210 REPORT— 1882. 



Japanese. It -was especially remarked at the time of the destructive 

 earthquake of 1854. 



In Yokohama, sixteen miles to the S.E., -where the high and low 

 ground has almost exactly the same topographical character as at Tokio, 

 the rule appears to be reversed. This was clearly evident in February 

 1880, when the shattering of chimneys, unroofing of houses, and destruc- 

 tion generally was almost wholly confined to the high ground. In 

 Hakodate the rule appears to be like that for Tokio — namely, that the 

 greatest disturbance is felt upon the low ground. 



In consequence of the great difference in motion observed in places 

 which are adjacent to each other, I have hitherto been unable to make 

 any satisfactory determination of the manner in which an earthquake 

 dies out as it radiates from its cpicentrum, although on many occasions I 

 have obtained a number of diagrams for the same earthquake from 

 distant stations. 



"V. Experiments on artificially produced Earthquakes. 



In 1881, in conjunction with Mr. Thomas Gray, I made experiments 

 upon a series of artificial earthquakes produced by allowing a heavy iron 

 ball to fall from various heights up to 35 feet. 



During the past year I continued these experiments on a larger and 

 more satisfactory scale, the disturbances being produced by charges of 

 dynamite exploded in bore-holes usually about 10 feet deep. In the first 

 two sets of experiments the vibrations resulting from the explosions were 

 simultaneously recorded upon moving glass plates at three stations. 



In consequence of the great increase in the intensity of the initial 

 disturbance as compared with that obtained from the falling ball, the result- 

 ing diagrams showing the backward and forward motions of the ground 

 were much larger, and, therefore, better fitted for analysis than those 

 which had been obtained pi-eviously. It also became possible to place 

 the observation stations at greater distances apart, and thus errors in the 

 calculation of velocity arising from inaccuracy in time observations were 

 considerably reduced. 



The results obtained were a confirmation of results which had been 

 previously obtained. They were, briefly — 



1. A graphic separation of normal and transverse vibrations. 



2. A determination of the relative amplitudes and periods of these 

 vibrations at various points. 



3. The determination of the manner in which these vibrations became 

 extinguished. 



4. The velocity with which these vibrations were propagated. 

 To these observations the following may be added : — 



5. Vibrations, especially those performed in a normal direction, take place 

 more rapidly near to the commencement of a disturbance than at the end. 



6. The greatest motion of the ground, as shown by the normal vibra- 

 tions, is inwards towards the origin of the disturbance. 



7. The direction in which the ground moves with the greatest velocity 

 is also imvards towards the origin of the disturbance. 



8. The motion does not appear to be simple harmonic. 



9. The vertical motion is not due to a direct shock, but to a surface 

 nndulation. 



10. The velocity of propagation of a disturbance is not constant, but 

 varies with the distance from the origin. 



The last observation led to three new sets of experiments being 

 undertaken, the chief object of which was to determine the velocity with 



