treated according to Van der Waalss Equation. 347 



(7) Weiss considers the critical temperature to be the tem- 

 perature at which the straight line of equation (3) is tangent 



to the curve m = cotb x at the origin, and hence the 



intensity ought to fall to zero at the critical temperature. 

 This is only true for residual magnetism when there is no 

 external field in action. But if there is an external field in 

 action the intensity does not, in fact, become zero at the 

 critical point, but diminishes asymptotically, and apparently 

 an infinitely high temperature would be required to annul 

 the intensity altogether. Weiss meets this latter case by 

 supposing that at some very low value of I, H becomes com- 

 parable with NI and is no longer negligible in the equation 



MH + NI) 

 x- RT , 



and then the theoretical relation of I to T above the critical 

 temperature has the appropriate hyperbolic shape. The 

 theory in its present form, however, does not lead to a 

 definite point where the intensity and field have critical 

 values, and the temperature which has been treated as the 

 critical temperature is really the temperature of disappear- 

 ance of magnetism when the field is zero. 



III. 



1. The relation of magnetic intensity to field-strength has 

 been treated so far in a general manner and without the 

 introduction of specific values. It remains to give numerical 

 values to the constants and to compare the results with 

 experimental facts. The expressions for the critical constants 

 from the ferromagnetic equation are 



rp _ _£_ a Ip . T _ 1-r . tt _ JL n n 2 



x c — iy- d| , x c — 'y x 0f ^c — £)n u ' I • 



From the equation to the critical temperature a value for a 

 has been derived *, and by using this value the intrinsic field 

 and the critical field can be calculated. Accepting the values 

 of a derived from the critical temperature, we find 



H,=0'82 x 10 6 for iron, and H c = 0'8y x 10 6 for nickel. 



If, however, curves of I = </>(H) are consulted the critical 



fields for these metals appear to be less than unity. In the 



same way calculated values for the field for instability and 



* Ashworth, Phil. Mag. vol. xxx. p. 723 (Nov. 1915). 



2 A 2 



