MINEEALOGY; WITH A CLASSIFICATION OF SILICATES. 31 



species wiil exhibit the same relations as those of the carbon series, and the prin.iple of 

 homology be greatly extended in its application. The history of mineral species affords 

 many instances of isomorphous silicates whose formulas differ by 71OM2, as the tourma- 

 lines, and the silicates of alumina and magnesia, while the latter, with many zeolites, 

 exhibit a similar différence of mO.H, [O in these formulas = 8]. The relation is in fact 

 that which exists between neutral, surbasic and hydrated salts." It was further 

 declared that the carbon-spars must be represented as polycarbonates, having not less than 

 from " twelve to eighteen equivalents of base replaceable so as to give rise to a great num- 

 ber of species ; " while the variations in the calculated atomic volumes of these carbon- 

 ates were said to "indicate the existence of several homologous genera, which are 

 isomorphous." 



§ 17. These conceptions of progressive series of more or less highly condensed mole- 

 cules of polycarbonates and polysilicates, and of similarity of volume for isomorphous 

 species, were developed more at length in a second paper published in the same year, 

 1853, on "The Constitution and Equivalent Volume of Mineral ï^pecies." ' It was therein 

 explained that the formulas of homologous bodies may be represented as series in arith- 

 metical progression, in which the first term may be like or unlike fhe common difference; 

 both cases being, it was shown, illustrated in the chemical history of mineral species, 

 including carbonates, silicates and oxyds. Similar views were also then extended to 

 nitrates and sulphates, as well as to chlorids and to sulphids. 



The simplest atomic formula of the carbonates being CMO;; (C = U and O = 8, accord- 

 ing to the equivalent weights then in use), the rhombohedral carbon-spars were referred 

 to three genera represented by «(CMOJ : namely, (1) calcite, w = 30 ; (2) dolomite, siderite 

 and diallogite, « = 36 ; and (3) smithsonite and magnesite, n = 40. For the prismatic species, 

 aragonite, like calcite, belonged to a genus with w = 80 ; while for strontianite, cerusite and 

 bromlite, w = 25 ; and for witherite, n = 22. The volumes of the rhombohedral species 

 deduced from these ec[uivalents were from 550 to 560, and for the prismatic species from 

 500 to 510. These arbitrary weights and volumes were at the time supported by com- 

 parisons with those deduced from the formulas of the rhombohedral red-silver ores and 

 the prismatic bournonite, and farther by the volume of the compound of glucose and 

 sodium-chlorid, regarded as homieomorphous with calcite, with a density of 1.563, which, 

 doubling its empirical formiila, gave a volume of 558.5. The various alums, if their for- 

 mulas be doubled, give in like manner, as was shown, volumes of from 543 to 561. 



\^ 18. Extending to the silicates the same notion of polymerism which had just been 

 applied to the carbonates, the existence of various polysilicates was admitted. Thus the 

 formulas of spodumene, diopside, hudsonite, and woUastonite, were described as present- 

 ing a homologous series of the first kind, in which the first term is- the same as the 

 common difference, "represented by «(SioMO,.), the respective values of n being 30, 

 26, 24 and 22." Spodumene was then, chiefly on crystallographic grounds, compared 

 with the pyroxenes. The excess of silica above the bisilicate ratio, met with in some 

 amphiboles, was referred to as an example of a homology of the the second kind, in which 

 the common difference is unlike the first term. To these species there was assigned an 

 equivalent volume approximating to 460. In support of this volume, it was noted that 



' Amer. Jour. Science, 1853 (xvi. 203-218) and iu abstract in the author's Chem. and Geol. Essays, p. 438, etc. 



