ar X iv : a st ro - p h / 05 01 31 4 v 1 1 5 Ja n 20 05 UV LUMINOSITY FUNCTION AT Z ∼ 4 , 3 , AND 2

We use very deep (Rlim=27) UnGRI imaging to study the evolution of the faint end of the UV-selected galaxy luminosity function from z∼4 to z∼2. We find that the number of sub-L∗ galaxies increases from z∼4 to z∼3 while the number of bright ones appears to remain constant. We find no evidence for continued evolution to lower redshift, z∼2. If real, this differential evolution of the luminosity function suggests that differentially comparing key diagnostics of dust, stellar populations, etc. as a function of z and L may let us isolate the key mechanisms that drive galaxy evolution at high redshift and we describe several such studies currently underway. 1. The Keck Deep Fields The shape of the galaxy luminosity function bears the imprint of galaxy formation and evolutionary processes and suggests that galaxies below L differ substantially from those above it in more than just luminosity. Our understanding of galaxy formation may profit from studying the evolution of not just the bright but also the faint component of the galaxy population at high redshift. To study the evolution of the galaxy luminosity function at high redshift, we have carried out a very deep imaging survey that uses the very same UnGRI filter set and color-color selection technique as used in the work of Steidel et al. (1999, 2003, 2004), but that reaches to R=27 — 1.5 magnitudes deeper and significantly below L at z=2–4 (Sawicki & Thompson, 2005). These Keck Deep Fields (KDF) were obtained with the LRIS imaging spectrograph on Keck I and represent 71 hours of integration split into five fields that are