22 Mr. H. Tomlinson on the Sudden Loss of the 



660° C, 680° 0., and 725° C. The author has himself ex- 

 amined the thermal expansion of iron at different tempera- 

 tures right up to and a little beyond that at which the metal 

 showed very conspicuously sudden loss of magnetic properties 

 without at the same time detecting anything like a sudden 

 change in its thermal expansibility. 



Evidently the critical points A and B, though occurring 

 at a much lower temperature in nickel than in iron, correspond 

 to similar molecular changes in the two metals, but the third 

 critical point C in iron has apparently no corresponding point 

 in nickel. There are at any rate no such sudden changes in 

 nickel when under stress and strain as occur in iron. 



It is, however, the critical point B in nickel and in iron 

 which concerns the present inquiry. Why does the temporary 

 magnetic susceptibility of both these metals diminish so 

 rapidly that, if an iron or nickel wire be surrounded by a 

 magnetizing solenoid always kept in action, and this again 

 be concentric with a secondary solenoid connected with a 

 galvanometer, a very sensible induced current can be observed 

 at the critical temperature both on heating and cooling*? 

 This question is a very important one, and if we can answer it 

 correctly much light will doubtless be thrown upon the nature 

 of magnetism. 



According to the author's view of the matter each molecule 

 of iron, say, contains within itself several magnetic molecules 

 (it may contain several hundreds or thousands). 



Let the circle in fig. 1 represent a molecule of iron, and 



Fig. 1. 



let mi, m 2 , and m 3 be the centres of the axes of three 

 magnetic molecules. Each of these magnetic molecules, like 



* Even when the wire is only under the influence of the vertical com- 

 ponent of the earth's magnetic force, the currents induced when the 

 magnetic properties are lost or restored are quite sensible, 



