— 3 — 



The height of the kuife-edge of the pendulum rested ou a steel plate about 

 ll'o CQis. square aud about 2 cms. thick. This plate was clamped tightly by 

 means of two screws to a heavy iron bar 75 cms. long, 75 cms. broad, and 2 

 cms. thick. This was placed on the pillar and kept in position by a large block 

 of stone resting upon it. The pendulum was passed through a hole in the middle 

 of the plate, which was first carefully levelled. The levelling of this steel plate 

 was quite a difficult task ; it was never accomplished in less than an hour, while 

 at times it took more than three or four. Levelling with two screws is indeed 

 a mere matter of chance, aud it would certainly be an improvement if the plate 

 were provided with three levelling screws staudiug in A''s, so that it might bo 

 clamped geometrically to the heavy iron bar hj the weight of the pendulum 

 itself. This plan had been thought of before the expedition, but want of time 

 prevented us carrying it out. 



The time of oscillation was measured with a break-circuit sidereal chro- 

 nometer of Xegus construction. Its rate was compared every morning by tele- 

 graphic communications with another Negus chronometer in the astronomical 

 observatory; and the rate of the latter was determined by Mr. Tanakadate from 

 transit observations made on every clear night during the progress of the ex- 

 periments. The pendulum usually continued to vibrate for twenty to forty 

 minutes, except in some few cases in which, from special circumstances, longer 

 continuation was necessary. The interval between the beginniug and cud of 

 each set of oscillations was measured in the way above described. Since the 

 pendulum had been previously adjusted so that its time of oscillation was very 

 nearly one second, the total number of seconds between the beginniug and end 

 of a set of oscillations was to a first approximation the same as the number of 

 oscillations during the whole interval. The correction applied to this estimate 

 could be readily deduced from Chronographie records by microscopic measure- 

 ment. 



The amplitudes of oscillations were measui-ed with a small miliiuictro 

 scale fixed horizontally near the lower end of the pendulum, aud readings 

 were taken at the beginning and end of each series of oscillations. 



The teuiperature of the room was measured with a good Salerou theruio- 

 meter, hung near the pendulum. 



'I'lie first series of expcrimcnls at Tokio having been completed, we started 

 for Xaha ou the seventeenth of July. 



During our stay at Kagoshima, wliili,' wailinL;- t'^r the despatch of slcauier, 

 wo found ample time to make a series of ex|)eriments. " 'J'iirough the kindness 

 of Mr. C. Watauabe, tlie lünjosltiina-jccn-nri, a small building in the kiiichu was 

 put at our disposal, aud liiere the experiments were conducted. 'J'lie cliron- 

 metcr and the chronograph were placed iu (he same building; aud astronomical 

 observations were made iu the garden just in front of the building. A largo 

 cylindur of stone aljoiit 1.5 metres long and 0.7 melrcs in diameter, was erected 

 upright ou tlie ground inside the building; and on this slono liio iron bar was 

 set. The adjustments aud measurements were carried out as usual. Owing to 



