268 REPORTS ON THE STATE OF SCIENCE, ETC. 



In the portion of the Summary dealt with during the past year, there are 

 many large earthquakes and numerous cases of deep focus. Notable 

 among the latter are the earthquakes of October 14, 1932, 31-6° N., 13-8° E., 

 where distant records are completely absent, and January 9, 1933, 36-5° N., 

 70-5° E., where there is a wealth of observations over a range varying in 

 epicentral distance from 4° to 80°. In the former case, the epicentre being 

 in the Pacific to the south of Japan, there were excellent observations of 

 P and S at 40 stations, all within a distance of 11-7°, but there were no 

 observations outside Japan. The focal depth (determined at Tokyo) was 

 300 km. The other epicentre, which is in Kafiristan near the north-west 

 frontier of India, is one to which 10 deep-focus earthquakes were assigned 

 in the years 1921 to 1930. 



From January 1933 onwards an attempt has been made to distinguish 

 in the Summary between compressional and dilatational longitudinal waves. 

 For a compressional wave, where the initial motion of P, PKP or PKKP is 

 away from the epicentre, the letter ' a ' (anaseism) is entered after the 

 reading. If the wave is dilatational, or towards the epicentre, the letter 

 ' k ' (kataseism) is used. This notation was adopted by the International 

 Seismological Association at Edinburgh in 1936, the use of the adjectives 

 anaseismic and kataseismic having been proposed by the Rev. E. Gherzi, S.J., 

 as long ago as 1924. 



If the components of displacement in the onset of P are recorded by the 

 observing station, the direction of initial motion is known, and the dis- 

 crimination between ' a ' and ' k ' can be made after the epicentre has been 

 determined. Particularly useful is the Z component, as an upward initial 

 motion always indicates an ' anaseism ' and knowledge of the position of the 

 epicentre is not required. A good many observatories are already providing 

 in their bulletins the necessary information with regard to the initial move- 

 ment of each earthquake. It is hoped that the practice will be adopted 

 generally. 



Work on Transmission Times and on Periodicity. 

 By Dr. Harold Jeffreys, F.R.S. 



The work on southern earthquakes and the core waves, which was in 

 progress at the time of the last Report, has been completed. For PKP 

 only readings at the most reliable stations with vertical component instru- 

 ments were used and the result was a symmetrical distribution of residuals 

 with a standard error of about 2 sec, nearly the same as for P. Accordingly 

 there is a high probability that the dangers of systematic error in PKP have 

 been removed. The summaries have a standard error of about 0-4 sec, 

 about the same as for P at most distances. The times of SKS have also 

 been rendered somewhat more accurate. Some of the earthquakes used 

 were found to show signs of multiplicity. There appeared to have been 

 two or three shocks at the same place, separated by intervals up to 10 sec, 

 and P had been read for the first, and S and SKS for a mixture, usually with 

 a preference for the later ones. This explains xnost of the ' Z ' phenomenon, 

 leaving no more than can be reasonably attributed to variations of focal 

 depth within the upper layers. Cases where the separation is larger have 

 already been considered by Stoneley and Tillotson and appear to provide 

 an explanation of most of the recorded cases of apparent ' high focus.' 



A study has been made of the frequency of after-shocks of the Tangp 

 (Japan) earthquake of March 7, 1927. They were found to agree with 

 a law, such that the chance of an earthquake in an interval dt is pro- 

 portional to dtj(t — a), where a is near the time of the main shock. Apart 



