82 Trans. Acad. Sci. of St. Louis. 



were mounted around the two spheres in their equatorial 

 planes. The ends of the wires were 14 cm. from the 

 axis of the tube, which was along the polar diameters 

 of the spheres. The wires on the two spheres curved in 

 opposite directions, so that each group would produce 

 whirling in an opposite direction. The device was hung 

 in a vertical position on a bundle of fibers of unspun 

 silk. From the lower sphere a small weight was hung 

 on a similar bundle of silk. 



One object to be secured was a symmetrical arrange- 

 ment of matter at the upper and lower ends of the whirl- 

 ing device. 



Vertical discharge rods were placed above and below 

 the whirl, so that sparks could be sent along the silk 

 fibres to the two balls. The tube was gently clamped in 

 a friction brake of hard rubber, the force being applied 

 by means of a flexible rubber band. This was necessary 

 by reason of the fact that local clouds of ionized air in 

 the room produced rotations of the whirl which were 

 sometimes in one direction and sometimes in the op- 

 posite direction. This device was much more sensitive 

 to differential point effects than the wire lying in the 

 glass tube. The air in the room has occasionally been 

 in such a condition that the frictional contacts could be 

 wholly removed, and the device would be in perfect equi- 

 librium while sparks were passing. Eeversing the dis- 

 charge gave the same result. 



Such stability has been maintained for intervals of 

 several minutes, and would only be disturbed by move- 

 ment of the obsem^er, or other causes producing a move- 

 ment of the air, such for example as gusts of wind. 



The apparatus shown in Fig. 1, has been varied in 

 length from 20 to 50 cm. The ends of the wire have been 

 tipped with small spheres of solder, in order to diminish 

 point discharges from the two ends into the surrounding 

 air. So far as could be observed, the creeping of the 

 wire was not affected by such changes. 



There have been conditions when no creeping of the 



