ON THE EARTHQUAKE AND VOLCANIC PHENOMENA OF JAPAN. 123 



movements of the writing-pointers, the latter quantities varying with 

 the amount of stability given to the horizontal pendulums. Immediately 

 after the earthquake Mr. F. Omori very kindly made such determinations 

 for a seismograph in the laboratory of the Imperial University, with the 

 result that the maximum slopes which this seismograph had recorded were 

 about one-third of a degree. Waves with these slopes, as shown on the 

 diagram, succeeded each other at intervals of about 2'2 seconds. The 

 vertical motion which was recorded was about 10 mm. ; but as ordinary 

 lever spring instruments, when the levers are not parallel to the wave- 

 fronts, are as sensitive to tilting as horizontal bracket or conical pendu- 

 lum seismographs, these measurements must be regarded as maximum 

 rather than actual values. Combining the maximum wave-slopes with 

 these records of vertical motion, we obtain certain values for the lengths 

 of the waves, which may be taken at 18 or 20 feet ; and as we know their 

 period, we may determine their velocity of propagation, which appears to 

 have been about 8 feet per second. This is exceedingly slow, but not- 

 withstanding the errors in the observation of vertical motion, I do not 

 think the velocity exceeded double this amount. The velocity of propa- 

 gation of more truly elastic vibrations will be referred to later. 



From these observations, which I think are made for the first time, 

 rather than concluding that modern seismographs are useless whenever 

 vertical motion occurs, we see that on such occasions they must be 

 regarded as angle-measurers. The action of any bracket seismograph 

 when recording horizontal motions depends greatly upon its inertia, but 

 to obtain the best measurements of tilting, any cause likely to produce 

 swinging should be minimised. To obtain a true measurement of vertical 

 motion, the method which first suggests itself is to have a number of 

 spring lever arrangements in different azimuths, the one which happened 

 to have its arm at right angles to the direction in which the wave 

 advanced being the one which would give the best results. 



Independently of any new instruments which may be devised to 

 measure tilting, we now know that the instruments we already possess 

 have a double function, not only naeasuring horizontal displacements, but 

 also measuring angles of tilting. In order to take advantage of this 

 second function, it is necessary that when a bracket or conical pendulum 

 instrument is once set up, experiments should be made to determine the 

 effects of tilting, otherwise, should it be tilted by an earthquake, its records 

 will not be measurable. 



An investigation of considerable importance in connection with the 

 intensity and direction of motion, which has been carried out by Mr. F. 

 Omori, relates to the overturning of bodies of various dimensions. At 

 all temples, which are more thickly distributed than the towns and 

 hamlets, there are stone lanterns, standing on circular or square pedestals, 

 whilst in the vicinity there are hundreds of gravestones, which are square 

 or rectangular in section, and stand freely on their ends. Applying the 

 overturning formula to some thousands of these which were overturned 

 in the Nagoya-Gifa Plain, average minima and maxima values for the 

 accelerations experienced at different points within the earthquake area 

 have been determined. Inasmuch as the results given by the formula, 

 which is due to Professor C. D. West, conform with the results obtained 

 by experiment, we have every confidence in the figures given in the 

 following table : — 



