Notes. Column 1: name of the primary protostar. Column 2: “Yes” or “No”: whether adding a disk-like Gaussian component improves the modeling at the respective frequency. Column 3: “Yes” if a disk-like component is detected and resolved by at least one of the two frequencies, “Marginally” if a disk-like component is detected at both frequencies but is only marginally resolved by our 231 GHz data, and “No” if a disk-like component is detected by at least one of the two frequencies but is not resolved by any. Column 4: disk radius. If a disk-like component is detected and resolved (at least one “Yes” in Cols. 2 and 3), it is the FWHM of the additional Gaussian component. When no disk component is detected at either frequency (two “No” in Cols. 2 and 3), we report the maximum disk-like component that can be added to the Plummer model. Column 5 reports either the flux of the disk-like component added to the envelope model at 231 GHz (if “Yes” in Col. 2 and “Yes” in Col. 4) or the maximum flux of the disk-like component that can be added to the best-fit envelope model (if “No” in Col. 2 and “No” in Col. 4), see text. Column 6: disk-to-envelope flux at 231 GHz (total envelope flux obtained from single-dish observations) for each source. ^(a)L1448-NB1 might be embedded within a ~200 au circumbinary structure together with L1448-NB2, see comments on this individual source in Appendix C.2, but the individual protostellar disk of NB1 cannot be stronger and larger than the remaining structure once this circumbinary structure and the surrounding Plummer envelope are subtracted. We here report these upper-limit values. ^(b)This value uses the single-dish envelope flux from Chini et al. (1997), but could be up to three times higher if the total envelope flux is closer to the lower value reported by Lefloch et al. (1998). See Appendix C.5 for further details. ^(c)The low S/N of binned visibilities for IRAM04191 and L1521F does not allow us to reliably determine the size from the disk-like component detected at 231 GHz in these two sources. Since all the error bars overlap at baselines > 300 kλ, we report this equivalent size as the maximum disk radius.