SURFACE REFLEXION OF EARTHQUAKE WAVES. 
arrive before S until A is 11,000 km., but for A 11,500 km. and more they arrive 
after S (according to Zoppretz’s curve). 
We now pass to consideration of the second phase. If we regard the start of the 
second phase S as representing the arrival of transversal waves we must resolve the 
disturbance into two: (l) transversal waves with direction of vibration at right 
angles to the plane containing the station, epicentre and earth’s centre, which I will 
call the azimuth plane; (2) transversal waves with vibration in the azimuth plane. 
For the waves vibrating at right angles to the azimuth plane, the vibration is 
entirely horizontal, there is no vertical component and the waves are reflected 
without change. 
For the waves vibrating in the azimuth plane the matter is more complicated and 
depends on the angle of impingence. For angles up to 56°, i.e., for A about 4000 km., 
the reflexion is complex. There will be a reflected transversal wave, but it may 
differ in type from the incident disturbance. There is no true reflected longitudinal 
wave, but a disturbance confined to the surface exists. I think there is little doubt 
that this latter disturbance is an important factor in the generation of “ Rayleigh 
waves. Fig. 3 shows the “ modulus” of the disturbance in dotted lines. We cannot 
speak of an angle of emergence for this region, but we note that the modulus of the 
reflected transversal, wave remains constant and = 1 from e — 0° up to e = 56°. The 
horizontal motion modulus is 0 at e = 0°, rises to a maximum at e — 22°, and falls 
again to zero at e = 45°, thereafter it rises rapidly to a value of 4'42 at e — 56°. 
The vertical motion modulus, always greater in this range than that of IT, reaches a 
maximum of 1'52 at e — 48° and then falls to '0 at 56°. 
As soon as we pass e = 56° there is a longitudinal as well as a transversal reflected 
disturbance without change of type. A 2 falls rapidly to 0'05 as e increases to 58°, 
and thereafter increases to 1 as e passes to 90°. A 3 falls from 27 7 at e = 56° to 
0 at 90°. The horizontal movement falls from 4'42 at e — 56° to about 1'8 at 
e — 63° and recovers to 2 at e = 90°. The vertical component rises to 0'92 at 
e = 60° and then falls to 0 at e = 90°. We may note that Y is always less than FI 
in this range, the least value of H/V being just under 2 at e — 63°. 
The very rapid change in value of Y and H from e = 45° to e = 65°, or A from 
2500 km. to 5500 km., is most remarkable, and suggests the necessity for a very 
thorough examination of records for such distances. 
S is often very far from sharp, and even in the cases where there is a sharp 8, it 
rapidly develops into a highly. periodic disturbance. A well-known feature is the 
comparatively small amount of vertical disturbance in S. We now find good reason 
for this. 
Little has been done in analysing S from actual records, and our results show how 
difficult the matter is without the guiding information provided by the phenomena 
of reflexion. Galitzin made some analyses of S in his book, but he assumed that 
the transversal wave arrived at the same angle as P. We have, however, a direct 
