of view of the Kinetic Theory of Matter. 



35 



tential. In order to carry out this experiment, the bulb shown 

 in figs. 9 and 10 was constructed. A spherical bulb, 10'8 ctn in 

 diameter, has pivoted on a needle point at its center, a system 

 of four very light spherical aluminum cups, held by four arms 

 going out from a steel pivot, as is shown in fig. 9. These 



10. 



cups are nearly hemi-spherical and are about *5 cm radius of 

 curvature. The needle point is connected by a platinum wire 

 to the outside of the tube. 



When the bulb is thoroughly exhausted by means of a 

 Sprengel pump, the movable system is put in connection with 

 one terminal of the two-part commutator for rectifying the 

 alternating potential-difference of the secondary of the induc- 

 tion coil as described in an earlier section. The cups are seen 

 to rotate, concave side forward. On disconnecting the electri- 

 cal source, and grasping the bulb in the hands, the cups im- 

 mediately begin to rotate in the opposite direction and con- 

 tinue to rotate for a short time before coming to rest. The 

 electrification of the walls is very troublesome, but when they 

 are carefully discharged after each operation, the above direc- 

 tion of rotation, upon connecting the movable system with a 

 source of high potential, has been observed, whether the po- 

 tential of the source be positive or negative in sign. 



§ 7. On considering the heat-flow lines from the surface of 

 a hot body immersed in a cool gas, it is clear that the shape 

 of the lines of flow will be affected by the distribution of 

 temperature on opposing or neighboring surfaces. Thus, in 

 fig. 7, if the spherical cup be placed in a closely fitting cylin- 

 der, whose walls in the vicinity of the cup, are at a tempera- 

 ture but slightly different from that of the spherical cup, the 



