R. B. Riggs — Composition of Tourmaline. 47 



As the outcome of his work, Eammelsberg concludes that 

 all tourmaline may be referred to the following silicate types : 

 E' 6 SiO B , E" 3 Si0 5 , and R" 9 SiO„ in which R'=H, Li, Na and 

 K, R"=Mg, Ca, Fe, and In and R'"=A1 and B, making bo- 

 ric acid and alumina equivalent. He groups them in two 

 classes: I, iron-magnesian tourmaline, represented by the 

 formula 



j R'SiO, ) j R" 3 Si0 5 ) 



j SR'^SiO J + >M 3R''' 2 SiO a j" 



E'"=:B : 2A1 and II, lithia tourmaline referred to the formula 



R' Si(X ) . I R",SiO R ) where R'"=B : 3A1, 

 8rf'" a SfO. f + w "{ 8R'",Si5 B f 



with the following special formulae : R' 2 R" 2 Al 4 B 2 Si 4 O 20 for the 

 more purely magnesian varieties ; E' 10 E" 4 Al 12 B 6 Si 12 O 60 for the 

 purer iron varieties, and for the lithia tourmaline E 7 6 Al 12 B 4 

 Si 9 4B , and other complications for those lithia varieties con- 

 taining iron. 



Attention has been called to the weak points in Eammels- 

 berg's analyses ; but, for the sake of emphasis, I refer again to 

 his method of determining the water and what depends on it. 

 He assumes that the fluorine is driven off quantitatively on ig- 

 nition, carrying with it silicon as the tetrafluoride. This 

 assumption is a questionable one. A half hour's blasting would 

 seem to be sufficient to drive off the water, but it was repeat- 

 edly observed that where the tourmaline contained fluorine, 

 further blasting, after the water was presumably all expelled, 

 was invariably accompanied by a small loss. This after loss 

 was probably fluorine in some of its silicon compounds, 

 as experiments made it clear that boric acid is driven off by 

 no amount of blasting. While making this correction for flu- 

 orine, the oxidation of the iron is wholly ignored. If anything 

 was noticeable in the ignition of tourmaline it was the appar- 

 ent ease with which the iron oxidized. In the above table the 

 loss on ignition is given in many cases. Comparing this loss 

 with the direct determinations of water, at the same time no- 

 ting the amounts of iron contained in the tourmaline in ques- 

 tion, one is forced to conclude that the oxidation of the iron 

 plays an all-important role here, in diminishing the loss by ig- 

 nition by the amount required to convert it into ferric oxide. 

 If we take Eammelsberg's ignition results* and make the more 

 justifiable correction for the oxidation of the iron, setting aside 

 the fluorine correction, we would have more concordant results 

 to deal with ; results, moreover, which agree closely with the 

 direct determinations of water. But if the percentages repre- 



, lxxx, 449; lxxxi, 1. 



