Chemical Composition of Amphibole. 25 



Before presenting the new analyses and entering upon their 

 discussion, it seems best to set forth certain premises which 

 may well be taken into consideration. 



In the first place, it may be noted that amphibole contains 

 an unusually large number of constituents. SiO„, A1 2 3 , Fe 2 3 , 

 FeO, MnO,' MgO, CaO, JS T a 2 0, K 2 0, H 2 and F, among which 

 are metallic elements having valences of three, two and one, 

 and if any reasonable formula is to be proposed there must be 

 some satisfactory disposition made of hydrogen and fluorine. 

 For a mineral having so many and such variable constituents, 

 and likewise having them in very variable proportions, it 

 seems reasonable to assume a complex chemical structure. 

 Compare amphibole for example with chrysolite, and a marked 

 difference is noted ; the formula of the latter is (Mg,Fe) 2 - 

 Si0 4 , and, disregarding impurities, if constituents other than 

 Si0 2 , MgO and FeO are noted in the analy ses, they are either 

 NiO or MnO, which are readily explained as isomorphous 

 with MgO. In what are generally regarded as simple chemi- 

 cal compounds only such constituents are to be found which 

 are alike in valence and in grouping according to the periodic 

 system, and thus conform to the ordinary law of isomorphism ; 

 for example, FeO and MnO ismorphous with MgO. It may 

 be noted also that in general Na 2 is not isomorphous with 

 either K 2 or CaO, nor CaO with MgO, such constituents 

 when occurring together generally forming double salts. In 

 amphibole, however, owing as it is believed to its complex 

 chemical structure, or to what one of the present writers has 

 called mass effect* such unlike constituents as H 2 0, Na 2 0, 

 CaO, MgO and A1 2 3 are in some way brought into a mole- 

 cule as isomorphous constituents or radicals, a result which we 

 do not meet with nor expect to meet with in the case of simple 

 chemical compounds. 



Again, when we consider the great diversity in chemical 

 composition of the various minerals of the amphibole group, 

 including glaucophane, riebeckite and arfvedsonite, all of them 

 being essentially alike in crystallization, in cleavage and in 

 prismatic angle, it must be assumed that there is some control- 

 ling factor responsible for this similarity in crystallization, and 

 it is believed that this controlling factor is the acid. From 

 the composition of the simplest minerals of the group, tremo- 

 lite and actinolite, it seems definitely proved that the amphi- 

 boles are salts of metasilicic acid, not necessarily H 2 Si0 3 , but 

 some multiple of this. Tschermak has suggested H 8 Si.,0 12 as 

 the stoichiometrical formula of the amphibole acid, and there 

 are reasons for believing that this, or some multiple of it, is 

 correct. It is to be regretted that at the present time there is 

 * This Journal, xiv, 211, 1902. 



