It is widely accepted that chemical compositions of volcanic arc lavas reflect the imprints of subducted components on their primary mantle sources, and that the extent of material transfer in subduction zones [SZs] depends on such factors as subduction geometry, thermal state of the slab, the type and amount of sediments subducted, and age and degree of alteration of the underthrust oceanic crust. To address the effects of progressive heating of a subducting slab and the release of its volatile constituents it is proposed to analyze primitive basaltic lavas from a cross-arc transect of the Quaternary southern Washington Cascades. This SZ is one of the warmest known, and this arc represents an important end member example where slab-derived contributions are minimal and characteristics of the mantle wedge may be more easily discerned. On the other hand, previously measured cross-arc variations in certain geochemical parameters (e.g., B/Zr, ?11B, Pb isotopes) indicate small but significant slab inputs below the forearc and volcanic front ranging to negligible contributions below the backarc region. This contrast affords an opportunity for quantitative definition of slab contributions to a well-defined wedge composition. In the proposed study, compositions of melt inclusions (MIs) in olivine phenocrysts will determined; such inclusions are more likely to retain original proportions of volatile components than extruded lavas, and thus more representative compositions of primitive magmas. Analyses will include major and trace element compositions, and volatile (H2O, CO2) contents. Work will focus on a small number of representative lavas, from locations spanning the width of the arc, that have been previously characterized for major and trace element and Sr-Nd-Pb isotopic compositions. This comprehensive body of data will allow us to address the relative retentivity of the analyzed elements in the Cascadia slab, and to evaluate effects on magma compositions due to geochemical fractionations associated with magma ascent through mantle wedge and crust. Our goal is to understand and quantify the spatial variability of volatile and fluid-mobile constituents in primitive magmas, and to evaluate the hypothesis that temperature variations in the slab (as calculated from models of conductive heating and mantle convection) largely dictate the composition and magnitudes of slab-derived contributions to arc magma sources.