THE CHANDRA VIEW OF NGC 4178: THE LOWEST MASS BLACK HOLE IN A BULGELESS DISK GALAXY?

There is mounting evidence that a significant fraction of supermassive black holes (SMBHs) reside in late-type galaxies, and that a classical bulge is not a requirement for an SMBH to form and grow (Filippenko & Ho 2003; Barth et al. 2004; Greene & Ho 2004, 2007; Satyapal et al. 2007, 2008, 2009; Dewangen et al. 2008; Ghosh et al. 2008; Mathur et al. 2008; Shields et al. 2008; Barth et al. 2009; Desroches & Ho 2009; Gliozzi et al. 2009; Jiang et al. 2011a, 2011b; McAlpine et al. 2011). In most late-type galaxies that host an active galactic nucleus (AGN), however, the galaxies have a pseudobulge component, characterized by an exponential surface brightness profile. And while classical bulges are believed to form through mergers, pseudobulges are thought to form through secular processes (Kormendy & Kennicutt 2004). Of the late-type, AGN-hosting galaxies, bulgeless (no evidence even for a pseudobulge) galaxies are by far the rarest. To date, there are only three such bulgeless disk galaxies that are confirmed to host SMBHs: NGC 4395 (Filippenko & Ho 2003; Shih et al. 2003; Peterson et al. 2005), NGC 1042 (Shields et al. 2008), and NGC 3621 (Satyapal et al. 2007, 2009; Barth et al. 2009; Gliozzi et al. 2009). While very large SMBHs (≳ 10⁶ M_☉) likely form through galaxy mergers (e.g., Kauffmann & Haehnelt 2000), leading to a tight correlation between the black hole mass, M_BH, and the host galaxy's bulge velocity dispersion σ (e.g., Gebhardt et al. 2000; Ferrarese & Merritt 2000; Haehnelt & Kauffmann 2002), it is still unclear how SMBHs form and grow in bulgeless galaxies. Central to this question is how SMBHs affect, or are affected by, their host galaxy properties. It has already been shown that the presence and properties of SMBHs likely do not correlate with galaxy disks (e.g., Kormendy et al. 2011; see also Ho 2007), although SMBHs may correlate with galaxy pseudobulges and bars (e.g., Hu 2008; Graham 2008; Graham et al. 2011). Interestingly, however, all three bulgeless disk galaxies with SMBHs have nuclear star clusters (NSCs), and there is growing evidence that suggests that the mass of SMBHs and NSCs may be correlated in galaxies that possess both (e.g., Seth et al. 2008; Graham & Spitler 2009).

Given their rarity, determining the properties of SMBHs in bulgeless disk galaxies is crucial to our understanding of the low end of the SMBH mass function and its relation to host galaxies. Observationally, the only viable method for finding SMBHs in bulgeless galaxies is through the search for AGNs. Since bulgeless galaxies are typically dusty, star-forming galaxies, a putative AGN is likely to be missed by optical surveys. X-ray observations are the ideal tool to search for AGNs in such galaxies since X-rays are generally only produced in the inner nuclear regions of an AGN, and hard X-rays are not substantially affected by absorption.

The goal of this paper is to investigate the X-ray properties of the putative SMBH that lurks at the center of NGC 4178, a bulgeless disk galaxy that was recently found to have prominent mid-IR [Ne v] emission associated with the nucleus (Satyapal et al. 2009). While the [Ne v] emission strongly suggests the presence of an AGN, the size of the Spitzer Infrared Spectrograph slit (47 × 113, at 14.3 μm) precludes us from confirming the nuclear origin of the emission. Thus, the presence of a significant X-ray point-source counterpart in high-resolution Chandra data would provide a confirmation of a nuclear SMBH, as X-ray emission associated with starburst activity is generally extended (e.g., Dudik et al. 2005; González-Martín et al. 2006; Flohic et al. 2006). If it does have an SMBH, NGC 4178 will be only the fourth known truly bulgeless disk galaxy with an SMBH.

NGC 4178 is a highly inclined (i ∼ 70°), SB(rs)dm galaxy (de Vaucouleurs et al. 1991) located within the Virgo Cluster at a distance of 16.8 Mpc (Tully & Shaya 1984). Based on its nuclear optical spectrum, it is classified as having an H ii nucleus (Ho et al. 1997) and contains an NSC of ∼5 × 10⁵ M_☉ (Böker et al. 1999; Satyapal et al. 2009). Other than some asymmetric, locally enhanced Hα emission near the outer parts of the disk, the Hα distribution of NGC 4178 is typical of that found in star-forming galaxies (Koopmann & Kenney 2004). The H i distribution is more extended than the optical part of the galaxy and shows no evidence of interactions (Cayatte et al. 1990). Apart from our previous Spitzer observations, there is no evidence for an AGN in this galaxy. With these considerations, the presence of an SMBH in this galaxy is highly unexpected.

This paper is structured as follows. In Section 2, we describe the Chandra observations and data reduction, as well as archival Very Large Array (VLA) data. We follow with a description of our results, including X-ray spectral modeling, in Section 3. In Section 4, we discuss constraints on the nuclear black hole using the bolometric luminosity by presenting an updated $L_{\mathrm{bol}}/L_{\mathrm{[\rm Ne\,\mathsc{v}]}}$ correlation. We compute the Eddington mass and compare our findings with mass estimates from other methods. In Section 5, we compare our results to other bulgeless disk galaxies, and we give a summary and our main conclusions in Section 6.

2.1. Chandra Data

We observed NGC 4178 with Chandra ACIS-S for 36 ks on 2011 February 19. The data were processed using CIAO v. 4.3 and we retained only events in the energy range 0.2–10 keV. We also checked that no background flaring events occurred during the observation.

We used the XSPEC v. 12.7.0 software package (Arnaud 1996; Dorman & Arnaud 2001) for the spectral analysis. For the bright off-nuclear source, we re-binned the spectrum in order to contain at least 15 counts per channel in order to use the χ² statistic. To compute the error (90% confidence) on the flux, we used the cflux model component available in XSPEC as a means to estimate fluxes and errors due to model components (Arnaud et al. 2012). For the other sources where low counts (≲ 50 counts) prevented a direct spectral fit, we employed X-ray hardness ratios as a rough estimator of spectral state. The hardness ratio we use is defined as

$$\frac{\mathrm{hard}}{\mathrm{soft}}\equiv {\frac{\mathrm{counts[\mathrm{2\hbox{--}10 \,keV}]}}{\mathrm{counts[\mathrm{0.2\hbox{--}2 \,keV}]}}}.$$

2.2. Archival VLA Data

As can be seen in Figure 1, Chandra clearly detects a nuclear X-ray source (source A), which appears to be situated symmetrically between two nearly mirrored infrared lobes and is located at ${\rm R.A.}=12^{\mathrm{h}}12^{\mathrm{m}}46\mbox{$.\!\!^{\mathrm s}$}32$, decl. = 10°51'5461. Centroid analysis reveals that the infrared lobes are at the same distance from the nuclear source, about 78, corresponding to a distance of 1.3 kpc at the distance of NGC 4178. These infrared lobes are likely associated with star formation regions, known from Hα studies to be associated with, and confined to, the bar (e.g., Martin & Kennicutt 2001). This is supported by the presence of Pa-α emission coincident with these lobes (Figure 2). Two weaker off-nuclear sources (B and C) are located at ${\rm R.A.}=12^{\mathrm{h}}12^{\mathrm{m}}47\mbox{$.\!\!^{\mathrm s}$}33$, decl. = 10°52'2252 32'' and R.A. = $12^{\mathrm{h}}12^{\mathrm{m}}48\mbox{$.\!\!^{\mathrm s}$}47$, decl. = 10°52'1103, 357 (5.2 kpc) and 36'' (5.9 kpc) from the nuclear source, respectively. A third, brighter off-nuclear source (D) is located at ${\rm R.A.}=12^{{\rm h}}12^{{\rm m}}44\mbox{$.\!\!^{\mathrm s}$}51$, decl. = 10°51'1364, 49'' (8 kpc) from the nuclear source. Only one of the off-nuclear sources, source C, appears to have a counterpart in any other band.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In order to determine if the nuclear X-ray source is consistent with the photocenter of the galaxy, the source coordinates were compared to the Two Micron All Sky Survey (2MASS) photocenter. The 2MASS photocenter was found to be at ${\rm R.A.}=12^{\mathrm{h}}12^{\mathrm{m}}46\mbox{$.\!\!^{\mathrm s}$}34$, decl. = 10°51'551, 06 ± 06 off from the X-ray source coordinates. Thus, we conclude that the nuclear X-ray source is coincident with the 2MASS photocenter, within the astrometric uncertainties. After spatially registering the Hubble Space Telescope (HST) H-band image with the 2MASS H-band image, we find that the Chandra source is coincident with the NSC (Figure 2).

3.1. The Nuclear Source

The nuclear X-ray source is shown in Figure 3. We detect 37 ± 7 (5.3σ) X-ray counts (0.2–10 keV) from the source. The nuclear source is strikingly soft, with 31 ± 6 counts in the 0.2–2 keV band and 5 ± 2 counts in the 2–10 keV band (for low counts, Poisson statistics are used to calculate the uncertainty, as described in the approach of Gehrels 1986, where the error corresponds to the 84.13% confidence limit), yielding a hardness ratio of 0.16^+0.12_{− 0.08}. Using a simplified phenomenological power-law (PL) model with a Galactic absorption of 1.91 × 10²⁰ cm⁻² (Dickey & Lockman 1990; Chandra Colden tool), and adopting the global intrinsic absorption for NGC 4178 of N_H ≃ 10²¹ cm⁻² (Cayatte et al. 1994), the observed hardness ratio can be replicated with a photon index Γ = 2.6^+0.6_{− 0.5}. Using these parameters, this model predicts an X-ray luminosity of L_{0.2–10 keV} = 4.6^+3.6_{− 1.4} × 10³⁸ erg s⁻¹ and the hard X-ray luminosity of L_2–10 keV = 7.9^+9.1_{− 4.9} × 10³⁸ erg s⁻¹, significantly lower than that implied by the observed [Ne v] luminosity.⁷ Indeed, the bolometric luminosity inferred by the [Ne v] luminosity is ∼10⁴³ erg s⁻¹ (see Section 4.1), five orders of magnitude higher than the hard X-ray luminosity predicted by this simplified phenomenological PL model. The bolometric correction factor is κ_2–10 keV ∼ 10⁵, much too high for any realistic AGN spectral energy distribution (SED; Vasudevan & Fabian 2009).⁸

Zoom In Zoom Out Reset image size

We therefore postulate that there is highly localized absorption around the central source. Indeed, analysis of data from Chandra Deep Field-North has shown that more complex spectra are characteristic of AGNs in local galaxies due to the ubiquity of heavy absorption, and a simple PL cannot reliably be estimated from hardness ratios (Brightman & Nandra 2012). With a covering fraction of 0.99 and absorption N_H = 5 × 10²⁴ cm⁻², a partially absorbed scenario combined with a simple PL can account for the hardness ratio with Γ = 2.3^+0.6_{− 0.5}. With these parameters, the intrinsic X-ray luminosity is L_{0.2–10 keV} = 3.1^+1.5_{− 0.5} × 10⁴⁰ erg s⁻¹ and the hard X-ray luminosity is L_2–10 keV = 8.6^+10.4_{− 5.5} × 10³⁹ erg s⁻¹. The partially absorbed scenario can be physically interpreted as a strongly accreting black hole imbedded in a heavy medium in which a few holes have been "punched out" by strong X-ray flux. This is in line with our understanding of black hole accretion and, our understanding of the interstellar environment of late-type spiral galaxies, and it is consistent with the bolometric luminosity.

We point out that the observed X-ray luminosity is low, and therefore by itself we cannot exclude the possibility that it is produced by X-ray binaries in the NSC. However, the [Ne v] luminosity is 8.23 × 10³⁷ erg s⁻¹, which is at the low end of the range observed in standard optically identified AGNs (Pereira-Santaella et al. 2010). Indeed, the [Ne v] luminosity in NGC 4178 exceeds that of NGC 3621, which has recently been confirmed to have an optical Seyfert 2 spectrum using high-resolution Keck observations (Barth et al. 2009). Moreover, the bolometric luminosity implied by the [Ne v] luminosity is two orders of magnitude greater than that of NGC 4395, which is indisputably an AGN (Filippenko & Ho 2003) (see Section 4.1). Therefore, the [Ne v] luminosity, combined with our X-ray results, strongly suggests that the X-ray source is due to an AGN.

3.2. Off-nuclear Sources

3.2.1. Source B

We detected 20 ± 5 net counts in the 0.2–10 keV range for source B. The hardness ratio is 0.41^+0.38_{− 0.22}. At a radial distance of 316, the known intrinsic absorption for NGC 4178 is N_H ≃ 10²¹ cm⁻² (Cayatte et al. 1994). With this absorption, source B's hardness ratio can be reproduced by an absorbed PL model with Γ = 1.7^+0.7_{− 0.6}. The corresponding X-ray luminosity is L_{0.2–10 keV} = 2.2^+1.5_{− 0.5} × 10³⁸ erg s⁻¹. In the hard X-ray band, the corresponding luminosity is L_2–10 keV = 1.2^+1.7_{− 0.8} × 10³⁸ erg s⁻¹.

3.2.2. Source C

The net counts in the 0.2–10 keV range for source C are 26 ± 5, with a hardness ratio of 0.20^+0.19_{− 0.12}. This relatively low hardness ratio can be reproduced by an absorbed PL model with Γ = 2.3^+0.9_{− 0.6} and N_H = 10²¹ cm⁻², giving an X-ray luminosity of L_{0.2–10 keV} = 6.2^+6.8_{− 1.2} × 10³⁸ erg s⁻¹ and a hard X-ray luminosity of L_2–10 keV = 1.7^+2.4_{− 1.2} × 10³⁸ erg s⁻¹.

3.2.3. Source D

Our brightest X-ray source, source D, had 575 ± 24 total counts in the 0.2–10 keV range, enough to directly fit a spectrum. We tested the variability of this source by extracting the light curve in the 0.2–10 keV range and applying a χ² test and found that the source flux was likely constant during the observation (χ²/dof = 22.584/39 and $P_{\chi ^{2}}=0.98$).

The spectrum is well fit (χ²_red = 0.997 for 33 degrees of freedom) using an absorbed PL model with Γ = 1.24 ± 0.12 and setting the intrinsic absorption N_H = 10²¹ cm⁻². This yields an X-ray luminosity of L_{0.2–10 keV} = (7.9 ± 0.8) × 10³⁹ erg s⁻¹ and a hard X-ray luminosity of L_2–10 keV = (5.9 ± 0.9) × 10³⁹ erg s⁻¹. If we allow the intrinsic absorption to vary, a comparable fit (χ²_red = 1.02 for 32 degrees of freedom) is achieved with N_H = (1.5 ± 0.14) × 10²⁰ cm² and Γ = 1.31 ± 0.23, yielding an intrinsic X-ray luminosity of L_{0.2–10 keV} = (8.0 ± 1.0) × 10³⁹ erg s⁻¹ and a hard X-ray luminosity of L_2–10 keV = (5.8 ± 0.7) × 10³⁹ erg s⁻¹.

This large luminosity gives us a second method of spectrally characterizing the source. Recent work has shown that the physically motivated Bulk Motion Comptonization (BMC) model may provide an effective means of estimating the mass of accreting black holes (Gliozzi et al. 2009, 2011; Shaposhnikov & Titarchuk 2009). This technique relies on the self-similarity of black holes and their accretion characteristics to relate the photon index Γ with the BMC model normalization N_BMC. In short, the BMC model convolves the inverse Comptonization of X-ray photons by thermalized electrons with the inverse Comptonization of X-ray photons by electrons with bulk relativistic motion (see Titarchuk et al. 1997 for details on the BMC model). The BMC model has four free parameters: the temperature kT, the spectral index α, related to the photon index Γ by α = Γ − 1, a parameter log A related to the fraction of Comptonized seed photons f by A = f/(f − 1), and the model normalization N_BMC.

The BMC model gives an excellent fit (χ²_red = 1.00 for 31 degrees of freedom) with the intrinsic absorption set to N_H = 5 × 10²⁰ cm⁻², kT = 0.48^+0.08_{− 0.07}, α = 0.77^+0.44_{− 0.25}, and log A = 7.1 (Figure 4). The normalization is N_BMC = 1.39 × 10⁻⁶. The corresponding X-ray luminosity is L_{0.2–10 keV} = (6.5 ± 0.4) × 10³⁹ erg s⁻¹, and the hard X-ray luminosity is L_2–10 keV = (5.1 ± 0.5) × 10³⁹ erg s⁻¹. Five different spectral patterns of Galactic black hole systems with mass and distance well constrained are provided in Gliozzi et al. (2011). These are the reference sources utilized in the X-ray scaling method. To be conservative in the estimate of the black hole mass of source D, we have used all five reference patterns and computed the average and standard deviation of the five M_BH values, resulting in 〈M_BH〉 = (6 ± 2) × 10³ M_☉.

Zoom In Zoom Out Reset image size

3.2.4. The Nature of the Off-nuclear Sources

With the exception of source C, the lack of noticeable counterparts in other bands (see Figure 1) suggests that these sources are not foreground objects. Source C is coincident with an extended (∼800 × 520 pc) region (Figure 5), especially evident in high-resolution HST data. This region appears to be an area of extensive star formation, though it is not associated with significant Pa-α emission. A priori, we cannot rule out that these sources are background AGNs that appear to be located within NGC 4178 by chance. With the known hydrogen column density of ∼10²¹ cm⁻², a background X-ray source would not be heavily absorbed by NGC 4178, even in the soft band. Using results from the Chandra Deep Field-South, we can calculate the likely number of hard X-ray sources in our field (∼4 × 10⁻⁴ deg⁻²) for a given flux that could be expected to occur by chance (e.g., Tozzi et al. 2001). For sources B and C, with hard X-ray fluxes on order of ∼10⁻¹⁵ erg s⁻¹ cm⁻², the expected number of sources is ∼5, so sources B and C may indeed be background objects. However, for source D, with hard X-ray flux on order of ∼10⁻¹³ erg s⁻¹ cm⁻², the expected number of sources is ∼0.1. We therefore conclude that source D is likely local to NGC 4178. Source D has a luminosity of ∼10⁴⁰ erg s⁻¹ and is therefore consistent with an ultraluminous X-ray source (ULX, e.g., Swartz et al. 2004; Winter et al. 2006; Berghea et al. 2008). Source D's spectrum is also well fit by the BMC model, suggesting that it may be an intermediate-mass black hole (IMBH). VLA observations by Cayatte et al. (1990) of neutral hydrogen in NGC 4178 show a heightened concentration (N_H = 1.48 × 10²² cm⁻²) of neutral hydrogen coincident with source D. Niklas et al. (1995) give the position of this source as R.A. ≃ 12^h12^m43^s, decl. ≃ 10°50'59''. This concentration is very diffuse, however, so it is uncertain that it is related to source D.

Zoom In Zoom Out Reset image size

4.1. Bolometric Luminosity and Eddington Mass Limit of the AGN

Our IR and X-ray observations allow us to derive a mass estimate for the nuclear black hole. In our previous work, we showed that the [Ne v] luminosity is tightly correlated with the bolometric luminosity of the AGN in a sample of optically identified AGNs. We can therefore use the [Ne v] luminosity to obtain an estimate of the Eddington mass of the black hole (Satyapal et al. 2007) and therefore set a lower mass limit on the AGN. Since the publication of our previous work, there have been a number of additional mid-IR [Ne v] fluxes available in the literature. We therefore update our previously published relation between the [Ne v] 14.32 μm luminosity and the AGN bolometric luminosity using the most up-to-date mid-IR line fluxes of AGNs observed by Spitzer (Haas et al. 2005; Weedman et al. 2005; Ogle et al. 2006; Dudik et al. 2007, 2009; Gorjian et al. 2007; Cleary et al. 2007; Armus et al. 2007; Deo et al. 2007; Tommasin et al. 2008, 2010; Dale et al. 2009; Veilleux et al. 2009; Pereira-Santaella et al. 2010) that have well-characterized nuclear SEDs. In Figure 6, we plot L_bol versus $L_{\mathrm{[\rm Ne\,\mathsc{v}]}}$, which shows a strong correlation. The Spearman rank correlation coefficient is 0.83, with a probability of chance correlation of 10⁻¹³. The bolometric luminosities for this updated sample ranged from ∼5 × 10⁴² erg s⁻¹ to ∼5 × 10⁴⁶ erg s⁻¹, and the black hole masses ranged from log M_BH = 6.15 to log M_BH = 9.56. The best-fit linear relation yields

$$\log {L_{\mathrm{bol}}}=0.615\log {L_{\mathrm{[\rm Ne\,\mathsc{v}]}}}+19.647 \,{\rm erg} \,{\rm s}^{-1}$$

with an rms scatter of 0.53 dex. Using the known [Ne v] 14.32 μm nuclear luminosity for NGC 4178 of 8.23 × 10³⁷ erg s⁻¹ (Satyapal et al. 2009), the predicted nuclear bolometric luminosity of the AGN is L_bol = 9.2⁺²²_{− 6.5} × 10⁴² erg s⁻¹. The Eddington mass limit for the nuclear black hole in NGC 4178 is then M_BH ⩾ 7.1⁺¹⁷_{− 5.0} × 10⁴ M_☉.

Zoom In Zoom Out Reset image size

4.2. Archival VLA Data and the Fundamental Plane

4.3. Other Considerations

With a solid X-ray source detection at the center of NGC 4178, we can add this galaxy to the growing collection of bulgeless, extremely late-type disk galaxies with confirmed AGNs. The best-studied definitively bulgeless disk galaxy with an AGN is NGC 4395, which shows the hallmark signatures of a Type 1 AGN (e.g., Filippenko & Ho 2003; Lira et al. 1999; Moran et al. 1999). The bolometric luminosity of the AGN is ∼10⁴⁰ erg s⁻¹ (Filippenko & Ho 2003), nearly three orders of magnitude lower than the estimated bolometric luminosity of the AGN in NGC 4178. The black hole mass of NGC 4395, determined by reverberation mapping, is M_BH = (3.6 ± 1.1) × 10⁵ M_☉ and does not appear to be radiating at a high Eddington ratio (Peterson et al. 2005). The bolometric luminosity of the AGN in NGC 1042, as estimated from Hα measurements, is ∼8 × 10³⁹ erg s⁻¹, and the central black hole is estimated to be between 60 M_☉ and 3 × 10⁶ M_☉ (Shields et al. 2008). With the updated $L_{\mathrm{bol}}/L_{\mathrm{[\rm Ne\,\mathsc{v}]}}$ relationship, the estimated bolometric luminosity of the AGN in NGC 3621 is 6.8⁺¹⁶_{− 4.8} × 10⁴² erg s⁻¹ and the Eddington mass is 5.2⁺¹²_{− 3.7} × 10⁴ M_☉, making NGC 4178 and NGC 3621 the most luminous AGNs in extremely late-type galaxies currently known. The likelihood that the AGN in NGC 4178 is heavily absorbed, combined with the high photon index, suggests that it is accreting at a high rate. The black hole mass of the AGN in NGC 4178 is ∼10⁴–10⁵ M_☉, possibly lower than the black hole in NGC 4395.

We have analyzed the X-ray characteristics of NGC 4178 from a 36 ks Chandra observation. The X-ray data, combined with considerations from the mid-IR and radio properties of the galaxy, have led us to the following results:

• 1.  
  There is a faint but statistically significant (5.3σ), unresolved X-ray source at the center of NGC 4178, confirming the presence of an AGN. The hardness ratio gives some clues about the spectral state, which is consistent with a scenario where the source is accreting at a high rate with Γ ≃ 2.3. The softness of this source, combined with the discrepancy between the [Ne v] luminosity and the observed X-ray luminosity, supports the scenario where NGC 4178 hosts an AGN embedded in a heavy absorber, accreting at a high rate.
• 2.  
  The bolometric luminosity of the AGN in NGC 4178 predicted by our mid-IR results is 9.2 × 10⁴² erg s⁻¹, significantly higher than that found in any other extremely late-type, bulgeless disk galaxy.
• 3.  
  The updated bolometric luminosity, combined with other lines of evidence such as the fundamental plane and the correlation between the mass of NSCs and their resident SMBHs, has led us to conclude that the AGN in NGC 4178 is powered by a black hole of ∼10⁴–10⁵ M_☉.
• 4.  
  Two weak off-nuclear sources found in NGC 4178 have X-ray luminosities consistent with very bright X-ray backgrounds or ULXs, although we cannot rule out the possibility that they are background objects. A third off-nuclear source is very bright and was directly fit with a PL model, showing that it is a ULX located in NGC 4178 about 8 kpc from the nuclear source. It was also directly fit with a BMC model, suggesting that it may be an IMBH of ∼6 × 10³ M_☉.

It is a pleasure to thank Tim Jordan for his invaluable help carefully compiling the fluxes used to construct Figure 6. We thank the referee for their very careful review and insightful comments that significantly improved the paper. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. N. S. and S. S. gratefully acknowledge support by the Chandra Guest Investigator Program under NASA Grant G01-12126X. Work by C. C. C. at NRL is supported in part by NASA DPR S-15633-Y.