Mid-infrared photometry provides a robust technique for identifying active galaxies. While the ultraviolet to mid-infrared (k P 5 m) continuum of stellar populations is dominated by the composite blackbody curve and peaks at approximately 1.6 m, the ultraviolet to mid-infrared continuum of active galactic nuclei (AGNs) is dominated by a power law. Consequently, with a sufficient wavelength baseline, one can easily distinguish AGNs from stellar populations. Mirroring the tendency of AGNs to be bluer than galaxies in the ultraviolet, where galaxies (and stars) sample the blue, rising portion of stellar spectra, AGNs tend to be redder than galaxies in the mid-infrared, where galaxies sample the red, falling portion of the stellar spectra. We report on Spitzer Space Telescope mid-infrared colors, derived from the IRAC Shallow Survey, of nearly 10,000 spectroscopically identified sources from the AGN and Galaxy Evolution Survey. On the basis of this spectroscopic sample, we find that simple mid-infrared color criteria provide remarkably robust separation of active galaxies from normal galaxies and Galactic stars, with over 80% completeness and less than 20% contamination. Considering only broad-lined AGNs, these mid-infrared color criteria identify over 90% of spectroscopically identified quasars and Seyfert 1 galaxies. Applying these color criteria to the full imaging data set, we discuss the implied surface density of AGNs and find evidence for a large population of optically obscured active galaxies.

We measure the evolution of the stellar mass function (SMF) from z = 0 − 1 using multi-wavelength imaging and spectroscopic redshifts from the PRism MUlti-object Survey (PRIMUS) and the Sloan Digital Sky Survey (SDSS). From PRIMUS we construct an i < 23 flux-limited sample of ∼ 40, 000 galaxies at z = 0.2 − 1.0 over five fields totaling ≈ 5.5 deg 2 , and from the SDSS we select ∼ 170, 000 galaxies at z = 0.01−0.2 that we analyze consistently with respect to PRIMUS to minimize systematic errors in our evolutionary measurements. We find that the SMF of all galaxies evolves relatively little since z = 1, although we do find evidence for mass assembly downsizing; we measure a ≈ 30% increase in the number density of ∼ 10 10 M ⊙ galaxies since z ≈ 0.6, and a 10% change in the number density of all 10 11 M ⊙ galaxies since z ≈ 1. Dividing the sample into star-forming and quiescent using an evolving cut in specific star-formation rate, we find that the number density of ∼ 10 10 M ⊙ star-forming galaxies stays relatively constant since z ≈ 0.6, whereas the space-density of 10 11 M ⊙ star-forming galaxies decreases by ≈ 50% between z ≈ 1 and z ≈ 0. Meanwhile, the number density of ∼ 10 10 M ⊙ quiescent galaxies increases steeply towards low redshift, by a factor of ∼ 2 − 3 since z ≈ 0.6, while the number of massive quiescent galaxies remains approximately constant since z ≈ 1. These results suggest that the rate at which star-forming galaxies are quenched increases with decreasing stellar mass, but that the bulk of the stellar mass buildup within the quiescent population occurs around ∼ 10 10.8 M ⊙ . In addition, we conclude that mergers do not appear to be a dominant channel for the stellar mass buildup of galaxies at z < 1, even among massive ( 10 11 M ⊙ ) quiescent galaxies.

We explore the connection between different classes of active galactic nuclei (AGNs) and the evolution of their host galaxies, by deriving host galaxy properties, clustering, and Eddington ratios of AGNs selected in the radio, X-ray, and infrared (IR) wavebands. We study a sample of 585 AGNs at 0.25 < z < 0.8 using redshifts from the AGN and Galaxy Evolution Survey (AGES). We select AGNs with observations in the radio at 1.4 GHz from the Westerbork Synthesis Radio Telescope, X-rays from the Chandra XBoötes Survey, and mid-IR from the Spitzer IRAC Shallow Survey. The radio, X-ray, and IR AGN samples show modest overlap, indicating that to the flux limits of the survey, they represent largely distinct classes of AGNs. We derive host galaxy colors and luminosities, as well as Eddington ratios, for obscured or optically faint AGNs. We also measure the two-point cross-correlation between AGNs and galaxies on scales of 0.3-10 h −1 Mpc, and derive typical dark matter halo masses. We find that: (1) radio AGNs are mainly found in luminous red sequence galaxies, are strongly clustered (with M halo ∼ 3 × 10 13 h −1 M ⊙ ), and have very low Eddington ratios (λ 10 −3 ); (2) X-rayselected AGNs are preferentially found in galaxies that lie in the "green valley" of color-magnitude space and are clustered similar to typical AGES galaxies (M halo ∼ 10 13 h −1 M ⊙ ), with 10 −3 λ 1; (3) IR AGNs reside in slightly bluer, slightly less luminous galaxies than X-ray AGNs, are weakly clustered (M halo 10 12 h −1 M ⊙ ), and have λ > 10 −2 . We interpret these results in terms of a simple model of AGN and galaxy evolution, whereby a "quasar" phase and the growth of the stellar bulge occurs when a galaxy's dark matter halo reaches a critical mass between ∼10 12 and 10 13 M ⊙ . After this event, star formation ceases and AGN accretion shifts from radiatively efficient (optical-and IR-bright) to radiatively inefficient (optically faint, radio-bright) modes.

We study the distribution of Eddington luminosity ratios, L bol /L Edd , of active galactic nuclei (AGNs) discovered in the AGN and Galaxy Evolution Survey (AGES). We combine H, Mg ii, and C iv line widths with continuum luminosities to estimate black hole ( BH ) masses in 407 AGNs, covering the redshift range z $ 0:3 Y 4 and the bolometric luminosity range L bol $ 10 45 Y10 47 ergs s À1. The sample consists of X-ray or mid-infrared (24 m) point sources with optical magnitude R 21:5 mag and optical emission-line spectra characteristic of AGNs. For the range of luminosity and redshift probed by AGES, the distribution of estimated Eddington ratios is well described as lognormal, with a peak at L bol /L Edd ' 1/4 and a dispersion of 0.3 dex. Since additional sources of scatter are minimal, this dispersion must account for contributions from the scatter between estimated and true BH mass and the scatter between estimated and true bolometric luminosity. Therefore, we conclude that (1) neither of these sources of error can contribute more than $0.3 dex rms, and (2) the true Eddington ratios of optically luminous AGNs are even more sharply peaked. Because the mass estimation errors must be smaller than $0.3 dex, we can also investigate the distribution of Eddington ratios at fixed BH mass. We show for the first time that the distribution of Eddington ratios at fixed BH mass is peaked, and that the dearth of AGNs at a factor of $10 below Eddington is real and not an artifact of sample selection. These results provide strong evidence that supermassive BHs gain most of their mass while radiating close to the Eddington limit, and they suggest that the fueling rates in luminous AGNs are ultimately determined by BH self-regulation of the accretion flow rather than galactic-scale dynamical disturbances.

We present evidence that the incidence of active galactic nuclei (AGNs) and the distribution of their accretion rates do not depend on the stellar masses of their host galaxies, contrary to previous studies. We use hard (2−10 keV) X-ray data from three extragalactic fields (XMM-LSS, COSMOS and ELAIS-S1) with redshifts from the Prism Multi-object Survey to identify 242 AGNs with L 2−10 keV = 10 42−44 erg s −1 within a parent sample of ∼25,000 galaxies at 0.2 < z < 1.0 over ∼ 3.4 deg 2 and to i ∼ 23. We find that although the fraction of galaxies hosting an AGN at fixed X-ray luminosity rises strongly with stellar mass, the distribution of X-ray luminosities is independent of mass. Furthermore, we show that the probability that a galaxy will host an AGN can be defined by a universal Eddington ratio distribution that is independent of the host galaxy stellar mass and has a power-law shape with slope −0.65. These results demonstrate that AGNs are prevalent at all stellar masses in the range 9.5 < log M * /M ⊙ < 12 and that the same physical processes regulate AGN activity in all galaxies in this stellar mass range. While a higher AGN fraction may be observed in massive galaxies, this is a selection effect related to the underlying Eddington ratio distribution. We also find that the AGN fraction drops rapidly between z ∼ 1 and the present day and is moderately enhanced (factor∼ 2) in galaxies with blue or green optical colors. Consequently, while AGN activity and star formation appear to be globally correlated, we do not find evidence that the presence of an AGN is related to the quenching of star formation or the color transformation of galaxies.