A surprising aspect of subduction zones is the juxtaposition of the heat sink of the cold underthrusting plate and the strong heat source of the adjacent volcanic arc. Where does the heat for the arc come from? The widely accepted and applied model is that of �corner flow�. Heat is pulled in from the deeper and landward hot asthenosphere. The underthrusting plate traction drags down the overlying mantle corner and the overlying mantle is pulled into the corner to replace it. A characteristic of most such models that the geotherms shallow approaching the arc and deepen landward in the backarc. Currie et al. (2006) showed that this characteristic is inconsistent with the observed nearly constant heat flow across the Cordillera and that an alternate model that does agree is vigorous small-scale convection that gives nearly constant (adiabatic) temperatures in the asthenosphere. A number of small-scale convection models have subsequently been proposed. We now have two new strong constraints that indicate a very constant depth to and temperature at the base of the lithosphere, which support the latter model. The first is remarkably constant seismic receiver function depths to the lithosphere-asthenosphere boundary (LAB) across the Cordillera, at ~65 km. The second is the depth of geochemical equilibration of recent backarc volcanic lavas, taken to be at the LAB, again averaging ~65 km and constant 1350C. This depth is proposed to be controlled by the spinel to garnet peridotite phase transition that occurs at this pressure (depth).