Key points are not available for this paper at this time.
We used the VLA to make deep images of two 7' X 7' fields at 8. 44 GHz with 10" resolution. With an rms noise of 3. 2 and 5. 1 microJy, respectively, in the two fields, we compiled a catalog of 82 sources. From the complete sample of 20 sources with S >= 14. 5 microJy, the differential 8. 44 GHz source count is dN (S) /dS = (-4. 6 +/- 0. 7) x S^-2. 3+/-0. 2^ Jy^-1^ sr^-1^ in the range 14. 5-1000 mJy. Analysis of statistical image fluctuations from weak sources (Fomalont et al. 1993) suggests that this slope remains unchanged at γ = 2. 310+/-0. 2 down to ~4 microJy. The normalized differential 8. 44 GHz counts are similar to those at 1. 41 and 4. 86 GHz. All show a similarly steep submillijansky slope, which is only somewhat flatter than that expected for a nonevolving Euclidean population (γ = 2. 5). Microjansky radio sources at 4. 86 GHz have been identified with faint blue galaxies (18 = 5", and the median is THETAₘed_~2. 6"+/-1. 4", or <~ 5-40 kpc at the expected median redshift. The extended steep-spectrum sources suggest synchrotron emission in distant galactic disks, while the extended flat-spectrum sources may indicate thermal bremsstrahlung from large-scale star formation, both occasionally with opaque radio cores. The estimated 31. 5 GHz sky brightness from nanojansky to jansky levels is <~ 36 microK (3 σ). Even if weak radio sources cluster on scales of degrees as faint galaxies do, their anisotropic contribution to the COBE DMR experiment (with 7ᵈeg^ FWHM beam) would not exceed ~ 1. 2 microK.
Windhorst et al. (Mon,) studied this question.