Deep Counts of Submillimeter Galaxies

We present the counts of luminous submillimeter (sub-mm) galaxies from an analysis of our survey of the distant Universe through lensing clusters, which uses massive cluster lenses with well-constrained mass models to magnify the background sky. This both increases the sensitivity and reduces the effects of source confusion. Accurate lens models are used to correct the observed sub-mm source counts. The associated uncertainties do not dominate the final errors. We derive counts by two independent methods: a direct inversion of the observed sources, corrected individually for lens amplification; and a Monte-Carlo simulation of our observations using a parametric model for the background counts, which is folded through the lens models and incompleteness estimates to determine best-fitting values of the count parameters. Both methods agree well and confirm the robustness of our analysis. Detections that are optically identified with galaxies in the lensing clusters are removed prior to our analysis, and the results are insensitive both to the details of the correction and to the redshift distribution of the detections. We present 850-micron counts between 0. 5 and 8 mJy. The most accurate count is determined at 1 mJy: 7900 +- 3000 deg^-2, including both Poisson and systematic errors. These are the deepest sub-mm counts published, and are not subject to source confusion because the detected galaxies are separated and magnified by the lens. Down to the 0. 5-mJy limit, the resolved 850-micron background radiation intensity is (0. 5 +- 0. 2) nW m^-2 sr^-1, comparable to the current COBE estimate of the background, and so the bulk of the 850-micron background radiation originated in distant ultraluminous galaxies.