Affinities of the Elements of Water. 303 
For the same reason the coils of wire which are used to give resistances beyond 
the range of the rheostat, are arranged in a glass jar filled with a mixture of 
alcohol and water, the latter alone being rather too good a conductor. This may 
seem a needless refinement; but I found a set of coils which gave the resistances 
1091 in air, sink to 1020 when thus managed, though the deflection was only 45°. 
By the same means the constant 7 is obtained: calling the resistance of the 
dividing wire w, and adding to r the resistance a before dividing the current, we 
have the two equations, 
pEXews A inka) Dawe 
i aan ae he 
from which 
SS SS Ee 3 
p—p—a 
and using several values of a we get a set of equations of condition from which it 
is found exactly. In these experiments it is 7.3. 
The apparatus which I used was constructed for other researches, requiring 
very powerful currents, and therefore the rheometer is not very delicate. Astatic 
needles would soon lose their quality in it, and numerous convolutions of fine wire 
introduce too much resistance. It consists of a rectangle of copper bar one- 
fourth inch square, six inches by three, with a needle three inches long supported 
in its centre on a point of steel and ruby cap. The circle is divided only to 
degrees ; but I estimate tenths of these pretty sure. I have said that I assume 
as the unit of current that which deflects this instrument to 45°; I may add, that 
this is able to afford, in five minutes, by the decomposition of water containing 
one-eighth of sulphuric acid, with electrodes of bright platinum, 6.57 cubic inches 
of gases at 60° thermometer, and 30°.0 barometer. 
The rheostat is analogous to the second form of that most valuable instrument 
described by Mr. Wheatstone ; its copper wire is one-twelfth diameter, and a 
revolution 8.3 inches. Though the moveable piece which connects its spiral 
with the circuit is kept in contact by a rather considerable weight, I find that its 
action is not perfectly uniform, and this, with a suspicion that even such thick 
wire may not be entirely secure from the effect of heat, has induced me to have 
one made ona different principle. In it the varying resistance is given by a pla- 
tina wire, whose effective length is varied by raising it out of mercury ; its 
VOL. XXI. 25 
