139 
should be made. Without any such experiment however 
we may safely say that the fourfold amount of energy indi- 
cated by the preceding formula, for a value of D yet twice 
as small, is very much greater than any estimate which our 
present knowledge allows us to accept for the heat of com- 
bination of zinc and copper. For something much less than 
the thermal equivalent of that amount of energy would melt 
the zinc and copper; and therefore if in combining they 
generated by their mutual attraction any such amount of 
energy, a mixture of zinc and copper filings would rush into 
combination (as the ingredients of gunpowder do) on being 
heated enough in any small part of the whole mass to melt 
together there. Hence we may infer that the electric 
attraction between metallically-connected plates of zinc and 
copper of only 40070- 0075-00 of a centimetre thickness, at a 
distance of only ro-07000700-0 of a centimetre asunder must 
be greatly less than that calculated from the magnitude of 
the force and the law of its variation observed for plates of 
measurable thickness, at measurable distances asunder. 
In other words, plates of zinc and copper so thin as a four- 
hundred millionth of a centimetre, and placed at as short a 
distance as a four-hundred-millionth of a centimetre from 
one another, form a mixture closely approaching to a mole- 
cular combination, if indeed plates so thin could be made 
without splitting atoms. 
Wishing to avoid complication, I have avoided hitherto 
noticing one important question as to the energy concerned 
in the electric attraction of metallically connected plates of 
zinc and copper. Is there not a change of temperature in 
molecularly thin strata of the two metals adjoining to the 
opposed surfaces, when they are allowed to approach one 
another, analogous to the heat produced by the condensation 
of a gas, the changes of temperature produced by the appli- 
cation of stresses to elastic solids which you have investi- 
gated experimentally, and the cooling effect I have proved 
