FORCE OF THE WESTON NORMAL CELL IN SEMI-ABSOLUTE VOLTS. 
167 
It will be seen from the table below that the means of the groups AFIN, CHKP, 
BEJM, DGLO, and also ABEF, IJMN, KLOP, CDGH have an average deviation 
from 8321'0 of less than 4 parts in 100,000. 
Table VI. 
AFIN 
give 
8320-8 
ABEF 
give 8320-6 
CHEF 
8321-1 
IJMN 
„ 8321-6 
BEJM 
8321-4 
KLOP 
„ 8321-2 
DGLO 
J) 
8320-8 
CDGH 
„ 8320*8 
1 
We have for the mean value of our ratio with the four windings of the small coil in 
parallel 
41a _ 100000-8321-04 
l, ~ 832P04 
11*01773, 
and hence I 2 /li = 2‘75443. 
It is important to record that Mr. R. O. King also made an accurate determination 
of this ratio, but unfortunately the notes containing his observations on “ d ” have 
been lost. However, assuming that the average d in his case was the same as that 
found for the dynamometer as it was set up when the writer first examined it in 1910, 
we calculated from his notes the mean value, 2L/I! = 5*5087 where one suspended coil 
is in parallel with the other, and hence 
yij = 2*75435. 
This figure was the mean for the eight positions which we have called I, J, K, L, 
M, N, O, P. 
The value of L/Ij is independent of the temperature of the dynamometer.* Hence 
if the ratio is expressed for the value 
d t = d 20 { 1W(£—20)}, 
where d t is the length of d at the temperature of observation t, and “ a' ” is the 
coefficient of expansion, we can use d 20 , x 20 , S 20 , and a 20 in our general formula instead 
of d t , x t , S t , and a t . 
* The current through the fixed coils produces no perceptible heating (tested by resistance measure¬ 
ments), and that through the suspended coils during the magnetometer readings is the same as in the 
deflection experiments, hence any error due to the extra heating of the suspended coils is eliminated in 
the final results. 
