JULIEN, DETERMINATION OF MINERAL CONSTITUTION 131 



into greater or prevailing proportion. Familiar examples of these transi- 

 tions are found in the endless variations of intermixture of quartz, even 

 within its crystals, with hyalite, iron-oxides, rutile, chlorite, etc.; the inter- 

 inclosure, intergrowth and inter-twinning of the feldspars in aggregates of 

 the most complex constitution, and the similar mutual envelopments of the 

 metallic sulphides. 



The possibility of even "ideal purity" of a mineral has been based 

 largely on results of examination of material selected for chemical analysis. 

 The precautions usually taken to insure freedom from impurities are proba- 

 bly shown fairly in those long ago described by Doelter.* The fragments 

 were first examined by the naked eye and then under a hand lens. A thin 

 section was prepared and inspected. Splinters and cleavage-plates in 

 different directions were then spread on a glass slide and examined by 

 transmitted light under a low magnifying power of the microscope. By 

 these means, it was believed, the visible purity of the material was insured, 

 or, if impurities could still be detected but not removed, they were identified 

 and allowance made for their amount in the reduction of the analytical 

 figures. 



In the light of present knowledge all these precautions appear insufficient 

 to insure purity. From the subtle revelations of existing intergrowths now 

 obtained through polarized light — the absolute concealment of all foreign 

 inclosures within subtranslucent and opaque specimens — and, in every 

 case, the escape of microscopic inclosures from observation, whatever their 

 abundance, whose minute dimensions fall below the resolving power of the 

 microscopic lens — the natural conclusion follows that the most effective 

 detection of inclosures must be sought through study of the relationships 

 of the chemical components of the mineral. 



2. The usual mode of application of purely hypothetical compounds in 

 rearrangement of components. 



Without questioning the propriety of their consideration in reconstruction 

 of an analysis, little seems to be gained toward real explanation of lacking 

 relationships, by excessive resort to imagined compounds, like Mg Fcj'" 

 SiOg, Fcj Fe/'' Sig O^j, and others, in pyroxene, which have never been dis- 

 covered in nature, in isomorphous interlocking with others, like CaMg SigOg, 

 whose co-existence as actual minerals is proved by optical behavior. In 

 such cases, a conviction of the extent of dissemination of existing minerals 

 as inclosures will lead rather to more persistent search for the latter, and, 

 I think, more satisfactory solution of difficult problems constantly presented 

 in recasting analyses. 



An analysis then is not the end, but it is only a step toward the discovery 

 of the existing mineral constitution. As the chemical composition of an 



1 Min. u. petr. Mitth., I, 49, 67, 373. 1878. 



