50 i?. B. Rig g 8 — Composition of Tourmaline. 



I. Lithia tourmaline 12Si0 2 , 3B 2 3 , 4H„0, SA1„0 3 , 2(NaLi) 2 0. 



II. Iron tourmaline 12SiO„", 3B o 3 , 4H„0, 7A1 2 3 , 4FeO, Na a O. 



III. Magnesian tourmaline i2SiO„," 3B 2 3 , 4H,0, 5A1„0 3 , 

 ^-MgO, |Na 2 0. 



Calculated. I. II. III. 



B o 3 11-00 10-18 10-90 



Si*0 o 37-70 34-89 37-38 



Al o 3 42-75 34-59 26'49 



FeO 13-95 



MgO 19-31 



Li„0 





1-57 

 3-21 



3-77 



Na o 3-21 2-90 2-18 



H„0 3-77 3-49 3-74 



100-00 100-00 100-00 



On comparing the theoretical composition of the above types 

 with the results of analysis, they will be found to agree as 

 closely as could be expected, at least in values of the constants. 

 The boric acid found invariably falls short of the theory. This 

 is to be expected. The analyses do not represent ideal com- 

 pounds, but are made of material more or less impure and the 

 case would be very exceptional where the impurity tended to 

 raise and not lower the percentage of boric acid. 



In some of the above formulas the group B0 2 has been as- 

 sumed because the oxygen ratio demanded it. As has already 

 been suggested the ease with which the iron oxidizes and the 

 mysterious manner in which this change takes place, under con- 

 ditions when we would suppose it to be impossible, possibly 

 point toward a higher degree of oxidation than the more com- 

 mon B 2 3 . As the result of a slow molecular rearrangement 

 the one is oxidized at the expense of the other. Such a change 

 is, I believe, not without its analogies. Certain borates, where 

 the assumption of an even higher oxide is thought necessary, 

 on being heated give borates of a lower order. 



Notes. — The question of color is an interesting one, particu- 

 larly when the varying colors of the lithia tourmaline are 

 concerned. For, while the color of the iron and magnesian 

 varieties depends on the amount of iron present and ranges from 

 the colorless Dekalb through all the shades of brown to the 

 Pierrepont black, the lithia tourmaline, containing more or less 

 manganese, give us the red, green and blue, as well as the color- 

 less varieties, the shades of color not depending on the absolute 

 amount of manganese present but rather on the ratios existing 

 between that element and the iron. When the ratios of Mn : 

 Fe approximate =1:1 we have the colorless, pink or very pale 

 green tourmaline. An excess of manganese produces the red 



