Differences with temperatures determined from spectroscopic measurements with uncertainties added in quadrature. For catalogs without reported errors, we assumed an average error of 50 K. Our temperatures are 22 K cooler (blue dashed line) with an rms scatter of 67 K, implying a mutual precision better than 1%. The scatter again increases at higher effective temperatures. We expect this behavior in this context as our new analysis procedure has larger differences in ﹩\mathrm{log}g﹩ (and hence ﹩{T}_{\mathrm{eff}}﹩) for hotter stars.