16 LEWIS, ON THE ELECTRIC LIGHT. 
condensed-discharge sparks rapidly enlarged, and burnt out 
the edges of the holes, altogether destroying their original 
forms. 
Observing that not infrequently sparks deviated from their 
direct course, in order to pass through adjacent portions of the 
paper which offered them less resistance, I tried some French 
insulating paper, but the heat of the sparks here melted 
the wax or composition with which the paper was saturated, 
so that it not only surrounded the holes in a thickened con- 
dition, but there was evidence, in most instances, of its hay- 
ing overflowed them, and thus materially interfering with the 
shape of their actual margins. The perforation of thin mi- 
croscopic glass was not more successful, for in every case the 
spark, on striking the surface, coursed along it for a consider- 
able distance before passing through, splintering or fusing it 
in such a way as to render it impossible to decide what was 
the real shape of the perforation. 
A Leyden jar was next charged in the ordinary way by 
means of a common cylinder machine, and the sparks from it 
were caused to pass through a card which was placed between 
the knob of the jar and that of the discharger. The disrup- 
tive effect of these was very violent, large burrs being raised 
over the holes, which entirely prevented their shape from 
being seen; thin sections were, however, made with a sharp 
microscopic dissecting-knife, and here again the five-sided 
character of every hole was clearly made out, notwithstanding 
the quantity of loose fibre which was strewed across them. 
Of the results of the foregoing experiments I have made 
careful drawings, by means of the camera-lucida (Pl. II); 
the micrometer scale will be found marked on each 
figure. In several instances the outlines of the fibre of the 
paper are shown, from which, I think, it will be evident that 
the peculiarity of shape is not due to the texture of the 
material. 
The figures show the effect of induction-sparks of various 
lengths ; some that of condensed-discharge induction-sparks ; 
while Fig. 6 shows that of discharge-sparks from a Leyden jar 
charged with frictional electricity. 
It must be borne in mind that in every experiment ex- 
cept that with the Leyden jar the paper or card was 
held by the hand, and was moved about between the termi- 
nals during the passage of a very rapid succession of sparks ; 
it is therefore most probable that several sparks passed through 
the greater number of holes, and, in consequence of the move- 
ment of the paper, the majority of these must have passed 
through in a more or less oblique direction. 
