276 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 13, No. 13 
Commerical hydrogen and nitrogen were used. Both were passed 
over hot copper to remove oxygen, and the nitrogen was in addition 
carefully dried. The purified hydrogen contained not over 0.2 per 
cent of nitrogen and allowance was made for this in the calculations. 
The gas mixtures were prepared by passing the hydrogen through 
wash bottles, containing distilled water, immersed in an electrically 
controlled thermostat. 
Electrically heated connecting tubes prevented the water from con- 
densing out of the gas mixture prior to its reaching the hot zone of the 
furnace. 
TABLE I.—EXPERIMENTS ON THE NATURE OF THE [RON PHASE 
| 
pment agen ete: | TIME, HOURS |RATIO H:O/H2 Papumiataans <6 REMARKS 
~ Initial Final 
1 750 6 0.476 100 99.98 Pure iron 
2 750 6 0.540 100 99.90 Pure iron 
3 750 6 0.33 99.78 99 .67 Piano wire 
4 980 6 0.44 100 99.86 Pure iron 
5 960 6.5 0.44 100 100.1 Pure iron 
6 960-80 6 0.44 98 .88 99 .84 Oxidized iron 
A earefully calibrated thermoelement and a potentiometer served 
for the temperature measurements. 
The initial and final products were analyzed for total iron. Both 
the usual permanganate method and the dichromate electrometric 
method were employed. 
The pure iron was an electrolytic product which had been reduced for 
many hours at 900°-1000°C. with dry hydrogen. Permanganate 
titrations indicated 99.90 and 100.0 per cent iron while dichromate 
gave 100.1 and 99.95 per cent. It contained no manganese and less 
than 0.05 per cent carbon. 
The results of the best controlled experiments are given in table I. 
The first five experiments do not indicate that pure iron will take 
up oxygen to form solid solutions. One might say, however, that such 
solutions are stable but that their rate of formation is such that they 
did not have time toform. This hypothesis seems improbable but was 
worth checking. Experiments at 750°C. afforded no information on 
this point. Even with dry hydrogen, a slightly oxidized iron sample 
3 Mr. Mulligan was unable to detect any difference between the oxidizing strengths of 
these solutions such as was recently reported to exist. Journ. Am. Chem. Soc. 44: 
2148. 1922. 
