BRUSH— A KINETIC THEORY OF GRAVITATION. 173 



part in 35,000. Thus a weight-mass difference effect appears fairly 

 well established and is impressive. 



Effect of Unequal Weight of Bobs. — In the foregoing experiments 

 the zinc cylinder or bob weighed about 7 grams less than the bismuth 

 (about y2 per cent.) ; and it was known that this would tend to 

 " speed up " the zinc and make the period difference or " zinc-bismuth 

 effect " appear less than if the bobs were equal in weight. But it 

 was thought that the effect of this weight difference was so small that 

 it would not be worth while to delay the experiments several days to 

 correct it, though computation would have shown differently. 



Last summer a review and extension of the pendulum experiments 

 was commenced, and the first thing done was correction of the weight 

 deficiency of the zinc cylinder. This was effected by placing a tightly 

 fitting equatorial band of thin sheet zinc on the cylinder and driving 

 it slightly one way or the other until perfect adjustment of center of 

 gravity was had, as and for the reason heretofore explained. Finally; 

 the zinc band was secured from accidental displacement by several 

 very small drops of solder at its edges. As it turned out, the banded 

 zinc, instead of being 7 grams lighter than the bismuth as before cor- 

 rection, was then about one tenth gram heavier than the bismuth. 

 This latter difference is trivial and was not corrected. 



The effect of equalizing the weight of the zinc and bismuth was 

 unexpectedly large. Instead of the bismuth gaining one beat in 7 

 hours 18 minutes, as shown in Table I., it now gained a beat in 

 5 hours II minutes. 



The reason for the large disturbing effect of any considerable 

 weight difference between the bobs is perfectly clear. It is due to 

 the distribution of weight in the equal pendulum rods. Obviously, 

 if we could have stretchless and weightless rods, even large weight 

 differences between the bobs, whatever they were made of, would 

 have no disturbing effects except those due to different diameters, 

 which would change their center of oscillation and air resistance. 



But the center of oscillation of a pendulum rod alone is very far 

 above that of its bob,, and it is the resultant center of oscillation of 

 the combined rod and bob that we have to deal with. Clearly, then, 

 if we reduce the weight of a bob, we raise the center of oscillation of 

 the combination and thus shorten its period. 



