156 Trans. Acad. Sci. of St. Louis. 



that could account for the peculiar discharge. The brass 

 wheel of a clock was selected, and all but one of its pointed 

 cogs were removed, with no better result. A return to the 

 weight gave a successful result, but continued repetition re- 

 sulted again in failure. A radial line was again drawn with 

 a different pencil from that formerly used, and with a suc- 

 cessful result, but this was again followed by a long list of 

 failures. It was finally found that the mark of a moistened 

 pencil would always } r ield the desired result. 



This threw discredit upon the idea which had begun to pre- 

 vail, that a definite frequency of oscillation was involved, and 

 which only a fortunate combination of adjustments could 

 secure. After a month of experimenting in this way the con- 

 ditions favorable to the immediate production of the phenom- 

 enon were found. 



The secondary circuit was discarded, and the metal disk 

 from which the discharge is to come was put in contact with 

 the negative knob of the machine. The disk was armed with 

 a radially placed needle point, which touched the sensitive 

 film. No condenser was used in the machine, and the knobs 

 were separated so that no visible discharge could occur be- 

 tween them. A needle point is also presented to the point 

 on the surface of the film. A very effective device consists of 

 two needles or pins with their eye or head ends lashed to- 

 gether, and attached with sealing wax to the end of a glass 

 tube serving as a handle. The needles form the arms of a T 

 of which the glass tube is the trunk. One point is held near 

 the point from which the balls are to issue upon the film. 

 The other point discharges upon the air. This device has 

 earned the name " teaser." It seldom fails to bring the ball 

 discharge at once. The mark of a moistened pencil upon the 

 film at the discharge point is sometimes needed. The teaser 

 may also be used to lead the balls into abnormal paths upon 

 the plate. When this device was hit upon, it was at once 

 used to determine whether ball discharges could be drawn 

 from the positive pole. The discharge point was placed at 

 the positive pole, with the teaser held in front of it. The 

 balls appeared, but they issued from the teaser and passed to 

 the positive pole. Several negatives were obtained in which 



