326 REPORT— 1893. 



As it is impossible here to desci-ibe all these cases in detail I refer 

 to the above paper ; only this may be mentioned, that in the case of the 

 great earthquake of Kumamoto two disturbances were observed, wliich 

 exactly agree with the supposition that the earthquake-wave travelled 

 round both sides of the globe with a velocity of 23 kilometres per 

 second. 



The time of a shock as obtained by this method is always rather un- 

 certain for two reasons. The first is tliat, owing to the slow rate of tbe 

 motion of the paper, which was 11 mm. per hour, the readings may be 

 2 or 3 minutes in error ; the second, that a shock very rarely begins 

 suddenly, but is generally preceded by smaller movements of the ground, 

 which make it impossible to decide wliich is the beginning of the eaith- 

 quake. In such cases, however, I have sometimes found the figures of 

 a disturbance to show remarkable coincidences. Thus, by comparing 

 the moments of sudden increase of motion at the two stations, which 

 occur in nearly every earthquake, several determinations of the difierence 

 of time may be obtained, the mean of which is affected by a smaller 

 probable error than a single determination. 



The velocities of propagation obtained from the above-mentioned 

 observations vary between 2 and 5 kilometres per second, which, how. 

 ever uncertain the single values may be, confirms the view that it is 

 impossible to speak of a constant of velocity in the distant propagation ot 

 earthquakes. 



If the first observations with the horizontal pendulum leave much to 

 be desired with respect to their application to the study of earthquakes, 

 they certainly show what might be done if only a small number ot 

 stations, well distributed all over the earth, were organised. It would 

 be a great satisfaction to the writer if this account should help to excite 

 an interest for the establishment of at least one such station in the far 

 West, say in the western part of the United States or Canada, and 

 another in the far East, in Japan, or rather in some other country 

 which is less subject to the eflFects of local earthquakes. "We may safely 

 predict that the comparison of horizontal pendulum curves obtained at 

 two such stations, with those at an intermediate European station, would 

 lead to most interesting results, not only from a seismological point of 

 view, but also with respect to the determination of the modulus of 

 elasticity of the upper strata of the earth. If, with this special object in 

 view, one should not lay any value on observing at the same time the 

 deflections of the plumb-line, the horizontal pendulum might be much 

 simplified, and perhaps an improved form of Milne's ' tromometer ' or 

 conical pendulum, when furnished with sufiiciently fine connections, 

 would be the cheapest and simplest instrument. The rate of the paper 

 ought to be much increased, and this could be done without raising the 

 expense, for it would not be necessary to employ a veiy high degree of 

 sensitiveness, the deflections of the pendulum would thus be reduced, and 

 a narrow strip of paper would be sufiicient where a broad sheet is now 

 required. 



(2) Earth Fulsations. — Earth pulsations, consisting of long, flat 

 waves, somewhat like the swell of the ocean, were for the first time 

 observed by myself at Potsdam on February 11, 1889. Owing to the 

 application of glycerine, as described above, the pendulum could not 

 swing, and a sharp black line was drawn by the light-point. The instru- 

 ment had been adjusted in the afternoon, but in consequence of some 



