48 BRUSH— KINETIC THEORY OF GRAVITATION. 



such as bismuth and zinc for instance, whose attractions for the 

 nearer silver ball are to be compared. The cylinders are carried on 

 the ends of a thin strip of wood D, which is pivoted at its center to, 

 and supported by, a thick disc of cast iron E whose upper face is 

 dressed flat and leveled. The height of E is such that a horizontal 

 plane midway between the upper and lower ends of cylinders B and 

 C is in the center plane of the balls A. The carriage D is covered 

 with tin-foil kept in metallic contact with £ by a brass-wire spring. 

 E is permanently grounded ; thus B and C are always grounded. 



The cylinders B and C are very carefully so placed on the carrier 

 D that when the latter is revolved i8o° and brought against a re- 

 movable stop-pin F , C will occupy exactly the same position in re- 

 spect to the balls A as did B before the reversal. 



All the metals experimented with are in the form of cylinders of 

 the same size, 4.9 cm. high and 6.1 cm. diameter. When in position, 

 the surface of a cylinder is 1.3 cm. from the center of the nearer 

 silver ball. 



The zinc cylinder weighs 1.014 kg., and the other cylinders weigh 

 more or less than this according to their several specific gravities. 



The zinc cylinder attracts the nearer silver ball with a force of 

 about one three hundred thousandth part of a dyne, and as the 

 oscillating system is very sensitive, having a free period of seven and 

 a half minutes, the excess of this attraction over that for the more 

 distant ball gives a scale deflection of about 4.2 cm., which is ample 

 for observation, because deflections are easily read to o.i mm. As 

 the mirror doubles the real deflection, the latter is 2.1 cm. at a dis- 

 tance of 2 meters. Hence the silver ball moves about 0.2 mm. 

 toward the attracting cylinder, where the attraction is about i per 

 cent, greater. This change in attracting force is approximately cor- 

 rected by so locating the cylinder B that the angle a h c \s slightly 

 obtijse at the start, and becoming more so as the ball advances 

 causes the attractive efifort to be less effective. Hence the deflection 

 as read by the telescope may be taken as a closely approximate 

 measure of the attraction of the cylinder for the ball. Of course, 

 the center of attraction in the cylinder does not lie in its axis, though 

 near it. But this does not matter, because its location is the same in 

 all the cylinders. 



