Boron contents were measured in predominantly mafic Quaternary lavas from the Central American Volcanic Arc to evaluate along-strike variations in subduction processes. Despite the significant range in B concentrations (∼2 to 46 ppm), B/La ratios vary in a systematic fashion along the arc: Higher values (>1) are typical between Guatemala and northern Costa Rica, whereas low values (most <0.5) typify central Costa Rica and western Panama. Because B/La is highly correlated with 10 Be/Be (r 2 = 0.94, excluding one sample), B/La may be a useful indicator of subduction contributions to the magma sources. If enrichments of both B and 10 Be are proportional to the flux of subducted sediment, along-strike variations in B/La suggest at least a twofold variation in this flux with maximum values below western Nicaragua and minimum values below Costa Rica and western Panama where the Cocos Ridge is being subducted. These data may also reflect significant differences in thermal state of the downgoing slab, which in turn differentially affect release patterns of fluids and fluid-mobile trace elements and possibly melting processes beneath different parts of the arc. The following scenario is suggested to explain the geochemical results. Steep subduction of cold slab beneath the northwestern arc favors more efficient subduction of fluid components to depths beneath the volcanic front. Their release provides fluid-mobile elements to the overlying mantle, which upon melting produces calcalkalic magmas. Shallow subduction of warmer slab beneath the southeastern arc favors shallow release of fluids and limits fluid-related metasomatism of sub-arc mantle beneath the volcanic front. Under such conditions, B/La and Ba/La ratios in the sub-arc mantle vary little from values seen in oceanic island basalts. Magmas in this part of the arc nevertheless display the highest La/Yb and lowest Ba/La and B/La ratios, which are consistent with prior light rare-earth-element enrichment in the source, significantly lower degrees of melting, or a combination thereof. Because some of the largest volcanoes occur in Costa Rica, and magma flux there is nearly an order of magnitude higher than elsewhere in the arc, source enrichment is considered to be the more plausible explanation. It is proposed that Quaternary magma production below Costa Rica involved lithospheric sources containing trapped or “stored” melt components, but this enrichment process is unlikely to have involved arc magmas or subduction fluids because we see no B-enrichment.