guns £9; 1.928 FERGUSON: OXIDES OF IRON 281 
Eastman has collected all the results to date on this equilibrium con- 
stant for various temperatures. A study of his graph of these shows 
that the results of Schriener and Grimnes give for the invariant point 
a temperature of 636°C. Our work places this temperature below 
577°C. and furnishes additional evidence that the constants of these 
workers are too large. It does not indicate, unfortunately, whether 
the values of Chaudron or of Eastman are to be preferred since one 
places the invariant point at 570° while the other gives it a value of 
565°C. 
Our stream method experiments give an upper limit for the constant. 
Since Chaudron was able to obtain good agreement between oxidation 
and reduction experiments at 785°C., we think that our value, which 
agrees with his, must be not far from the true value for the particular 
temperature. 
The discrepancy between this value and that calculated by Eastman 
from the constants of the (CO,CO,,Fe,FeO) system arises from the 
fact that he gave great weight to the work of Matsubara. Chaudron’s 
results on the two systems are in excellent agreement at 785°C. as 
evidenced by the fact that from them he obtained a value for the con- 
stant of the water-gas reaction of 0.87, Haber’s value for this being 
0.86. In view of this, one is somewhat at a loss as to how to interpret 
these results of Matsubara. Perhaps his iron phase was not the same 
as that found by us in the other system. His conclusion, that the iron 
phase contains much ferrous oxide in solution, though the most obvious 
one, is not the only possible explanation of his results and would seem 
to be in marked conflict with our own observations. 
SUMMARY 
1. The iron phase in the system (H2,H.O,Fe,FeO) does not con- 
tain appreciable quantities of oxygen. 
2. The transition temperature of the ferrous oxide phase appears 
to be lowered by solution of ferro-ferric oxide in the ferrous oxide, 
The quadruple point lies below 577°C. 
3. A value for the equilibrium constant at 750°C. was obtained 
by the stream method. It agrees with the value obtained by the 
interpolation from the results of Chaudron and furnishes a check 
on the latter. 
10 The temperature of the quadruple point in the system (Fe, O,H) may be considered 
for our purpose as identical with that of the invariant point in the system (Fe, O). 
