330 Professor John Milne [Feb. 12, 



seen to gradually change into double waves. The velocity of propa- 

 gation evidently increased with the intensity of the initial impulse ; 

 it was greater for vertical and normal than for transverse waves, and 

 vibrations generally were propagated more rapidly to stations near 

 an origin than between stations at some distance from the same. 

 These and many other results were confirmed and extended by 

 records obtained from a series of nine seisraometric stations situated 

 on a plot of ground the area of which was only a few acres. In these 

 investigations the records, which were drawn upon the surfaces of 

 smoked plates, were those of real earthquakes. The motion on one 

 side of this ground was invariably so much greater than it was 

 900 feet distant upon the other side, that it offered an explanation 

 for the peculiar distribution of ruin so often observed in a city after 

 it has been shaken by an earthquake. The houses in one street 

 may stand, whilst others possibly not more than 100 feet distant, also 

 standing on alluvium, but somewhat softer in character, may be 

 shattered. From the survey of a field, seismic investigations were 

 extended to the survey of Tokio, and then to the survey of the 

 northern half of Japan. At this point the Government came to 

 the assistance of private observers, and took under its control the 

 survey of the whole empire, embracing an area of 140,000 square 

 miles, within which there are now close on 1000 stations at which 

 earthquakes are recorded. 



The results of this undertaking are not at present fully known. 

 "What we have learned is that during the last six years the average 

 number of shocks have been about three per day, a frequency which 

 is greater than that which is usually given for the whole world. 



If we take the well-marked earthquake districts of the world and 

 give to them frequencies one-third of that in Japan, it would not be 

 an over-estimate to say that 10,000 movements sufficiently strong to 

 be felt and shake considerable areas of our planet occur every year. 

 Five thousand of these come from the home of our deep-sea cables. 



The Japan earthquakes, like those of South America, mostly 

 originate on the side of the country which slopes steeply down beneath 

 the Pacific Ocean. In fact, it may be taken as a rule that whenever 

 ground over a considerable distance, which I will take at 120 geogra- 

 phical miles, has an average slope greater than 1 in 50, in such 

 districts under the influence of bradyseismical bending or of secular 

 crush round the base of the continental domes, earthquakes are 

 frequent. From Japan to beneath the Pacific, slopes of 1 in 25 

 occur, whilst on the coast of Peru slopes as great as 1 in 16 may 

 be found. The volcanic districts of Japan which, like those of South 

 America, are found along the upper part of a bradyseismic fold, are 

 singularly free from earthquakes, and the times of seismic and 

 volcanic activity show no marked connection. 



The analyses of the Japan records, as a whole, as with the analysis 

 of the records of most other countries, show a marked annual and 

 semi-annual periodicity. The former of these, which shows a winter 



