76 
Proceedings of the Royal Society of Edinburgh. [Sess. 
red oxide powder held in a copper boat, and for a bar of Bacares ore. This 
method of treatment appears to develop greater fluctuations in the magnetic 
state of the bar than the second method, in which the bar is not raised above 
the critical temperature of magnetite. There is a certain amount of 
similarity in the course of the two curves, and the magnetic state tends 
to become steady after two hours’ total heating at 850°. It is remarkable 
Double 
Deflecl'i'or ? 
Heating in Gas Furnace at 850° C. for intervals of 10 minutes. 
that the first minimum for the haematite bar is on the zero line ; also 
that in each of the curves the first maximum is followed by a higher 
one, as in Curves 2, 3, 5, 6, 7, and 9. The method is unlikely to lead 
to numerical results, but it exhibits the gradual change in the magnetic 
state of a bar subjected to short intervals of heating at a temperature 
above the critical point of iron. 
II. Ferric Oxide Bars heated in Hydrogen. 
The hydrogen used in this part of the investigation was produced by 
the action of hydrochloric acid on granulated zinc, and the gas was bubbled 
through strong sulphuric acid, to remove moisture, before it was led into 
the furnace. 
1. Heat Cycles. (See Curve 7.) 
The two curves given refer to the behaviour of a specimen of red oxide 
and one of brown oxide, each held in a copper tube. From the curves it 
may be seen that magnetic quality begins to appear a little above 300°, 
increases to a high maximum at 575°, and falls to a minimum at 600° or 
625°. In one case the heating was discontinued at 800°. In the second 
case another minimum was reached at 900°. In both cases the powder was 
magnetic during cooling, and the cooling curves point to a separation of 
