246 Mr. Scrymgeour's Experiments to determine the Influence 



sequently retarded the motion and caused the time to be slower. 

 The friction beyond the escaping point being equal in ascent 

 and descent, occasions no difference in time. 



I next put a lighter maintaining power on the clock, which 

 reduced the extent of vibration to 2°, its extent before it was 

 cleaned ; but it still gained 5 seconds in 24 hours. This result 

 shows that the spring was such as to cause the vibrations to 

 be equal in time, and that the loss in its rate before cleaning 

 was occasioned by friction. 



I was now desirous of ascertaining the rate of the pendu- 

 lum of this clock in a detached state. For this purpose, I 

 suspended a detached pendulum, with an adjusted suspending 

 spring, close in front of the clock, and adjusted it so as to 

 vibrate with the clock pendulum. I then detached the pallets 

 from the crown-wheel. The collet to which the pallets were 

 riveted being fixed to the arbor by a screw, they could thus 

 be put out of action, clear of the wheel, and fixed again upon 

 the arbor. In this construction, the pallets, with the arbor 

 and back-fork, form part of the pendulum. 



The pendulum was put in motion by the hand, a little be- 

 yond its extent of vibration, so that the mean extent in the 

 detached state should be the same as when attached to the 

 clock. The experiment showed that it gained upon the de- 

 tached pendulum. I then adjusted it to the detached pendu- 

 lum, and put it in action with the work as before. It now 

 lost about 2 \ seconds in 24 hours upon its former rate. This 

 experiment, therefore, showed the influence of the escape- 

 ment upon the time of the vibrations, which was occasioned, 

 conformably to what has been already stated, by the remain- 

 ing friction upon the pallets in the descent, in consequence of 

 the impulse being given in the ascent. 



'It is proper to mention here, that the bob of this clock tra- 

 versed close to the back of the case; and the resistance thus 

 occasioned, accounts for part of the loss of 2 \ seconds, as will 

 be shown hereafter. 



The same experiment was next tried with the clock having 

 the mercurial pendulum, of about 13J pounds in weight, and 

 furnished with the strong adjusted suspending spring. The 

 result showed a loss amounting to within 2 or 3 tenths of a 

 second of the loss of rate exhibited in the former experiment. 



In order to give the impulse in the descent instead of the 

 ascent, the following means were adopted. The flanches of 

 the pallets were hollowed by forming them thus : / with 



a small cylindrical lap of about £ of an inch in diameter. 

 These pallets were made separate, and being jointed upon their 

 axes of motion, the distance between them was regulated by 



