VAN ORSTRAND AND WRIGHT: MINERAL ANALYSES 515 



between the different methods which have been proposed and thus 

 to ascertain which method is the best. It may be stated in 

 anticipation of the final results that the methods of dealing di- 

 rectly with the data of observation are superior both from a 

 theoretical and a practical standpoint to the new m.ethods which 

 have been suggested. Too strong emphasis cannot be placed 

 on the importance of the direct comparison of the data of obser- 

 vation. Division or multiplication of the weight percentages 

 of the analysis by different numbers automatically assigns differ- 

 ent weights to these values; by this procedure a really large error 

 can be made to appear small and practically negligible and vice , 

 versa, a small error relatively large and important. This fact 

 was emphasized by the writers in a recent article^ but a more de- 

 tailed discussion of certain phases of the subject seems now desir- 

 able in view of the statements in the paper by Wells cited above. 



The chemical analysis of a mineral presents in quantitative 

 fonn the weight percentages of the various chemical elements 

 of which it has been found to consist. In mineral analytical 

 work chemists consider carefully a number of factors among 

 which the following are the more important: Purity of sub- 

 stance to be analyzed (to insure proper purity microscopic 

 examination and selection of the individual mineral grains is 

 often necessary); fineness of grinding of material; purity of 

 chemical reagents; precision of chemical balance and weights; 

 accuracy of atomic weights used; contamination from utensils 

 em.ployed; accuracy of the chemical methods for separating and 

 for detemiining the chemical elements present; personal skill 

 of the analyst. Most of these factors can be either eliminated 

 or their influence definitely determined and proper correction 

 made for them by the chemist. In addition to these and other 

 analytical factors the presence of foreign elements in solution 

 in the crystal state, is always to be considered in the final re- 

 sult; this may at times be so serious as to veil completely the 

 simple chemical formula of the mineral compound analyzed. 



Systematic and accidental errors. In the chemical analysis 

 of a mineral the actual weight percentage relations between 



' J. Wash. Acad. Sci., 3: 223-231. 1913. 



