138 
at the earth’s surface, to do the same amount of work, we 
have 
NADp/i. = 2 x 10 _l ° x Ng> 
which gives 
7 2 x 10- 10 
h ~ pD 2 ’ 
or, as o=8, nearly enough for the present rough estimate, 
h - 1 . 
(200000D) 2 
Hence if 
D = 2 o-oVoo- centimetre, 
h— 1 centimetre. 
The amount of energy thus calculated is not so great as to 
afford any argument against the conclusion which general 
knowledge of divisibility, electric conductivity, and other 
properties of matter indicates as probable : that, down to 
thicknesses of tooVoo of a centimetre for the metal plates 
and intervening spaces, the contact electrification, and the 
attraction due to it, follow with but little if any sensible 
deviation the laws proved by experiment for plates of mea- 
surable thickness with measurable intervals between them. 
But let D be a two-hundred-millionth of a centimetre. If 
the preceding formulae were applicable to plates and spaces 
of this degree of thinness we should have 
h = 1,000,000 centimetres or 10 kilometres. 
The thermal equivalent of the work thus represented is 
about 248 times the quantity of heat required to warm the 
whole mass (composed of equal masses of zinc and copper) 
by 1° cent. This is probably much more than the whole 
heat of combination of equal masses of zinc and copper 
melted together. For it is not probable that the compound 
metal when dissolved in an acid would show anything ap- 
proaching to so great a deficiency in the heat evolved below 
that evolved when the metallic constituents are separately 
dissolved,* and their solutions mixed ; but the experiment 
* Will you try this experiment ? You would easily make a good thing of it. 
