160 Trans. Acad. Sci. of St. Louis. 



In many cases such attempts to form branch tracks prove 

 abortive, but all of the tracks which are maintained for a 

 sufficient time show most elaborate branches, of the most 

 beautiful form. This is especially true when the photographic 

 plate is not too near the machine, and when the teaser is only 

 used to start a ball, which is then allowed to wander over the 

 plate. The most interesting ball discharge yet obtained was 

 found on a plate inclosed in a paste-board box in which pho- 

 tographic plates are packed. The discharge was disruptive 

 in character. The negative terminal was a small knob in con- 

 tact with the coin, from which a wire passed through the 

 center of the cover, aud was held in place by sealing-wax. 

 The ball discharges wandered over the entire plate. They 

 even branched off towards the coin and two such branches 

 ran under the coin itself. The X-ray was playing upon the 

 plate during the exposure, but this was apparently not an 

 essential feature. The walls of the box undoubtedly did 

 have some influence. 



The essential differences between the three kinds of dis- 

 charge described are well shown in Fig. 1, wmich contains 

 them all. Ordinary visible disruptive sparks are shown 

 most sharply in that part acted upon by the light during the 

 discharge. The} 7 are not sharply defined, and are curiously 

 bent near their outer ends. When first appearing on this 

 negative, they were dark. As the development was pushed, 

 they reversed. The dark corona on the negative immediately 

 around the brass weight was produced by radiations along the 

 lines of force. They were straight lines or nearly so, and 

 the discharges which produce them are invisible. Within 

 this coronal discharge in the shaded part of the plate will be 

 found the track of a ball discharge, which on an ordinary 

 silver print comes out very sharply, under a strong lens, but 

 the reproduction will not bear very much of magnifying. 



Fig. 9 shows a ball lightning discharge from a plate 

 which was pushed somewhat in the developing bath. It would 

 have been very nearly as good if it had been fixed without 

 developing at all. It is, however, somewhat better to de- 

 velop, if the arrangement has been such that disruptive sparks 

 have not passed over the plate. 



