Mixing and unmixing of magmas

Abstract

Linear compositional variations of calc-alkaline rocks are difficult to explain in terms of mixing of two genetically unrelated magmas. At least two processes must be involved, a deep-seated one responsible for the long-term variations between suites of rocks erupted in successive episodes and a shallow-level one by which magma columns become vertically zoned during periods of repose between eruptions. The first process appears to involve resorption of differing proportions of mafic residue in a rising batch of magma. The second results from fractionation and collection of a buoyant liquid under the roof of a high-level reservoir. The effect of cooling and crystallization of a calc-alkaline intrusion is to produce a compositional change and decrease of density in the fractionated liquid that has a greater effect on the boundary layer than thermal contraction. The same effect may result from melting of felsic wall rocks or absorption of water. Many of the features that are commonly observed in calc-alkaline rocks can be explained in terms of these two processes without appealing to physically implausible mechanisms involving mixing of liquids of differing densities.