1: % Version 8; preprint format; revised after referee's report; last minor 

2: % corrections from Jenny

3: %\documentclass[12pt,preprint]{aastex}

4: \documentclass[preprint]{aastex}

5: \usepackage{emulateapj5}

6: \usepackage{apjfonts}

7: \usepackage{natbib}

8: 

9: %\documentstyle[12pt,aasms4]{article}

10: %\documentstyle[11pt,aaspp4]{article}

11: %\documentstyle[11pt,/home/lho/tex/emulateapj]{article}

12: %\documentstyle[/home/java/lho/tex/aastex/aaspptwo]{article}

13: \input psfig.tex

14: 

15: %Luis's definitions

16: \def\aa{{A\&A}}

17: \def\aas{{ A\&AS}}

18: \def\aj{{AJ}}

19: \def\al{$\alpha$}

20: \def\bet{$\beta$}

21: \def\amin{$^\prime$}

22: \def\annrev{{ARA\&A}}

23: \def\apj{{ApJ}}

24: \def\apjs{{ApJS}}

25: \def\asec{$^{\prime\prime}$}

26: \def\baas{{BAAS}}

27: \def\cc{cm$^{-3}$}

28: \def\deg{$^{\circ}$}

29: \def\ddeg{{\rlap.}$^{\circ}$}

30: \def\dsec{{\rlap.}$^{\prime\prime}$}

31: \def\cc{cm$^{-3}$}

32: \def\e#1{$\times$10$^{#1}$}

33: \def\etal{{et al. }}

34: \def\flamb{ergs s$^{-1}$ cm$^{-2}$ \AA$^{-1}$}

35: \def\flux{ergs s$^{-1}$ cm$^{-2}$}

36: \def\fnu{ergs s$^{-1}$ cm$^{-2}$ Hz$^{-1}$}

37: \def\hal{H$\alpha$}

38: \def\hst{{\it HST}}

39: \def\kms{km s$^{-1}$}

40: \def\lamb{$\lambda$}

41: \def\lax{{$\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}$}}

42: \def\gax{{$\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}$}}

43: \def\simlt{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}

44: \def\simgt{\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}

45: \def\lum{ergs s$^{-1}$}

46: \def\mbh{{$M_{\rm BH}$}}

47: \def\micron{{$\mu$m}}

48: \def\mnras{{MNRAS}}

49: \def\nat{{Nature}}

50: \def\pasp{{PASP}}

51: \def\perang{\AA$^{-1}$}

52: \def\percm2{cm$^{-2}$}

53: \def\peryr{yr$^{-1}$}

54: \def\pp{\parshape 2 0truein 6.1truein .3truein 5.5truein}

55: \def\reference{\noindent\pp}

56: \def\refindent{\par\noindent\parskip=2pt\hangindent=3pc\hangafter=1 }

57: \def\solum{$L_\odot$}

58: \def\solmass{$M_\odot$}

59: %\def\ion#1#2{\setcounter{ctr}{#2}#1$\;${\small\rm\Roman{ctr}}\relax}

60: %\def\oii{[O~{\src II}]}

61: %\def\neii{[Ne~{\src II}]}

62: %\def\neiii{[Ne~{\src III}]}

63: \def\heii{\ion{He}{2}}

64: \def\hi{\ion{H}{1}}

65: \def\hii{\ion{H}{2}}

66: \def\oii{[\ion{O}{2}]}

67: \def\oiii{[\ion{O}{3}]}

68: \def\ni{[\ion{N}{1}]}

69: \def\oi{[\ion{O}{1}]}

70: \def\nii{[\ion{N}{2}]}

71: \def\neii{[\ion{Ne}{2}]}

72: \def\neiii{[\ion{Ne}{3}]}

73: \def\hei{\ion{He}{1}}

74: \def\sii{[\ion{S}{2}]}

75: \def\siii{[\ion{S}{3}]}

76: 

77: \def\lhal{$L_{{\rm H}\alpha}$}

78: \def\lbol{$L_{{\rm bol}}$}

79: \def\ledd{$L_{{\rm Edd}}$}

80: 

81: \def\vmax{$V_{{\rm max}}$}

82: \def\sig{$\sigma_*$}

83: %\def\mlbulge{$M_\bullet-L_{\rm bul}$}

84: %\def\msigma{$M_\bullet-\sigma_*$}

85: \def\mlbulge{$M_{\rm BH}-L_{\rm bul}$}

86: \def\msigma{$M_{\rm BH}-\sigma_*$}

87: \def\mhi{$M_{{\rm H~I}}$}

88: \def\lb{$L_B$}

89: \def\vc{${\upsilon_c}$}

90: \def\vm{${\upsilon_m}$}

91: \def\vrot{$\upsilon_{{\rm rot}}$}

92: \def\vmsig{$\upsilon_m-\sigma_*$}

93: \def\mvm{$M_{\rm BH}-\upsilon_m$}

94: 

95: \slugcomment{To appear in {\it The Astrophysical Journal}.}

96: %\lefthead{Ho et al.}

97: %\righthead{\hi\ in Active Galaxies}

98: \shorttitle{\hi\ in ACTIVE GALAXIES}

99: \shortauthors{HO, DARLING, \& GREENE}

100: 

101: \begin{document}

102: 

103: \title{Properties of Active Galaxies Deduced from \hi\ Observations}

104: 

105: \author{Luis C. Ho\altaffilmark{1}, Jeremy Darling\altaffilmark{2}, and

106: Jenny E. Greene\altaffilmark{3,4}}

107: 

108: \altaffiltext{1}{The Observatories of the Carnegie Institution of Washington,

109: 813 Santa Barbara St., Pasadena, CA 91101.}

110: \altaffiltext{2}{Center for Astrophysics and Space Astronomy, Department of

111:         Astrophysical and Planetary Sciences, University of Colorado, 389 UCB, 

112:         Boulder, CO 80309-0389.}

113: \altaffiltext{3}{Princeton Observatory, Peyton Hall, Princeton University, 

114: Princeton, NJ 08544-1001.}

115: \altaffiltext{4}{Hubble Fellow and Carnegie-Princeton Fellow.}

116: 

117: \begin{abstract}

118: We have completed a new survey for \hi\ emission for a large, well-defined 

119: sample of 154 nearby ($z$ \lax\ 0.1) galaxies with type~1 (broad-line) active 

120: galactic nuclei (AGNs).  We make use of the extensive database of \hi\ and 

121: optical parameters, presented in a companion paper, to perform a comprehensive 

122: appraisal of the cold gas content in active galaxies and to seek new 

123: strategies to investigate the global properties of the host galaxies and their 

124: relationship to their central black holes.  After excluding objects with 

125: kinematically anomalous line profiles, which occur with high frequency in the 

126: sample, we show that the black hole mass obeys a strong, roughly linear 

127: relation with the host galaxy's dynamical mass, calculated by combining the 

128: \hi\ line width and the optical size of the galaxy.  Black hole mass follows 

129: a looser, though still highly significant, correlation with the maximum 

130: rotation velocity of the galaxy, as expected from the known scaling between 

131: rotation velocity and central velocity dispersion.  Neither of these \hi-based 

132: correlations is as tight as the more familiar relations between black hole mass 

133: and bulge luminosity or velocity dispersion, but they offer the advantage of 

134: being insensitive to the glare of the nucleus and therefore are promising new 

135: tools for probing the host galaxies of both nearby and distant AGNs.  We 

136: present evidence for substantial ongoing black hole growth in the most actively 

137: accreting AGNs.  In these nearby systems, black hole growth appears to be 

138: delayed with respect to the assembly of the host galaxy but otherwise has 

139: left no detectable perturbation to its mass-to-light ratio, as judged from the 

140: Tully-Fisher relation, or its global gas content.  The host galaxies of type~1 

141: AGNs, including those luminous enough to qualify as quasars, are generally 

142: gas-rich systems, possessing a cold interstellar medium reservoir at least as 

143: abundant as that in inactive galaxies of the same morphological type.  This 

144: calls into question current implementations of AGN feedback in models of 

145: galaxy formation that predict strong cold gas depletion in unobscured AGNs.

146: %From the radial velocity offset between the \oiii\ \lamb 5007 line and \hi, we 

147: %show that the narrow-line region contains an outflowing component whose 

148: %strength varies systematically with accretion rate.

149: \end{abstract}

150: \keywords{galaxies: active --- galaxies: bulges --- galaxies: ISM ---

151: galaxies: kinematics and dynamics --- galaxies: nuclei --- galaxies: Seyfert}

152: 

153: \section{Introduction}

154: 

155: Dynamical studies of the central regions of nearby inactive galaxies have 

156: revealed that supermassive black holes (BHs; \mbh\ $\approx 10^6-10^9$ 

157: \solmass) are ubiquitous, and that their masses are strongly coupled to the 

158: host galaxy's bulge luminosity (Kormendy \& Richstone 1995; Magorrian et al. 

159: 1998) and stellar velocity dispersion (\msigma\ relation: Gebhardt et al. 

160: 2000a; Ferrarese \& Merritt 2000; Tremaine et al. 2002). The occupation 

161: fraction of central BHs in bulgeless or very late-type galaxies is poorly 

162: constrained, but at least some such systems have been found to contain BHs 

163: with masses as low as $\sim 10^5$ \solmass\ (Filippenko \& Ho 2003; Barth et 

164: al. 2004; Greene \& Ho 2004, 2007b, 2007c), which continue to obey the 

165: \msigma\ relation (Barth et al. 2005; Greene \& Ho 2006b).  These empirical 

166: correlations, which strongly suggest that BH growth is closely coupled to 

167: galaxy formation and evolution, have inspired considerable observational and 

168: theoretical attention in the last few years (see, e.g., reviews in Ho 2004a).  

169: BH growth, its observational manifestation as nuclear activity, and the 

170: consequences of feedback from active galactic nuclei (AGNs) are now widely 

171: viewed as unavoidable pieces of the overall puzzle of cosmological structural 

172: formation (e.g., Granato et al.  2004; Springel et al. 2005; Hopkins et al. 

173: 2006).

174: 

175: The BH-bulge correlations beg several pressing, unanswered questions.  Do BHs

176: grow predominantly by radiatively efficient accretion  during the luminous (Yu

177: \& Tremaine 2002) or obscured (Fabian 1999) quasar phase, by low accretion

178: rates in moderately luminous AGNs at recent times (e.g., Cowie et al. 2003),

179: or by mergers (e.g., Yoo \& Miralda-Escud\'e 2004)?  Which came first, BH or

180: galaxy?  Must the growth of the BH and its host be finely synchronized so as

181: to preserve the small intrinsic scatter observed in the local BH-host scaling

182: relations?

183: 

184: These issues can be observationally tackled in two steps: first by extending

185: the BH-host scaling relations to {\it active}\ galaxies, wherein the BHs are

186: still growing, and second by probing the evolution of the scaling relations by

187: stepping back in redshift.  To this end, two ingredients are needed---BH

188: masses and host galaxy parameters.  Fortunately, \mbh\ can be estimated

189: readily in broad-line (type 1) AGNs, solely from basic spectroscopic data 

190: using the mass-luminosity-line width relation (also known as the ``virial 

191: method''; Kaspi et al. 2000; Greene \& Ho 2005b; see Peterson 2007, and 

192: references therein).  Based on such mass estimates, local AGNs do seem to obey 

193: roughly the same BH-host scaling relations as in inactive galaxies (Gebhardt 

194: et al. 2000b; Ferrarese et al. 2001; McLure \& Dunlop 2001; Nelson et al. 

195: 2004; Onken et al. 2004; Greene \& Ho 2006b; Shen et al. 2008).  It is crucial 

196: to recognize, however, that at the moment this claim is based on {\it very}\ 

197: limited, and possibly biased, data.  The bright AGN core renders measurement 

198: of bulge luminosity and, in particular, central stellar velocity dispersion 

199: extremely challenging (see detailed discussion in Greene \& Ho 2006a).  

200: Notwithstanding these worries, many people now routinely assume that the 

201: BH-host scaling relations can be directly applied to AGNs, of arbitrarily high 

202: luminosity and redshift.

203: 

204: Here we propose a new angle to investigate the relationship between BH masses

205: and the host galaxies of AGNs.  In normal, inactive galaxies, it is well known

206: that the stellar velocity dispersion of the bulge tracks the maximum rotation

207: velocity of the disk, roughly as \vm\ = $(\sqrt{2}-\sqrt{3}) \times \sigma_*$ 

208: (Whitmore et al. 1979).  Although the theoretical 

209: underpinnings of this correlation are still murky, and the \vmsig\ relation 

210: is not as tight as has been claimed (Ferrarese 2002; Baes et al. 2003; 

211: Pizzella et al. 2005), nevertheless an empirical relation between \vm\ and 

212: \sig\ {\it does}\ exist (Courteau et al. 2007; Ho 2007a), which implies that 

213: \sig\ can be estimated from \vm.  Now, \vm\ can be measured straightforwardly 

214: through \hi\ observations for galaxies that are sufficiently gas-rich.  In the 

215: absence of a resolved rotation curve, \vm\ can be estimated from a single-dish 

216: measurement of the \hi\ profile, provided that we have some constraint on the 

217: inclination angle of the disk.  With the rotation velocity in hand, we can 

218: deduce immediately two physical quantities of interest: the total galaxy 

219: luminosity through the Tully-Fisher relation (Tully \& Fisher 1977), and, 

220: given an estimate of the size of the disk (e.g., from optical imaging), the 

221: dynamical mass of the system.  The advantages of this approach are clear.  

222: Since the \hi\ is distributed mostly at large radii (e.g., Broeils \& Rhee 

223: 1997), it should be largely ``blind'' to the AGN core.  This allows us to 

224: circumvent the problems encountered in the optical, where attempts to 

225: disentangle the bulge from the active nucleus are maximally impacted.

226: Applying this principle, Ho (2007b) has recently used the integrated profile 

227: of the rotational CO line to infer certain properties of high-redshift quasar 

228: host galaxies.  

229:  

230: Apart from dynamical constraints, the \hi\ observations yield another piece

231: of information of significant interest---a first-order measurement of the cold

232: gas content.  As the raw material responsible for fueling not only the growth

233: of the BH but ultimately also the galaxy, one can hardly think of a more

234: fundamental quantity to ascertain.  How does the gas content of AGN host

235: galaxies compare with that in inactive galaxies?  Depending on one's view on

236: the evolutionary path of AGNs and the efficacy of gas removal by AGN feedback,

237: active galaxies may be more or less gas-rich than inactive galaxies.  Does

238: the gas content scale with the degree of nuclear activity?  Does it vary with

239: the nebular properties of the central source?  One of the major goals of this

240: paper is to try to address some of these basic questions, which to date

241: have been largely unanswered.

242: 

243: 

244: \section{The Database}

245: 

246: The analysis in this paper is based on the database presented in the companion

247: paper by Ho et al. (2008), to which the reader is referred for full details.

248: In brief, a comprehensive sample of 154 nearby type~1 (broad-lined) AGNs was 

249: assembled, consisting of new Arecibo\footnote{The Arecibo Observatory

250: is part of the National Astronomy and Ionosphere Center, which is operated by

251: Cornell University under a cooperative agreement with the National Science

252: Foundation.} observations of 101 sources with $z$ \lax\ 

253: 0.11, mostly selected from the Fourth Data Release of the Sloan Digital Sky 

254: Survey (SDSS; Adelman-McCarthy et al. 2006), and a supplementary sample of 53 

255: other nearby sources collected from the literature.  In addition to basic \hi\ 

256: properties (line fluxes, line widths, and radial velocities), we also 

257: assembled optical data for the AGN (emission-line strengths, line widths, line 

258: centroid, nuclear luminosity) and the host galaxy (image, concentration index, 

259: size, axial ratio, total magnitude, central stellar velocity dispersion).  

260: From this material a number of important physical parameters were derived, 

261: including BH mass, Eddington ratio, morphological type, inclination angle, 

262: deprojected rotation velocity, \hi\ mass, dynamical mass, estimated host galaxy 

263: luminosity, and certain rudimentary properties of the narrow-line region.

264: Distance-dependent quantities were calculated assuming $H_0$ = 70 

265: \kms~Mpc$^{-1}$, $\Omega_{m} = 0.3$, and $\Omega_{\Lambda} = 0.7$. 

266: 

267: \section{Analysis}

268: 

269: \subsection{The \vmsig\ Relation in Active Galaxies}

270: 

271: To set the stage of using \vm\ as a surrogate dynamical variable to 

272: investigate BH-host scaling relations, we first examine the correlation 

273: between \vm\ and \sig\ for the active galaxies in our sample.  Including all 

274: objects that have measurements of both quantities, the scatter is 

275: discouragingly large (Fig.~1{\it a}).  Closer inspection, however, reveals 

276: that many of the extreme outliers correspond to objects that have potentially

277: untrustworthy deprojected rotation velocities because of uncertain inclination 

278: corrections, as well as a handful of sources whose central stellar velocity 

279: dispersions were estimated indirectly from the width of the \oii\ \lamb3727 

280: emission line (following Greene \& Ho 2005a).  Removing these objects improves 

281: the correlation.  As further discussed in \S 3.5, a sizable fraction of the 

282: objects in our sample contain non-classical \hi\ velocity profiles, which are 

283: either single-peaked, highly asymmetric, or both, indicative of strongly 

284: perturbed or dynamically unrelaxed \hi\ distributions.  If we further remove 

285: these cases---a cut that, unfortunately, drastically reduces the sample to 

286: only 40 objects---a much cleaner correlation between \vm\ and \sig\ emerges 

287: (Fig.~1{\it b}).  Because of the small number of objects and their limited 

288: dynamic range, the present sample is not well suited to define the \vm--\sig\ 

289: relation.  Nevertheless, the present sample still reveals a statistically 

290: significant correlation between \vm\ and \sig; the Kendall's $\tau$ 

291: correlation coefficient is $r = 0.53$, which rejects the null hypothesis 

292: of no correlation with a probability of $P$ = 98.2\%.  We overplot on the 

293: figure the \vm--\sig\ relation obtained from the 550 ``kinematically normal,'' 

294: inactive spiral galaxies from the study of Ho (2007a).  The ordinary 

295: least-square bisector fit for the inactive objects, 

296: 

297: \begin{equation}

298: \log \upsilon_m = (0.80\pm0.029) \log \sigma_* + (0.62\pm0.062), 

299: \end{equation}

300: 

301: \noindent

302: provides a reasonably good match to the AGN sample.  According to the 

303: Kolmogorov-Smirnov test, the probability of rejecting the null hypothesis that 

304: the AGN and non-AGN samples are drawn from the same parent population is $P$ = 

305: 69.2\%.  We conclude that the two populations are not significantly different.

306: The scatter is still substantial, but recall that the zeropoint of the \vmsig\ 

307: relation depends on morphological type, such that early-type systems have a 

308: lower value of \vm/\sig\ than late-type spirals (\vm/\sig\ $\approx$ $1.2-1.4$ 

309: for E and S0, compared with \vm/\sig\ $\approx$ $1.6-1.8$ for spirals of type 

310: Sc and later; see Ho 2007a).  That the zeropoint for inactive spirals seems to 

311: roughly match the current sample of AGNs suggests that their host galaxies are 

312: mostly disk galaxies with modest bulge components, consistent with the actual 

313: estimated BH masses (median \mbh\ = $1.6\times10^7$ \solmass) and 

314: morphological types (median $T$ = 2.5, corresponding to Sab) of the objects.

315: 

316: Since local AGN host galaxies obey the \msigma\ relation, and we have just 

317: shown that they roughly follow the same 

318: 

319: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

320: \vskip 0.3cm

321: %%BoundingBox: 30 10 430 800

322: \figurenum{1}

323: \begin{figure*}[t]

324: \centerline{\psfig{file=f1.eps,width=19.5cm,angle=-90}}

325: \figcaption[fig1.ps]{

326: Correlation between the central stellar velocity dispersion and the maximum

327: rotation velocity for the active galaxies in our sample.  Panel ({\it a})

328: plots all available data, including those that have uncertain inclination

329: corrections ({\it open circles}) and \oii-based estimates of the central

330: velocity dispersion ({\it open triangles}); objects with

331: both reliable rotation velocities and stellar velocity dispersions are plotted

332: as filled circles.  No selection has been done with regards to \hi\ profile

333: type.  Panel ({\it b}) excludes all objects with single-peaked and/or

334: asymmetric \hi\ profiles, as well as those with uncertain rotation velocities

335: and \oii-based estimates of the central velocity dispersion.  Note the

336: significant reduction in scatter.  The dashed line represents the fit to the

337: \vmsig\ relation for the 550 ``kinematically normal'' inactive spiral

338: galaxies from the sample of Ho (2007a); see text.

339: \label{fig1}}

340: \end{figure*}

341: \vskip 0.3cm

342: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

343: 

344: \noindent

345: asymmetric \hi\ profiles (panel {\it a})\footnote{For the rest of the 

346: $\upsilon_m-\sigma_*$ relation as in 

347: inactive galaxies, we expect \mbh\ to be correlated with \vm.  Not 

348: surprisingly, this is confirmed in Figure~2, which additionally reemphasizes 

349: that a tighter relation results after removing objects with single-peaked 

350: and/or paper, we have also excluded two objects from the literature sample that 

351: have suspiciously low values of \vm\ (Mrk 359, \vm\ = 38 \kms; Mrk 493, \vm\ = 

352: 24.7 \kms).  Closer inspection of the original \hi\ data shows that Mrk 359 

353: has a highly unusual profile consisting of a narrow peak and a very broad base 

354: (Springob et al. 2005); moreover, the inclination angle given in Hyperleda, 

355: $i$ = 39.5\deg, seems inconsistent with the nearly face-on appearance of its 

356: Digital Sky Survey image (see Appendix in Ho 2007a for a discussion on the 

357: uncertainties associated with inclination angles listed in Hyperleda), 

358: suggesting that we have seriously underestimated the inclination correction.  

359: Mrk 493 may suffer from the same problem; the optical (SDSS) image of the 

360: object appears to be much more face-on than Hyperleda's value of $i$ = 

361: 45\deg.}.  From the Kendall's $\tau$ test, $r = 0.58$ and $P$ = 99.9\%.

362: Overplotted on the figure is the relation expected from combining equation (1) 

363: with the \msigma\ relation for local AGNs obtained by Greene \& Ho (2006b),

364: $\log (M_{\rm BH}/M_\odot) = 4.02 \log (\sigma_*/200\,{\rm km~s^{-1}}) + 7.96$:

365: 

366: \begin{equation}

367: \log (M_{\rm BH}/M_\odot) = 5.1 \log \upsilon_m - 4.4.

368: \end{equation}

369: 

370: \noindent

371: As with Figure~1, we ran a Kolmogorov-Smirnov test to see whether the present 

372: sample of AGNs is drawn from the same population of objects used to 

373: define the predicted correlation between $M_{\rm BH}$ and \vm; the probability 

374: of rejecting the null hypothesis of no correlation is $P$ = 68.6\%.  Again, 

375: the two populations are not significantly different.

376: 

377: Nevertheless, the \mvm\ relation still 

378: contains significant scatter, even after omitting the kinematically peculiar 

379: objects.  What is responsible for this?  We believe that most of the scatter 

380: is intrinsic.  The virial BH mass estimates for nearby AGNs, based on broad 

381: H\al\ or H\bet, currently have an uncertainty of $\sim 0.3-0.5$ dex (Greene \& 

382: Ho 2006b; Peterson 2007), and errors associated with the inclination 

383: correction applied to the \hi\ line widths do not seem to be the dominant 

384: source of scatter (Fig.~2{\it b}).  Part of the scatter can be attributed 

385: to an effect related to Hubble type.  As mentioned above and 

386: discussed at length in Ho (2007a), \vm/\sig\ varies systematically with galaxy 

387: concentration or morphological type, and because our sample contains a wide 

388: range of morphological types (see Ho et al. 2008), we expect the zeropoint of 

389: the \mvm\ relation to be smeared by the actual mixture of galaxy types.  

390: This is illustrated in Figure~2{\it c}, wherein the sample is divided into 

391: four broad bins in Hubble type.  The small subsamples make it difficult to see 

392: the trend in detail, but comparison of the E+S0 galaxies with the later-type 

393: spirals clearly shows that the two subgroups are offset from each other, in 

394: the sense expected from the variation of \vm/\sig\ with morphological type.   

395: 

396: Taken at face value, another contribution to the scatter seems to be 

397: connected to variations in AGN properties, as shown in Figures~2{\it d} and 

398: 2{\it e}, where we examine possible dependences on broad-line region line 

399: width and Eddington ratio.  Objects with broad H\al\ or H\bet\ full-width at 

400: half maximum (FWHM) $\leq 2000$ \kms, commonly designated ``narrow-line'' 

401: Seyfert 1 (NLS1) galaxies, have a slight tendency to be offset toward lower 

402: \mbh\ for a given \vm\ (Fig.~2{\it d}).  Since many NLS1 galaxies tend to have 

403: elevated accretion rates (e.g., Collin \& Kawaguchi 2004), the weak trend 

404: with FWHM may be more clearly discerned by dividing the sample according to 

405: Eddington ratio.  Indeed this expectation seems to hold, as shown in 

406: Figure~2{\it e}, where, within the limited statistics, there appears to be a 

407: monotonic decrease of \mbh/\vm\ with increasing \lbol/\ledd.  The trend with 

408: Eddington 

409: 

410: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

411: \vskip 0.3cm

412: %%BoundingBox: 10 50 490 780

413: \figurenum{2}

414: \begin{figure*}[t]

415: \centerline{\psfig{file=f2.eps,width=19.5cm,angle=-90}}

416: \figcaption[fig2.ps]{

417: Correlation between BH mass and the maximum rotation velocity deduced from

418: \hi\ observations.  The different panels test for possible dependences on

419: ({\it a}) \hi\ profile type, ({\it b}) inclination angle correction, ({\it c})

420: Hubble type, ({\it d}) FWHM of broad H\al\ or H\bet, and ({\it e}) Eddington

421: ratio.  In panel ({\it a}), objects with uncertain inclination corrections

422: were removed; in panels ({\it b})--({\it e}), objects with single and/or

423: asymmetric \hi\ profiles were additionally removed.  Panel ({\it f}) is the

424: same as panel ({\it e}), but the different Hubble types have been adjusted to

425: a common zeropoint by shifting \vm.  The solid line represents the predicted

426: correlation for the whole sample, equation (2), given the \msigma\ relation of

427: Greene \& Ho (2006b) and the \vmsig\ relation (eq. 1); the dotted and

428: dashed lines show the zeropoint offset (in \mbh) for the objects with

429: $L_{\rm bol}/L_{\rm Edd} \leq 0.1$ and $L_{\rm bol}/L_{\rm Edd} > 0.1$,

430: respectively.

431: \label{fig2}}

432: \end{figure*}

433: \vskip 0.3cm

434: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

435: 

436: \noindent 

437: ratio is stronger than that with luminosity alone (not shown). Could 

438: the apparent variation of \mbh/\vm\ with FWHM and \lbol/\ledd\ be a secondary 

439: effect related to the dependence of \mbh/\vm\ on morphological type?  The most 

440: active AGNs in the nearby Universe live in moderate-mass, relatively late-type 

441: disk galaxies (e.g., Heckman et al. 2004; Greene \& Ho 2007b); these are 

442: precisely the hosts of high-\lbol/\ledd\ systems, such as NLS1s.   We attempt 

443: to remove the Hubble type dependence by shifting \vm\ for the different Hubble 

444: type subgroups to a common reference zeropoint (defined by the entire sample).

445: As expected, the scatter goes down, but, interestingly, the trend with 

446: \lbol/\ledd\ persists (Fig.~2{\it f}).

447: 

448: \subsection{Correlation between \mbh\ and Galaxy Dynamical Mass}

449: 

450: Although our \hi\ observations yield no information on the spatial distribution

451: of the gas, we can still obtain a rough estimate of the characteristic 

452: dynamical mass of the galaxy by combining the deprojected rotation velocity 

453: from the \hi\ line width with the optical diameter.  This approach is justified

454: for the following reason.  The size of the \hi\ disk in spiral galaxies over 

455: a wide range of Hubble types and luminosities tightly scales with the size 

456: of the optical disk; within 30\%--40\%, $D_{\rm H~{\tiny I}}/D_{\rm 25} 

457: \approx 1.7$ (Broeils \& Rhee 1997; Noordermeer et al. 2005), where 

458: $D_{\rm 25}$ is the optical isophotal diameter at a surface brightness level 

459: of $\mu_B = 25$ mag~arcsec$^{-2}$.  Thus, even if the absolute dynamical mass 

460: may be uncertain because of our ignorance of the size of the \hi-emitting 

461: disk, the {\it relative}\ masses should be reasonably accurate.  Following 

462: Casertano \& Shostak (1980), Ho et al. (2008) calculated dynamical masses for 

463: our AGN sample, which, interestingly, show a fairly strong correlation with 

464: BH mass (Fig.~3).  In particular, BH mass correlates more strongly with 

465: $M_{\rm dyn}$, which scales as $D_{\rm 25} \upsilon_m^2$, than with 

466: $\upsilon_m$ alone.  An ordinary least-squares bisector fit to the entire 

467: sample, after making the quality cut as above, gives

468: 

469: \begin{equation}

470: \log (M_{\rm BH}/M_\odot) = (1.29\pm0.11) \log (M_{\rm dyn}/M_\odot) - 

471: (6.96\pm1.17),

472: \end{equation}

473: 

474: \noindent

475: with an rms scatter of 0.61 dex.  The slope is formally steeper than unity, 

476: but this result should be regarded as provisional considering that most of the 

477: objects span a relatively narrow range of $M_{\rm dyn}$.  Closer inspection 

478: reveals that the zeropoint of the $M_{\rm BH}-M_{\rm dyn}$ correlation also 

479: depends on Hubble type (Fig.~3{\it a}).  At a given $M_{\rm dyn}$, early-type 

480: galaxies have a higher \mbh\ than late-type galaxies.  Relative to the 

481: best-fitting line for the whole sample, at a fixed log~\mbh\ the shift in 

482: $\log M_{\rm dyn}$ is $+0.19$, $-0.05$, $-0.10$, and $-0.59$ for galaxies of 

483: type E/S0, Sa, Sb, and Sc/Sd, 

484: 

485: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

486: \vskip 0.3cm

487: \figurenum{3}

488: \begin{figure*}[t]

489: %%BoundingBox: 30 70 430 730

490: \centerline{\psfig{file=f3.eps,width=19.5cm,angle=-90}}

491: \figcaption[fig3.ps]{

492: Correlation between BH mass and dynamical mass of the host galaxy.  Objects

493: with uncertain inclination corrections and that have single and/or asymmetric

494: \hi\ profiles are excluded.  Panel ({\it a}) divides the sample into Hubble

495: types, and panel ({\it b}) bins the sample by Eddington ratio, after adjusting

496: the different Hubble types to a common zeropoint by shifting $M_{\rm dyn}$.

497: The solid line shows the ordinary least-squares bisector fit given in equation

498: (3) for the entire sample; the dotted and dashed lines show the zeropoint

499: offset (in \mbh) for the objects with $L_{\rm bol}/L_{\rm Edd} \leq 0.1$ and

500: $L_{\rm bol}/L_{\rm Edd} > 0.1$, respectively.

501: \label{fig3}}

502: \end{figure*}

503: \vskip 0.3cm

504: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

505: 

506: \noindent

507: respectively.  If we adjust the zeropoint of 

508: these groups by their corresponding offsets (in $M_{\rm dyn}$), the rms scatter 

509: of the resulting correlation decreases to 0.55 dex (Fig.~3{\it b}).  As in the 

510: case of the \mvm\ diagram, we find that at a given $M_{\rm dyn}$ objects with 

511: higher \lbol/\ledd\ tend to have smaller \mbh.

512: 

513: \subsection{The Tully-Fisher Relation for Active Galaxies}

514: 

515: The empirical correlation between \vm\ and total galaxy luminosity, first

516: introduced by Tully \& Fisher (1977), provides yet another avenue to assess

517: potential differences between active and inactive galaxies.  As described in 

518: Ho et al. (2008), approximate optical host galaxy luminosities can be obtained 

519: by comparing the integrated photometry with the nuclear spectroscopy.  Figure~4

520: plots the host galaxy absolute magnitude (after removing the AGN contribution) 

521: converted to the $B$ band, versus the maximum rotation velocity; overlaid for 

522: comparison is the corresponding Tully-Fisher relation for inactive (spiral) 

523: galaxies in the Ursa Major cluster (Verheijen 2001).  Considering first the 

524: entire sample (Fig.~4{\it a}), $M_{B,{\rm host}}$ loosely traces \vm, but 

525: the scatter is again considerable and the apparent correlation is formally

526: statistically insignificant ($r = -0.23$, $P$ = 93.3\%).  Pruning, as before, 

527: the objects with unreliable inclination corrections and kinematically peculiar 

528: profiles filters out most of the egregious outliers, such that the remaining 

529: sample falls mostly within the locus of inactive spirals (Figs.~4{\it b} and 

530: 4{\it c}).  The formal correlation between $M_{B,{\rm host}}$ and \vm\ is 

531: still low ($r = -0.30$) and only marginally significant ($P$ = 95.0\%), 

532: probably because of the small sample size and limited dynamic range 

533: in \vm\ (90\% of the sample has \vm\ = 200$\pm$100 \kms).

534: Apart from the larger scatter exhibited by the active sample, an 

535: effect that can be attributed, at least partly, to its broad mixture of Hubble 

536: types, the approximate nature in which the host luminosities were estimated, 

537: and the sensitivity of the blue bandpass to extinction and stellar population 

538: variations (e.g., De~Rijcke et al. 2007), there are no other gross differences 

539: between the Tully-Fisher relation of active and inactive galaxies.  Different 

540: Hubble types define slightly offset, parallel sequences (Fig.~4{\it b}), an 

541: effect that has been well-documented in the $B$ band.  For a given $B$-band 

542: luminosity, an early-type spiral has a larger characteristic rotation 

543: amplitude than a late-type spiral (e.g., Rubin et al. 1985; De~Rijcke et al. 

544: 2007).  To separate out this effect, we applied small corrections to the 

545: zeropoints (in host galaxy luminosity) of each Hubble type bin.  The adjusted 

546: distribution shows no obvious segregation by AGN properties, such as Eddington 

547: ratio (Fig.~4{\it c}).  

548: 

549: \subsection{\hi\ Content}

550: 

551: We next examine the \hi\ content of our sample, with special emphasis on 

552: whether AGN hosts differ in any noticeable way from inactive galaxies.  In 

553: absolute terms, our survey objects are quite gas-rich in neutral atomic 

554: hydrogen.  The \hi\ masses for the detected sources range from \mhi\

555: $\approx\,10^9$ to $4\times10^{10}$ \solmass\ (Ho et al. 2008).  Taking into 

556: account the upper limits for the nondetections, the Kaplan-Meier product-limit 

557: estimator (Feigelson 

558: 

559: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

560: \vskip 0.3cm

561: %%BoundingBox: 120 30 490 760

562: \figurenum{4}

563: \begin{figure*}[t]

564: \centerline{\psfig{file=f4.eps,width=19.5cm,angle=-90}}

565: \figcaption[fig4.ps]{

566: The $B$-band Tully-Fisher relation for the active galaxies in our sample.

567: In panel ({\it a}), solid symbols denote reliable deprojected rotation

568: velocities, while less reliable values are marked with open symbols; crosses

569: indicate lower limits on \vm\ that lack estimates of the inclination angle.

570: In panel ({\it b}), the sample is divided according to Hubble type (open 

571: circles = E/S0; triangles = Sa; squares = Sb; filled circles = Sc/Sd), and

572: panel ({\it c}) bins the sample by Eddington ratio (solid, open, and star 

573: symbols denote objects with $L_{\rm bol}/L_{\rm Edd} \leq 0.1$, $0.1-0.3$, 

574: and $> 0.3$, respectively), after adjusting the different Hubble types to a 

575: common zeropoint by shifting $M_{B,{\rm host}}$.  Objects with uncertain 

576: deprojected rotation velocities and single-peaked and/or asymmetric \hi\ 

577: profiles were removed from panels ({\it b}) and ({\it c}).  The {\it solid}\ 

578: line represents the $B$-band Tully-Fisher relation for inactive spiral 

579: galaxies, as derived by Verheijen (2001) for galaxies in the Ursa Major 

580: cluster; the {\it dashed}\ lines mark the region that has twice the rms 

581: scatter of the Ursa Major sample.

582: \label{fig4}}

583: \end{figure*}

584: \vskip 0.3cm

585: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

586: 

587: \noindent

588: \& Nelson 1985) yields a mean of \mhi\ = 

589: $(7.0\pm0.66)\times10^{9}$ \solmass\ and a median of \mhi\ = 

590: $5.0\times10^{9}$ \solmass.   This is almost identical to the total 

591: \hi\ mass of the Milky Way ($5.5 \times 10^9$ \solmass; Hartmann \& Burton 

592: 1997), and comparable to the average value for Sb spirals (Roberts \& Haynes 

593: 1994).  The above statistics pertain just to the newly surveyed SDSS sources. 

594: The literature sample is both heterogeneous and possibly biased because the 

595: database from which the \hi\ data were compiled, 

596: Hyperleda\footnote{{\tt http://leda.univ-lyon1.fr/}}, does not report 

597: upper limits.  Nonetheless, we confirm that combining both samples does not

598: significantly alter these statistics.

599: 

600: The \hi\ content of galaxies varies greatly and systematically across the 

601: Hubble sequence (Haynes \& Giovanelli 1984; Roberts \& Haynes 1994).  Thus, 

602: in order to conduct a meaningful comparison of the \hi\ budget of active 

603: and inactive galaxies, we must know their morphological types.  Among the 

604: 154 sources in our sample that have \hi\ detections or meaningful upper 

605: limits, 148 have estimates of both their morphological type and host 

606: galaxy optical luminosity.  Figure~5 shows the distribution of \hi\ masses

607: normalized to the $B$-band luminosity of the host galaxy, subdivided into six 

608: bins of Hubble types.  For reference, we computed the distribution of 

609: $M_{{\rm H~I}}/L_{B,{\rm host}}$ for 13,262 inactive galaxies culled from 

610: Hyperleda. These represent all galaxies in the database, reported to be 

611: current up to the end of 2003, that have reliable entries of \hi\ flux, 

612: morphological type, and $B$-band magnitude; known AGNs were excluded, as 

613: discussed in Ho et al. (2008; Appendix).  As previously mentioned, Hyperleda 

614: does not record upper limits for \hi\ nondetections, so these distributions 

615: should be viewed strictly as upper bounds.  Since the majority of the galaxies 

616: in Hyperleda are relatively bright and nearby, the detection rate of \hi\ 

617: among the spirals should be very high ($\sim$90\%; Haynes \& Giovanelli 

618: 1984), such that the observed distribution can be regarded as being quite 

619: close to the true distribution.  The situation for early-type galaxies, 

620: however, is quite different, because the detection rate is only $\sim$15\% for 

621: ellipticals (Knapp et al. 1985) and $\sim$30\% for S0s (Wardle \& Knapp 1986), 

622: and so the distributions shown in the figure are highly biased.

623: 

624: Comparison of the active and inactive distributions reveals two interesting

625: points.  Among mid- to late-type spirals (Sb and later), the frequency 

626: distribution of $M_{{\rm H~I}}/L_{B,{\rm host}}$ is roughly the same for the 

627: two populations.  However, for the bulk of the sample, which comprise 

628: bulge-dominated Sa spirals and S0s, active galaxies appear to be {\it more}\ 

629: gas-rich than their inactive counterparts.  The most dramatic manifestation of 

630: this effect shows up among the ellipticals, although here we 

631: are handicapped somewhat by the small number of sources (9) in the active 

632: sample.  Note that among the E and S0 subgroups, the difference between 

633: the active and inactive samples is far greater than portrayed (although 

634: we have no rigorous statistical way to quantify this) because, as mentioned 

635: above, the inactive distribution is highly biased by the omission of upper 

636: limits.  (As noted, the literature data for the active sample are likely 

637: biased as well, but we verified that our conclusions are essentially 

638: unaffected if we exclude these data from the AGN sample.) The host galaxies of 

639: nearby type~1 AGNs across all Hubble types are at least as gas-rich as 

640: inactive galaxies, and among early-type systems their 

641: 

642: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

643: \vskip 0.3cm

644: \figurenum{5}

645: %BoundingBox:  50 30 580 750

646: \psfig{file=f5.eps,width=8.5cm,angle=0}

647: \figcaption[fig5.ps]{

648: The distribution of \hi\ masses normalized to the $B$-band luminosity of the

649: host galaxy, as a function of its Hubble type.  The black histograms show the

650: distributions for 13,262 inactive galaxies with \hi\ detections listed in

651: Hyperleda.  The AGN sample from our study is plotted in red histograms, with

652: upper limits indicated by dotted lines.  The entire sample of 154 objects is

653: plotted in the penultimate panel from the bottom, while the top panels show

654: the sample sorted by Hubble type.  The number of inactive galaxies in

655: each group is labeled, with the number of active objects shown in

656: parentheses.  The bottom-most panel isolates the quasars.  The red

657: histograms correspond to objects from the current \hi\ sample; the blue

658: histograms correspond to PG quasars observed in CO (Ho 2005a); the green

659: histograms correspond to $z < 0.5$ PG quasars with dust masses (Haas et al.

660: 2003); see text for details.

661: \label{fig5}}

662: \vskip 0.3cm

663: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

664: 

665: \noindent

666: gas content appears to 

667: be markedly enhanced.

668: 

669: %\vskip 1.5cm

670: \subsection{Environmental Effects} 

671: 

672: Crude information on the spatial distribution or dynamical state of the gas 

673: can be ascertained from the degree to which the observed \hi\ line profiles 

674: deviate from the classical double-horned signature of an optically thin 

675: rotating disk.  Experience with nearby galaxies indicates that tidally 

676: disturbed systems often exhibit asymmetric, single-peaked, or otherwise highly 

677: irregular line profiles (e.g., Gallagher et al. 1981; Haynes et al. 1998).   

678: Given the modest signal-to-noise ratio of our data, Ho et al. (2008) decided 

679: not to implement a rigorous scheme to quantify the line morphology, but rather 

680: adopted a qualitative classification scheme in which obviously non-classical 

681: profiles were simply labeled as ``A'' (asymmetric), ``S'' single-peaked, or 

682: ``AS'' (combination of both).  Among the 66 objects detected in our new survey, 

683: eight (12\%) are classified as ``A,'' five (8\%) are classified as ``S,'' 

684: and 16 (24\%) are classified as ``AS,'' for an overall frequency of 

685: non-classical profiles of 44\%.  Ho et al. (2008) performed a systematic 

686: census of physical neighbors within a search radius of 7\farcm5 around each 

687: object.  There seems to be no clear association between profile peculiarity 

688: and the presence of nearby neighbors.  While some objects with single-peaked 

689: and/or asymmetric profiles have plausible nearby companions, many do not;

690: at the same time, a number of objects with apparent companions exhibit 

691: seemingly regular line profiles.  The presence or absence of kinematic 

692: irregularity also appears to be uncorrelated with any of the global or AGN 

693: parameters that we have at our disposal.  We searched for possible differences 

694: in the following quantities, but found none that was statistically 

695: significant: morphology type, galaxy luminosity, total and relative \hi\ mass, 

696: AGN luminosity, broad-line region FWHM, BH mass, and Eddington ratio.  

697: 

698: 

699: \subsection{Connections with AGN Properties}

700: 

701: The availability of optical data affords us an opportunity to explore possible

702: connections between the \hi\ and AGN properties of our sample.  Concentrating

703: on the SDSS objects, for which we have homogeneous optical spectroscopic 

704: measurements, we find that their narrow emission-line ratios place the 

705: majority of them in the territory of Seyfert nuclei (Fig.~6).  This is not 

706: surprising, since most of the SDSS-selected type~1 AGNs tend to have 

707: relatively high accretion rates (Greene \& Ho 2007b), which generally 

708: correspond to high-ionization sources (Ho 2004b).  The objects detected in 

709: \hi\ do not stand out in any noticeable way from the nondetections.  The 

710: same holds for the electron densities of the narrow-line region, as can be 

711: inferred from the line ratio \sii\ \lamb6716/\sii\ \lamb6731: the \hi\ 

712: detections are statistically indistinguishable from the \hi\ nondetections 

713: (Fig.~7).  Next, we searched for a possible dependence of AGN 

714: luminosity or Eddington ratio on \hi\ content, either in absolute (\mhi) or 

715: relative ($M_{{\rm H~I}}/L_{B,{\rm host}}$) terms, but again found 

716: none (Fig.~8).   There is, at best, a mild trend of increasing H\al\ luminosity 

717: with increasing \hi\ mass (Fig.~8{\it a}), but given the mutual dependence of 

718: the two quantities on distance, we regard this result as highly suspect.  

719: 

720: \section{Discussion}

721: 

722: \subsection{Black Hole-Host Galaxy Scaling Relations}

723: 

724: The relative ease with which integrated \hi\ emission can be detected in 

725: nearby AGNs opens up the possibility of using the global \hi\ line width as 

726: a new dynamical variable to investigate the scaling between BH mass and host 

727: galaxy potential.  A quantity that can be readily measured from single-dish 

728: spectra, the \hi\ line width has been widely used in a variety of 

729: extragalactic contexts as an effective shortcut to estimate \vm, the maximum 

730: rotation velocity of galaxy disks.  If \vm\ correlates with bulge velocity 

731: dispersion, \sig, as originally suggested by Whitmore et al. (1979), \vm\ can 

732: be used in place of \sig\ to define a \mvm\ relation to substitute for the more 

733: traditional \msigma\ relation. A number of authors have suggested that 

734: galaxies over a wide range in morphological types obey a tight correlation 

735: between \vm\ and \sig, to the point that \vm\ can actually replace, and 

736: perhaps should be regarded as more fundamental than, \sig\ (Ferrarese 2002; 

737: Baes et al. 2003; Pizzella et al.  2005).  This result has been challenged by 

738: Courteau et al. (2007) and Ho (2007a), who found, using much larger samples, 

739: that the ratio \vm/\sig\ shows significant intrinsic scatter and systematic 

740: variation across the Hubble sequence.  Nevertheless, a \vmsig\ relation 

741: {\it does}\ exist, at least statistically, and in circumstances when \sig\ is 

742: difficult or impossible to measure (e.g., in very bright or very distant 

743: AGNs), \vm\ may offer the best or possibly the {\it only}\ means of 

744: constraining the host galaxy.  Such was the motivation behind the recent study 

745: of Ho (2007b; see also 

746: 

747: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

748: \vskip 0.3cm

749: %%BoundingBox: 60 30 410 760

750: \figurenum{6}

751: \begin{figure*}[t]

752: \centerline{\psfig{file=f6.eps,width=19.5cm,angle=-90}}

753: \figcaption[fig6.ps]{

754: The location of the \hi\ sample on the line intensity ratio diagrams

755: ({\it a}) log \oiii\ \lamb5007/H\bet\ vs. log \nii\ \lamb6583/H\al\ and

756: ({\it b}) log \oiii\ \lamb5007/H\bet\ vs. log \sii\ \lamb\lamb6716, 6731/H\al.

757: Objects detected and undetected in \hi\ are plotted as solid circles and

758: triangles, respectively.  The crosses represent \hii\ nuclei and the open

759: squares AGNs from the Palomar survey of nearby galaxies (Ho et al. 1997a).

760: The \hi\ sample mostly occupies the locus of Seyfert galaxies, with no 

761: apparent variation with \hi\ detection.

762: \label{fig6}}

763: \end{figure*}

764: \vskip 0.3cm

765: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

766: 

767: \noindent

768: Shields et al. 2006), who used the width of the 

769: rotational CO line, locally calibrated against \hi, to infer the masses of the 

770: host galaxies of high-redshift quasars.   The line width method, however, can 

771: prove to be useful even under less extreme conditions.  As discussed at length 

772: in Greene \& Ho (2006a), a variety of factors conspire to make measurement of 

773: \sig\ very challenging in galaxies containing bright AGNs, regardless of their 

774: redshift.  To bypass this difficulty, many studies resort to using gas velocity 

775: dispersions measured from narrow nebular lines, but this shortcut has its own 

776: set of complications (Greene \& Ho 2005a).   Others skip the kinematical route 

777: altogether and instead use the host galaxy (or bulge, if available) luminosity 

778: as the variable to relate to the BH mass.  However, cleanly 

779: 

780: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

781: \vskip 0.3cm

782: \figurenum{7}

783: %%BoundingBox: 20 200 590 630

784: \psfig{file=f7.eps,width=8.5cm,angle=0}

785: \figcaption[fig7.ps]{

786: Distribution of the line intensity ratio

787: \sii\ \lamb6716/\sii\ \lamb6731, which is inversely proportional to the

788: electron density, for objects detected (hashed histograms) and

789: undetected (open, dashed histograms) in \hi.

790: \label{fig7}}

791: \vskip 0.3cm

792: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

793: 

794: \noindent

795: decomposing the 

796: underlying galaxy, much less its bulge, in the presence of a bright AGN core 

797: is often a nontrivial task, even under the most ideal conditions (e.g., Kim 

798: et al. 2007).

799: 

800: Within this backdrop, we explored the correlation between BH mass and \vm.    

801: As expected, a loose correlation exists between these two quantities, but the 

802: scatter is enormous, $\sim 1$ dex (Fig.~2{\it a}).   Many of the extreme 

803: outliers correspond to objects with single-peaked and/or clearly asymmetric 

804: \hi\ profiles, while others have especially questionable inclination 

805: corrections; removing these produces a cleaner correlation.

806: Note that apart from being inefficient (roughly 40\% of the sample was 

807: rejected), this step imposes no serious complication, since classical 

808: double-horned line profiles are easy to recognize.  Nonetheless, even after 

809: this filtering, the scatter is still substantial (0.9 dex; Fig.~2{\it b}).  We 

810: identified a number of parameters associated with the AGN that seem to 

811: contribute to the scatter (FWHM of the broad emission lines, nuclear 

812: luminosity, and Eddington ratio), but an important contribution comes from

813: the morphological type variation within the sample (Fig.~2{\it c}).  As 

814: discussed in Courteau et al.  (2007) and Ho (2007a), the zeropoint of the 

815: \vmsig\ relation varies systematically with galaxy concentration (which is 

816: loosely related to galaxy morphology or bulge-to-disk ratio).  If we 

817: arbitrarily shift the zeropoints of the different Hubble type bins to the 

818: average relation for all Hubble types, the scatter decreases somewhat to 

819: $\sim 0.7$ dex (Fig.~2{\it f}).

820: 

821: More promising still seems to be the correlation between BH mass and 

822: galaxy dynamical mass, which requires knowledge of one additional parameter, 

823: namely the galaxy's optical diameter.  Retaining, as before, only the objects

824: with robust inclination corrections and kinematically normal \hi\ profiles, 

825: the $M_{\rm BH}-M_{\rm dyn}$ has an rms scatter of 0.61 dex, which further 

826: improves to 0.55 dex after shifting the different Hubble types to a common 

827: reference point (Fig.~3).  While this scatter is larger 

828: 

829: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

830: \vskip 0.3cm

831: %%BoundingBox:  20 100 600 670

832: \figurenum{8}

833: \begin{figure*}[t]

834: \centerline{\psfig{file=f8.eps,width=19.5cm,angle=0}}

835: \figcaption[fig8.ps]{

836: The variation of nuclear H\al\ luminosity (panels {\it a}\ and {\it b})

837: and Eddington ratio (panels {\it c}\ and {\it d}) with \hi\ content.  Objects

838: detected in \hi\ are plotted as filled symbols, while the nondetections are

839: marked as open symbols with arrows.

840: \label{fig8}}

841: \end{figure*}

842: \vskip 0.3cm

843: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

844: 

845: \noindent

846: than that of the 

847: local AGN \msigma\ relation (0.4 dex; Greene \& Ho 2006b), our estimates 

848: of the dynamical masses leave much room for improvement, both in terms of 

849: getting higher signal-to-noise ratio line profiles and deeper optical images, 

850: which will allow more accurate measurements of inclination angles and 

851: isophotal diameters and better estimates of morphological types.

852: 

853: \vskip 1.0cm

854: \subsection{The Chicken or the Egg}

855: 

856: One of the most intriguing results from our analysis is the tentative 

857: detection, both in the \mvm\ (Fig.~2{\it f}) and $M_{\rm BH}-M_{\rm dyn}$ 

858: (Fig.~3{\it b}) diagrams, of differential growth between the central BH and 

859: the host galaxy.  For a given host galaxy potential (\vm\ or $M_{\rm dyn}$), 

860: AGNs with higher accretion rates (Eddington ratios) have systematically 

861: {\it less}\ massive BHs, implying that in these systems the central BH, still 

862: vigorously accreting near its maximum rate, has yet to reach its final mass, 

863: which, from observations of inactive systems, we know to be a well-defined 

864: constant fraction of the total bulge luminosity or mass (Kormendy \& Richstone 

865: 1995; Magorrian et al. 1998; H\"aring \& Rix 2004).  Many (but not all) of the 

866: objects with still-growing BHs turn out to have narrower broad Balmer lines 

867: (FWHM \lax\ 2000 \kms) because NLS1 galaxies tend to have higher accretion 

868: rates.  At a fixed \vm, the AGNs in our sample with \lbol/\ledd\ \gax\ 0.1 on 

869: average have BHs 0.19 dex (factor of 1.5) less massive than those with 

870: \lbol/\ledd\ \lax\ 0.1; in terms of fixed $M_{\rm dyn}$, the difference in BH 

871: masses between the low- and high-accretion rate subgroups is $\sim$0.41 

872: dex (factor of 2.6).  In their study of type 1 AGNs with stellar velocity 

873: dispersion measurements, Greene \& Ho (2006a) also noticed that the 

874: best-fitting \msigma\ relation for AGNs has a small zeropoint offset of 

875: $-$0.17 dex relative to Tremaine et al.'s (2002) fit of the \msigma\ relation 

876: for inactive galaxies.  A result similar to that of Greene \& Ho (2006a) has 

877: been reported by Shen et al.  (2008), who additionally note that the amplitude 

878: of the \msigma\ relation for AGNs depends on Eddington ratio, in the same 

879: sense that we find in our study.  The overall qualitative agreement among the 

880: three independent studies lends credence to the idea that the most highly 

881: accreting BHs are still actively growing.

882: 

883: Mathur et al. (2001; see also Grupe \& Mathur 2004; Mathur \& Grupe 2005), 

884: using the width of the \oiii\ \lamb 5007 line as a substitute for \sig, a 

885: shortcut that allows large numbers of AGNs to be placed on the \msigma\ 

886: relation, proposed that NLS1s contain BHs that are undermassive with respect 

887: to their bulges, by as much as an order of magnitude.  Employing the same 

888: technique, Bian \& Zhao (2004) arrived at a similar result, as did Wandel 

889: (2002), who used bulge luminosities as a reference.  The claim that NLS1s 

890: are ``young'' AGNs, however, has not been universally embraced, in part 

891: because different authors, using the same methods, have arrived at conflicting

892: conclusions (e.g., Wang \& Lu 2001), because of worries concerning the 

893: reliability of the \oiii-based velocity dispersions (Botte et al. 2004; Greene 

894: \& Ho 2005a), and, most seriously, because direct measurements of stellar 

895: \sig\ in NLS1s find no compelling evidence that they deviate from the \msigma\ 

896: relation (Barth et al. 2005; Botte et al. 2005; Greene \& Ho 2006b).  Part of 

897: the confusion stems from the very definition of the NLS1 class.  Strictly 

898: speaking, NLS1s are classified on the basis of optical spectroscopic criteria 

899: (e.g., Pogge 2000), which, apart from the (fairly arbitrary) line width limit 

900: of FWHM $\leq$ 2000 \kms, are neither rigorously defined nor universally 

901: followed.  That many cataloged and well-studied NLS1s have high accretion 

902: rates (e.g., Pounds et al. 1995; Collin \& Kawaguchi 2004) is a statistical 

903: reflection of the fact that bright AGNs in the local Universe tend to have 

904: moderately low-mass BHs (Heckman et al. 2004; Greene \& Ho 2007b), and that 

905: soft X-ray selection preferentially favors systems in their high-state (Greene 

906: \& Ho 2007a).  There is no requirement that NLS1s must have high accretion 

907: rates.  The present sample, in fact, provides two useful examples.  By their 

908: line widths alone, both NGC 4051 and NGC 4395 qualify as NLS1s, and yet the 

909: former has \lbol/\ledd\ = $6.3\times10^{-2}$ while the latter, with 

910: \lbol/\ledd\ = $2.7\times10^{-3}$, formally has the lowest Eddington ratio in 

911: the sample.

912: 

913: That \mbh/\vm\ and \mbh/$M_{\rm dyn}$ decrease {\it systematically}\ with

914: increasing \lbol/\ledd\ suggests not only that the growth of the BH and the

915: growth of the galaxy are not finely synchronized, but also that there is a

916: preferential sequence to their coevolution: during any particular common

917: growth event, star formation {\it precedes}\ BH growth.  Thus, at least for 

918: the local sample under consideration, we have a tentative solution to the 

919: proverbial ``the chicken or the egg problem.''  Phenomenologically, starburst 

920: activity comes first, followed by AGN activity---not the other way around, and 

921: probably not concurrently.  This resembles the scenario envisioned by Ho 

922: (2005a, 2005b; see also Kim et al. 2006) in his analysis of the star formation 

923: rate in nearby quasars.  Interestingly, the sequence of events seems to be 

924: reversed at high redshifts, where the best evidence to date suggests that 

925: galaxy growth actually {\it lags}\ BH growth (Peng et al. 2006a, 2006b; Shields

926: et al. 2006; Ho 2007b).  One might speculate that the pathway for BH growth 

927: and its interdependence on the host galaxy could be radically different for 

928: luminous quasars in massive, early-type galaxies, compared to the much more 

929: modest AGNs under consideration here, which reside mostly in later type, disk 

930: systems (e.g., Hopkins \& Hernquist 2006).  On the other hand, we note that 

931: moderate-redshift ($z=0.36$) AGNs, mostly Seyfert galaxies not too dissimilar 

932: from those in our sample, also follow the trend of high-redshift quasars in 

933: showing a higher ratio of BH mass to host galaxy (bulge) mass (Woo et al. 

934: 2006; Treu et al. 2007).  We cannot, at the moment, reconcile these 

935: conflicting results, which urgently need to be verified.

936: 

937: We end the discussion on BH growth with some cautionary remarks.  Throughout 

938: this study we take at face value that the virial BH masses contain no major 

939: sources of systematic error, other than the $\sim$0.5 dex uncertainty in the 

940: zeropoint (Greene \& Ho 2005b; Peterson 2007).  As discussed in Greene \& Ho 

941: (2007c; Appendix), however, our assumption that the virial mass formalism is 

942: invariant with respect to AGN properties is, at the moment, more an article of 

943: faith than an indisputable fact.  The size-luminosity relation on which the 

944: virial formalism depends continues to undergo revision and is especially poorly 

945: constrained at the low-luminosity end (Bentz et al. 2006).  The very structure

946: of the broad-line region, and hence the geometric (``$f$'') factor of the 

947: virial relation, may depend on accretion rate or Eddington ratio, but 

948: presently we have no reliable means to quantify this effect (Greene \& Ho 

949: 2007c).  Strictly speaking, therefore, the trend of decreasing \mbh/\vm\ and 

950: \mbh/$M_{\rm dyn}$ with increasing \lbol/\ledd\ can be interpreted as evidence 

951: that \mbh\ is systematically underestimated in high-\lbol/\ledd\ AGNs.  

952: Resolving this degeneracy lies beyond the scope of this work.  At the 

953: moment we also cannot prove that the identified trends are not the result of 

954: subtle mass-dependent selection effects.  In their investigation of the local 

955: BH mass function of nearby AGNs, Greene \& Ho (2007b) note that low-mass BHs, 

956: which even at their Eddington limit are still relatively faint AGNs, 

957: preferentially make it into a magnitude-limited survey such as SDSS if they 

958: reside in more luminous (more massive) galaxies.  

959: 

960: \subsection{Host Galaxies and Environment}

961: 

962: Even though most sizable nearby galaxies harbor central BHs, they exhibit 

963: very feeble traces of nuclear activity (Ho et al. 1997b).  Only a tiny 

964: fraction of the local galaxy population accrete at rates high enough to 

965: qualify them as respectable AGNs (Ho 2004b; Greene \& Ho 2007b).  Why is this?

966: The low space density of luminous AGNs in the local Universe surely reflects 

967: some combination of the overall reduced gas supply in present-day galaxies 

968: as well as their more placid dynamical environments, but precisely which

969: physical parameters---internal or external to the system---actual trigger the 

970: onset of nuclear activity in any given galaxy remains largely an unsolved 

971: problem (Ho et al. 1997c, 2003; Ho 2004b).  In a similar vein, and especially

972: given the recent discussion of the purported link between BH and galaxy 

973: growth, it is of interest to know what impact AGN activity actually has on the

974: properties of the host galaxy.  

975: 

976: Our new \hi\ survey offers an opportunity to examine these issues from some 

977: fresh angles.  Despite the limited quality of the data, our visual examination 

978: the optical images do not give the impression that the galaxies or their 

979: immediate environments are particularly unusual.  Some clear cases of 

980: interacting or binary systems exist within a separation of 50 kpc (Ho et al. 

981: 2008; see Fig.~3{\it g}, SDSS~J165601.61+211241.2; Fig.~4{\it b}, 

982: SDSS~J130241.53+040738.6), but overall the optical morphologies of the hosts

983: appear undisturbed and large (i.e. nearly equal-mass) physical companions are 

984: rare.  There are also no discernible differences between the objects 

985: detected and undetected in \hi\ (cf. Fig.~3 vs. Fig.~4 in Ho et al. 2008).

986: 

987: Excluding the objects with anomalous \hi\ profiles, the rest of the sample is 

988: not just morphologically normal but also dynamically normal.  We draw this 

989: inference from our analysis of the Tully-Fisher relation 

990: for the active galaxies, which, besides the anticipated larger scatter due 

991: to various measurement uncertainties, reveals no striking differences compared 

992: to regular disk galaxies (Fig.~4).  As an aside, it is worth making an obvious 

993: point: the fact that our objects obey the Tully-Fisher relation implies that 

994: the \hi\ must be roughly regularly distributed, at least regular enough so

995: that its integrated line width traces the flat part of the galaxy rotation 

996: curve.  Hutchings et al. (1987) and Hutchings (1989) previously failed to 

997: see a clear Tully-Fisher relation for their sample of AGNs because of the 

998: high frequency of asymmetric \hi\ profiles.  By contrast, the more extensive 

999: study by Whittle (1992) found that Seyfert galaxies do define a Tully-Fisher 

1000: relation, but one that is offset toward lower velocities compared to normal 

1001: spiral galaxies.  He interpreted this as an indication that Seyfert galaxies 

1002: have lower mass-to-light ratios (by a factor of $1.5-2$), possibly as a 

1003: result of enhanced star formation.  Our results do not agree with Whittle's.  

1004: We have verified that the discrepancy between the two studies does not stem 

1005: from the different fiducial relations chosen for inactive spirals.  While 

1006: Whittle's reference Tully-Fisher relation has a much steeper slope than the 

1007: one we adopted (from Verheijen 2001), the two zeropoints are very similar.  We 

1008: speculate, but cannot prove, that the low-velocity offset seen by Whittle may 

1009: be caused by contamination from objects with kinematically peculiar, 

1010: preferentially narrow \hi\ profiles.  Inspection of our data (Fig.~4) reveals 

1011: that prior to excluding the objects with asymmetric or single-peaked profiles 

1012: our sample also has a mild excess of low-velocity points.  Indeed, a 

1013: nonnegligible fraction of all galaxies, irrespective of their level of nuclear 

1014: activity, show this behavior, which Ho (2007a) attributes to dynamically 

1015: unrelaxed gas acquired either through minor mergers or primordial accretion.  

1016: 

1017: Enhanced star formation frequently seems to precede nuclear activity (Kauffmann 

1018: et al. 2003; see discussion in Ho 2005b), but at least for the AGN luminosities

1019: probed in this study, the extent of the young stellar population has left no 

1020: visible imprint on the mass-to-light ratio of host galaxies, insofar as 

1021: we can gauge from the Tully-Fisher relation.

1022: 

1023: Do AGN host galaxies possess a higher fraction of kinematically anomalous 

1024: \hi\ profiles than inactive galaxies?  Our study, as in previous ones 

1025: (Mirabel \& Wilson 1984; Hutchings et al. 1987), tentatively suggests that the

1026: answer is yes, but a quantitative comparison is difficult.  Single-dish 

1027: surveys of isolated, (mostly) inactive disk galaxies also report frequent 

1028: detections of asymmetric \hi\ line profiles (e.g., Baldwin et al. 1980; Lewis 

1029: et al.  1985; Matthews et al. 1998), which are attributed to noncircular 

1030: motions, unresolved companions, or genuine perturbations in the spatial 

1031: distribution of the gas. Richter \& Sancisi (1994) and Haynes et al. (1998), 

1032: for example, find asymmetric profiles in roughly 50\% of the spiral galaxies 

1033: they surveyed.  Similarly, Lewis (1987) studied a large sample of face-on 

1034: spiral galaxies and found narrow \hi\ profiles to be rare.  He interprets this 

1035: result to imply that a substantial fraction of the galaxies must have 

1036: distorted \hi\ distributions, whose large-scale motion is misaligned with 

1037: respect to the plane of the stellar disk.  Thus, in absolute terms the 

1038: frequency of non-classical \hi\ profiles in our sample (44\%; \S3.5) is very 

1039: similar to the frequency reported for inactive galaxies, but we are wary to

1040: draw any firm conclusions from this because the signal-to-noise ratio of our

1041: spectra is generally much lower than those of the control samples, and 

1042: because of the qualitative nature of our profile classification.  

1043: 

1044: The kinematically anomalous objects account for 17\% of the 792 galaxies studied 

1045: by Ho (2007a), but this fraction depends on Hubble type, being more common 

1046: among early-type systems.  Among S0 galaxies, the fraction reaches 40\% (Ho 

1047: 2007a), nearly identical to the frequency found among the active systems 

1048: studied here, which, although strictly not all S0 galaxies, nonetheless tend 

1049: to be bulge-dominated disk galaxies.  As in the case of inactive galaxies 

1050: (Ho 2007a), the kinematic peculiarity for the AGN sample cannot be attributed 

1051: to tidal interactions or comparable-sized nearby neighbors. 

1052: 

1053: \vspace{0.3cm}

1054: \subsection{Gas Content and Implications for AGN Feedback Models}

1055: 

1056: Insofar as their \hi\ content is concerned, the host galaxies of the AGNs in 

1057: our sample are endowed with plenty of gas.  The most meaningful metric is the 

1058: specific rather than the absolute gas mass, and because this quantity spans

1059: a wide range across the Hubble sequence (Roberts \& Haynes 1994), we can 

1060: sharpen the comparison even further by specifying the morphological type of

1061: the galaxy.  This exercise (Fig.~5) convinces us that type~1 AGNs are {\it at 

1062: least}\ as gas-rich as inactive galaxies, and among early-type hosts, 

1063: beginning with Sa spirals and certainly by the time we reach S0s and Es, AGN 

1064: hosts appear to be even {\it more}\ gas-rich than their inactive counterparts.

1065: For example, the sensitive survey of 12 inactive E and S0 galaxies by Morganti 

1066: et al. (2006) detected \hi\ emission in nine objects, of which only one has 

1067: $M_{{\rm H~I}}/L_{B,{\rm host}} > 0.1$.  By contrast, among the 30 active E 

1068: and S0 galaxies in our survey that were detected in \hi, 26 (87\%) have 

1069: $M_{{\rm H~I}}/L_{B,{\rm host}} > 0.1$.  Our results are qualitatively 

1070: consistent with those of Bieging \& Biermann (1983) and Mirabel \& Wilson 

1071: (1984).  Because of the nature of the control sample of inactive objects (only 

1072: detections are available), we hesitate to be more specific, but E and S0 

1073: galaxies hosting AGNs may be overabundant in \hi\ by as much as an order of 

1074: magnitude.   Previous studies of nearby elliptical and S0 galaxies have hinted 

1075: of a possible statistical connection between \hi-richness and radio activity 

1076: (Dressel et al. 1982; Jenkins 1983; Morganti et al. 2006), two well-known 

1077: examples being the low-power radio galaxies NGC~1052 (Knapp et al. 1978) and 

1078: NGC~4278 (Raimond et al. 1981).  The validity of our conclusion, of course, 

1079: critically depends on the accuracy of the morphological types.  However, to 

1080: bring the average $M_{{\rm H~I}}/L_{B,{\rm host}}$ ratios of the E and S0 

1081: subgroups to conform to the average values of inactive galaxies, we would have 

1082: to shift the morphological types to as late as Sb or even Sc spirals.  It 

1083: seems unlikely that the morphologies could be so blatantly wrong.  

1084: Alternatively, perhaps we have grossly miscalculated the host galaxy optical 

1085: luminosities.  Without improved optical imaging (\S5), however, these 

1086: alternative explanations remain purely hypothetical.  

1087: 

1088: The global gas content bears no relationship to the level of AGN activity, 

1089: either in terms of absolute luminosity or Eddington ratio (Fig.~8).  Ho et al. 

1090: (2003) had come to the same conclusion for nearby LINERs and Seyferts, but now 

1091: the same holds for AGNs on average 2--3 orders of magnitude more luminous.  

1092: Whether \hi\ is detected or not also makes no impact whatsoever on the optical 

1093: spectrum (Figs.~6 and 7).  Perhaps none of this should come as a surprise. 

1094: The \hi\ data, after all, probe spatial scales vastly disproportionate to those 

1095: relevant for the AGN central engine.  Moreover, the accretion rates required

1096: to power the observed activity are quite modest.  The H\al\ luminosities of 

1097: our sample range from $\sim 10^{40}$ to $10^{44}$ \lum, with a median value of 

1098: $4\times 10^{41}$ \lum.  Adopting the H\al\ bolometric correction given in 

1099: Greene \& Ho (2007b) and a radiative efficiency of 0.1, this corresponds 

1100: to a mass accretion rate of merely $\sim 2\times 10^{-2}$ \solmass\ \peryr.  

1101: 

1102: Many current models of galaxy formation envision AGN feedback to play a 

1103: pivotal role in controlling the joint evolution of central BHs and their host 

1104: galaxies. During the major merger of two gas-rich galaxies, each initially 

1105: seeded with its own BH, gravitational torques drive a large fraction of the 

1106: cold gas toward the central region of the resulting merger remnant.  Most of 

1107: the gas forms stars with high efficiency in a nuclear starburst, at the 

1108: same time feeding the central BH at an Eddington-limited rate.  This process 

1109: consumes a large amount of the original cold gas reservoir.  The combined 

1110: energy generated from supernova explosions and the central engine wrecks havoc 

1111: on the rest of the interstellar medium in the galaxy, shocking it to high 

1112: temperatures and redistributing it to large scales, thereby shutting off 

1113: further star formation and accretion.  While the specific formulation of the 

1114: problem and the methodology for solving it may vary from study to study (e.g., 

1115: Granato et al. 2004; Springel et al. 2005; Croton et al. 2006; Hopkins et al. 

1116: 2006; Hopkins \& Hernquist 2006; Sijacki et al. 2007; Di~Matteo et al. 2008), 

1117: one generic prediction runs constant: the onset of nuclear activity, 

1118: especially during the peak phase of accretion when the central object unveils 

1119: itself as an optically visible AGN, in concert with supernova feedback from 

1120: the accompanying starburst, liberates so much energy on such a short timescale 

1121: that the bulk of the cold gas gets expelled from the galaxy.  None of the 

1122: existing models makes very precise predictions about the cold gas content 

1123: during the evolution of the system, but it seems clear, regardless of the 

1124: details, that during the optically visible (unobscured) phase of the AGN the 

1125: host galaxy should be {\it deficient}\ in cold gas.  This expectation is 

1126: inescapable if AGN feedback is to have as dramatic an effect on the evolution 

1127: of the host galaxy as has been suggested.

1128: 

1129: Our observations present a challenge to the framework of AGN feedback just 

1130: outlined.  Far from being gas-deficient, the host galaxies of optically 

1131: selected type~1 AGNs are, if anything, unusually gas-rich, and at the very 

1132: least we can state with confidence that their gas content is normal.  To be 

1133: fair, the vast majority of our sample consists of Seyfert galaxies.  They are 

1134: hosted in disk (S0 and spiral) galaxies, which probably never experienced many 

1135: major mergers, and their nuclei fall far short of luminosity threshold of 

1136: quasars, which, although rare, dominate the BH mass density in the Universe 

1137: (e.g., Yu \& Tremaine 2002).  These objects probably fall outside of the 

1138: purview of major merger-driven models (Hopkins \& Hernquist 2006).  Our 

1139: sample, on the other hand, {\it does}\ contain seven objects that formally 

1140: satisfy the luminosity criterion of quasars, of which four (PG~0844+349, 

1141: PG~1426+015, PG~2130+099, and RX~J0608.0+3058) are detected in \hi.   All four 

1142: are very gas-rich, with \hi\ masses ranging from \mhi\ = $5 \times10^9$ to 

1143: $2\times10^{10}$ \solmass\ and $M_{{\rm H~I}}/L_{B,{\rm host}}$ values that 

1144: are not obviously low.  

1145: 

1146: Beyond the present set of observations, two other lines of evidence contradict 

1147: the notion that optically selected quasars are gas-deficient.  Scoville et al. 

1148: (2003) performed an unbiased search for CO emission from all $z < 0.1$ quasars 

1149: from the ultraviolet-selected sample of Palomar-Green (PG) sources (Schmidt \& 

1150: Green 1983), finding that the majority of them ($\sim$75\%) contain abundant 

1151: molecular gas, ranging from $M_{{\rm H_2}} \approx 10^{9}$ to $10^{10}$ 

1152: \solmass; those that were undetected have upper limits consistent with the 

1153: detections.  Bertram et al. (2007) report very similar statistics from their 

1154: recent CO study of low-luminosity quasars selected from the Hamburg/ESO 

1155: survey.  These molecular gas masses are comparable to, if not greater than, 

1156: those typically found in mid- to late-type spirals.  Ho's (2005a) compilation 

1157: of CO observations of other PG quasars further reinforces this point, and even 

1158: more extreme molecular gas masses ($M_{{\rm H_2}} \approx 10^{10}-10^{11}$ 

1159: \solmass) have been detected in high-redshift quasars (Solomon \& Vanden~Bout 

1160: 2005, and references therein).  The Milky Way, for reference, has 

1161: $M_{{\rm H_2}} = 2\times10^9$ \solmass\ (Scoville \& Good 1989).  Of the three 

1162: PG quasars that overlap with our \hi\ sample, for instance, PG~1426+015 has 

1163: $M_{{\rm H_2}} = 6.5\times10^9$ \solmass, PG~2130+099 has $M_{{\rm H_2}} = 

1164: 3\times10^9$ \solmass, and PG~0844+349 has an upper limit of $M_{{\rm H_2}} = 

1165: 1\times10^9$ \solmass.  Combining this with the \hi\ masses, all three objects 

1166: have a total (\hi\ $+\,{\rm H}_2$) neutral gas mass in excess of $10^{10}$ 

1167: \solmass.  Yet another way to recognize that PG quasars have plenty of cold 

1168: gas comes from noting that substantial amounts of dust have been detected in 

1169: them.  Haas et al. (2003) appraised the dust content in a sample of 51 PG 

1170: quasars by combining far-infrared continuum observations along with millimeter 

1171: and submillimeter data.  Their derived dust masses range from $M_{{\rm dust}} 

1172: =7\times10^5$ to $3\times10^8$ \solmass, with a median value of $8.8\times10^6$ 

1173: \solmass.  For a Galactic molecular gas-to-dust ratio of 150, this corresponds 

1174: to $M_{{\rm H_2}} = 1.3\times10^9$ \solmass; for a higher gas-to-dust ratio of 

1175: 600 (Young \& Scoville 1991), $M_{{\rm H_2}} = 5.3\times10^9$ \solmass.  These

1176: indirect estimates of molecular gas mass are consistent with those derived 

1177: from CO observations.

1178: 

1179: To graphically compare the gas masses of PG quasars side-by-side with those 

1180: of the lower luminosity AGNs in our sample, we converted the molecular and 

1181: dust masses into approximate \hi\ masses.  We adopt, for the sake of 

1182: illustration, $M_{{\rm H_2}}$/\mhi\ = 3, a value characteristic of early-type 

1183: spirals (Young \& Scoville 1991), and $M_{{\rm H_2}}/M_{{\rm dust}} = 600$.

1184: We calculated BH masses for the sources using the H\bet\ line widths from 

1185: Bororon \& Green (1992) and optical continuum luminosities from the 

1186: spectrophotometry of Neugebauer et al. (1987), and then inferred $B$-band host 

1187: galaxy luminosities using the correlation between BH mass and bulge luminosity 

1188: given in Kormendy \& Gebhardt (2001), for simplicity assigning all the 

1189: luminosity to the bulge.  As summarized in the bottom panel of Figure~5, PG 

1190: quasars span almost 4 orders of magnitude in $M_{{\rm H~I}}/L_{B,{\rm host}}$, 

1191: from objects as gas-poor as the most extreme ellipticals to as gas-rich as the 

1192: most gas-dominated dwarf irregulars.  Although this exercise becomes 

1193: progressively more uncertain for high-redshift objects, we can attempt a 

1194: similar order-of-magnitude calculation for the nine high-redshift quasars 

1195: listed in Ho (2007b), using BH masses listed therein and the molecular gas 

1196: masses given in Solomon \& Vanden~Bout (2005).  Under the same assumptions 

1197: adopted for the PG quasars, we find an average value of 

1198: $\log \langle M_{{\rm H~I}}/L_{B,{\rm host}} \rangle \approx -0.9$, in the 

1199: middle of the range observed for PG quasars (Fig.~5).  In fact, this estimate, 

1200: which assumes a local value for the ratio of BH mass to bulge luminosity, is 

1201: most likely a lower limit because high-redshift quasars 

1202: in general (Peng et al. 2006a, 2006b) and this subset of objects in particular 

1203: (Shields et al. 2006; Ho 2007b) seem to exhibit a higher ratio (by a factor of 

1204: a few) of BH mass to host galaxy mass.  

1205: 

1206: Given the lack of specific theoretical predictions, we cannot attempt a 

1207: quantitative comparison of our results with AGN feedback models.  Nonetheless, 

1208: we wish to stress that the typical unobscured quasar possesses a gas reservoir 

1209: normal for its stellar content and does {\it not}\ appear to be depleted in 

1210: cold gas.  Whether this poses a difficulty or not for AGN feedback models 

1211: deserves further consideration.

1212: 

1213: \section{Future Directions}

1214: 

1215: The pilot study presented here can be improved and extended in several 

1216: ways.  A number of the \hi\ observations would benefit from longer 

1217: integrations.  Roughly $1/3$ of our sample was undetected in the current 

1218: experiment, and the upper limits on their \hi\ masses remain relatively high;

1219: nearly half of the nondetections have limits above \mhi\ $=\,10^{10}$ 

1220: \solmass. It would be useful to improve these limits by a factor of a few, 

1221: down to \mhi\ $\approx 5\times 10^{9}$ \solmass, the scale of the Milky Way.

1222: Higher signal-to-noise ratio spectra of the detected sources would allow us 

1223: to better study the \hi\ line profiles, especially to quantify the apparent 

1224: excess in line asymmetry and its implications for the dynamical state of 

1225: the gas and the circumgalactic environment of the host.  In terms of new 

1226: observations, it would be particularly worthwhile to observe a larger sample 

1227: of high-Eddington ratio systems in order to verify the trends with Eddington 

1228: ratio identified in this study.  This can be easily accomplished using the 

1229: current SDSS database (Greene \& Ho 2007b, 2007c).  To put more stringent 

1230: constraints on AGN feedback scenarios, it would be highly desirable to 

1231: increase the number of high-luminosity AGNs (quasars) in the survey.  Given 

1232: the rarity of nearby quasars and the redshift limit of Arecibo, this will be a 

1233: difficult task, but some progress can still be made.   

1234: 

1235: The current project represents a major initiative in characterizing the cold 

1236: gas content in type~1 AGNs.  We have demonstrated that \hi\ observations can 

1237: yield many useful insights into the nature of AGNs, BHs, and their host 

1238: galaxies.  This tool should be applied to other classes of AGNs.  A useful 

1239: extension of this program would be to observe a well-matched sample of type~2 

1240: (narrow-line) AGNs in order to test their relationship to type~1 objects.  Kim 

1241: et al. (2006; see also Ho 2005b and Lal \& Ho 2007), for example, have argued 

1242: that type~2 quasars are the evolutionary precursors, as opposed to simply 

1243: misoriented counterparts, of type~1 quasars.  If true, type~2 sources are 

1244: likely to be more gas-rich than their type~1 counterparts, or perhaps their 

1245: interstellar medium might be less dynamically relaxed, differences that can 

1246: potentially be discerned through \hi\ observations.   

1247: 

1248: The analysis presented here critically depends on a number of quantities 

1249: derived from optical images of the host galaxy, namely its inclination angle, 

1250: isophotal diameter, luminosity, morphological type, and, to a lesser extent, 

1251: local environment.  All of these measurements can be greatly improved 

1252: with the help of better optical imaging.  In many instances, the SDSS images 

1253: (Ho et al. 2008) simply lack the necessary depth or resolution to yield 

1254: clear-cut measurements of these parameters.  High-resolution (e.g., 

1255: {\it Hubble Space Telescope}) images would be particularly valuable, as they 

1256: would simultaneously yield a reliable decomposition of the AGN core from the 

1257: host galaxy, as well as a quantitative measurement of the bulge component. 

1258: All of the correlations presented in this study should be reexamined once 

1259: such data become available, to see if their scatter can be reduced.

1260: 

1261: We have stressed the utility of \hi\ observations, but, of course, a full

1262: inventory of the cold interstellar medium should include the molecular 

1263: component as well, which, depending on the galaxy type, can dominate over the 

1264: atomic component.  A systematic CO survey of the objects in our \hi\ program 

1265: would be highly desirable, not only to complete the gas inventory but also to 

1266: bootstrap the kinematical analysis to the CO line width.  Unlike the current 

1267: limitations of \hi\ observations, the rotational transitions of CO can be 

1268: observed to almost arbitrarily high redshifts (Solomon \& Vanden~Bout 2005), 

1269: a necessary ingredient to access AGN host galaxies at early epochs.  Ho 

1270: (2007b) discusses the use of the CO Tully-Fisher relation as a powerful 

1271: tool to investigate the host galaxies of high-redshift quasars.

1272: 

1273: \acknowledgements

1274: The work of L.~C.~H. was supported by the Carnegie Institution of Washington

1275: and by NASA grants HST-GO-10149.02 and HST-AR-10969 from the Space Telescope

1276: Science Institute, which is operated by the Association of Universities for

1277: Research in Astronomy, Inc., for NASA, under contract NAS5-26555.  Support for

1278: J.~D. and J.~E.~G. was provided by NASA through Hubble Fellowship grants

1279: HF-01183.01-A and HF-01196, respectively, awarded by the Space Telescope

1280: Science Institute.  We thank Paul Martini for useful correspondence and 

1281: the anonymous referee for helpful suggestions.

1282: 

1283: \begin{thebibliography}{}

1284: 

1285: \bibitem[]{}

1286: Adelman-McCarthy, J. K., et al. 2006, \apjs, 162, 38

1287: 

1288: \bibitem[]{}

1289: Baes, M., Buyle, P., Hau, G.~K.~T., \& Dejonghe, H. 2003, \mnras, 341, L44

1290: 

1291: \bibitem[]{}

1292: Baldwin, J.~E., Lynden-Bell, D., \& Sancisi, R. 1980, \mnras, 193, 313

1293: 

1294: \bibitem[]{}

1295: Barth, A.~J., Greene, J. E., \& Ho, L.~C. 2005, \apj, 619, L151

1296: 

1297: \bibitem[]{}

1298: Barth, A.~J., Ho, L.~C., Rutledge, R. E., \& Sargent, W.~L.~W. 2004, \apj,

1299: 607, 90

1300: 

1301: \bibitem[]{}

1302: Bentz, M. C., Peterson, B. M., Pogge, R. W., Vestergaard, M., \& Onken, C. A. 

1303: 2006, \apj, 644, 133

1304: 

1305: \bibitem[]{}

1306: Bertram, T., Eckart, A., Fischer, S., Zuther, J., Straubmeier, C., Wisotzki, 

1307: L., \& Krips, M. 2007, \aa, 470, 571

1308: 

1309: \bibitem[]{}

1310: Bian, W., \& Zhao, Y. 2004, \mnras, 352, 823

1311: 

1312: \bibitem[]{}

1313: Bieging, J.~H., \& Biermann, P. 1983, AJ, 88, 161

1314: 

1315: \bibitem[]{}

1316: Boroson, T.~A., \& Green, R.~F. 1992, \apjs, 80, 109

1317: 

1318: \bibitem[]{}

1319: Botte, V., Ciroi, S., Di Mille, F., Rafanelli, P., \& Romano, A. 2005, \mnras,

1320: 356, 789

1321: 

1322: \bibitem[]{}

1323: Botte, V., Ciroi, S., Rafanelli, P., \& Di Mille, F. 2004, \aj, 127, 3168

1324: 

1325: \bibitem[]{}

1326: Broeils, A.~H., \& Rhee, M.-H. 1997, \aa, 324, 877

1327: 

1328: \bibitem[]{}

1329: Casertano, S. P. R., \& Shostak, G. S. 1980, \aa, 81, 371

1330: 

1331: \bibitem[]{}

1332: Collin, S., \& Kawaguchi, T. 2004, \aa, 426, 797

1333: 

1334: \bibitem[]{}

1335: Courteau, S., McDonald, M., Widrow, L. M., \& Holtzman, J. 2007, \apj, 655, L21

1336: 

1337: \bibitem[]{}

1338: Cowie, L.~L., et al.  2003, ApJ, 584, L57

1339: 

1340: \bibitem[]{}

1341: Croton, D. J., et al. 2006, MNRAS, 365, 11

1342: 

1343: \bibitem[]{}

1344: De Rijcke, S., Zeilinger, W.~W., Kau, G. K. T., Prugniel, P., \& Dejonghe, H. 

1345: 2007, \apj, 659, 1172

1346: 

1347: \bibitem[]{}

1348: Di Matteo, T., Colberg, J., Springel, V., Hernquist, L., \& Sijacki, D. 

1349: 2008, \apj, in press (astro-ph/0705.2269)

1350: 

1351: \bibitem[]{}

1352: Dressel, L.~L., Bania, T.~M., \& O'Connell, R.~W. 1982, \apj, 259, 55

1353: 

1354: \bibitem[]{}

1355: Jenkins, C.~R. 1983, \mnras, 205, 1321

1356: 

1357: \bibitem[]{}

1358: Fabian, A.~C. 1999, MNRAS, 308, L39

1359: 

1360: \bibitem[]{}

1361: Feigelson, E.~D., \& Nelson, P.~I. 1985, \apj, 293, 192

1362: 

1363: \bibitem[]{}

1364: Ferrarese, L. 2002, \apj, 578, 90

1365: 

1366: \bibitem[]{}

1367: Ferrarese, L., \& Merritt, D. 2000, \apj, 539, L9

1368: 

1369: \bibitem[]{}

1370: Ferrarese, L., Pogge, R.~W., Peterson, B.~M., Merritt, D., Wandel, A., \& 

1371: Joseph, C.~L. 2001, \apj, 555, L79

1372: 

1373: \bibitem[]{}

1374: Filippenko, A.~V., \& Ho, L.~C. 2003, \apj, 588, L13

1375: 

1376: \bibitem[]{}

1377: Gallagher, J.~S., Knapp, G.~R., \& Faber, S.~M. 1981, AJ, 86, 1781

1378: 

1379: \bibitem[]{}

1380: Gebhardt, K., et al. 2000a, \apj, 539, L13

1381: 

1382: \bibitem[]{}

1383: ------. 2000b, \apj, 543, L5

1384: 

1385: \bibitem[]{}

1386: Granato, G.~L., De Zotti, G., Silva, L., Bressan, A., \& Danese, L. 2004,

1387: \apj, 600, 580

1388: 

1389: \bibitem[]{}

1390: Greene, J. E., \& Ho, L. C. 2004, ApJ, 610, 722

1391: 

1392: \bibitem[]{}

1393: ------. 2005a, ApJ, 627, 721

1394: 

1395: \bibitem[]{}

1396: ------. 2005b, ApJ, 630, 122

1397: 

1398: \bibitem[]{}

1399: ------. 2006a, ApJ, 641, 117

1400: 

1401: \bibitem[]{}

1402: ------. 2006b, ApJ, 641, L21

1403: 

1404: \bibitem[]{}

1405: ------. 2007a, ApJ, 656, 84

1406: 

1407: \bibitem[]{}

1408: ------. 2007b, ApJ, 667, 131

1409: 

1410: \bibitem[]{}

1411: ------. 2007c, ApJ, 670, 92

1412: 

1413: \bibitem[]{}

1414: Grupe, D., \& Mathur, S. 2004, \apj, 606, L41

1415: 

1416: \bibitem[]{}

1417: Haas, M., et al. 2003, \aa, 402, 87

1418: 

1419: \bibitem[]{}

1420: H\"aring, N., \&  Rix, H.-W. 2004, \apj, 604, L89

1421: 

1422: \bibitem[]{}

1423: Hartmann, D., \& Burton, W.~B. 1997, Atlas of Galactic Neutral Hydrogen

1424: (Cambridge: Cambridge Univ. Press)

1425: 

1426: \bibitem[]{}

1427: Haynes, M. P. \& Giovanelli, R.  1984, \aj, 89, 758

1428: 

1429: \bibitem[]{}

1430: Haynes, M.~P., van Zee, L., Hogg, D. E., Roberts, M. S., \& Maddalena, R. J. 

1431: 1998, AJ, 115, 62

1432: 

1433: \bibitem[]{}

1434: Heckman, T.~M., Kauffmann, G., Brinchmann, J., Charlot, S., Tremonti, C.,

1435: \& White, S. D. M. 2004, \apj, 613, 109

1436: 

1437: %\bibitem[]{}

1438: %Ho, L.~C. 1996, in The Physics of LINERs in View of

1439: %Recent Observations, ed. M. Eracleous et al. (San Francisco: ASP), 103

1440: 

1441: \bibitem[]{}

1442: Ho, L. C. 2004a, ed., Carnegie Observatories Astrophysics Series, Vol. 1:

1443: Coevolution of Black Holes and Galaxies (Cambridge: Cambridge Univ. Press)

1444: 

1445: \bibitem[]{}

1446: ------. 2004b, in Carnegie Observatories Astrophysics Series, Vol. 1:

1447: Coevolution of Black Holes and Galaxies, ed. L. C. Ho (Cambridge: Cambridge

1448: Univ. Press), 293

1449: 

1450: \bibitem[]{}

1451: ------. 2005a, \apj, 629, 680

1452: 

1453: \bibitem[]{}

1454: ------. 2005b, in Proceedings of Extreme Starbursts (astro-ph/0511157)

1455: 

1456: \bibitem[]{}

1457: ------. 2007a, \apj, 668, 94

1458: 

1459: \bibitem[]{}

1460: ------. 2007b, \apj, 669, 821

1461: 

1462: \bibitem[]{}

1463: Ho, L. C., Darling, J., \& Greene, J. E. 2008, \apjs, in press

1464: 

1465: \bibitem[]{}

1466: Ho, L.~C., Filippenko, A.~V., \& Sargent, W.~L.~W. 1997a, \apjs, 112, 315

1467: 

1468: \bibitem[]{}

1469: ------. 1997b, \apj, 487, 568

1470: 

1471: \bibitem[]{}

1472: ------. 1997c, \apj, 487, 591

1473: 

1474: \bibitem[]{}

1475: ------. 2003, \apj, 583, 159

1476: 

1477: \bibitem[]{}

1478: Hopkins, P. F., \& Hernquist, L. 2006, \apjs, 166, 1

1479: 

1480: \bibitem[]{}

1481: Hopkins, P. F., Hernquist, L., Cox, T. J., Di Matteo, T.,  Martini, P., 

1482: Robertson, B., \& Springel, V. 2006, \apjs, 163, 1

1483: 

1484: \bibitem[]{}

1485: Hutchings, J.~B. 1989, \aj, 98, 524

1486: 

1487: \bibitem[]{}

1488: Hutchings, J.~B., Gower, A.~C., Price, R. 1987, \aj, 93, 6

1489: 

1490: \bibitem[]{}

1491: Kaspi, S., Smith, P.~S., Netzer, H., Maoz, D., Jannuzi, B.~T., \&

1492: Giveon, U. 2000, \apj, 533, 631

1493: 

1494: \bibitem[]{}

1495: Kauffmann, G., et al. 2003, \mnras, 346, 1055

1496: 

1497: \bibitem[]{}

1498: Kim, M., Ho, L. C., \& Im, M. 2006, \apj, 642, 702

1499: 

1500: \bibitem[]{}

1501: Kim, M., Ho, L. C., Peng, C. Y., \& Im, M. 2007, \apj, 658, 107

1502: 

1503: \bibitem[]{}

1504: Knapp, G.~R., Faber, S.~M., \& Gallagher, J.~S. 1978, \aj, 83, 139

1505: 

1506: \bibitem[]{}

1507: Knapp, G.~R., Turner, E.~L., \& Cunniffe, P.~E. 1985, \aj, 90, 454

1508: 

1509: \bibitem[]{}

1510: Kormendy, J., \& Gebhardt, K. 2001, in The 20th Texas Symposium on Relativistic Astrophysics, ed. H. Martel \& J.~C. Wheeler (Melville: AIP), 363

1511: 

1512: \bibitem[]{}

1513: Kormendy, J., \& Richstone, D.~O. 1995, \annrev, 33, 581

1514: 

1515: \bibitem[]{}

1516: Lal, D. V., \& Ho, L. C. 2007, in The Central Engine of Active Galactic Nuclei,

1517: ed. L. C. Ho \& J.-M.  Wang (San Francisco: ASP), 479

1518: 

1519: \bibitem[]{}

1520: Lewis, B.~M. 1987, \apjs, 63, 515

1521: 

1522: \bibitem[]{}

1523: Lewis, B.~M., Helou, G., \& Salpeter, E. E. 1985, \apjs, 59, 151

1524: 

1525: \bibitem[]{}

1526: Magorrian, J., et al.  1998, \aj, 115, 2285

1527: 

1528: \bibitem[]{}

1529: Mathur, S., \& Grupe, D. 2005, \aa, 432, 463

1530: 

1531: \bibitem[]{}

1532: Mathur, S., Kuraszkiewicz, J., \& Czerny, B. 2001, NewA, 6, 321

1533: 

1534: \bibitem[]{}

1535: Matthews, L.~D., van Driel, W., \& Gallagher, J.~S. 1998, \aj, 116, 1169

1536: 

1537: \bibitem[]{}

1538: McLure, R.~J., \& Dunlop, J.~S. 2001, MNRAS, 327, 199

1539: 

1540: \bibitem[]{}

1541: Mirabel, I.~F., \& Wilson, A.~S. 1984, \apj, 277, 92

1542: 

1543: \bibitem[]{}

1544: Morganti, R., et al. 2006, \mnras, 371, 157

1545: 

1546: \bibitem[]{}

1547: Nelson, C.~H., Green, R. F., Bower, G., Gebhardt, K., \& Weistrop, D. 2004,

1548: \apj, 615, 652

1549: 

1550: \bibitem[]{}

1551: Neugebauer, G., Green, R.~F., Matthews, K., Schmidt, M., Soifer, B.~T., \&

1552: Bennett, J. 1987, \apjs, 63, 615

1553: 

1554: \bibitem[]{}

1555: Noordermeer, E., van der Hulst, J. M., Sancisi, R., Swaters, R. A., \& van 

1556: Albada, T. S. 2005, \aa, 442, 137

1557: 

1558: \bibitem[]{}

1559: Onken, C.~A., Ferrarese, L., Merritt, D., Peterson, B.~M., Pogge, R. W.,

1560: Vestergaard, M., \& Wandel, A. 2004, \apj, 615, 645

1561: 

1562: \bibitem[]{}

1563: Peng, C.~Y., Impey, C.~D., Ho, L.~C., Barton, E. J., \& Rix, H.-W. 2006a, 

1564: \apj, 640, 114

1565: 

1566: \bibitem[]{}

1567: Peng, C.~Y., Impey, C.~D., Rix, H.-W., Kochaneck, C.~S., Keeton, C. R., Falco, 

1568: E. E., Leh\'ar, J., \& McLeod, B.~A. 2006b, \apj, 649, 616

1569: 

1570: \bibitem[]{}

1571: Peterson, B.~M. 2007, in The Central Engine of Active Galactic Nuclei,

1572: ed. L. C. Ho \& J.-M.  Wang (San Francisco: ASP), 3

1573: 

1574: \bibitem[]{}

1575: Pizzella, A., Corsini, E., Dalla Bont\'a, E., Sarzi, M., Coccato, L., \& 

1576: Bertola, F. 2005, \apj, 631, 785

1577: 

1578: \bibitem[]{}

1579: Pogge, R.~W. 2000, NewAR, 44, 381

1580: 

1581: \bibitem[]{}

1582: Pounds, K.~A., Done, C., \& Osborne, J.~P. 1995, \mnras, 277, L5

1583: 

1584: \bibitem[]{}

1585: Raimond, E., Faber, S.~M., Gallagher, J.~S., III, \& Knapp, G.~R.  1981, 

1586: \apj, 246, 708

1587: 

1588: \bibitem[]{}

1589: Richter, O.-G., \& Sancisi, R. 1994, \aa, 290, L9

1590: 

1591: \bibitem[]{}

1592: Roberts, M.~S., \& Haynes, M.~P. 1994, \annrev, 32, 115

1593: 

1594: \bibitem[]{}

1595: Rubin, V.~C., Burstein, D., Ford, W.~K., Jr., \& Thonnard, N. 1985, \apj,

1596: 289, 81

1597: 

1598: \bibitem[]{}

1599: Schmidt, M., \& Green, R.~F. 1983, \apj, 269, 352

1600: 

1601: \bibitem[]{}

1602: Scoville, N.~Z., Frayer, D.~T., Schinnerer, E., \& Christopher, M. 2003,

1603: \apj, 585, L105

1604: 

1605: \bibitem[]{}

1606: Scoville, N.~Z., \& Good, J. C. 1989, \apj, 339, 149

1607: 

1608: \bibitem[]{}

1609: Shen, J., Vanden Berk, D. E., Schneider, D. P., \& Hall, P. B. 2008, \aj, 135, 928

1610: 

1611: \bibitem[]{}

1612: Shields, G.~A., Menezes, K. L., Massart, C. A., \& Vanden Bout, P. 2006, 

1613: \apj, 641, 683

1614: 

1615: \bibitem[]{}

1616: Sijacki, D., Springel, V., Di Matteo, T., \& Hernquist, L. 2007, \mnras, 380,

1617: 877

1618: 

1619: \bibitem[]{}

1620: Solomon, P.~M., \& Vanden Bout, P. 2005, \annrev, 43, 677

1621: 

1622: \bibitem[]{}

1623: Springel, V., Di Matteo, T., \& Hernquist, L. 2005, \mnras, 361, 776

1624: 

1625: \bibitem[]{}

1626: Springob, C. M., Haynes, M. P., Giovanelli, R., \& Kent, B. R. 2005, 

1627: \apjs, 160, 149

1628: 

1629: \bibitem[]{}

1630: Tremaine, S., et al. 2002, \apj, 574, 740

1631: 

1632: \bibitem[]{}

1633: Treu, T., Woo, J.-H., Malkan, M. A., \& Blandford, R. D. 2007, \apj, 667, 117

1634: 

1635: \bibitem[]{}

1636: Tully, R.~B., \& Fisher, J.~R. 1977, \aa, 54, 661

1637: 

1638: \bibitem[]{}

1639: Verheijen, M.~A.~W. 2001, \apj, 563, 694

1640: 

1641: \bibitem[]{}

1642: Wandel, A. 2002, \apj, 565, 762

1643: 

1644: \bibitem[]{}

1645: Wang, T.-G., \& Lu, Y.-J. 2001, \aa, 377, 52

1646: 

1647: \bibitem[]{}

1648: Wardle, M., \& Knapp, G.~R. 1986, \aj, 91, 23

1649: 

1650: \bibitem[]{}

1651: Whitmore, B.~C., Schechter, P. L., \& Kirshner, R. P. 1979, \apj, 234, 68

1652: 

1653: %\bibitem[]{}

1654: %Whittle, M. 1985, \mnras, 213, 33

1655: 

1656: \bibitem[]{}

1657: Whittle, M. 1992, \apj, 387, 121

1658: 

1659: \bibitem[]{}

1660: Woo, J.-H., Treu, T., Malkan, M. A., \& Blandford, R. D. 2006, \apj, 645, 900

1661: 

1662: \bibitem[]{}

1663: Yoo, J., \& Miralda-Escud\'e, J. 2004, ApJ, 614, L25

1664: 

1665: \bibitem[]{}

1666: Young, J.~S., \& Scoville, N.~Z. 1991, \annrev, 29, 581

1667: 

1668: \bibitem[]{}

1669: Yu, Q., \& Tremaine, S. 2002, MNRAS, 335, 965

1670: \end{thebibliography}

1671: 

1672: \end{document}

1673: