ox THE EARTHQUAKE AND VOLCANIC PHENOMENA OK JAPAN. 183 



sometimes reaching ^ incli. This continued until the 25th, when, sus- 

 pecting that the cause might be due to air currents resulting from rapid 

 desiccation, I removed the calcium chloride, and the movements ceased. 



1 have repeated the experiment several times, with the result tliat when 

 the calcium chloride is introduced, movements are produced, which in 

 the developed tilm have the appeai-ance of a violent tremor storm, and 

 when it is taken away the diagram is a clear straight line. We have 

 here a cause of air currents which has not yet received serious atten- 

 tion. Another cause of movement, whicli is easily verified by experi- 

 ments on a very light pendulum beneath a glass covering, may be due 

 to the unequal heating of the surrounding walls. If portions of the tremors 

 which have been recorded are due to causes, such as these which act 

 within the casings, then it is understood why extremely small and light 

 pendulums have shown more movement than those which are compara- 

 tively large and heavy. After my last twenty installations, and those 

 which preceded them, the horizontal pendulums might be arranged 

 according to their sensibilities as follows. The most sensitive are those 

 with booms from one quarter to 3 or 4 inches in length, the next are those 

 like R, and the one at Shide T, where the boom is of reed or straw about 



2 ft. 6 in. in length, following which are booms about 5 feet in length of 

 bamboo like A, and lastly, as the least sensitive, are the somewhat shorter 

 and comparatively heavier booms of brass or aluminium used at the 

 remaining stations. The only exceptions in the last group were the 

 instruments E and F in the undergi'ound chamber, whicli recorded 

 ti-emors almost equally as large as A. In this chamber, although there 

 was but little appreciable daily change in temperature, the ventilation 

 was good, and therefore there may have been considerable changes in the 

 hygrometric state of the atmosphere entering the covering cases, which 

 were of wood resting upon a floor of asphalt. A ditference in the rate 

 at which moisture was absorbed or evaporated from the walls of this 

 casing might possibly give rise to air currents. The live instruments in 

 badly ventilated caves may have failed to show tremors, partly on 

 account of their inertia, and partly perhaps because there was neither 

 any sensible change in temperature nor in the dryness or wetness of the 

 atmosphere. That tremors were practically absent from all the instru- 

 ments on the surface, can only be attributed to their inertia, or to the fact 

 that they were so well ventilated that no difference in temperature within 

 their coverings was possible ; but as the huts and casings were similar to 

 that of R, the most probable explanation is the former. 



The most difficult things which require explanation respecting earth 

 tremors, assuming them to result from air currents due to differences in 

 temperature or desiccation within the walls that inclose the instruments, 

 are the facts that tremors have in all cases but one been most pro- 

 nounced between 6 a.m. and 9 a.m., and during the night, and that they 

 accompany certain meteorological conditions already formulated. 



Before attempting tliese explanations it would be advisaT)le to com- 

 pare the movements of two light pendulums standing on the same column, 

 one having walls varying in character, and the other, if possible, in 

 vacuum. 



