262 
MESSRS. R. T. GLAZEBROOK AND E, B, SARGANT ON THE 
The box was used as a shunt to the 20-unit arm, and adjusted as before till the 
galvanometer showed no deflection, the effective resistance of this arm was found to 
be 19'0444 units, while at the same temperature the value of G is '99887 unit. 
Substituting the value of R at the temperature of the observation and reducing our 
result to the temperature 12°, we find 
S 0 =3059'37 units 
A second series of observations on a different occasion gave 
S 0 =3059-86 B.A. units 
The mean is 
S 0 =3059-62 B.A. units 
and this is the value we have used in our calculations.* 
The coils were in all cases placed in water baths and allowed to stay for some hours 
in them. Before making the observations the temperature was read by thermometers 
graduated to fifths of a degree, which were compared with each other. 
To determine then the value of the B.A. unit we must use the value of R u = 160'520, 
found in this series of measurements. Now the values of S 0 and R 0 found from the 
boxes were respectively 
3060'80 and 100’5S6 
These differ from the values we have obtained in our last observations by l 1 18 and 
•066 unit respectively, or rather more than 1 in 3000. This difference would 
correspond to an error of about 1° in the measurement of the temperature. 
Considering then the uncertainty which must attach to the temperature of the coils 
inside the box, it seemed best to take our last values rather than the mean of the two 
—we would rather regard the first series as a check upon any large error. But though 
this difference in the value of S will affect to the amount of 1 in 3000 the value of 
R 0 in absolute units, it does not affect at all sensibly the value of the B.A. unit, for 
this latter depends on the ratio of S 0 /R 0 ; taking the values of S 0 and R 0 from the 
boxes, we have 
194)602 
while, if we use the last values obtained for S 0 and R 0 , we get 
^=19-0607 
l«o 
the difference is only about 1 in 40,000 and does not concern us. We shall therefore 
put R 0 =160’520 ohms, the temperature being 12°. 
* The value 3059'89 used on page 255 is obtained from this by applying a temperatui’e correction to the 
60 units, the resistance of the galvanometer, which was at 13°'5. 
