36 JPenjield and Stanley—- 



Si0 2 : K m 2 3 :R"0 + F J SiO a :E"0 + F a 



Tremolite 



'957 : 



•014 : 



•958 



1 : 



1-00 I 



983 : 



•018 : 



•954 



1 



•991 



938 : 



•017 : 



•935 



1 : 



•997 



•914 : 



•041 : 



•908 



1 : 



•993 



•880 : 



•059 : 



•874 



1 : 



•993 



•876 : 



•045 : 



•923 



1 



1 -05 



I. Richville 



II. Lee 



f III. Greiner 



. . ... I IV. Russell 

 Actinolite < -.t- t ^ 



j V . Kragero 



[ VI. Pierrepont 



In the first five analyses the ratios of Si0 2 : R' l O + F 2 are 

 almost exactly as 1:1, and attention ma} 7 be called to the fact 

 that such close approximations to an exact ratio are seldom 

 met with in mineral analyses. It may be concluded therefore 

 from the ratios that the materials were very pure, the analyses 

 exceptionally good, and that both water and fluorine, which 

 have been generally disregarded in previous calculations, must 

 be taken into consideration. Five such results preclude the 

 possibility that the close agreement of the ratios to 1 : 1 is a 

 matter of accident. Analysis VI is irregular in that it shows 

 an excess of (R u O + F„) over Si0 2 . This may be due to defects 

 in the analysis, to possible impurities in the material, or this 

 special variety may be a transition between actinolite and horn- 

 blende, the latter, as will be shown later, being characterized 

 by having an excess of (R"0 + F 2 ) over Si0 2 . Analyses IV, V 

 VI, the ratios derived from them, and the percentages of the 

 several constituents are so nearly alike, that it seems best to 

 class the mineral from Pierrepont as actinolite. The analysis 

 needs revision. 



Except as regards fluorine, the ratios derived from the six 

 analyses confirm the theory of Rammelsberg, namely, that the 

 'composition is RSi0 3 + R 2 3 , the RSiO s including MgSiO s , 

 FeSi0 3 , CaSiOg and Na 2 Si0 3 , while considerable H 2 Si0 3 , which 

 Rammelsberg left out of consideration, must also be included. 

 The ratios also confirm in a general way the theory of Tscher- 

 mak except that considerable hydrogen would have to be 

 brought into the formulas and some provision made for 

 fluorine. Moreover the ratios indicate clearly that the idea 

 advanced by Tschermak, that sodium is present as a molecule, 

 jSTa 2 Al 2 Si 4 12 , is quite untenable, since for every JSTa 2 there 

 would have to be deducted 1A1 2 3 and 4Si0 2 , which would 

 deplete the total silica and destroy the 1 : 1 ratio. 



In a large number of minerals it has been shown that fluorine 

 and the isomorphous hydroxyl unite with metallic elements to 

 form radicals, thus (MgF)', or (MgOH)', univalent, and (A1F)" 

 or (AlOH)" bivalent, and it seems probable that fluorine, and 

 to some extent hydroxyl, enter in some way into the amphibole 

 molecule in combination with trivalent aluminium and iron. 

 It is possible in one way to account for the presence of both 

 R 2 3 and fluorine without destroying the 1 : 1 ratio shown by 



