622 
PROFESSOR J. A. EWING ON EXPERIMENTAL 
gives augmentation of magnetism if this value is exceeded. Sir William Thomson 
calls the value of the magnetising force at which this reversal of effect occurs, the 
“ Villari critical value.” It is clear (as he has also pointed out) that the “ Vlllari 
critical value ” depends on the particular amount of load alternately applied and 
removed, and also upon the mean value of the load. 
An examination of my results shows that the effects of stress on magnetism are, 
on the whole, not greatly different whether the magnetism in question is residual or 
induced, and I am therefore disposed to regard the Villari reversal as depending 
rather on the value of the magnetisation than on the value of the magnetising force. 
This view receives strong confirmation from experiments which I have described 
in a paper, “ On the production of transient electric currents in iron and steel 
conductors by twisting them when magnetised, &c.” # I have shown there that 
the inclined pull and push which make up torsional stress give effects on a veiy 
strongly magnetised wire which have the same sign as the effects of direct pull and 
push on a feebly magnetised wire; which is to be explained by the fact that these 
act on 3 cos 3 q; or only one-half of the whole magnetism, an amount which does not 
exceed the “ Villari critical value ” of 3, although one-half of <§ may be made to 
exceed very largely the value of the magnetising force sufficient to cause the Villari 
reversal when pull is applied in the direction of magnetisation. 
I shall therefore apply the term “ Villari critical point” to name that value of 
the magnetisation 3 at which reversal occurs in the sign of the magnetic difference 
produced by two (assigned) states of longitudinal stress. 
This point may be determined in two ways, which (on account of hysteresis) will 
lead to different results. We may (with Thomson) cause a repeated alternation to 
take place from one to the other of the two assigned states of stress, and find the 
value of 3 at which that alternation causes no magnetic change, or we may take 
two separate curves of magnetisation (that is, curves of the relation to 3 to «§) in 
the two assigned states of stress, and see where these curves cross each other. The 
value of 3 at the crossing point is the critical value for the two given states of stress. 
Several of the figures which have been given show values of the “ Villari critical 
point ” according to this last mode of definition. 
We may, however, conveniently limit the term “Villari critical point” to the 
case where one of the two assigned loads is zero; that is to say, we may use the term 
only when we are comparing the magnetic condition reached under a given load with 
that reached under no load. With this restriction the Villari critical value of 3 i § 
(for any load) that value at which the curve of magnetisation under that load crosses 
the normal or no-load curve. Thus, in fig. 47 (§ 94) the curves for 12, 14, 16'2, and 
18'5 kilos, all cross the normal curve within the range of the observations, and so 
determine the Villari critical points (as just defined) for these particular loads. 
* Proc. Roy. Soc., Yol. 36 (1883), p. 117. 
