222 REPORT— 1893. 



projecting downwards, as in an ordinary balance, is relatively deflected 

 through an angle corresponding to the tilt. This pointer moves a hori- 

 zontal lever, at the outer extremity of which a sliding needle writes its 

 record on a smoked-glass plate. Two such pieces of apparatus at right 

 angles to each other, writing on the same surface, constitute a complete 

 instrument. 



In one apparatus the balance-arm with its weights is replaced by 

 a heavy metal disc, supported in a vertical plane by knife edges at its 

 centre. 



Already one or two earthquakes have been recorded, and as these are 

 the first written records of earth- waves, a portion of one of them is here 

 reproduced :^ 



Fig. 2. 



It shows the E. and W. tilting during a small portion of an earth- 

 •quake which occurred at 5.40 p.m. on January 8, 1893. The numbers 

 indicate successive seconds, from which we see that the period of the 

 waves varied from I- to ^ second. The average angular deflection was 

 about 2' 40", and the smallest about 1' 30". 



The movement continued over at least 20 seconds, dying out with 

 hardly perceptible waves having periods of about ^ second. The N. and 

 ■S. component of tilting was exceedingly small. The direction in which 

 the waves were propagated was approximately E.N".E. to W.S.W. 



Inasmuch as tilting apparently occurs whenever we have vertical 

 motion, an unpleasant conclusion — which, however, is not expressed for 

 the first time — is that all the records hitherto published in Japan where 

 vertical motion has been recorded are of but little value. Not only may 

 the horizontal motion have been exaggerated, but the records of vertical 

 motion have also suffered distortion, this being greatest when the arm of 

 the lever seismograph has been parallel to the direction of the wave- 

 slope. The disturbances in which the vertical component has been 

 mai'ked form about 10 per cent, of what should be our most important 

 records. 



What we require to know, for example, as an assistance in investi- 

 gations relating to construction is the configuration, dimensions, and 

 rapidity of recurrence of these earth- waves. 



The varying slope of the waves, their period, and their direction of 

 advance, may be measured by the apparatus described. 



As an attempt to measure the vertical component of these waves, four 

 lever seismographs have been arranged with their arms at 45° to each 

 other, it being assumed that the record from the instrument with its arm 

 most nearly at right angles to the direction of the advancing wave will 

 be tbe one which will most closely measure the vertical motion. 



Another possible method of measuring this element of earthquake 

 motion would be to avoid errors consequent on tilting by arranging a 

 vertical lever seismograph on gimbals. 



