RESEARCHES IN MAGNETISM. 
563 
(3) Roughly speaking, the intensity of magnetisation at which this maximum in 
the ratio appears is in the region of what Wiedemann has called the 
“ Wendepunct,” or point at which the ratio of 3 to is a maximum (the 
point at which a tangent drawn through the origin meets the curve 
connecting 3 and .§.) 
(4) The value of the maximum in the ratio of residual to induced magnetism is 
from 0 - 84 to 0'93 in annealed iron, nearly as great in annealed, hardened, or 
tempered steel, but much less in hard-drawn iron, where its value is more 
like 0'6. 
(5) As the magnetisation is further increased the residual magnetism approaches 
saturation more rapidly than the induced, and consequently this ratio 
diminishes—slightly in soft iron, more in steel, and much more in hard- 
drawn iron. 
By induced magnetism, in the above statements, is meant the value of 3 which is 
induced when a magnetising force is applied gradually, and without mechanical 
disturbance, the piece under test being initially in a neutral state. 
By residual magnetism is meant the value of 3 which remains when, after being so 
applied, the magnetising, force is removed gradually, and without mechanical dis¬ 
turbance. 
No magnetising force must act on the piece except the assumed force <£>, and 
consequently the conditions of the experiment are realised only when we use either 
endless magnets or magnets of such length that the field produced by the ends is 
sensibly inoperative throughout the greater part of their length. These conditions 
were fulfilled in the experiments which have been described, as far as it seems 
practically possible to fulfil them. Any actual sample of iron must be only imperfectly 
homogeneous, and consequently its magnetisation, in a uniform field, cannot be 
uniform, apart from the action of the ends. Probably no form of specimen can be 
found which is more nearly homogeneous, and at the same time uniform in section, 
than the form chosen in these experiments—a drawn wire with a diameter of about 
one millimetre. 
At the same time, the magnetisation of these wires, though perhaps as nearly 
uniform in the central part of their length as is practicable in any experiment, could 
not have been absolutely uniform, nor were they quite free from self-demagnetising 
forces when the field due to the solenoid was removed. When we consider that 
under the necessarily imperfect conditions of the experiments as much as 90 and 
even 93 per cent, of the magnetisation of soft iron was retained on the removal of 
the solenoid’s magnetising force, it seems far from unlikely that soft iron, if tested 
under ideally perfect conditions, would be found to have perfect retentiveness, that is 
to say, that none of the induced magnetisation would disappear on the gradual 
removal of all inducing force. This view does not admit of experimental verification 
or disproval: the experiments show that it is at least not untenable. 
