Orogenic andesite magmas probably evolve at 1100 to 900°C, have between 1 and 5 wt. % H 2O and f O 2 above the NiNiO buffer during phenocryst precipitation and may be saturated with a second fluid phase prior to eruption. Consequently, many solid/liquid trace element partition coefficients appropriate to andesites are higher than for basalts, perhaps due to the presence of fewer octahedrally coordinated sites in andesite liquids as well as to lower temperatures. Simple quantitative interpretation of trace element concentration data for orogenic andesites in general or for most specific suites is not possible using available partition coefficients. Fractional crystallization models involving magnetite or amphibole or both seem less at odds with available data than do models invoking equilibration with subducted oceanic crust or overlying peridotite. Critical choices between simple models, or identification of supplementary or more complex processes, require a more rigorous data base than presently available.