130 PHILOSOPHICAL TRANSACTIONS. [aNNO 1726. 



degrees of heat, as nearly as could be determined ; conceiving it would not be 

 very difficult, by making use of two sorts of metal, differing considerably in 

 their degrees of expansion and contraction, to remedy, in great measure, the 

 irregularities to which common pendulums are subject. But although it is easily 

 discoverable, that all these metals suflTer a sensible alteration of their dimensions 

 by heat and cold ; yet he found their differences, in quantity from each other, 

 were so small, as gave no hopes of succeeding this way ; and made him leave 

 off prosecuting this affair any farther at that time. In the beginning of 

 December, 1721, having occasion for an exact level, besides other materials 

 Mr. G. made trial of, quicksilver was one; which, though he found it was by 

 no means proper for a level, yet the extraordinary degree of expansion ob- 

 served in it, when placed near the fire, beyond what he had conceived to 

 be in so dense a fluid, immediately suggested the use that might be made 

 of it, by applying it to a pendulum. 



In a few days after, he made the experiment, but with much too long a 

 column of quicksilver, the clock going slower with an increase of cold, con- 

 trary to the common pendulum ; however, it was a greater confirmation of the 

 advantage to be expected from it, since it was easy to shorten the column in 

 any degree required. The only doubt entertained, was, lest there should not 

 be a proportional expansion and contraction between the quicksilver, and the 

 rod of the pendulum, through the various degrees of heat and cold, from the 

 one extreme to the other. To make this experiment the more convincing, he 

 placed the clock in a part of the house the most exposed of any to the changes 

 of heat and cold, the room having no fire in it in the winter, and exposed to a 

 south sun, with leads above it, which in the summer made it extremely hot. 

 He hung a thermometer by it, and had also another clock at no greater distance 

 from it than was necessary to keep the cases from touching each other. This 

 clock had been made some years before, with extraordinary care, having a pen- 

 dulum above 6o pounds in weight, and not vibrating above one degree and a 

 half from the perpendicular ; and which, in a more temperate situation, had 

 not altered above 12 or 14 seconds in 24 hours, between winter and summer ; 

 but in this place it altered 30^ a day, between the hottest and coldest weather, 

 in the year 17'22, a year no way remarkable for either extremes. But this great 

 alteration was owing to the situation abovementioned, and which was chosen to 

 make the experiment the more sensible. 



The two clocks being firmly screwed to a party-wall, Mr. G. began to make 

 the first trial of this kind of pendulum, December 18, 17'21 ; and by Jan. 3, 

 perceiving the column of quicksilver considerably too long, he procured a 

 shorter glass, which he made use of till the beginning of June following: by 



