296 SOSMAN AND HOSTETTER: REDUCTION OF IRON OXIDES 



from the furnace. The amount of iron absorbed, as calculated 

 from the amount of oxygen evolved and remaining in the appa- 

 ratus after coohng to room temperature, making no allowance 

 for any permanent thermal dissociation of the magnetite, is 

 found to be 0.9 mg. This is one-half of the total gain in 

 weight. 



Magnetite is also reduced by platinum at high temperatures 

 in the open air. 6.608 grams of Merck's reagent Fe203 was 

 heated to a maximum of 1612° in an open platinum crucible 

 (No. 509) in a 25 mm vertical-tube platinum wire furnace. The 

 charge melted down at about 1582°, yielding black crystalline 

 "magnetite" on cooling. The crucible, which "originally weighed 

 9517.3 mg, increased in weight to 9521.2 mg, a gain of 3.9 mg. 

 Certain points on the bottom of the crucible attracted a light 

 magnetic needle. 



Reduction of ferric oxide. Products containing more ferric 

 oxide than magnetite, as well as magnetite itself, seem to be re- 

 duced by platinum under a low pressure of oxygen. This action 

 is evident even in products which are practically pure Fe203. 

 One crucible (No. 2) used in our work on the system Fe203-Fe304, 

 after 18 series of measurements, had increased in weight from 

 2851.8 mg to 2863.0 mg, a gain of 11.2 mg or 0.39 per cent. 

 About one-fourth of this gain is found to be due to iron absorbed. 

 At the close of the measurements this crucible showed the usual 

 iron coloration, produced by cleaning and heating for a few min- 

 utes in a flame. Various points on the crucible also attracted a 

 light magnetic needle. 



Under 10.2 mm oxygen pressure at 1200°, under which con- 

 ditions the magnetite is nearly completely oxidized to Fe203, 

 •this crucible gained 1.6 mg. Another similar crucible contain- 

 ing Fe203 gained 1.0 mg under 18.3 mm oxygen pressure at 1200°. 

 A part, at least, of these gains must be due to absorbed iron. In 

 order to estabhsh definitely the presence of iron in the crucibles, 

 and to determine its amount, we have made both chemical 

 analyses and magnetic tests on the crucibles in question. 



Methods of analysis. For the analysis of one portion of crucible 

 No. 2 the platinum metals were separated from the iron by pre- 



