the Light of Bodies in a State of Comhufion. 203 
in fulated fubftance. I brought the ball within the {hiking 
■diftance of my conductor, and the {parkin palling from the con- 
ductor to the ball appeared very brilliant ; but the whole length 
of the filver thread appeared faintly luminous at the fame in- 
ffant. In other words, when the fpark was confined within 
the dimenfions of a fphere one-eighth of an inch in diameter, 
it was bright, but, when dfjtufed over the furface of air which 
received it from the thread, its light became fo faint as to be 
feen only in a dark room. If I leflened the furface of air 
which received the fpark by fhortening the thread, I never 
failed to increafe the brightnefs of the appearance. 
exp. ix. To prove that the faintnefs of the eleCtric light in 
vacuo depends on the enlarged dimenfions of the {pace through 
which it is diffufed, we have nothing more to do, than to in- 
troduce two pointed wires into the vacuum, fo that the fluid 
may pafs from the point of the one to the point of the other, 
when the diftance between them is not more than the one-tenth 
of an inch. In this cafe we fhall find a brilliancy as great as 
in the open air. 
exp. x. Into a Torricellian vacuum, 36 inches in length, I 
conveyed as much air as would have filled two inches only of the 
exhaufted tube, if it were inverted in water. This quantity of 
air afforded refinance enough to condenfe the fluid as it pafled 
through the tube into a fpark 38 inches in length. The bril- 
liancy of the fpark in condenfed air, in water, and in all fub- 
ftances through which it pafles with difficulty, depends on 
principles, flmilar to thofe which account for the preceding fads. 
I would now proceed to {hew, 
5. That in the appearances of electricity, as well as in thofe 
of burning bodies, there are cafes in which all the rays of light 
do not efcape ; and that the molt refrangible rays are thofe 
D d 2 v which 
