Constitution of the Natural Silicates. 305 



species previously soluble could be thus split up into a soluble 

 aud an insoluble part. In a number of cases ignition of a min- 

 eral caused the liberation of silica, which could afterwards be 

 dissolved out with soda solution and quantitatively determined. 

 For this purpose a solution of sodium carbonate, 250 grams 

 to the liter, was employed. Finally, in almost every case the 

 nature of the water in a mineral was investigated, by a study 

 of the temperatures at which its several fractions were expelled. 

 For low temperatures, an ordinary air-bath was used ; for 

 higher temperatures the minerals were heated in a stream of 

 dry air, between indicators of known melting points, just as in 

 the treatment with gaseous hydrochloric acid. 



So much for the methods of research, which will be more 

 clearly understood from a study of the details to be given pres- 

 ently. On the theoretical side we have adhered to the work- 

 ing hypothesis that the more complex silicates are substitution 

 derivatives of normal salts, as.has been suggested by one of us in 

 several earlier publications. From this point of view the nor- 

 mal ortho-silicate of magnesium is the fittest starting point for 

 discussion, and this salt is approximately represented by olivine. 

 The purer forsterite was not available. 



1. Olivine. 



For this species the only material at hand was a supply of 

 the chrysolite pebbles from near Fort Wingate, New Mexico. 

 The mineral from this locality, we believe, hasnot been pre- 

 viously analyzed ; and as it is rich enough in color and is suffi- 

 ciently clear it is often cut as a gem. It has the characteristic 

 color of the peridot, and is apparently quite free from inclu- 

 sions. The analysis gave the following results : 



Si0 2 



Fe 2 3 



FeO 



NiO 



MnO 



M<jO 



H 2 ..... 



100-11 



Only 0*05 per cent of the water was lost at 105°. 



1*1027 grams of the mineral were heated in gaseous hydro- 

 chloric acid for twenty-two hours, and gained in weight 0*0157 

 grm. On leaching with water and a drop of nitric acid the 

 following percentages of material went into solution : 



MgO 1-47 



FeO -43 



41 



•98 





•51 



5 



•71 





•42 





•10 



51 



•11 





•28 



