The long-standing assumption that calc-alkaline magmatism requires contemporaneous subduction is critically reviewed. Two of many widely spread fields of Tertiary calc-alkaline magmatism in the Pacific Northwest, one Eocene and one Miocene, provide clear examples of the association between calc-alkaline magmatism and lithospheric extension. We suggest that the calc-alkaline magmatism of these fields is better explained by decompression melting of the subcontinental lithospheric crust and mantle, enriched by previous subduction events, than by contemporaneous Eocene subduction. This model proposes that extension will cause calc-alkaline-dominated magmatism in areas of continental crust with a history of previous subduction, as along the western margin of North America. Currently popular models for the tectonic/magmatic evolution of the Basin and Range province, in which subduction-driven calc-alkaline magmatism is followed by extension-related bimodal magmatism, may need modification; much more of this magmatism may result from extension. In the Pacific Northwest, we suggest that subduction beneath the Idaho batholith ceased at the end of the Cretaceous, blocked by the accreted terraces, to be reactivated in its present position beneath the Cascades during the Eocene. The various isolated Eocene volcanic fields spread across the Pacific Northwest were due to lithospheric extension, not to the presence of a subducting plate immediately below. ¿ American Geophysical Union 1995