electrical phenomena exhibited in vacuo. 67 
to the glass. This, and another circumstance, namely, that 
when the leg in which the torricellian vacuum was made was 
15 or 16 inches long, the phenomenon was very rarely per- 
ceptible, and always disappeared when the tube was inverted, 
and the mercury made to strike the top with some force, led 
me to conclude that the minute space was really filled with 
the vapour of mercury ; the attraction of the particles of the 
fluid mercury for each other preventing their actual contact 
with the glass, except when this contact was forcibly made 
by mechanical means ; and I soon proved that this was the 
case : for by causing the mercury, when its column was 
short, to descend into the more perfect from the less perfect 
vacuum, with more or less velocity, I could make the space 
more or less, or cause its disappearance altogether, in which 
last case the cohesion between the mercury and the glass 
was always extremely strong. 
I found that in all cases when the mercurial vacuum was 
perfect, it was permeable to electricity, and was rendered 
luminous by either the common spark, or the shock from a 
Leyden jar, and the coated glass surrounding it became 
charged ; but the degree of intensity of these phenomena 
depended upon the temperature : when the tube was very 
hot, the electric light appeared in the vapour of a bright green 
colour, and of great density; as the temperature diminished, 
it lost its vividness ; and when it was artificially cooled to 
20 0 below zero of Fahrenheit, it was so faint as to require 
considerable darkness to be perceptible. 
The charge likewise communicated to the tin or platinum 
foil was higher the higher the temperature ; which, like the 
other phenomenon, must depend upon the different density of 
