Mineral Sulphides of Iron. 177 



the lower half of the apparatus. In use, this joint is greased 

 with vaseline and the rings are held in place by two spring 

 clamps which are not shown in the figure. The two vessels 

 D and E are used to contain, respectively, hydrochloric acid 

 and boiled water cooled in carbon dioxide, with which two 

 liquids the ground sulphide in B is successively washed. The 

 pressure of the carbon dioxide which enters the apparatus at 

 H is sufficient to drive either liquid over into B when the 

 proper stop-cocks are opened. G opens into a filter flask 

 filled nearly full of water, connected with the water pump. 

 Suction is carefully regulated by a pinch-cock. The appara- 

 tus is successively evacuated by the pump and filled with car- 

 bon dioxide before using. In drying the material in the 

 vacuum desiccator, instead of using the water pump as Stokes 

 did, we have evacuated by an oil or mercury pump to a frac- 

 tion of a millimeter. At the end of a half hour the sulphide 

 is not dry but it does not contain enough water to affect the 

 result essentially. In the determination of the ferrous iron, 

 we have employed a weight burette instead of a volume 

 burette. 



Results with natural pyrite and marcasite. — The minerals 

 employed were pyrite from Elba* and marcasite from Joplin, 

 Mo. They were purified with great care and then analyzed. f 

 The only impurities found were small quantities of silica and 

 a minute trace of copper in the marcasite. 



Marcasite Pyrite *' Cal. for FeS 2 



Fe 46-53 46'49 46'56 



S 53-30 53-49 53'44 



SiO„ -20 -04 



100*03 100-02 100-00 



The oxidation coefficients obtained by us were 56 for pyrite 

 and 14 for marcasite, while Stokes found 60 and 18 respect- 

 ively. The differences have not been entirely accounted for, 

 though we have taken somewhat greater precautions in our 

 work. However, and this is the point to be emphasized, the 

 results of each are probably consistent among themselves. 



Determination of the relative quantities of pyrite and 

 marcasite in mixtures. — By grinding together the two min- 

 erals in different proportions and then determining the oxida- 

 tion number for the mixture, Stokes constructed a curve repre- 



*In some of the later experiments, a pyrite from Leadville, Col., con- 

 taining 0*1 per cent copper, was used. 



f For the method of analysis, see Allen and Johnston, Zs. anorg. Chem., 

 lxix, 102, 1911. 



