ON THE EARTHQUAKE PHENOMENA OF JAPAN. 247 



From this last table we see that, although the period of motion at 

 Station II. is slow in consequence of the very large amplitude usually 

 experienced at that station, the maximum velocity, and more markedly 

 the maximum acceleration, which may be taken as a measure of the 

 intensity or destructive power of a disturbance, have been much greater 

 than at Stations I. and III. 



One of the most remai-kable earthquakes in the series was that of March 

 26. Although the amplitude of this was sufficiently great to constitute 

 a destructive shock, the period was so long that the disturbance almost 

 escaped attention. Several persons observed lamps and pendulums 

 swinging, and thought there might be an earthquake occurring, but I only 

 found one or two persons who detected any motion of the ground or 

 bnilding. 



Speaking generally about these observations, it may be said that had 

 three independent observers been placed at the three stations which are 

 only 800 feet apart, and each had been provided with similar instruments, 

 they could not have failed in giving very different accounts of the same 

 earthquake, both as to its period, its duration, and, I may add, its direc- 

 tion. A result of practical interest that is dependent in the records which 

 I have obtained, is the benefit to be derived by engineers and architects 

 by making a systematic seismic survey of the ground, on which they 

 intend to erect important structures in earthquake-shaken districts. 



Observations with the Gray-Milne Seismograph. 



As this instrument has been described and illustrated in the ' Quarterly 

 Journal of the Geological Society of London ' (vol. xxxix. p. 218) , and in other 

 publications, I will not describe the details of its construction. It consists of 

 a pair of conical pendulum seismographs, which record upon the smoked 

 surface of a drum, two mutually rectangular components of the horizontal 

 motion of the earth. The drum is kept continuously in motion by clock- 

 work. The vertical motion is described by a spring lever seismograph. 

 At a certain part of the earthquake, a mark is made on the drum, simul- 

 taneously with which time is recorded from a specially arranged time- 

 piece. By this means the time can be calculated at which any particular 

 vibration of an earthquake occurred. 



As the instrument is designed more for the systematic observation of 

 earthquakes, rather than for experimental purposes, I entered into corre- 

 spondence with the Meteorological Department of this country to admit it 

 into their department as an instrument for regular observation. This 

 Mr. Arai Ikunosuke, the director of the Meteorological Department, has 

 kindly done. After repairing slight damage, which it suffered in its 

 transit, it was exhibited to His Imperial Majesty the Emperor of this 

 country. Since then a heavy brick column on a massive concrete foun- 

 dation has been built for its instalment, and it has been put in charge of a 

 regular observer. During the early part of this year, although several 

 earthquakes were experienced, no results were obtained. This was due to 

 the pointers of the conical or horizontal pendulums, which are extremely 

 sensitive to slight changes in level, slowly wandering to the right and left 

 of their normal position on the revolving drum. The consequence of 

 this was, that instead of simply tracing in the smoked surface a single 

 line, they made a path sometimes an inch in breadth, and when the earth- 



