ON THE KARTIIQUAKE AM) VOLCANIC I'lIEXOMEXA 01' JAI'AX. 113 



frequent from the end of June until the end of November, and accompany 

 winds having a velocity of from 50 to 100 km. per hour, the barometer 

 being low. During the remaining months of the year although slight 

 tremors are frequent it is seldom that they are pronounced. When winds 

 are fairly strong, reaching 30 or 40 km., unless these are continuous over 

 several days the tremors are slight. Cases occur when tremors are fairly 

 frequent with a low barometer (757 mm.) whilst there was but little wind. 

 On the other hand, with a wind not exceeding 30 km., there have been 

 decided tremors with a high barometer (763 mm.). The conclusion to be 

 derived from these records, which, however, do not show the state of the 

 bai-ometric gradient, is that tremors are frequent with high winds, with 

 winds of moderate intensity if these are continuous over several days, and 

 at times when the barometer is low. Because they are sometimes absent 

 when a wind of moderate intensity is blowing, it would seem that their 

 appearance may be more closely connected with changes in barometrical 

 pressure than with the mechanical effects of wind upon surface irregulari- 

 ties. They do not appear to be connected with daily changes in tempera- 

 ture, the hygrometric state of the atmosphere, or with the occuri'ence of 

 earthquakes. 



Since writing the above Father M. Saderra, S.J., Director of the Obser- 

 vatory at Manila, writes me as follows : 'Evidently the results are influenced 

 by atmospheric currents, but in an indirect manner, as is proved by the- 

 fact that the hours of greatest wind force and* greatest movements of the 

 tromometer do not always coincide. The influence is verified by means 

 of the vibratory motion which the wind produces in the earth itself by 

 impinging against mountains. It now remains to observe and study this 

 influence and endeavour to ascertain which wind exerts most influence. 

 Hitherto this has not been feasible, but now, God helping, I will set 

 about doing it.' 



Although these observations throw a certain amount of new light upon 

 the occurrence of tremors, and possibly explain their morning frequency, 

 the causes producing tremor storms are yet obscure In Japan, at least, 

 they appear to be a surface phenomenon which exhibits itself in different 

 degrees in different localities. Although the longer period motions of hori- 

 zontal pendulums show that changes in barometrical pressure may be suf- 

 ficient to produce changes in level, it does not seem unlikely that rapid 

 alterations in barometrical pressure over an area, the yielding of diflTerent 

 portions of which are unequal, may be sufficient to create irregular 

 mechanical distui'bances in such a district, and the motion once started as 

 with t* severe earthquake may continue for several hours after the initiat- 

 ing vo.use has ceased. Whatever may be the cause of these ubiquitous 

 phenomena, because they interfere with so many delicate physical opera- 

 tions, they certainly demand serious attention. For the assistance of 

 those who desire to scrutinise the analyses of which I have only given 

 the results or to make new investigations, the following list of meteoro- 

 logical conditions during the period which has been considered is here 

 appended. (See Appendix, p. 182.) 



(i) Meteorological Tables for Tokio, October 13, 1894, to January 1895. 



The following tables have been extracted or computed from informa- 

 tion given in the weather maps which are issued three times per day by 

 the Central Meteorological Office in Tokio : — 



