The ‘orogenic andesite’ puzzle of C. E. Tilley: 2—Some exploratory experiments and their possible implications

The five principal mechanisms proposed so far as possible contributors to the production of andesites from basalts in the ‘orogenic’ (i.e. volcanic arc) regime are fractional crystallisation/crystal subtraction, partial melting, magma mixing, crustal contamination and volatile loss. The process of volatile loss, while long recognised, has been given little credence as a significant contributor to modification of volcanic melt compositions. However, advances in the understanding of volcanic exhalative processes in the formation of many important ore deposits and their volcanic–chemical–sedimentary associates, i.e. in the capacity of the volatile phase to abstract and transport significant quantities of material from the volcanic melt, suggest that the observed major-element abundance relationships might be accounted for by a combination of fractional crystallisation and volcanic volatile loss. Using an experimental technique, an exploratory investigation confirms the findings of previous investigators with respect to Fe and, in addition, shows that the behaviour of Ca is remarkably similar to that of Fe both in pattern and amount. Mg is also markedly reduced. Reduction of the three elements, determined by analysis of the post-experimental residues, is of the order of 25–40%. Loss of this order by Fe, Ca and Mg in the volcanic environment would be more than sufficient to convert a basalt to an andesite or dacite. Analysis of the residues of five natural lavas, following runs of 21 days at 750°C and 6.0 kbar with an Ni–NiO buffer in water vapour, shows that the procedure has induced, concomitantly with the reduction in Fe, Ca and Mg, the conversion of a mafic basalt into a felsic basalt; an island arc basalt into an andesite; a basaltic andesite to an andesite; an andesite into a dacite; and a dacite into a more felsic variety. Within the restricted subsolidus conditions, it appears to open up the possibility that the ‘orogenic andesites’ may be derived from basalts by the above combination of processes. On this basis, it is proposed that the ‘hyperfusibles’ play a role in modifying melt composition not only by influencing the course of crystallisation and hence the identities and compositions of the crystals subtracted from the melt during fractional crystallisation but also, as an ancillary process, by inducing direct differential loss of materials in the vapour phase during and following eruption. If this were the case, the materials of the volcanic island arc andesite–dacite–rhyolite spectrum, including those of, and approximating to, the calc-alkaline series, may be seen as substantially superficial, localised, ‘scar tissue’ induced by the progressive degassing—i.e. high-temperature gaseous stripping—of basalts already modified by fractional crystallisation and subtraction.