SOSMAN AND HOSTETTER: REDUCTION OF IRON OXIDES 297 



cipitating the former with hydrogen in sulphuric acid solution.^ 

 For the other portion of crucible No. 2, and for crucible No. 1, 

 the method of MyHus and MazzuccheUi,'* by which all metals 

 present could be determined, was employed. In addition to 

 iron, this method showed the presence of rhodium in both cruci- 

 bles, and this latter metal was also determined. The order of 

 agreement between the two methods for iron is shown in Table 

 III, in which the results of the analyses are given. To test the 

 determination of iron by the method of Mylius and Mazzuc- 

 chelli, a solution containing 977 mg platinum and 5.00 mg iron 

 was subjected to analysis; 4.83 mg iron was recovered. 



TABLE III 



Magnetic properties of iron-bearing platinum. The apparatus 

 for the magnetic tests was suggested by Dr. C. W. Burrows, of 

 the Bureau of Standards, to whom we are greatly indebted for 

 advice and information in connection with these tests. The 

 procedure consists essentially in weighing the pull exerted 

 on the specimen by a strong electromagnet. In paramagnetic 

 and diamagnetic bodies the force is practically independent of 

 the shape of the specimen. 



Since we wished to obtain only comparative figures, we did not 

 attempt to determine the field strength or field gradient. With 

 our apparatus, as set up, 870.2 mg of ferrous ammonium sulphate 

 gave a pull of 1.74 mg, equivalent to 2.00 mg per gram. The 

 pull can be determined to about 0.02 mg by the method of swings. 



^ For the details of this method and for many helpful suggestions, we are 

 indebted to Dr. E. T. Allen of this Laboratory. 

 ^ Zs. anorg. Chem., 89: 1-38. 1914. 



