24 Penfield and Stanley — 



theory is as follow : To tremolite, the simplest form of am phi- 

 hole, the formula CaMg 3 Si 4 12 is assigned ; ordinary green 

 amphibole, actinolite, corresponds to Ca(Mg,Fe) 3 Si 4 12 ; while 

 in varieties containing considerable amounts of sesquioxides, 

 there is assumed the existence of an alumo-silicate mole- 

 cule (Ca,Mg) 2 Al 4 Si 2 12 and a corresponding one containing 

 ferric iron, (Ca,Mg) 2 Fe 4 Si. 2 )2 . The last two molecules have 

 only a theoretical existence ; they are supposed to be isomor- 

 phous with Ca(Mg,Fe) 3 Si 4 ]2 , but no pure chemical combina- 

 tions corresponding to them have ever been observed. The 

 presence of alkalies is ascribed to the existence of the meta- 

 silicate molecules Na„Al 2 Si 4 12 and K 2 Al 2 Si 4 ]2 , the former 

 being analogous to jadeite, NaAlSi 2 O . Water is supposed to 

 be due to alteration and there is no provision in the formulas 

 for fluorine, which is an almost unfailing constituent of amphi- 

 bole. In Tschermak's latest publication* the formulas sug- 

 gested for amphibole are CaMg 3 Si 4 12 , CaFe 3 Si 4 ]2 , and two 

 alu mo-silicates, CaMg 2 Al 2 Si 3 12 and ]STa 2 Al 2 Si 4 12 . 



It is true that by a combination of the several molecules 

 suggested by Tschermak theoretical compositions may be cal- 

 culated which correspond rather closely to some of the analyses, 

 but the formulas are in general found to be wanting in many 

 respects when subjected to careful comparison with existing 

 analyses, and Tschermak freely admits that the problem is one 

 presenting numerous difficulties. 



Rammelsbergf has pointed out that amphibole analyses con- 

 form to the formula E // Si0 3 ,E = Ca,Mg,Fe,Mn,]Sra 2 and K 2 , 

 plus varying proportions of A1 2 3 (and Fe 2 3 ), which for- 

 mulas, however, do not apply to arfvedsonite and related 

 minerals of the amphibole group, rich in soda and sesquioxides. 

 In their essential features the formulas of Tschermak and 

 Rammelsberg are alike, Ca(Mg,Fe) 3 Si 4 12 of Tschermak being 

 equivalent to 4ESiO s , and the alumo-silicate molecules (Ca,Mg) 2 - 

 Al 4 Si 2 12 to 2PtSi0 3 + 2Al 2 3 . They differ in that Tschermak 

 regards the alkalies as belonging to molecules like JSJa 2 Al 2 Si 4 12 , 

 respectively JSTa 2 Fe 2 Si 4 ]2 , while Rammelsberg recognized 

 Na 2 SiO s as isomorphous with CaSi0 3 and MgSi0 3 . Neither 

 of them take into account water and fluorine, which, as will 

 be shown, are essential and often prominent constituents of 

 amphibole and should not be disregarded. 



Numerous other formulas have been suggested by various 

 investigators, but these in general refer only to analyses of 

 single specimens of amphibole, and it does not seem necessary 

 to discuss them at the present time. 



* Lehrbtich der Mineralogie, 1894, p. 458. 

 f Mineralcliemie, 1875, page 394. 



