an be produced, it would be a very
curious matter indeed to know what its relation to electrical phenomena
would be; and as shell-lac and metal are directly opposed to each other,
whether a vacuum would be opposed to them both, and allow neither of
induction or conduction across it. Mr. Morgan[A] has said that a vacuum
does not conduct. Sir H. Davy concluded from his investigations, that as
perfect a vacuum as could be made[B] did conduct, but does not consider the
prepared spaces which he used as absolute vacua. In such experiments I
think I have observed the luminous discharge to be principally on the inner
surface of the glass; and it does not appear at all unlikely, that, if the
vacuum refused to conduct, still the surface of glass next it might carry
on that action.

[A] Philosophical Transactions, 1785, p. 272

[B] Ibid. 1822, p. 64.

1614. At one time, when I thought inductive force was exerted in right
lines, I hoped to illustrate this important question by making experiments
on induction with metallic mirrors (used only as conducting vessels)
exposed towards a very clear sky at night time, and of such concavity that
nothing but the firmament could be visible from the lowest part of the
concave _n_, fig. 143. Such mirrors, when electrified, as by connexion with
a Leyden jar, and examined by a carrier ball, readily gave electricity at
the lowest part of their concavity if in a room; but I was in hopes of
finding that, circumstanced as before stated, they would give little or
none at the same spot, if the atmosphere above really terminated in a
vacuum. I was disappointed in the conclusion, for I obtained as much
electricity there as before; but on discovering the action of induction in
curved lines (1231.), found a full and satisfactory explanation of the
result.

1615. My theory, as far as I have ventured it, does not pretend to decide
upon the consequences of a vacuum. It is not at present limited
sufficiently, or rendered precise enough, either by experiments relating to
spaces void of matter, or those of other kinds, to indicate what would
happen in the vacuum case. I have only as yet endeavoured to establish,
what all the facts seem to prove, that when electrical phenomena, as those
of induction, conduction, insulation and discharge occur, they depend on,
and are produced by the action of _contiguous_ particles of matter, the
next existing particle being considered as the contiguous one; and I have
fu