HORIZONTAL PENDULUMS. 137 



20. The earthquake of the 23rd April, 1898. This was 

 a large earthquake, whose centre was not very distant from the 

 observing station, being under the ooean about 400 km. NK by 

 N of Tokvo, The area of di^;turbanco, within which the shock 

 was strong enongh to l)e felt without instrumental aid, was veiy 

 extensive, its longest and shortest axes being respectively about 

 050 and 380 km. Along the east coast of the northern part of 

 the Main Island the shock was sufficiently strong to damage 

 some buildings. In Tokyo itself the earthquake was felt as mild 

 shakings of long duration. 



The NS component diagram of the earthquake is shown in 

 fig. 16. (The EW component apparatus was then not ready). 

 The shock l)egan at 8. 37.0 a.m., and, after about oO seconds of 

 preliminary tremor, the motion became large, till at 1 m. 14 s. 

 from the commencement the pointer got off (at ^4) the record- 

 receiver. At 8 ra. 41 s. from the commencement, T put (at B) the 

 pointer again on the smoked paper. The motion was then still 

 pretty large, but standing perfectly still, T could feel absolutely 

 nothing of it. The subsequent motion was recorded to the end, 

 the pointer having been purposely displaced from F", to E' and 

 again at H in order to avoid the overlapping and confusion ot 

 several lines. The total duration of the earthquake was about 

 2 hours. 



The earthquake motion may l)e considered to have consisted 

 of two kinds of vibrations : firstly, of waves whose periods were 

 shoit and included between a fraction of a second and about two 

 seconds ; and secondly, of waves whose periods were long and 

 varied from 5 seconds to about 12 seconds. Ordinary seismo- 

 graphs would have recorded the waves of the first order, but would 

 have been incapable of recording those of the second. 



