RESEARCHES IN MAGNETISM. 
609 
The presence of stress in an annealed wire tends to round off the outlines of the 
curve of magnetisation, so that it resembles somewhat the curve described in § 33 
as characteristic of a wire which has been stretched beyond its limit of elasticity, and 
this happens although the stress is too small to give the wire any permanent set, or to 
harden it appreciably. 
§ 92. Plate 64, fig. 45 represents the results of the same experiment in a different 
manner. In it the values of magnetisation reached by applying a particular mag¬ 
netising force while the load was kept constant, are plotted in relation to the constant 
loads (2, 4, and 6 kilos.) which were kept on during the several magnetisations, and a 
curve is drawn through the points so obtained. This gives a series of curves, each 
referring to some particular value of the magnetising force, which is entered in the 
figure alongside of the curve. The curves may be said to represent the relation which 
the susceptibility to magnetisation (by the several assigned magnetising forces) bears to 
the stress upon the wire. They are closely analogous to curves of the type of figs. 31 
to 43, but with this important distinction, that in the present case the load was on 
before the magnetising field was applied, while in them the magnetising field was 
applied first and then the load. Were there no hysteresis in the changes of magnet¬ 
isation the curves of fig. 45 should agree exactly with curves obtained by applying 
to an annealed wire a constant magnetising force first and then varying the stress. 
But we know that there is hysteresis in the relation of magnetisation to magnetising 
force, under constant load. We also know that there is hysteresis in the relation of 
magnetisation to stress, when stress is changed in a constant field. There is, there¬ 
fore, a double reason why the curves obtained in the way now exemplified should 
differ from those of fig. 43, where, as here, the iron was dealt with in the soft annealed 
state. 
Fig. 45 shows that at low magnetisations the susceptibility is increased by the 
presence of longitudinal pull. At higher magnetisations we find a maximum of 
susceptibility, which occurs with lower and lower values of the stress the stronger 
the magnetisation is, until finally, when we approach the region of saturation this 
maximum disappears, as it were, to the left of the figure, and we then find the sus¬ 
ceptibility greater without than with the load. 
This is quite in agreement with the results already obtained by the other method, 
—the method, namely, of varying stress in a constant field (see figs. 36-38, and 43, 
§§76-82, 89). 
§ 93. Stretched Iron Wire. —The effects of the presence of stress on magnetic sus¬ 
ceptibility become much more conspicuous when we are dealing with wire which has 
been stretched beyond its elastic limit. 
The wire of the above experiment was stretched by applying a load of 15^ kilos., 
which brought it to the state already mentioned in § 87. In addition to the experi¬ 
ments of the old type described there, observations were made upon it (May 4, 1882) 
by the plan of magnetising under constant load. The loads used were 0, 1, 2, 4, 6, 
