Estimate of Φ0(MG, (B − R)), analogous to that in Figure 7, but for four different choices of ﹩{\overline{{\rm{S}}/{\rm{N}}}}_{\varpi ,\min }﹩: 5, 10, 40, and 80. Despite the sample size differences of a factor of 5, the inferred Φ0(MG, (B − R)) are mutually consistent. High values of ﹩{\overline{{\rm{S}}/{\rm{N}}}}_{\varpi ,\min }﹩ lead to small samples but better detail in the high-density structure of Φ0(MG, (B − R)); low values of ﹩{\overline{{\rm{S}}/{\rm{N}}}}_{\varpi ,\min }﹩ lead to larger samples and a better sampling of the (MG, (B − R)) plane, but at reduced resolution, as the uncertainties in MG become noticeable.