Liquid Immiscibility in a Silicate Melt

Abstract

THE formation of immiscible liquids in silicate melts has been suggested as one of the possible methods by which natural magmas may produce a variety of rock types. This theory has, however, not found much favour among petrologists, because of the lack of observational and experimental evidence. Bowen¹ pointed out that direct evidence of this phenomenon, if it existed in rock melts, should be found in glassy lavas such as obsidian; in these lavas would be found discrete spheres of one glass embedded in another glass of different composition. No such phenomena have been observed. Greig² conducted extensive laboratory investigations on immiscibility in binary and ternary silicate systems. The areas of liquid immiscibility which he found lay well away from those representing even abnormal rock compositions. More recently, however, Roedder³ found an extensive region of low-temperature immiscibility in the system FeO—K₂O—A1₂O₃—SiO₂. This new region of immiscibility fell much closer to the compositions of some abnormal rock types. Direct evidence of the existence of silicate immiscibility in Nature has still been lacking, however.