Penfield and Foote — Composition of Tourmaline, 99 



At this time it was shown that all varieties of tourmaline con- 

 tained chemically combined water, and the amount of water was 

 estimated from the earlier determinations of the loss on igni- 

 tion after making certain corrections for volatilization of sili- 

 con fluoride. He found that the iron was chiefly if not wholly 

 ferrous, and recalculated accordingly his earlier results. Six 

 direct determinations of boric oxide were made, and it is 

 pointed out that these amounts correspond closely with the 

 indirect determinations by difference. As a result of the 

 revision, Rammelsberg reached the conclusion that all tourma- 

 lines are derived from the acid H 6 Si0 5 . In this he considered 

 the hydrogen atoms to be replaced by metals of different val- 

 ences, or, in other words, he regarded tourmaline as composed 

 of a mixture of the following molecules : 



R' 6 SiO & R' = Na, K, Li and H. 



R" 3 Si0 5 R" = Fe, Mg, Mn and Ca. 



R"' 2 Si0 5 R '" = Al and B. 



Furthermore he decided that certain varieties correspond 

 closely to the special formulas 



t j R' 3 Al 2 BSi s O,„ and n jR'. 



Al w B 1 Si.O« 



■Al.ASi.O, 



while he regarded others as mixtures of these two molecules. 

 By substituting hydrogen atoms for the metals and boron of 

 these special formulas, the acids become respectively H 24 Si 4 O 20 

 and H 64 Si 9 45 , both of which are multiples of H 6 Si0 5 . He 

 concluded that the Si0 2 and B 2 3 are not present in a definite 

 molecular proportion, but that boron plays the part of a metal 

 and is isomorphous with aluminium. 



In 1888 Riggs* published the results of twenty analyses of 

 various types of tourmaline from American localities. The 

 analyses were executed in the laboratory of the U. S. Geologi- 

 cal Survey at Washington, and bear every evidence of being 

 made with the precision and care characteristic of the analyti- 

 cal work of that laboratory. Boric oxide, water and ferrous 

 oxide were determined directly by reliable methods, and a high 

 degree of accuracy is claimed for the analyses. A careful 

 description of the quality of the material analyzed is not 

 given, and although it is to be supposed that great care was 

 taken in its selection, the following statement has left this in 

 doubt : " The analyses do not represent ideal compounds, but 

 are made of material more or less impure . . . . ' Riggs 

 concludes that the analyses give "as a general tourmaline 

 formula the simple boro-orthosilicate R 9 B0 2 2Si0 4 " which is 

 expressed graphically as follows : 



* This Journal, IIT, xxxv, p. 35. 



