70 Scientific Proceedings, Royal Dublin Society. 



separated from free quartz, orthoclase, and tourmaline, by Thoulet's 

 specific gravity solution. The crystals thus obtained were freed of 

 large mixed fragments and very fine particles by descent through 

 a long column of still water ; — a method described by me elsewhere 

 (Proc. B. D. S., vol. iv., p. 291). The large rough fragments so 

 removed were found on microscopic examination to be very im- 

 pure, the medium-sized crystals fairly pure, the fine dust very 

 heterogeneous and impure. A few of the larger of the medium- 

 sized crystals being sorted out for photographing and mounting, the 

 remainder, weighing about two decigrammes, was devoted to 

 analysis. 



The percentages obtained in the analysis hardly approximate 

 to the numbers for iolite given in the left-hand column. This was 

 to be expected from the impurity of the sample. The presence of 

 4 per cent, of MgO is, however, important, as there was probably 

 no source of impurity present capable of affording an appreciable 

 weight of that body. I think it highly probable that glucina may 

 enter into the composition of these crystals, replacing the magnesia, 

 the elements Be and Mg being isomorphous. Grlucina was not 

 looked for in the analysis. It is remarkable that these crystals 

 seem present only in orthoclase, thus intermixed with beryl. In 

 other sections of Grlencullen orthoclase, as well as in sections of 

 granite 1 from Co. Cavan, Co. Mayo, Newry, Killiney, Warrenpoint, 

 and from the Mourne Mountains, I could detect none of these 

 crystals. As in the case of their felspathic matrix, I have little 

 doubt of the correctness of my diagnosis of the mineral species. 

 Whether perfectly normal in chemical composition or not, can 

 however only be decided by further and more careful analysis. 



The formula to be deduced from the analysis, such as it is, is — 



10 SiO, . 2 (A1A . FeO) . MgO . H a O, 



which, be it observed, affords a bisilicate oxygen ratio, instead of a 

 unisilioate ratio. 



As a microscopical mineral, this iolite will be recognized by its 

 basal angles of 150°, 120°, or 60° ; its generally symmetrical extinc- 

 tion on elongated rectangular sections, and the transverse cleavage 

 on such sections. The foliation, or plating on O, and the oblique 



1 Kindly lent by Professor Hull. 



