November ii, 1920] 



NATURE 



349 



the individual waves remained possible. The use 

 of a "dug-out " in a pit bank was secured, distant 

 two miles from the home station, in a direction 

 17° west of north, and one of the instruments in- 

 stalled there. Both instruments were arranged in 

 the same azimuth, with the same period, damping, 

 and ratio of magnification. The clocks could be 

 synchronised only by motoring between the two 

 stations with a watch that varfed about one-fifth 

 of a second per hour. Comparisons were made 





VVVW^r»VV^V^'/^V\^M^W^V\N«-«vsV^^V^^ 



^\^-^^wviAaAAAWyvV\i 



•i-" , . I ,'•0 . J 



2*0 



• Fio. I. — Earthquake and microseisms. 



hiith w.iys, and the precision was calculated by 

 interpolation. On favourable occasions the clocks 

 were set alike within about one-tenth of a second. 

 It was at once seen that there would be no diffi- 

 culty in identifying the waves at stations two miles 

 apart (see Fig. 2). The films were timed by a short 



Home. 11 3 30 

 Oufii OUT. 



HOME. 14 2 ZO 



Dy« OUT . 



Fio. 9. — ComiHirative trares. 



<:• ii; -r <if the light every minuit. In the first 

 attempt the clocks were compared once daily, and 

 ;lie films timed by measuring from an eclipse to 

 lie nearest apex in the trace. By this method 

 iiffcrcnces of as much as two seconds were noted, 

 whiTcas one second was the maximum expected 

 from the known rate of surface wave trans- 

 mission. It was evident that a higher precision 

 NO, 2663, VOL. I06l 



was necessary. The errors were chiefly traced to a 

 lack of uniformity in the peripheral speeds of the 

 recording drums. In a second series of observa- 

 tions the clocks were compared twice a day, 

 and comparisons of the films limited to the point 

 marked by the eclipse. For the success of this 

 device it was necessary to obtain a record of the 

 wave during the eclipse. This was achieved by 

 cutting a fine slit across the shutter, so that whilst 

 the beginning or end of the eclipse was still 

 clearly defined, the leakage of 

 light through the slit produced a 

 "ghost" during the eclipse. 



With the beginning and end of 

 an excursion, and also the time- 

 •Tiark being defined, it remained 

 only to resolve the harmonic 

 motion to obtain the phase of the 

 wave at each station at the in- 

 stant the shutters closed, and so 

 to deduce the time difference be- 

 tween the two observatories. By 

 the first method there were read- 

 ings ±1-13 sec. from the mean. 

 Hv the second, this was reduced 

 ii» ±0-3 sec. ; but it is worthy of 

 note that by cither method the 

 difference between the two means 

 was only 004 sec. for the 

 series. 



It was noted that the wave 

 period increased with the ampli- 

 tude, and the amplitude increases 

 generally with the daily move- 

 air. There was, however, one 

 marked exception on March 10, when a very 

 moderate air movement coincided with nearly the 

 maximum amplitude recorded. At this time rough 

 weather was being experienced around the north 

 of Scotland. 



The chief objective in these proceedings was 

 a comparison of wave direction with wind direc- 

 tion, and perhaps in its complete failure lies the 

 greatest value of these observations, U>r it was 

 discovered, it is believed for the first time, that 

 there is practically no change in the wave direc- 

 tion, whatever the meteorological conditions may 

 be. The microseisms always came, more or less, 

 from the north — they always arrived at the "dug- 

 out" first by about 08 sec. The small variations 

 in time difference shown in column 2 of the table 

 may or may not indicate a varying azimuth. It 

 is equally probable that they are the result of 

 personal or instrumental error. 



To reduce the difficulties and shortcomings in 

 the above experiments, it is proposed- with im- 

 proved timing facilities and three machines placed 

 at the corners of a lo-milc triangle — to continue 

 this investigation, when it is hoped not only to 

 obtain more precisely the rate of propaga- 

 tion, but also to determine to what extent the 

 azimuth is constant. It would be of value 

 if observers in other countries were to pursue 

 the study of microseisms on this new system 

 in order to determine whether this unidirectional 

 character of the phenomenon is universal. 



ment of the 



