A Black Hole Mass Determination for the Compact Galaxy Mrk 1216
Jonelle L. Walsh, Remco C.E. van den Bosch, Karl Gebhardt, Akın Yıldırım, Kayhan Gültekin, Bernd Husemann, Douglas O. Richstone
aa r X i v : . [ a s t r o - ph . GA ] D ec Draft version March 13, 2018
Preprint typeset using L A TEX style emulateapj v. 5/2/11
A BLACK HOLE MASS DETERMINATION FOR THE COMPACT GALAXY MRK 1216
Jonelle L. Walsh , Remco C. E. van den Bosch , Karl Gebhardt , Akın Yıldırım , , Kayhan G¨ultekin , BerndHusemann , , and Douglas O. Richstone George P. and Cynthia Woods Mitchell Institute for Fundamental Physics and Astronomy, Department of Physics and Astronomy,Texas A&M University, 4242 TAMU, College Station, TX 77843, USA; [email protected] Max-Planck-Institut f¨ur Astronomie, K¨onigstuhl 17, D-69117 Heidelberg, Germany Department of Astronomy, The University of Texas at Austin, 2515 Speedway, Stop C1400, Austin, TX 78712, USA Max-Planck-Institut f¨ur Astrophysik, Karl-Schwarzschild-Str. 1, 85741 Garching, Germany Department of Astronomy, University of Michigan, 1085 S. University Ave., Ann Arbor, MI 48109, USA European Southern Observatory, Karl-Schwarzschild-Str. 2, 85748 Garching, Germany
Draft version March 13, 2018
ABSTRACTMrk 1216 is a nearby, early-type galaxy with a small effective radius of 2.8 kpc and a large stellarvelocity dispersion of 308 km s − for its K -band luminosity of 1 . × L ⊙ . Using integral-fieldspectroscopy assisted by adaptive optics from Gemini North, we measure spatially resolved stellarkinematics within ∼
450 pc of the galaxy nucleus. The galaxy exhibits regular rotation with velocitiesof ±
180 km s − and a sharply peaked velocity dispersion profile that reaches 425 km s − at thecenter. We fit axisymmetric, orbit-based dynamical models to the combination of these high angularresolution kinematics, large-scale kinematics extending to roughly three effective radii, and HubbleSpace Telescope imaging, resulting in a constraint of the mass of the central black hole in Mrk 1216.After exploring several possible sources of systematics that commonly affect stellar-dynamical blackhole mass measurements, we find a black hole mass of M BH = (4 . ± . × M ⊙ and a H -band stellarmass-to-light ratio of Υ H = 1 . ± . ⊙ (1 σ uncertainties). Mrk 1216 is consistent with the local blackhole mass – stellar velocity dispersion relation, but is a factor of ∼ −
10 larger than expectationsfrom the black hole mass – bulge luminosity and black hole mass – bulge mass correlations whenconservatively using the galaxy’s total luminosity or stellar mass. This behavior is quite similar to theextensively studied compact galaxy NGC 1277. Resembling the z ∼ Subject headings: galaxies: elliptical and lenticular, cD – galaxies: individual (Mrk 1216) – galaxies:kinematics and dynamics – galaxies: nuclei – black hole physics INTRODUCTION
Our understanding of the connection between super-massive black holes and their host galaxies is anchoredby ∼
100 dynamical black hole mass ( M BH ) measure-ments that have been made over the past two decades(e.g., Kormendy & Ho 2013; van den Bosch 2016, andreferences therein). Strong correlations have emergedbetween M BH and large-scale galaxy properties, likethe bulge luminosity ( L bul ; e.g., Kormendy & Richstone1995; Marconi & Hunt 2003; Kormendy & Ho 2013) ormass ( M bul ; e.g., H¨aring & Rix 2004; Sani et al. 2011;McConnell & Ma 2013), and the stellar velocity disper-sion ( σ ⋆ ; e.g., Ferrarese & Merritt 2000; Gebhardt et al.2000; G¨ultekin et al. 2009). These relations are con-nected to each other, and the search for the mostfundamental one is still ongoing (Beifiori et al. 2012;Saglia et al. 2016; van den Bosch 2016). The empiricalrelationships imply that black holes are key componentsof galaxies and regulate galaxy properties via feedbackmechanisms (Silk & Rees 1998; Fabian 1999), althougha non-causal origin in which black holes do not activelyshape their host galaxies is also possible (Peng 2007;Jahnke & Macci`o 2011). Establishing the exact role ofblack holes in galaxy evolution and accurately inferringthe black hole mass function requires increasing the num- ber of M BH measurements, specifically targeting galax-ies with diverse properties that have experienced variedgrowth channels.As we begin to examine a broader range of blackhole masses and hosts, galaxies with different structuralproperties show surprises in the scaling relations. Re-cent progress detecting high-mass black holes in Bright-est Cluster Galaxies (BCGs) and other large early-typegalaxies hint that these objects may be positive outliersfrom M BH − σ ⋆ and M BH − L bul (e.g., McConnell et al.2012; Rusli et al. 2013; Thomas et al. 2016), but thereare still too few measurements to firmly characterize thescaling relations at M BH & M ⊙ . The uncertaintiesare equally severe at the opposite end, where spiral galax-ies with low-mass black holes ( M BH . M ⊙ ) mea-sured from water megamaser disks exhibit substantialscatter below the global black hole – host galaxy relations(e.g., Greene et al. 2010; L¨asker et al. 2016; Greene et al.2016). There are also new observations of compactgalaxies, whose black holes are a remarkably large frac-tion of the galaxy’s stellar mass (e.g., Seth et al. 2014;Walsh et al. 2015, 2016; Saglia et al. 2016).NGC 1277 and NGC 1271 are two such compactgalaxies, with NGC 1277 being widely studied overthe last few years (e.g., van den Bosch et al. 2012;Emsellem 2013; Walsh et al. 2016; Scharw¨achter et al.2016; Graham et al. 2016a). Both were origi-nally discovered by the HET Massive Galaxy Sur-vey (van den Bosch et al. 2015) and share consider-able similarities with the z ∼ z ∼ R e ∼ − M ⋆ ∼ M ⊙ , and stellar mass surface density profilesthat are elevated at the center and drop off steeply atlarger radii compared to low-redshift early-type galax-ies. In addition, NGC 1277 and NGC 1271 are rotating,consistent with the disk-like flattened structures of the z ∼ ∼
10 Gyr) ex-tending out to several R e . The red nuggets are thoughtto grow in size and moderately in mass through merg-ers to produce the present-day massive galaxies (e.g.,van Dokkum et al. 2010), with a small fraction experi-encing passive evolution since z ∼ M BH measure-ments, NGC 1277 and NGC 1271 are most similar toNGC 4342 (Cretton & van den Bosch 1999) and NGC1332. All are nearby galaxies that are flattened and ro-tating, with small effective radii and large stellar veloc-ity dispersions for their luminosities. They contain blackholes that are more massive than the predictions from M BH − L bul , yet are consistent with M BH − σ ⋆ . Themagnitude of the offset from M BH − L bul depends onthe adopted bulge luminosity (e.g., Walsh et al. 2015,2016; Graham et al. 2016a,b), although we note thatNGC 1332 may very well be consistent with the blackhole relations given the uncertainties associated withboth M BH (Rusli et al. 2011; Barth et al. 2016a,b) andthe bulge component (Rusli et al. 2011; Kormendy & Ho2013; Savorgnan & Graham 2016; Saglia et al. 2016).Clearly, additional mass measurements of black holes inNGC 1277-like galaxies are needed. A significant sampleof local analogs to the z ∼ M BH − L bul would suggest that thisblack hole scaling relation did not apply at earlier times,galaxies instead harbored over-massive black holes, andsubsequent galaxy growth had yet to occur.Beyond the connections to the z ∼ M BH relation-ships, and their properties are quite distinct from theBCGs and giant ellipticals that are expected to house themost massive black holes in the Universe. Currently, atthe high-mass end, the differing behaviors of M BH − σ ⋆ and M BH − L bul , and the poorly characterized scatterin M BH for fixed σ ⋆ or L bul , lead to strongly divergentpredictions of the black hole mass function (Lauer et al.2007a). This in turn affects inferences about black holegrowth histories and constraints of the mean radiativeefficiency of black hole accretion, the duty cycle of activegalactic nuclei, and the redshift evolution of the scalingrelations (e.g., Marconi et al. 2004; Lauer et al. 2007b;Shankar et al. 2009; Robertson et al. 2006).In this paper, we examine Mrk 1216, an early-type, compact ( R e = 2 . M ⋆ = 1 . × M ⊙ ), high-dispersion ( σ ⋆ = 308 km s − ) galaxy found through theHET Massive Galaxy Survey. Yıldırım et al. (2015) pre-sented wide-field integral-field spectroscopy of Mrk 1216,and constructed axisymmetric Schwarschild models in or-der to learn about the galaxy’s dynamical stellar mass-to-light ratio and dark matter halo. The spatial reso-lution, however, was insufficient to pin down the blackhole mass, and Yıldırım et al. (2015) set an upper-limitof M BH < × M ⊙ . Here, we use Gemini North ob-servations assisted by adaptive optics (AO) to resolvethe region where the black hole dominates the gravi-tational potential (the black hole sphere of influence; r sphere = GM BH /σ ⋆ ), thereby obtaining a secure stellar-dynamical M BH measurement. We assume a distance of94 Mpc to Mrk 1216. This is the same distance adoptedby Yıldırım et al. (2015) and is the Virgo + Great At-tractor + Shapley Supercluster Infall value (Mould et al.2000) for H = 70 . − Mpc − , Ω M = 0 .
27 andΩ Λ = 0 .
73. At this distance, 1 ′′ corresponds to 456 pc.The paper is structured as follows. We review theimaging observations and the luminous mass model inSection 2. In Sections 3 and 4, we describe the high an-gular resolution and the large-scale spectroscopic obser-vations, the measured stellar kinematics, and the point-spread function (PSF) characterizations. An overview ofthe stellar-dynamical models and the results from thosemodels, including an examination of the black hole masserror budget, is provided in Section 5. We study thegalaxy’s orbital structure in 6, and discuss the locationof Mrk 1216 on the black hole mass – host galaxy rela-tionships, as well as the implications, in Sections 7 and8. Concluding remarks are provided in Section 9. HST IMAGING
We obtained an
HST
Wide-Field Camera 3 (WFC3) F W image of Mrk 1216. The WFC3/IR observa-tions were executed under program GO-13050, and in-cluded dithered full array images and brief subarray ex-posures with a total integration time of 1354 s. Thedata were reduced, and the flattened, calibrated im-ages were corrected for geometric distortions, cleaned,and combined using AstroDrizzle (Gonzaga et al. 2012)to produce a super-sampled image with a scale of 0 . ′′ − . After masking the foreground stars, we de-scribed the galaxy’s stellar surface brightness distribu-tion as the sum of two-dimensional (2D) Gaussians.Such a Multi-Gaussian Expansion (MGE; Monnet et al.1992; Emsellem et al. 1994) is able to recover the surfacebrightness profiles of realistic multi-component galaxieswhile also allowing for the intrinsic luminosity density tobe determined through an analytical deprojection. Dur-ing the MGE fit, we took into account the WFC3 PSFfrom van der Wel et al. (2012). The PSF was generatedwith TinyTim (Krist & Hook 2004) for the F W fil-ter and a G2 V star at the center of the WFC3 detector,then drizzled to produce the same scale as our final Mrk1216 image.The Mrk 1216 MGE is composed of 10 Gaussianswith dispersions, measured along the major axis, of0 . ′′ − . ′′
76, and projected axis ratios between 0.52 and0.99. The components have the same center, and a po-sition angle of 70.2 ◦ east of north. The final parame-ter values, after correction for Galactic extinction usingthe Schlafly & Finkbeiner (2011) WFC3 F W valueof 0.017 mag and assuming an H -band absolute solarmagnitude of 3.32 (Binney & Merrifield 1998), are givenin Yıldırım et al. (2015). The MGE fits the HST imagevery well within the central ∼ ′′ , and allows us to accu-rately infer the stellar gravitational potential. We referthe reader to Yıldırım et al. (2015) for additional detailsregarding the imaging observations, data reduction, andconstruction of the MGE model. NIFS OBSERVATIONS AND MEASUREMENTS
In addition to the luminous mass model, stellarkinematics on scales comparable to the black holesphere of influence are crucial inputs into the dynam-ical models. We therefore observed Mrk 1216 withthe Near-infrared Integral Field Spectrometer (NIFS;McGregor et al. 2003) aided by the ALTtitude conjugateAdaptive optics for the InfraRed (Herriot et al. 2000;Boccas et al. 2006) system on Gemini North. The ob-servations were taken on 21 Dec 2013 under programGN-2013A-Q-1 in laser guide star (LGS) mode using an R = 13 . ′′ from the galaxy nu-cleus as the tip-tilt reference. Four blocks of consecu-tive Object-Sky-Object sequences with 600 s exposureswere recorded. The observations were acquired with the H + K filter and the K grating centered on 2 . µ m. Weobserved the tip-tilt star once during the night for anestimate of the PSF and an A0 V star for telluric cor-rection. The normal baseline calibrations consisting ofdark frames, flat fields, Argon/Xeon arc lamp exposures,and a Ronchi mask (to establish the spatial rectification)were taken as well.We processed the NIFS data using PyRAF and theGemini data reduction package version 1.11, followingthe steps in the NIFS example scripts for calibration,telluric, and science exposures. For the galaxy, the ba-sic procedure consisted of preparing the raw images forprocessing within the NIFS data reduction package, sub-tracting sky frames from adjacent object exposures, flatfielding, removing bad pixels and cosmic rays, wave-length calibration, and spatial rectification. We cor-rected for telluric features using an A0 V star, whosespectrum had been divided by a blackbody with a tem-perature of 9480 K after interpolating over the Br γ ab-sorption line. We then assembled data cubes having x and y spatial dimensions, each with a scale of 0 . ′′ − , and a wavelength axis. We determined the spa-tial offsets between individual galaxy cubes by summingover the wavelength dimension and cross-correlating theimages. These eight cubes, corresponding to a total of1.3 hours on-source, were aligned and combined to pro-duce the final Mrk 1216 data cube. We followed similarsteps to reduce the NIFS observation of the PSF star. Stellar Kinematics
From the reduced Mrk 1216 data cube, we measuredthe stellar kinematics as a function of spatial loca-tion. Specifically, we extracted the line-of-sight veloc-ity distribution (LOSVD), parameterized by the first PyRAF is a product of the Space Telescope Science Institute,which is operated by AURA for NASA four Gauss-Hermite moments, in 67 Voronoi spatial bins(Cappellari & Copin 2003) using the penalized pixel fit-ting (pPXF) code of Cappellari & Emsellem (2004). Thespatial bins were chosen so that the galaxy spectra hada signal-to-noise ratio (S/N) &
40, where the S/N wasmeasured as the median flux divided by the standard de-viation of the pPXF model residuals. Such a high S/Nspectrum is required in order to measure the LOSVD’sdeviation from a Gaussian (e.g., van der Marel & Franx1993; Bender et al. 1994).We provided pPXF with a velocity template librarycomposed of 12 stars (K0 − M5 giant stars and two late-type supergiants), which were observed with NIFS in the K band. The stars are a subset of those presented inWinge et al. (2009), but we have reduced the data our-selves (Walsh et al. 2016), starting with the raw framesand their calibration files retrieved from the Gemini Sci-ence Archive. During the fit with pPXF, we correctedfor slight differences in the continuum shape and equiva-lent width between the LOSVD-convolved template starsand the observed galaxy spectra via an additive con-stant and a multiplicative Legendre polynomial of de-gree 1. The LOSVD was largely constrained by thestrong CO(2 −
0) and CO(3 −
1) bandheads, whichwere contained within the 2 . − . µ m fitting re-gion. We masked the Ca I absorption line because ourtemplate library does not include cool dwarf stars (e.g.,Krajnovi´c et al. 2009), and further excluded a few arti-ficial features, likely the result of imperfect sky subtrac-tion or telluric correction. Example fits with pPXF tothe observed galaxy spectra located at the nucleus, in anintermediate region, and in one of the outermost spatialbins are given in Figure 1.After an initial fit to the galaxy spectrum in each spa-tial bin, we ran a Monte Carlo simulation with 100 it-erations. During each realization, we generated a syn-thetic spectrum by taking the best-fit model and addingrandom Gaussian noise based upon the standard devia-tion of the model residuals. We re-fit the spectrum us-ing pPXF with the penalization turned off. From theresulting distribution for each Gauss-Hermite moment,we took the mean to be the kinematic value and thestandard deviation to be the 1 σ uncertainty. Finally, wepoint-symmetrize the kinematics using the method de-scribed in van den Bosch & de Zeeuw (2010), which alsoremoves the systematic offsets in the odd Gauss-Hermitemoments.The resulting radial velocity ( V ) map shows that Mrk1216 is rotating, such that the southwest side of thegalaxy is blueshifted and the northeast side is redshiftedwith values of ±
180 km s − . The velocity dispersion ( σ )rises from 230 km s − at a projected radius of ∼ ′′ to 425km s − at the nucleus. The third Gauss-Hermite moment( h ), or skewness, falls between ± h − V anti-correlation, which is a common for rotating,axisymmetric systems (e.g., Fisher 1997). The map ofthe fourth Gauss-Hermite moment ( h ), or the kurtosis,has a slight peak at the nucleus to a value of 0.08. Thekinematics have median errors of 15 km s − , 18 km s − ,0.04, and 0.04 for V , σ , h , and h , respectively. Table 1provides the extracted Gauss-Hermite moments for eachNIFS spatial bin, and the 2D velocity fields are shown inFigure 2. TABLE 1NIFS Kinematics x y V ∆ V σ ∆ σ h ∆ h h ∆ h ( ′′ ) ( ′′ ) (km s − ) (km s − ) (km s − ) (km s − ) (7) (8) (9) (10)-0.055 0.003 -1.205 12.708 425.376 17.927 -0.010 0.024 0.075 0.0310.012 -0.069 54.401 14.221 424.049 20.599 -0.023 0.028 0.082 0.0340.013 0.079 -52.747 14.410 420.777 20.519 0.024 0.028 0.083 0.0330.087 -0.012 17.161 13.271 415.750 18.381 -0.000 0.026 0.072 0.0340.107 0.092 -46.839 13.620 369.604 17.681 0.031 0.027 0.063 0.033 Note . — The first two columns provide the x and y locations of the Voronoi bin generators, whilethe remaining columns present the point-symmetrized NIFS kinematics and their uncertainties. Theposition angle is 283.94 ◦ , defined counterclockwise from the galaxy’s major axis to x . This table isavailable in its entirety in machine-readable form. Fig. 1.—
Example fits with pPXF to the observed Mrk 1216 spec-tra located at the nucleus (top), at an intermediate location (mid-dle), and in one of the outermost spatial bins (bottom). The redline shows the optimal stellar template convolved with the LOSVD,whose shape is further adjusted with an additive constant anda first-degree multiplicative Legendre polynomial. Several wave-length regions shown in gray were masked during fit, including theCa I absorption line and artifacts from imperfect data reduction.The green dots are the model residuals that have been shifted up-ward by a constant, arbitrary amount.
We also examined the robustness of the stellar kine-matics by changing how the measurements were madewith pPXF. We modified the fitting region to be longer(2 . − . µ m) and shorter (2 . − . µ m), includedthe Ca I absorption line in the fit, required that the rel-ative mix of template stars remain the same betweenspatial bins, and tested different combinations of degree0 − σ compared to our default fitting ap-proach. We note that there were a number of bins whosekinematics were inconsistent at the 1 σ level, with thenumber of discrepant bins ranging from 0 −
15 depend- ing on the fitting method being adopted. In Section 5.2,we test the effect on M BH if the kinematics from an alter-native fitting approach in which a multiplicative degree2 polynomial and no additive term is used instead. PSF Model
We described the NIFS PSF as the sum of two con-centric, circular Gaussians. Following past work (e.g.,Krajnovi´c et al. 2009; Seth et al. 2014), the MGE in Sec-tion 2 was convolved with the NIFS PSF and comparedto the Mrk 1216 data cube, after collapsing along thewavelength axis. This resulted in best-fit values of 0 . ′′ . ′′
36 for the dispersions, and 0.56 and 0.44 for theweights of the inner and and outer Gaussian components,respectively. The method further allowed us to determinethe center of the NIFS aperture. Attempts to fit a three-Gaussian PSF model produced a negligible componentthat contributed 1% to the total flux. Our characteriza-tion of the NIFS PSF is consistent with expectations ofthe Gemini ALTAIR system (e.g., Gebhardt et al. 2011;Onken et al. 2014; Drehmer et al. 2015).We also measured the PSF using an LGS AO NIFSobservation of the galaxy’s tip-tilt star. With the 2Dimage decomposition package Galfit (Peng et al. 2010),we found that the sum of three circular Gaussians fit thecollapsed NIFS data cube of the star well. The Gaussianshave dispersions of 0 . ′′
07, 0 . ′′
14, and 0 . ′′
37 with weights of0.39, 0.21, and 0.40. Due to the temporal variabilityof the AO correction, and because the observations ofthe star were conducted on-axis in contrast to the off-axis observations of the galaxy, we view this second PSFdetermination as a rough estimate. Nevertheless, we usethis result to test how sensitive the inferred M BH is tothe assumed PSF in Section 5.1. LARGE-SCALE SPECTROSCOPY
The NIFS kinematics are complemented by large-scalespectroscopic observations that provide important con-straints on the stellar mass-to-light ratio and the orbitaldistribution (e.g., Shapiro et al. 2006). The large-scalespectra were obtained with the Potsdam Multi Aper-ture Spectrograph (PMAS; Roth et al. 2005) in the Pmasfiber PAcK (PPAK; Verheijen et al. 2004; Kelz et al.2006) mode from the 3.5 m telescope at Calar Alto Obser-vatory, and from the Marcario Low-Resolution Spectro-graph (LRS; Hill et al. 1998) on the Hobby-Eberly Tele-scope at McDonald Observatory. Yıldırım et al. (2015)and van den Bosch et al. (2015) presented the PPAKand HET data, but we provide a brief summary below.The PPAK integration time was 1.5 hours on-source,
Fig. 2.—
The Mrk 1216 stellar kinematics from NIFS assisted by LGS AO. The measurements are shown on the scale given by the colorbar to the right, and the range of values is provided at the top of the V , σ , h , and h maps. Mrk 1216 exhibits regular rotation, a centralrise in the stellar velocity dispersion, a h − V anti-correlation, and a slight peak in h at the nucleus. The blueshifted velocities correspondto the southwest side of the galaxy. with two 900 s exposures taken at three dither positionsto fully sample the 331 2 . ′′ − R ∼ V , σ , h , and h in 41 Voronoi spatial bins using pPXF, the Indo-U.S. Li-brary of Coud´e Feed Stellar Spectra (Valdes et al. 2004),and an additive Legendre polynomial of degree 15. As afinal step, we point-symmetrized the stellar kinematics.The PSF was reconstructed by comparing the collapsedPPAK data cube to the Mrk 1216 MGE. The PSF has aninner Gaussian component with a dispersion of 1 . ′′
24 thatcontributes 77% to the total flux, while the second Gaus-sian component has a weight of 23% and a dispersion of3 . ′′ ′′ -wide slit aligned withthe galaxy major axis. The g2 grating and 2 × − − . After the ini-tial data processing, we constructed 21 spatial bins andmeasured the stellar kinematics with pPXF and theMILES template library (S´anchez-Bl´azquez et al. 2006;Falc´on-Barroso et al. 2011). Measurements of V and σ were made in each of the spatial bins, with h and h be-ing extracted from the inner 10 bins. The HET data haveslightly better spatial resolution than the PPAK observa-tions, and the PSF is given by the sum of two Gaussianswith dispersions of 1 . ′′
19 and 3 . ′′
39, each weighted by 0.55and 0.45, respectively.The PPAK and HET kinematics extend out to ∼ R e and show features that are very similar to those seen fromNIFS. The large-scale kinematics reveal that the galaxyis rotating with redshifted velocities of ∼
160 km s − tothe northeast, a peak in the central velocity dispersionto ∼
350 km s − , and a clear anti-correlation between h and V . The PPAK and HET kinematics are con-sistent with the NIFS kinematics over the radial extentthey share in common, after accounting for differences inspatial resolution and binning. STELLAR-DYNAMICAL MODELS
In order to constrain the mass of the central black holein Mrk 1216, we calculated three-integral, orbit-baseddynamical models using the triaxial Schwarzschild codeof van den Bosch et al. (2008). We ran the code in theaxisymmetric limit, meaning that triaxial orbit families(e.g., box orbits) are included in the orbital libraries butwe adopt a nearly oblate axisymmetric shape with anintermediate to long axis ratio of 0.99. The assump-tion of axisymmetry is justified by the galaxy’s rotationand the absence of isophotal and kinematic twists in the
HST image and the NIFS/PPAK data. We deprojectedthe MGE in Section 2 using an inclination angle of 70 ◦ .The same inclination was used by Yıldırım et al. (2015),and is mid-way between the range of angles for whichthe MGE can be deprojected given the apparent axisratio of the flattest Gaussian component. Since Mrk1216 does not contain a nuclear dust disk, we are un-able to derive an independent estimate of the inclinationangle, as has been possible for a few nearby galaxies (e.g.,van den Bosch et al. 2012; Yıldırım et al. 2016a).The stellar potential is combined with the gravitationalpotential due to a black hole and a Navarro-Frenk-White(NFW; Navarro et al. 1996) dark matter halo. We cre-ated an orbit library that samples 32 equipotential shellswith logarithmically spaced radii beginning at 0 . ′′ H -band stellar mass-to-light ratio (Υ H ), theconcentration ( c ) of the NFW halo, and the fraction ofdark matter relative to the stellar mass ( f DM ). The stel-lar mass-to-light ratio is assumed to be constant with ra-dius, which is supported by the lack of color gradient in HST
WFC3 F814W and F160W images (Yıldırım et al.2015). We generated model grids that sampled 41 valuesof M BH with 8 . ≤ log( M BH /M ⊙ ) ≤ .
5, 28 values ofΥ H between 0.3 and 3 . ⊙ , and 24 NFW halos with c = 5 , ,
15 and log( f DM ) = 0 . − .
0. Models without adark halo were run as well. While the results of our fidu-cial model grid presented in Section 5.1 were obtained byfitting to the NIFS+PPAK kinematics, in Section 5.2 wealso test fitting NIFS-only and NIFS+HET kinematics.With four Gauss-Hermite moments in 108 NIFS+PPAKspatial bins, there are 432 observables. Ultimately, thebest-fit model is the one with the lowest χ ( χ ), andthe 1 σ statistical uncertainty for a given parameter isset by marginalizing over the other free parameters andsearching for where the change in χ (∆ χ ≡ χ − χ )is 1.0. The 3 σ model fitting error is taken to be where∆ χ = 9 . Modeling Results
We present the results of our fiducial model grid in theleft panel of Figure 3, and provide a comparison betweenthe best-fit model and NIFS/PPAK kinematics in Figure4. We find that M BH = 4 . × M ⊙ , Υ H = 1 . ⊙ , c = 10, and log( f DM ) = 3 .
5, which translates to a halovirial mass of 5 × M ⊙ . This model reproduces theobserved kinematics well, and has a reduced χ of 0.6.The 1 σ statistical uncertainties on M BH and Υ H corre-spond to (4 . +0 . − . ) × M ⊙ and 1 . ± . ⊙ , respec-tively, whereas the 3 σ statistical uncertainties translateto M BH = (4 . +1 . − . ) × M ⊙ and Υ H = 1 . ± . ⊙ .In contrast, the dark halo parameters are not well con-strained. All three values of c are allowed within 3 σ andlog( f DM ) > .
5. Models without a dark halo are clearlyruled out, as χ = 320 for the models without a darkhalo, which corresponds an increase of 65 relative to thebest-fit model incorporating an NFW dark matter halo.Moreover, we build into the M BH and Υ H error bud-gets the effect of adopting a different NIFS PSF model,using unsymmetrized kinematics, and assuming a differ-ent inclination angle. Due to the difficulties in measuringthe AO PSF, it is important to test how other reasonablePSF characterizations might affect M BH . Our fiducialmodel above utilizes a two-Gaussian description deter-mined by comparing the galaxy’s MGE to the Mrk 1216collapsed data cube. If instead the NIFS PSF is taken tobe the sum of three concentric, circular Gaussians mea-sured from the NIFS observation of the galaxy’s tip-tiltstar, we find M BH = 5 . × M ⊙ and Υ H = 1 . ⊙ .A similar change occurs when the observed stellar kine-matics are not forced to be point-symmetric. The fidu-cial model was fit to NIFS and PPAK kinematics thatwere averaged in a two-fold manner over the major andminor axes in order to reduce noise in the kinematic mea-surements. When only the systematic offsets in the oddGauss-Hermite moments are subtracted off, but no otheradjustments are made, we find that M BH = 5 . × M ⊙ and Υ H = 1 . ⊙ .In addition to fitting to the point-symmetricNIFS+PPAK kinematics, the fiducial model was run foran inclination angle of 70 ◦ . Often Schwarzschild mod-els are calculated for a single viewing orientation, asit is computationally expensive to sample over M BH ,Υ H , two dark halo parameters, and the inclination (orthree angles in the case of triaxiality). In the few caseswhere inclination was allowed to vary, the parameterwas not well constrained by the 2D line-of-sight kinemat-ics (Krajnovi´c et al. 2005; van den Bosch & van de Ven2009; Walsh et al. 2012). Therefore, we determined theeffect on M BH and Υ H if a near edge-on angle of 85 ◦ is used instead. We find that M BH decreases by 22% to3 . × M ⊙ and Υ H changes by 23% to 1 . ⊙ .By adding in quadrature the 1 σ formal model fit-ting uncertainty and the percent change in the best-fit values relative to the fiducial model above, we ul-timately find that M BH = (4 . ± . × M ⊙ andΥ H = 1 . ± . ⊙ for Mrk 1216. These values are consis-tent with models fit to only the (bi-symmetrized) PPAKkinematics – Yıldırım et al. (2015) found a black holemass upper-limit of 1 . × M ⊙ and Υ H = 1 . +0 . − . Υ ⊙ ,along with an NFW halo parameterized by c = 10 andlog( f DM ) = 2 . +1 . − . (3 σ statistical uncertainties). In ad-dition, our dynamical H -band mass-to-light ratio is inagreement with expectations from stellar population syn-thesis models for both a Kroupa (1 . ⊙ ) and a Salpeter(1 . ⊙ ) initial mass function, assuming solar metallic-ity and a ∼
13 Gyr age (Vazdekis et al. 1996). Given theblack hole mass of 4 . × M ⊙ and the bulge stellar ve-locity dispersion of 308 km s − (see Section 7), the NIFSdata, with a central 0 . ′′ . ′′
49 black hole sphere of influence.
Additional Models
The PPAK data cube provides an increase in 2D spatialcoverage, more spatial bins, smaller uncertainties on theextracted kinematics, and better spectral resolution thanthe HET long-slit spectroscopy. Thus, we use the PPAKkinematics in place of the HET kinematics when con-structing the dynamical models. If instead Schwarzschildmodels are fit to the NIFS+HET kinematics, we recoverthe same results, with M BH = (4 . +0 . − . ) × M ⊙ andΥ H = 1 . ± . ⊙ (1 σ ).Likewise, we don’t see significant changes in M BH and Υ H if models are constrained by only the NIFSkinematics. By fitting to the NIFS data alone, theblack hole mass is less susceptible to systematic ef-fects that commonly plague stellar-dynamical models,such as assumptions about the dark matter halo (e.g.,Gebhardt & Thomas 2009; Schulze & Gebhardt 2011;Rusli et al. 2013) and the radial form of the stellar mass-to-light ratio (e.g., McConnell et al. 2013). This comesat the expense of larger M BH statistical uncertaintiesdue to the poor constraint on Υ H . Results of fittingorbit-based models to the NIFS kinematics alone areshown in the right panel of Figure 3, and we find that M BH = (5 . +1 . − . ) × M ⊙ and Υ H = 1 . +0 . − . Υ ⊙ (1 σ ).Finally, the NIFS kinematics were measured using anadditive constant and a degree 1 multiplicative polyno-mial to account for differences in continuum shape be-tween the velocity template library and the observedgalaxy spectra. We consider these NIFS kinematics ro-bust, as changes to how pPXF is run (see Section 3.1)produce similar kinematics. However, slight adjustmentsto the degree of the additive/multiplicative polynomialscan cause inconsistent kinematics at the 1 σ level (all areconsistent within 2 σ ). In particular, one of the largestdifferences is seen when running pPXF with a multiplica-tive degree 2 polynomial (with no additive component),which produces 11 spatial bins in which the kinemat-ics differ by more than 1 σ relative to our adopted setof NIFS kinematics. If instead we use the NIFS kine-matics extracted with pPXF and a multiplicative degree2 polynomial, we infer M BH = (3 . +0 . − . ) × M ⊙ and Fig. 3.—
Contours of χ for various stellar-dynamical models (gray dots) with different combinations of black hole mass and H -bandstellar mass-to-light ratio after marginalizing over the dark halo parameters. The red square is the best-fit model, the red contour indicateswhere ∆ χ = 2 .
3, and the subsequent black contours correspond to ∆ χ = 6 . .
8, respectively. These ∆ χ values correspond to 1 σ ,2 σ , and 3 σ confidence regions for two parameters. The results are shown for dynamical models fit to the combination of NIFS and PPAKdata sets (left) and for models fit to only the NIFS kinematics (right). The two grids produce consistent results. Υ H = 1 . ± . ⊙ (1 σ ). We note that this second set ofNIFS kinematics show systematically smaller dispersionscompared to the central PPAK kinematics, hence we per-form this test as a sanity check but do not incorporatethe results into our M BH and Υ H error budgets.The results from each of the three model grids aboveare in agreement with our final black hole mass and mass-to-light ratio measurements for Mrk 1216 of M BH =(4 . ± . × M ⊙ and Υ H = 1 . ± . ⊙ . THE ORBITAL STRUCTURE OF MRK 1216
Not only do the Schwarzschild models provide us withconstraints on the black hole mass and the stellar mass-to-light ratio, but they also allow for an examination ofthe galaxy’s orbital structure. Figure 5 illustrates theamount of anisotropy and the orbit type as a function ofradius. Using the best-fit model from Section 5.1, we plotthe ratio σ r / σ t , where the tangential velocity dispersionis given by σ t = ( σ φ + σ θ ) / r, θ, φ ) are sphericalcoordinates. In order to gain an idea of the uncertain-ties in σ r / σ t , we show all the models within ∆ χ = 1from our fiducial grid, and the best-fit models from thegrid searches in which we assumed a different NIFS PSF,unsymmetrized kinematics, and a near edge-on inclina-tion angle. We find that Mrk 1216 is roughly isotropicwithin the black hole sphere of influence, but becomes ra-dially anisotropic with σ r / σ t ∼ . ′′ . Instead, short-axis tube or-bits dominate our best-fit model, contributing ∼ − THE BLACK HOLE – HOST GALAXY RELATIONS
Placing Mrk 1216 on the black hole–host galaxy re-lations further requires identifying the bulge compo-nent, and currently there is a broad range of measure-ments in the literature. From Galfit models of the
HST
F160W image, Yıldırım et al. (2015) find an upper limiton the H -band bulge luminosity and effective radiusof L H, bul = 8 . × L ⊙ and R e , bul = 3 . ′′
42, and alower limit of L H, bul = 1 . × L ⊙ and R e , bul =1 . ′′ B/T = 0 .
69 and 0 .
13. The former characterization in-cludes a centrally concentrated component with a Sersicindex of 3.61 and a projected axis ratio of 0.56. Sincethis double Sersic model produced pronounced residu-als, Yıldırım et al. (2015) increased the number of Sersicfunctions to four to obtain a good fit to the galaxy’scomplex structure. Each of the four components, how-ever, have rather low Sersic indices between 0.99 and1.61, which complicates a morphological classification. Adynamical decomposition using orbital weights from theYıldırım et al. (2015) best-fit Schwarzschild model sup-port the picture depicted by the 4-component photomet-ric decomposition. In contrast, Savorgnan & Graham(2016) argue that Mrk 1216’s bulge component has an H -band luminosity of ∼ . × L ⊙ based on a one-dimensional (1D) multi-component Sersic fit to the HST
F160W surface brightness brightness profile. Their fitincludes an intermediate-scale disk, in addition to a nu-clear exponential disk and a spheroidal component. BothYıldırım et al. (2015) and Savorgnan & Graham (2016)find that more than one Sersic component is required tomatch the surface brightness despite the elliptical galaxyclassification in the NASA/IPAC Extragalactic Databaseand Hyperleda.
Fig. 4.—
The observed NIFS (left) and PPAK (right) kinematics, plotted as a function of projected radial distance from the nucleus, arecompared to the best-fit stellar dynamical model (red) with M BH = 4 . × M ⊙ and Υ H = 1 . ⊙ . The data are folded and multipleposition angles are depicted. The best-fit model reproduces the kinematic features well, and has a reduced χ of 0.6. Fig. 5.—
Mrk 1216’s orbital structure over the radial extent of theNIFS and PPAK kinematics. The anisotropy (top) and orbit type(bottom) are determined from the best-fit Schwarzschild model(solid lines) and a rough estimate of the uncertainties is depictedby the shaded regions. The galaxy is essentially isotropic (graydashed horizontal line) within the black hole sphere of influence(gray dot-dashed vertical line), and becomes radially anisotropicat larger distances from the nucleus. The orbits are composedmainly of short-axis tubes (red), although a small fraction of boxorbits (blue) are also present.
For Mrk 1216, we conservatively adopt limits that ex-tend from the total quantities down to the smallest bulgemeasurements in the literature, with values set to the re-sult from a 2-component 2D Sersic fit (Yıldırım et al.2015). The galaxy’s total luminosity was determinedfrom the MGE model in Section 2, which also agreeswith a single Sersic Galfit model of the
HST
F160Wimage. After assuming H − K = 0 . K -bandabsolute solar magnitude of 3 .
28 (Binney & Merrifield1998), we establish a bulge luminosity for Mrk 1216 of L K, bul = (9 . +4 . − . ) × L ⊙ . For comparison, the upperbound on L K, bul derived in this manner is 36% smallerthan the growth curve analysis of Two Micron All SkySurvey (Skrutskie et al. 2006) images by van den Bosch(2016). We determine the bulge mass using Υ H =1 . ⊙ from the best-fit model in Section 5.1. Al-though a similar compact, high-dispersion galaxy, NGC1277, showed evidence for a radially varying V -band stel-lar mass-to-light ratio (e.g., Mart´ın-Navarro et al. 2015),there is no HST
WFC3 F814W − F160W color gradi-ent observed for Mrk 1216 (Yıldırım et al. 2015) anda similar Υ H from dynamical models fit to only thesmall-scale NIFS data in Section 5.2 and models fit toonly the large-scale PPAK data (Yıldırım et al. 2015)are found. Hence, adopting Υ H = 1 . ⊙ is justifiedand we find that M bul = (1 . +0 . − . ) × M ⊙ for Mrk1216. Finally, we calculate σ ⋆ for bulge effective radiiof R e , bul =1 . ′′
34, 3 . ′′
42, and 6 . ′′
34, which correspond tothe measurements from a 4-component, a 2-component,and a 1-component Sersic Galfit model (Yıldırım et al.2015). We use our best-fit stellar-dynamical model topredict the luminosity-weighted second moment withina circular aperture whose radius equals R e , bul , while alsoexcluding the region within the black hole sphere of influ-ence (e.g., Gebhardt et al. 2011; McConnell & Ma 2013).Thus, we determine that σ ⋆ = 308 +16 − km s − for Mrk1216, which agrees well with the previous measurement ofthe PPAK velocity dispersion within a circular aperturethat contains half of the light by Yıldırım et al. (2015).As can be seen in Figure 6, Mrk 1216 is an outlier fromthe M BH − L K, bul and M BH − M bul relations, but is con-sistent with M BH − σ ⋆ . Even when using the galaxy’stotal luminosity or total stellar mass, Mrk 1216 falls afactor of ∼ ∼
10 above the Kormendy & Ho (2013)and L¨asker et al. (2014) M BH − L K, bul correlations, aswell as a factor of ∼ −
10 above M BH − M bul de-pending on whether the relation from McConnell & Ma(2013), Kormendy & Ho (2013), Savorgnan et al. (2016),Saglia et al. (2016) is assumed. Thus, using the total lu-minosity (stellar mass) makes Mrk 1216 a 2 . − . σ out-lier (1 . − . σ outlier) from the M BH − L K, bul ( M BH − M bul ) relation given the various calibrations and scatterof the correlations. In Figure 7, we show predictions ofa model with a 5 . × M ⊙ black hole, which is themass expected from M BH − M bul (Saglia et al. 2016) forMrk 1216’s total stellar mass of 1 . × M ⊙ . These σ and h predictions are compared to the NIFS obser-vations and the best-fit model from Section 5.1 with M BH = 4 . × M ⊙ . Our best-fit model exhibits a sim-ilar central velocity dispersion peak and elevated central h values as the observed kinematics, whereas the modelbased on M BH − M bul cannot reproduce these features.For comparison, the two other compact, high-dispersion galaxies from the HET Massive Galaxy Sur-vey are shown in Figure 6. We follow the same con-ventions for characterizing the NGC 1277 and NGC1271 bulge quantities as was used for Mrk 1216. Inparticular, from the MGE descriptions of HST images(van den Bosch et al. 2012; Walsh et al. 2015) and thebest-fit stellar mass-to-light ratios from dynamical mod-els fit to AO observations (Walsh et al. 2015, 2016),we measure total luminosities of L V = 1 . × L ⊙ and L H = 7 . × L ⊙ , and total stellar massesof 1 . × M ⊙ and 1 . × M ⊙ for NGC 1277and NGC 1271, respectively. These values are simi-lar to the total luminosities and masses reported byEmsellem (2013) for NGC 1277 and Graham et al.(2016b) for NGC 1271. Graham et al. (2016a) cal-culate a larger total luminosity for NGC 1277 basedon modeling the 1D light profile, however their MGEmodel suggests a smaller total luminosity than the oneadopted here (their MGE model has 43% less lightthan their 1D component analysis). Despite using to-tal properties, NGC 1277 and NGC 1271 remain out-liers from M BH − L K, bul by 2 . − . σ (Kormendy & Ho2013; L¨asker et al. 2014) and M BH − M bul by 1 . − . σ (McConnell & Ma 2013; Kormendy & Ho 2013;Savorgnan et al. 2016; Saglia et al. 2016). The twogalaxies are in good agreement with the expectations from M BH − σ ⋆ (McConnell & Ma 2013; Kormendy & Ho2013; Saglia et al. 2016; van den Bosch 2016).van den Bosch (2016) conclude that M BH − σ ⋆ is thebest empirical relationship available and that a multi-variate scaling relation between M BH , L K and R e is aprojection of M BH − σ ⋆ with an equal amount of in-trinsic scatter. Previous attempts to explore whetherthe inclusion of an additional parameter leads to tighterscaling relations (e.g., Beifiori et al. 2012; Saglia et al.2016) have also generally found no significant decreasesin the amount of intrinsic scatter compared to the single-parameter M BH − σ ⋆ relation. Given the arguments ofvan den Bosch (2016), it is not surprising that thesecompact galaxies are consistent with M BH − σ ⋆ but withtheir small sizes are outliers from M BH − L bul . Thethree HET compact galaxies are in the compilation ofvan den Bosch (2016). With this new black hole massand luminosities derived from the HST images, we findthat Mrk 1216 and NGC 1271, are within ∼ . σ of theblack hole mass – galaxy luminosity – galaxy size relationcorrelation, and that NGC 1277 is consistent with the re-lation, given the intrinsic scatter. Future work on multi-variate scaling relations requires overcoming a samplingbias in the known M BH hosts; currently there is a verylimited spread in effective radii at a given K -band galaxyluminosity (van den Bosch et al. 2015). DISCUSSION
Mrk 1216 hosts one of the largest black holes dynami-cally detected to date, naturally leading to the questionof how such a massive black hole ended up in a relativelymodest galaxy. One interesting explanation is that Mrk1216 is a relic of the z ∼ z ∼ R e (e.g., Trujillo et al. 2014; Ferr´e-Mateu et al.2015; Mart´ın-Navarro et al. 2015; Yıldırım et al. 2016b,in prep). Mrk 1216 is akin to NGC 1277 and NGC 1271,and is similar to the z ∼ R e = 2 . M BH − L bul and M BH − M bul relationsis that we simply do not have enough measurementsat the upper-end of the correlations. With the lim-ited number of objects in this high-mass regime, nei-0 Fig. 6.—
Location of Mrk 1216 (red square) on the black hole –host galaxy relations. We show multiple calibrations of the correla-tions, but for clarity only display the intrinsic scatter (gray) mea-sured by Saglia et al. (2016) for the black hole mass – stellar veloc-ity dispersion (top) and the black hole mass – bulge mass (bottom)relationships. The intrinsic scatter from both Kormendy & Ho(2013) and L¨asker et al. (2014) are shown for the black hole mass– K -band bulge luminosity (middle) relationship. NGC 1277 andNGC 1271 (blue asterisks) are two compact galaxies similar to Mrk1216 from the HET Massive Galaxy Survey with M BH measure-ments. Due to uncertainties in the bulge components, we showlimits that extend from the total quantities (upper bound of thehorizontal solid line) to the smallest bulge estimates (lower boundof the horizontal dotted line) in the bottom two panels. Even whenusing the total luminosity/stellar mass, Mrk 1216, NGC 1277, andNGC 1271 are outliers from M BH − L K, bul and M BH − M bul , yetare consistent with M BH − σ ⋆ . ther the form of the correlations nor the magnitudeand distribution of the scatter are well determined (e.g.,McConnell & Ma 2013). Therefore, Mrk 1216 could fallin the tails of a distribution between black hole massand galaxy properties that still needs to be fully flushedout. We note that many of the compact galaxies in theHET Massive Galaxy Survey have nuclear dust disks,indicating the presence of cleanly rotating gas (Ho et al.2002; Alatalo et al. 2013), from which independent blackhole mass measurements can be derived for compari-son to the stellar-dynamical determinations. Such cross-checks between mass measurement methods is essentialfor establishing the amount of intrinsic scatter in theblack hole correlations, and eventually for assessing howstrongly Mrk 1216, NGC 1277, and NGC 1271 devi-ate from the relations. Presently, a majority of themeaningful comparisons between the stellar and gas-dynamical methods have led to discrepancies where thestellar-dynamical M BH exceeds the gas-dynamical massby factors of 2 − M BH − M ⋆ correlation,but the simulation is most sensitive to galaxies with M BH ∼ M ⊙ and M ⋆ ∼ M ⊙ . Due to the limitedbox size, predictions cannot be made for more massiveNGC 1277-like galaxies (Barber et al. 2016). Contraryto NGC 1277 and NGC 1271, which are members of thePerseus cluster, Mrk 1216 is an isolated galaxy in thefield, with only two other galaxies within 1 Mpc at itsdistance (Yıldırım et al. 2015). Combined with the reg-ular isophotes and lack of tidal signatures in the HST image, the idea that Mrk 1216 was once the center of amore massive galaxy seems unlikely. CONCLUSION
In summary, we measured the 2D stellar kinematicsof the compact, high-dispersion galaxy Mrk 1216 usingnewly acquired AO NIFS observations that probe withinthe black hole sphere of influence. Mrk 1216 is rotating,has a distinct rise in the stellar velocity dispersion at thenucleus, exhibits the expected anti-correlation between V and h , and has elevated central h values. The high an-gular resolution NIFS kinematics, along with large-scalekinematic measurements and the luminous mass modelfrom an HST image, are fit with stellar-dynamical mod-els based upon the Schwarzschild superposition method.We constrain the mass of the central black hole in Mrk1216 to be (4 . ± . × M ⊙ and the H -band stellarmass-to-light ratio to be 1 . ± . ⊙ . The error bud-get incorporates some possible systematic effects and theformal 1 σ model fitting uncertainties.With σ ⋆ = 308 km s − , Mrk 1216 is consistent with the M BH − σ ⋆ relationship, but is a surprising positive outlierfrom M BH − L bul and M BH − M bul , even when conserva-tively using the galaxy’s total luminosity and stellar mass1 Fig. 7.—
Comparison of the stellar kinematics measured from NIFS (left) to the best-fit model with a 4 . × M ⊙ black hole (middle)and a model with a 5 . × M ⊙ black hole (right), which is the mass expected from M BH − M bul (Saglia et al. 2016) when conservativelyusing Mrk 1216’s total stellar mass of 1 . × M ⊙ . When generating models for the 5 . × M ⊙ black hole, we sampled over a rangestellar mass-to-light ratios and dark matter halos and present the model with the lowest χ . The best-fit model is a good match to thedata, while the model with a smaller black hole cannot reproduce the sharp rise in the velocity dispersion (top) or the elevated h values(bottom) at the nucleus. of L K = 1 . × L ⊙ and M ⋆ = 1 . × M ⊙ . Mrk1216 is similar to NGC 1277 and NGC 1271 – the twocompact, high-dispersion galaxies from the HET MassiveGalaxy Survey that have prior stellar-dynamical M BH measurements from AO observations. All three galax-ies resemble the quiescent galaxies at z ∼ z ∼ M BH − L bul and M BH − M bul relationswere higher at earlier times. In other words, perhapsblack hole growth precedes that of its host galaxy. An-other possibility is that the galaxies are simply unusualand are in the tail of a distribution between M BH andgalaxy properties that still needs to be firmly established.Distinguishing between the two scenarios requires obtain-ing a more complete census of local black holes in a widerange of galaxies with diverse evolutionary histories.Based on observations obtained at the Gemini Ob-servatory acquired through the Gemini Science Archiveand processed using the Gemini IRAF package, whichis operated by the Association of Universities for Re-search in Astronomy, Inc., under a cooperative agree-ment with the NSF on behalf of the Gemini partnership:the National Science Foundation (United States), the Na-tional Research Council (Canada), CONICYT (Chile),the Australian Research Council (Australia), Minist´erioda Ciˆencia, Tecnologia e Inova¸c˜ao (Brazil) and Ministe- rio de Ciencia, Tecnolog´ıa e Innovaci´on Productiva (Ar-gentina), under program GN-2013A-Q-1. Also based onobservations made with the NASA/ESA Hubble SpaceTelescope, obtained at the Space Telescope Science In-stitute, which is operated by the Association of Univer-sities for Research in Astronomy, Inc., under NASA con-tract NAS 5-26555. These observations are associatedwith program REFERENCESAlatalo, K., Davis, T. A., Bureau, M., et al. 2013, MNRAS, 432,1796Barber, C., Schaye, J., Bower, R. G., et al. 2016, MNRAS, 460,1147Barth, A. J., Darling, J., Boizelle, B. D., et al. 2016a, ApJ, 823, 51 Barth, A. J., Boizelle, B. D., Darling, J., et al. 2016b, ApJ, 822,L28Beifiori, A., Courteau, S., Corsini, E. M., & Zhu, Y. 2012,MNRAS, 419, 2497 Bender, R., Saglia, R. P., & Gerhard, O. E. 1994, MNRAS, 269,785Binney, J., & Merrifield, M. 1998, Galactic Astronomy. PrincetonUniv. Press, Princeton, NJBoccas, M., Rigaut, F., Bec, M., et al. 2006, Proc. SPIE, 6272,114Cappellari, M., & Copin, Y. 2003, MNRAS, 342, 345Cappellari, M., & Emsellem, E. 2004, PASP, 116, 138Crain, R. A., Schaye, J., Bower, R. G., et al. 2015, MNRAS, 450,1937Cretton, N., & van den Bosch, F. C. 1999, ApJ, 514, 704de Francesco, G., Capetti, A., & Marconi, A. 2006, A&A, 460, 439Drehmer, D. A., Storchi-Bergmann, T., Ferrari, F., Cappellari,M., & Riffel, R. A. 2015, MNRAS, 450, 128Emsellem, E., Monnet, G., & Bacon, R. 1994, A&A, 285, 723Emsellem, E. 2013, MNRAS, 433, 1862Fabian, A. C. 1999, MNRAS, 308, L39Falc´on-Barroso, J., S´anchez-Bl´azquez, P., Vazdekis, A., et al.2011, A&A, 532, 95Ferrarese, L., & Merritt, D. 2000, ApJ, 539, L9Ferr´e-Mateu, A., Mezcua, M., Trujillo, I., Balcells, M., & van denBosch, R. C. E. 2015, ApJ, 808, 79Fisher, D. 1997, AJ, 113, 950Gebhardt, K., Bender, R., Bower, G., et al. 2000, ApJL, 539, L13Gebhardt, K., & Thomas, J. 2009, ApJ, 700, 1690Gebhardt, K., Adams, J., Richstone, D., et al. 2011, ApJ, 729, 119Gonzaga, S., Hack, W., Fruchter, A, & Mack, J., 2012, TheDrizzlePac Handbook (Baltimore, STScI)Graham, A. W., Durre´e, M., Savorgnan, G. A. D., et al. 2016a,ApJ, 819, 43Graham, A. W., Ciambur, B. C., & Savorgnan, G. A. D. 2016b,accepted by ApJ, arXiv:1608.00711Greene J. E., Peng, C. Y., Kim, M., et al. 2010, ApJ, 721, 26Greene J. E., Seth, A., Kim, M., et al. 2016, ApJ, 826, L32G¨ultekin, K., Richstone, D. O., Gebhardt, K., et al. 2009, ApJ,698, 198H¨aring, N., & Rix, H.-W. 2004, ApJ, 604, L89Herriot, G., Morris, S., Anthony, A., et al. 2000, Proc. SPIE,4007, 115Hill, G. J., Nicklas, H. E., MacQueen, P. J., et al. 1998, Proc.SPIE, 3355, 375Ho, L. C., Sarzi, M., Rix, H.-W., et al. 2002, PASP, 114, 137Husemann, B., Jahnke, K., S´anchez, S. F., et al. 2013, A&A, 549,87Jahnke, K., & Macci`o, A. V. 2011, ApJ, 734, 92Kelz, A., Verheijen, M. A. W., Roth, M. M., et al. 2006, PASP,118, 129Kormendy, J., & Richstone, D. 1995, ARA&A, 33, 581Kormendy, J., & Ho, L. C. 2013, ARA&A, 51, 511Krajnovi´c, D., Cappellari, M., Emsellem, E., McDermid, R. M.,& de Zeeuw, P. T. 2005, MNRAS, 357, 1113Krajnovi´c, D., McDermid, R. M., Cappellari, M., & Davies, R. L.2009, MNRAS, 399, 1839Krist, J., & Hook, R. 2004, The Tiny Tim User’s Guide(Baltimore:STScI)Lauer, T. R., Faber, S. M., Richstone, D., et al. 2007a, ApJ, 662,808Lauer, T. R., Tremaine, S., Richstone, D., & Faber, S. M. 2007b,ApJ, 670, 249L¨asker, R., Ferrarese, L., van de Ven, G., & Shankar, F. 2014,ApJ, 780, 70L¨asker, R., Greene, J. E., Seth, A., et al. 2016, ApJ, 825, 3Marconi, A., & Hunt, L. K. 2003, ApJ, 589, L21Marconi, A., Risaliti, G., Gilli, R., et al. 2004, MNRAS, 351, 169Mart´ın-Navarro, I., La Barbera, F., Vazdekis, A., et al. 2015,MNRAS, 451, 1081McConnell, N. J., Ma, C.-P., Murphy, J. D., et al. 2012, ApJ,756, 179McConnell, N. J., & Ma, C.-P. 2013, ApJ, 764, 184McConnell, N. J., Chen, S.-F. S, Ma, C.-P., et al. 2013, ApJ, 768,21McGregor, P. J., Hart, J., Conroy, P. G., et al. 2003, Proc. SPIE,4841, 1581Monnet, G., Bacon, R., & Emsellem, E. 1992, A&A, 253, 366Mould, J. R., Huchra, J. P., Freedman, W. L., et al. 2000, ApJ,529, 786 Navarro, J. F., Frenk, C. S., & White, S. D. M. 1996, ApJ, 462,563Onken, C. A., Valluri, M., Brown, J. S., et al. 2014, ApJ, 791, 37Peng, C. Y. 2007, ApJ, 671, 1098Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2010, AJ,139, 2097Robertson, B., Hernquist, L., Cox, T. J., et al. 2006, ApJ, 641, 90Roth, M. M., Kelz, A., Fechner, T., et al. 2005, PASP, 117, 620Rusli, S. P., Thomas, J., Erwin, P., et al. 2011, MNRAS, 410,1223Rusli, S. P., Thomas, J., Saglia, R. P., et al. 2013, AJ, 146, 45Saglia, R. P., Opitsch, M., Erwin, P., et al. 2016, ApJ, 818, 47S´anchez, S. F., Kennicutt, R. C., Gil de Paz, A., et al. 2012,A&A, 538, 8S´anchez-Bl´azquez, P., Peletier, R. F., Jim´enez-Vicente, J., et al.2006, MNRAS, 371, 703Sani, E., Marconi, A., Hunt, L. K., & Risaliti, G. 2011, MNRAS,413, 1479Savorgnan, G. A. D., Graham, A. W., Marconi, A., & Sani, E.2016, ApJ, 817, 21Savorgnan, G. A. D., & Graham, A. W. 2016, MNRAS, 457, 320Scharw¨achter, J., Combes, F., Salom´e, P., Sun, M., & Krips, M.2016, MNRAS, 457, 4272Schaye, J., Crain, R. A., Bower, R. G., et al. 2015, MNRAS, 446,521Schulze, A., & Gebhardt, K. 2011, ApJ, 729, 21Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103Seth, A. C., van den Bosch, R., Mieske, S., et al. 2014, Natur,513, 398Shankar, F., Weinberg, D. H., & Miralda-Escud´e, J. 2009, ApJ,690, 20Shapiro, K. L., Cappellari, M., de Zeeuw, T., et al. 2006,MNRAS, 370, 559Silk, J., & Rees, M. J. 1998, A&A, 331, L1Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131,1163Thomas, J., Ma, C.-P., McConnell, N. J., et al. 2016, Natur, 532,340Trujillo, I., Cenarro, A. J., de Lorenzo-C´aceres, A., et al. 2009,ApJL, 692, L118Trujillo, I., Ferr´e-Mateu, A., Balcells, M., Vazdekis, A., &S´anchez-Bl´azquez, P. 2014, ApJL, 780, L20Valdes, F., Gupta, R., Rose, J. A., Singh, H. P., & Bell, D. J.2004, ApJS, 152, 251van den Bosch, R. C. E., van de Ven, G., Verolme, E. K.,Cappellari, M., & de Zeeuw, P. T. 2008, MNRAS, 385, 647van den Bosch, R. C. E., & van de Ven, G. 2009, MNRAS, 398,1117van den Bosch, R. C. E., & de Zeeuw, P. T. 2010, MNRAS, 401,1770van den Bosch, R. C. E., Gebhardt, K., G¨ultekin, K., et al. 2012,Nature, 491, 729van den Bosch, R. C. E., Gebhardt, K., G¨ultekin, K., Yıldırım,A., & Walsh, J. L. 2015, ApJS, 218, 10van den Bosch, R. C. E. 2016, ApJ, 831, 134van der Marel, R. P., & Franx, M. 1993, ApJ, 407, 525van der Wel, A., Rix, H.-W., Wuyts, S., et al. 2011, ApJ, 730, 38van der Wel, A., Bell, E. F., H¨aussler, B., et al. 2012, ApJS, 203,24van Dokkum, P. G., Whitaker, K. E., Brammer, G., et al. 2010,ApJ, 709, 1018Vazdekis, A., Casuso, E., Peletier, R. F., & Beckman, J. E. 1996,ApJS, 106, 307Verheijen, M. A. W., Bershady, M. A., Andersen, D. R., et al.2004, AN, 325, 151Walsh, J. L., van den Bosch, R. C. E., Barth, A. J., & Sarzi, M.2012, ApJ, 753, 79Walsh, J. L., Barth, A. J., Ho, L. C., & Sarzi, M. 2013, ApJ, 770,86Walsh, J. L., van den Bosch, R. C. E., Gebhardt, K., et al. 2015,ApJ, 808, 183Walsh, J. L., van den Bosch, R. C. E., Gebhardt, K., et al. 2016,ApJ, 817, 2Wellons, S., Torrey, P., Ma, C.-P., et al. 2016, MNRAS, 456, 1030Winge, C., Riffel, R. A., & Storchi-Bergmann, T. 2009, ApJS,185, 1863