PHYSICAL PEOPEETIES OP VOLCANIC EOCKS. 89 



quantities of alkali (soda and potash) may have a larger percentage of 

 silica without excess, than rocks containing more of lime, magnesia, and 

 iron and less of alkali. Thus trachytes, which have a comparatively large 

 proportion of soda and potash, and very little lime and iron, seldom show 

 any evidence of excess of silica unless the percentage exceeds 68 per cent, 

 and then, as the silica increases, they graduate into rhyolites. On the other 

 hand, such rocks as propylite and andesite, which contain an abundance of 

 lime and iron, begin to show evidence of an excess of silica when the percent- 

 age of it exceeds 62 per cent, or sometimes even 60 per cent. The reason for 

 this is not far to seek. The alkalies are capable of forming definite combi- 

 nations with a much higher percentage of silica than are lime, magnesia, 

 and iron. The alkalies give rise to the acid feldspars, albite, and orthoclase, 

 while the lime gives rise to the basic feldspar, anorthite, and iron and mag- 

 nesia to the equally basic minerals of the pyroxenic, hornblendic, and olivin 

 groups. 



On the other hand, the alkalies sometimes form basic minerals, such as 

 leucite and nephelin. This happens whenever these bases are present in 

 quantities in excess of those required to form feldspar, or, what amounts to 

 the same thing, when the ratio of silicate of alumina to soda or potash is 

 less than that required to form albite or orthoclase. Hence, in basic rocks 

 rich in potash, we find leucite, and when they are rich in soda, nephelin, 

 either or both replacing feldspar. 



Turning now to the magnesian minerals, the same kind of correlation 

 is seen. Where the quantity of magnesia relatively to the silica is very 

 great olivin is formed abundantly. This is the most basic mineral occurring 

 in eruptive rocks, and is found only in rocks which are least siliceous. 

 Where the quantity of magnesia is less, augite and hornblende are 

 formed. In the two latter minerals it appears that lime, magnesia, and 

 iron protoxide largely replace each other, lime predominating in augite, 

 and magnesia in hornblende. They are moderately basic, but less so 

 than olivin. In the more acid rocks magnesia takes frequently the form 

 of mica (biotite), in which the quantity of protoxide base is still less than 

 in hornblende. 



With regard to alumina, it is somewhat remarkable that although the 



