the effects of compro- 
mise due to the determination of the epicentre with the faulty tables. 
In the fourth and succeeding columns of Table VI. the effects of 
applying the newly-suggested corrections are shown. First, under the 
heading ‘ Old Tables’ are given the tabular values for P, followed by the 
differences 8P,;. Under the heading ‘New’ are given the tables with 
corrections as from PR, and the differences 5P,. Now it is seen that 
these differences 6P) are approximately constant from A=29°-0 to 
A = 46°-5, whereas 8P, drops steadily. We are reminded of the phe- 
nomenon shown by the Pulkovo angles of emergence. (See p. 54 of 
Dr. G. W. Walker’s “ Modern Seismology ’). In the last report Dr. Walker 
recalled attention to these observations, emphasising Galitzin’s belief 
in them, and suggesting in explanation a focal depth of the order of 1,300 
km. No support for this hypothesis has been otherwise forthcoming, 
and it seems possible that a simpler explanation may be found in errors 
of the tables; for a comparatively slight further modification would 
give angles of emergence in fair accord with the Pulkovo observations. 
Such a modification is shown in the column of Table VI. headed ‘ possible.’ 
Its differences 8P3 show a fall to A = 20°, then a rise to about A =28°, 
and then a fall again, which is all that is required. We proceed to show 
that this possibility is independently suggested by a study of the Y 
phenomenon or polychord. 
The Polychord or Y Phenomenon. 
Is it simply PR; ? 
In the 21st Report the following approximate times were deduced 
from the earthquake of 1913, March 14, for the Y phenomenon :-— 
pS Sale 100° 105° 110° 
Time for Y = 1420s. 1470s. 1490s. 1520s. 
Average vel. 14-9s, 14-7s, 14-23, 13-8s. 
ee ee ee 
