Composition of Staurolite. 87 



visible in the fragments, when examined with the microscope ; 

 from the results of the analyses, however, it is evident that 

 they were not wholly eliminated. To test this point more 

 carefully, the following experiments were made on some of 

 the finely powdered minerals left over from the regular 

 analyses. After digesting with cold, strong hydrofluoric acid 

 for twelve hours and washing, silica determinations were made, 

 which are given below, along with the determinations from 

 the previous analyses. 



St. Gothard. Windham, Me. Lisbon, N. H. 

 Si0 2 after treatment with HF., 27'52 27'36 27-15 



" from regular analyses. .. . 27*73 27*84 27*81 



It will be observed that hydrofluoric acid has removed some 

 silica, but still the percentages are higher than the theory. 

 We should infer, therefore, that quartz is an impurity in the 

 mineral and that it is present as very minute inclusions. If, 

 for example, the inclusions are as fine or finer than the acicu- 

 lar crystals of rutile in quartz, they could not be removed by 

 a specific gravity separation, nor, being enclosed in the stauro- 

 lite, would they be wholly accessible to the action of hydro- 

 fluoric acid. That the formula suggested by Groth is correct 

 is well established by our analyses and, surely, its simplicity is 

 one of the strongest arguments that can be advanced for its 

 acceptance. 



On the regular arrangement of inclusions in staurolite 

 crystals. — In examining orientated thin sections of crystals 

 from Lisbon, ]S r . H., it was observed that they all contained 

 dark inclusions, arranged in certain definite planes, resembling 

 the phenomena so common in andalusite. That the inclusions 

 are carbonaceous material was proved by the fact that, on 

 separating the pulverized mineral by specific gravity, the dark 

 portion was found to be lighter than the clear staurolite, and 

 on igniting it in a current of air, purified by passing over 

 caustic potash, carbon dioxide was abundantly evolved. These 

 inclusions can only be clearly seen in plates ground sufficiently 

 thin to be transparent and can best be studied in basal sections. 



Figures 1 to 4 represent the arrangement of the inclusions 

 in plates cut from a simple prismatic crystal 50 mm in length by 

 ll mm broad from which eleven basal sections were cut. Near 

 the ends the impurities are arranged as in fig. 1 ; at the middle 



