62 sosman: problems of the oxides of iron 



therefore, before all the adsorbed oxygen is removed from the 

 surface. Mathematically speaking, we may say that the dis- 

 sociation pressure curve is not asymptotic to the axis of ordinates, 

 drawn at the composition Fe20 3 , but crosses this axis at a small 

 angle. This second conception of the case gives, I believe, the 

 better interpretation of the facts. 



MAGNETIC PROPERTIES AND POLYMORPHISM OF THE OXIDES 



The problems so far considered are all problems which are 

 common to nearly all the chemical elements and their compounds. 

 But there is one field of physics in which iron holds at present the 

 first place in interest, namely, the field of magnetism. Progress 

 in the experimental study of magnetic properties has lagged 

 behind experimental progress in electricity, chiefly because we 

 possess no magnetic insulator. As has been well stated by Mr. 

 Sanford at a recent meeting of the Philosophical Society of 

 Washington, "Measuring magnetic properties is like measuring 

 electrical resistance with the Wheatstone bridge and all the 

 connections immersed in a conducting liquid." Nevertheless, 

 rapid progress is now being made on several lines of magnetic 

 research. 



The oxides of iron have been, from this standpoint, rather 

 neglected. Measurements will be found here and there in con- 

 nection with studies of the salts of iron, crystalline or dissolved. 

 From these it is clear that the oxides are not to be classed mag- 

 netically with the majority of the iron compounds. Ferric 

 oxide is abnormally low in susceptibility, in comparison with the 

 common salts of iron. Ferrosoferric oxide, on the other hand, 

 belongs, as is well known, among the highly ferromagnetic 

 substances. Ferrous oxide, again, seems to be in a class with 

 ferric oxide. Intermediate oxides possessing a variety of com- 

 binations of magnetic properties also exist. 



The inversions in ferric oxide first attract our attention. 

 Honda 11 has found that at liquid air temperatures ferric oxide 

 has only about two-thirds as great a susceptibility as at room 

 temperature, 1 and that a change occurs rather suddenly in the 



11 Honda, K., and Sone, T. Sci. Rep. Tohoku (Sendai), 3: 223-234. 1914. 



