82 
Proceedings of the Royal Society of Edinburgh. [Sess. 
of the temperature at which the magnetic effect reaches a maximum, viz. 
near 550° C. Further, there are signs of three maxima on each curve. 
These occur at the temperatures 425°, 575°, and 750° (800°) respectively, 
and are supposed by us to correspond with the appearances of Fe 3 0 4 , 
FeO, and Fe, or possibly with the sequence: magnetic ferric oxide, Fe 3 0 4 , 
and Fe. The cooling curves are both iron curves, and not magnetite, since 
they proceed from a minimum at about 900° C. 
The reduction of volume of the powder which takes place when the 
ferric oxide is heated above 500° C. in hydrogen or air makes it impossible 
to add a true scale of intensity or susceptibility to the curves. 
Summary of Results. 
The experimental results may be briefly summarised as follows : — 
1. When pure artificial ferric oxide and impure haematite ore were 
heated in air in a magnetic field of 43 C.G.S., they became magnetic 
with rise of temperature, and were magnetised, more or less permanently, 
when cooled down from 1000° C. The seat of magnetisation in the case 
of the ore, as indicated by magnetic methods, was ferrosoferric oxide. 
(Curves 1, 2, 3.) 
2. When ferric oxide was heated in hydrogen, the magnetic effect 
usually appeared later and at a higher temperature than when air was 
employed. Also, the temperature at which magnetism began to appear 
was more definite, viz. 300° C., and the intensity of magnetisation was 
much greater than with air. (Curves 1 and 7.) 
3. The temperature-deflection curves for ferric oxide heated in hydrogen 
have maxima at temperatures 425°, 575°, and 750° C.,* and the cooling curve 
is that for iron. These three temperatures are therefore regarded as the 
optimum temperatures for the production of Fe 3 0 4 , FeO, and Fe respectively, 
or for the sequence : magnetic ferric oxide, Fe 3 0 4 , and Fe. (Curve 7.) 
4. The experiments, so far as they go, support the view that magnetic 
oxide may be formed at comparatively low temperatures in rocks which 
contain haematite. 
Note . — In a short appended statement, references are made to the work 
of Moissan and others on the chemical transformation of ferric oxide when 
continuously heated in a reducing gas. The results seem to the authors 
to have an intimate bearing on the interpretation of the experimental results 
described above. 
* See p. 77. 
