212 Proceedings of the Royal Society of Edinburgh. [Sess. 
gives a decrease of resistance of 65 in 10,000, produces when superposed 
cyclically upon the steadily acting longitudinal field a much greater diminu- 
tion of resistance, namely, 124. As a glance through the table will show, the 
combination T(H) is of the same sign as T, and is, with the single exception 
of I. 1, always greater, in most cases much greater, than the corresponding 
T. There is nothing here at all in harmony with the view first suggested 
as to the relations of H and H(T). In the relations of T and T(H), the steadily 
maintained longitudinal field, which by itself invariably decreases the 
conductance, becomes when associated with the cyclically applied transverse 
field the cause of a greater increase in the conductance than is obtained by 
the transverse field acting alone. Instead of the steady longitudinal field 
being a restraint upon the effect of the superposed transverse field, it makes 
the nickel, as regards its change of resistance, even more sensitive to the 
influence of the transverse field. 
The effect of the longitudinal field acting alone on the resistance 
10,000 is to change it to 10,000 + H; and the effect of the transverse field 
acting alone is to change 10,000 to 10,000 + T, where in these experiments 
T is always negative. 
Now T(H) is the result of the transverse field superposed on the 
condition H. Hence the original resistance 10,000 becomes, under the 
combined fields, 
(10,000 + H) 10 ’°| ) o 0 + - ^ S 
= I 0 - 000 + H + T(H) + hw- ) - 
On account of the comparative smallness of H and T(H) compared to 
10,000, the last term may be neglected. 
Similarly when the longitudinal field is superposed on the transverse 
field, producing a change of resistance H(T) per 10,000, the original 
resistance 10,000 becomes 
10,000 + T + H(T) + T x H(T ) , 
v ' 10,000 ’ 
where as in the previous case the last term is negligible. 
Thus the total changes per 10,000 of resistance of the nickel under the 
combined fields are given by the sums 
H + T(H) and T + H(T). 
If the effects of the combined fields were simply superposable, these sums 
should be equal to H + T. A brief inspection of the table will show that, 
with the single exception of the case I. 1, the sums H + T(H) and T + H(T) 
