2$ Description of a new Compensation Pendulum. 



determinate increase of heat, from the quantity that is ex- 

 tended by the hammer ; much depends on the degree of 

 curvature and polish of the stake and hammer, and probably 

 on the heating of the bar at the time ; for it is necessary to 

 heat it a little hotter than boiling water, otherwise it will 

 crack in hammering. 



In all these experiments it is to be understood that the 

 ball of the pendulum was suspended by its centre ; but if 

 the ball be made to rest on its lower edge, the expansion 

 and contraction of it must be taken into consideration. 



It has been the opinion of some mechanists that zinc is 

 an unfit substance for a compensation pendulum, because 

 they have thought it too soft for the purpose, and that after 

 being heated or cooled to a considerable degree, it does not 

 return to its original dimensions. If that was really the case, 

 uo doubt but it would be a general one, common to all me- 

 tals in a greater or less degree ; but from the experiments 

 and observations I have made on zinc pendulums, 1 am fully 

 satisfied there is no foundation whatever for such an opi- 

 nion. Some time in the latter part of last summer, I how- 

 ever noticed a circumstance that made me doubt the matter 

 — for when J first used any zinc pendulum, I could never 

 bring the clock to keep the same rate two days together, 

 but it was continually retarded, whether I used the lamp 

 or not ; and had I not before observed a similar effect on a 

 lever pendulum that was made of brass and steel, I should 

 have ascribed the cause wholly to the softness of the zinc 

 bar; but by constantly comparing its daily rate with one 

 that had been going a longer time, I found this retarding 

 property gradually wore oft', and in less than a month would 

 become quite settled to the rate that it would afterwards 

 keep. By subsequent experiments with the lamp too, I 

 have constantly found that all the pendulums I have hi- 

 therto tried kept precisely the same rate, both during the 

 time they were heated (provided they were properly adjusted) 

 and afterwards, as they had done before. The cause of this 

 retardation appears to me to be, that the points of contact 

 of the different pieces, which compose the pendulum, are 

 more closely connected after a little time than they are at 



first. 



