Notes. Entries are at ℓ = 3000 in μK² and are given at the effective band centre for each component. Uncertainties on the “power-law” model are statistical errors propagated from uncertainties in the Mocanu et al. (2013) source-count data. Priors on the dust component have formally been calculated only for the Hillipop mask, but they are repeated for the other masks, for which they are accurate to better than 1%. The results from different codes, to which these predictions should be compared, use TTℓ> 50 data with a prior of τ = 0.07 ± 0.02. For Mspec about 90% of the dusty contribution is cleaned out at the map level, hence the measured values above are in some cases far less than prior value.