2000AJ....119..524C

We have carried out an independent quantitative evaluation of the galaxy component of the COSMOS/UKST southern sky object catalog (SSC) and the APM/UKST J Catalogue (APM). Using CCD observations, our results corroborate the accuracy of the photometry of both catalogs, which have an overall dispersion of about 0.2 mag in the range 17≤b_J≤21.5. The SSC presents externally calibrated galaxy magnitudes that follow a linear relation, while the APM instrumental magnitudes of galaxies, calibrated only internally by the use of stellar profiles, require second-order corrections. The completeness of both catalogs in a general field falls rapidly fainter than b_J_=20.0, being slightly better for APM. The 90% completeness level of the SSC is reached between b_J=19.5 and 20.0, while for APM this happens between b_J=20.5 and 21.0. Both SSC and APM are found to be less complete in a galaxy cluster field, where completeness reachs 90% in the ranges b_J=19.0-19.5 and b_J=19.5-20.0, respectively. Galaxies misclassified as stars in the SSC receive an incorrect magnitude because the stellar ones take saturation into account, besides using a different calibration curve. In both cases, the misclassified galaxies show a large diversity of colors, which range from typical colors of early types to those of blue star-forming galaxies. A possible explanation for this effect is that it results from the combination of low-sampling resolutions with properties of the image classifier for objects with characteristic sizes close to the instrumental resolution. We find that the overall contamination by stars misclassified as galaxies is <5% to b_J=20.5, as originally estimated for both catalogs. Although our results come from small areas of the sky, they are extracted from two different plates and are based on the comparison with two independent data sets. We conclude that both the SSC and APM can be a particularly valuable resource for extragalactic studies in the southern hemisphere once their limitations are taken into account.