292 Proceedings of Eoyal Society of Edinburgh. [sess. 
in the case of the iron and steel, the apparent average permeabilities 
being equal, — above 16 in a field of 585 C. G. S. units. 
In low fields, the results obtained with the nickel tubes do not 
quite agree with those obtained with the iron and steel. There are 
indications, however, that somewhat similar conditions are fulfilled, 
as the following numbers may serve to show ; — 
Induction in Low Fields. 
Field. 
Tube I. 
Tube II. 
Tube III. 
Tube lY. 
4 
85 
113*6 
103*6 
7 
179 
234 
229 
207 
22 
850 
935 
1050 
1043 
30 
1210 
1340 
1490 
1486 
585 
4600 
6000 
6900 
6920 
Thus, in sufficiently low fields, Ho. III. has a distinctly greater 
induction than Ho. lY., and Ho. II. than Ho. III. ; hut there is no 
indication of Ho. I. similarly exceeding Ho. II. The tendency 
seems to he in the other direction ; for the induction in Ho. I. is 
proportionately smaller than that in Ho. II., the lower the field. 
There are several reasons why the explanatioii mentioned above 
should fail to apply to the case of the nickel. 
The mathematical theory of magnetic induction is based upon 
certain assumptions which are approximately realised when iron 
and steel are subjected to small magnetising forces. Thus, in iron 
tubes, the residual magnetism is barely appreciable in low fields, 
and the magnetisation is proportional to the magnetising force. 
Hence the relation given by theory is fairly applicable, namely. 
1 = 
Ic 
l+hP 
H, 
where I is the magnetisation, H the field, 7c the susceptibility, and P 
the “ demagnetising factor,” as Du Bois terms it — a quantity whose 
value depends on the form of the body. Its value can be calcu- 
