The ‘orogenic andesite’ puzzle of C. E. Tilley: 1—Is the volatile phase the missing piece?

Abstract

Some 65 years ago, C. E. Tilley, in his 1950 Presidential Address to the Geological Society of London, brought into sharp focus the then perhaps-not-fully-recognised problem of the origin of what he termed the ‘orogenic andesites’—and, in particular, their commonly constituent calc-alkaline lineage of lavas and pyroclastic materials. In spite of much investigation and progress since, the problem remains not fully resolved. The two most conspicuous features of such orogenic—i.e. island arc—volcanism are its copious, explosive, degassing and, as Tilley noted, its copious development of andesites and their more felsic associates. If these rock types were derived from basalt, the principal relevant changes in major element composition are reductions in Fe, Ca and Mg, and increases in Si, Al and (Na + K). Four principal mechanisms for inducing these changes have been proposed: (1) magmatic differentiation induced by fractional crystallisation, (2) magma mixing, (3) partial melting of an ultramafic-derived basaltic parent, and (4) crustal contamination. A fifth postulate—long recognised but only occasionally seriously considered—is that loss in the volatile phase during eruption and later degassing may make a significant contribution to the overall process. On the basis that Tilley's (and others’) orogenic basalt–andesite–dacite–rhyolite series constitutes a linear arrangement on the conventional FeOt–MgO–(Na2O + K2O) diagram it is shown on numerical grounds that fractional crystallisation alone cannot account for the observed compositional trend—that, while accumulates are neither uncommon nor insubstantial in amount in modern arcs, they are olivine-rich, not magnetite-rich, and that numerical, petrological and field evidence indicates that the high oxidation/magnetite subtraction hypothesis of orogenic andesite genesis can no longer be seriously entertained. The partial melting hypothesis encounters contrary numerical evidence, closely analogous with that adduced for fractional crystallisation. Magma mixing eventually encounters the difficulty of the nature of the felsic material required for rhyolite formation, and assimilation of the preservation of primitive isotope ratios as lava evolution progresses. A fifth mechanism—volatile loss—may make a significant contribution to the relevant compositional change, perhaps in combination with one or more of the other proposed processes. Geochemical, petrological and field evidence pertaining to this is reviewed and examined, and it is postulated that indeed a combination of fractional crystallisation as the principal mechanism, volatile loss as a contributing factor, provides a plausible process. A case is made for an exploratory experimental investigation of the possibility of significant loss in a gas/vapour phase.