458 H: S. Washington — Rhyolites of Lipari. 



about twenty specimens. It is also quite in harmony with 

 the generally domalkalic characters of the lavas of the 

 Aegina-Methana-Milos-Santorini-Nisyros row of vol- 

 canoes, the rocks of which are mainly dacites and 

 andesites, with some basalts. 



It is of interest to note the chemical correspondence 

 between the Milos obsidian and one from Nisyros, pub- 

 lished by Martelll (No. 2 above). The two are very 

 closely alike, except for the somewhat higher potash in 

 the second and the higher ferric oxide as compared with 

 ferrous. The Nisyros obsidian, it is to be noted, is not 

 a pure glass, but contains phenocrysts of quartz, andesine, 

 orthoclase, and pyroxene ; so that here also we find the 

 general rule as to iron oxides and crystallinity observed. 



The Relation of Iron Oxides and Glass Content. 



Attention has already been called to a feature shown 

 by the rhyolites of Lipari which seems to merit some 

 further discussion, because of its apparently common 

 occurrence elsewhere, and because it may possibly throw 

 some light on the nature of the gases that are present in 

 lavas as they issue from the depths. This is, that in 

 many effusive lavas, ranging from rhyolites to basalts, 

 ferrous oxide dominates very greatly over ferric oxide 

 in the glassy forms, while ferric dominates over ferrous 

 oxide, or at least is in much greater relative amount, in 

 the docrystalline or holocrystalline forms. That is, it 

 would appear that when a lava cools rapidly, other things 

 being equal, the iron is less oxidized than if it cools more 

 slowly. This seems to be true of the lavas of several 

 Italian volcanoes that I have studied chemically, and also 

 seems to be often true of those of other regions. It is a 

 point to which I have already called attention several 

 times, 18 but which has apparently not been noticed 

 otherwise. 



To illustrate this relation there are given in Table V 

 data from analyses by me of lavas from various Italian 

 volcanoes, selected so as to eliminate, so far as possible 

 with the data at hand, disturbing factors due to differ- 

 ences in chemical composition. Many other analyses 

 might be cited as well, but it would seem best here to 

 restrict our observation to the lavas of one general 

 region. The analyses given are confined to rhyolites 



18 H. S. Washington, this Journal, 27, 148, 1909; 36, 583, 1913; Jour. 

 Geoi., 22, 707, 1914. 



