We present a robust measurement and analysis of the rest-frame ultraviolet (UV) luminosity functions at z = 4 to 8. We use deep Hubble Space Telescope imaging over the CANDELS/GOODS fields, the Hubble Ultra Deep Field and the Hubble Frontier Field deep parallel observations near the Abell 2744 and MACS J0416.1-2403 clusters. The combination of these surveys provides an effective volume of 0.6-1.2 ×10 6 Mpc 3 over this epoch, allowing us to perform a robust search for faint (M UV = −18) and bright (M UV < −21) highredshift galaxies. We select candidate galaxies using a well-tested photometric redshift technique with careful screening of contaminants, finding a sample of 7446 candidate galaxies at 3.5 < z < 8.5, with >1000 galaxies at z ≈ 6 -8. We measure both a stepwise luminosity function for candidate galaxies in our redshift samples, as well as a Schechter function, using a Markov Chain Monte Carlo analysis to measure robust uncertainties. At the faint end our UV luminosity functions agree with previous studies, yet we find a higher abundance of UV-bright candidate galaxies at z ≥ 6. Our best-fit value of the characteristic magnitude M * UV is consistent with −21 at z ≥ 5, different than that inferred based on previous trends at lower redshift, and brighter at ∼2σ significance than previous measures at z = 6 and 7 (Bouwens et al. 2007(Bouwens et al. , 2011b. At z = 8, a single powerlaw provides an equally good fit to the UV luminosity function, while at z = 6 and 7, an exponential cutoff at the bright end is moderately preferred. We compare our luminosity functions to semi-analytical models, and find that the lack of evolution in M * UV is consistent with models where the impact of dust attenuation on the bright end of the luminosity function decreases at higher redshift, though a decreasing impact of feedback may also be possible. We measure the evolution of the cosmic star-formation rate (SFR) density by integrating our observed luminosity functions to M UV = −17, correcting for dust attenuation, and find that the SFR density declines proportionally to (1+z) −4.3±0.5 at z > 4, consistent with observations at z ≥ 9. Our observed luminosity functions are consistent with a reionization history that starts at z 10, completes at z > 6, and reaches a midpoint (x HII = 0.5) at 6.7 < z < 9.4. Finally, using a constant cumulative number density selection and an empirically derived rising star-formation history, our observations predict that the abundance of bright z = 9 galaxies is likely higher than previous constraints, though consistent with recent estimates of bright z ∼ 10 galaxies.

We study the evolution of galaxy rest-frame ultraviolet (UV) colors in the epoch 4 z 8. We use new wide-field near-infrared data in the Great Observatories Origins Deep Survey -South field from the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey, Hubble Ultra Deep Field 2009 and Early Release Science programs to select galaxies via photometric redshift measurements. Our sample consists of 2812 candidate galaxies at z 3.5, including 113 at z ≃ 7-8. We fit the observed spectral energy distribution (SED) to a suite of synthetic stellar population models, and measure the value of the UV spectral slope (β) from the best-fit model spectrum. We run simulations to show that this measurement technique results in a smaller scatter on β than other methods, as well as a reduced number of galaxies with catastrophically incorrect β measurements (i.e., ∆β > 1). We find that the median value of β evolves significantly from −1.82 +0.00 −0.04 at z = 4, to −2.37 +0.26 −0.06 at z = 7. Additionally, we find that faint galaxies at z = 7 have β = −2.68 +0.39 −0.24 (∼ −2.4 after correcting for observational bias); this is redder than previous claims in the literature, and does not require "exotic" stellar populations (e.g., very-low metallicities or top-heavy initial mass functions) to explain their colors. This evolution can be explained by an increase in dust extinction, from low amounts at z = 7, to A V ∼ 0.5 mag at z = 4. The timescale for this increase is consistent with low-mass AGB stars forming the bulk of the dust. We find no significant (< 2σ) correlation between β and M UV when measuring M UV at a consistent restframe wavelength of 1500 Å. This is particularly true at bright magnitudes, though our results do show evidence for a weak correlation at faint magnitudes when galaxies in the HUDF are considered separately, hinting that dynamic range in sample luminosities may play a role. We do find a strong correlation between β and the stellar mass at all redshifts, in that more massive galaxies exhibit redder colors. The most massive galaxies in our sample have similarly red colors at each redshift, implying that dust can build up quickly in massive galaxies, and that feedback is likely removing dust from low-mass galaxies at z ≥ 7. Thus the stellar-massmetallicity relation, previously observed up to z ∼ 3, may extend out to z = 7 -8.

We explore star formation histories (SFHs) of galaxies based on the evolution of the star formation rate stellar mass relation (SFR-M * ). Using data from the FourStar Galaxy Evolution Survey (ZFOURGE) in combination with far-IR imaging from the Spitzer and Herschel observatories we measure the SFR-M * relation at 0.5 < z < 4. Similar to recent works we find that the average infrared spectral energy distributions of galaxies are roughly consistent with a single infrared template across a broad range of redshifts and stellar masses, with evidence for only weak deviations. We find that the SFR-M * relation is not consistent with a single power law of the form M SFR * µ a at any redshift; it has a power law slope of α ∼ 1 at low masses, and becomes shallower above a turnover mass (M 0 ) that ranges from 10 9.5 to 10 10.8 M e , with evidence that M 0 increases with redshift. We compare our measurements to results from state-of-the-art cosmological simulations, and find general agreement in the slope of the SFR-M * relation albeit with systematic offsets. We use the evolving SFR-M * sequence to generate SFHs, finding that typical SFRs of individual galaxies rise at early times and decline after reaching a peak. This peak occurs earlier for more massive galaxies. We integrate these SFHs to generate mass growth histories and compare to the implied mass growth from the evolution of the stellar mass function (SMF). We find that these two estimates are in broad qualitative agreement, but that there is room for improvement at a more detailed level. At early times the SFHs suggest mass growth rates that are as much as 10× higher than inferred from the SMF. However, at later times the SFHs under-predict the inferred evolution, as is expected in the case of additional growth due to mergers.