434 EEPORT — 1888. 



and therefore in Italy as in Japan tremors are accompanied by wind, 

 although the wind may not be blowing at some particular place where 

 the tremors are observed. Certainly tremors often occur with a low 

 barometer, but the greater frequency of tremors apparently happens 

 when the gradient is steep, no matter whether the barometer is high or 

 whether it is low, and cases may therefore be observed of low barometer 

 unaccompanied by tremors. OlDservations like these have inclined me to 

 the opinion that tremors are more closely connected with wind than with 

 barometric pi'essure. An examination of the observations which have 

 been made shows : — 



1. That there are 80 cases of well-pronounced tremors having occurred 

 with strong winds blowing in Central Japan. In several instances tremors 

 were observed in Tokio some hours ahead of the wind, which was blow- 

 ing heavily to the S.W., and travelling up the country towards Tokio. 

 By Central Japan is meant all the country within 200 or 300 miles of 

 Tokio. 



2. There have been 40 cases of strong wind and no tremors. In 34 

 of these cases the wind has been local or of short duration ; that is to 

 say, it was only observed in Tokio, or it was only observed at one of the 

 tri-daily observations. In the remaining six cases, if tremors are the 

 result of wind they ought to have been observed. 



3. With no wind and no tremors there are 79 cases. 



4. With no wind and small tremors there are 63 cases. In 39 of these 

 cases the record on one band of paper showed no tremors, and therefore 

 these 39 cases might have been classified in the preceding group. On 

 the other band tremors were barely visible. In 8 of the remaining 

 26 cases the tremors observed were immediately in advance of a heavy 

 wind, or were tremors continuing after a heavy wind had passed, at which 

 time tremors had been well pronounced. There are therefore only 18 

 cases (26-8) where tremors can be said to have occurred where there was 

 no wind, and these tremors were slight. The above results may be tabu- 

 lated as follows : — 



No. of cases. 



1. Strong wind and well-pronounced tremors . . . . . 80 



2. Strong wind and no tremors, 40 cases which may be subdivided 



into : 

 (ff) Cases where tremors ought to have occurred ... 6 

 (J) Cases where the wind was local or of short duration, anditis 



doubtful whether tremors should have been recorded . , 34 



3. With no wind and no tremors 79 



4. With no wind and slight tremors, 63 cases, which may be sub- 



divided into : 

 («) Cases where tremors were so small that they were only 



recorded on one band of paper, 39 + a possible 16 = . . .55 

 (J) Cases which may have been due to wind .... 8 



Total 262 



The conclusion then is, that out of 86 cases of strong wind there are only 

 six cases where tremors were not observed, while when there was no wind there 

 were no tremors, or, at most^ tremors so slight that they were harely recorded. 

 These results agree ivith the results obtained in jirevious years. 



For three months an automatic spark record was kept of tremors 

 which might be due to vertical motion, but as these only occurred when 

 tremors were recorded by the ordinary tromometer, and were extremely 



