The circular pattern of anisotropic fast-axis orientations of split SKS arrivals observed in the western U.S. cannot be attributed reasonably to either preexisting lithospheric fabric or to asthenospheric strain related to global-scale plate motion. A plume origin for this pattern accounts more successfully for the anisotropy field, but little evidence exists for an active plume beneath central Nevada. We suggest that mantle flow around the edge of the sinking Gorda–Juan de Fuca slab is responsible for creating the observed anisotropy. Seismic images and kinematic reconstructions of Gorda–Juan de Fuca plate subduction have the southern edge of this plate extending from the Mendocino triple junction to beneath central Nevada, and flow models of narrow subducted slabs produce a strong toroidal flow field around the edge of the slab, consistent with the observed pattern of anisotropy. This flow may enhance uplift, extension, and magmatism of the northern Basin and Range while inhibiting extension of the southern Basin and Range.