394 NATURAL SCIENCE. j UNE , 



tracing it from Nishi-katabiri to Haku-san, a distance of 40 miles, 

 and there is good reason for believing that it extends as far as Fukui, 

 i.e., for a total length of 70 miles. Proceeding in a general south-east 

 and north west direction, it cuts through hills and plains alike. With 

 one exception, the ground on the north-east side is relatively depressed, 

 and it is also constantly shifted by as much as 5 or 6 feet towards the 

 north-west. Where the vertical displacement is small, a mark was 

 left like the track of a ploughshare or of a gigantic mole. In other 

 places, where the fault crosses flat ground, it has the appearance from 

 a distance of a railway embankment. This is the case at Midori, in 

 the Neo Valley, where the bed of the valley is split longitudinally, and 

 an abrupt step formed, from 18 to 20 feet in height. When we think 

 of the remarkable length of this fault, the magnitude of its throw, and 

 the friction that must have resulted from the rapid displacement of 

 such enormous masses of rock, we can entertain but little doubt as to 

 the correctness of Professor Koto's conclusion that the formation of 

 the fault-scarp was the cause of the great earthquake (1). 



Distribution of After-Shocks in Time. — Gifu, the capital of Mino, 

 lies about 4^ miles from the nearest point of the fault-scarp. The 

 seismograph at the meteorological observatory there recorded the 

 first half-dozen vibrations, and then the building fell and the 

 instrument was buried. At 2 p.m., i.e., about seven and a half hours 

 after the earthquake, it was again in working order, and the registra- 

 tion of the after-shocks commenced. From this time until the end of 

 December, 1893, or within a little more than two years, the total 

 number of after-shocks recorded is 3,365. Of these, ten are described 

 as violent, ninety-seven strong, 1,808 weak, and 1,041 feeble; the 

 remaining 409 being merely sounds unaccompanied by any per- 

 ceptible motion. 



The relation between the number of after-shocks and the time 

 since the great earthquake is shown graphically in Fig. 3. The dots 

 denote by their distance from the horizontal line the total number of 

 shocks recorded during each successive month from November, 1891, 

 to December, 1893. The continuous curve, which passes through 

 these dots, exhibits very clearly the rapid decline of frequency 

 during the first two months and the much slower rate of decrease at 

 the end of five or six months. For instance, in November and 

 December, 1891, there were 1,087 an ^ 4*6 after-shocks respectively ; 

 while in December, 1892, there were only thirty-nine, and in the 

 December following not more than sixteen. At the same time, this 

 decline is not continuous, there being several maxima and minima, 

 though not of great importance. Smoothing away these irregu- 

 larities, we get the dotted curve in the figure, and this, it will be 

 seen, approximates very closely to the form of a rectangular 

 hyperbola. If this curve represents the true law of decline in 

 frequency, it follows that the monthly number of shocks must go on 

 diminishing, until, after a lapse of about forty years, the ground will 



