SEISMOLOGICAL INVESTIGATIONS 269 



about 25 sees. Recently a further case of ' high focus ' has been noticed. 

 The earthquake of 1932 September, 26^. i9h. 20"!. 42s., with epicentre 

 39° -8 N., 23° -8 E., in N.W. Aegean Sea, just outside the Gulf of Salonica, 

 is very well recorded over a range of distances from 5° (Italian stations) to 

 158° (New Zealand stations), and 120 P readings are in approximate agree- 

 ment. However, a consistent difference exists in residuals between the 

 European group of stations and the North American, both in the same 

 azimuth, and the only explanation that seems to fit the circumstances is 

 that the European stations are at the distance at which travel times are not 

 much affected by focal depth (or height) whereas at the distances of the 

 American stations about the observed amount of delay would be expected. 

 On the Turner scale of A corrections for abnormal focus a ' height ' of 

 o 0075 has been applied ; this brings into line the otherwise discordant times 

 at different distances in the same azimuth. 



There were numerous deep focus earthquakes in 1 93 2 . The large Japanese 

 earthquake of April 4^ from epicentre 30° -6 N., 139° 5 E. off the coast, 

 is a useful illustration of the good fit of the Turner deep focus correction 

 tables. It is not necessary to claim that the fraction used to define the depth 

 must be exact as a proportion of the earth's radius, but it does constitute a 

 relative standard by which depths of focus in different earthquakes may be 

 compared. In the present case the depth is nominally 0-065 of the earth's 

 radius or 400 km., and the Turner corrections account for the differences of 

 observed A from A = 2° • 5 where the figure is + 1° • 8 to nearly 90° where 

 it is - 8°o. 



Another important Japanese deep focus earthquake occurred on 1932 

 November 13d at about 43° -6 N., 137° -3 E., with focal depth about 0050. 

 The details of this are not yet finally settled, but it is one of the most widely 

 recorded shocks and will furnish data from stations at all distances. 



The Mongolian Earthquake of 193 i, August 10. 

 By Dr. R. Stoneley. 



In last year's B.A. Report Mr. Hughes commented on a peculiar feature 

 of this earthquake ; the P observations determine the epicentre as 46° -9 N., 

 90° -o E., accurate within 0° • i , but the S readings are then about 30 sec. late. 

 According to the old I.S.S. routine this would probably have indicated a 

 ' high focus ' shock. 



One possible explanation immediately suggests itself. In the absence of 

 microseisms, P can be picked out without difficulty, even if of very small 

 amplitude, whereas S has to be read against an already disturbed ' back- 

 ground.' It is fortunate for seismology that S is generally of relatively 

 larger amplitude than P. If a rather weak shock is followed a few seconds 

 later by a decidedly stronger shock, distant stations will tend to read the P 

 of the first shock and the S of the second. This explanation seems to apply 

 to the earthquake under consideration. 



By the kindness of Mr. Hughes and Miss Bellamy I have been able to 

 examine the seismograms of Oxford and Dyce and to make use of a pre- 

 liminary proof of the I.S.S. In both records the earthquake begins as an 

 eP, followed 3 or 4 sec. later by an iP. The calculated time falls between 

 the eP and the iP. There is a very clear sudden onset for which O-C is 

 about 33 sec. for Oxford and 32 sec. for Dyce ; these onsets cannot be ex- 

 plained as reflected waves. The Kew records, for the loan of which I am 

 indebted to Dr. Whipple, are quite convincing ; the proininent onset some 



