552 
Proceedings of the Royal Society 
When € is small, 
To find F^, which is the deviating force on the magnetometer — 
Let ABCDA'B'C'D' be a horizontal projection of the frame; and 
let the centre of the magnetometer be excentrically placed at the 
point (Sa, 8c). 
By supposing the frame to be divided into two parts by a vertical 
plane, c?<;dc'( where Ad = 2Sa = A'd'), and imagining pairs of equal and 
opposite currents to flow in this plane up and down, we may resolve 
the circuit into two pairs of coils, dcCD, d'cCT>', and ABcd, A'B'cd'. 
Fig. 5. 
The former pair, being symmetrically placed with respect to the 
assumed position of the magnetometer, will give no deviating force ; 
and in finding F^ we have only to consider the remaining pair, 
namely, ABc(i, A'B'c'tl'. Again, suppose two vertical planes, EF 
and E'F', to be drawn parallel to the planes of the coils, making 
EB = 8c* = E'B'. In this way the remaining pair is again divided 
into two pairs, AEFc/, A' E'F'<^', and EBcF, E'B c'F' ; of these 
the former, being placed symmetrically with regard to the mag- 
netometer, will give no deviating force. 
We are thus left with a pair of narrow magnetic shells EBcF and 
E'B'c'F', the breadth of each shell being 28a, height 25, and thickness 
* [Evidently this should be 25c, not 5c. The mistake, which I have noticed 
only in reading the proof, does not affect the accuracy of Mr Tanakadate’s con- 
clusions as to the proper proportions of the coils ; and the equations which 
follow, as well as the numerical values given in fig. 8, need not be altered, if we 
assume the excen tricity of the magnetometer in the direction NS to be |5c 
instead of 5c as in the text. — J, A. E.]. 
