54 
MR. GEORGE W. WALKER ON THE PROBLEM OF 
Table II. 
j 
A kins. 
e. 
I. 
1/5-63 
secs. 
Z 
secs. 
Z - 1/5-63 
0 
O 
90 
0 
0 
0 
0 
500 
79 
25 
4 
69 
65 
1,000 
69 
95 
17 
136 
119 
1.500 
61 
203 
36 
199 
163 
2,000 
54 
344 
61 
257 
196 
2,500 
48 
512 
91 
310 
219 
3,000 
44 
699 
124 
358 
234 
3,500 
43 
894 
159 
402 
243 
4,000 
42 
1,096 
195 
442 
247 
4,500 
43 
1,299 
231 
478 
247 
5,000 
44 
1,497 
266 
512 
246 
5,500 
46 
1,688 
300 
542 
242 
6,000 
48 
1,871 
332 
572 
240 
6,500 
51 
2,044 
363 
601 
238 
7,000 
54 
2,205 
392 
631 
239 
7,500 
58 
2,351 
418 
660 
242 
8,000 
62 
2.481 
441 
688 
247 
8,500 
65 
2,596 
461 
716 
255 
9,000 
67 
2,700 
480 
743 
263 
9,500 
68 
2,797 
497 
769 
272 
10,000 
70 
2,888 
513 
795 
282 
10,500 
71 
2,972 
528 
820 
292 
11,000 
72 
3,055 
543 
844 
301 
11.500 
72 
3,133 
556 
867 
311 
12.000 
73 
3,209 
570 
888 
318 
12,500 
73 
3,283 
583 
909 
326 
13,000 
74 
3,355 
596 
929 
333 
We now proceed to show how a direct test may be applied to the Pulkovo data, and one 
which will give a determination of A and V 2 . 
If e is the emergence angle of a ray, then the Pulkovo data gives us two distances, 
say, J 2 and d 2 , for which e is the same. We hence infer that a PR wave will be reflected 
at Jj and pass to ejjicentral distance 2 A 1 A r d 2 , and another PR wave will be reflected 
at J 2 and pass to distance 2d 2 + ^i- E.g., e = 48 degrees gives = 2500, and A 2 = 
6000, from which we get distances to station 11,000 km. and 14,500 km. 
In this way fig. 8 has been determined directly from the Pulkovo data. It shows that 
the least distance at which PR occurs is 11,000 km., and for greater distances there are 
two PR waves for a given epicentral distance. The curve is in very close agreement 
with the theoretical curve in fig. 2. 
The test of the validity of the Pulkovo data is, then, whether for A > 11,000 km. we 
can identify the two PR waves on the seismogram. The Pulkovo Bulletins for 1913-1914 
show quite a number of records for A > 11,000 km., and it would seem desirable that 
a careful study of the seismograms for such distances should be made. Should the 
search prove successful the curve fig. 8 will then give two distances, A 1 and A 2 , for which 
