MINEEALOGY ; WITU A CLASSIFICATION OF SILICATES. 63 



homœomorpliouis with epidote, and is often spoken of as a ceriiim-epidote, to the atomic 

 ratios of which some analyses apparently conform. A farther study of the group of 

 minerals commonly inchxded under the name of allauite or orthite is required. The great 

 differences in density ; the facts that some resist the action of a^ids, while others are 

 attacked thereby, that some are anhydroiis, while others are more or less highly hydrated, — 

 all lead to the conclusion that several species are here included. We have already sepa- 

 rated therefrom the so-called xanthorthite, as a cerium-zeolitoid, and it is probable that 

 besides one or two hydrous species, and a true adamantoid, there will be found at least one 

 intermediate spathoid species. The alumina in the allanites is often in part replaced by 

 ferric oxyd. A pure alumiua-allauite, with the garnet-ratio, in which the protoxyd bases 

 are ec^ually divided between cerous and ferroirs oxyds and lime, gives the value for V as 

 calculated in the table. 



^i 81. The gluciuic species, beryl, is generally regarded as having the atomic ratio, 

 1:1:4, and has a volume near to garnet. The late analyses of Penfield ' have, however, 

 shown that beryl contains a small and variable amount of alkalies replacing glucina, 

 besides a portion of water varying from 1.50 to 2.50 per cent. He finds that the com- 

 position of the mineral is best expressed by the more complex formula (gl-,al,isi^o)Oi.j+4ac[,, 

 a change which, however, afiects very slightly the A'alues calculated in the table, that of V 

 being thereby changed to 5.48. 



Euclase, though closely related to beryl in composition and, like it, hydrated, shows 

 a much greater condensation. Ardeunite, which presents the atomic ratio of euclase, and 

 is also hydrated, is essentially a manganese-ahxraina silicate with some magnesia and 

 lime, besides a small portion of vanadate, more or less completely replaced, in some 

 instances, by arsenate. These latter elements are probably comparable, in their relations, 

 to the sulphates in nosite and hauyne. Abstracting them, we find for the silicates 

 essentially the formula given in the table, which can, however, only be regarded as 

 approximate. 



The species axinite is noticable for containing some boric oxyd. The formula which 

 we have deduced in the table, in which one eighth of the silica is thus replaced, and one 

 third of the sesquioxyd is ferric, is, also, but an approximation. The composition of this, 

 like that of beryl, of ardennite and a great number of polysilicates, cannot be accurately 

 represented by such simple formulas, which, however, suffice to show, with sufficient 

 exactness, the atomic volume and the place of the species in the system. 



§ 82. We come next to epidote, the composition of which presents many variations, 

 due in part to a greater or less replacement of alumina by ferric oxyd, and in the so-called 

 piedmontite, by manganese sesquioxyd. The presence of a small amount of water, equal 

 to about 2.0 per cent, seems, as in beryl, euclase and ardennite, to be essential to the 

 composition of the species. The atomic formula for a pure lime-alumina epidote, as 

 imagined by Eammelsberg, is (ca,al.,si;,)o,., ; but such an epidote is unknown in nature, 

 and we have, for the purpose of determining the volume of the species, selected a variety 

 in which one third of the sescjuioxyd is ferric. The formula, moreover, takes no note of 

 the small amount of water present in the species. 



Zoisite is essentially a lime-alumina silicate, seldom containing over five or six 



Amer. .Tour. Science, 1884, xxviii. 25. 



