-±60 Hi S. Washington — Rhyolites of Lipari. 



Basalts. 



Submarine Eruptions and Pantelleria. 



, Hyaline > r— Crystalline— >, 



SiO, 48.97 44.83 45.72 46.40 46.22 



Fe 2 6 3 1.33 1.35 1.57 4.09 4.91 



FeO 8.56 11.79 12.01 8.22 7.71 



^r 14.88 18.22 16.70 4.38 3.45 



Fe 2 3 



Symbol TI.5.3.4 III.5.3.4 III.5.3.4 III.5.3.4 III.5.3.4 



The figures given in Table V would seem to show quite 

 conclusively that there is a direct relation between the 

 ratios of the iron oxides and the degree of crystallinity 

 in the rhyolites and the basalts of the Italian volcanoes, 

 in the sense that ferrous oxide dominates over ferric in 

 the glassy forms, while the reverse is true in the crystal- 

 line forms, of lavas that are otherwise of almost identical 

 composition and are derived from the same magma. 

 The same relation holds good for rhyolites and other 

 such highly silicic lavas at other volcanoes, as in the 

 Auvergne, the Yellowstone Park, at Pelee, and else- 

 where, the data for which are not given here. The same 

 relation is also apparent, though possibly not so invari- 

 able, among the lavas of intermediate composition, such 

 as dacites, trachytes, and andesites, and is also to be 

 seen among the basaltic lavas, though hyaline forms are 

 less commonly met with among basalts, while analyses of 

 them are still more rare. 



An adequate explanation of this state of affairs is 

 somewhat difficult to find, and the subject requires further 

 study. It may reasonably be assumed that the hyaline 

 forms represent more closely than the crystalline . the 

 chemical conditions as to the iron oxides that obtain in 

 the liquid magma itself, and the relations may presum- 

 ably be connected with the reducing or oxidizing char- 

 acters of the magmatic gases. As to the exact nature 

 of these gases as they exist in solution in the magma we 

 have but little knowledge. The earlier analyses, such 

 as those of Deville, Fouque, and Brun, cast no light on 

 the problem, partly because the gases were collected 

 after issuing from the lava and thus having been 

 subject to mixture with and oxidation by the air, and also 

 because of the very imperfect methods of analysis that 

 were used in their study. The studies of the gases from 



