Determination of Pyrite and Marcasite. 381 



reckoned in weight per cent, but before we consider in detail 

 the meaning and accuracy of the results, it will be best to put 

 the method on a safer footing by examining all other possible 

 sources of error which we could discover. 



3 Other sources of error in the Stokes method. 



(a) The ratio of the surfaces of the two minerals in iveight 

 percentage mixtures. — In the first place if the active mass of 

 each sulphide depends on its surface rather than its weight, as 

 it should for solids, it will be obvious to the reader that the 

 relation between (x—a) the iron dissolved in the Stokes re- 

 action, and y, the percentage of pyrite in the sulphide mixture, 

 cannot be strictly linear in weight percentage mixtures, be- 

 cause the minerals naturally differ a little in density. They 

 are of nearly the same hardness, and assuming that both are 

 brought to the same degree of fineness when they are ground 

 together, the percentage of surface of the marcasite, the 

 lighter mineral, would be somewhat greater than its per- 

 centage by weight. Suppose that a mixture containing 50 

 per cent of each mineral by weight when ground in water,* 

 consists of spherical particles of uniform size ; then 



— = ^ = r7 r- , where n, = the number of particles of mar- 

 n 2 S p 4-89' 



casite, n 9 = the number of particles of pyrite, and S m and S p 

 the total surface of the two minerals respectively. The 

 ground mixture of the minerals in equal quantities would 

 therefore contain only 49*3 per cent instead of 50 per cent 

 pyrite surface. This error, however, is approximately com- 

 pensated, under the conditions described, by the more rapid 

 action of the ferric solution on the marcasite particles. If all 

 the particles are of the same size, those of each mineral must 

 be uniformly attacked, and assuming the action of each is in- 

 dependent of the other, the final surface of the marcasite 

 would be s m = 7^4 7TB, 2 , and the final surface of the pyrite 

 would be s p = ?iAtt R, 2 , where R, and E 2 are the radii of the 

 particles of the two minerals at the end of the oxidation. It 

 follows from the above that the final weights of the two 

 minerals in a 50 per cent mixture where 1 g. was taken would 

 be: 



0*155 

 0-500 — =0423 for the marcasite, and 



A 



0-500 - () ^=0-468 for the pyrite. (See also p. 373.) 



* To insure a more uniform size of grain as well as a more uniform mix- 

 ture, each mineral was screened within similar limits before mixing, after 

 which the mixture was ground in water. 



