27<5 FHILOSOPHICAL TRANSACTIONS. [aNNO 1751. 



is once well adjusted, it will be found to lengthen or shorten the pendulum to as 

 great a degree of exactness, as any other method whatever. 



The method he took for adjusting the longer arms of the levers of the pen- 

 dulum to the shorter ones, is represented in fig. 4. To a strong post, fixed to 

 the wall, is fastened a small shelf, supported by two brackets a, b. In the middle 

 of this shelf is fastened a wire, by the screw e; to the end of which the pen- 

 dulum is to be hung. Below this shelf, at the distance of about 40 inches, is 

 placed the index cd, turning freely on a centre; the length of the index is 50 

 inches. At the distance of half an inch, on a part of the index produced be- 

 yond the centre, is placed a steel pm; and in the back of the pendulum, as near 

 the centre of oscillation as may be, is drilled a hole to receive this pin ; when the 

 pendulum is hung on the wire against the post, and the wire is screwed higher 

 or lower by the screw e, till the pin resting against the upper part of the hole 

 (which is filed into a proper shape for that purpose) keeps the index nearly in a 

 horizontal position. Below the bottom of the pendulum is placed a second index 

 fg, exactly like the former, except that it is kept in a horizontal position by the 

 screw k, bearing against the end of the iron rod. When the experiment is to 

 be made, the pendulum is first put into a box, and gradually heated by a large 

 fire, to a considerable degree, being often turned, that every part may be equally 

 exposed to the fire. And having continued shut up in the box for some time 

 after it is removed from the fire, that the two rods may be heated as uniformly 

 to the same degree as possible, the pendulum is hung on the wire, and the two 

 indexes made to stand nearly in a horizontal position. The two graduated 

 plates h, i, are then slid on a wire, till the divisions in each marked o are pointed 

 to by the indexes. As the pendulum cools, the lower index will be seen gradu- 

 ally to descend; but if the ends of the two screws, in the ball of the pendulum, 

 act on proper parts of the levers, the upper index will continue in the same 

 place. If the ends of the screws be either too far ofi^, or too near the centres 

 of the levers, the index will either rise or descend; and by comparing the number 

 of divisions it has varied, with those which the lower index has varied, a near 

 estimate may be made, how much the screws require to be altered; and, in a 

 very few trials, they may easily be adjusted to a very great exactness. In order 

 to make an actual trial how far this contrivance of the pendulum will answer the 

 end proposed, it is necessary, that the clock, to which the pendulum is fitted, 

 be made with great exactness, and entirely to be depended on; for otherwise the 

 experiments will be verv uncertain, as he found in the clock he first made use 

 of. In order to render this clock as perfect as possible, he made it in several 

 respects difl^erent from the common ones, in hopes of removing some imperfec- 

 tions he apprehended they were liable to. But as in this attempt he fell into an 



