236 REPORTS ON THE STATE OF SCIENCE, ETC. 
down feeling. After 3 minutes of this, gradual subsidence occurred for 
24 to 1 minute, and then calm. (A very slight tremor occurred again at 
about 3.35 P.M., duration about to secs. or less.) One expected the earth 
to give way, but no cracks are visible. ‘The lines of the building resembled 
a jerky shaking outline. The house appeared to sway and rattle itself to 
pieces the whole time. (I liken it to a rough-haired dog shaking itself.) 
People say they remember nothing like it. Several brick buildings a mile 
and more away are damaged also. 
‘The rumbling appeared to come from south-west but may have been 
house noises or the galvanised iron roof and only apparent the first half 
minute I was in the house. I watched proceedings a safe distance out of 
doors. One felt giddy 15 minutes afterwards and there may have been a 
gentle wave motion of the earth after the 3 minute tremor.’ 
British Earthquakes—There was no considerable earthquake in the 
British Isles during the year, but small disturbances, some of which may 
have been due to the collapse of old workings in mines, were reported by 
the newspapers as occurring on the following dates : 
1933, October 28. Nottingham. 
1934, February 10. Roslin, near Edinburgh. 
1934, March 17. Coasts of the Bristol Channel. 
1934, April 23. Elvington, near Dover. 
1934, June 8. Dufftown, Banffshire. 
1934, August 16. Dingwall and Cromarty. 
Earthquake Prediction. 
It has always been the desire of seismologists to be able to give warning 
of earthquakes. A letter from the Director of the Observatory at Manila, 
communicated by Mr. F. Hope-Jones, suggests a new line of research. 
The Observatory has been equipped with a synchronome ‘ Shortt’ Free 
Pendulum, of the same construction as the Standard Siderial Pendulum at 
Greenwich. This type of pendulum has a variation of not more than 
about three seconds per annum. The Director’s letter contains the 
following passage : 
‘ Another interesting feature which I have noticed in your Synchronome 
(and also to a slight extent in the two Rieflers which I have) is the sensi- 
bility to a tilt in the land. On more than ten occasions some three or four 
days before a local earthquake the rate of the clock has changed very abruptly, 
varying as much as a tenth of a second within twenty-four hours, and then 
suddenly assuming a very slight rate which it keeps until the ’quake comes. 
After the ’quake the clock generally resumes its old rate, or one very near 
to it. ‘There may be other explanations for this strange performance, but 
it may also prove to be a helpful hint in the apparently impossible solution 
of a method of predicting earthquakes of tectonic origin.’ 
Periodicity of Earthquakes —The question whether earthquakes are more 
likely when the moon is in one position or another is frequently asked. 
An answer is to be found in a paper by Dr. C. Davison in the Philosophical 
Magazine for April 1934, ‘The Lunar Periodicity of Earthquakes.’ The 
earthquakes included in a number of catalogues have been investigated. 
The frequency of earthquakes tends in some cases to a maximum at the 
times of new and full moon but in other cases to a minimum. In the 
former category are such earthquakes in Japan as have their foci under the 
land, but Japanese earthquakes with foci under the sea are least frequent at 
“new and full moon, as are the volcanic earthquakes of Honolulu. In the 
catalogues including earthquakes in all localities the minimum frequency is 
