416 wells: interpretation of mineral analyses 



CHEMISTRY. — The interpretation of mineral analyses.'^ Roger 

 C. "Wells, Geological Survey. 



In a recent number of this Journal W. T, Schaller proposed an 

 improvement in the usual method of reducing a mineral analysis 

 to a chemical formula which seems to have some advantages. ^ 

 A little later F. E. Wright and C. E. Van Orstrand published a 

 paper on the determination of the order of agreement between 

 observation and theory in mineral analyses in which they state 

 that Schaller's method is in error in certain details of principle 

 and discuss several methods of making comparisons between 

 analytical data and theoretical formulas based upon the method 

 of least squares.^ Without doubt both of these papers represent 

 certain advances, but as they stand there is too much variance 

 between them. The subject is one of general interest and fun- 

 damental importance so that some further observations may not 

 be inappropriate. 



The excellent analysis of pearceite upon which the discussion 

 in these papers is based was published some time ago by F. R. 

 Van Horn and C. W. Cook.^ At present only methods of com- 

 paring the analytical data with theory are in question. Van Horn 

 and Cook obtained from their analysis the atom numbers of sul- 

 fur, arsenic and silver 10.80, 2.000 and 7.886 from which they 

 easily deduced the correct formula, 8Ag2S.As2S3. The real ques- 

 tion decided concerned only one atom in 15, that is between the 

 formulas 9Ag2S.As2S3 and 8Ag2S.As2S3. The analysis was con- 

 siderably more accurate than necessary to decide this point. 



The improvement proposed in the calculation by Schaller is 

 suggestive altho not essential to the evaluation in this case and 

 consists in the use of an arithmetic mean instead of a single value 

 of the greatest common divisor by which percentages found in 

 an analysis are reduced to numbers of atoms. The details of 

 carrying out this calculation may be found in his paper. He thus 

 obtained the numbers 10.92, 2.02 and 7.97 which are closer to 



^ Published with the permission of the Director of the U. S. Geological Survey. 

 ^ Journ. Wash. Acad. Sci., 3: 97. 1913. 

 3 Journ. Wash. Acad. Sci., 3: 223. 1913. 

 " Amer. Journ. Sci., (4) 31: 518. 1911. 



