14 Trans. Acad. Sci. of St. Louis. 
A after fixing was not observable during the developing 
of the film when it was viewed from above. It was seen 
on turning the plate over so as to expose the under side. 
This fogging had therefore been produced through two 
glass plates. 
Plate VIII shows two photographic plates exposed 
back to back in the same manner, when the negative dis- 
charge was passed through the upper wire. From the 
film of the lower plate B a negative discharge passed to 
the grounded wire below. This downward discharge pro- 
ceeded from an area which was coincident with that of 
the blackened area on the upper film A. This is revealed 
by the presence of small black points here and there to- 
wards which fine discharge lines proceed, and from 
which the discharge passed to the wire below. The form 
of these lines seems to have been somewhat affected by 
electro-magnetic induction from the discharge wire across 
the upper film. The fogging effect on a film from which 
electricity passes to a conductor, is much less than that 
caused by a like discharge of electricity against the film. 
It is this difference, which has always been ascribed to a 
difference between positive and negative discharges. 
The exposure of plates like those shown in Plates III 
and V, where the grounded wire was in place, was varied 
as follows: 
A blast of air from a large tank of 800 liters capacity 
and maintained at constant pressure of two and a half 
atmospheres, was blown across the end of the grounded 
wire below the photographic plate. The blast swept 
through the gap between this end and the photographic 
plate. This was done with both positive and negative dis- 
charge. The blast was also directed along the discharge 
wire in contact with the upper film. Not a trace of any 
effect on the discharge lines could be detected, although 
the blast was maintained throughout the entire exposure. 
The results thus far described seem to show conclu- 
sively that the apparent emission of positive electricity 
from the positive terminal of the influence machine, is 
