CLARK CELL AS A STANDARD OF ELECTROMOTIVE FORCE. 
575 
Let V be the resistance taken out of the Box I. when the ends of the resistance Pt 
are connected to L and M. Then as the current varies Y varies also, being always 
proportional to i. During the time the current was passing a series of observations 
of the value of Y, regulated to produce a balance, was made and the corresponding 
times noted. These observations were plotted on squared paper, and a curve formed 
with the times for abscissae and the values of Y for ordinates; the area of this curve 
is proportional to the quantity of electricity which has passed, and it is easy to 
calculate from the curve the mean value of the current and the corresponding value 
of Y. This method was followed in the observations on July 31, August 5, and 
August 7. 
On July 31 the value of Y changed considerably, viz., from 3876 to 3917'5. With 
one exception, however, the observations lie on a very regular curve, and the mean 
value, 390P5, got from the curve is probably accurate to 1 or 2 ohms. 
On August 5 the changes were very small, and, as shown by the corresponding 
curve for the Clark cells, are almost entirely due to alterations in the Leclanche cells. 
The curves are given in Plate 16, tigs. 1, 2, 3. 
In all the other experiments an adjustable mercury resistance was introduced into 
the main circuit with the voltameters. The plugs in the boxes were adjusted until 
on making the galvanometer key there was no deflection. The galvanometer circuit 
was then kept closed, and the mercury resistance adjusted so that no deflection should 
take place. Occasionally it happened that the changes needed in the resistance of 
the main circuit to maintain i constant were outside the limits of the mercury 
rheostat. When this was the case a suitable change was made in the plugs in I. 
and II. and in the rheostat simultaneously, and the time noted. The change needed, 
except on the morning of August 12, was always small, some 4 or 5 ohms at most, 
and the mean value of Y was readily determined with all the accuracy required. On 
August 12, after the current had been flowing steadily for 25 minutes, a sudden drop of 
over 1 per cent, took place. As during the next five minutes the current continued 
somewhat unsteady, the experiment was concluded at the end of 30 minutes 
45 seconds, a much shorter time than in any of the other experiments. 
In this second method of making the observations the very slight heating effect 
produced in the strip coil when the bath was not stirred was just observable. A 
balance was obtained, the bath being unstirred for a few moments. On moving the 
stirrer the spot was very slightly deflected in the direction indicating a fall in the 
resistance R, and the rheostat required a slight readjustment to restore the balance. 
The readjustment necessary, however, was extremely slight, being equivalent to a 
change of only a small fraction of an ohm in the Box I. As 1 ohm in the box means 
an alteration of ‘00024 volt in the E.M.F., the error produced by heating in the strip 
was quite negligible. 
