298 sosMAN AND hostetter: reduction of iron oxides 



The mean of two sets taken with reversal of the current ehminates 

 a small correction due to the steel parts of the balance. 



One-half of crucible No. 2 had already been dissolved for 

 analysis. The other half was tested magnetically and gave a 

 pull of 0.20 mg. This was so much smaller than the effect which 

 we expected on the basis of preliminary results obtained by 

 Burrows and Burgess at the Bureau of Standards, that we experi- 



TABLE IV 



mented further on the magnetic effects produced by known 

 percentages of iron added to pure platinum. For this purpose 

 a new 10 x 18 mm platinum crucible (Heraus ''extra rein") was 

 first tested and found to give no measurable magnetic pull. 

 Iron was deposited in the crucible electrolytically from oxalate 

 solution, and it was then heated in the vacuum furnace, first at 

 1200°, later at 1400°, in order to dissolve the iron in the platinum. 

 The results are recorded in Table IV. 



