Differences with temperatures determined from infrared flux measurements with uncertainties added in quadrature. The two different IRFM methods had an offset of ∼35 K from each other, which when combined led to an offset 7 K hotter than our temperatures with an rms scatter of 72 K. For clarity, the plot above has removed 16 points from Gonzalez Hernandez & Bonifacio (2009) with uncertainties larger than 3%, which leaves the offset unchanged and the scatter at a nearly identical 71 K. Separating the two shows that the temperatures of Casagrande et al. (2010) are 11 K hotter than ours, and those of Gonzalez Hernandez & Bonifacio (2009) are 24 K cooler.