SEISMOLOGICAL INVESTIGATIONS, 



219 



The duration of tlie eclipse was 4"7 seconds in each case, so that opening or 

 closing were equally serviceable as datum points. Therefore a new method of 

 comparing the films was devised as follows : — 



The eclipsing shutter was provided with a narrow slit through which a small 

 percentage of light could pass when the shutter was closed. This feeble beam 

 produced a ghost-like trace during the interval of each eclipse. 



In making comparisons instances were chosen where the amplitude was not 

 only large but also where the shutter had opened or closed near the middle 

 or zero position of the wave. 



The change of intensity of the trace was sharp and easily measured, whilst 

 the extremity of the excursion could be seen in the ghost. 



The period of the wave and its phase at the datum point having been deter- 

 mined, it wae then possible to resolve the harmonic motion, and so obtain the 

 difference in time to one-tenth of a second. 



It is interesting to note that by either method the average difference was 

 0"8 second, but the second method gave much more consistent readings. 



A further object was to note to what extent the direction of propagation, 

 the amplitude, or the period were affected by meteorological conditions, particu- 

 larly the direction and force of the wind. 



We were indebted to A. J. Kelly, Esq., Director of the Birmingham and 

 Midland Institute Observatory (four miles distant), for his help in this matter. 



The force of the wind and the amplitude did appear to be co-related, inasmuch 

 that the microseisms were small during calm spells and vice versa, but there was 

 a notable exception on March 10. During March 9 and 10 the air movement 

 had been small, 178 and 272 miles in each 24 hours respectiveily, yet on the 

 evening of the 10th nearly the largest waves of the series were recorded. 



Within a period of 24 hours, March 12 to 13, the velocity of the wind 

 ranged from 37 to 12 and back to 37 miles -per hour in three nearly equal periods, 

 but there was no corresponding fluctuation in the amplitude of the microseisms. 

 Similar fluctuations on other dates were equally ineffective to produce sudden 

 change in the ground movement. 



There was little variation in period. It was usually 6 to 7 seconds. On 

 a few occasions it fell to 45 seconds, but never exceeded 8 seconds. It will 

 l)e observed that the period appears to increase with the amplitude. 



The outstanding, and we venture to think important, discovery was that 

 the microseismic waves always arrived from the same direction. On every film 

 they were seen to arrive at the ' dug-out ' or northerly station first. 



During the period of observation the wind blew from all points, except 

 north to east, but no quarter seemed to affect the regularity with which the 

 waves arrived from the north. 



Column two in the following table gives the time in seconds by which the 

 waves arrived at the dug-out first : — 



By First Method. 



