Search

In the widely accepted ‘unified model’1solution of the classification puzzle of active galactic nuclei, the orientation of a dusty accretion torus around the central black hole dominates their appearance. In ‘type-1’ systems, the bright nucleus is visible at the centre of a face-on torus. In ‘type-2’ systems the thick, nearly edge-on torus hides the central engine. Later studies suggested evolutionary effects2and added dusty clumps and polar winds3but left the basic picture intact. However, recent high-resolution images4of the archetypal type-2 galaxy NGC 10685,6, suggested a more radical revision. The images displayed a ring-like emission feature that was proposed to be hot dust surrounding the black hole at the radius where the radiation from the central engine evaporates the dust. That ring is too thin and too far tilted from edge-on to hide the central engine, and ad hoc foreground extinction is needed to explain the type-2 classification. These images quickly generated reinterpretations of the dichotomy between types 1 and 27,8. Here we present new multi-band mid-infrared images of NGC 1068 that detail the dust temperature distribution and reaffirm the original model. Combined with radio data (J.F.G. and C.M.V.I., manuscript in preparation), our maps locate the central engine that is below the previously reported ring and obscured by a thick, nearly edge-on disk, as predicted by the unified model. We also identify emission from polar flows and absorbing dust that is mineralogically distinct from that towards the Milky Way centre.

We explore the rate and impact of galaxy mergers on the massive galaxy population using the amplitude of the two-point correlation function on small scales for M * > 5 x 1010 M galaxies from the COSMOS and COMBO-17 surveys. Using a pair fraction derived from the Sloan Digital Sky Survey as a low-redshift benchmark, the large survey area at intermediate redshifts allows us to determine the evolution of the close-pair fraction with unprecedented accuracy for a mass-selected sample: we find that the fraction of galaxies more massive than 5 x 1010 M in pairs separated by less than 30 kpc in three-dimensional space evolves as F(z) = (0.0130 +- 0.0019) x (1 + z)1.21+- 0.25 between z = 0 and z = 1.2. Assuming a merger timescale of 0.5 Gyr, the inferred merger rate is such that galaxies with mass in excess of 1011 M have undergone, on average, 0.5 (0.7) mergers involving progenitor galaxies both more massive than 5 x 1010 M since z = 0.6 (1.2). We also study the number density evolution of massive red sequence galaxies using published luminosity functions and constraints on the M/L B evolution from the fundamental plane. Moreover, we demonstrate that the measured merger rate of massive galaxies is sufficient to explain this observed number density evolution in massive red sequence galaxies since z = 1.

We present a study of galaxy mergers and the influence of environment in the Abell 901/902 supercluster at z[?] 0.165, based on 893 bright (R Vega[?] 24) intermediate-mass (M * [?] 109 M ) galaxies. We use HST ACS F606W data from the Space Telescope A901/902 Galaxy Evolution Survey, COMBO-17, Spitzer 24 mm, and XMM-Newton X-ray data. Our analysis utilizes both a physically driven visual classification system and quantitative CAS parameters to identify systems which show evidence of a recent or ongoing merger of mass ratio >1/10 (i.e., major and minor mergers). Our results are (1) after visual classification and minimizing the contamination from false projection pairs, we find that the merger fraction f merge is 0.023 +- 0.007. The estimated fractions of likely major mergers, likely minor mergers, and ambiguous cases are 0.01 +- 0.004, 0.006 +- 0.003, and 0.007 +- 0.003, respectively. (2) All the mergers lie outside the cluster core of radius R< 0.25 Mpc: the lack of mergers in the core is likely due to the large galaxy velocity dispersion in the core. The mergers, instead, populate the region (0.25 Mpc

Both observations and simulations show that major tidal interactions or mergers between gas-rich galaxies can lead to intense bursts of star formation. Yet, the average enhancement in star formation rate (SFR) in major mergers and the contribution of such events to the cosmic SFR are not well estimated. Here we use photometric redshifts, stellar masses, and UV SFRs from COMBO-17, 24 mm SFRs from Spitzer, and morphologies from two deep Hubble Space Telescope (HST) cosmological survey fields (ECDFS/GEMS and A901/STAGES) to study the enhancement in SFR as a function of projected galaxy separation. We apply two-point projected correlation function techniques, which we augment with morphologically selected very close pairs (separation <2'') and merger remnants from the HST imaging. Our analysis confirms that the most intensely star-forming systems are indeed interacting or merging. Yet, for massive (M * [?] 1010 M ) star-forming galaxies at 0.4 < z < 0.8, we find that the SFRs of galaxies undergoing a major interaction (mass ratios [?] 1:4 and separations [?] 40 kpc) are only 1.80 +- 0.30 times higher than the SFRs of non-interacting galaxies when averaged over all interactions and all stages of the interaction, in good agreement with other observational works. Our results also agree with hydrodynamical simulations of galaxy interactions, which produce some mergers with large bursts of star formation on [?]100 Myr timescales, but only a modest SFR enhancement when averaged over the entire merger timescale. We demonstrate that these results imply that only [?]10% of star formation at 0.4 [?] z [?] 0.8 is triggered directly by major mergers and interactions; these events are not important factors in the build-up of stellar mass since z = 1.

We present a study of bar and host disk evolution in a dense cluster environment, based on a sample of [?]800 bright (M V [?] -18) galaxies in the Abell 901/2 supercluster at z[?] 0.165. We use Hubble Space Telescope (HST) Advanced Camera for Surveys (ACS) F606W imaging from the STAGES survey, and data from Spitzer, XMM-Newton, and COMBO-17. We identify and characterize bars through ellipse-fitting, and other morphological features through visual classification. We find the following results. (1) To define the optical fraction of barred disk galaxies, we explore three commonly used methods for selecting disk galaxies. We find 625, 485, and 353 disk galaxies, respectively, via visual classification, a single component Sersic cut (n [?] 2.5), and a blue-cloud cut. In cluster environments, the latter two methods suffer from serious limitations, and miss 31% and 51%, respectively, of visually identified disks, particularly the many red, bulge-dominated disk galaxies in clusters. (2) For moderately inclined disks, the three methods of disk selection, however, yield a similar global optical bar fraction (f bar-opt) of 34%+10% -3% (115/340), 31%+10% -3% (58/189), and 30%+10% -3% (72/241), respectively. (3) We explore f bar-opt as a function of host galaxy properties and find that it rises in brighter galaxies and those which appear to have no significant bulge component. Within a given absolute magnitude bin, f bar-opt is higher in visually selected disk galaxies that have no bulge as opposed to those with bulges. Conversely, for a given visual morphological class, f bar-opt rises at higher luminosities. Both results are similar to trends found in the field. (4) For bright early-types, as well as faint late-type systems with no evident bulge, the optical bar fraction in the Abell 901/2 clusters is comparable within a factor of 1.1-1.4 to that of field galaxies at lower redshifts (z < 0.04). (5) Between the core and the virial radius of the cluster (R[?] 0.25-1.2 Mpc) at intermediate environmental densities (log(S10) [?] 1.7-2.3), the optical bar fraction does not appear to depend strongly on the local environment density tracers (k, S10, and intracluster medium (ICM) density), and varies at most by a factor of [?]1.3. Inside the cluster core, we are limited by number statistics, projection effects, and different trends from different indicators, but overall f bar-opt does not show evidence for a variation larger than a factor of 1.5. We discuss the implications of our results for the evolution of bars and disks in dense environments.

We perform a comprehensive estimate of the frequency of galaxy mergers and their impact on star formation over z[?] 0.24-0.80 (lookback time T b[?] 3-7 Gyr) using [?]3600 (M[?] 1 x 109 M ) galaxies with GEMS Hubble Space Telescope, COMBO-17, and Spitzer data. Our results are as follows. (1) Among [?]790 high-mass (M[?] 2.5 x 1010 M ) galaxies, the visually based merger fraction over z[?] 0.24-0.80, ranges from 9% +- 5% to 8% +- 2%. Lower limits on the major merger and minor merger fraction over this interval range from 1.1% to 3.5%, and 3.6% to 7.5%, respectively. This is the first, albeit approximate, empirical estimate of the frequency of minor mergers over the last 7 Gyr. Assuming a visibility timescale of [?]0.5 Gyr, it follows that over T b[?] 3-7 Gyr, [?]68% of high-mass systems have undergone a merger of mass ratio >1/10, with [?]16%, 45%, and 7% of these corresponding respectively to major, minor, and ambiguous "major or minor" mergers. The average merger rate is [?] a few x10-4 galaxies Gyr-1 Mpc-3. Among [?]2840 blue-cloud galaxies of mass M[?] 1.0 x 109 M , similar results hold. (2) We compare the empirical merger fraction and merger rate for high-mass galaxies to three L cold dark matter-based models: halo occupation distribution models, semi-analytic models, and hydrodynamic SPH simulations. We find qualitative agreement between observations and models such that the (major+minor) merger fraction or rate from different models bracket the observations, and show a factor of 5 dispersion. Near-future improvements can now start to rule out certain merger scenarios. (3) Among [?]3698 M[?] 1.0 x 109 M galaxies, we find that the mean star formation rate (SFR) of visibly merging systems is only modestly enhanced compared to non-interacting galaxies over z[?] 0.24-0.80. Visibly merging systems only account for a small fraction (<30%) of the cosmic SFR density over T b[?] 3-7 Gyr. This complements the results of Wolf et al. over a shorter time interval of T b[?] 6.2-6.8 Gyr, and suggests that the behavior of the cosmic SFR density over the last 7 Gyr is predominantly shaped by non-interacting galaxies.

We explore the amount of obscured star formation as a function of environment in the Abell 901/902 (A901/902) supercluster at z = 0.165 in conjunction with a field sample drawn from the A901 and CDFS fields, imaged with the Hubble Space Telescope as part of the Space Telescope A901/902 Galaxy Evolution Survey and Galaxy Evolution from Morphology and Spectral Energy Distributions (SEDs) Survey. We combine the COMBO-17 near-UV/optical SED with Spitzer 24 m photometry to estimate both the unobscured and obscured star formation in galaxies with M * > 1010 M . We find that the star formation activity in massive galaxies is suppressed in dense environments, in agreement with previous studies. Yet, nearly 40% of the star-forming (SF) galaxies have red optical colors at intermediate and high densities. These red systems are not starbursting; they have star formation rates (SFRs) per unit stellar mass similar to or lower than blue SF galaxies. More than half of the red SF galaxies have low infrared-to-ultraviolet (IR-to-UV) luminosity ratios, relatively high Sersic indices, and they are equally abundant at all densities. They might be gradually quenching their star formation, possibly but not necessarily under the influence of gas-removing environmental processes. The other [?]40% of the red SF galaxies have high IR-to-UV luminosity ratios, indicative of high dust obscuration. They have relatively high specific SFRs and are more abundant at intermediate densities. Our results indicate that while there is an overall suppression in the SF galaxy fraction with density, the small amount of star formation surviving the cluster environment is to a large extent obscured, suggesting that environmental interactions trigger a phase of obscured star formation, before complete quenching.

Surveys of distant galaxies with the Hubble Space Telescope and from the ground have shown that there is only mild evolution in the relationship between radial size and stellar mass for galactic disks from z [?] 1 to the present day. Using a sample of nearby disk-dominated galaxies from the Sloan Digital Sky Survey (SDSS) and high-redshift data from the GEMS (Galaxy Evolution from Morphology and SEDs) survey, we investigate whether this result is consistent with theoretical expectations within the hierarchical paradigm of structure formation. The relationship between virial radius and mass for dark matter halos in the LCDM model evolves by about a factor of 2 over this interval. However, N-body simulations have shown that halos of a given mass have less centrally concentrated mass profiles at high redshift. When we compute the expected disk size-stellar mass distribution, accounting for this evolution in the internal structure of dark matter halos and the adiabatic contraction of the dark matter by the self-gravity of the collapsing baryons, we find that the predicted evolution in the mean size at fixed stellar mass since z [?] 1 is about 15%-20%, in good agreement with the observational constraints from GEMS. At redshift z [?] 2, the model predicts that disks at fixed stellar mass were on average only 60% as large as they are today. Similarly, we predict that the rotation velocity at a given stellar mass (essentially the zero point of the Tully-Fisher relation) is only about 10% larger at z [?] 1 (20% at z [?] 2) than at the present day.

Recent observations have demonstrated a significant growth in the integrated stellar mass of the red sequence since z = 1, dominated by a steadily increasing number of galaxies with stellar masses M* < 1011 M. In this paper, we use the COMBO-17 photometric redshift survey in conjunction with deep Spitzer 24 mm data to explore the relationship between star formation and the growth of stellar mass. We calculate star formation rate functions in four different redshift slices, also splitting them into contributions from the red sequence and blue cloud for the first time. We find that the growth of stellar mass since z = 1 is consistent with the integrated star formation rate. Yet, most of the stars formed are in blue cloud galaxies. If the stellar mass already in, and formed in, z < 1 blue cloud galaxies were to stay in the blue cloud, the total stellar mass in blue galaxies would be dramatically overproduced. We explore the expected evolution of stellar mass functions, finding that in this picture the number of massive M* > 3 x 1010 M blue galaxies would also be overproduced; i.e., most of the new stars formed in blue cloud galaxies are in the massive galaxies. We explore a simple truncation scenario in which these "extra" blue galaxies have their star formation suppressed by an unspecified mechanism or mechanisms; simple cessation of star formation in these extra blue galaxies is approximately sufficient to build up the red sequence at M* [?] 1011 M.

We calculate the projected two-point correlation function for samples of luminous and massive galaxies in the COMBO-17 photometric redshift survey, focusing particularly on the amplitude of the correlation function at small projected radii and exploring the constraints such measurements can place on the galaxy merger rate. For nearly volume-limited samples with 0.4 < z < 0.8, we find that 4% +- 1% of luminous MB < -20 galaxies are in close physical pairs (with real-space separation of <30 proper kpc). The corresponding fraction for massive galaxies with M* > 2.5 x 1010 M is 5% +- 1%. Incorporating close pair fractions from the literature, the 2dFGRS and the SDSS, we find a fairly rapid evolution of the merger fraction of massive galaxies between z = 0.8 and the present day. Assuming that the major merger timescale is of order the dynamical timescale for close massive galaxy pairs, we tentatively infer that ~50% (70%) of all galaxies with present-day masses M* > 5 x 1010 M (remnants of mergers between galaxies with M* > 2.5 x 1010 M) have undergone a major merger since z = 0.8(1): major mergers between massive galaxies are a significant driver of galaxy evolution over the last eight billion years.

The jet in 3C 273 is a high-power quasar jet with radio, optical, and X-ray emission, whose size and brightness allow a detailed study of the emission processes acting in it. We present deep Chandra observations of this jet and analyze the spectral properties of the jet emission from radio through X-rays. We find that the X-ray spectra are significantly softer than the radio spectra in all regions of the bright part of the jet, except for the first bright "knot A," ruling out a model in which the X-ray emission from the entire jet arises from beamed inverse Compton scattering of cosmic microwave background photons in a single-zone jet flow. Within two-zone jet models, we find that a synchrotron origin for the jet's X-rays requires fewer additional assumptions than an inverse Compton model, especially if velocity shear leads to efficient particle acceleration in jet flows.

We present a study of the colors, structural properties, and star formation histories for a sample of ~1600 dwarfs over look-back times of ~3 Gyr (z = 0.002-0.25). The sample consists of 401 distant dwarfs drawn from the Galaxy Evolution from Morphologies and SEDs (GEMS) survey, which provides high-resolution Hubble Space Telescope (HST) Advanced Camera for Surveys (ACS) images and accurate redshifts, and also 1291 dwarfs at 10-90 Mpc compiled from the Sloan Digital Sky Survey (SDSS). The sample is complete down to an effective surface brightness of 22 mag arcsec-2 in z and includes dwarfs with Mg = -18.5 to -14 mag. Rest-frame luminosities in Johnson UBV and SDSS ugr filters are provided by the COMBO-17 survey, and structural parameters have been determined by Sersic fits. We find that the GEMS dwarfs are bluer than the SDSS dwarfs by ~0.13 mag in g - r, which is consistent with the color evolution over ~2 Gyr of star formation histories involving moderate starbursts and long periods of continuous star formation. The full color range of the samples cannot be reproduced by single starbursts of different masses or long periods of continuous star formation alone. Furthermore, an estimate of the mechanical luminosities needed for the gas in the GEMS dwarfs to be completely removed from the galaxies shows that a significant number of low-luminosity dwarfs are susceptible to such a complete gas loss, if they would experience a starburst. On the other hand, a large fraction of more luminous dwarfs is likely to retain their gas. We also estimate the star formation rates per unit area for the GEMS dwarfs and find good agreement with the values for local dwarfs.

We have used the 28' x 28' Hubble Space Telescope image mosaic from the GEMS (Galaxy Evolution from Morphology and SEDs) survey in conjunction with the COMBO-17 photometric redshift survey to constrain the incidence of major mergers between spheroid-dominated galaxies with little cold gas (dry mergers) since z = 0.7. A set of N-body merger simulations was used to explore the morphological signatures of such interactions: they are recognizable either as <5 kpc separation close pairs or because of broad, low surface brightness tidal features and asymmetries. Data with the depth and resolution of GEMS are sensitive to dry mergers between galaxies with MV [?] -20.5 for z [?] 0.7; dry mergers at higher redshifts are not easily recovered in single-orbit HST imaging. Six dry mergers (12 galaxies) with luminosity ratios between 1 : 1 and 4 : 1 were found from a sample of 379 red early-type galaxies with MV < -20.5 and 0.1 < z < 0.7. The simulations suggest that the morphological signatures of dry merging are visible for ~150 Myr, and we use this timescale to convert the observed merger incidence into a rate. On this basis we find that present-day spheroidal galaxies with MV < -20.5 on average have undergone between 0.5 and 2 major dry mergers since z ~ 0.7. We have compared this result with the predictions of a cold dark matter-based semianalytic galaxy formation model. The model reproduces the observed declining major merger fraction of bright galaxies and the space density of luminous early-type galaxies reasonably well. The predicted dry merger fraction is consistent with our observational result. Hence, hierarchical models predict and observations now show that major dry mergers are an important driver of the evolution of massive early-type galaxies in recent epochs.

We combine HST imaging from the GEMS (Galaxy Evolution from Morphologies and SEDs) survey with photometric redshifts from COMBO-17 to explore the evolution of disk-dominated galaxies since z [?] 1.1. The sample is composed of all GEMS galaxies with Sersic indices n < 2.5, derived from fits to the galaxy images. We account fully for selection effects through careful analysis of image simulations; we are limited by the depth of the redshift and HST data to the study of galaxies with MV [?] -20, or equivalently, log img1.gif [?] 10. We find strong evolution in the magnitude-size scaling relation for galaxies with MV [?] -20, corresponding to a brightening of ~1 mag arcsec-2 in rest-frame V band by z ~ 1. Yet disks at a given absolute magnitude are bluer and have lower stellar mass-to-light ratios at z ~ 1 than at the present day. As a result, our findings indicate weak or no evolution in the relation between stellar mass and effective disk size for galaxies with log img1.gif [?] 10 over the same time interval. This is strongly inconsistent with the most naive theoretical expectation, in which disk size scales in proportion to the halo virial radius, which would predict that disks are a factor of 2 denser at fixed mass at z ~ 1. The lack of evolution in the stellar mass-size relation is consistent with an "inside-out" growth of galaxy disks on average (galaxies increasing in size as they grow more massive), although we cannot rule out more complex evolutionary scenarios.

We combine imaging from the Hubble Space Telescope Advanced Camera for Surveys, as part of the GEMS (Galaxy Evolution from Morphologies and SEDs) survey, with redshifts and rest-frame quantities from COMBO-17 to study the evolution of morphologically early-type galaxies with red colors since z = 1. From 0.deg5 x 0.deg5 imaging, we draw a large sample of 728 galaxies with centrally concentrated radial profiles (i.e., n [?] 2.5 from Sersic fits) and rest-frame (U - V) colors on the red sequence. We explore how the correlations of rest-frame V-band luminosity and of stellar mass with intrinsic half-light size change over the last half of cosmic time. By appropriate comparison with the well-defined local relations from the Sloan Digital Sky Survey, we find that the luminosity-size and stellar mass-size relations evolve in a manner that is consistent with the passive aging of ancient stellar populations. By itself, this result is consistent with a completely passive evolution of the red early-type galaxy population. If instead, as demonstrated by a number of recent surveys, the early-type galaxy population builds up in mass by roughly a factor of 2 since z ~ 1, our results imply that new additions to the early-type galaxy population follow similar luminosity-size and stellar mass-size correlations, compared to the older subset of early-type galaxies. Adding early-type galaxies to the red sequence through the fading of previously prominent disks appears to be consistent with the data. Through comparison with models, the role of dissipationless merging is limited to <1 major merger on average since z = 1 for the most massive galaxies. Predictions from models of gas-rich mergers are not yet mature enough to allow a detailed comparison to our observations. We find tentative evidence that the amount of luminosity evolution depends on galaxy stellar mass, such that the least massive galaxies show stronger luminosity evolution compared to more massive early types. This could reflect a different origin of low-mass early-type galaxies and/or younger stellar populations; the present data are insufficient to discriminate between these possibilities.

We present a first analysis of deep 24 mm observations with the Spitzer Space Telescope of a sample of nearly 1500 galaxies in a thin redshift slice, 0.65 [?] z < 0.75. We combine the infrared data with redshifts, rest-frame luminosities, and colors from COMBO-17 and with morphologies from Hubble Space Telescope images collected by the Galaxy Evolution from Morphology and SEDs (GEMS) and Great Observatories Origins Deep Survey (GOODS) projects. To characterize the decline in star formation rate (SFR) since z ~ 0.7, we estimate the total thermal IR luminosities, SFRs, and stellar masses for the galaxies in this sample. At z ~ 0.7, nearly 40% of intermediate- and high-mass galaxies (with stellar masses [?]2 x 1010 M) are undergoing a period of intense star formation above their past-averaged SFR. In contrast, less than 1% of equally massive galaxies in the local universe have similarly intense star formation activity. Morphologically undisturbed galaxies dominate the total infrared luminosity density and SFR density: at z ~ 0.7, more than half of the intensely star-forming galaxies have spiral morphologies, whereas less than ~30% are strongly interacting. Thus, a decline in major merger rate is not the underlying cause of the rapid decline in cosmic SFR since z ~ 0.7. Physical properties that do not strongly affect galaxy morphology--for example, gas consumption and weak interactions with small satellite galaxies--appear to be responsible.

We present the rest-frame colors and luminosities of ~25,000 mR [?] 24 galaxies in the redshift range 0.2 < z [?] 1.1 drawn from 0.78 deg2 of the COMBO-17 survey (Classifying Objects by Medium-Band Observations in 17 Filters). We find that the rest-frame color distribution of these galaxies is bimodal at all redshifts out to z ~ 1. This bimodality permits a model-independent definition of red early-type galaxies and blue late-type galaxies at any given redshift. The colors of the blue peak become redder toward the present day, and the number density of blue luminous galaxies has dropped strongly since z ~ 1. Focusing on the red galaxies, we find that they populate a color-magnitude relation. Such red sequences have been identified in galaxy cluster environments, but our data show that such a sequence exists over this redshift range even when averaging over all environments. The mean color of the red galaxy sequence evolves with redshift in a way that is consistent with the aging of an ancient stellar population. The rest-frame B-band luminosity density in red galaxies evolves only mildly with redshift in a L-dominated cold dark matter universe. When we account for the change in stellar mass-to-light ratio implied by the redshift evolution in red galaxy colors, the COMBO-17 data indicate an increase in stellar mass on the red sequence by a factor of 2 since z ~ 1. The largest source of uncertainty is large-scale structure, implying that considerably larger surveys are necessary to further refine this result. We explore mechanisms that may drive this evolution in the red galaxy population, finding that both galaxy merging and truncation of star formation in some fraction of the blue star-forming population are required to fully explain the properties of these galaxies.

We present the results of an extensive Fabry-Perot investigation of the extended line-emission gas surrounding powerful radio galaxies. High spatial (14) and kinematical (400 km s-1) resolution observations of the [O II] ?3727 emission line of the powerful, double radio source 3C 34 (z= 0.689) are given. We present evidence that anisotropic radiation emanating from a hidden AGN is responsible for photoionizing a line-emission region extending more than 120 kpc. This is indicated by the distinctively biconical morphology of 3C 34's [O II] gas. A number of line-emission components may also show various degrees of "shadowing" outward from the central ionizing source. A simple photoionization model is used to show that this interpretation is energetically viable on these length scales. The luminosity of the hidden central AGN, necessary to account for the observed [O II] luminosity, is compatible with that of a typical 3CR quasar at a similar redshift. Although this interpretation can account for the excitation and most of the unique shape of the warm gas, it is insufficient to explain the velocity and line width structures observed with our Fabry-Perot. Therefore, we also propose that the illuminated medium surrounding 3C 34 is the result of gas swept up by the lateral expansion and backflow of the radio source lobes.

We investigate the evolution of the galaxy population since redshift 2 with a focus on the color bimodality and mass density of the red sequence. We obtain precise and reliable photometric redshifts up to z = 2 by supplementing the optical survey COMBO-17 with observations in four near-infrared bands on 0.2 deg2 of the COMBO-17 A901-field. Our results are based on an H-band-selected catalog of 10,692 galaxies complete to H = 21.[?]7. We measure the rest-frame color (U 280-V) of each galaxy, which across the redshift range of our interest requires no extrapolation and is robust against moderate redshift errors by staying clear of the 4000 A break. We measure the color-magnitude relation of the red sequence as a function of look-back time from the peak in a color-error-weighted histogram, and thus trace the galaxy bimodality out to z [?] 1.65. The (U 280-V) of the red sequence is found to evolve almost linearly with look-back time. At high redshift, we find massive galaxies in both the red and the blue population. Red-sequence galaxies with log M */M >11 increase in mass density by a factor of [?]4 from z [?] 2 to 1 and remain nearly constant at z < 1. However, some galaxies as massive as log M */M = 11.5 are already in place at z [?] 2.