1: %\documentclass[preprint]{aastex}

2: \documentclass{emulateapj}

3: \usepackage{apjfonts}

4: \usepackage[]{natbib}

5: \usepackage{graphics}

6:      

7: \newcommand{\etal}{et al.}  

8: \newcommand{\per}{\ensuremath{^{-1}}}

9: \newcommand{\persq}{\ensuremath{^{-2}}}

10: \newcommand{\percu}{\ensuremath{^{-3}}}

11: \newcommand{\percucm}{cm\ensuremath{^{-3}}}

12: \newcommand{\hal}{H\ensuremath{\alpha}}

13: \newcommand{\hbeta}{H\ensuremath{\beta}} 

14: \newcommand{\hst}{\emph{HST}}

15: \newcommand{\mstar}{\ensuremath{M_{\star}}}

16: \newcommand{\msun}{\ensuremath{M_{\odot}}}

17: \newcommand{\lsun}{\ensuremath{L_{\odot}}}

18: \newcommand{\kms}{km s\ensuremath{^{-1}}} 

19: \newcommand{\ebv}{\ensuremath{E(B-V)}}

20: \newcommand{\ncrit}{\ensuremath{n_{crit}}}

21: \newcommand{\lam}[1]{\ensuremath{\lambda}#1}

22: \newcommand{\mbh}{\ensuremath{M_\mathrm{BH}}}

23: \newcommand{\hn}{\hal+[\ion{N}{2}]}

24: \newcommand{\chisq}{\ensuremath{\chi^2}}

25: \newcommand{\sigmastar}{\ensuremath{\sigma_\star}}

26: \newcommand{\mbul}{\ensuremath{M_\mathrm{bulge}}}

27: \newcommand{\msigma}{\ensuremath{\mbh-\sigmastar}}

28: \newcommand{\rblr}{\ensuremath{r_{\mathrm{BLR}}}}

29: \newcommand{\vfwhm}{\ensuremath{v_{\mathrm{FWHM}}}}

30: \newcommand{\mgb}{Mg\ensuremath{b}}

31: \newcommand{\rosat}{\emph{ROSAT}}

32: \newcommand{\lbol}{\ensuremath{L_{\mathrm{bol}}}}

33: \newcommand{\lbul}{\ensuremath{L_{\mathrm{bul}}}}

34: \newcommand{\ledd}{\ensuremath{L_{\mathrm{Edd}}}}

35: \newcommand{\dn}{\ensuremath{D_n(4000)}}

36: \newcommand{\hdeltaa}{H\ensuremath{\delta_A}}

37: \newcommand{\lstar}{\ensuremath{L^\star}}

38: \newcommand{\fwhmo}{\ensuremath{\mathrm{FWHM}_\mathrm{[OIII]}}}

39: \newcommand{\fwhms}{\ensuremath{\mathrm{FWHM}_\mathrm{[SII]}}}

40: \newcommand{\sigmao}{\ensuremath{\sigma_\mathrm{[OIII]}}}

41: \newcommand{\lothree}{\ensuremath{L_\mathrm{[OIII]}}}

42: 

43: 

44: 

45: 

46: \shorttitle{LOW-MASS SEYFERT 2 GALAXIES} 

47: \shortauthors{BARTH ET AL.}

48: 

49: \begin{document} 

50: 

51: \title{Low-Mass Seyfert 2 Galaxies in the Sloan Digital Sky Survey}

52: 

53: \author{Aaron J. Barth}

54: \affil{Department of Physics and Astronomy, 4129 Frederick

55:   Reines Hall, University of California, Irvine, CA 92697-4575}

56: \author{Jenny E. Greene\altaffilmark{1}}

57: \affil{Department of Astrophysical Sciences, Princeton

58:   University, Princeton, NJ, 08544}

59: \altaffiltext{1}{Hubble Fellow}

60: \and

61: \author{Luis C. Ho}

62: \affil{The Observatories of the Carnegie Institution of

63:   Washington, 813 Santa Barbara Street, Pasadena, CA 91101}

64: 

65: 

66: \begin{abstract}

67: 

68: We describe a sample of low-mass Seyfert 2 galaxies selected from the

69: Sloan Digital Sky Survey, having a median absolute magnitude of $M_g =

70: -19.0$ mag.  These galaxies are Type 2 counterparts to the Seyfert 1

71: galaxies with intermediate-mass black holes identified by Greene \& Ho

72: (2004).  Spectra obtained with the \emph{Echellette Spectrograph and

73: Imager} at the Keck Observatory are used to determine the central

74: stellar velocity dispersions and to examine the emission-line

75: properties.  Overall, the stellar velocity dispersions are low

76: ($\sim40-90$ \kms), and we find 12 objects having $\sigmastar < 60$

77: \kms, a range where very few Seyfert 2 galaxies were previously known.

78: The sample follows the correlation between stellar velocity dispersion

79: and FWHM([\ion{O}{3}]) seen in more massive Seyfert galaxies,

80: indicating that the narrow-line FWHM values are largely determined by

81: virial motion of gas in the central regions of the host galaxies, but

82: the [\ion{O}{3}] emission lines exhibit a higher incidence of redward

83: asymmetries and double-peaked profiles than what is found in typical

84: Seyfert samples.  Using estimates of the black hole masses and AGN

85: bolometric luminosities, we find that these galaxies are typically

86: radiating at a high fraction of their Eddington rate, with a median

87: $\lbol/\ledd = 0.4$.  We identify one galaxy, SDSS

88: J110912.40+612346.7, as a Type 2 analog of the nearby dwarf Seyfert 1

89: galaxy NGC 4395, with a nearly identical narrow-line spectrum and a

90: dwarf spiral host of only $M_g = -16.8$ mag.  The close similarities

91: between these two objects suggest that the obscuring torus of AGN

92: unification models may persist even at the lowest luminosities seen

93: among Seyfert galaxies, below $\lbol = 10^{41}$ ergs s\per.

94: Spectropolarimetry observations of four objects do not reveal any

95: evidence for polarized broad-line emission, but SDSS

96: J110912.40+612346.7 has a continuum polarization significantly in

97: excess of the expected Galactic foreground polarization, possibly

98: indicative of scattered light from a hidden nucleus.  Forthcoming

99: observations of this sample, including X-ray and mid-infrared

100: spectroscopy, can provide new tests of the obscuring torus model for

101: active galaxies at low luminosities.

102: 

103: \end{abstract}

104: 

105: \keywords{galaxies: active --- galaxies: kinematics and dynamics ---

106: galaxies: nuclei --- galaxies: Seyfert}

107: 

108: 

109: 

110: \section{Introduction}

111: 

112: 

113: The majority of active galactic nuclei (AGNs) are found in giant

114: galaxies with substantial bulges, and there is a dramatic drop in the

115: AGN fraction for late Hubble types \citep{hfs97} and for host galaxies

116: with stellar masses below $\sim10^{10}$ \msun\ \citep{kau03agn}.  This

117: trend is consistent with the generally accepted scenario in which

118: black hole growth and bulge growth are closely coupled, as expected

119: from the correlations between black hole mass and bulge mass

120: \citep{kr95}, and between black hole mass and stellar velocity

121: dispersion \citep[the \msigma\ relation;][]{fm00, geb00}.  The low AGN

122: fraction in late-type and low-mass galaxies arises from a combination

123: of factors.  Perhaps most important, the black hole occupation

124: fraction in low-mass galaxies is apparently below unity, as

125: demonstrated by the stellar-dynamical non-detections of central black

126: holes in the Local Group galaxies M33 \citep{geb01, mfj01} and NGC 205

127: \citep{val05}.  For those late-type galaxies that do contain a central

128: black hole, shallow central gravitational potential wells may lead to

129: a low efficiency for fueling black hole accretion.  A low-mass black

130: hole ($\lesssim10^6$ \msun), even if radiating at its Eddington rate,

131: cannot produce a very luminous AGN, and such objects can only be

132: readily detected as AGNs if they are fairly nearby.  In optical

133: surveys, dust extinction and blending with circumnuclear star-forming

134: regions can further hinder the detection of low-luminosity active

135: nuclei in late-type spirals.

136: 

137: If the \msigma\ relation continues toward low masses, then a black

138: hole of $10^6$ \msun\ would correspond to a host galaxy velocity

139: dispersion of $\sigmastar\approx60$ \kms\ \citep{tre02}.  Thus, to

140: explore the intermediate-mass regime for black holes, it is

141: particularly interesting to search for AGN host galaxies having

142: $\sigmastar<60$ \kms.  Such objects are known to exist, but they are

143: rare.  Prior to the Sloan Digital Sky Survey (SDSS), there were only

144: two nearby examples of dwarf galaxies known to contain Seyfert 1

145: nuclei: POX 52, a dE galaxy with $\sigmastar=36\pm5$ \kms\

146: \citep{ksb87, bar04}, and NGC 4395, an Sd-type spiral with a central

147: velocity dispersion of $\sigmastar\lesssim30$ \kms\ \citep{fs89,

148: fh03}.  The availability of SDSS has made it possible for the first

149: time to search systematically for more examples of AGNs with low-mass

150: black holes and small central velocity dispersions.  Greene \& Ho

151: (2004; hereinafter GH04) carried out the first such survey using Data

152: Release 1 of the SDSS, identifying 19 Seyfert 1 galaxies as likely

153: candidates for having $\mbh < 10^6$ \msun.  Although host galaxy

154: properties were not considered in the sample selection, the hosts for

155: this sample turned out to be galaxies of relatively low luminosity, on

156: average about 1 mag fainter than \lstar.  \citet{bgh05} measured

157: stellar velocity dispersions for many of the galaxies in this sample

158: and found that they are also low ($\sim35-80$ \kms) and fall close to

159: the local \msigma\ relation extrapolated toward lower masses.

160: 

161: Most previous searches for low-mass AGNs have concentrated on Type 1

162: (broad-lined) objects (Greene \& Ho 2004, 2007; Dong \etal\ 2006)

163: because the broad-line widths and AGN continuum luminosity can be used

164: to estimate the black hole mass.  However, at low luminosities, most

165: Seyferts are Type 2 objects \citep{hfs97, hao05b}, and a full

166: determination of the demographics of AGNs with low-mass black holes

167: must take into account the Type 2 population.  Furthermore, without

168: the glare of a bright AGN point source, the host galaxies of Seyfert

169: 2s are more easily studied.  This can make it possible to examine the

170: central stellar populations and star formation histories of Seyfert 2

171: host galaxies in a level of detail that would be impossible for Type 1

172: AGNs \citep{kau03agn, hec04}.  Since searches for broad-lined AGNs

173: require both the flux and the width of a broad emission line (such as

174: \hal) to be above some detection threshold, surveys for Type 2 AGNs

175: based on measurements of narrow emission lines can potentially probe

176: lower AGN luminosities, and therefore may be sensitive to AGNs in

177: smaller host galaxies, or containing black holes of lower mass than

178: those that can be found in Seyfert 1 surveys.  Optical spectroscopic

179: surveys such as the Palomar survey \citep{hfs97} have previously

180: identified some examples of very low-luminosity Seyfert 2 nuclei in

181: late-type galaxies, such as the Sc galaxy NGC 1058, which has a

182: central velocity dispersion of only $31\pm6$ \kms\ \citep{bhs02}.

183: Other strategies for finding obscured AGNs, such as mid-infrared

184: searches for high-ionization coronal line emission from obscured

185: active nuclei \citep{sat07}, have contributed further evidence for

186: black holes in late-type spirals.  Since stellar-dynamical searches

187: for low-mass black holes are limited to very nearby galaxies (within a

188: few Mpc at best), AGN surveys remain the best way to examine the

189: demographics of these objects, and to determine the properties of the

190: host galaxies in which low-mass black holes form and grow.

191: 

192: This paper describes new observations of a set of Seyfert 2 galaxies

193: selected from Data Release 2 of the Sloan Digital Sky Survey

194: \citep[SDSS DR2]{aba04} to have host galaxies with relatively low

195: luminosities and small central velocity dispersions.  We select

196: galaxies spectroscopically identified as having Seyfert 2 nuclei

197: (rather than LINER or AGN/starburst transition or composite types), in

198: order to focus on objects having nuclear emission-line spectra that

199: are clearly and unambiguously dominated by an accretion-powered AGN.

200: These galaxies are in many respects Type 2 counterparts of the

201: low-mass Seyfert 1 sample found by GH04, although the two samples are

202: selected by very different criteria.  Most of the objects in both of

203: these samples are not small or faint enough to be considered dwarf

204: galaxies, but the objects described here are sub-\lstar\ galaxies that

205: occupy ranges in mass, luminosity, and central velocity dispersion

206: where very few Type 2 AGNs were known before SDSS.

207: 

208: We present measurements of the properties of these galaxies from

209: high-resolution spectra taken at the Keck Observatory.  The

210: small-aperture, high-resolution Keck spectra make it possible to

211: confirm the AGN classification of these galaxies, to detect weak

212: emission-line features and search for faint broad-line components, to

213: measure stellar velocity dispersions, and to fully resolve the

214: profiles of the narrow emission lines and measure accurate linewidths.

215: We discuss the properties of these galaxies and the Seyfert 1s from

216: the GH04 sample in the context of unified models of AGNs.  We

217: also describe a search for polarized broad-line and continuum emission

218: in four of these low-mass Seyfert 2 galaxies.  In this paper,

219: distance-dependent quantities are calculated assuming $H_0 = 72$ km

220: s\per\ Mpc\per, $\Omega_m$ = 0.3, and $\Omega_\Lambda = 0.7$.

221: 

222: 

223: \section{Sample Selection}

224: 

225: Our goal for this project was to identify Seyfert 2 galaxies having

226: low velocity dispersions, as Type 2 counterparts to the Seyfert 1

227: galaxies identified by GH04.  We initially searched the SDSS DR2

228: archives for narrow emission-line galaxies at $z<0.1$ with

229: emission-line ratios matching a Seyfert 2 classification (as described

230: below), and host galaxies with absolute magnitude fainter than $M_g =

231: -20$ mag.  In the GH04 Seyfert 1 sample, the host galaxy absolute

232: magnitudes ranged from $M_g = -17.8$ to $-20.6$ mag with a median of

233: $-18.9$ mag.  

234: 

235: The line-ratio selection was based on two criteria:

236: \begin{displaymath}

237: \log([\mathrm{O~III}]/\mathrm{H}\beta) > 0.61 /

238:     \{\log([\mathrm{N~II}]/\mathrm{H}\alpha) - 0.47\} + 1.19,

239: \end{displaymath}

240: and

241: \begin{displaymath}

242: [\mathrm{O~III}] / \mathrm{H}\beta >3,

243: \end{displaymath}

244: where [\ion{O}{3}] and [\ion{N}{2}] refer to the fluxes of the

245: [\ion{O}{3}] $\lambda5007$ and [\ion{N}{2}] $\lambda6583$ emission

246: lines, respectively.  The first criterion represents the ``maximum

247: starburst line'' of \citet{kew01}, which separates the \ion{H}{2}

248: region and AGN branches of a \citet{bpt81} ``BPT'' diagnostic diagram.

249: The location of this maximum starburst line was slightly modified in

250: later work by \citet{kau03agn} and \citet{kew06}, but our sample is

251: not affected by this change.  The second criterion was used to

252: identify high-excitation Seyfert galaxies and exclude LINERs or

253: objects with LINER/\ion{H}{2} transition-type spectra, following

254: traditional classification criteria \citep{hfs97}.

255: 

256: After we had begun some initial Keck observations, the SDSS DR2 galaxy

257: catalogs of G. Kauffmann and collaborators (the MPA/JHU catalogs),

258: based on work described by \citet{kau03sfr, kau03agn}, became

259: available to the public.\footnote{Updated versions of the MPA/JHU

260: catalogs described by Kauffmann \etal\ are available at

261: http://www.mpa-garching.mpg.de/SDSS~.}  The MPA/JHU AGN catalogs

262: include narrow emission-line galaxies falling above and to the right

263: of the maximum starburst line in the [\ion{O}{3}]/\hbeta\ vs.

264: [\ion{N}{2}]/\hal\ diagram, so their AGN samples contain both Seyferts

265: and LINERs.  For a description of their AGN sample selection and

266: classification criteria, see \citet{kau03agn}, which was based on the

267: SDSS DR1 sample.  We then used these catalogs to select further

268: candidates, and most of the objects we had initially selected from the

269: SDSS archives were included in the MPA/JHU DR2 AGN sample.

270: 

271: The MPA/JHU catalogs are an excellent resource for statistical studies

272: of nearby narrow-line AGN populations, because of the very large

273: sample size (33,589 AGNs in DR2 and 88,178 in the later DR4 catalog)

274: and the inclusion of additional data on the host galaxies including

275: stellar masses, stellar velocity dispersions, and the absorption-line

276: indices \hdeltaa\ and \dn, which are sensitive to the galaxy's star

277: formation history \citep[see][for details]{kau03sfr,kau03agn}.  To

278: ensure that our sample contained galaxies with well-determined AGN

279: classifications, we further selected only galaxies having measurements

280: of [\ion{O}{3}] flux with S/N$>10$ based on the DR2 emission line

281: catalog.  Several of the objects in our sample have also been

282: previously identified as Seyferts based on SDSS spectra by

283: \citet{hao05a}, but a literature search revealed that only two

284: galaxies in our sample (1440+0247 and 1032+6502) were identified as

285: AGNs prior to the SDSS.  A few cases of galaxies with possible broad

286: \hal\ emission lines, that otherwise met our selection criteria, were

287: included in the sample.  As described below, we used the Keck spectra

288: to test the reality of the broad-line emission.

289: 

290: A total of 29 galaxies meeting these selection criteria were observed

291: at Keck.  The sample properties are described in Table \ref{sample},

292: and Figure \ref{images} displays SDSS images of each object.  The

293: redshifts range from $z=0.0056$ to 0.0712, with a median of 0.042.

294: Since the size of this Keck sample was determined by the available

295: telescope time and the practical need to observe objects distributed

296: over a range of right ascension, the set of objects in this study is

297: not a statistically complete sample, but they should be generally

298: representative of the small population of Seyfert 2 nuclei in

299: sub-\lstar\ host galaxies in SDSS.  In Table \ref{sample}, the

300: absolute $g$-band magnitudes are $K$-corrected to $z=0$ using

301: M. Blanton's \textit{kcorrect} IDL code \citep[version

302: 4\_1\_4;][]{br07} and corrected for Galactic extinction using the dust

303: maps of \citet{sfd98}.

304: 

305: We note that the absolute magnitudes listed in Table \ref{sample} are

306: based on the SDSS Petrosian magnitudes \citep[see][]{str02}.  For most

307: objects in this sample, the Petrosian aperture magnitudes agreed to

308: within $\sim0.1$ mag with the ``model'' magnitudes based on the

309: best-fitting de Vaucouleurs or exponential model.  The magnitudes are

310: not corrected for internal extinction, since for most objects in the

311: sample there is insufficient morphological information to make such a

312: correction.  

313: 

314: For the objects selected from the MPA/JHU DR2 catalog, the host galaxy

315: stellar masses ranged from $\log(\mstar/\msun) = 8.1$ to 10.5, with a

316: median value of 9.8.  Although the MPA/JHU DR2 and DR4 AGN catalogs

317: contain a substantial number of AGN candidates with extremely low

318: stellar masses (i.e., below $10^9$ \msun), we found that the vast

319: majority of these objects do not satisfy the selection criteria given

320: above.  The MPA/JHU DR2 AGN catalog contained 90 objects with

321: $\log(\mstar/\msun) < 9.5$ and 17 with $\log(\mstar/\msun)<9.0$.

322: Inspection of the SDSS spectra of these objects revealed that most of

323: them have LINER or transition-type spectra, or very weak emission

324: lines with such low equivalent width that the AGN classification is

325: somewhat ambiguous.  There were also a few cases of nearby giant

326: galaxies (including NGC 5775 and NGC 3166) being listed in the AGN

327: catalog with very low host galaxy masses of $\log(M_\star/\msun) < 9$.

328: These appear to be objects for which the SDSS catalog magnitudes

329: include only the galaxy's nucleus, resulting in a severe underestimate

330: of the total stellar mass.  The situation for the DR4 AGN catalog is

331: similar: at $\log(M_\star/\msun) < 9$, there are 52 objects in this

332: catalog, most of which are either LINER or transition-type objects,

333: objects with emission lines of very low equivalent width, or large

334: nearby galaxies with incorrect total magnitudes.  Only a small

335: fraction of the objects in this mass range ($\lesssim10\%$) would

336: satisfy our Seyfert selection criteria.  Thus, the population

337: statistics of AGNs with extremely low-mass host galaxies in these

338: catalogs should be treated with some caution.  

339: 

340: 

341: 

342: \section{Observations}

343: \label{sectionobservations}

344: 

345: \subsection{ESI Spectroscopy}

346: 

347: The spectroscopic observations described here were obtained with the

348: ESI spectrograph \citep{she02} on the Keck-II telescope.  Most of the

349: this sample was observed during 2005 May 16--17, with a few objects

350: observed during runs on 2003 November 23--24 and 2004 October 9--10.

351: We used the ESI echelle mode with a 0\farcs75-wide slit, resulting in

352: an instrumental dispersion of $\sigma_i \approx 22$ \kms\ and a total

353: wavelength coverage of 3850--11000 \AA.  The pixel scale in the

354: dispersion direction is 11.5 \kms\ pix\per.  The slit was held at a

355: fixed position angle during each observation, corresponding

356: approximately to the parallactic angle for the midpoint of the

357: exposure.  Exposure times for the galaxies ranged from 900 to 3600 s

358: (Table \ref{esidata}).  Seeing was mostly in the range

359: 0\farcs7--1\farcs0.  One or more flux standards were observed during

360: each night, and several velocity template stars of spectral type

361: ranging from G8III to K4III, and also a few A0V stars, were observed

362: during twilight.  For wavelength calibration, exposures of HgNe, Xe,

363: and CuAr comparison lamp spectra were taken.

364: 

365: The spectra were bias-subtracted, flattened, and extracted with a

366: fixed extraction width of 1\arcsec.  Standard (not optimally weighted)

367: extractions were used, because optimal extraction routines tend to

368: truncate the peaks of emission-line profiles when the emission-line

369: regions have a different spatial extent from the stellar continuum.

370: The extracted spectra for the individual echelle orders were

371: wavelength-calibrated and flux-calibrated, and corrected for telluric

372: absorption by dividing by a normalized spectrum of a white dwarf star

373: observed on the same night.  Error spectra were also extracted and

374: propagated through the full sequence of calibrations.  Finally, the 10

375: echelle orders were combined into a single spectrum for each galaxy,

376: using a weighted average to combine the overlap regions between

377: orders.  Systematic offsets between the flux scales of adjacent orders

378: within the overlap regions were typically smaller than 2.5\%.  The

379: combined spectra were binned to a logarithmic wavelength scale with a

380: binsize of $\Delta[\log(\lambda/$\AA$)] = 1.665\times10^{-5}$.  Figure

381: \ref{spectra} displays the central portion of the ESI spectra.

382: 

383: \subsection{Spectropolarimetry}

384: 

385: We obtained spectropolarimetry observations of four objects using the

386: LRIS polarimeter \citep[LRISp;][]{gcp95} on the Keck I telescope on

387: the night of 2005 December 30, in clear conditions with 0\farcs7

388: seeing.  The instrumental setup consisted of a 400 lines mm\per\ grism

389: on the blue side of the spectrograph, covering 3200--5700 \AA\ at 1.09

390: \AA\ pixel\per, and a 600 lines mm\per\ grating on the red side,

391: covering 5500--8000 \AA\ at 1.28 \AA\ pixel\per.  A 1\arcsec-wide slit

392: was used, and the slit was oriented along the parallactic angle for

393: the midpoint of the exposure sequence.  Observations of each object

394: consisted of a sequence of four exposures with the half-wave plate

395: oriented at 0, 45, 22.5, and 67.5 degrees.  For three galaxies

396: (0119+0037, 0214-0016, and 0947+5349), one exposure set was taken, for

397: a total on-source exposure of one hour, and for the galaxy 1109+6123

398: we obtained two exposure sequences for a total of two hours of

399: integration.  Calibrations were performed using observations of

400: polarized and unpolarized standards \citep{mf70,ct90,tur90}, and flux

401: standard stars observed during twilight.  Data reduction and

402: polarization analysis followed the methods outlined by \citet{mrg88},

403: using a 3\arcsec-wide spectral extraction.  The null standard star HD

404: 57702 was found to be unpolarized to within $0.08\%$ in both the blue

405: and red side extractions, in close agreement with other recent

406: measurements done with LRISp \citep{leo02}.

407: 

408: \section{Results and Discussion}

409: 

410: \subsection{Emission-Line Diagnostics}

411: 

412: Before measurement of the emission lines from the ESI spectra, we

413: applied a simple starlight subtraction procedure to remove the stellar

414: continuum from each spectrum. The fitting routine used a linear

415: combination of two stellar templates, a K giant and an A0V star, from

416: the library of stellar spectra observed on the same nights as the

417: galaxies.  A featureless, power-law continuum component was added, and

418: the continuum was velocity-broadened by convolution with a Gaussian

419: kernel and fitted to the galaxy spectrum using a Levenberg-Marquardt

420: minimization routine.  The free parameters in the fit included the

421: normalization of the two stellar spectra and the featureless

422: continuum, the stellar velocity dispersion, the power-law index of the

423: featureless continuum, and the reddening (assuming a Galactic

424: extinction law).  Spectral regions containing emission lines were

425: masked out in the fit.  This procedure was repeated for each galaxy

426: using several different late-type giant stars and A0V stars to find

427: the best-fitting combination.  While the starlight-subtracted spectra

428: of some galaxies still contained broad residual features or small

429: jumps at the boundaries between echelle orders, this procedure

430: generally yielded a good fit to the continuum shape and the strong

431: stellar features over the region $\sim4000-7000$ \AA, as illustrated

432: in Figure \ref{starsub}.  Although these fits give an estimate of the

433: stellar velocity dispersion, better measurements of the velocity

434: dispersions are determined from fits to smaller spectral windows in

435: individual echelle orders containing strong absorption-line features

436: (as described below), and we used these global fits primarily for

437: purposes of starlight subtraction to obtain the emission-line spectra.

438: 

439: The fluxes of the \hal, \hbeta, [\ion{O}{3}] $\lambda5007$,

440: [\ion{O}{1}] $\lambda6300$, [\ion{N}{2}] $\lambda\lambda6548, 6583$,

441: and [\ion{S}{2}] $\lambda\lambda6716, 6731$ emission lines were

442: measured by Gaussian fitting to the line profiles in the

443: starlight-subtracted spectra.  Each line was first fit with a single

444: Gaussian component, and if there were significant systematic

445: residuals, a second Gaussian component was added.  For the

446: [\ion{N}{2}] and [\ion{S}{2}] doublets, the wavelength separations

447: between each component were fixed to their laboratory values, and the

448: velocity widths of the corresponding components for each line were

449: held to be equal.  The flux ratio of the [\ion{N}{2}] $\lambda6583$

450: and 6548 lines was set to 2.96:1.

451: 

452: From inspection of the SDSS spectra of these galaxies, four objects

453: appeared to be likely or possible candidates for having broad \hal\

454: emission, and we used a single Gaussian component to fit the broad

455: line (Figure \ref{broadhal}).  Two of the broad \hal\ candidates were

456: the two objects that were not selected from the MPA/JHU catalog:

457: 1032+6502 and 1440+0247.  In the case of 1032+6502, inspection of the

458: starlight-subtracted ESI spectrum clearly confirms the reality of the

459: broad \hal\ feature, and the broad component in the ESI data has FWHM

460: = $3020\pm33$ \kms.  The result of the model fit for 1440+0247 is

461: somewhat more ambiguous.  A fit using a double-Gaussian narrow-line

462: model left very weak residuals in the wings of \hal\, but the

463: residuals only extended in wavelength as far as the gaps between \hal\

464: and the [\ion{N}{2}] lines on either side.  Adding a broad \hal\

465: component marginally improved the quality of the fit, and the

466: best-fitting model had a broad component width of FWHM = $1040\pm20$

467: \kms.  Given the narrowness and low amplitude of this possible broad

468: component, it is not clear whether this represents emission from a

469: distinct broad-line region (BLR), and we consider it a tentative

470: detection at most.

471: 

472: The other two broad \hal\ candidates were selected from the MPA/JHU

473: catalog.  The first, 0110+0026 has an obvious and strong broad \hal\

474: emission line that is clearly visible in the SDSS spectrum and

475: confirmed in the ESI data, with FWHM = $5170$ \kms.  The other

476: candidate is 1629+4254; for this object both the SDSS and ESI spectra

477: show broad bases to all of the narrow emission lines.  Fitting the

478: lines with a narrow core and broad redshifted base for each narrow

479: line, we were able to reproduce the overall shape of the

480: \hal+[\ion{N}{2}] blend adequately.  Although the double-Gaussian

481: model does not perfectly reproduce the narrow-line profiles, the broad

482: bases are present on all of the forbidden lines and there does not

483: appear to be any need to include a separate broad \hal\ component in

484: the fit.  For the remaining objects, no broad-line component was

485: required to fit the \hal+[\ion{N}{2}] spectral region in the ESI

486: spectra.

487: 

488: Figure \ref{bptdiagrams} shows the location of the galaxies in this

489: sample (measured from the Keck ESI spectra) on line-ratio diagnostic

490: diagrams, as well as the GH04 Seyfert 1 sample and the nearby AGNs NGC

491: 4395 and POX 52.  The AGN classifications are essentially unchanged

492: between the SDSS 3\arcsec-diameter fiber measurements and the Keck

493: $0\farcs75\times1\arcsec$ aperture extractions.  In the smaller Keck

494: aperture, only one object falls just slightly below our

495: [\ion{O}{3}]/\hbeta\ selection threshold.

496: 

497: In the diagram of [\ion{O}{3}]/\hbeta\ vs.\ [\ion{N}{2}]/\hal, the

498: galaxies in this sample, as well as the GH04 sample, deviate

499: systematically away from the main locus of SDSS Seyfert 2 galaxies,

500: toward lower values of [\ion{N}{2}]/\hal.  NGC 4395 and POX 52 also

501: lie in this same sparsely populated region of the diagram, around

502: log([\ion{N}{2}]/\hal)$\approx -0.5$ and

503: log([\ion{O}{3}]/\hbeta)$\approx1$.  Photoionization modeling by

504: \citet{kra99} and \citet{gro06} has shown that this region of the

505: diagnostic diagram is occupied by Seyferts having lower narrow-line

506: region (NLR) metallicity than in typical Seyfert nuclei; for

507: high-excitation AGNs, decreasing the metallicity moves the line ratios

508: systematically leftward on the diagram.  This region of the diagram is

509: nearly empty, relative to the main Seyfert branch, as a result of the

510: low incidence of active nuclei in low-metallicity galaxies.  In the

511: models of \citet{gro06}, the majority of Seyferts have NLR metallicity

512: between $\sim2$ and 4$Z_\odot$, while NGC 4395 and POX 52, and the

513: more extreme objects in our sample and the GH04 sample, fall in the

514: region of the diagram corresponding to $Z \approx Z_\odot$.  At still

515: lower metallicities, below $0.5Z_\odot$, the AGN and \ion{H}{2}

516: sequences overlap, and additional diagnostics would be needed to

517: distinguish AGN from star-forming nuclei.  However, we note that the

518: multicomponent photoionization models for NGC 4395 described by

519: \citet{kra99} found a good match to the narrow-line spectrum for a

520: metallicity of $0.5Z_\odot$ overall and an N/H ratio of 1/6 solar.

521: Regardless of the exact value of the metal abundances in these

522: galaxies, the work of \citet{gro06} clearly demonstrates that AGN

523: selection based on low host galaxy mass efficiently identifies AGNs

524: with lower than average metallicity, and our sample follows this

525: trend.  In our sample, the object that deviates most from the normal

526: Seyfert branch in the BPT diagram is 1109+6123, which is also the

527: faintest and least massive galaxy in the sample.

528: 

529: We searched for high-ionization coronal line emission in the spectra,

530: as additional indicators of AGN activity.  The galaxy 1440+0247 shows

531: a range of coronal lines including [\ion{Fe}{6}] $\lambda5177$,

532: [\ion{Fe}{7}] ($\lambda5158$, 5720, and 6087 \AA), [\ion{Fe}{10}]

533: $\lambda6374$, and [\ion{Fe}{11}] $\lambda7802$ (Figure

534: \ref{1440lines}).  Its spectrum is extremely similar to that of NGC

535: 4395 \citep{fs89, kra99} aside from the relative weakness of the broad

536: \hal\ emission.  The Type 1 object 1032+6502 also has weak

537: [\ion{Fe}{7}] and [\ion{Fe}{10}] emission.  Among the remaining Type 2

538: objects in the sample, these high-excitation Fe features are seen in

539: only one object, 0214$-$0016.  The LRISp spectra extend farther to the

540: blue than the ESI data, and three of the four LRISp targets

541: ($0214-0016$, 0947+5349, and 1109+6123) have [\ion{Ne}{5}]

542: $\lambda3426$ emission in their spectra.

543: 

544: Since the individual ESI spectra have rather low S/N at the blue and

545: red ends, we created a composite Seyfert 2 spectrum by summing the

546: individual spectra (excluding the three objects with definite or

547: possible broad \hal\ emission), corrected for Galactic extinction and

548: weighted by the S/N of each spectrum.  Figure \ref{composite} shows

549: the composite spectrum, with the starlight subtraction procedure

550: applied as described above.  The best-fitting K+A spectral model for

551: the composite spectrum shows deep Balmer absorption, and reveals

552: emission lines that are difficult to detect clearly in many of the

553: noisy individual spectra, including [\ion{Ne}{3}]

554: $\lambda\lambda3868,3967$, [\ion{Fe}{7}] $\lambda6086$, and the

555: [\ion{S}{3}] $\lambda\lambda$9069, 9532 lines at the red end.  We

556: measured emission line fluxes from the composite spectrum; when

557: necessary, in regions where the starlight subtraction left significant

558: residuals a spline was first fitted to the local continuum and

559: subtracted in order to flatten the continuum.  Emission-line fluxes

560: measured from the composite spectrum are given in Table

561: \ref{compositelines}.

562: 

563: 

564: \subsection{[\ion{O}{3}] Profiles}

565: \label{forbiddenlines}

566: 

567: Following methods similar to those described by \citet{gh05}, we

568: measured the linewidths of strong emission lines from fits of single

569: or double-Gaussian models to each line.  We focus on the [\ion{O}{3}]

570: $\lambda5007$ emission line since this is the strongest forbidden line

571: in the spectra and it is uncontaminated by underlying stellar

572: absorption or blending with other emission lines.  For three galaxies

573: in the sample, the [\ion{O}{3}] profiles were adequately modeled by

574: single Gaussian components with no strong systematic residuals.  In

575: all other cases, two Gaussian components were used to model each

576: narrow emission line.  This type of line-profile decomposition is

577: often interpreted in terms of a core component dominated by virial

578: motions and a wind-dominated wing component \citep[e.g.,][]{gh05},

579: although some objects in our sample have more complex profiles that do

580: not fit this simple qualitative description.  Figure \ref{o3} shows

581: the [\ion{O}{3}] $\lambda5007$ profiles for the sample.

582: 

583: The [\ion{O}{3}] profiles exhibit a greater degree of diversity than

584: seen in previous SDSS samples.  It has long been known that blueward

585: asymmetries on [\ion{O}{3}] profiles are much more common than redward

586: asymmetries \citep[e.g.,][]{om86}.  In the large sample of 1749 SDSS

587: Seyfert 2 spectra studied by \citet{gh05}, most of the narrow-line

588: profiles were well described by a narrow core component plus a

589: low-amplitude, blueshifted ``wing'' component.  The wing component was

590: found to be redshifted relative to the core in only 6\% of the

591: objects.  In contrast, our sample seems to have an unusually large

592: fraction of objects with redward asymmetries.  Eight objects, or 28\%

593: of the sample, have red-asymmetric profiles, in the sense of having a

594: single dominant peak with an excess of flux on the red side of the

595: peak relative to the blue side.  Some other objects (such as 1112+5529

596: and 1432+0046) have profiles that are well fit by the double-Gaussian

597: model but for which the model components cannot be individually

598: labeled as core or wing components, because the broader component

599: contains the majority of the line flux and the velocity difference

600: between the two components is small.  Either blueward or redward

601: asymmetries can be interpreted as resulting from radial motion (inflow

602: or outflow) of clouds combined with the presence of dust, either

603: within the clouds themselves or in a surrounding medium.  In the more

604: common situation of blueward asymmetries, a likely explanation is

605: outflowing motion of clouds in a dusty medium, in which the redshifted

606: clouds on the far side of the galaxy nucleus are behind a larger

607: absorbing column of dust \citep[e.g.,][]{heck81}.  However, there are

608: possible geometric arrangements of an outflow and dust screen that

609: would lead to a redward asymmetry as well, if the outflow direction is

610: not restricted to be perpendicular to the plane of the dust screen.

611: Alternatively, if the dust is contained within the NLR clouds

612: themselves, then the emission from a cloud would be strongest in the

613: direction facing the central ionizing source, and in an outflow we

614: would have a more direct view of the ionized faces of the redshifted

615: clouds on the far side of the nucleus, rather than the ionized faces

616: of the blueshifted clouds, leading to a redward asymmetry on the line

617: profiles.  As discussed by \citet{whit85}, the line profiles do not

618: contain sufficient information to uniquely determine the kinematic

619: state of the NLR or the distribution of dust within it.

620:  

621: More surprising is the detection of double-peaked forbidden-line

622: profiles in three objects.  Two galaxies, 0343$-$0735 and 1006+4456,

623: have fairly symmetric double-peaked [\ion{O}{3}] profiles with

624: well-resolved peaks.  The peak-to-peak velocity separation of the two

625: components in these objects is 115 and 275 \kms, respectively.  The

626: most spectacular example is 1629+4254, which contains a narrow core

627: component having FWHM = 78 \kms\ and a very broad, redshifted wing

628: component with a peak velocity separation of 223 \kms\ and FWHM = 589

629: \kms.  Additionally, a marginally resolved second peak is visible on

630: the red shoulder of the [\ion{O}{3}] line profile in 0914+0238.  For

631: comparison, the \citet{gh05} study found double-peaked profiles in

632: only about 1\% of SDSS Seyfert 2 spectra.  The relatively high

633: incidence of double-peaked profiles in this sample likely results from

634: a combination of the high spectral resolution of ESI, the smaller

635: spectroscopic aperture of the Keck observations, and the low intrinsic

636: linewidths for the individual velocity components in these low-mass

637: galaxies.  In the SDSS spectra, velocity splittings much smaller then

638: the instrumental FWHM of $\sim170$ \kms\ would be unresolved, and only

639: one of these four objects (1006+4456) exhibits double-peaked lines in

640: the lower-resolution SDSS data.  The SDSS spectrum for 1629+4254 shows

641: a single-peaked but strongly red-asymmetric [\ion{O}{3}] profile.

642: 

643: Double-peaked narrow emission lines can arise from a disklike NLR,

644: from outflows, or possibly from a binary AGN with two distinct NLRs

645: \citep[e.g.,][]{zhou04}.  While a fast outflow is the most likely

646: explanation for the extremely broad and asymmetric profile seen in

647: 1629+4254, the galaxies $0343-0735$ and 1006+4456 have more symmetric

648: double-peaked [\ion{O}{3}] profiles suggestive of a disklike emitting

649: region.  The possibility of a disklike NLR would be difficult to test

650: directly by spatially resolved kinematics or morphology, however,

651: since the diameter of such a disk would be too small to resolve

652: directly in these galaxies.

653: 

654: 

655: The linewidths for [\ion{O}{3}] were characterized by two different

656: quantities: the line dispersion (that is, the second moment of the

657: line profile) and the full-width at half-maximum.  Given the limited

658: S/N in the profile wings in some objects, we chose to measure the

659: linewidth parameters from the fitted model profile rather than from

660: the data itself, following methods similar to those used by

661: \citet{gh05}.  The line dispersion is given by

662: \begin{displaymath}

663: \sigma_\mathrm{line}^2 = \left(\frac{c}{\lambda_0}\right)^2 

664: \frac{\int{(\lambda - \lambda_0)^2 f_\lambda

665:     d\lambda}}{\int{f_\lambda d\lambda}},

666: \end{displaymath}

667: where $\lambda_0$ is the flux-weighted mean wavelength of the emission

668: line.  The FWHM was measured as the full velocity width at half of the

669: peak flux density of the modeled line profile.  In the case of a

670: Gaussian profile, $\sigma=$FWHM/2.35.  For line profiles having the

671: prototypical core/wing structure, the FWHM is primarily sensitive to

672: the width of the narrow core component, and the ratio of $\sigma$ to

673: $(\mathrm{FWHM}/2.35)$ gives a measure of the relative prominence of

674: the line wings.

675: 

676: The instrumental dispersion was removed by subtracting it in

677: quadrature from the measured line dispersions, according to

678: $\sigma_\mathrm{line}^2 = \sigma_\mathrm{obs}^2 -

679: \sigma_\mathrm{inst}^2$, and for the FWHM values we subtracted

680: ($2.35\sigma_\mathrm{inst}$) in quadrature from the measured values.

681: Uncertainties on FWHM and $\sigma$ were determined by Monte Carlo

682: simulations.  Starting with the single or double-Gaussian fit

683: parameters for each galaxy, we created 1000 realizations of the line

684: profile, drawing the amplitude, width, and centroid of each Gaussian

685: component randomly from a normal distribution based on the fitting

686: uncertainties from the original profile fit.  Noise was then added to

687: match the S/N of the original spectrum, and the simulated profile was

688: fitted in the same manner as the original spectrum.  The uncertainty

689: on the line dispersion or FWHM was then taken to be the standard

690: deviation of linewidth values measured from the set of simulated

691: profiles.  The results are listed in Table \ref{esidata}.  The FWHM

692: values for the [\ion{S}{2}] lines, measured in the same manner, are

693: also listed.

694: 

695: As expected for galaxies of relatively low mass, the narrow emission

696: lines in these galaxies are much narrower than those of more luminous,

697: classical Seyfert 2 galaxies.  Excluding the three strongly

698: double-peaked objects, the median values of \fwhmo\ and \sigmao\ are

699: 132 and 87 \kms, respectively, and the narrowest-lined object,

700: 1109+6123, has \fwhmo\ = 66 \kms.

701: 

702: 

703: \subsection{Stellar Velocity Dispersions}

704: \label{veldisp}

705: 

706: Stellar velocity dispersions (\sigmastar) were measured from the ESI

707: spectra by direct fitting of velocity-broadened stellar templates,

708: following the method described by \citet{bhs02}. Although the

709: starlight subtraction procedure described above in

710: \S\ref{forbiddenlines} yielded an estimate of the stellar velocity

711: dispersion, the starlight subtraction fits were performed over the

712: entire wavelength range of the data. For accurate measurement of

713: velocity dispersions it is critical to perform the measurements over a

714: restricted wavelength range containing strong stellar absorption

715: features.  It is also critical to mask out wavelengths containing

716: narrow emission lines so that they do not affect the dispersion

717: measurements.  To eliminate any potential issues of mismatch in

718: spectral resolution between echelle orders, we carried out the

719: velocity dispersion measurements using spectra from a single echelle

720: order.

721: 

722: Measurements were performed independently over two wavelength ranges,

723: as illustrated in Figure \ref{dispersions}: a blue region (5130--5470)

724: including the strong Mg$b$ and Fe 5270 lines, and the \ion{Ca}{2}

725: triplet region (8450--8700 \AA).  When visible in the spectra or in

726: the fit residuals, emission lines such as [\ion{N}{1}]

727: $\lambda\lambda5199,5201$ were masked out from the fits.  For the

728: velocity dispersion measurements, the model fits consisted of a

729: template star broadened by a Gaussian in velocity, diluted by addition

730: of a featureless continuum (modeled by a quadratic polynomial in this

731: case), and multiplied by another quadratic polynomial that allows for

732: reddening or systematic flux calibration differences between the

733: galaxy and template star observations.  The velocity dispersion was

734: measured using template stars observed on the same night as each

735: galaxy.  Uncertainties on the velocity dispersions were determined as

736: the sum in quadrature of the formal fitting errors from the

737: best-fitting model and the standard deviation of the velocity

738: dispersions found from all available templates.  The final velocity

739: dispersion values listed in Table \ref{esidata} are the weighted

740: averages of the red and blue measurements.  As expected for galaxies

741: of low stellar mass, the velocity dispersions for this sample are low,

742: ranging from 39 to 94 \kms\ with a median value of 63 \kms, and with

743: 12 objects having $\sigmastar<60$ \kms.  For two objects, 1053+0410

744: and 1109+6123, the continuum S/N was too low to permit a measurement

745: of $\sigmastar$.

746: 

747: These measurements give the dispersion of the line-of-sight velocity

748: profile for the integrated galaxy light falling in the ESI slit and

749: extraction aperture, and represent a combination of light from bulges

750: or pseudobulges, disks, and possibly central star clusters.

751: Higher-resolution images would be needed in order to determine the

752: relative contributions of starlight from the various structural

753: components.  Using bulge-disk decompositions of the GH04 Seyfert 1

754: sample from \emph{Hubble Space Telescope} (\hst) imaging,

755: \citet{ghb08} found that the measured dispersions from \citet{bgh05}

756: were dominated by bulge light for those systems with detected bulges,

757: but that in some host galaxies without detected bulges the measured

758: dispersion was likely dominated by disk starlight.  In either case,

759: the fact that the GH04 Seyfert 1 galaxies fall close to the \msigma\

760: relation extrapolated to low masses means that the central stellar

761: velocity dispersion measured in a $\sim1\arcsec$ aperture, regardless

762: of its origin, is still reasonably well correlated with black hole

763: mass for these low-mass AGNs.

764: 

765: \subsection{Comparison between Stellar and Gas Linewidths}

766: 

767: With measurements of both stellar velocity dispersions and

768: forbidden-line widths, we can examine whether the correlation between

769: \sigmastar\ and [\ion{O}{3}] linewidth, well established for typical

770: Seyfert 2 galaxies \citep{nw96}, continues to hold for these

771: low-dispersion Seyferts.  This issue has attracted considerable recent

772: attention as a result of the potential utility of the [\ion{O}{3}]

773: linewidth as a substitute for \sigmastar\ in studies of the \msigma\

774: relation for certain subclasses of AGNs for which \sigmastar\ can be

775: difficult or impossible to measure directly

776: \citep{wl01,shi03,gm04,bz04,sal07,byz06}.  For a large sample of

777: nearby Seyferts (mostly of Type 2), \citet{nw96} found that the mean

778: value of $\fwhmo/(2.35\sigmastar)$ was very close to unity, but that

779: for individual objects there was considerable scatter in the

780: relationship between stellar and gas linewidths.  Using SDSS spectra

781: of a large sample of low-redshift Seyfert 2 galaxies, \citet{gh05}

782: confirmed this general trend. Additionally, they found that Seyfert 2

783: galaxies tend to have an ``excess'' emission-line velocity dispersion,

784: over and above the stellar velocity dispersion, that increases as a

785: function of $\lbol/\ledd$, and that the [\ion{O}{3}] ``core''

786: component width (based on a double-Gaussian decomposition) is a better

787: predictor of \sigmastar\ than the dispersion of the full [\ion{O}{3}]

788: profile.  These trends are consistent with a scenario in which the

789: linewidth of the NLR core component is primarily set by virial

790: motion of gas in the host galaxy's bulge, while the increasing excess

791: linewidth with Eddington ratio reflects an increasing prevalence of

792: winds or outflows in highly accreting objects.  Since the SDSS spectra

793: themselves cannot be used to measure stellar or emission-line widths

794: for features with $\sigma \lesssim 70$ \kms, our sample allows us to

795: extend these relationships to galaxies with lower velocity

796: dispersions.

797: 

798: Figure \ref{linewidths} shows the comparison between stellar and

799: [\ion{O}{3}] linewidths, excluding the two galaxies for which

800: \sigmastar\ could not be measured as well as the three objects with

801: strongly double-peaked [\ion{O}{3}] profiles.  We examine both

802: \sigmao, which is sensitive to the profile wings, as well as \fwhmo.

803: In our sample, the distinction between core and wing

804: components is not always obvious from the line-profile fits, and we

805: cannot uniformly select a core component from the double-Gaussian

806: profile fits, but the FWHM values essentially reflect the width of the

807: line core. The measured values of \sigmao\ are systematically higher

808: than \sigmastar, while the values of FWHM([\ion{O}{3}])/2.35 track

809: \sigmastar\ much more closely.  To quantify this comparison, following

810: \citet{nw96} and \citet{gh05}, we measure the quantities $\Delta W =

811: \log(\fwhmo)- \log(2.35\sigmastar)$, and $\Delta\sigma = \log\sigmao -

812: \log \sigmastar$.  For our sample, we find a mean value of

813: $\langle\Delta\sigma\rangle = 0.14\pm0.15$, or alternatively

814: $\langle\sigmao/\sigmastar\rangle = 1.48\pm0.56$.  This is very

815: similar to the excess [\ion{O}{3}] linewidth in higher-luminosity

816: samples: \citet{gh05} found $\langle\sigmao/\sigmastar\rangle =

817: 1.43\pm0.75$ for Seyfert 2 galaxies overall in the MPA/JHU DR2 catalog

818: with $\sigmastar>70$ \kms.  On the other hand, the \fwhmo\ values do

819: not reveal any strong excess linewidth for the core component.  The

820: mean value of $\Delta W$ for our sample is $\langle \Delta W \rangle =

821: -0.05 \pm 0.12$, indicating that on average the [\ion{O}{3}] FWHM

822: values track the stellar velocity dispersion (albeit with substantial

823: scatter), similar to results found by \citet{nw96} for more massive

824: Seyfert galaxies.  The [\ion{S}{2}] lines show a similar trend

825: relative to the stellar velocity dispersions, with $\langle \Delta

826: W_\mathrm{[SII]} \rangle = -0.05 \pm 0.15$.

827: 

828: Overall, these results demonstrate that the relationships between

829: [\ion{O}{3}] linewidth and stellar velocity dispersion continue to

830: hold for Seyfert 2 galaxies with stellar velocity dispersions of

831: $\sim40-80$ \kms, and that the \fwhmo\ values are primarily driven by

832: virial motion in the NLR.  Nevertheless, the excess linewidth seen in

833: the [\ion{O}{3}] line wings, and in the \sigmao\ values, indicates

834: that even for these relatively low-luminosity, low-mass AGNs,

835: nongravitational motion (presumably outflows driven by the central

836: engine) does play a significant role in the NLR kinematics.

837: 

838: \subsection{Polarization}

839: 

840: The spectropolarimetry results for the wavelength region surrounding

841: the \hal\ emission line in the red-side spectra are shown in Figure

842: \ref{polplot}.  Each panel displays the total flux spectrum, the

843: fractional polarization $p$ given as the ``rotated Stokes parameter''

844: \citep{mrg88}, and the Stokes flux, which is given by $p \times

845: f_\lambda$.  The clearest signature of a hidden broad line region

846: would be a broad ``bump'' in the $p$ spectrum at the \hal\ emission

847: line, which results in a Stokes flux spectrum resembling that of a

848: Seyfert 1 galaxy \citep{am85}.  Continuum polarization is another

849: possible sign of a hidden active nucleus, but some care must be taken

850: to distinguish it from polarization by transmission through aligned

851: dust grains, either in the host galaxy or of Galactic origin.

852: 

853: In the continuum near \hal, the S/N in the total flux spectra ranges

854: from $\sim40$ to 160 per pixel for these observations (Table

855: \ref{poldata}).  Since the spectra are strongly dominated by

856: starlight, only relatively strong polarized broad lines ($p \gtrsim

857: 1-2\%$ above the continuum level) could potentially be detected in

858: observations of this quality, and none of these four objects exhibits

859: any obvious polarization feature at \hal.  The significance of any

860: possible features in the Stokes $q$ or $u$ spectra was quantified

861: using a method similar to that described by \citet{bfm99}.  For each

862: object, we fit quadratic polynomial models to the continuum in the

863: red-side $q$ and $u$ spectra, excluding a 150-\AA\ wide region around

864: \hal\ from the fit.  Over this 150-\AA\ wide region centered on \hal,

865: the excess polarization in the $q$ Stokes parameter relative to the

866: continuum polarization is then $\delta_q = \langle q_\mathrm{data} -

867: q_\mathrm{model} \rangle$.  To evaluate the uncertainty on this

868: measurement, we calculated the same quantity over a large number of

869: randomly placed 150-\AA\ wide bins in the continuum regions of the

870: spectrum over $5600-7600$ \AA.  The standard deviation of these

871: residual differences was taken to be the $1\sigma$ uncertainty

872: $\epsilon_q$ for the measurement at \hal.  This generally exceeds the

873: error level from the propagated photon-counting statistics and

874: detector readnoise because of small but unavoidable systematic errors

875: in spectral extractions such as errors in interpolating fractional

876: pixel values at the edges of the extraction apertures for spatially

877: extended objects \citep{bfm99}.  These errors result in broad

878: systematic wiggles in the $p$ spectra, as seen in Figure

879: \ref{polplot}.  The significance of any polarization excess at \hal\

880: relative to the fitted continuum level is then $\Delta_q = \delta_q /

881: \epsilon_q$, and similarly for the $u$ parameter.  The measurements

882: confirm that no significant \hal\ line polarization is detected.  The

883: maximum deviation from the continuum polarization level is only at the

884: $2\sigma$ level for the $u$ Stokes parameter in SDSS 0947+5349.  For

885: all other objects the Stokes parameters at \hal\ have $|\Delta_q| < 1$

886: and $|\Delta_u| < 1$.

887: 

888: The continuum polarization and polarization angle were measured over

889: wavelength bins $4400-5500$ \AA\ on the blue side and $6000-7500$ \AA\

890: on the red side.  All measurements were performed on the Stokes

891: parameter spectra, with the results converted to $p$ and $\theta$ as

892: the final step.  The results are reported in Table \ref{poldata},

893: along with the Galactic reddening $E(B-V)$ for these objects

894: \citep{sfd98} and the maximum expected level of Galactic foreground

895: polarization, $p_\mathrm{max} (\%) = 9.0 \times E(B-V)$ \citep{smf75}.

896: For the null standard HD 57702, we find blue and red-side

897: polarizations of $(0.07\pm0.01)\%$ and $(0.08\pm0.01)\%$,

898: respectively, so any systematic instrumental polarization should be at

899: or below this level.  The galaxies' continuum polarizations are $0.2$

900: to $0.76\%$, significantly larger than that of the null standard.

901: 

902: For three of the objects, the continuum polarization is only

903: marginally higher (by $\lesssim 0.2\%$) than the maximum expected

904: Galactic foreground polarization.  The only object exhibiting a

905: continuum polarization signal significantly above the Galactic

906: foreground level is SDSS 1109+6123, which has $p=(0.76\pm0.10)\%$ and

907: ($0.72\pm0.10$)\% on the red and blue sides, respectively, while the

908: Galactic foreground $p_\mathrm{max}$ along this sightline is only

909: $0.1\%$.  Thus, the detected polarization almost certainly arises

910: within the host galaxy, but the limited S/N and lack of any polarized

911: line features does not permit a clear interpretation of its physical

912: origin.  Electron scattering or dust scattering of a hidden AGN

913: continuum could be responsible for the observed polarization.  If the

914: electron scattering mirror were located interior to the narrow-line

915: region then the narrow emission lines should have a lower polarization

916: than the surrounding continuum.  This is not observed in our data, but

917: would be testable with higher S/N observations.

918: 

919: Polarization by transmission through foreground dust in the host

920: galaxy is an alternative possibility, and is the dominant polarization

921: mechanism in some low-luminosity AGNs with high internal extinction

922: such as NGC 3718 \citep{bfm99}.  For SDSS 1109+6123, the Balmer

923: decrement \hal/\hbeta\ is $2.94\pm0.09$ in the SDSS spectrum and

924: $2.87\pm0.10$ in the LRISp spectrum, consistent with the Case B

925: recombination value of 2.85 and smaller than the typical AGN

926: narrow-line region value of $\sim3.1$ \citep[e.g.,][]{of06}, so the

927: reddening toward the NLR is too low to be determined accurately.  If

928: the dust transmission polarization within the host galaxy follows the

929: Galactic relationship between reddening and maximum polarization,

930: $p_\mathrm{max} = 9.0\times E(B-V)$ \citep{smf75}, then the observed

931: polarization would require $E(B-V) \gtrsim0.08$ mag, much higher than

932: the Galactic reddening of $E(B-V) = 0.011$ mag \citep{sfd98}.  For

933: Case B recombination this reddening would modify the observed Balmer

934: decrement to \hal/\hbeta$\gtrsim3.07$.  Given the uncertainties on the

935: measured Balmer decrements, we cannot fully rule out the possibility

936: that the observed continuum polarization is dominated by dust

937: transmission polarization within the host galaxy.  Scattered light

938: from a hidden AGN remains an intriguing possibility for this object,

939: but very long spectropolarimetry exposures would be needed in order to

940: achieve sufficient S/N to place interesting limits on the presence of

941: polarized broad \hal\ emission.

942: 

943: Nondetections of polarized emission lines or continuum in Seyfert 2

944: galaxies are unfortunately difficult to interpret; possible

945: explanations include lack of a BLR, lack of a suitable scattering

946: mirror, foreground dust absorption covering the host galaxy's nucleus,

947: or contamination by starlight that dilutes the strength of a

948: scattered-light polarization signal.  In starlight-dominated AGNs with

949: hidden BLRs, the broad-line polarization is often $\lesssim1\%$ above

950: the continuum polarization \citep{bfm99,mor00}, and the S/N of these

951: observations is too low to detect such weak polarized line features.

952: 

953: \subsection{Radio and X-ray Counterparts}

954: 

955: We searched for counterparts to these sources in the online source

956: catalogs of the FIRST \citep{bwh95} and NVSS \citep{con98} radio

957: surveys, and the \rosat\ All-Sky Survey \citep{vog99}.  Only two

958: objects were matched with sources in the radio catalogs: 1629+4254 is

959: detected in the FIRST survey with $f_\nu = 1.79$ mJy, and 1032+6502 is

960: an NVSS source with a flux density of 7.6 mJy.

961: 

962: The only two objects in this sample detected in the \rosat\ All-Sky

963: Survey are the two previously known AGNs with weak broad \hal\

964: emission, 1032+6502 and 1440+0247.  In both cases, the association

965: with the \rosat\ source has previously been noted: 1032+6502

966: \citep{bol92,mhh96} with a soft X-ray flux of $6.2\times10^{-13}$ ergs

967: cm\persq\ s\per, and 1440+0247 with a flux of $2.8\times10^{-13}$ ergs

968: cm\persq\ s\per\ in the \rosat\ 0.1--2.4 keV band \citep{and03}.

969: 

970: The lack of additional radio and X-ray detections for other objects in

971: this sample is not surprising.  Among the Type 1 objects found by

972: GH04, only one was detected in deep VLA imaging at 6 cm, and the

973: entire GH04 sample is extremely radio quiet \citep{ghu06}.  Similarly,

974: the GH04 sample is faint in X-rays, with only 30\% of the objects

975: detected in the \rosat\ All-Sky Survey.  If the Type 2 objects in our

976: sample are intrinsically similar objects with hidden broad-line

977: regions that are obscured either by a compact torus or by larger-scale

978: dust lanes in the host galaxy, the soft X-ray flux would likely be

979: strongly attenuated as well.  Deep \emph{XMM-Newton} observations of a

980: few objects in this sample have recently been obtained in order to

981: search for nuclear X-ray sources and to constrain the amount of

982: obscuring material toward the nuclei; the results of these

983: observations will be presented in a forthcoming paper.

984: 

985: \subsection{Host Galaxy Morphologies and Stellar Populations}

986: 

987: The host galaxies of AGNs with low-mass black holes exhibit a

988: surprising range of morphologies.  NGC 4395 is an essentially

989: bulgeless, late-type Sd spiral \citep{fs89}, and serves as the best

990: demonstration that black holes can form in some galaxies that lack a

991: substantial central bulge.  On the other hand, the structure of POX 52

992: is best described as a dwarf elliptical \citep{bar04, tho08}, although

993: with a host galaxy absolute magnitude of $M_V=-17.6$ and S\'ersic

994: index of $\sim4$ it is both more luminous and more centrally

995: concentrated than typical dE galaxies \citep{bj98}.  All of the host

996: galaxies in the GH04 Seyfert 1 sample are more luminous than these two

997: nearby objects, but analysis of \hst\ images shows that they also

998: exhibit a range of morphologies, with about 60\% being disk dominated

999: galaxies that contain a bulge or pseudobulge, and the remainder being

1000: compact galaxies with spheroidal morphologies similar to POX 52

1001: \citep{ghb08}.

1002: 

1003: From examination of the SDSS images for this sample, about half of the

1004: objects appear to have definite or likely disks, while the remaining

1005: objects apear compact without obvious disks.  These may be objects

1006: with POX 52- type host galaxies, but higher-resolution imaging would

1007: be needed to determine the morphological types.  The SDSS

1008: concentration index $C$, defined as the ratio of the radii containing

1009: 90\% and 50\% of the total galaxy light in the $r$ band, gives a rough

1010: quantitative measure of morphology \citep{str01}, as there is a

1011: general separation between early-type, bulge-dominated galaxies with

1012: $C\gtrsim2.6$ and late-type, disk-dominated galaxies with $C<2.6$.

1013: The concentration values for our sample, taken from the MPA/JHU

1014: catalog and SDSS archives, are listed in Table \ref{sample}.  The

1015: median concentration for our sample is 2.59, only slightly lower than

1016: the median concentration for the full AGN sample in the DR2 MPA/JHU

1017: catalog overall, and the distribution of concentration values is

1018: consistent with a mixed population including some elliptical or

1019: spheroidal hosts.  Figure \ref{cmd} shows a

1020: color-magnitude diagram for the sample, using the extinction-corrected

1021: and $k$-corrected SDSS photometric data in the $g$ and $r$ bands, and

1022: compared with standard galaxy colors from \citet{fuku95}.  The hosts

1023: span a wide range in color, corresponding to typical colors ranging

1024: from those of ellipticals to late-type spirals, with all but one

1025: object having $g-r$ between 0.4 and 0.8.  

1026: 

1027: Only one galaxy in this sample, 1109+6123, appears to be a promising

1028: candidate for a very late-type, nearly bulgeless ($\sim$Sd) disk

1029: galaxy similar to NGC 4395 (see $\S$\ref{individualobjects}).  This

1030: late-type disk galaxy is a notable outlier in this sample, both in

1031: terms of its low luminosity ($M_g = -16.8$ mag) and its blue color

1032: ($g-r=0.28$ mag).  Even with SDSS, there are remarkably few known

1033: examples of Seyfert nuclei in very late-type, low-mass disk galaxies.

1034: The small number of NGC 4395 analogs in SDSS is largely a selection

1035: effect: galaxies of this low luminosity are only sampled in the SDSS

1036: spectroscopic survey out to $z\approx0.02-0.03$ due to the apparent

1037: magnitude limit of $r=17.77$ of the spectroscopic survey.  Since SDSS

1038: probes only a very small volume for these NGC 4395-type objects, our

1039: sample is biased toward somewhat more luminous host galaxies and more

1040: luminous AGNs.

1041: 

1042: Some limited information on stellar populations and star formation

1043: history can be obtained from the \dn\ and \hdeltaa\ indices in the

1044: MPA/JHU catalog.  The measurement of these parameters is described in

1045: detail by \citet{kau03sfr}.  The \dn\ index measures the strength of

1046: the 4000 \AA\ break, which increases monotonically with population age

1047: for a single-burst population.  The \hdeltaa\ index is a measure of

1048: the H$\delta$ absorption-line strength, which, for a single-burst

1049: population, is strongest for ages of $\sim0.1-1$ Gyr.

1050: \citet{kau03sfr} showed that population synthesis models for galaxies

1051: incorporating exponentially declining star formation rates and random

1052: bursts of star formation were generally able to track the locus of

1053: SDSS galaxies in a plot of \hdeltaa\ vs.\ \dn.  To illustrate the

1054: range of stellar population ages in our sample, Figure \ref{hostage}

1055: displays the values of \hdeltaa\ and \dn\ from the MPA/JHU catalogs

1056: for our sample, in comparison with the overall sample of all galaxies

1057: in the catalog having $\log(\mstar/\msun)<10$.  The AGN hosts exhibit

1058: a wide range in both stellar indices, but avoid the largest values of

1059: \dn\ (corresponding to the oldest stellar populations).  In this

1060: respect, they follow the same general trends as the overall population

1061: of Seyfert 2 galaxies in SDSS, as shown by \citet{kew06}: the Seyfert

1062: 2s tend to be found in hosts with $\dn$ in the range $\sim1.2-1.8$,

1063: while LINERs are found in systematically older host galaxies with \dn\

1064: up to $\sim2.2$ \citep{kew06}.  

1065: 

1066: 

1067: \subsection{Black Hole Mass and Eddington Ratio}

1068: 

1069: For Seyfert 2 galaxies, there is no direct way to measure the black

1070: hole mass from the optical spectrum of the AGN, and the best available

1071: alternative is to use the stellar velocity dispersion and the \msigma\

1072: relationship to obtain a rough estimate of \mbh.  As shown by

1073: \citet{bgh05}, the low-mass Seyfert 1 galaxies from GH04 lie fairly

1074: close to the extrapolated \msigma\ relation, on average having

1075: slightly higher black hole masses than would be predicted from their

1076: velocity dispersions.  The mean offset between the estimated black

1077: hole masses for these galaxies and the \citet{tre02} \msigma\ relation

1078: is 0.23 dex.  This offset is likely to reflect a flattening in the

1079: slope of the \msigma\ relation at low masses \citep{w06,gh06}, but,

1080: lacking a more detailed direct measurement of the change in the

1081: \msigma\ slope as a function of \sigmastar, we simply assume a uniform

1082: offset.  Since the Seyfert 2 galaxies in this sample fall in the same

1083: range of stellar velocity dispersions as the GH04 Seyfert 1 galaxies,

1084: we use the \sigmastar\ measurements to estimate \mbh, assuming an

1085: offset of 0.23 dex relative to the Tremaine \etal\ \msigma\ relation.

1086: For the objects not having direct measurements of \sigmastar, we use

1087: \fwhmo/2.35 in place of \sigmastar\ to estimate \mbh.  With this

1088: prescription, the black hole mass estimates for this sample are in the

1089: range $4.7 < \log(\mbh/\msun) < 6.8$, with a median value of

1090: $\log(\mbh/\msun) = 6.1$.  The uncertainty in any individual mass

1091: estimate is likely to be at least 0.3 dex, however, due to the scatter

1092: in the \msigma\ relation and the poorly constrained \msigma\ slope at

1093: low masses.

1094: 

1095: Similarly, the bolometric luminosities can only be estimated

1096: indirectly.  The best available indicator of the AGN luminosity in

1097: these objects is the [\ion{O}{3}] $\lambda5007$ luminosity, which we

1098: assume to arise predominantly from the AGN rather than from

1099: star-forming regions, based on the locations of these galaxies in the

1100: BPT diagnostic diagrams.  (As a caveat, we note that for host galaxies

1101: of relatively low metallicity, the contribution from star-forming

1102: regions to the [\ion{O}{3}] luminosity could be greater than in

1103: high-metallicity galaxies, but we have no direct diagnostics of the

1104: star-formation rate in the nuclei.)  From a sample of low-redshift

1105: Seyfert 1 galaxies and quasars, \citet{hec04} found an [\ion{O}{3}]

1106: bolometric correction of $\lbol/L_\mathrm{[OIII]} \approx 3500$, with

1107: a typical scatter of 0.38 dex.  We adopt this prescription for our

1108: sample, but note as an additional caveat that there is evidence that

1109: the $L_\mathrm{[OIII]}/L_X$ ratio decreases systematically with AGN

1110: luminosity \citep{net06}, implying that there is a luminosity

1111: dependence to the [\ion{O}{3}] bolometric correction.  The

1112: [\ion{O}{3}] luminosities for our sample are listed in Table

1113: \ref{sample}; most are in the range $6.0 < \log(\lothree/\lsun) <

1114: 7.5$, implying bolometric luminosities of $43 < \log[\lbol/$(ergs

1115: s\per)]$<44.6$ for most objects.  The two very low-redshift galaxies

1116: with late-type spiral hosts, 1032+6502 and 1109+6123, are much lower

1117: in luminosity, both having $\log(\lothree/\lsun) \approx 4.9$ and

1118: $\lbol \approx 10^{42}$ ergs s\per.

1119: 

1120: Combining the black hole masses and luminosities, we can obtain rough

1121: estimates of \lbol/\ledd.  The distribution of \lbol/\ledd\ estimates

1122: is shown in Figure \ref{lbol}; the median value of $\log(\lbol/\ledd)$

1123: for the sample is $-0.4$.  Although these values of \lbol/\ledd\ will

1124: be extremely uncertain for individual galaxies, the estimates suggest

1125: that many of these galaxies contain black holes that are undergoing a

1126: significant episode of accretion.  \citet{gh07} found that in their

1127: DR4 sample of Seyfert 1 galaxies with $\mbh < 2\times10^6$ \msun, the

1128: distribution of Eddington ratio is peaked at $\log(\lbol/\ledd)

1129: \approx -0.5$, highlighting the similarities between the two samples.

1130: 

1131: Alternatively, assuming the same [\ion{O}{3}] bolometric correction,

1132: we can estimate the minimum black hole mass that would be required to

1133: support the luminosity in each object if all objects are radiating at

1134: $\lbol/\ledd<1$.  Figure \ref{bhmin} shows the resulting lower limits

1135: to \mbh\ as a function of \sigmastar\ (again using \fwhmo/2.35 in

1136: place of \sigmastar\ for the two objects without measured velocity

1137: dispersions).  The minimum required black hole masses for the sample

1138: are generally low, and 2/3 of the galaxies have \mbh(min)$<10^6$

1139: \msun.  The results are consistent with most objects having black hole

1140: masses that fall close to or below the low-mass extrapolation of the

1141: \msigma\ relation from \citet{tre02}; eight galaxies in the sample

1142: have minimum \mbh\ values that lie above the \msigma\ relation.  The

1143: largest deviation above the extrapolated \msigma\ relation is found

1144: for 1440+0247, which lies approximately an order of magnitude higher

1145: than the Tremaine \etal\ \msigma\ relation extrapolated to

1146: $\sigmastar=45$ \kms.  Given the substantial uncertainties in the

1147: [\ion{O}{3}] bolometric correction, these lower limits to \mbh\ should

1148: be viewed cautiously on a case-by-case basis, but on the whole these

1149: results are consistent with expectations for a sample of black holes

1150: of relatively low mass in which many of the black holes are undergoing

1151: a significant accretion episode.

1152: 

1153: \subsection{Unified Models and the Obscuring Torus}

1154: 

1155: These Seyfert 2 galaxies and the GH04 Seyfert 1 objects do not

1156: constitute perfectly matched samples in terms of their nuclear or host

1157: galaxy properties, as a result of the very different selection

1158: criteria that were used for the two samples.  Nevertheless, the two

1159: samples do occupy similar and overlapping ranges in some key

1160: parameters.  As shown in Figure \ref{hostcompare}, the distributions

1161: of $g$-band absolute magnitudes are overlapping but not identical; the

1162: Seyfert 2s were selected based on an absolute magnitude cut and all

1163: have $M_g \geq -19.7$ mag, while the GH04 objects were selected

1164: without regard to host galaxy properties, and their distribution

1165: extends to slightly more luminous galaxies, up to $M_g = -20.6$.  The

1166: GH04 sample also does not contain any host galaxies as faint as the

1167: least luminous Seyfert 2 galaxy in this sample, 1109+6123, with $M_g =

1168: -16.8$.  In terms of [\ion{O}{3}] luminosity, which is the best

1169: available indicator of total AGN luminosity, the two samples occupy

1170: overlapping ranges with most galaxies in the range $\lothree \sim

1171: 10^6-10^8$ \lsun; again, the Type 1 sample extends to slightly higher

1172: luminosities and the Type 2 sample extends to lower luminosities, with

1173: the galaxy 1109+6123 falling at the low-luminosity end of the

1174: distribution.

1175: 

1176: Although the narrow-line ratios of the two samples occupy overlapping

1177: ranges, as seen in Figure \ref{bptdiagrams}, there are some key

1178: differences between the Type 1 and Type 2 objects.  From inspection of

1179: Figure \ref{bptdiagrams}, it is evident that the Seyfert 2 galaxies

1180: from this sample have somewhat lower values of [\ion{O}{3}]/\hbeta\ on

1181: average than the GH04 sample.  the median values of

1182: [\ion{O}{3}]/\hbeta\ are 7.9 and 4.9 for the Type 1 and Type 2 samples

1183: respectively, and a Kolmogorov-Smirnov (K-S) test confirms that the

1184: two distributions are significantly different, with a 1.4\%

1185: probability of being drawn from the same parent population. We also

1186: examine the [\ion{O}{2}]$\lambda3727$/[\ion{O}{3}]$\lambda5007$ ratio

1187: for the two samples, measured from the SDSS spectra since the ESI

1188: spectra do not cover [\ion{O}{2}] $\lambda3727$ for most objects (see

1189: Figure \ref{hostcompare}).  The Type 1 and Type 2 objects have median

1190: values for [\ion{O}{2}]/[\ion{O}{3}] of 0.23 and 0.61, respectively.

1191: A K-S test shows that the two distributions of

1192: [\ion{O}{2}]/[\ion{O}{3}] have only a $3\times10^{-5}$ probability of

1193: being drawn from the same parent population.

1194: 

1195: The higher [\ion{O}{2}]/[\ion{O}{3}] ratio in the Type 2 objects is

1196: intriguing in light of similar trends found recently among

1197: high-luminosity AGNs.  \citet{ho05} showed that among the PG quasar

1198: sample, the [\ion{O}{2}]/[\ion{O}{3}] ratio is generally consistent

1199: with that expected from AGN photoionization, with no evidence for an

1200: additional contribution to the [\ion{O}{2}] emission from star-forming

1201: regions.  Following up on this study, \citet{khi06} found that

1202: luminous Type 2 AGNs from SDSS such as those found by \citet{zak03}

1203: have enhanced levels of [\ion{O}{2}] emission relative to what is seen

1204: in a corresponding Type 1 AGN sample from SDSS, possibly indicating an

1205: elevated rate of star formation in the Type 2 objects compared to the

1206: Type 1 objects.  This observation presents a strong challenge to the

1207: simplest unified models, in which Type 1 and Type 2 AGNs differ only

1208: by orientation and obscuration of the central engine.  The low-mass

1209: Seyfert 2 hosts evidently follow the same trend of enhanced

1210: [\ion{O}{2}] emission: the median values of log([\ion{O}{3}]/\hbeta)

1211: and log([\ion{O}{2}]/[\ion{O}{3}]) for the Type 2 objects in this

1212: sample are 0.69 and $-0.21$, respectively, which fall remarkably close

1213: to the values for these line ratios measured from the composite Type 2

1214: quasar spectrum of \citet{zak03}, as displayed in Figure 7 of

1215: \citet{khi06}.  The lower [\ion{O}{3}]/\hbeta\ ratios among the

1216: Seyfert 2 sample could be another manifestation of a larger

1217: contribution of star-forming regions to the spectra, although another

1218: possibility is that the Seyfert 2 galaxies have slightly lower NLR

1219: ionization parameter than the GH04 Seyfert 1 galaxies.

1220: 

1221: Aside from the question of the relative amounts of star formation in

1222: the Type 1 and Type 2 samples, the more fundamental unification

1223: question remains: do they all have intrinsically similar central

1224: engines, in which case the presence or absence of broad emission lines

1225: would depend primarily on their orientation and the degree of

1226: obscuration along our line of sight?  Or, alternatively, are some of

1227: the Type 2 objects ``true'' Type 2 AGNs that intrinsically lack a BLR?

1228: There is growing observational evidence that at low luminosities,

1229: there is a population of Seyfert 2 galaxies that do not exhibit

1230: polarized broad-line emission or X-ray absorption

1231: \citep[e.g.][]{pap01, tran03, nmm03, bian08}, suggesting that the line

1232: of sight to the nucleus is unobscured and the BLR is not present.  One

1233: scenario to explain true Seyfert 2 nuclei has been proposed by

1234: \citet{nic00}; in this model, the BLR consists of an accretion disk

1235: wind and when \lbol/\ledd\ falls below a critical value of

1236: $\sim10^{-3}$, the luminosity is insufficient for the disk to form a

1237: radiation-pressure dominated region from which to launch the wind.  In

1238: another model proposed by \citet{laor03}, there is a maximum BLR

1239: linewidth of $\sim25,000$ \kms\ above which the BLR clouds do not

1240: survive, and there is a critical luminosity below which the BLR radius

1241: shrinks to a size where the corresponding Keplerian velocities exceed

1242: this maximum linewidth.  At present it is not completely clear whether

1243: \lbol\ or \lbol/\ledd\ is the primary parameter that determines

1244: whether a BLR forms (or whether there may be some additional key

1245: parameter), and further exploration of the properties of the least

1246: luminous Seyfert 2 galaxies is needed in order to resolve this

1247: question.  As described above, most of our sample consists of objects

1248: with relatively high \lbol/\ledd, and with \lbol\ in the range

1249: $10^{43}-10^{44.6}$ ergs s\per; in both the \citet{nic00} and

1250: \citet{laor03} scenarios these objects would be above the critical

1251: luminosity or Eddington ratio needed to support BLRs.

1252: 

1253: A closely related question is the nature of the obscuring torus in

1254: unified models of AGNs and whether the properties or existence of the

1255: torus depends on AGN luminosity or other properties.  One attractive

1256: possibility for the origin of the obscuring torus is that it may be

1257: the dusty component of a wind driven from the surface of the accretion

1258: disk \citep[e.g.,][]{kk94}.  \citet{es06} have recently examined the

1259: dynamics of hydromagnetic disk winds and the dependence of outflow

1260: properties on the AGN luminosity.  A key prediction of their model is

1261: that for AGN luminosities below $\sim10^{42}$ ergs s\per, the AGN is

1262: not powerful enough to sustain the torus wind, and the torus should

1263: therefore not be present in very low-luminosity AGNs.  While most of

1264: our sample is more luminous than this threshold, the least luminous

1265: object in our sample (1109+6123) presents a particularly interesting

1266: test case.  Its narrow-line spectrum and host galaxy are nearly

1267: identical to NGC 4395 (\S\ref{individualobjects}), and it is the best

1268: Type 2 counterpart to NGC 4395 known.  NGC 4395 has a bolometric

1269: luminosity of $\sim5\times10^{40}$ ergs s\per\ \citep{mor99, pet05},

1270: more than an order of magnitude below the typical cutoff luminosity

1271: for a wind-driven torus in the Elitzur \& Shlosman model.  While the

1272: bolometric luminosity of SDSS 1109+6123 is unknown, its [\ion{O}{3}]

1273: luminosity is about 40\% smaller than that of NGC 4395.

1274: 

1275: This galaxy and NGC 4395 together pose an interesting challenge to

1276: models for the obscuring torus: \emph{the Type 1/Type 2 dichotomy

1277: apparently extends down to the very lowest luminosities seen among

1278: nearby Seyfert galaxies}, below $\lbol=10^{41}$ ergs s\per.  While

1279: this is already well established for massive host galaxies

1280: \citep{ho08}, the discovery of a near-perfect Type 2 analog to NGC

1281: 4395 raises the possibility that the obscuring torus might still be

1282: present even at these very low masses and luminosities.  The detection

1283: of continuum polarization in 1109+6123 gives an intriguing hint of an

1284: obscured nucleus, and further searches for the presence of an

1285: obscuring torus and obscured BLR in 1109+6123, including deeper

1286: spectropolarimetry, X-ray observations, and mid-infrared spectroscopy,

1287: may provide critical new tests for wind-driven torus models.

1288: Alternatively, 1109+6123 might be a ``true'' Seyfert 2 that

1289: intrinsically lacks a BLR.  In this situation, the question remains:

1290: why would a BLR occur in NGC 4395, but not in 1109+6123, which has a

1291: similar [\ion{O}{3}] luminosity, and presumably a similar black hole

1292: mass as well?  If 1109+6123 is a true unobscured Seyfert 2, this would

1293: imply that the formation or lack of a BLR must depend on some other

1294: properties of the AGN in addition to \lbol\ or \lbol/\ledd.

1295: 

1296: Obscuration by larger-scale dust lanes within the host galaxies is

1297: likely to play some role as well in determining the relative numbers

1298: of Type 1 and Type 2 AGNs in low-mass disk galaxies.  The SDSS images

1299: show that three objects in our sample are nearly edge-on disk

1300: galaxies, in which host galaxy-scale obscuration could easily explain

1301: the lack of broad emission lines in the optical spectrum.  In

1302: contrast, in the GH04 Seyfert 1 sample, none of the spiral host

1303: galaxies are close to edge-on \citep{ghb08}.  This is consistent with

1304: trends seen in the general Seyfert population; Seyfert 1 nuclei are

1305: deficient in edge-on host galaxies relative to face-on hosts

1306: \citep{keel80,mr95}.  Similarly, in deep X-ray surveys, the difference

1307: between ``optically dull'' X-ray sources and broad-lined AGNs can be

1308: attributed at least in part to host galaxy-scale obscuration related

1309: to host inclination \citep{rig06}.  A search for nuclear dust lanes in

1310: high-resolution images would provide some constraints on the impact of

1311: host-galaxy scale obscuring material \citep[e.g.,][]{mgt98}.

1312: 

1313: \subsection{Notes on Individual Objects}

1314: \label{individualobjects}

1315: 

1316: \emph{0110+0026}: This galaxy was selected from the MPA/JHU SDSS

1317: catalog of narrow-line AGNs, but it has a very prominent broad \hal\

1318: emission line visible in both the SDSS and ESI spectra.

1319: 

1320: \emph{0914+0238}: A possible companion is located about 25\arcsec\ to

1321: the east of this galaxy, but its redshift is unknown (not catalogued

1322: in NED).

1323: 

1324: \emph{0947+5349}: This galaxy appears to be part of an interacting

1325: triplet, labeled in NED as CGCG 265--039.  The redshifts of the two

1326: larger galaxies just south of it in the SDSS image are unknown.

1327: 

1328: \emph{1032+6502}: Also known as NGC 3259, this is a nearby SBbc

1329: galaxy.  It was previously identified as a ROSAT source by

1330: \citet{bol92} and \citet{mhh96}, and as a Seyfert by \citet{hao05a} from

1331: the SDSS spectrum.  Our Keck spectrum confirms the presence of a weak

1332: broad \hal\ emission line.  Multicolor \hst\ images show that it has a

1333: substantial bulge, with $M \approx 3\times10^8$ \msun\ and a

1334: predominantly old stellar population \citep{car07}.  It was also

1335: recently noted as an example of a galaxy containing both a nuclear

1336: star cluster and an AGN by \citet{seth08}.

1337: 

1338: \emph{1109+6123}: Also known as UGC 06192 or MCG +10-16-069, this

1339: galaxy was not previously identified as hosting an AGN prior to SDSS.

1340: It has an unusual high-excitation narrow-line spectrum with the lowest

1341: [\ion{N}{2}]/\hal\ ratio among our sample ([\ion{N}{2}]/\hal$=0.18$),

1342: characteristic of low-metallicity AGNs and very similar to the

1343: spectrum of NGC 4395 \citep{kra99}, but without any hint of broad

1344: emission lines.  Its [\ion{O}{3}] luminosity, from the MPA/JHU

1345: catalog, is $\log(L_\mathrm{[OIII]}/\lsun) = 4.94$.  This is very

1346: similar to the narrow-line luminosity of NGC 4395, which has

1347: $\log(L_\mathrm{[OIII]}/\lsun) = 5.09$ based on the emission-line data

1348: from \citet{kra99} and a distance of 4.3 Mpc \citep{thim04}.  With

1349: $\log(\mstar/\msun) = 8.07$, this galaxy has the lowest stellar mass

1350: of any unambiguous Seyfert 2 galaxy in either the DR2 or DR4 MPA/JHU

1351: AGN catalogs.  From examination of the SDSS images, the host galaxy is

1352: a low surface brightness, late-type spiral, probably best classified

1353: as Scd or Sd.  A surface brightness profile measured from the SDSS

1354: $i$-band image reveals that the host galaxy is well described by an

1355: exponential disk with a scale length of 7\farcs9 ($\approx 1.1$ kpc),

1356: with a slight central excess over the exponential profile in the inner

1357: $\sim5\arcsec$ that may indicate the presence of a small bulge or

1358: pseudobulge.  Overall, the properties of this galaxy make it a

1359: near-twin of NGC 4395, aside from the lack of broad emission lines in

1360: its spectrum.  Deeper and higher-resolution imaging with \hst\ would

1361: be particularly valuable to definitively detect or constrain the

1362: presence of a bulge in this galaxy and to determine whether a compact

1363: central star cluster is present.

1364: 

1365: \emph{1432+0046}: The SDSS image shows an apparent close companion

1366: just to the east of this galaxy, but it is a chance projection; the

1367: disk galaxy is SDSS J143232.24+004617.4, at $z=0.034$.

1368: 

1369: \emph{1440+0247}: Also known as Tol 1437+030, this object was

1370: identified as an AGN by \citet{bfw78}.  It has also been noted as an

1371: AGN from the SDSS spectrum by \citet{and03}, \citet{kni04}, and

1372: \citet{hao05a}.  Its spectrum is very similar to those of NGC 4395 and

1373: POX 52 in terms of the narrow-line ratios as well as the detection of

1374: high-ionization lines (up to [\ion{Fe}{10}] and [\ion{Fe}{11}]), but

1375: broad \hal\ emission is only marginally detectable in the ESI

1376: spectrum.

1377: 

1378: \emph{1629+4254} : At first glance this appears to be a likely Seyfert

1379: 1 galaxy because of the broad base to the \hal+[\ion{N}{2}] emission

1380: blend, but our line profile fits show that the broad bases are present

1381: on all of the forbidden lines.  The [\ion{O}{3}] profile has an

1382: extraordinary broad and redshifted component (FWHM = 589 \kms, $\Delta

1383: v = +223$ \kms\ relative to the core component), while the core

1384: component is fairly narrow (FWHM = $78\pm1$ \kms).  From the FIRST

1385: radio detection, the radio power of this source at 20 cm is

1386: $5\times10^{21}$ W Hz\per.  Multicomponent [\ion{O}{3}] profiles are

1387: often found in objects with linear radio sources \citep{whi88}, and

1388: the detection of a radio counterpart in the FIRST survey makes this an

1389: unusual object relative to the very radio-quiet GH04 sample.  To

1390: compare with the radio properties of the GH04 objects, we follow

1391: \citet{ghu06} and adopt the radio loudness parameter $R =

1392: f_\mathrm{6~cm}/f_\mathrm{4400~\AA}$.  We estimate the flux density of

1393: the (presumably obscured) optical nonstellar continuum by using the

1394: [\ion{O}{3}] flux as a proxy, assuming the median ratio of

1395: [\ion{O}{3}] luminosity to 4400 \AA\ continuum luminosity for the GH04

1396: sample \citep{ghu06}.  This yields $f$(4400 \AA) $\approx 0.6$ mJy.

1397: Based on the FIRST 20 cm flux and an assumed flat spectral index

1398: between 6 and 20 cm, which is typical of low-luminosity Seyfert 2

1399: galaxies \citep{uh01}, we find $R \approx 3$.  This is formally in the

1400: radio-quiet regime (since $R<10$), but this value of $R$ is greater

1401: than most of the upper limits found for the GH04 sample \citep{ghu06}.

1402: 

1403: \section{Summary and Conclusions}

1404: 

1405: We have presented an initial, exploratory study of the properties of

1406: Seyfert 2 galaxies with sub-\lstar\ luminosities selected from the

1407: Sloan Digital Sky Survey.  Our measurements reveal very low central

1408: stellar velocity dispersions (\sigmastar $<60$ \kms) in 12 objects;

1409: these are among the smallest velocity dispersions found in any AGN

1410: host galaxies, and imply that these galaxies contain some of the least

1411: massive black holes in any known AGNs.  We have also identified one of

1412: the very few known examples of a high-excitation Seyfert 2 nucleus in

1413: a late-type, dwarf spiral host galaxy similar to NGC 4395.  The

1414: correlations between [\ion{O}{3}] linewidth and \sigmastar\

1415: established for higher-mass Seyferts continue to hold in this low-mass

1416: regime, while the low-mass Seyfert 2 galaxies exhibit an unusually

1417: high incidence of peculiarities (such as redward asymmetries and

1418: double-peaked profiles) in their forbidden emission lines.  Using

1419: rough estimates of black hole mass and bolometric luminosity, we find

1420: that the median value of $\lbol/\ledd$ for this sample is 0.4,

1421: indicating that many of these are objects in which the black hole is

1422: undergoing a major growth phase.  Future work on these objects will

1423: include \emph{XMM-Newton} observations and \emph{Spitzer} mid-infrared

1424: spectroscopy, in order to search for evidence of nuclear obscuration

1425: and to test whether the obscuring torus of AGN unified models is

1426: present in this low-mass, low-luminosity regime.

1427: 

1428: 

1429: \acknowledgments 

1430: 

1431: We are very grateful to Guinevere Kauffmann and her collaborators for

1432: making their AGN and stellar mass catalogs available to the community.

1433: We thank George Djorgovski for obtaining some preliminary ESI spectra

1434: for this project, Craig Markwardt and Mike Blanton for their excellent

1435: and useful IDL software, Michael Strauss for providing plotting

1436: scripts, and an anonymous referee for suggestions that improved the

1437: manuscript.  Research by A.J.B. was supported by the UC Irvine

1438: Physical Sciences Innovation Fund and by NSF grant AST-0548198.  Data

1439: presented herein were obtained at the W.M. Keck Observatory, which is

1440: operated as a scientific partnership among Caltech, the University of

1441: California, and NASA. The Observatory was made possible by the

1442: generous financial support of the W.M. Keck Foundation.  The authors

1443: wish to recognize and acknowledge the very significant cultural role

1444: and reverence that the summit of Mauna Kea has always had within the

1445: indigenous Hawaiian community. Funding for the SDSS and SDSS-II has

1446: been provided by the Alfred P. Sloan Foundation, the Participating

1447: Institutions, the National Science Foundation, the U.S. Department of

1448: Energy, the National Aeronautics and Space Administration, the

1449: Japanese Monbukagakusho, the Max Planck Society, and the Higher

1450: Education Funding Council for England. The SDSS Web Site is

1451: http://www.sdss.org/.

1452: 

1453: %\emph{Facilities:}  \facility{Keck:II}

1454: 

1455: \begin{thebibliography}{}

1456: 

1457: \bibitem[Abazajian et al.(2004)]{aba04} Abazajian, K., et al.\ 2004,

1458: \aj, 128, 502

1459: 

1460: \bibitem[Anderson et al.(2003)]{and03} Anderson, S.~F., et 

1461: al.\ 2003, \aj, 126, 2209 

1462: 

1463: \bibitem[Antonucci \& Miller(1985)]{am85} Antonucci, R.~R.~J., \&

1464: Miller, J.~S.\ 1985, \apj, 297, 621

1465: 

1466: 

1467: \bibitem[Baldwin et al.(1981)]{bpt81} Baldwin, J.~A., 

1468: Phillips, M.~M., \& Terlevich, R.\ 1981, \pasp, 93, 5 

1469: 

1470: \bibitem[Barth et al.(1999)]{bfm99} Barth, A.~J., Filippenko, 

1471: A.~V., \& Moran, E.~C.\ 1999, \apj, 525, 673 

1472: 

1473: \bibitem[Barth et al.(2005)]{bgh05} Barth, A.~J., Greene, 

1474: J.~E., \& Ho, L.~C.\ 2005, \apjl, 619, L151 

1475: 

1476: \bibitem[Barth \etal(2004)]{bar04} Barth, A.~J., Ho, L.~C., Rutledge,

1477: R.~E., \& Sargent, W.~L.~W.\ 2004, \apj, 607, 90

1478: 

1479: \bibitem[Barth \etal(2002)]{bhs02} Barth, A.~J., 

1480: Ho, L.~C., \& Sargent, W.~L.~W.\ 2002, \aj, 124, 2607 

1481: 

1482: \bibitem[Becker et al.(1995)]{bwh95} Becker, R.~H., White, 

1483: R.~L., \& Helfand, D.~J.\ 1995, \apj, 450, 559 

1484: 

1485: \bibitem[Bian \& Zhao(2004)]{bz04} Bian, W., \& Zhao, Y.\ 

1486: 2004, \mnras, 347, 607 

1487: 

1488: \bibitem[Bian et al.(2006)]{byz06} Bian, W., Yuan, Q., \& 

1489: Zhao, Y.\ 2006, \mnras, 367, 860 

1490: 

1491: \bibitem[Bianchi et al.(2008)]{bian08} Bianchi, S., Corral, 

1492: A., Panessa, F., Barcons, X., Matt, G., Bassani, L., Carrera, F.~J., 

1493: \& Jim{\'e}nez-Bail{\'o}n, E.\ 2008, \mnras, 108 

1494: 

1495: \bibitem[Binggeli \& Jerjen(1998)]{bj98} Binggeli, B., \& 

1496: Jerjen, H.\ 1998, \aap, 333, 17 

1497: 

1498: \bibitem[Blanton \& Roweis(2007)]{br07} Blanton, M.~R., \& Roweis, S.\

1499: 2007, \aj, 133, 734

1500: 

1501: \bibitem[Bohuski et al.(1978)]{bfw78} Bohuski, T.~J., 

1502: Fairall, A.~P., \& Weedman, D.~W.\ 1978, \apj, 221, 776 

1503: 

1504: \bibitem[Boller et al.(1992)]{bol92} Boller, T., Meurs, 

1505: E.~J.~A., Brinkmann, W., Fink, H., Zimmermann, U., \& Adorf, H.-M.\ 1992, 

1506: \aap, 261, 57 

1507: 

1508: \bibitem[Carollo et al.(2007)]{car07} Carollo, C.~M., Scarlata, C.,

1509: Stiavelli, M., Wyse, R.~F.~G., \& Mayer, L.\ 2007, \apj, 658, 960

1510: 

1511: \bibitem[Clemens \& Tapia(1990)]{ct90} Clemens, D.~P., \& 

1512: Tapia, S.\ 1990, \pasp, 102, 179 

1513: 

1514: \bibitem[Condon et al.(1998)]{con98} Condon, J.~J., Cotton, 

1515: W.~D., Greisen, E.~W., Yin, Q.~F., Perley, R.~A., Taylor, G.~B., \& 

1516: Broderick, J.~J.\ 1998, \aj, 115, 1693 

1517: 

1518: \bibitem[Dong et al.(2006)]{dong06} Dong, X., et al.\ 2007, 

1519: \apj, 657, 700 

1520: 

1521: \bibitem[Elitzur \& Shlosman(2006)]{es06} Elitzur, M., \& 

1522: Shlosman, I.\ 2006, \apjl, 648, L101 

1523: 

1524: 

1525: \bibitem[Ferrarese \& Merritt(2000)]{fm00} Ferrarese, L.~\& Merritt,

1526: D.\ 2000, \apjl, 539, L9

1527: 

1528: \bibitem[Ferrarese et al.(2001)]{fer01} Ferrarese, L., Pogge, 

1529: R.~W., Peterson, B.~M., Merritt, D., Wandel, A., \& Joseph, C.~L.\ 2001, 

1530: \apjl, 555, L79 

1531: 

1532: \bibitem[Filippenko \& Ho(2003)]{fh03} Filippenko, A. V., \& Ho,

1533:   L. C. 2003, \apj, 588, L13

1534: 

1535: \bibitem[Filippenko \& Sargent(1989)]{fs89} Filippenko, 

1536: A.~V., \& Sargent, W.~L.~W.\ 1989, \apjl, 342, L11 

1537: 

1538: \bibitem[Fukugita et al.(1995)]{fuku95} Fukugita, M., 

1539: Shimasaku, K., \& Ichikawa, T.\ 1995, \pasp, 107, 945 

1540: 

1541: \bibitem[Gebhardt et al.(2000)]{geb00} Gebhardt, K., \etal\ 2000,

1542: \apjl, 539, L13

1543: 

1544: \bibitem[Gebhardt et al.(2001)]{geb01} Gebhardt, K., et al.\ 2001,

1545: \aj, 122, 2469

1546: 

1547: \bibitem[Goodrich et al.(1995)]{gcp95} Goodrich, R.~W., 

1548: Cohen, M.~H., \& Putney, A.\ 1995, \pasp, 107, 179 

1549: 

1550: \bibitem[Greene \& Ho(2004)]{gh04} Greene, J. E., \& Ho, L. C. 2004,

1551:   \apj, 610, 622 (GH)

1552: 

1553: \bibitem[Greene \& Ho(2005)]{gh05} Greene, J.~E., \& Ho, L.~C.\ 2005,

1554: \apj, 627, 721

1555: 

1556: \bibitem[Greene \& Ho(2006)]{gh06} Greene, J.~E., \& Ho, 

1557: L.~C.\ 2006, \apjl, 641, L21 

1558: 

1559: \bibitem[Greene et al.(2006)]{ghu06} Greene, J.~E., Ho, 

1560: L.~C., \& Ulvestad, J.~S.\ 2006, \apj, 636, 56 

1561: 

1562: \bibitem[Greene, Ho, \& Barth(2008)]{ghb08} Greene, J. E., Ho, L. C.,

1563:   \& Barth, A. J.  2008, \apj, submitted

1564: 

1565: \bibitem[Greene \& Ho(2007)]{gh07} Greene, J.~E., \& Ho, L.~C.\ 2007,

1566: \apj, 670, 92 

1567: 

1568: \bibitem[Groves et al.(2006)]{gro06} Groves, B.~A., Heckman, 

1569: T.~M., \& Kauffmann, G.\ 2006, \mnras, 371, 1559 

1570: 

1571: \bibitem[Grupe \& Mathur(2004)]{gm04} Grupe, D., \& Mathur, S.\ 2004,

1572: \apjl, 606, L41

1573: 

1574: \bibitem[Hao et al.(2005a)]{hao05a} Hao, L., et al.\ 2005a, \aj, 

1575: 129, 1783 

1576: 

1577: \bibitem[Hao et al.(2005b)]{hao05b} Hao, L., et al.\ 2005b, \aj, 

1578: 129, 1795 

1579: 

1580: \bibitem[Heckman et al.(1981)]{heck81} Heckman, T.~M., Miley, 

1581: G.~K., van Breugel, W.~J.~M., \& Butcher, H.~R.\ 1981, \apj, 247, 403 

1582: 

1583: \bibitem[Heckman et al.(2004)]{hec04} Heckman, T.~M., 

1584: Kauffmann, G., Brinchmann, J., Charlot, S., Tremonti, C., \& White, 

1585: S.~D.~M.\ 2004, \apj, 613, 109 

1586: 

1587: \bibitem[Ho(2008)]{ho08} Ho, L. C. 2008, \araa, in press

1588: 

1589: \bibitem[Ho \etal(1997)]{hfs97} Ho, L. C.,

1590:   Filippenko, A. V., \& Sargent, W. L. W. 1997, \apjs, 112, 315

1591: 

1592: \bibitem[Ho(2005)]{ho05} Ho, L.~C.\ 2005, \apj, 629, 680 

1593: 

1594: % star formation histories:

1595: \bibitem[Kauffmann et al.(2003a)]{kau03sfr} Kauffmann, G., et 

1596: al.\ 2003a, \mnras, 341, 33 

1597: 

1598: % agn selection:

1599: \bibitem[Kauffmann et al.(2003b)]{kau03agn} Kauffmann, G., et 

1600: al.\ 2003b, \mnras, 346, 1055 

1601: 

1602: \bibitem[Keel(1980)]{keel80} Keel, W.~C.\ 1980, \aj, 85, 198 

1603: 

1604: \bibitem[Kewley \etal(2001)]{kew01} Kewley, L. J., Heisler, C. A.,

1605:   Dopita, M. A., \& Lumsden, S. 2001, \apjs, 132, 37

1606: 

1607: \bibitem[Kewley et al.(2006)]{kew06} Kewley, L.~J., Groves, 

1608: B., Kauffmann, G., \& Heckman, T.\ 2006, \mnras, 372, 961 

1609: 

1610: \bibitem[Kim et al.(2006)]{khi06} Kim, M., Ho, L.~C., \& Im, 

1611: M.\ 2006, \apj, 642, 702 

1612: 

1613: \bibitem[Kniazev et al.(2004)]{kni04} Kniazev, A.~Y., 

1614: Pustilnik, S.~A., Grebel, E.~K., Lee, H., \& Pramskij, A.~G.\ 2004, \apjs, 

1615: 153, 429 

1616:  

1617: \bibitem[Konigl \& Kartje(1994)]{kk94} Konigl, A., \& 

1618: Kartje, J.~F.\ 1994, \apj, 434, 446 

1619: 

1620: 

1621: \bibitem[Kormendy \& Richstone(1995)]{kr95} Kormendy, J., \&

1622: Richstone, D.\ 1995, \araa, 33, 581

1623: 

1624: 

1625: \bibitem[Kraemer et al.(1999)]{kra99} Kraemer, S.~B., Ho, 

1626: L.~C., Crenshaw, D.~M., Shields, J.~C., \& Filippenko, A.~V.\ 1999, \apj, 

1627: 520, 564 

1628: 

1629: \bibitem[Kunth et al.(1987)]{ksb87} Kunth, D., Sargent, 

1630: W.~L.~W., \& Bothun, G.~D.\ 1987, \aj, 93, 29 

1631: 

1632: \bibitem[Laor(2003)]{laor03} Laor, A.\ 2003, \apj, 590, 86 

1633: 

1634: \bibitem[Leonard et al.(2002)]{leo02} Leonard, D.~C., 

1635: Filippenko, A.~V., Chornock, R., \& Foley, R.~J.\ 2002, \pasp, 114, 1333 

1636: 

1637: \bibitem[Maiolino \& Rieke(1995)]{mr95} Maiolino, R., \& 

1638: Rieke, G.~H.\ 1995, \apj, 454, 95 

1639: 

1640: \bibitem[Malkan et al.(1998)]{mgt98} Malkan, M.~A., Gorjian, 

1641: V., \& Tam, R.\ 1998, \apjs, 117, 25 

1642: 

1643: \bibitem[Mathewson \& Ford(1970)]{mf70} Mathewson, D.~S., \& 

1644: Ford, V.~L.\ 1970, \memras, 74, 139 

1645: 

1646: \bibitem[Merritt et al.(2001)]{mfj01} Merritt, D., Ferrarese, 

1647: L., \& Joseph, C.~L.\ 2001, Science, 293, 1116 

1648: 

1649: \bibitem[Miller et al.(1988)]{mrg88} Miller, J.~S., Robinson, L.~B.,

1650: \& Goodrich, R.~W.\ 1988, Instrumentation for Ground-Based Optical

1651: Astronomy, Present and Future, ed.  L.B.~Robinson (New York:

1652: Springer-Verlag), 157

1653: 

1654: \bibitem[Moran et al.(1996)]{mhh96} Moran, E.~C., Halpern, 

1655: J.~P., \& Helfand, D.~J.\ 1996, \apjs, 106, 341 

1656: 

1657: \bibitem[Moran et al.(1999)]{mor99} Moran, E.~C., Filippenko, 

1658: A.~V., Ho, L.~C., Shields, J.~C., Belloni, T., Comastri, A., Snowden, 

1659: S.~L., \& Sramek, R.~A.\ 1999, \pasp, 111, 801 

1660: 

1661: \bibitem[Moran et al.(2000)]{mor00} Moran, E.~C., Barth, 

1662: A.~J., Kay, L.~E., \& Filippenko, A.~V.\ 2000, \apjl, 540, L73 

1663: 

1664: \bibitem[Nelson \& Whittle(1996)]{nw96} Nelson, C.~H., \& 

1665: Whittle, M.\ 1996, \apj, 465, 96 

1666: 

1667: \bibitem[Netzer et al.(2006)]{net06} Netzer, H., Mainieri, V., Rosati,

1668: P., \& Trakhtenbrot, B.\ 2006, \aap, 453, 525

1669: 

1670: \bibitem[Nicastro(2000)]{nic00} Nicastro, F.\ 2000, \apjl, 

1671: 530, L65

1672: 

1673: \bibitem[Nicastro et al.(2003)]{nmm03} Nicastro, F., 

1674: Martocchia, A., \& Matt, G.\ 2003, \apjl, 589, L13 

1675: 

1676: \bibitem[Osterbrock \& Ferland(2006)]{of06} Osterbrock, D.~E., \&

1677: Ferland, G.~J.\ 2006, Astrophysics of Gaseous Nebulae and Active

1678: Galactic Nuclei, 2nd.~ed. (Sausalito, CA: University Science Books)

1679: 

1680: \bibitem[Osterbrock \& Mathews(1986)]{om86} Osterbrock, 

1681: D.~E., \& Mathews, W.~G.\ 1986, \araa, 24, 171 

1682: 

1683: \bibitem[Pappa et al.(2001)]{pap01} Pappa, A., Georgantopoulos, I.,

1684: Stewart, G.~C., \& Zezas, A.~L.\ 2001, \mnras, 326, 995

1685: 

1686: \bibitem[Peterson et al.(2005)]{pet05} Peterson, B.~M., et 

1687: al.\ 2005, \apj, 632, 799 

1688: 

1689: \bibitem[Rigby et al.(2006)]{rig06} Rigby, J.~R., Rieke, 

1690: G.~H., Donley, J.~L., Alonso-Herrero, A., \& P{\'e}rez-Gonz{\'a}lez, P.~G.\ 

1691: 2006, \apj, 645, 115 

1692: 

1693: \bibitem[Salviander et al.(2007)]{sal07} Salviander, S., 

1694: Shields, G.~A., Gebhardt, K., \& Bonning, E.~W.\ 2007, \apj, 662, 131 

1695: 

1696: 

1697: \bibitem[Satyapal et al.(2007)]{sat07} Satyapal, S., Vega, 

1698: D., Heckman, T., O'Halloran, B., \& Dudik, R.\ 2007, \apjl, 663, L9 

1699: 

1700: \bibitem[Schlegel et al.(1998)]{sfd98} Schlegel, D.~J., 

1701: Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525 

1702: 

1703: \bibitem[Serkowski et al.(1975)]{smf75} Serkowski, K., 

1704: Mathewson, D.~L., \& Ford, V.~L.\ 1975, \apj, 196, 261 

1705: 

1706: \bibitem[Seth \etal(2008)]{seth08} Seth, A., Agueros, M., Lee, D., \&

1707:   Basu-Zych, A.  2008, ApJ, in press (arXiv:0801.0439)

1708: 

1709: \bibitem[Sheinis et al.(2002)]{she02} Sheinis, A.~I., Bolte, M., Epps,

1710: H.~W., Kibrick, R.~I., Miller, J.~S., Radovan, M.~V., Bigelow, B.~C.,

1711: \& Sutin, B.~M.\ 2002, \pasp, 114, 851

1712: 

1713: \bibitem[Shields et al.(2003)]{shi03} Shields, G.~A., 

1714: Gebhardt, K., Salviander, S., Wills, B.~J., Xie, B., Brotherton, M.~S., 

1715: Yuan, J., \& Dietrich, M.\ 2003, \apj, 583, 124 

1716: 

1717: \bibitem[Strateva et al.(2001)]{str01} Strateva, I., et al.\ 2001,

1718: \aj, 122, 1861

1719: 

1720: \bibitem[Strauss et al.(2002)]{str02} Strauss, M.~A., et al.\ 

1721: 2002, \aj, 124, 1810 

1722: 

1723: \bibitem[Thim et al.(2004)]{thim04} Thim, F., Hoessel, J.~G., 

1724: Saha, A., Claver, J., Dolphin, A., \& Tammann, G.~A.\ 2004, \aj, 127, 2322 

1725: 

1726: \bibitem[Thornton et al.(2008)]{tho08} Thornton, C. E., Barth, A. J.,

1727:   Ho, L. C., Rutledge, R. E., \& Greene, J. E. 2008, \apj, submitted

1728: 

1729: \bibitem[Tran(2003)]{tran03} Tran, H.~D.\ 2003, \apj, 583, 632 

1730: 

1731: \bibitem[Tremaine \etal(2002)]{tre02} Tremaine, S., et al.\ 2002,

1732: \apj, 574, 740

1733: 

1734: \bibitem[Turnshek et al.(1990)]{tur90} Turnshek, D.~A., 

1735: Bohlin, R.~C., Williamson, R.~L., II, Lupie, O.~L., Koornneef, J., \& 

1736: Morgan, D.~H.\ 1990, \aj, 99, 1243 

1737: 

1738: \bibitem[Ulvestad \& Ho(2001)]{uh01} Ulvestad, J.~S., \& Ho, L.~C.\

1739: 2001, \apj, 558, 561

1740: 

1741: \bibitem[Valluri et al.(2005)]{val05} Valluri, M., Ferrarese, L.,

1742: Merritt, D., \& Joseph, C.~L.\ 2005, \apj, 628, 137

1743: 

1744: \bibitem[Voges et al.(1999)]{vog99} Voges, W., et al.\ 1999, \aap,

1745: 349, 389

1746: 

1747: 

1748: \bibitem[Wang \& Lu(2001)]{wl01} Wang, T., \& Lu, Y.\ 2001, \aap, 377,

1749: 52

1750: 

1751: \bibitem[Whittle(1985)]{whit85} Whittle, M.\ 1985, \mnras, 

1752: 213, 1 

1753: 

1754: \bibitem[Whittle et al.(1988)]{whi88} Whittle, M., Pedlar, A., Meurs,

1755: E.~J.~A., Unger, S.~W., Axon, D.~J., \& Ward, M.~J.\ 1988, \apj, 326,

1756: 125

1757: 

1758: \bibitem[Wyithe(2006)]{w06} Wyithe, J.~S.~B.\ 2006, \mnras, 

1759: 365, 1082 

1760: 

1761: \bibitem[Zakamska et al.(2003)]{zak03} Zakamska, N.~L., et 

1762: al.\ 2003, \aj, 126, 2125 

1763: 

1764: 

1765: \bibitem[Zhou et al.(2004)]{zhou04} Zhou, H., Wang, T., Zhang, 

1766: X., Dong, X., \& Li, C.\ 2004, \apjl, 604, L33 

1767: 

1768: \end{thebibliography}

1769: 

1770: \clearpage

1771: 

1772: \begin{deluxetable}{lccccccccc}

1773: \tablewidth{7in} \tablecaption{Sample Properties}

1774: 

1775: \tablehead{ \colhead{Galaxy} & \colhead{Plate/MJD/Fiber} & \colhead{z} &

1776:      \colhead{$M_g$} & \colhead{$\log{M_\star}$} & \colhead{log $L$([O III])} &

1777:      \colhead{$C$} & \colhead{References}  \\

1778:      \colhead{} & \colhead{} &  \colhead{} & \colhead{(mag)} &

1779:      \colhead{(\msun)} &

1780:      \colhead{($L_\odot$)} & & }

1781: 

1782: \startdata

1783: \object{SDSS J010053.57+152728.1} & 421.51821.461 & 0.0404 & $-19.5$ & 9.75 & 5.83 & 2.59 & a,b \\

1784: \object{SDSS J011059.31+002601.1} & 397.51794.474 & 0.0188 & $-18.7$ & 9.81 & 6.23 & 2.46 & a \\

1785: \object{SDSS J011905.14+003745.0} & 398.51789.585 & 0.0327 & $-17.6$ & 9.16 & 6.64 & 3.26 & a,b \\

1786: \object{SDSS J021405.91-001637.0} & 405.51816.168 & 0.0373 & $-18.9$ & 9.72 & 7.37 & 2.74 & a,b \\

1787: \object{SDSS J034330.26-073507.4} & 462.51909.044 & 0.0357 & $-18.7$ & 9.61 & 5.81 & 2.84 & a,b \\

1788: \object{SDSS J091414.34+023801.8} & 567.52252.259 & 0.0735 & $-19.7$ & 9.97 & 7.56 & 2.78 & a \\

1789: \object{SDSS J091608.50+502126.9} & 766.52247.197 & 0.0495 & $-18.8$ & 9.90 & 6.39 & 2.64 & a \\

1790: \object{SDSS J094716.13+534944.9} & 769.52282.240 & 0.0383 & $-18.5$ & 9.56 & 6.78 & 2.85 & a \\

1791: \object{SDSS J100654.83+445642.8} & 943.52376.222 & 0.0419 & $-19.0$ & 10.1 & 6.99 & 3.20 & a \\

1792: \object{SDSS J102332.00+645240.2} & 489.51930.308 & 0.0404 & $-19.2$ & 9.78 & 6.17 & 2.76 & a,b \\

1793: \object{SDSS J103126.56+624648.6} & 772.52375.548 & 0.0529 & $-19.4$ & 10.50 & 7.55 & 2.49 & a \\

1794: \object{SDSS J103234.85+650227.9} & 489.51930.193 & 0.0056 & $-18.6$ & \nodata & 4.85 & 2.29 & b,c  \\

1795: \object{SDSS J105308.99+041036.2} & 579.52338.107 & 0.0431 & $-18.5$ & 9.54 & 6.61 & 2.61 & a \\

1796: \object{SDSS J110912.40+612346.7} & 775.52295.309 & 0.0067 & $-16.8$ & 8.07 & 4.94 & 1.95 & a \\

1797: \object{SDSS J111255.26+552928.8} & 909.52379.249 & 0.0498 & $-19.1$ & 9.86 & 6.53 & 2.53 & a \\

1798: \object{SDSS J114315.36+631108.2} & 776.52319.604 & 0.0622 & $-19.4$ & 9.68 & 7.16 & 2.37 & a \\

1799: \object{SDSS J120815.94+512325.7} & 882.52370.012 & 0.0331 & $-18.3$ & 9.60 & 6.54 & 2.45 & a \\

1800: \object{SDSS J130525.75+642121.5} & 602.52072.172 & 0.0527 & $-19.5$ & 10.41 & 7.37 & 2.41 & a,b \\

1801: \object{SDSS J142151.63+033121.2} & 584.52049.218 & 0.0552 & $-18.9$ & 9.59 & 6.66 & 2.76 & a \\

1802: \object{SDSS J143231.15+004614.4} & 535.51999.049 & 0.0712 & $-19.5$ & 9.97 & 7.65 & 2.50 & a,b \\

1803: \object{SDSS J143534.96+591658.3} & 790.52441.409 & 0.0682 & $-19.4$ & 10.41 & 7.51 & 2.48 & a \\

1804: \object{SDSS J144012.70+024743.5} & 536.52024.575 & 0.0298 & $-19.0$ & \nodata & 7.43 & 2.72 & b,d,e  \\ 

1805: \object{SDSS J151135.60+023807.0} & 540.51996.524 & 0.0389 & $-19.2$ & 9.71 & 6.89 & 2.83 & a,b \\

1806: \object{SDSS J155404.40+545708.2} & 619.52056.448 & 0.0457 & $-18.8$ & 9.98 & 6.46 & 2.45 & a,b\\

1807: \object{SDSS J160428.50-010435.7} & 344.51693.053 & 0.0296 & $-19.4$ & 9.78 & 6.45 & 2.74 & a \\

1808: \object{SDSS J160510.73+472901.5} & 813.52354.558 & 0.0435 & $-19.0$ & 9.85 & 6.08 & 2.85 & a \\

1809: \object{SDSS J161038.31+522334.2} & 623.52051.381 & 0.0617 & $-19.4$ & 10.04 & 6.51 & 1.95 &a,b  \\

1810: \object{SDSS J162917.44+425439.8} & 815.52374.032 & 0.0353 & $-18.1$ & 9.61 & 7.44 & 2.24 & a \\

1811: \object{SDSS J172344.17+583032.2} & 366.52017.274 & 0.0780 & $-19.7$ & 10.16 & 7.61 & 2.55 & a,b

1812: \enddata

1813: \tablecomments{``Plate/MJD/Fiber'' gives the reference to the SDSS

1814:   spectroscopic observation of each object, for cross-identification

1815:   in the SDSS archives.  References for previous AGN identifications

1816:   are: (a) \citet{kau03agn} with updated data from

1817:   http://www.mpa-garching.mpg.de/SDSS ; (b) \citet{hao05a}; (c)

1818:   \citet{mhh96}; (d) \citet{bfw78}; (e) \citet{and03}.  Stellar

1819:   masses, [\ion{O}{3}] luminosities, velocity dispersions, and

1820:   concentration indices $C$ are from the Kauffmann \etal\ online

1821:   catalog, except for 1032+6502 and 1440+0247; for these two objects

1822:   the concentration indices are from the SDSS database and

1823:   $L$([\ion{O}{3}]) is determined from the extinction-corrected

1824:   [\ion{O}{3}] flux and the galaxy distance.  

1825:   Absolute $g$-band magnitudes are corrected for Galactic

1826:   extinction using the \citet{sfd98} extinction maps and $k$-corrected

1827:   using M.\ Blanton's $kcorrect$ code.  }

1828: \label{sample}

1829: \end{deluxetable}

1830: 

1831: 

1832: \clearpage

1833: 

1834: 

1835: \begin{deluxetable}{lcccccc}

1836: \tablewidth{7in} \tablecaption{Velocity Dispersions and Linewidths}

1837: 

1838: \tablehead{ \colhead{Galaxy} & \colhead{Exp. Time} & \colhead{S/N} &

1839:   \colhead{$\sigmastar$} & \colhead{$\sigma$([O~III])} & 

1840:      \colhead{FWHM([O~III])} & \colhead{FWHM([S~II])} \\ \colhead{} & \colhead{(s)} &

1841:      \colhead{} & \colhead{(\kms)} & \colhead{(\kms)} & \colhead{(\kms)} & \colhead{(\kms)}  }

1842: 

1843: \startdata

1844: J010053.57$+$152728.1 & 900  & 10.9 & $52\pm4$ &  $36\pm2$  &$86\pm4$  & $77\pm9$ \\

1845: J011059.31$+$002601.1 & 3600 & 40.4 & $63\pm5$ &  $88\pm3$  &$143\pm11$ & $157\pm2$ \\

1846: J011905.14$+$003745.0 & 900  & 8.4  & $39\pm8$ &  $66\pm2$  &$77\pm3$  & $90\pm3$ \\

1847: J021405.91$-$001637.0 & 3600 & 35.6 & $59\pm5$ &  $181\pm1$ &$159\pm1$ & $142\pm2$ \\

1848: J034330.26$-$073507.4 & 3600 & 45.8 & $94\pm3$ &  $75\pm1$  & $223\pm3$\tablenotemark{a} & \nodata\tablenotemark{a} \\

1849: J091414.34$+$023801.8 & 1800 & 6.1  & $65\pm12$&  $68\pm1$  &$152\pm4$\tablenotemark{a} & $164\pm13$ \\

1850: J091608.50$+$502126.9 & 1800 & 11.6 & $48\pm3$ &  $75\pm3$  &$96\pm8$ & $75\pm2$  \\

1851: J094716.13$+$534944.9 & 1800 & 14.5 & $55\pm3$ &  $77\pm1$  &$147\pm1$ & $129\pm1$ \\

1852: J100654.83$+$445642.8 & 1800 & 12.0 & $65\pm5$ &  $147\pm13$ &$428\pm7$\tablenotemark{a} & \nodata\tablenotemark{a} \\

1853: J102332.00$+$645240.2 & 1800 & 16.8 & $56\pm5$ &  $57\pm1$  &$145\pm3$  & $161\pm5$ \\

1854: J103126.56$+$624648.6 & 1800 & 7.1  & $72\pm9$ &  $85\pm4$  &$138\pm13$ & $236\pm5$  \\

1855: J103234.85$+$650227.9 & 1800 & 25.2 & $43\pm4$ &  $69\pm1$  &$71\pm4$  & $53\pm2$ \\

1856: J105308.99$+$041036.2 & 1800 & 2.6  & \nodata  &  $84\pm8$  &$134\pm24$ & $125\pm18$  \\

1857: J110912.40$+$612346.7 & 1700 & 3.4  & \nodata  &  $28\pm1$  &$66\pm1$ & $63\pm2$ \\

1858: J111255.26$+$552928.8 & 1800 & 12.7 & $67\pm4$ &  $69\pm1$  &$109\pm4$ & $128\pm5$  \\

1859: J114315.36$+$631108.2 & 1800 & 14.7 & $70\pm9$ &  $123\pm2$ &$168\pm3$ & $152\pm8$ \\

1860: J120815.94$+$512325.7 & 1800 & 4.6  & $47\pm8$ &  $74\pm2$  &$174\pm8$ & $153\pm7$ \\

1861: J130525.75$+$642121.5 & 1800 & 9.9  & $77\pm4$ &  $70\pm3$  &$166\pm8$ & $218\pm4$ \\

1862: J142151.63$+$033121.2 & 1800 & 18.4 & $50\pm4$ &  $62\pm1$  &$71\pm2$  & $66\pm2$ \\

1863: J143231.15$+$004614.4 & 1800 & 7.9  & $65\pm8$ &  $75\pm6$  &$138\pm14$ & $118\pm9$ \\

1864: J143534.96$+$591658.3 & 1800 & 11.1 & $91\pm6$ &  $96\pm4$  &$162\pm10$ & $172\pm7$ \\

1865: J144012.70$+$024743.5 & 3600 & 32.0 & $45\pm4$ &  $122\pm1$ &$108\pm1$ & $105\pm1$ \\ 

1866: J151135.60$+$023807.0 & 1800 & 29.3 & $65\pm5$ &  $95\pm2$  &$131\pm4$ & $135\pm6$ \\

1867: J155404.40$+$545708.2 & 1800 & 25.0 & $73\pm3$ &  $62\pm1$  &$72\pm1$  & $63\pm3$ \\

1868: J160428.50$-$010435.7 & 1800 & 20.2 & $56\pm4$ &  $97\pm2$  &$94\pm4$  & $103\pm2$ \\

1869: J160510.73$+$472901.5 & 1800 & 19.7 & $67\pm3$ &  $149\pm5$ &$217\pm14$ & $197\pm6$ \\

1870: J161038.31$+$522334.2 & 1800 & 8.1  & $59\pm7$ &  $88\pm4$  &$152\pm12$ & $158\pm9$  \\

1871: J162917.44$+$425439.8 & 1800 & 17.3 & $61\pm4$ &  $238\pm1$ &$90\pm2$\tablenotemark{a} & $141\pm2$\tablenotemark{a,b} \\

1872: J172344.17$+$583032.2 & 1800 & 15.0 & $66\pm6$ &  $105\pm3$ &$193\pm6$ & $154\pm3$

1873: \enddata

1874: 

1875: \tablecomments{S/N is the mean signal-to-noise ratio per pixel in the

1876:   extracted ESI spectrum over the wavelength range 5200--5600 \AA. }

1877:   \tablenotetext{a}{[O III] profile is double-peaked.}

1878: \label{esidata}

1879: \end{deluxetable}

1880: 

1881: 

1882: \clearpage

1883: 

1884: 

1885: \begin{deluxetable}{lc}

1886: \tablewidth{3in}

1887: \tablecaption{Emission Lines in the Composite Seyfert 2 Spectrum}

1888: \tablehead{ \colhead{Line} & \colhead{$f/f(\hbeta)$} }

1889: \startdata

1890: {}[\ion{Ne}{3}] $\lambda$3868   &     $0.55  \pm    0.04$ \\

1891: {}[\ion{Ne}{3}] $\lambda$3967   &     $0.15  \pm    0.01$ \\

1892: H$\epsilon$                      &     $0.15  \pm    0.01$ \\

1893: {}[\ion{S}{2}] $\lambda$4068    &     $0.08  \pm    0.01$ \\

1894: H$\delta$                        &     $0.22  \pm    0.02$ \\

1895: H$\gamma$                        &     $0.41  \pm    0.02$ \\

1896: {}[\ion{O}{3}]    $\lambda$4363 &     $0.07  \pm    0.01$ \\

1897: \ion{He}{1}  $\lambda$4471       &     $0.03  \pm    0.01$ \\

1898: \ion{He}{2} $\lambda$4686        &     $0.20  \pm    0.02$ \\

1899: {}[\ion{Ar}{4}] $\lambda$4711   &     $0.04  \pm    0.01$ \\

1900: {}[\ion{Ar}{4}] $\lambda$4740   &     $0.04  \pm    0.01$ \\

1901: \hbeta\                          &     $1.00  \pm    0.03$ \\

1902: {}[OIII] $\lambda$4959          &     $2.07  \pm    0.06$ \\

1903: {}[OIII] $\lambda$5007          &     $6.47  \pm    0.16$ \\

1904: {}[N I] $\lambda$5199,5201      &     $0.12  \pm    0.01$ \\

1905: \ion{He}{1} $\lambda$5876        &     $0.12  \pm    0.01$ \\

1906: {}[\ion{Fe}{7}] $\lambda$6086   &     $0.04  \pm    0.01$ \\

1907: {}[\ion{O}{1}] $\lambda$6300    &     $0.52  \pm    0.02$ \\

1908: {}[\ion{S}{3}] $\lambda$6312    &     $0.05  \pm    0.01$ \\

1909: {}[\ion{O}{1}] $\lambda$6363    &     $0.22  \pm    0.02$ \\

1910: {}[\ion{N}{2}] $\lambda$6548    &     $0.95  \pm    0.11$ \\

1911: \hal\                            &     $4.63  \pm    0.15$ \\

1912: {}[\ion{N}{2}] $\lambda$6583    &     $2.64  \pm    0.12$ \\

1913: \ion{He}{1} $\lambda$6678        &     $0.02  \pm    0.01$ \\

1914: {}[\ion{S}{2}] $\lambda$6716    &     $1.18  \pm    0.03$ \\

1915: {}[\ion{S}{2}] $\lambda$6731    &     $0.99  \pm    0.03$ \\

1916: {}[\ion{Ar}{3}] $\lambda$7135   &     $0.19  \pm    0.01$ \\

1917: {}[\ion{O}{2}] $\lambda$7310    &     $0.09  \pm    0.01$ \\

1918: {}[\ion{O}{2}] $\lambda$7331    &     $0.10  \pm    0.01$ \\

1919: {}[\ion{S}{3}] $\lambda$9069    &     $0.74  \pm    0.05$ \\

1920: {}[\ion{S}{3}] $\lambda$9532    &     $1.93  \pm    0.12$ \\

1921: \enddata

1922: \tablecomments{ Line strengths measured from the composite Seyfert 2

1923:  spectrum are given relative to \hbeta.  The spectrum is corrected for

1924:  Galactic extinction but not for internal extinction within the host

1925:  galaxies.}

1926: \label{compositelines}

1927: \end{deluxetable}

1928: 

1929: \clearpage

1930: 

1931: 

1932: 

1933: \begin{deluxetable}{lccccccc}

1934: \tablewidth{7in} \tablecaption{Polarization Data}

1935: 

1936: \tablehead{ \colhead{Galaxy} & \colhead{S/N} & \colhead{$p_b$} & \colhead{$\theta_b$}

1937:     & \colhead{$p_r$} & \colhead{$\theta_r$} & \colhead{$E(B-V)$} &

1938:     \colhead{$p_\mathrm{max}\mathrm{(Galactic)}$}

1939: \\ \colhead{} & \colhead{} & 

1940:     \colhead{(\%)} & \colhead{(degrees)} & \colhead{(\%)} &

1941:     \colhead{(degrees)} & \colhead{(mag)} & \colhead{($\%$)}

1942: }

1943: 

1944: \startdata

1945: J011905.14$+$003745.0 & 80 & $0.55\pm0.06$ & $111\pm3$ & $0.29\pm0.05$ & $119\pm6$ & 0.032 & 0.29 \\

1946: J021405.91$-$001637.0 & 160 & $0.50\pm0.03$ & $160\pm2$ & $0.52\pm0.03$ & $159\pm2$ & 0.038 & 0.34 \\

1947: J094716.13$+$534944.9 & 105 & $0.22\pm0.04$ & $33\pm5$ & $0.20\pm0.04$ & $25\pm5$ & 0.014 & 0.13 \\

1948: J110912.40$+$612346.7 & 42 & $0.76\pm0.10$ & $69\pm4$ & $0.72\pm0.10$ & $72\pm4$ & 0.011 & 0.10 \\

1949: %HD 14069 (null) & $0.19\pm0.01$ & $132\pm1$ & $0.18\pm0.01$ & $136\pm1$ \\

1950: HD 57702 (null) & 1060 & $0.07\pm0.01$ & $161\pm1$ & $0.08\pm0.01$ & $150\pm2$ \\

1951: %HD 109055 (null) & $0.15\pm0.01$ & $94\pm1$ & $0.15\pm0.01$ & $78\pm1$

1952: \enddata 

1953: \tablecomments{S/N is the signal-to-noise per pixel over the

1954: continuum wavelength range 6380--6480 \AA\ in the total flux

1955: spectrum. Polarization on the blue and red sides ($p_b$, $p_r$) and

1956: polarization angles ($\theta_b$, $\theta_r$) are measured over the

1957: wavelength regions 4000--5500 and 6000--7500 \AA.  Galactic reddening estimates

1958: are from \citet{sfd98}, and the maximum estimated Galactic foreground

1959: polarization is $p_\mathrm{max} (\%) = 9.0\times E(B-V)$

1960: \citep{smf75}.

1961: \label{poldata}}

1962: \end{deluxetable}

1963: 

1964: \clearpage

1965: 

1966: 

1967: \begin{figure}

1968: \begin{center}

1969: \plotone{f1small.eps}

1970: \end{center}

1971: \caption{SDSS color composite images of the Seyfert galaxy sample.

1972: The image size is $1\arcmin\times1\arcmin$.  North is up and east is

1973: to the left.

1974: \label{images}}

1975: \end{figure}

1976: 

1977: 

1978: 

1979: \begin{figure}

1980: \plotone{f2asmall.eps}

1981: \caption{Central portions of the Keck ESI spectra of the SDSS Seyfert

1982: sample.  The spectra are binned to 0.5 \AA\ pixel\per\ for clarity.

1983: \label{spectra}}

1984: \end{figure}

1985: 

1986: \addtocounter{figure}{-1}

1987: \begin{figure}

1988: \plotone{f2bsmall.eps}

1989: \caption{(Continued) Keck ESI spectra of the SDSS Seyfert sample.}

1990: \end{figure}

1991: 

1992: 

1993: \begin{figure}

1994: \plotone{f3small.eps}

1995: \caption{Examples of starlight subtraction.  In each panel, the galaxy

1996:   spectrum is in black, the best-fitting model is in red, and the

1997:   starlight-subtracted spectrum is in blue.  The model spectra are

1998:   shifted downward by a constant offset for clarity.

1999: \label{starsub}}

2000: \end{figure}

2001: 

2002: 

2003: 

2004: \begin{figure}

2005: \plotone{f4.eps}

2006: \caption{Multi-Gaussian fits to the \hal, [\ion{N}{2}], and

2007:   [\ion{S}{2}] lines for four objects, after starlight subtraction.

2008:   In each plot, the narrow-line components are plotted in blue, the

2009:   broad components in magenta, and the sum of all model components in

2010:   red.  For the galaxy 1629+4254, all of the narrow lines contain the

2011:   redshifted wing component, and the \hal\ profile does not require an

2012:   additional broad component.

2013: \label{broadhal}

2014: }

2015: \end{figure}

2016: 

2017: 

2018: \begin{figure}

2019: \begin{center}

2020: \rotatebox{-90}{\scalebox{0.7}{\includegraphics{f5.eps}}}

2021: \end{center}

2022: \caption{Line-ratio diagrams for [O III] $\lambda5007$ / \hbeta\

2023:   versus [N II] $\lambda6583$ / \hal, [S II] $\lambda6716,6731$ /

2024:   \hal, and [O I] $\lambda6300$ / \hal.  Grey contours represent

2025:   emission-line galaxies from \cite{kau03agn} with S/N $>6$ in

2026:   \hal, [\ion{O}{3}], and [\ion{N}{2}], and S/N$>3$ in

2027:   [\ion{O}{1}] flux.  Blue circles are the Seyfert 1 galaxies from

2028:   \citet{gh04}, and red crosses are measurements from the Keck data

2029:   for the Seyfert sample described here.  NGC 4395 and POX 52 are

2030:   shown as a green triangle and square.  Dashed curves are the

2031:   ``maximum starburst'' lines from \citet{kew06}, and the horizontal

2032:   dotted line shows the selection cut at [\ion{O}{3}]/\hbeta\ $=3$.

2033: \label{bptdiagrams}}

2034: \end{figure}

2035: 

2036: 

2037: \begin{figure}

2038: \plotone{f6.eps}

2039: \caption{Two sections of the starlight-subtracted ESI spectrum of SDSS

2040:   J144012.70+024743.5, showing coronal-line emission including lines

2041:   of [\ion{Fe}{7}] and [\ion{Fe}{10}].  [\ion{Fe}{11}] emission is

2042:   also present at 7892 \AA.

2043: \label{1440lines}

2044: }

2045: \end{figure}

2046: 

2047: 

2048: \begin{figure}

2049: \begin{center}

2050: \scalebox{0.7}{\rotatebox{-90}{\includegraphics{f7.eps}}}

2051: \end{center}

2052: \caption{Composite Seyfert 2 spectrum.  \emph{Upper panel:}  Composite

2053:   spectrum (in black), model fit from the starlight subtraction

2054:   routine (red), and starlight-subtracted spectrum (blue).

2055:   \emph{Middle panel:}  Expanded view of the emission-line spectrum

2056:   after starlight subtraction.  \emph{Bottom panels:}  Expanded views

2057:   of the region including \hbeta\ and the Mg$b$ lines (left) and the

2058:   \ion{Ca}{2} triplet and [\ion{S}{3}] $\lambda\lambda$9069, 9532 \AA\

2059:   emission lines.  The model spectra (red) are shifted downward by a

2060:   constant offset for clarity.

2061: \label{composite}}

2062: \end{figure}

2063: 

2064: 

2065: \begin{figure}

2066: \plotone{f8.eps}

2067: \caption{[O III] $\lambda5007$ profiles with single or double-Gaussian

2068:   model fits. Individual Gaussian components are shown in blue and the

2069:   summed model is in red.

2070: \label{o3}}

2071: \end{figure}

2072: 

2073: 

2074: \begin{figure}

2075: \begin{center}

2076: \scalebox{0.5}{\rotatebox{-90}{\includegraphics{f9.eps}}}

2077: \end{center}

2078: \caption{Examples of fits from velocity dispersion measurements,

2079:   illustrating the range of data quality for the Keck observations.

2080:   Upper panels show fits to the \mgb\ spectral region and lower panels

2081:   show the \ion{Ca}{2} triplet region.  The spectrum in red is the

2082:   best-fitting broadened K-giant template star.

2083: \label{dispersions}

2084: }

2085: \end{figure}

2086: 

2087: 

2088: \begin{figure}

2089: \plotone{f10.eps}

2090: \caption{Comparison of stellar and emission-line velocity dispersions.

2091:   The upper panel shows the line dispersion \sigmao, and the lower

2092:   panel shows \fwhmo/2.35.  The solid line in each panel represents a

2093:   1:1 correspondence between stellar and gas linewidths.  Objects with

2094:   double-peaked profiles are not included.

2095: \label{linewidths}

2096: }

2097: \end{figure}

2098: 

2099: \clearpage

2100: 

2101: 

2102: 

2103: \begin{figure}

2104: \begin{center}

2105: \scalebox{0.3}{\includegraphics{f11a.eps}}

2106: \hspace*{0.3in}

2107: \scalebox{0.3}{\includegraphics{f11b.eps}}

2108: 

2109: \scalebox{0.3}{\includegraphics{f11c.eps}}

2110: \hspace*{0.3in}

2111: \scalebox{0.3}{\includegraphics{f11d.eps}}

2112: \end{center}

2113: \caption{Keck LRIS polarimetry data for four objects in the spectral

2114:   region surrounding the \hal\ emission line.  In each plot, the top

2115:   panel is the total flux spectrum in units of $10^{-16}$ erg s\per\

2116:   cm\persq\ \AA\per, the middle panel is the percent polarization

2117:   given in the form of the rotated Stokes parameter, and the bottom

2118:   panel is the Stokes flux $p \times f_\lambda$.

2119:   \label{polplot}

2120: }

2121: \end{figure}

2122: 

2123: \begin{figure}

2124: \plotone{f12.eps}

2125: \caption{Color-magnitude diagram for the galaxies in this sample,

2126:   based on SDSS photometry.  Standard colors for different Hubble

2127:   types are from \citet{fuku95}.

2128:   \label{cmd}

2129: }

2130: \end{figure}

2131: 

2132: 

2133: \begin{figure}

2134: \plotone{f13.eps}

2135: \caption{The \dn\ and \hdeltaa\ indices for this sample (red crosses).

2136:   Grey contours and small dots represent the distribution of galaxies

2137:   from the \citet{kau03agn} sample within the mass range $9.0 <

2138:   \log(M/\msun) < 10.0$.

2139: \label{hostage}

2140: }

2141: \end{figure}

2142: 

2143: \begin{figure}

2144: \plotone{f14.eps}

2145: \caption{Eddington ratio $\lbol/\ledd$ for our sample, based on the

2146:   assumptions of an [\ion{O}{3}] bolometric correction of

2147:   $\lbol/L$([\ion{O}{3}])$=3500$ \citep{hec04} and an offest of 0.23

2148:   dex from the \citet{tre02} \msigma\ relation.

2149: \label{lbol}

2150: }

2151: \end{figure}

2152: 

2153: \begin{figure}

2154: \plotone{f15.eps}

2155: \caption{Minimum black hole mass vs. stellar velocity dispersion,

2156:   assuming that the black holes are radiating at $\lbol/\ledd<1$, and

2157:   assuming a bolometric correction of $\lbol/L$([\ion{O}{3}])$=3500$.

2158:   For the two galaxies lacking measurements of \sigmastar, the

2159:   quantity FWHM([\ion{O}{3}])$/2.35$ has been used in place of the

2160:   stellar velocity dispersion.  The solid line is the \msigma\

2161:   relation from \citet{tre02}.

2162: \label{bhmin}}

2163: \end{figure}

2164: 

2165: 

2166: \begin{figure}

2167: \begin{center}

2168: \rotatebox{-90}{\scalebox{0.5}{\includegraphics{f16.eps}}}

2169: \end{center}

2170: \caption{Histograms of host galaxy absolute magnitude $M_g$,

2171:   [\ion{O}{3}] luminosity, and $\log($[\ion{O}{2}] $\lambda3727 / $

2172:   [\ion{O}{3}] $\lambda5007)$ for the \citet{gh04} Seyfert 1 sample

2173:   (upper panels) and the Seyfert 2 sample described in this paper

2174:   (lower panels).

2175: \label{hostcompare}

2176: }

2177: \end{figure}

2178: 

2179: 

2180: \end{document}

2181:  

2182: