178 MR. A. NORMAN SHAW: A DETERMINATION OF THE ELECTROMOTIVE 
interchanged. The following mean values for the two cylinders were obtained and 
repeated without a variation of more than three in the last figure :— 
Mass of cylinder 1.196’1797 gr. 
„ „ II.196-1711 „ 
392-3508 „ 
(f) Further Remarks on the Accuracy of the Method. 
An examination of the possible effects of small asymmetries appears to show 
conclusively that no error as large as 1 part in 100,000 could occur in these measure¬ 
ments for the determination of B. Let us consider— 
(1) The effect produced by a small deviation in the position of the tube and 
cylinders as a whole, which might occur initially or vary during readjustment. (It 
must be noted, however, that this is quite a different matter from a variation of the 
cylinder with respect to the tube. In the final observations the tube was never 
readjusted between determinations for a “ Tj” and a “T 2 ” which were to be taken 
together.) 
(2) The effect produced by the deviation of the centre of mass of the two cylinders, 
in either the internal or external position, from the desired position on the axis of 
rotation of the suspended system. This effect might be produced by the effect 
mentioned in (l) and also by a variation of the position of the cylinder with respect 
to the tube. 
(3) The effect produced by the slight difference in weight. 
Suppose (a) that the centre of the tube is at a small distance x from the axis of 
the rotation; ( b ) that the centre of mass of the cylinders, at the external position, is 
at a small distance y x from the centre of the tube , and, at the internal position, at a 
small distance y 2 ; (c) that m 1 —m 2 is not zero, where m 1 is the weight of one cylinder 
and m 2 that of the other. 
Let 
x + y x — z x and x + y 2 = z 2 , 
then it can be shown that 
K.-K, = M + + (zA-zA), 
provided that we take z as lying in the direction of the axis of the tube. The 
possible displacements in other directions would be far less. Now d 1 and d 2 , involving 
y 1 and y 2 , could be reset to within a few thousandths of a millimetre; the centre of 
mass of the two cylinders could be reset to within a few hundredths of a millimetre; 
and m l —m 2 was 0’0086 gr. Substitution in the above formula shows that E^—K 2 
