Calc-alkaline volcanic suites from circum-oceanic island-arc regions should be treated separately from the volcanic and plutonic assemblages of continental margins when developing petrogenetic hypotheses. Chemical plots for volcanic suites from the West Indies and Solomon Islands indicate that crystal fractionation of high-alumina basalt magmas could account for the basalt-andesitedacite suite. The separated minerals responsible for this control are likely to differ from those (chiefly magnesian olivine) in the tholeiitic suite.
Experimental studies on a variety of calc-alkaline volcanic rocks in vacuo have shown unexpectedly high liquidus temperatures for the andesites and dacites (> 1250° C), and a general irregularity in the relationships between melting temperatures and usually accepted chemical-composition parameters, despite measures taken to eliminate residual grains and to achieve nucleation from the glasses. Reconnaissance runs at P H2O = 2 kb have helped to resolve some of these problems. They lead to the conclusion that the calc-alkaline volcanic suite differs from the tholeiitic and alkaline chiefly in that dissolved water is needed to produce calc-alkaline andesitic and dacitic melts at reasonable temperatures.
Plutonic nodules from the West Indies lavas carry an assemblage of cumulate crystals notably rich in amphiboles and anorthite, with interstitial vesicular basalt. This supports the concept that magma was generated by partial melting of either a locally hydrous layer of the upper mantle, or a depressed wedge of hydrous basaltic rock. Fractional crystallization under conditions in which P H2O and oxygen fugacity were constant or increased thus led to derivative andesitic and dacitic liquids. The removal of amphibole is more likely than the removal of iron oxides in producing such derivative liquids.