| We use very deep optical and near-infrared imaging surveys to set constraints on galaxy evolution models, based on the numbers, colors, and morphologies of infrared-excess galaxies that are found in the field. We concentrate on a curious population of extremely faint (K > 20) infrared-excess galaxies whose blue-optical colors are not consistent with the expectations of any passive evolution models. These "faint red-outlier galaxies" (FROGs) are distinct from the redder and brighter "extremely red objects" (EROs; R-K≃6 ) . In a concerted effort to identify a good sample of FROG s and to study their properties, we used Keck/NIRC to image several independent high-galactic latitude fields. Much of the analysis presented here is drawn from a very deep 3.24 arcmin2 K-band ( Klim≈24 at 3sigma) mosaic in the deep Westphal HST/WFPC2 pointing of the Groth Survey Strip, for which F606W and F814W data were publically available. The surface density of FROGs is found to be ≈ 3.3 +/- 1 arcmin-2, more than ten times that of EROs. Whereas reliable photometric redshifts are not forthcoming without the development of more relevant models, the colors are broadly consistent with the redshift range 1.2 < z < 2.3. If placed at z ≈ 1.5, FROGs occur at space densities of about 10% of the local f* space density of K-selected galaxies.; To map wavelength-dependent morphologies of two FROGs, we observed a portion of our main survey field with HST/NICMOS imaging through the F160W (1.6 mum) filter. The target was resolved into two r≃0''.6 objects with similar colors, separated by ≈0'' .7. This is suggestive of old and dynamically-relaxed systems. The colors of FROGs are not satisfactorily fit by dust-reddened Bruzual-Charlot models at any redshift and for a broad range of assumed star formation histories. The best possible fits are consistent with very large amounts of reddening, E(B -- V) ≈ 1.; If the infrared-excess in EROs and FROGs is taken to be entirely due to the effects of dust, then we can estimate the upper limit of their contribution to the far infrared background. We use a range of star formation histories at different ages and representative reddening values. The integral contribution of EROs and FROGs is calculated to be no greater than ∼ 1 -- 10 nW m-2 sr-1. At lambda ∼ 100mum, 1 nW m-2 sr-1 corresponds to ∼ 10% of the total measured infrared background. |
