CLARK CELL AS A STANDARD OF ELECTROMOTIVE FORCE. 
577 
were repeated on the following day, the bowls having usually been left for the interval 
in the balance case. In some cases during the interval the bowls were again heated. 
In none of the observations was there any difference sufficient to affect the result 
between the weighings. 
Temperature corrections had to be introduced into the comparison between the 
bottle cell and the standard, for these could not conveniently be put into the same 
bath, and consequently differed slightly in temperature. 
This was done by the aid of Lord "Rayleigh's value for the coefficient, viz., ’00078 
—a value which our own experiments (see p. 615) sufficiently confirmed—in the 
following way. Since the plugs required out of the Box I. to balance the Clark were 
about 5800 ohms, a change of ’00078 in the E.M.F. of the Clark will mean a 
change in the resistance of 5800 X ’00078, or about 4’5 ohms in the box. Thus an 
increase of 1° in the temperature of the cell means a fall of 4’5 ohms in the resistance 
required to balance it. The actual change in E.M.F. corresponding to this will 
be 1’43 X ’00078, or ’00112 volt, and the change corresponding to one ohm of 
the box is found by dividing this by 4’5. This gives ’00025 volt. 
The temperature of the bottle cell and of the Leclanches varied slightly during 
the progress of each experiment, and part of the variations observed in the ratios of 
the two are no doubt due to temperature changes. 
§11. Details of Experiments. 
We proceed now to the details of each experiment. Most of these can be best 
given in tabular form, and this is done in Table I. (p. 581). 
Some explanation is needed, however, of the method by which the values of some 
of the tabulated numbers, specially those in columns 3, 6, 10, and 15, are arrived at, 
and of the notation employed. 
Let V be the resistance out of the box required to balance the difference of 
potential Bf between the ends of the strip coil, W that required to balance the 
Clark, E the E.M.F. of the Clark which was compared with Bi at the temperature of 
the observation. 
B is, as above, the resistance of the strip, i the current. Let M be the mass of 
silver deposited, T the time the current has passed, and y the electrochemical equi¬ 
valent of silver in grammes per ampere per second. 
Then 
y = ’001118. 
Then we have 
M = fyT, 
therefore, 
E/Bi = W/V, 
E = B. 
W 
V 
4 E 
M 
7 • T ’ 
MDCCCXCII.—A 
