VOL. XXXIV.] PHILOSOPHICAL TRANSACTIONS. 131 



which time he was well satisfied of the advantage of the contrivance, though 

 both these pendulums were but rudely executed, and this last had the pillar of 

 quicksilver too short, but much nearer the true length than the first. This 

 encouraged him to provide another glass, a little longer than the last, and to 

 bestow more care on all the parts of the pendulum that required exactness. 

 This being finished, by the Qth of June, he began then to observe the motion 

 of the clock, by the transits of the fixed stars, as often as the weather permitted, 

 making use of a telescope which moved in the plane of the meridian; with this 

 instrument he could be sure of not erring above 1 seconds in time. The clock 

 was kept constantly going, without having either the hands or pendulum altered, 

 from the Qth of June, \712, to the 1 4th of October, {"Jib, being 3 years and 

 4 months. 



For the first year, Mr. G. wrote down every day, the difference between the 

 two clocks, with the height of the thermometer, not omitting the transits of 

 the stars, as often as it was clear. The result of all the observations was this, 

 that the irregularity of the clock with the quicksilver pendulum, compared with 

 the transits of the stars, exceeded not, when greatest, a 6th part of that of the 

 other clock with the common pendulum ; but for the greatest part of the year, 

 not above an 8th or Qth part ; and even this quantity would have been lessened, 

 had the column of mercury been a little shorter; for it differed a little the con- 

 trary way from the other clock, going faster with heat, and slower with cold ; 

 but no alteration was made in the length, to avoid an interruption of the ob- 

 servations. To confirm this experiment the more, about the beginning of July, 

 1723, Mr. G. took off the heavy pendulum from the other clock, and made 

 another with quicksilver, but with this difference, that instead of a glass tube, 

 he made use of brass, and varnished the inside, to secure it from being injured 

 by the mercury. This pendulum he has made use of ever since, and finds it 

 about the same degree of exactness as the other. The reason why this kind of 

 pendulum is more exact than the common sort, will be evident to any one who 

 considers, that as heat lengthens the rod of the pendulum, at the same time it 

 increases the length of the column of quicksilver, and its centre of gravity is 

 moved upwards : and when by cold, the rod of the pendulum is shortened, the 

 column of quicksilver is likewise shortened, and its centre of gravity carried 

 downwards: by this means, if the column of quicksilver be of a proper length, 

 the distance between the point of suspension and the centre of oscillation of 

 the pendulum, will be always nearly the same, on which the exact motion of a 

 clock principally depends. Were the pendulum of a clock to remain invariably 

 of the same length, yet some little inequalities would appear in its motion, from 

 the difference of friction, arising from the iaiperfectioiis of the materials, as 



s 2 



