Using the MOSFIRE near-infrared multi-slit spectrograph on the Keck 1 Telescope, we have secured high signal-to-noise ratio absorption line spectra for six massive galaxies with redshift 2 < z < 2.5. Five of these galaxies lie on the red sequence and show signatures of passive stellar populations in their rest-frame optical spectra. By fitting broadened spectral templates we have determined stellar velocity dispersions and, with broad-band Hubble Space Telescope and Spitzer photometry and imaging, stellar masses and effective radii. Using this enlarged sample of galaxies, we confirm earlier suggestions that quiescent galaxies at z > 2 have small sizes and large velocity dispersions compared to local galaxies of similar stellar mass. The dynamical masses are in very good agreement with stellar masses (log M-*/M-dyn = -0.02 +/- 0.03), although the average stellar-to-dynamical mass ratio is larger than that found at lower redshift (-0.23 +/- 0.05). By assuming evolution at fixed velocity dispersion, not only do we confirm a surprisingly rapid rate of size growth but we also consider the necessary evolutionary track on the mass-size plane and find a slope alpha = d log R-e/d log M-* greater than or similar to 2 inconsistent with most numerical simulations of minor mergers. Both results suggest an additional mechanism may be required to explain the size growth of early galaxies.

We analyze the stellar populations of a sample of 62 massive (log M-*/M-circle dot > 10.7) galaxies in the redshift range 1 < z < 1.6, with the main goal of investigating the role of recent quenching in the size growth of quiescent galaxies. We demonstrate that our sample is not biased toward bright, compact, or young galaxies, and thus is representative of the overall quiescent population. Our high signal-to-noise ratio Keck/LRIS spectra probe the rest-frame Balmer break region that contains important absorption line diagnostics of recent star formation activity. We obtain improved measures of the various stellar population parameters, including the star formation timescale tau, age, and dust extinction, by fitting templates jointly to both our spectroscopic and broadband photometric data. We identify which quiescent galaxies were recently quenched and backtrack their individual evolving trajectories on the UVJ color-color plane finding evidence for two distinct quenching routes. By using sizes measured in the previous paper of this series, we confirm that the largest galaxies are indeed among the youngest at a given redshift. This is consistent with some contribution to the apparent growth from recent arrivals, an effect often called progenitor bias. However, we calculate that recently quenched objects can only be responsible for about half the increase in average size of quiescent galaxies over a 1.5 Gyr period, corresponding to the redshift interval 1.25 < z < 2. The remainder of the observed size evolution arises from a genuine growth of long-standing quiescent galaxies.

We report the discovery of RG1M0150, a massive, recently quenched galaxy at z = 2.636 that is multiply imaged by the cluster MACSJ0150.3-1005. We derive a stellar mass of log M-* = 11.49(-0.16)(+0.10) and a half-light radius of R-e,R-maj = 1.8 +/- 0.4 kpc. Taking advantage of the lensing magnification, we are able to spatially resolve a remarkably massive yet compact quiescent galaxy at z > 2 in ground-based near-infrared spectroscopic observations using Magellan/FIRE and Keck/MOSFIRE. We find no gradient in the strength of the Balmer absorption lines over 0.6R(e)-1.6R(e), which are consistent with an age of 760 Myr. Gas emission in [N II] broadly traces the spatial distribution of the stars and is coupled with weak H alpha emission (log [N II]/H alpha = 0.6 +/- 0.2), indicating that OB stars are not the primary ionizing source. The velocity dispersion within the effective radius is sigma(e,stars) = 271 +/- 41 km s(-1). We detect rotation in the stellar absorption lines for the first time beyond z similar to 1. Using a two-integral Jeans model that accounts for observational effects, we measure a dynamical mass of log M-dyn = 11.24 +/- 0.14 and V/sigma = 0.70 +/- 0.21. This is a high degree of rotation considering the modest observed ellipticity of 0.12 +/- 0.08, but it is consistent with predictions from dissipational merger simulations that produce compact remnants. The mass of RG1M0150 implies that it is likely to become a slowly rotating elliptical. If it is typical, this suggests that the progenitors of massive ellipticals retain significant net angular momentum after quenching which later declines, perhaps through accretion of satellites.

Observations of strong gravitational lensing, stellar kinematics, and larger-scale tracers enable accurate measures of the distribution of dark matter (DM) and baryons in massive early-type galaxies (ETGs). While such techniques have been applied to galaxy-scale and cluster-scale lenses, the paucity of intermediate-mass systems with highquality data has precluded a uniform analysis of mass-dependent trends. With the aim of bridging this gap, we present new observations and analyses of 10 group-scale lenses at < z > = 0.36, characterized by Einstein radii theta(Ein) = 2.'' 5 - 5.'' 1 and a mean halo mass of M-200 = 10(14.0) M-circle dot. We measure a mean concentration C200 = 5.0 0.8 consistent with unmodified cold dark matter halos. By combining our data with other lens samples, we analyze the mass structure of ETGs in 10(13) M-circle dot-10(15) M-circle dot halos using homogeneous techniques. We show that the slope of the total density profile gamma(tot) within the effective radius depends on the stellar surface density, as demonstrated previously, but also on the halo mass. We analyze these trends using halo occupation models and resolved stellar kinematics with the goal of testing the universality of the DM profile. Whereas the central galaxies of clusters require a shallow inner DM density profile, group-scale lenses are consistent with a Navarro Frenk White profile or one that is slightly contracted. The largest uncertainties arise from the sample size and likely radial gradients in stellar populations. We conclude that the net effect of baryons on the DM distribution may not be universal, but more likely varies with halo mass due to underlying trends in star formation efficiency and assembly history.

We report the discovery of a massive log(M/M-circle dot) = 10.74 (+0.18)(-0.16) galaxy at the same redshift as a carbon-monoxide-bearing sub-damped Ly alpha absorber (sub-DLA) seen in the spectrum of QSO J1439+1117. The galaxy, J1439B, is located 4.'' 7 from the QSO sightline, a projected distance of 38 physical kpc at z = 2.4189, and exhibits broad optical emission lines (sigma([O) (III]) = 303 +/- 12 km s(-1) with ratios characteristic of excitation by an active galactic nucleus (AGN). The galaxy has a factor of similar to 9 lower star formation than is typical of star-forming galaxies of the same mass and redshift. The nearby sub-DLA is highly enriched, suggesting its galactic counterpart must be massive if it follows the z similar to 2 mass-metallicity relationship. Metallic absorption within the circumgalactic medium of the sub-DLA and J1439B is spread over a velocity range Delta v > 1000 km s(-1), suggesting an energetic origin. We explore the possibility that a different galaxy could be responsible for the rare absorber, and conclude that it is unlikely based on imaging, integral-field spectroscopy, and high-z massive galaxy pair statistics. We argue that the gas seen in absorption against the QSO was likely ejected from the galaxy J1439B and therefore provides a unique observational probe of AGN feedback in the distant universe.

We present new observations of the three nearest early-type galaxy (ETG) strong lenses discovered in the SINFONI Nearby Elliptical Lens Locator Survey (SNELLS). Based on their lensing masses, these ETGs were inferred to have a stellar initial mass function ( IMF) consistent with that of the Milky Way, not the bottom-heavy IMF that has been reported as typical for high-sigma ETGs based on lensing, dynamical, and stellar population synthesis techniques. We use these unique systems to test the consistency of IMF estimates derived from different methods. We first estimate the stellar M*/L using lensing and stellar dynamics. We then fit high-quality optical spectra of the lenses using an updated version of the stellar population synthesis models developed by Conroy & van Dokkum. When examined individually, we find good agreement among these methods for one galaxy. The other two galaxies show 2-3 sigma tension with lensing estimates, depending on the dark matter contribution, when considering IMFs that extend to 0.08M(circle dot). Allowing a variable low-mass cutoff or a nonparametric form of the IMF reduces the tension among the IMF estimates to <2 sigma. There is moderate evidence for a reduced number of low-mass stars in the SNELLS spectra, but no such evidence in a composite spectrum of matched-sigma ETGs drawn from the SDSS. Such variation in the form of the IMF at low stellar masses (m less than or similar to 0.3M(circle dot)), if present, could reconcile lensing/dynamical and spectroscopic IMF estimates for the SNELLS lenses and account for their lighter M*/L relative to the mean matched-sigma ETG. We provide the spectra used in this study to facilitate future comparisons.

Quiescent galaxies at z greater than or similar to 2 are compact and have weak or absent emission lines, making it difficult to spatially resolve their kinematics and stellar populations using ground-based spectroscopy. Gravitationally lensed examples provide a promising route forward, but such objects are very rare. We describe a search in the fields of 232 galaxy clusters that has uncovered five bright (H-AB < 20) lensed galaxies with red near-infrared colors. These include MRG-M0138, which is the brightest lensed galaxy known in the near-infrared. Analysis of near-infrared spectra and multiband photometry confirms that all are quiescent galaxies at z = 1.95-2.64 with stellar ages of 0.5-1.4. Gyr (corresponding to formation epochs z(form)similar or equal to 3-4) and stellar masses of 10(11.6-12.8) mu(-1) M-circle dot, where mu is the magnification. In three cases, we derive lens models and reconstruct the source structure; these galaxies are massive (M-* greater than or similar to 10(11.0) M-circle dot) and follow the mass-size relation defined by unlensed samples. In two of these three galaxies, the main structural component is an inclined disk. Weak emission lines are detected in four of five galaxies with high ratios [N II]/H alpha similar or equal to 2-6 that are inconsistent with a star formation origin. Based on the line ratios, the Ha equivalent widths, and the distribution and kinematics of the gas, we infer that shocks are likely to be present in at least two galaxies and could be present in all of the line emitters. We speculate that these could be analogs of local galaxies in which AGN jet-driven outflows are thought to heat the interstellar medium and suppress star formation. In further papers, we will present spatially resolved measurements of the stellar populations and kinematics of this unique sample.

We present deep near-infrared spectra for a sample of 24 quiescent galaxies in the redshift range 1.5 < z < 2.5 obtained with the MOSFIRE spectrograph at the W. M. Keck Observatory. In conjunction with a similar data set we obtained in the range1< z < 1.5 with the LRIS spectrograph, we analyze the kinematic and structural properties for 80 quiescent galaxies, the largest homogeneously selected sample to date spanning 3 Gyr of early cosmic history. Analysis of our Keck spectra together with measurements derived from associated Hubble Space Telescope images reveals increasingly larger stellar velocity dispersions and smaller sizes to redshifts beyond z similar to 2. By classifying our sample according to Sersic indices, we find that among disk-like systems the flatter ones show a higher dynamical to stellar mass ratio compared to their rounder counterparts, which we interpret as evidence for a significant contribution of rotational motion. For this subset of disk-like systems, we estimate that V/sigma, the ratio of the circular velocity to the intrinsic velocity dispersion, is a factor of two larger than for present-day disky quiescent galaxies. We use the velocity dispersion measurements also to explore the redshift evolution of the dynamical to stellar mass ratio, and to measure for the first time the physical size growth rate of individual systems over two distinct redshift ranges, finding a faster evolution at earlier times. We discuss the physical origin of this time-dependent growth in size in the context of the associated reduction of the systematic rotation.

Stellar kinematics provide insights into the masses and formation histories of galaxies. At high redshifts, spatially resolving the stellar kinematics of quiescent galaxies is challenging due to their compact sizes. Using deep near-infrared spectroscopy, we have measured the resolved stellar kinematics of four quiescent galaxies at z = 1.95-2.64, introduced in Paper. I, that are gravitationally lensed by galaxy clusters. Analyses of two of these have previously been reported individually by Newman et al. and Toft et al., and for the latter, we present new observations. All four galaxies show significant rotation and can be classified as "fast rotators." In the three systems for which the lensing constraints permit a reconstruction of the source, we find that all are likely to be highly flattened (intrinsic ellipticities of approximate to 0.75-0.85) disk-dominated galaxies with rapid rotation speeds of V-max = 290-352. km s(-1) and predominantly rotational support, as indicated by the ratio (V/sigma)(Re) 1.7 2.3. (s) = Compared to coeval star-forming galaxies of similar mass, the quiescent galaxies have smaller V/sigma. Given their high masses, M-dyn greater than or similar to 2 x 10(11) M-circle dot, we argue that these galaxies are likely to evolve into "slow rotator" elliptical galaxies whose specific angular momentum is reduced by a factor of 5-10. This provides strong evidence for merger-driven evolution of massive galaxies after quenching. Consistent with indirect evidence from earlier morphological studies, our small but unique sample suggests that the kinematic transformations that produced round, dispersion-supported elliptical galaxies were not generally coincident with quenching. Such galaxies probably emerged later via mergers that increased their masses and sizes while also eroding their rotational support.

Dark-matter-only simulations predict that dark matter halos have cusp-like inner density profiles, while observations of low-mass galaxies have found a range of inner slopes that are typically much shallower. It is still not well established whether this discrepancy can be explained by baryonic feedback or if it may require modified dark matter models. To better understand the diversity of dark matter profiles in dwarf galaxies, we undertook a survey of 26 low-mass galaxies (logM(*)/M-circle dot = 8.4-9.8, v(max) = 50-140 km s(-1)) within 30 Mpc using the Palomar Cosmic Web Imager, which is among the largest integral field spectroscopic surveys of its type. In this paper, we derive H alpha velocity fields for the full sample with a typical spatial resolution of similar to 160 pc. We extract rotation curves and verify their robustness to several choices in the analysis. We present a method for improving the velocity precision obtained from image slicing spectrographs using narrowband H alpha images. For 11 galaxies, we compare the H alpha velocity fields to CO kinematics measured using CARMA, finding the maps to be in good agreement. The standard deviation of the difference is typically similar to 7 km s(-1), comparable to the level of turbulence in the interstellar medium, showing that the two tracers have substantially the same bulk kinematics. In a companion paper, we will use the rotation curves produced here to construct mass models of the galaxies and determine their dark matter density profiles.