On tlie Position of Lievrite in the Mineral Series. 349 



Apart from this want of concordance, these formulae do not 

 serve to connect our mineral with other species of kindred cha- 

 racter. In other words, they fail to present any satisfactory in- 

 dication of the place which the Lievrite should occupy amongst 

 the silicates generally. Reflecting upon this, it occurred to me 

 that the true composition of the mineral might be arrived at by 

 tracing out its mineralogical affinities. If this seem paradoxical, 

 it must be remembered that purely mineralogical considerations 

 have taken the initiative on more than one occasion in the solu- 

 tion of obscure questions connected with mineral chemistry. 

 Whilst, for example, the various garnets, the different varieties 

 of pyroxene, &c, were still kept apart by the chemist who ad- 

 hered to chemistry alone, mineralogy insisted upon their union, 

 and thus led the way to the recognition of isomorphism. If the 

 relationship of Lievrite to a mineral of known composition can 

 be clearly shown, a great assistance will at least be afforded 

 towards the deduction of its true atomic character. 



On comparing this mineral with other silicates, one cannot 

 help being struck by the remarkable correspondence existing 

 between it and chrysolite, at least as regards the so-called Fayal- 

 ite and other iron-holding varieties of the latter. It is curious 

 that this coincidence should hitherto have escaped attention. 

 The gelatinization of the silica in acids, a very peculiar character 

 in the case of anhydrous silicates, is exhibited by the two species 

 in common. Their conditions of occurrence are also more or 

 less identical ; the form in each is trimetric, with axial relations 

 in part corresponding \ and each contains an unusually low ave- 

 rage of silica. This amounts in hyalosiderite and Eayalite to 

 about 30 per cent. The same also in Lievrite. Dana places the 

 latter mineral (though doubtfully) in his Andalusite Group, with 

 Andalusite, topaz, and staurolite ; but a collocation of this kind 

 is an exceedingly forced one. Geological relations (an element 

 in mineralogical classification of the highest importance, although 

 hitherto strangely overlooked), composition, and general charac- 

 ters are all opposed to it. The form, it is true, is trimetric, with 

 some remote analogy, as shown by Dana, to that of Andalusite ; 

 but since we find such opposite minerals as augite and borax, 

 for example, exhibiting an identity of crystallization, no great 

 stress can be placed on this character alone. There is an equal 

 amount of crystallographic correspondence, moreover, between 

 Lievrite and chrysolite, and in other respects the two present 

 a close agreement. In chrysolite several vertical prisms are 

 known. If we denote the protaxial prism by the symbol V, and 

 make the macrodiagonal unity, these forms give for the brachy- 

 diagonal the following values: — 0*4660 (=V; prism-angle, 

 130° £') ; 0-9484 (=V 2 ; prism-angle, 93° 3') ; 1*397 ( = Y 3 ; 



