98 Penfield and Foote — Composition of Tourmaline. 



He was the first to point out that silica and boric oxide are 

 present in the definite molecular proportion 4 : 1. He errone- 

 ously decided that tourmaline contained carbon dioxide, the 

 reasons being as follows : It was generally believed at that 

 time that the mineral contained no water, as stated by Hermann 

 " die Turmaline Jceine Spur davon enthalten" and it is true 

 that when fragments are tested by the usual method of heating 

 to redness in a closed tube no water is obtained. It is only 

 when the material is heated intensely, best as fine powder, that 

 hydroxyl is decomposed and water given off. When Hermann 

 dissolved fragments in a borax bead he observed that a gas was 

 evolved, and, since it was believed that this could not be water 

 vapor, he supposed that it must be carbon dioxide. Pains 

 were taken to fuse some of the mineral in a tube with borax, 

 and to conduct the gas into lime water, by which treatment a 

 precipitate was obtained which effervesced with acids, but it is 

 safe to assume that the carbon dioxide thus detected was de- 

 rived from improperly purified air and not from the mineral. 



In 1850 Rammelsberg* published the results of the analyses 

 of thirty varieties of tourmaline. The execution of such a 

 large number of analyses must be regarded as a very great 

 undertaking, since at that time gas and many facilities of our 

 modern laboratories were not available, and many methods of 

 analysis were not perfected. Special evidence is given that 

 great care was taken in the selection of material for analysis 

 and in the analytical methods, which appear to have been well 

 chosen and reliable in character. The analyses, however, were 

 defective in several important particulars. Thus the iron was 

 regarded chiefly as ferric. Believing, like Hermann, that 

 tourmaline contained no water, and having detected the pres- 

 ence of fluorine in some varieties of the mineral, he supposed 

 that the considerable loss on ignition which occurred was due 

 to the volatilization of silicon fluoride, and from this loss he 

 estimated fluorine to be present in amounts varying from 1*30 

 to 2*51 per cent. Direct determinations of boric oxide were 

 made in three cases only, and in the remaining analyses this 

 important constituent was estimated by difference. Although 

 the analyses led to no satisfactory formula, they indicated cer- 

 tain prominent characteristics of the mineral, namely, the great 

 variation in the relative amounts of aluminium, iron, magne- 

 sium and alkalies, and the nearly uniform amounts of silica 

 and boric oxide. 



Fully realizing certain defects in his earlier analyses, Ram- 

 melsbergf published in 1870 a revision of his former paper. 



* Ann. der Phys. u. Chem., lxxx, p. 449 and lxxxi, p. 1. 

 f Ann. der Phys. u. Chem., ccxv, pp. 379 and 547. 



