64 THOMAS STERRY HUNT ON A NATUEAL SYSTEM IN 



hundredths of ferrie oxyd, aud often traces only. It is not improbable that the true ratio 

 of the protoxyd and sesquioxyd bases in these two species, as in ineionite, with which 

 they have been paralleled, may be I'epresented by 4 : 9, rather than by 1:2. We note 

 next the more siliceoirs jadeite, whose formula, as already pointed out (§ 38), is related to 

 that of zoisite as dipyre is to meionite (§ 31). While zoisite is essentially a calcic species, 

 seldom containing OA'er three or four hundredths of soda, jadeite is sodic, and it 

 appears, like the compact zoisite or saussurite, to be anhydrous. The atomic volume of 

 zoisite and of jadeite, as shown in the table, appears to be less than that of garnet and 

 epidote, showing a more condensed molecule. 



§ 83. We have next to notice three remarkable adamantoids, closely related to those 

 just mentioned, and also to the spathoid ilvaite. In garnet, axinite, ex>idote and keilhau- 

 tite, the sesquioxyd may be in large part ferric, and in schorlomite and ilvaite it is 

 entirely so, the protoxyd bases in these being chieily lime, magnesia and ferrous oxyd. 

 We have in œgirite, arfvedsouite and acmite, three well-defined protopersilicates in 

 which the sescjuioxyd is entirely ferric and the protoxyd in large part sodic. These three 

 species, which have hitherto been little understood, will be seen from tlie table to be 

 related respectively, tegirite to garnet, acmite to epidote, and arfvedsonite to euclase, and 

 to haA-e a common value for V A'ery near to that of garnet and epidote. 



The presence, in each of these ferric species of large amounts of soda, ec[ual to ten or 

 twelve hundredths, is the more remarkable since the aluminous silicates with which we 

 have compared them contain little or no alkali. This association recalls the highly 

 alkaliferous character of another iron-silicate, glauconite. While these three homoeomor- 

 phous species, all ferric unisilicates with soda, having very different ratios between 

 protoxyds and sesquioxyds, are, from their condensed molecule, and their indifference to 

 acids, assigned a place among adamantoids, the related ilvaite, with a larger volume, has 

 been placed among the spathoids. It is possible, from the analysis of Rammelsberg, that 

 babingtonite may be a ferric species belonging to the one or other of these tribes, but 

 without farther analyses it would be premature to fix its place. 



§ 84. We come next to spodumene, a lithia-alumina species with the atomic ratio, 

 1 : 4 : 10, remarkable for its great condensation and its volume of 4'88. It is instructive 

 to compare it with the still more siliceous lithia-alumina silicate, petalite, which, with its 

 lower density, has a volume of 633, and takes its i^lace among the spathoids. The rela- 

 tions between these two silicates are analogous to those between zoisite or jadeite and a 

 scapolite like marialite. While these two lithia-bearing species, with the ratio of pro- 

 toxyd to alumina of 1 : 4, are among the most siliceous known, sapphiriue, which has 

 the same ratio, is the most basic, and with its atomic formula of (mgial,si)o,„ serves 

 to connect the silicates with the spinellids, while, by its great condensation, it takes a 

 place by the side of spodumene. 



Staurolite is essentially an aluminous double silicate, with the ratios of 1 : 4 : 2|, 

 the protoxyds being ferrous oxyd with a little magnesia, and, rarely, a portion of oxyd of 

 zinc. In one variety it woïild seem that manganese-sesquioxyd replaces a portion of 

 alumina, and a small portion of water appears to be an essential element. Omitting the 

 water we get a volume of 5'01. 



§ 85. We come now to the tourmalines, a family of silicates which, perhaps, miglit be 

 called a subtribe, since the five distinct species, representing as many genera, differ from 



