62 Trans. Acad. Sci. of St. Louis. 



In obtaining such ionization of the air around the 

 grounded anode terminal, the plates of the machine 

 should be turned with extreme slowness. There should 

 be a gap of a couple of mm. at the machine terminal, and 

 the rotation should cease for a moment, if it is thought 

 that a spark may pass across the plate. After half a 

 minute the velocity of rotation may be increased and the 

 spark allowed to pass over the film. This method of pro- 

 cedure will give time for the ionization to be brought 

 about, and will occasionally give desired results. 



In Fig. C of Plate XXII the pin soldered to the end of 

 the terminal wire was so bent, that the rounded head 

 made contact at the central point of its rounded head with 

 the photographic film. The form of the shadow shows 

 that the cause for the fogging effect around this pin-head 

 lies in a very thin layer of air at the film. 



Fig. C of Plate XXII also shows that the origin of the 

 agency which produces the fogging effect is at the muzzle 

 of the discharge channel about midway between the ter- 

 minals. Fig. D is a print from Fig. C. Hundreds of 

 plates have been used in securing a comparatively few 

 specimens which show the dispersion of the disruptive 

 discharge at some point between the metallic terminals. 

 In no case has such a dispersion area of ionized air ex- 

 tended to the negative pin-head. These ionization ef- 

 fects always have their origin at the grounded pin-head, 

 although the presence of the negative pin-head is of 

 course necessary in order to make the grounded pin-head 

 effective. 



Fig. A of Plate XXIII shows one of two exposures 

 which have been obtained, in which the disruptive part of 

 the discharge is a very small part of the distance be- 

 tween the pin-heads. 



In both cases there is evidence that there was a dis- 

 charge towards the negative terminal such as might be 

 produced by an issuance of positively (or, perhaps, nega- 

 tively) charged particles from the negative end of the 

 spark channel. There is slight evidence of such dis- 

 charge in other plates, which show the negative discharge 



