VALUE OF THE BRITISH ASSOCIATION UNIT OF RESISTANCE. 
259 
Hence 
P_ . u—ft+x—y 
Q~ + 5~ 
_ 1 oL—fi—x' + y' 
-) and such terms. 
neglecting 
Now the values of the 5 units being known, we know a and ft, and we find at the 
temperature of the observation 
a — 3 
—'0061 
Hence 
P-Q=Qj- 0061 +^| 
P-Q=q{^-' 006 i} 
The values of x, y, &c., actually obtained were such as fully to justify the neglect 
of"j^fj and such terms. By this means the values of the 500, 1000, 1000j and 2000- 
unit coils in the Elliott box were determined. 
Having thus determined the values of the resistances of the box, they were used to 
determine that of If in the following manner. 
B was connected with one end of the wire of the bridge, and the two boxes in 
multiple arc with the other, two 30-unit coils forming as before the third and fourth 
arms. 
170 units were then taken out of the Warden box, and the Elliott box adjusted 
until the difference between It and the total resistance of the compound circuit formed 
by the two boxes could be measured in terms of the bridge-wire. This was the case 
when 2920 units were out of the box. The actual value of these resistances at 12° is 
known from our table of resistances of the box. Making the correction for temperature 
and for the difference between It and the multiple arc resistance, we find 
11^= 160*602 B.A. units. 
A second determination, in which 180 units were out of one box and 14.90 in the 
other, gave 
ft 0 = 100*570 B.A. units. 
Each of these results is the mean of several experiments. 
We may take, therefore, as the value of K 0 the mean 
B 0 = 160*586 B.A. units. 
The value of S, the resistance of the circuit through which the fractional part of the 
2 l 2 
