478 Prof. W. F. Barrett on the Relationship 



near low red beat, and remains positive till near the melting-point 

 of iron, where it appears possible from some of my experiments 

 that it may again change sign." And further on Professor Tait 

 suggests the idea "that iron becomes, as it were, a different 

 metal on being raised above a certain temperature. This may 

 possibly have some connexion with the ferricum and ferrosum of 

 the chemist, with the change of magnetic properties of iron, and 

 of its electric resistance at high temperatures." Professor Tait 

 adds the interesting fact that he has found an anomalous thermo- 

 electric behaviour in nickel similar to that in iron, and, as one 

 might venture to anticipate, at a much lower temperature. 



Thus two separate lines of inquiry have converged on the 

 same point — namely, that a profound molecular disturbance 

 takes place in iron at a low red heat. In connexion with future 

 theories of magnetism, this fact is likely to be of considerable 

 importance, inasmuch as it seems probable that this disturb- 

 ance is confined to the magnetic metals, and that it occurs at or 

 about the temperature when they leave or reenter this condition. 



LXIII. On the Relationship of the Magnetic Metals, 

 By W. F. Barrett, F.C.S* 



THE remarkable similarity in the chemical and physical pro- 

 perties of the magnetic metals has no doubt often attracted 

 attention ; but I am not aware that any definite collation of these 

 properties has ever been made. This I propose briefly to do 

 in the following paper. The extraordinary homology these 

 metals are thus seen to exhibit furnishes instructive evidence 

 concerning the molecular state of a magnet. 



By magnetic metals I mean those metals which possess 

 magnetic properties under ordinary circumstances— namely, 

 iron, nickel, and cobalt. 



First we will compare their physical characteristics. The 

 specific gravity of the thirty-eight known metals ranges from 

 lithium 0*59, to platinum 21*5, a difference of nearly 21; 

 whereas the specific gravity of iron is 7*8, nickel 8'3, and cobalt 

 8*5, an extreme difference of only 07. The specific heat of 

 these three metals is also nearly identical ; and their atomic heat 

 is the same. Their conductivity for sound is almost absolutely 

 the same ; and so far as their heat and electric conductivitv have 

 been determined they are also alike. Their dilatation by heat is 

 the same, and so also is the amount they lengthen by mechanical 

 strain. They belong, I believe, to the same system of crys- 

 tallization, namely the monometric, though too little is known 



* Communicated by the Author, having been read before the British 

 Association at Bradford, September 1873. 



