ON THE EARTHQUAKE AND VOLCANIC PHENOMENA OF JAPAN. 109 



large that earthquakes have not occurred. The tremor storms which were 

 numei'ous in tlie early j^art of the year have no doubt obliterated many of 

 the unfelt earthquakes to which tlie pendulums were sensitive. Notwith- 

 standing this, there are a considerable number of disturbances on the 

 traces, the record of which must be left for a future report. The most 

 remarkable of these was recorded by pendulum F, which at the time was 

 steady and producing a clear sharp straight line (see fig. 3). This was 

 on June 3, when at 4.36 p.m. the pendulum commenced to move from side 

 to side, and, with the exception of two or three intervals of about five 

 minutes, it continued to move until 10 P.M. Although 14 points of 

 maximum may be counted, the photograph represents what is practically 

 a continuous earthquake of 5 hours 24 minutes' duration. The picture is 

 that of a series of small flat cones, each inverted, with their axes in one 

 straight line. No displacement of the pendulum took place, and after the 



Fig. 3. 



About half actual size. 

 The gap at or near noon represents an interval of one hour. 



disturbance it continued to draw the same thin line. I do not know 

 where this or the other unfelt earthquakes originated. The rate at which 

 the decided movements are propagated is from 2'5 to 4 kilometres per 

 second, and there are reasons for believing that many of them, like that 

 of March 22, originated beneath the bed of the ocean. 



(o) Tremors. 



In the extracts from the Journal (pp. 96 to 99) it will be seen that 

 tremors or earth pulsations have often been recorded, and that some- 

 times these were greater underground than on the surface. During the 

 last two months they have been greater on the surface. Previous analyses 

 have shown that they nearly always accompany a steep barometric gradient. 

 They are sometimes marked when the daily curve is barely visible, but 

 small tremors at least usually accompany these waves, and' they are more 

 pronounced during the night and early morning, when the rate at which 

 a pendulum is being displaced is relatively slow. The fact that small 

 tremors were produced at the time the well was emptied is a fact not to 

 be overlooked when considering their origin. 



I regret to say that a more careful examination of the tremor records 

 must be left for a future report. 



{p) Observations at Yokohama and Kanagawa. 



As already stated, the instruments at Kanagawa (i) and (g and h) at 

 Yokohama are underground, and stand on short brick columns rising from 

 soft tuff rock. The softness of this rock may be judged of from the fact that 

 when a person stands near one of the columns, the boom of the pendulum is 

 deflected from 5 to 17 mm., from which it appears that, as a foundation to 

 resist loading eflfects, the tuff rock is no better than a slab of conci-ete on 

 the alluvium in Tokio. Owing to the collapse of the roof of the Yoko- 

 hama cave, which caused a delay of two weeks, and owing to the fact that 

 the clocks have been continually stopping, and good clocks cannot be found 



