- Eocperiments made at Ceylon. 



411 



The marginal dia- 

 gram will explain our 

 method of putting the 

 pendulum in motion, 

 which our observations 

 proved to us must be .^^^ 



aoiJOifui 'tot irrio'i fBomon 



B' 



J^ . 



,i'Ki{*lUlJ 83 viol 



III 



iiibnijp jV'io rrojjolo8 •^ririo gnfiocq 

 > 'io.a:J/iOf;.?rft')oo'Drff ^loot o?foxlw 16 



:\s^«,\$1o?:iiii :tiifjpf>oidf/o 



tiVbflH ^YlOadi ^ lil J lu < KJ f ^<i.UOKib odT 



done with great pre-^ .j^iy \o\^Am ^di m) an-d ydi smas.i8^*5 

 cision. At the cir4, jjyv«:, /n* lo °" " " 't r .T. „ 

 cumference of the circle : 

 over which the pen- 

 dulum was to oscillate, 

 a box A was made sta- j^«'<tsfral\ 

 tionary by means of K\L 



weights. On its upper «apfp¥ 



surface, at the extremity nearest the circle, was placed a small 

 block of hard wood, C, which served for adjusting the proper 

 amount of elevation, and for supporting the thread which held 

 the ball B by means of a noose, and was stretched over the block 

 C and secured to the other end of the box A. The ball of the 

 pendulum being brought to a state of perfect rest in a point of 

 suspension vertically over the selected line of oscillation, the 

 thread was severed on the block by the vertical pressure of a 

 sharp knife : this mode of releasing the ball was attended with 

 less chance of extraneous motion than by burning the thread, or 

 by any other means we could devise. 



Observations were recorded every hour. When the motion of 

 the pendulum was renewed, or if ellipticity was not observed, it 

 was allowed to continue for another hour. i 



Previous to the experiment here recorded, many attempts 

 were made with balls of less weight, thicker wires, different 

 modes of suspension and starting, &c. ; but the ellipticity of the 

 pendulum's motion was in all of them so considerable, that the^ 

 results were very unsatisfactory and are not here introduced. 



^limMqy 



I 



«8di Miiii^i. 



Direction of the initial motion 

 witli reference to the mag- 

 netic meridian. 



N. and S. 



N.E. and S.W. 



. ■cMUiSlishi]. 



E. ana^^iio 

 N.W.andS.E. 



Angular variation observed 



after the lapse of intervals 



of one hour each. 





Jl 





70 



h \ih-\ 



Mean variation per hour 1*87 

 Calculated do 1-8111 



>iyN. 





)hm '^ >4Xixad 



/Jji fsibi 



