Page:The Journal of geology (IA journalofgeology11893univ).pdf/191

igneous rocks already cited. There is nearly the same relation between the alkalies and alumina, the soda being still more in excess of potash, and both increasing gradually toward the more silicious end of the series. The alumina maintains a high position, gradually decreasing. The lime, iron-oxide and magnesia decrease rapidly from the less silicious to the more silicious end of the series, and lie somewhat close together. In one instance there is a marked drop in the magnesia.

In Table II are given the analyses of lavas from the Andes south of Colombia, with one exception, as they have been recorded in Roth's tables of rock analyses. They present a somewhat wider range of silica percentages, from 52.02 to 73.61, but whether they have all been made from unaltered rocks is not known to the writer. An analysis of "obsidian" from Colombia is omitted, since it is extremely crude, given 75 per cent. of silica and 3 per cent. of magnesia, with no lime. Diagram 2 expresses the molecular variations among the rocks included in this grouping. They have the same general nature as those just described, especially as to the alkalies and alumina; but the lime, iron-oxide and magnesia are more variable, which may represent the true condition of the case, or may be due to imperfect methods of analysis.

It is evident from these diagrams that the lavas of the Andes represent various phases of the differentiation of a magma which is chemically similar to that which has furnished the lavas of the Great Basin and extreme western Cordilleras in the United States, and that they belong to similar petrographical provinces.