724 REPORT — 1885. 
‘electrolytes are either simple binary compounds, or are capable of being 
formed from them by double decomposition, the ions being the radicles so. 
exchanged’ (Wiedemann); &c. Such statements may be true, and if 
trne may be important chemical generalisations, but they are not 
physical definitions or tests of electrolytic conductivity. 
As a physical definition of an electrolyte we have only to say, an 
electrolyte is a substance which conducts electrolytically. 
A statement like this only helps by fixing our attention on the real 
point in question, viz. the difference between electrolytic and metallic 
conduction—a true physical distinction, which is capable of definite 
examination, and may be capable of precise statement. 
Whether we are able thus sharply to divide off electrolytes from other 
conductors will depend on whether any substance be found which can 
conduct both like a metal and like an electrolyte. Such a substance is 
not yet, I believe, certainly known, though it has been often suspected ; 
it has even been suspected that all electrolytes have a trace of metallic 
conductivity. Though for the present this is quite unproved, and may 
be regarded as against the weight of evidence, it would be rash to insert 
the word wholly into the above definition; it must be left as meaning 
that in so far as a body conducts electrolytically, so far it is an electro- 
lyte. 
‘A Now electrolytic conduction differs from metallic conduction in 
several ways. Metallic conduction appears to be a true passage of 
electricity through matter; it is unaccompanied by any reversible or 
chemical process, it simply generates heat. 
Electrolytic conduction is accompanied by certain reversible chemical 
processes, and appears to be of the nature of a convection of electricity 
by the atoms of matter. 
A substance which so conducts is an electrolyte, and whether it be a 
good or a bad conductor is foreign to the inquiry. 
It is important to notice this because it is often sought to deny that 
(say) water is an electrolyte by showing that it is a bad conductor. Of 
course if it does not conduct at all it is no electrolyte, but a dielectric. 
Our definition says first that an electrolyte must conduct, and it then 
proceeds to say how it must conduct. 
Water, alcohol, turpentine, glass, either conduct a little or they do 
not. If they do not, they are dielectrics pure and simple; but experiments 
on leaks have always shown a more rapid leak if water, alcohol, or tur- 
pentine be used to replace air; and as to some kinds of glass, it seems to 
be a mere question of temperature whether they conduct or not. If it be 
proved that any of these things, even when pure, conduct, be it ever so 
badly, they are not simple dielectrics, and it remains only to consider 
whether their conduction is metallic or electrolytic. 
But it may happen that some of these bodies, or perhaps all liquids, 
behave as dielectrics to rapidly intermittent or alternating E.M.F.s, but 
as electrolytes to slow and long-continued E.M.F.; just as pitch is elastic 
to rapid vibrations (it can transmit sound for instance), but is viscous 
and essentially fluid to long-continned forces. Whether this is so or not 
is a vital question in relation to the electromagnetic theory of light, 
wherein the transparency of conducting liquids has only been provisionally 
and tentatively explained. 
