voz REPORT—1885. 
out properly connected with the existence of liquid and solid contact, or 
surface-action. 
But how does this theory explain migration? For the atoms in a 
molecule, if they electrify each other by contact, necessarily do so with 
equal quantities; and this is where the theory differs from Quincke’s,. 
Wiedemann explains the unequal rate of travel postulated by Hittorf, 
not by inequality in the propelling force, but (more satisfactorily as it 
seems to me) by a difference in the resistance met with. He would say 
that a hydrogen atom slips through the liquid more easily than an oxy- 
gen atom, and so gets along faster ; moreover, he has conjectured, and 
experimentally verified within certain limits, that the ease of travelling of 
a given ion is inversely as the ordinary viscosity of the liquid; so that 
conductivity and viscosity are inverse to one another. He has further 
established the important and convenient fact that migration and endos- 
mose are totally distinct things, having apparently no sort of relation to 
one another. 
Wiedemann’s theory thus chimes in beautifully with that of Kohlrausch, 
who postulates a specific velocity for every ion: a velocity which depends 
only on the nature of the liquid in which it has to travel, and the 
dV which drives it. 
dx 
That part of his theory which asserts a connection between conduc- 
tivity and limpidity is curiously illustrated by the behaviour of water (or 
saline or acid solutions) at different temperatures flowing through a 
capillary tube. It is known that hot water flows some five times as fast 
as cold water through a given tube. It is also known that hot water 
conducts much better than cold, Further, it is known that if terminals 
connected to an electrometer are immersed in the liquid at either end of 
such a tube, an E.M.F. is discovered between them, depending on the 
rate of flow. 
Now J. W. Clark has discovered that, notwithstanding its vastly 
greater rate of flow, the E.M.F. of hot water is almost identical with 
that of cold. His untimely death has prevented his publishing this 
result, but when he told it and showed it me, some year or so ago, I con- 
jectured that the observed E.M.F. must be a sort of residue of the whole 
generated E.M.F., being the part unable to leak back through the liquid ; 
and I accordingly hunted up data to see whether the extra leakiness or 
conductivity of hot water might so nearly neutralise the H.M.F. gene- 
rated by its more rapid flow as to give the same residue of E.M.F. 
Whether this be the true account of the matter or no, the fact is so. 
The empirical formula in Naumann giving the conductivity of the liquor 
used in Clark’s experiment (it was salt and water I think), at different 
temperatures, contains practically the same co-efficients as that giving 
the viscosity of the same liquid, measured by its flow throngh a capillary 
tube. 
The observation thus agrees perfectly with the theory of Wiedemann, 
connecting electrolytic conductivity with mechanical limpidity. 
The decrease in liquid viscosity with increase of temperature is re- 
markable. On any kinetic theory of viscosity, as due to diffusion one 
would have expected it necessarily to increase with rising temperature. 
In a gas it does so, as is well known, but in a liquid it does just the 
reverse. Iam unable to suggest any reason for this. 
