of minute Forces on the Rate of the Seconds Pendulum. 245 



about J of tin inch in length. I was so fortunate as to obtain 

 my object on the first trial ; for when the vibrations were 

 made to vary from 1°*5 to 2°\5, on each side of the point of 

 rest, they were equal to within less than half a second in 

 24 hours. 



I next adjusted the suspending spring of the other clock, 

 which had a mercurial pendulum, and a suspending spring 

 nearly \ of an inch in breadth, and T £ n of an inch in thick- 

 ness. A moveable clamp-piece was made to fasten near the 

 top of the pendulum rod, so as to be shifted up and down 

 at pleasure, and to clasp the suspending spring. By various 

 trials I found, that for this strength of spring and weight of 

 pendulum, the acting part required to be about ^V of an inch 

 in length. 



After both clocks had been going for some time, I was sur- 

 prised to find that each of them experienced a losing rate, but 

 particularly the shop time-piece with the light bob. Since 

 that period, which is about four years, I have had ample op- 

 portunities of satisfying myself that this loss of rate originated 

 in the weakening of the spring. 



The clamping or fastening of the spring in the manner de- 

 scribed, without any shortening of the absolute length of the 

 pendulum, occasioned the clock having the light pendulum to 

 gain about 3 minutes in 24 hours. I therefore lowered the 

 pendulum bob, and brought it to time. About a month after- 

 wards, I found it losing on its rate, and screwed up the pen- 

 dulum bob. It is now (at the distance of nearly four years) 

 screwed up to about half of the distance it was lowered at 

 first; a circumstance which shows how much the spring has 

 weakened since that time. The clock frequently maintained a 

 uniform rate for three or four weeks successively; after which 

 it began to lose on its rate, particularly when any sudden 

 change of temperature occurred. 



Pendulums with thin suspending springs (which most of 

 the best clocks have) will not be subject to this variation, after 

 having gone for some time, at least as far as my experience 

 enables me to judge. 



The clock just mentioned was lately cleaned. Before clean- 

 ing, it lost about a second in 24 hours, and the extent of its 

 vibrations was reduced to 2°; after cleaning, the extent of its 

 vibrations was 2 0, 5, its original extent when clean ; it then 

 gained 5 seconds in 24 hours, making a difference of 6 se- 

 conds all together. The cause of this difference was, less fric- 

 tion on the pallets. 



The impulse being given in the ascent occasioned all the 

 friction that affected the time to be in the descent, which con- 



