Is there a Population of Weak Radio Sources that Contributes to the Excess Sky Temperature Observed by Arcade 2?

The ARCADE 2 balloon experiment has reported a 3 GHz sky brightness that is 65 +/- 8 mK in excess of the CMB which is much greater than the ~13 mK that can be explained by the known population of radio sources. We have used the Karl Jansky Very Large Array to image one primary beam at 3 GHz with 8 arcsecond resolution and 1.0 μJy per beam rms noise near the pointing center. Our image is confusion limited with an rms confusion noise of ~1.2 μJy per beam. Modeling the distribution of confusion noise suggests that any excess background must be very smooth, so any hypothetical population of discrete sources that can account for the ARCADE result must be very numerous and weaker than ~ 30 nanoJy. If the ARCADE results are correct, the number of weak radio sources would be several orders of magnitude greater than the faintest galaxies in the Hubble Ultra Deep Field.