44^ jbefcr'iptktt of an Apparatus, a>i3 



MM, the obfervcf caufes them to move through a confiderable portion of a femlcircle, they 

 will arrive at a fccond pofition, in which their joint a£lion will move the arm in the con-, 

 trary direclion. Thefe weights are prevented from ftrikiiig the inftrument by pieces of 

 wood faftened to the wall of the building. 



- The filuation of the arm is determined by flips of ivory placed within the cafe near each 

 extremity of the arm. Thefe flips bear a fet of divifions of the inch into twenty parts, to 

 which a fmall ftip at each end of the arm applies, without contact, a vernier that affords a 

 fubdivifion into hundredth parts. Eflimation by the eye looliing through the fhort telefcopes 

 T and T, gives- the pofition of the arm to (till greater nicety. The lamps L and L throvf 

 light on tlie divifions by convex glafles, and no other light is admitted into the room. 



t" K is a wooden rod, which, by means of an endlefs fcrcw, turns round the fupport to 

 which the wire ^ / is faftened, and thereby enables the obferver to turn round the wire tiU< 

 the arm fettles in the middle of the cafe without danger of touching either fide. The wire 

 g L'li fattened to its fupport at top, and to the centre of the arni at bottom, by brafs dips, ia 

 which it is pinched by fctews. The different parts are drawn nearly in the proper propor- 

 tion to each other. 



Suppcfe the arm to be at reft and its pofition to be obferved, let the weights be thea 

 moved ; the arm wi! not only be drawn afide thereby, but It will be made to vibrate, and 

 its vibrations will continue a great while ; fo that, in order to determine how much the arm 

 is drawn afide, it is necefiary to obferve the extreme points of the vibrations, and from 

 thence to determine the point which it would reft at if its motion were deftroycd, or the 

 point of reft. To do this, Mr. Cavendilh obferves three fucceffive extreme points of vibra- 

 tion, and takes the mean between the firft and third of thefe points as the extreme point 

 of vibration in one direclion, and then affumes the mean between this and the fecond ex- 

 treme as the point of reft; for, as the vibrations are continually diminiftiing, it is evident 

 that the mean between two extreme points will not give the true point of reft. 



It may, he adds, be thought more exa6l to obferve many extreme points of vibration, 

 fb as to find the point of reft by different fets of three extremes, and to take the mean re- 

 fult; but it muft be cbferved, that, notwithftanding the pains taken to prevent any dif- 

 turbing force, the arm will feldom remain perfeflly at reft for an hour together ; for which 

 teafon it is beft to determine the point of reft from obfervations made as foon after the mo- 

 tion of the weights as poffible. 



The next thing to be determined is the time of vibration, which he finds in this man- 

 ner : He obferves the two extreme points of a vibration, and alfo the times at which the 

 arm arrives at two given divifions between thefe extremes, taking care, as well as he can 

 guefs, that thefe divifions ftiallbe on different fides of the middle point, and not very far 

 from it. He then completes the middle point of the vibration, and by proportion finds the 

 time at which the arm comes to this middle point. He then, after a number of vibrations,, 

 repeats this operation, and divides the interval of time betweenthe coming of the arm to 

 thefe two middle points by the number of vibrations, which gives the time of one vibra- 

 tion. The propriety of this method is more fully explained in the Memoir itfelf. 



In Mr. Cavendifti's firft experiments the wire by which the arm was fuf^ended was 39^- 

 inches long. It was of copper filvered, one foot of which weighed a/^ths grains ; and its 

 ftiffhefs was fuch as to make the arm perforraone vibration in about 15 minutes. It was 



imraediatel}r. 



