488 Prof. S. J. Shand — Saturation in Petrography. 



Here is the crux of the whole matter. Anybody who can produce 

 a system in which chemical and mineralogical characters shall be 

 correlated in a satisfactory manner will have solved the great 

 problem of petrography. For myself, having studied the attempts 

 that have been made up to the present time to correlate these 

 properties of rocks, and having myself made very numerous and 

 laborious efforts to accomplish the feat, I have become convinced that 

 nothing is to be gained, but much to be lost, by juggling with these 

 two factors in a hybrid classification. 



How can one talk of correlating mineral and chemical composition 

 when one of the commonest of rock minerals shows such a range of 

 composition as the following ? — 



The figures are extracted from Iddings,^ and the various materials 

 to which the analyses refer are all described as ' hornblende'. 



Nephelite is a very definite species to the observer at the eyepiece 

 of the microscope, yet it may hold from 14 to 18 per cent of soda, 

 and from 1 to 7 per cent of potash. The black or brown garnet, 

 which the microscopist identifies as melanite, may range in composition 

 from andradite to schorlomite, and may liold 10 per cent of Mn or 

 AI2 O3, or 20 per cent of Ti Oj, which can only be detected by 

 chemical analysis. 



It is clearly impossible for us to isolate and to analyse separately 

 every constituent of a given rock ; hence I say that until we discover 

 a means of deducing the chemical composition of such minerals as 

 these from their optical properties, with as much certainty as is 

 attainable in the case of the felspars, no satisfactory expression of the 

 chemical composition of a rock can be made in terms of minerals, or 

 vice versa. 



The question whether mineral and chemical composition can be 

 correlated in a satisfactory way is surely answered by the failure of 

 every attempt to perform the operation. Osann's method- of building 

 a chemical superstructure upon a mineralogical basis is severely 

 condemned by the American chemical school of petrography,^ and 

 has not found favour in Britain or France. Iddings, in his most 

 recent work,* employs a simple mineral-textural method of grouping 

 and appends to each group a discussion of the chemical composition 

 of the rocks comprised within it. But to do this he must practically 

 make a separate statement of composition for each rock ; no general 



^ Rock Minerals, 1911. 



^ Tschermaks Min. Pet. Mitt., xix et seq. 



^ Cf . Cross in The Quantitative Classification of Igneous Bocks. 



* Igneous Rocks, vol. ii, 1913. 



