310 



PROCEEDINGS OP THE AMERICAN ACADEMY. 



{ 



-^H& 



^ 



) 



h 



Dl 



the use of the metal balls, an occasional slight adhesion of the ball to 

 one side of the hole from the bottom of which it was released (see top of 



Fig. 1), due to the accidental presence there of 

 a bit of wax intended for use on the suspending 

 thread. This source of error was discovered 

 April 17, 1902, and wax was thereafter avoided. 

 Moreover, as each ball, in spite of considerable 

 care given to making it clean before dropping, 

 might possibly leave some small amount of 

 grease adhering to its circular line of contact 

 with the release plate, this part of the plate was, 

 in all later work, wiped after the dropping of 

 each ball. The usual instrument employed for 

 this purpose was a button of cork fastened at 

 the lower end of a wire. 



Previously I had attributed the occasional 

 very bad behavior of a ball to a jerk given by 

 a stray fibre of tlie silk suspending thread 

 catching on some little roughness of the neigh- 

 boring metal ; and it is probable that some of 

 the irregularities observed were due to this 

 cause. More than once I heard a sharp clitik 

 from the pull of a fibre when a suspended ball 

 was slowly revolving. It was to prevent such 

 complications that wax had been used on the 

 thread. Fine copper wire was for a time used 

 in place of silk thread for the suspension. It 

 soon appeared, however, that the time required 

 for the melting off of the wire, though short, was long enough to allow 

 the ball, before full release, to swing off a trifle from the side on which 

 the flame was applied to the wire. On April 23 and thereafter, fine 

 cotton thread was used for the suspension, iti the hope that its fibres 

 would prove less troublesome than those of the silk ; but, as an addi- 

 tional precaution, that part of the thread from which trouble was feared 

 was singed, just before the ball was put into position, by passhig it very 

 quickly through a Bunsen flame. 



After these changes of practice had been made, cases of exceedingly 

 bad behavior on the part of the balls, such as occurred often before, 

 were very infrequent, and I was henceforth able to make no great 

 improvement in their performance. It has already been stated that 



A 



Tf 



Figure 1. 



