lOS 



REPORT — 1895. 



For several days the pointer of the pendulum had been fairly steady, 

 pointing at division 70 on the scale of millimetres. What happened when 

 the well was emptied is given in the table below. 



The photographic trace with interruptions in it Avhen the light was 

 removed is shown in fig. 7, Plate II. The movement of the pointer from 

 70 to 79 indicates a tilt of l"-36 and the direction of motion was as if a. 

 load had been taken away on the well side, and the ground on that side 

 had therefore risen. This may be explained by the fact that as the water 

 came to the surface it was run into a gutter to flow away quickly down a 

 hill. The pendulum remained between 77 and 82 until May 27, when the 

 experiment was repeated. It started at 80, and in 6 hours and 40 minutes 

 it reached 86, and here it has remained with a tendency to get higher but 

 not to return. 



Not only was tilting produced by these operations, but as seen in the 

 photograph tremors were induced. 



It might have been anticipated that by emptying the well and the 

 subsequent inflow of water to refill the same — if in consequence of this 

 operation a superficial movement took place — this would have assumed 

 the form of a quaquaversal dip towards the well. What happened was 

 exactly the reverse, from which it may be inferred that the motion 

 of the pendulum was due to the removal of a weight rather than to the 

 movement of the subterranean water. 



(n) Earthquakes. 



In the last column of the table showing the wandering of pendulums, 

 the number of earthquakes which occurred on various days is given. 

 These are the earthquakes which were recorded by seismographs in Tokio, 

 and it is only one or two of these like the disturbances of March 22, 

 when earth waves were produced, that are recorded by the pendulums. As 

 already stated when speaking of the Kamakura records, although it is 

 probable that most of these shocks were of local origin, this fact cannot 

 be ascertained until the records accumulated at the Meteorological 

 Department have been analysed. Two things, however, are very remark- 

 able, the first being that at about the time of nearly all the shocks, 

 pendulum A has shown abnormally large movements, and secondly there 

 are only three occasions when the movements of A have been moderately 



