516 A TREATISE ON METAMORPHISM. 



From the above table it appears that the order of loss, reckoned by 

 quantity, beginning with the greatest, is: Lime, magnesia, soda, potash, 

 silica, iron oxide, alumina. Of the two alkalies, more of the soda is dis- 

 solved than the potash. Of the alkaline earths, while on the whole the lime 

 is dissolved in greater amount, in some rocks more magnesia is dissolved 

 than lime. One of the unexpected things shown by this table is the fact 

 that the percentag-e loss in silica is more than one-half as g-reat as the average 

 losses of the alkalies and alkaline earths. When it is remembered that the 

 average amount of silica in the original rocks is 48.7 per cent, and that 

 the average of the alkalies and alkaline earths together is but 17.5 per cent 

 of the total, it is clear that the loss of the silica is about the same as the 

 entire losses of the alkalies and the alkaline earths together. 



This is a very interesting conclusion, since it shows that a very 

 erroneous impression prevails in reference to silica. It has been generally 

 supposed that the rocks in the belt of weathering are not much depleted in 

 this compound. On account of the extreme insolubility of crystallized 

 silica of quartz, it has been supposed that the silica of a rock is exceedingly 

 insoluble. For instance, Merrill says: "Silica, even in its most soluble form, 

 requires 10,000 times its weight of water for solution."" Doubtless, under 

 ordinary circumstances, quartz is but sparingly soluble, although, as 

 shown in another place (see p. 848), even this form of silica is dissolved 

 to some extent. But in the majority of instances a large part of the silica 

 is present as a silicate, although of course quartz is also abundant. When 

 the silicates are broken up by carbonation there is every reason to believe 

 that the greater portion of the freed silica separates as silicic hydroxide 

 or colloidal silicic acid, which is readily soluble. That this reaction 

 takes place has already been shown (see p. 480) and that the silicic 

 hydroxide is dissolved and abstracted from the rocks is shown by the 

 tables already given. In the rocks in which the silica is largely as 

 silicate and the alumina is taken as constant the calculated losses of 

 silica are high — in gneiss, 52.45 per cent; in syenite, 62.18 per cent; in 

 decomposed diabase and basalt, 18.03 to 65.56 per cent; in diorite, 37.31 

 per cent; in pyroxenite, 43.58 per cent, and in argyllite, 57.57 per cent. 

 The average of these numbers is 48.14 per cent, It should also be 

 remembered that these are underestimates, because they are based upon the 



a Merrill, George P., Rocks, rock-weathering, and soils, Macmillan Co., New York, 1897, p. 238. 



