184 MR. A. NORMAN SHAW: A DETERMINATION OF THE ELECTROMOTIVE 
eight and nine volts. b 2 represents a large main battery of 100 volts. By proper 
adjustment of r lt it was possible to supply a current of between five and sis amperes 
for several hours with a constancy of better than 1 part in 100,000. It was 
arranged that b 2 supplied the larger part of the current, while b x served as a means 
of maintaining a constant voltage. 
While taking deflection measurements r 4 and r 5 were kept adjusted so that there 
was only a small deflection in the galvanometer G P A change of 1 part in 
100,000 was represented by seven scale divisions; r 2 and r 3 were kept adjusted 
so that the deflection in G 2 would be small, and this deflection could be eliminated 
and interpreted by the proper choice of the point H (see below). 
It was arranged that there was a drop of a millionth of a volt per millimetre along 
EF ; one scale division of G 2 corresponded to one millimetre of the wire. 
(d) The Potentiometer Correction. 
A complete record of the deflection measurements for this investigation is given in 
Table XVIII. The electrical connections are explained in detail above, and it will be 
seen that the current which passed through the suspended coils also passed through a 
standard resistance in an oil bath. It was arranged that the drop in potential across 
this differed from the electromotive force of a normal Weston cell at 25° C. by only a 
small number of microvolts. When opposing these two potential differences and 
measuring their difference on an ordinary potentiometer with a sensitive galvano¬ 
meter in the circuit, it was possible to adjust the current to a constant value. It was 
not, however, necessary always to adjust the current in order to obtain a given 
reading, for if the difference on the potentiometer was read, the value of the current 
could be easily expressed in terms of that current which would produce a zero 
difference between the E.M.F. of the standard cell and the drop in potential across 
the resistance. Observations of the deflection were taken simultaneously with a 
potentiometer reading which never exceeded 700 microvolts and averaged about 
250 microvolts. As this variation is so small and was taken to within five microvolts 
of the true balance, it was always possible to apply to the deflection a correction 
which could not be in error by as much as 0‘001 cm. on the scale. (For example if 
the drop in potential across the standard resistance was 100 microvolts greater than 
the electromotive force of the Weston cell, then the current would have to be cut 
down by only 1 part in 10,180 to get the value which would produce the same drop 
in potential as the cell. As the circuits connecting the fixed and suspended coils of 
the dynamometer were in parallel and as the deflection varies with the product of the 
two currents which are kept in a fixed ratio we see that the deflection must be 
reduced by 2 parts in 10,180 or, in our case, where the readings lie between 42'3 and 
42'4 cm. the correction would be — 0'0083. We could thus determine this more closely 
than we could read the deflection and were able to save considerable time and much 
