An Empirical Limit on Extremely High Redshift Galaxies

We apply the Lyman absorption signature to search for galaxies at redshifts z \~ 6 - 17 using optical and infrared images of the Hubble Deep Field. The infrared images are sensitive to a point source 5 sigma detection threshold of AB(22,000) = 23.8, which adopting plausible assumptions to relate rest-frame ultraviolet flux densities to unobscured star formation rates is easily sufficient to detect the star formation rates expected for massive elliptical galaxy formation to quite high redshifts. For q_0 = 0.5, the infrared images are sensitive to an unobscured star formation rate of 100 h^-2 solar masses per year to redshifts as large as z = 17, and for q_0 = 0, the infrared images are sensitive to an unobscured star formation rate of 300 h^-2 solar masses per year to redshifts as large as z = 14. The primary result of the analysis is that only one extremely high redshift galaxy candidate is identified at the 5 sigma level of significance (and four at the 4 sigma level). This implies a strict upper limit to the surface density of extremely high redshift galaxies of < 1.5 arcmin^-2 to a limiting magnitude threshold AB(22,000) = 23.8. This also implies a strict upper limit to the volume density of extremely high redshift galaxies if (and only if) such galaxies are not highly obscured by dust.