MINEKALOGV; WJTH A CLASSIFICATION OF SILICATES. 35 



zeolites, in which appear well crystallized species, having the atomic ratios (excluding 

 the water) of the intermediate feldspars, and also from the evidences of species like hyalo- 

 phane and lerxcite. The late observations by Tschermak as to the action of acids on 

 various intermediate scapolites, to be noticed further on (§ '75), go far to show that these 

 are not admixtures but integral compounds. 



§ 26. In concluding the paper of 1854, which I have here reviewed, it was said 

 with reference to the problem of a natural system in mineralogy, then, as now, before 

 the writer : — ■ 



" No mineralogical classiiication can be comi^lete which does not take into account 

 both the chemical and the physical characters of species ; and the connection between 

 these, which is shown in the relation of equivalent weight to specific gravity, must consti- 

 tute an important element in a natural system. Gruided by their physical characters and 

 composition, we bring together such homœomorphous species as belong to one chemical 

 sub-type, and from the densities fix their formulas and comparative equivalent weights. 

 From the comparison of the formulas and the associations of these different minerals, we 

 must also decide which are to be considered as mixtures and which are true species. 

 Until we shall have determined with certainty the comparative volumes of dissimilar 

 crj'^stalline forms, the relations of species differing in this respect must be decided by their 

 affinities, and their places in a homologous series must remain undetermined. In this 

 way we may hope to arrive at a mineralogical classification which shall satisfy alike the 

 chemist and the naturalist." 



§ 2*7. Before going farther, it seems proper to advert to the history of the notion of 

 polymerism in silicates and carbonates, which enters into the views maintained in the 

 author's papers of 1853 and 1854 ; and to show its relation to the views previously put 

 forward by Auguste Laurent. He had already in 184*7 proposed to reduce all natural 

 silicates to a small number of types corresponding to the observed atomic ratios. These 

 ^àeld both neutral and basic salts, according to Laurent, who, moreover, in his notation 

 admitted, in order to explain the complex results of chemical analysis, a divisibility of 

 molecules to which he assigned no limit, and supposed that protoxyds and sesquioxyds 

 might within certain limits replace each other indefinitely. He also extended a similar 

 to the borates. ' In a subsequent memoir, in 1849, Laurent criticised the arbitrary 

 formulas proposed by chemical mineralogists, and showed that the relations therein set 

 forth were often but approximations. It was pointed out by him that in many related 

 species, as for example in the various micas, the atomic relations between sesquioxyds and 

 protoxyds were not constant, and it was argued that these two classes of bases, and water, 

 were capable of replacing each other mutually within certain limits, in ratios which, as 

 represented by liim in atomic formulas, seemed to be indefinite. He also insisted on the 

 importance in silicates of small portions of water, which, though generally neglected in 

 the formulas, ought not to be regarded as accidental. This later paper - however, while 

 reflecting the perplexed state of chemical mineralogy, fails to propose any solution of the 

 difficulties. The reader will note the broad distinction between the simple formulas with 



' Comptes Rendus des Trav. de Chimie, July, 1847, from Comptes Rendus de l'Acad. xxiii. 1050, and xxiv. 94. 

 For an analysis of Laurent's memoir by the writer, see Amer. Jour. Science, 1848, v. 405. 

 Snr le.s Silicates ; Comptes Rendus des Trav. de Chimie, 1849, pp. 256-288. 



