1: %%

2: %% Beginning of file 'ms.tex'

3: %%

4: %% Modified 06 Mar 11

5: %%

6: %% This is a sample manuscript marked up using the

7: %% AASTeX v5.x LaTeX 2e macros.

8: 

9: %% The first piece of markup in an AASTeX v5.x document

10: %% is the \documentclass command. LaTeX will ignore

11: %% any data that comes before this command.

12: 

13: %% The command below calls the preprint style

14: %% which will produce a one-column, single-spaced document.

15: %% Examples of commands for other substyles follow. Use

16: %% whichever is most appropriate for your purposes.

17: 

18: % format + spacing preferred by jnewman.

19: %\documentclass[11pt,preprint]{aastex}

20: %\renewcommand{\baselinestretch}{1.75}

21: % format preferred by apj

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

23: 

24: %\documentclass[manuscript]{aastex}

25: 

26: % compact format preferred by mcc.

27: \documentclass{emulateapj}

28: 

29: %\setlength{\oddsidemargin}{-0.2 in}

30: %\setlength{\evensidemargin}{-0.2 in}

31: \setlength{\topmargin}{0.25 in}

32: %\setlength{\textheight}{9.5 in}

33: %\setlength{\textwidth}{7.0 in}

34: %\setlength{\marginparwidth}{0 in}

35: 

36: %\usepackage{emulateapj5}

37: %\usepakcage{apjfonts}

38: %\usepackage{onecolfloat}

39: %\usepackage{epsfig}

40: %\usepackage{amsmath}

41: 

42: % use natbib and aa for apj.

43: \usepackage{natbib}

44: %\citestyle{aa}

45: 

46: %% manuscript produces a one-column, double-spaced document:

47: % \documentclass[manuscript]{aastex}

48: %% preprint2 produces a double-column, single-spaced document:

49: % \documentclass[preprint2]{aastex}

50: 

51: %% If you want to create your own macros, you can do so

52: %% using \newcommand. Your macros should appear before

53: %% the \begin{document} command.

54: %%

55: %% If you are submitting to a journal that translates manuscripts

56: %% into SGML, you need to follow certain guidelines when preparing

57: %% your macros. See the AASTeX v5.x Author Guide

58: %% for information.

59: 

60: %\newcommand{\vdag}{(v)^\dagger}

61: %\newcommand{\myemail}{cooper@astron.berkeley.edu}

62: 

63: %% You can insert a short comment on the title page using the command below.

64: 

65: %\slugcomment{}

66: 

67: %% If you wish, you may supply running head information, although

68: %% this information may be modified by the editorial offices.

69: %% The left head contains a list of authors,

70: %% usually a maximum of three (otherwise use et al.).  The right

71: %% head is a modified title of up to roughly 44 characters.  Running heads

72: %% will not print in the manuscript style.

73: 

74: % more apj smack...

75: \shorttitle{Metallicity and Environment in the SDSS}

76: \shortauthors{Cooper et al.}

77: 

78: %% This is the end of the preamble.  Indicate the beginning of the

79: %% paper itself with \begin{document}.

80: 

81: \begin{document}

82: 

83: %% LaTeX will automatically break titles if they run longer than

84: %% one line. However, you may use \\ to force a line break if

85: %% you desire.

86: 

87: %\title{The Relationship between Environment and Metallicity: Environment's

88: %  Role in the Mass--Metallicity Relation}

89: \title{The Role of Environment in the Mass--Metallicity Relation}

90: 

91: %The Relationship between Star

92: %  Formation and Galaxy Environment at $\lowercase{z} < 1$}

93: 

94: %% Use \author, \affil, and the \and command to format

95: %% author and affiliation information.

96: %% Note that \email has replaced the old \authoremail command

97: %% from AASTeX v4.0. You can use \email to mark an email address

98: %% anywhere in the paper, not just in the front matter.

99: %% As in the title, you can use \\ to force line breaks.

100: 

101: \author{

102: Michael C.\ Cooper\altaffilmark{1,2},

103: Christy A.\ Tremonti\altaffilmark{1,3},

104: Jeffrey A.\ Newman\altaffilmark{4},

105: Ann I.\ Zabludoff\altaffilmark{1}

106: }

107: 

108: \altaffiltext{1}{Steward Observatory, University of Arizona, 

109: 933 N.\ Cherry Avenue, Tucson, AZ 85721 USA; 

110: cooper@as.arizona.edu, tremonti@as.arizona.edu, azabludoff@as.arizona.edu}

111: 

112: \altaffiltext{2}{Spitzer Fellow}

113: 

114: \altaffiltext{3}{Hubble Fellow}

115: 

116: \altaffiltext{4}{Department of Physics and Astronomy, University of

117:   Pittsburgh, 401--C Allen Hall, 3941 O'Hara Street, Pittsburgh, PA 15260

118:   USA; janewman@pitt.edu}

119: 

120: 

121: 

122: 

123: \begin{abstract}

124: 

125:   Using a sample of 57,377 star--forming galaxies drawn from the Sloan

126:   Digital Sky Survey, we study the relationship between gas--phase oxygen

127:   abundance and environment in the local Universe. We find that there is a

128:   strong relationship between metallicity and environment such that more

129:   metal--rich galaxies favor regions of higher overdensity. Furthermore,

130:   this metallicity--density relation is comparable in strength to the

131:   color--density relation along the blue cloud. After removing the mean

132:   dependence of environment on color and luminosity, we find a significant

133:   residual trend between metallicity and environment that is largely driven

134:   by galaxies in high--density regions, such as groups and clusters. We

135:   discuss the potential source of this relationship between metallicity and

136:   local galaxy density in the context of feedback models, with special

137:   attention paid to quantifying the impact of environment on the scatter in

138:   the mass--metallicity relation. We find that environment is a

139:   non--negligible source of scatter in this fundamental relation, with

140:   $\gtrsim \! 15\%$ of the measured scatter correlated with environment.

141: 

142: \end{abstract}

143: 

144: \keywords{galaxies:evolution, galaxies:statistics, galaxies:

145:   abundances, galaxies:fundamental parameters, large--scale structure of

146:   universe}

147: 

148: 

149: 

150: \section{Introduction}

151: \label{sec_intro}

152: 

153: 

154: Gas--phase metallicity is one of the most fundamental characteristics of a

155: galaxy, affecting the evolution of its stellar population and the

156: composition of its interstellar medium (ISM). Moreover, metallicity

157: indirectly traces a galaxy's star--formation history and reflects the

158: balance of several important physical processes: the release of metals into

159: the interstellar medium via supernovae and stellar winds, the ejection of

160: gas via galactic outflows, and the accretion of gas onto the galaxy from

161: the surrounding environs. Understanding how metallicity evolves, especially

162: in relation to other fundamental galaxy properties, is essential in

163: isolating the physical mechanisms that drive star formation and, more

164: generally, galaxy evolution.

165: 

166: 

167: As first observed by \citet{lequeux79}, metallicity is strongly correlated

168: with galaxy stellar mass, such that more massive galaxies are more

169: metal--rich in composition. Due to the relative ease of measuring

170: luminosities versus stellar masses, many subsequent studies extended this

171: early work to larger samples of galaxies by studying the correlation

172: between luminosity and metallicity \citep[e.g.,][]{skillman89, brodie91,

173:   zaritsky94, garnett02, kobulnicky03, lamareille04}. Using large data sets

174: from surveys such as the Sloan Digital Sky Survey \cite[SDSS,][]{york00},

175: more recent analyses have brought measurements of the luminosity-- and

176: mass--metallicity relations on par with each other, measuring relations

177: that span more than ten magnitudes in optical luminosity and six orders

178: of magnitude in stellar mass, ranging from dwarf galaxies up to the most

179: massive star--forming systems \citep[e.g.,][]{pilyugin00, lee03,

180:   tremonti04, shapley05, erb06, lee06}.

181: 

182: Both the luminosity--metallicity and mass--metallicity relations show

183: significant scatter, with only half the observed spread in the metallicity

184: distribution at fixed stellar mass being due to observational error and an

185: even greater ($\sim \! 50\%$ greater) scatter measured for the

186: luminosity--metallicity relation \citep{tremonti04}.  Various studies have

187: pointed to physical sources of the scatter in these fundamental

188: relations. For example, studying a sample of UV--selected galaxies at $z <

189: 0.4$, \citet{contini02} find that these systems are offset from the

190: luminosity--metallicity relation due to a recent starburst that has

191: enriched their ISM and decreased their mass--to--light ratios, moving them

192: off of the median trend. As illustrated by \citet{tremonti04}, however,

193: these results suggest that the relationship between metallicity and stellar

194: mass (and not luminosity) is more fundamental; even when accounting for

195: variations in mass--to--light ratio due to dust attenuation and observing

196: at redder wavelengths so as to minimize the impact of newly formed stars on

197: the measured luminosity, the scatter in the luminosity--metallicity

198: relation is still greater than that observed between stellar mass and

199: metallicity.

200: 

201: By analyzing the correlations between the scatter in the mass--metallicity

202: relation and other galaxy properties (e.g., rest--frame color, inclination,

203: photometric concentration, etc.), \citet{tremonti04} point to a potential

204: connection with stellar surface mass density, $\mu_*$, such that galaxies

205: with higher surface densities are more metal--rich relative to galaxies of

206: similar stellar mass \citep[see also][]{ellison08a}. This trend is

207: potentially explained by a scenario where galaxies with higher surface

208: densities have converted more of their gas reservoirs into stars and

209: thereby elevated their metallicity. In conflict with this picture, however,

210: \citet{tremonti04} find no significant correlation between scatter in the

211: mass--metallicity relation and morphology (as traced by the concentration).

212: 

213: 

214: Local galaxy density (i.e., the local ``environment'') could act as an

215: alternate source of the scatter in the mass--metallicity

216: relation. Supernovae are predicted to enrich the intergalactic medium (IGM)

217: over roughly Mpc scales \citep[e.g.,][]{adelberger05}, which would impact

218: the metallicity of nearby galaxies in high--density

219: environments. Similarly, in clusters of galaxies, intracluster supernovae

220: may inject a significant quantity of metals into the intracluster gas

221: \citep{domainko04}, which is subsequently accreted onto cluster members,

222: thereby raising their metallicity.

223: 

224: Galaxies in high--density regions should collapse and form stars earlier

225: than their counterparts in low--density environs; studies of the

226: color--density relation show that galaxies with older stellar populations

227: favor higher--density environments at $z \sim 0$ \citep[e.g.,][]{

228:   balogh04a, balogh04b, blanton05a} and at $z \sim 1$

229: \citep[e.g.,][]{cooper06, smith05}. Thus, galaxies might be expected to

230: become more metal--rich sooner in high--density regions.

231: 

232: Direct evidence for the potential role of environment in shaping the

233: metallicity of a galaxy is found in observational work by \citet{kewley06},

234: which shows that galaxy interactions, common in galaxy pairs and groups

235: \citep{cavaliere92}, may lead to inflows that drag metal--poor gas to the

236: galaxy center, decreasing the gas--phase metallicity in such systems

237: \citep[see also][]{ellison08b}. The analysis of \citet{kewley06}, however,

238: probes a limited range of extreme environments (focusing on pairs at close

239: separations), which provides a vastly incomplete view of the role of

240: environment. Similarly, analysis of 41 metal--rich, low--mass galaxies by

241: \citet{peeples08}, finds that such outliers on the mass--metallicity

242: relation tend to be isolated and undisturbed systems (i.e., reside in

243: low--density environments). Though, this work is clearly limited by its

244: small sample size and the restricted mass range probed.

245: 

246: 

247: In this paper, we utilize data from the Sloan Digital Sky Survey to study

248: the relationship between metallicity and environment among the nearby,

249: star--forming galaxy population. Specifically, we inspect the correlation

250: between metallicity and environment in comparison to well--established

251: correlations between environment and properties such as rest--frame

252: color. In addition, we examine the potential impact of environment on the

253: scatter in the mass--metallicity relation. In \S \ref{sec_data}, we outline

254: the data set used in this analysis. In \S \ref{sec_res1}, \S \ref{sec_rem},

255: and \S \ref{sec_res2}, we present our results on the relationship between

256: metallicity and environment at $z \sim 0.1$. We then endeavor to quantify

257: the role of environment in driving the scatter in the mass--metallicity

258: relation in \S \ref{sec_scatter}. In \S \ref{sec_disc} and \S

259: \ref{sec_sum}, the results of this analysis are then discussed and

260: summarized. Unless otherwise noted, all work in this paper employs a flat,

261: $\Omega_{\Lambda} = 0.7$, $\Omega_{m} = 0.3$, $h=1$ cosmology.

262: 

263: 

264: 

265: 

266: \section{Data Sample}

267: \label{sec_data}

268: 

269: To study the relationship between local galaxy environment and various

270: galaxy properties, including metallicity, we utilize data drawn from the

271: SDSS public data release 4 \citep[DR4,][]{adelman06}, as contained in the

272: NYU Value--Added Galaxy Catalog \citep[NYU--VAGC,][]{blanton05b}. We

273: restrict our analysis to the redshift regime $0.05 < z < 0.15$ in an effort

274: to probe a broad range in galaxy luminosity, with large sample size, while

275: minimizing aperture effects related to the finite size (3'') of the SDSS

276: fibers. In addition, we limit our sample to SDSS fiber plates for which the

277: redshift success rate for targets in the main spectroscopic survey is 80\%

278: or greater.

279: 

280: 

281: \subsection{Measurements of Local Galaxy Environments}

282: \label{sec_data_environ}

283: 

284: We estimate the local galaxy overdensity, or ``environment'', in the SDSS

285: using measurements of the projected $3^{\rm rd}$--nearest--neighbor surface

286: density $(\Sigma_3)$ about each galaxy, where the surface density depends

287: on the projected distance to the $3^{\rm rd}$--nearest neighbor, $D_{p,3}$,

288: as $\Sigma_3 = 3 / (\pi D_{p,3}^2)$. In computing $\Sigma_3$, a velocity

289: window of $\pm 1000\ {\rm km}/{\rm s}$ is employed to exclude foreground

290: and background galaxies along the line--of--sight. Tests by

291: \citet{cooper05} found this environment estimator to be a robust indicator

292: of local galaxy density within deep surveys.

293: 

294: 

295: To correct for the redshift dependence of the SDSS sampling rate, each

296: surface density value is divided by the median $\Sigma_3$ of galaxies at

297: that redshift within a window of $\Delta z = 0.02$; this converts the

298: $\Sigma_3$ values into measures of overdensity relative to the median

299: density (given by the notation $1 + \delta_3$ here) and effectively

300: accounts for redshift variations in the selection rate \citep{cooper05}.

301: Finally, to minimize the effects of edges and holes in the SDSS survey

302: geometry, we exclude all galaxies within $1\ h^{-1}$ Mpc (comoving) of a

303: survey boundary. For further details regarding the computation of galaxy

304: environments in the SDSS, we direct the reader to \citet{cooper06} and

305: \citet{cooper08}.

306: 

307: 

308: 

309: 

310: \subsection{Measurements of Rest--frame Color, Absolute Magnitude, and

311:   Stellar Mass}

312: \label{sec_data_photo}

313: 

314: We compute rest--frame $g-r$ colors, absolute $r$--band magnitudes $(M_r)$,

315: and stellar masses from the apparent, petrosian $ugriz$ magnitudes in the

316: SDSS DR4, using the {\it kcorrect} K--correction code (version v4\_1\_2) of

317: \citet[][see also \citealt{blanton03b}]{blanton07}. The template SEDs

318: employed by {\it kcorrect} are based on those of \citet{bc03}. To estimate

319: stellar masses, the best--fit SED given the observed $ugriz$ photometry and

320: spectroscopic redshift is used to directly compute the stellar

321: mass--to--light ratio $({\rm M}_{*}/L)$, assuming a \citet{charbrier03}

322: initial mass function. We have also employed the stellar mass estimates of

323: \citet{kauffmann03a}, which do not rely on fitting SEDs to the SDSS

324: photometry; instead they have been derived by fitting to stellar

325: absorption--line indices, measured from the observed SDSS spectra, while

326: also attempting to correct for attenuation due to dust. Using these

327: alternate stellar mass values produces no significant changes in the

328: results of our analyses. Finally, all magnitudes within this paper are

329: given in the AB system \citep{oke83}.

330: % Except for those in the nearly--AB SDSS system, all magnitudes in this

331: % paper are given in the AB system.

332: 

333: 

334: \begin{figure}[h!]

335: \centering

336: \plotone{f1.eps}

337: \caption{The color--magnitude distribution for all 246,242 galaxies within

338:   $0.05 < z < 0.15$ in the SDSS NYU--VAGC DR4 catalog (\emph{black solid

339:     lines}) and for the 57,377 star--forming galaxies with accurate

340:   environment and metallicity measurements in the redshift range $0.05 < z

341:   < 0.15$ (\emph{red points and dashed lines}). The dominant impact of the

342:   cuts made in selecting our star--forming sample is the exclusion of

343:   quiescent galaxies and AGN, which preferentially reside at the red end of

344:   the blue cloud or on the red sequence.}

345: \label{cmd_fig}

346: \end{figure}

347: 

348: 

349: 

350: \subsection{Measurements of Spectral Properties: 

351: Metallicity and Star--Formation Rate}

352: 

353: To study the metallicities of the SDSS galaxies, we utilize oxygen

354: abundances, $12 + \log_{10}({\rm O}/{\rm H})$, from \citet{tremonti04},

355: which have been derived by statistically comparing the fits of nebular

356: emission lines in the SDSS spectra to the models of \citet{charlot01}. The

357: sample is limited to only those sources with H$\beta$, H$\alpha$, and [N

358: {\small II}] $\lambda6584$ all detected at a $5\sigma$ level. Furthermore,

359: we constrain our analysis to the star--forming galaxy sample, excluding

360: those objects hosting an active galactic nucleus (AGN) according to the

361: conservative line--diagnostic criteria of \citet{kauffmann03b}. By also

362: requiring accurate environment measures, as described above, we arrive at a

363: final star--forming galaxy sample including 57,377 sources at $0.05 < z <

364: 0.15$. A distribution of the sample in color--magnitude space is shown in

365: Figure \ref{cmd_fig}. By excluding quiescent galaxies and active galactic

366: nuclei (AGN), the star--forming sample is biased against galaxies residing

367: on the red end of the blue cloud or the red sequence. In Figure

368: \ref{delta3_fig}, we also show the distribution of environment measures for

369: the star--forming sample relative to that for the full SDSS sample. While

370: the star--forming galaxies are biased towards lower overdensities

371: (consistent with a sample dominated by blue galaxies), the sample still

372: spans a full range of environments, from voids to clusters.

373: 

374: 

375: \begin{figure}[h!]

376: \centering

377: \plotone{f2.eps}

378: \caption{The distribution of the logarithm of the local ovderdensities,

379:   $\log_{10}(1+\delta_3)$, for all 232,882 SDSS galaxies within $0.05 < z <

380:   0.15$ and with accurate environment measurements (\emph{solid line}) and

381:   for the 57,377 star--forming galaxies with accurate environment and

382:   metallicity measurements (\emph{dashed line}). Here, we scale the two

383:   histograms so that their integrals are equal. The star--forming galaxies

384:   are biased towards lower overdensities, though the sample still spans the

385:   full range of environments probed by the SDSS.}

386: \label{delta3_fig}

387: \end{figure}

388: 

389: 

390: 

391: To probe the ongoing star--formation activity in this sample, we employ the

392: aperture--corrected star--formation rates (SFR) of \citet{brinchmann04},

393: which are estimated by fitting models to the nebular emission features in

394: the SDSS spectra. For the star--forming galaxy population, these SFRs show

395: excellent agreement with UV--based star--formation rate estimates

396: \citep{salim07}. Note that the SFR values of \citet{brinchmann04} are

397: estimated using $h=0.7$ rather than $h=1$.

398: 

399: 

400: \section{The Dependence of Mean Environment on Metallicity}

401: \label{sec_res1}

402: 

403: A wide variety of galaxy properties at low and intermediate redshift have

404: been shown to correlate with environment. For instance, at $z < 1$, blue,

405: star--forming galaxies are found to reside in regions of lower galaxy

406: density in comparison to red and dead systems \citep[e.g.,][]{balogh98,

407:   kauffmann04, cooper06, cucciati06, capak07}. Moving beyond direct studies

408: of the color--density or morphology--density relations, \citet{blanton05a}

409: analyzed the relationship between environment and the luminosities, surface

410: brightnesses, rest--frame colors, and structural characteristics

411: (S\'{e}rsic indices) of nearby galaxies in the SDSS sample. Among this set

412: of galaxy properties, they found that color and luminosity are the pair

413: that prove to be most predictive of the local environment. That is,

414: rest--frame color and luminosity are the two characteristics most closely

415: related to the galaxy density, as measured on small ($\sim 1\ h^{-1}$ Mpc)

416: scales. Furthermore, at fixed color and luminosity, they found no

417: significant trend between local galaxy density and surface brightness or

418: S\'{e}rsic index among the star--forming population --- although, for the

419: full SDSS sample, there is some residual correlation observed at high

420: luminosities, likely driven by rare, very luminous, red systems in dense

421: environments such as brightest cluster galaxies \citep{blanton05a}.

422: 

423: Like surface brightness and S\'{e}rsic index, metallicity is strongly

424: correlated with color and luminosity, such that brighter and redder sources

425: on the blue cloud tend to have higher metal abundances. This trend is

426: clearly evident in the top panel of Figure \ref{metal_fig}, where we show

427: the mean gas--phase oxygen abundance, $12 + \log_{10}({\rm O}/{\rm H})$, as

428: a function of rest--frame color and absolute magnitude for the SDSS

429: star--forming sample. Not surprisingly, when we examine the relationship

430: between metallicity and environment, we find a strong trend that includes

431: contributions from the correlations between (\emph{a}) metallicity, color,

432: and luminosity and (\emph{b}) color, luminosity, and environment. As shown

433: in the bottom portion of Figure \ref{metal_fig}, the typical environment

434: increases in overdensity for galaxies with higher

435: metallicities.\footnote{The local environment is thought to influence

436:   galaxy properties, such that galaxy properties are typically studied as a

437:   function of environment. In Figure \ref{metal_fig}b, however, we plot the

438:   dependence of mean environment on metallicity and not vice versa for one

439:   significant reason: measurements of environment are significantly more

440:   uncertain than measures of metallicity. Thus, binning galaxies according

441:   to local overdensity would yield significant correlation between

442:   neighboring environment bins, which would consequently smear out the

443:   underlying correlation between metallicity and local galaxy overdensity.}

444: This metallicity--environment relation agrees with the well--established

445: color--density relation along the blue cloud \citep{hogg03, blanton05a},

446: where the mean galaxy density increases with color. 

447: 

448: 

449: 

450: \begin{figure}

451: \plotone{f3.eps}

452: \caption{(\emph{Top}) We plot the mean gas--phase metallicity, $12 +

453:   \log_{10}({\rm O}/{\rm H})$, as a function of rest--frame color and

454:   absolute magnitude, computed in a sliding box of width $\Delta M_r = 0.2$

455:   and height $\Delta (g-r) = 0.05$, as shown in the upper left corner. The

456:   mean metallicity depends on both color and luminosity, with more luminous

457:   and redder galaxies tending to have greater metal

458:   content. (\emph{Bottom}) We plot the mean galaxy overdensity as a

459:   function of gas--phase metallicity for the star--forming population. The

460:   dashed black line and grey shaded region show the mean and 1$\sigma$

461:   uncertainty in the mean overdensity computed in a sliding box of width

462:   $\Delta (12 + \log_{10}({\rm O}/{\rm H})) = 0.1$. The points and

463:   corresponding error bars give the mean and $1\sigma$ error in the mean in

464:   discrete bins of metallicity. We compute overdensities using the full

465:   SDSS galaxy sample (i.e., not just the star--forming population), thus

466:   the mean values plotted here are generally less than zero (in the

467:   logarithm), as expected from the color--density relation.}

468: \label{metal_fig}

469: \end{figure}

470: 

471: While the trend evident in Figure \ref{metal_fig}b may not be surprising,

472: the \emph{strength} of this environment--metallicity relation is very

473: striking when compared to that seen between environment and color,

474: luminosity, stellar mass, or star--formation rate. As shown in Figure

475: \ref{metal_fig}b and Figure \ref{clss_fig}, the metallicity--density

476: relation is roughly comparable in strength to the color--density relation

477: amongst the star--forming population. In addition, the dependence of mean

478: overdensity on luminosity, stellar mass, and SFR are all weaker than that

479: observed with metallicity. Along the blue cloud, there is clearly a strong

480: relationship between gas--phase oxygen abundance and the local galaxy

481: environment.

482: 

483: 

484: \begin{figure*}[tb]

485: \plotone{f4.eps}

486: \caption{(\emph{Left}) For the star--forming population, we plot the

487:   dependence of mean overdensity, $\log_{10}(1+\delta_3)$, on rest--frame

488:   color and absolute magnitude, as given by the black circles plus solid

489:   line and red diamonds plus dashed line, respectively. (\emph{Right})

490:   Similar to the plot on the left, but for stellar mass, ${\rm M}_{*}$,

491:   (black circles and solid line) and star--formation rate (red diamonds and

492:   dashed line). Within the star--forming sample, the

493:   metallicity--environment trend is as strong as the color--density

494:   relation, which is the strongest of the relations plotted here.}

495: \label{clss_fig}

496: \end{figure*}

497: 

498: 

499: \section{Removing the Mean Color--Luminosity--Environment Relation}

500: \label{sec_rem}

501: 

502: 

503: Given the relationships between metallicity, color, and luminosity, it

504: would be reasonable to expect that the strong relationship between local

505: galaxy density and metallicity is entirely contained in the

506: color--luminosity--environment relation (within the precision of our

507: measurements), such that there is no residual trend between metallicity and

508: environment at fixed color and luminosity --- similar to the findings of

509: \citet{blanton05a} for surface brightness and S\'{e}rsic index. To probe

510: the dependence of environment on metallicity at fixed color and luminosity,

511: we fit and remove (subtract) the dependence of mean environment on

512: rest--frame color and absolute magnitude. Figure \ref{mden_fig}a shows the

513: mean overdensity as a function of rest--frame $g-r$ color and $r$--band

514: absolute magnitude, or $< \log_{10}(1 + \delta_3)[g-r, M_r] >$, for the

515: SDSS star--forming sample. There is a clear color--density trend, where the

516: mean overdensity increases with color along the blue cloud. To remove this

517: relationship of environment to color and luminosity, we subtract the mean

518: overdensity at the color and luminosity of each galaxy from the measured

519: overdensity:

520: %

521: \begin{equation}

522:   \Delta_3 = \log_{10}(1 + \delta_3) - < \log_{10}(1 + \delta_3)[g-r, M_r]

523:   > ,

524: \label{D3_eqn}

525: \end{equation}

526: %

527: where the distribution of mean environment with color and absolute

528: magnitude, $ < \log_{10}(1 + \delta_3)[g-r, M_r] > $ (see Fig.\

529: \ref{mden_fig}a), is median smoothed on $\Delta(g-r) = 0.15$ and $\Delta

530: M_{r} = 0.6$ scales prior to subtraction.

531: 

532: 

533: \begin{figure*}[tb]

534: \plotone{f5.eps}

535: \caption{(\emph{Left}) For the star--forming sample, we show the mean

536:   galaxy overdensity, $\log_{10}(1+\delta_3)$, as a function of rest--frame

537:   galaxy color, $g-r$, and absolute magnitude, $M_r$, computed in a sliding

538:   box of width $\Delta M_r = 0.2$ and height $\Delta (g-r) = 0.05$. The

539:   size and shape of the box are illustrated in the upper left corner of the

540:   plot. (\emph{Middle}) The mean residual environment, $\Delta_3$, as a

541:   function of color and magnitude, computed in the same sliding

542:   box. (\emph{Right}) We plot the distribution of mean residual environment

543:   for all regions where the sliding box contains 20 or more galaxies.)}

544: \label{mden_fig}

545: \end{figure*}

546: 

547: 

548: 

549: An alternate method of effectively removing the

550: color--luminosity--environment relation from our analysis would be to study

551: the metallicity--environment relation in bins of rest--frame color and

552: absolute magnitude (or in bins of stellar mass). This approach, however,

553: can be far less sensitive, since dividing the sample into such restricted

554: subsets reduces the signal--to--noise ratio of any trend that occurs across

555: the entire color--magnitude distribution (i.e., spans the blue cloud). In

556: \S \ref{sec_disc_2}, we return to this point in comparison to other recent,

557: related analyses.

558: 

559: 

560: The ``residual'' environment, $\Delta_3$, quantifies the overdensity about

561: a galaxy relative to galaxies of similar color and luminosity, where values

562: of $\Delta_3$ greater than zero correspond to galaxies in environments more

563: overdense than the typical galaxy with like star--formation history (that

564: is, like $g-r$ and $M_r$). Figure \ref{mden_fig}b shows the dependence of

565: mean $\Delta_3$ on color and luminosity; no significant color or luminosity

566: dependence is evident. Furthermore, Figure \ref{mden_fig}c displays the

567: distribution of $< \! \Delta_3 \! >$ values from Fig.\ \ref{mden_fig}b,

568: illustrating that deviations from $< \! \Delta_3 \! > = 0$ are small.

569: 

570: 

571: While the $\Delta_3$ statistic effectively removes the mean color--density

572: and luminosity--density relations from the data set, this measure of the

573: residual environment is only a small perturbation to the ``absolute''

574: overdensity, $\log_{10}(1+\delta_3)$. As shown in Figure \ref{diff_fig},

575: the $\Delta_3$ value for each galaxy in our sample is still strongly

576: correlated with the corresponding $\log_{10}(1+\delta_3)$ measurement. This

577: close correlation is, at least in part, due to the large uncertainty in

578: individual overdensity, $\log_{10}(1+\delta_3)$, measures. The bias towards

579: $\Delta_3 > \log_{10}(1 + \delta_3)$ is a product of the color--density

580: relation and the inclusion of red--sequence galaxies in the measurement of

581: galaxy overdensities (see \S \ref{sec_data_environ}).

582: 

583: 

584: 

585: 

586: \begin{figure}[h]

587: \plotone{f6.eps}

588: \caption{For the 57,377 galaxies in the star--forming population, we plot

589:   the relationship between the ``residual'' environment, $\Delta_3$, and

590:   the ``absolute'' environment, $\log_{10}(1+\delta_3)$. For a definiton of

591:   the $\Delta_3$ statistic, refer to Equation \ref{D3_eqn}.}

592: \label{diff_fig}

593: \end{figure}

594: 

595: 

596: 

597: 

598: \section{The Residual Dependence of Environment on Metallicity}

599: \label{sec_res2}

600: 

601: By studying the dependence of residual environment, $\Delta_3$, on various

602: galaxy properties, we can determine whether there is any excess trend with

603: environment beyond that contained in the color--luminosity--environment

604: relation. As a sanity check, in Figure \ref{resid_fig1}a we examine the

605: dependence of mean $\Delta_3$ on color or absolute magnitude, and confirm

606: that there is no trend with these properties, as expected. We likewise test

607: for any dependence of residual environment on stellar mass or

608: star--formation rate (see Figure \ref{resid_fig1}b). 

609: 

610: We find no signigicant trend of $\Delta_3$ with color, luminosity, stellar

611: mass, or SFR. This result is clearly to be expected for color and

612: luminosity, by construction. Given the relatively tight relationship

613: between the combination of $g-r$ and $M_r$ with ${\rm M}_{*}$

614: \citep[e.g.,][]{kauffmann03a, cooper07}, it is also not surprising to find

615: no residual trend with stellar mass, as an additional test. When using the

616: stellar mass values of \citet{kauffmann03a}, which were derived from fits

617: to stellar absorption features in the SDSS spectra rather than computed

618: directly from the SDSS photometry (see \S \ref{sec_data_photo}), we find a

619: similar lack of any trend. For star--formation rate, which exhibits a

620: weaker correlation with absolute environment, $\log_{10}(1+\delta_3)$, we

621: find no evidence for a relationship with residual environment, $\Delta_3$,

622: much like the lack of secondary environment--dependencies on surface

623: brightness or S\'{e}rsic index found by \citet{blanton05a}.

624: 

625: 

626: 

627: 

628: 

629: \begin{figure*}[tb]

630: \plotone{f7.eps}

631: \caption{(\emph{Left}) The dependence of mean residual environment,

632:   $\Delta_3$, on rest--frame color and absolute magnitude, as given by the

633:   black circles plus solid line and red diamonds plus dashed line,

634:   respectively. (\emph{Right}) Similar to the plot on the left, but for

635:   stellar mass, ${\rm M}_{*}$, (black circles and solid line) and

636:   star--formation rate (red diamonds and dashed line). After removing the

637:   mean dependence of environment on color and luminosity, we find no

638:   significant residual trend with color, luminosity, stellar mass, or

639:   star--formation rate.}

640: \label{resid_fig1}

641: \end{figure*}

642: 

643: 

644: 

645: Turning our attention towards metallicity, we examine the dependence of

646: mean residual environment, $\Delta_3$, on gas--phase oxygen abundance; as

647: shown in Figure \ref{resid_fig2}, there is a striking trend such that more

648: metal--rich galaxies typically reside in more overdense environments

649: relative to galaxies of like color and luminosity (i.e., of like stellar

650: mass). While the residual environment statistic, $\Delta_3$, has no

651: relationship with color, luminosity, stellar mass, or SFR, it is strongly

652: related to metallicity. In particular, this trend seems to be most

653: significant among the most metal--rich galaxies $(12 + \log_{10}({\rm

654:   O}/{\rm H}) > 9.1)$.

655: 

656: 

657: 

658: \begin{figure}[h!]

659: \plotone{f8.eps}

660: \caption{The dependence of mean residual environment, $\Delta_3$, on

661:   metallicity. We find a strong trend with metal--rich galaxies being

662:   found, on average, in regions of higher overdensity relative to galaxies

663:   of like color and luminosity.}

664: \label{resid_fig2}

665: \end{figure}

666: 

667: 

668: Given that the residual environment closely traces the absolute overdensity

669: measurement (see Fig.\ \ref{diff_fig}), it is interesting to examine this

670: residual metallicity--environment relation from the opposite

671: perspective. Figure \ref{resid_fig3} shows the dependence of mean

672: gas--phase oxygen abundance on the residual environment within the SDSS

673: star--forming sample. While studying mean relations from this perspective

674: is physically intuitive, binning galaxies according to environment

675: (residual or absolute) introduces significant correlation between

676: neighboring environment bins, due to the significant uncertainties in

677: measuring local galaxy densities ($\sigma_{\log(1+\delta_3)} \sim 0.5$

678: versus $\sigma_{12 + \log({\rm O}/{\rm H})} \sim 0.1$), which can therefore

679: weaken or erase any underlying trends. Despite this smearing effect, we

680: still find a strong trend, where the mean metallicity increases

681: dramatically in higher density regions $(\Delta_3 \gtrsim 1)$; this

682: suggests that the residual metallicity--environment relation is dominated

683: by phenomena occurring in overdense regions (such as groups and clusters),

684: rather than underdense environments.

685: 

686: 

687: \begin{figure}[h]

688: \plotone{f9.eps}

689: \caption{The dependence of mean (black points and solid line) and median

690:   (red dashed line) metallicity on the residual environment, $\Delta_3$. A

691:   strong correlation is found between residual environment and metallicity

692:   in overdense regions. This trend is evident, despite smearing effects

693:   related to the relatively large uncertainty in individual environment

694:   measures (see text) and the statistical dominance of galaxies with

695:   metallicities at $12 + \log_{10}({\rm O}/{\rm H}) \sim 9$, where the

696:   metallicity--environment relation is weak (see Fig.\

697:   \ref{resid_fig2}). Note that in this figure the median relation has been

698:   offset by $-0.03$ in $12 + \log_{10}({\rm O}/{\rm H})$ to facilitate

699:   display.}

700: \label{resid_fig3}

701: \end{figure}

702: 

703: \section{Scatter in the Mass--Metallicity Relation}

704: \label{sec_scatter}

705: 

706: 

707: The excess correlation between metallicity and environment, beyond that

708: contained in the color--luminosity--environment relation (or stellar

709: mass--environment relation), strongly suggests that the shape or

710: normalization of the mass--metallicity relation must depend on local galaxy

711: environment. This suggestion is confirmed in Figure \ref{mz_fig}, where we

712: show fits to the mass--metallicity relation, computed using galaxies in the

713: extreme quintiles of the environment distribution. Over the entire range of

714: stellar masses probed by the SDSS sample, the mass--metallicity relation is

715: biased towards higher metallicities in higher--density regions.

716: 

717: 

718: 

719: \begin{figure}[h]

720: \plotone{f10.eps}

721: \caption{We plot the gas--phase oxygen abundance versus stellar mass for

722:   the star--forming sample and overplot the fits to the mass--metallicity

723:   relation in the extreme quintiles of the residual overdensity

724:   distribution. The contours correspond to galaxy numbers of $N_{\rm

725:     galaxy} = 50, 200, 500, 1000, 1500, 2000$, while the solid black and

726:   dashed red lines show the median mass--metallicity relation for galaxies

727:   residing in high-- and low--density regions, respectively. The lines

728:   follow the median metallicity values, computed in discrete bins of

729:   stellar mass. At all masses, the median metallicity of galaxies in

730:   high--density regions is greater than that of galaxies in low--density

731:   regions.}

732: \label{mz_fig}

733: \end{figure}

734: 

735: 

736: While Figure \ref{mz_fig} clearly illustrates the environment dependence of

737: the mass--metallicity relation, showing an offset towards higher

738: metallicity in higher--density regions, it does not \emph{quantify} the

739: level to which environment contributes to the scatter in this fundamental

740: relationship. To this end, we examine the correlation between environment

741: and the residual metallicity, $\Delta_{({\rm O}/{\rm H})} = 12 +

742: \log_{10}({\rm O}/{\rm H}) - f({\rm M}_{*})$, measured relative to the

743: median mass--metallicity relation, $f({\rm M}_{*})$, as determined by the

744: full star--forming sample.

745: 

746: 

747: As shown in Figure \ref{mz_fig2}, the average residual metallicity

748: exhibits a clear dependence on environment, such that galaxies in overdense

749: regions are biased towards higher metallicities than galaxies of like

750: stellar mass. This result is effectively a rephrasing of the trend shown in

751: Fig.\ \ref{resid_fig3} and Fig.\ \ref{mz_fig}, except that in this form we

752: are able to subtract the average offset in the mass--metallicity relation

753: due to environment, yielding a quantity

754: %

755: \begin{equation}

756:   \epsilon = 12 + \log_{10}({\rm O}/{\rm H}) - < \Delta_{({\rm

757:       O}/{\rm H})}[\Delta_{3}] >,

758: \end{equation}

759: %

760: which gives the metallicity corrected for the observed environment

761: dependence. 

762: 

763: Subtracting (in quadrature) the measured scatter in the mass--$\epsilon$

764: relation from the scatter in the mass--metallicity relation, we find that

765: environment is correlated with $\gtrsim \! 15\%$ of the observed scatter in

766: the mass--metallicity relation. This environment--dependence is evident,

767: with comparable strength, at all stellar masses. As discussed in \S 4 and

768: \S 5, the relatively large uncertainties in the environment measurements

769: can smear out the underlying correlation between metallicity and

770: environment, thereby weakening the measured contribution of environment to

771: the scatter in the mass--metallicity relation. Thus, local environment is

772: correlated with \emph{at least} $15\%$ of the observed scatter, which

773: represents a non--negligible contribution to the total intrinsic scatter.

774: 

775: 

776: 

777: While we find a significant offset in the normalization of the

778: mass--metallicity relation in different environments, we do not detect any

779: environment--dependent variation in the intrinsic scatter. As shown in

780: Figure \ref{mz_fig3}, the measured root--mean-square (RMS) scatter in the

781: mass--metallicity relation is independent of environment, at a constant

782: level of roughly $\sigma_{{\rm O}/{\rm H}} \sim 0.1$. This suggests that

783: whatever is dominating the intrinsic scatter in the mass--metallicity

784: relation is independent of local galaxy overdensity.

785: 

786: 

787: \begin{figure}[h]

788: \plotone{f11.eps}

789: \caption{The median residual metallicity, relative to the median

790:   mass--metallicity relation, as a function of environment. We find a

791:   significant offset in metallicity (relative to the median

792:   mass--metallicity relation) as a function of galaxy overdensity. The

793:   dashed red line is the smoothed relation used to compute $\epsilon$. Note

794:   that the dependence of the mean residual metallicity on environment

795:   closely follows the relation shown for the median residual metallicity.}

796: \label{mz_fig2}

797: \end{figure}

798: 

799: 

800: \begin{figure}[h]

801: \plotone{f12.eps}

802: \caption{The root--mean-square of the devations in the mass--metallicity

803:   relation as a function of environment. The intrinsic scatter (or

804:   ``puffiness'') of the mass--metallicity relation shows no variation with

805:   environment. Note that the errors on $\sigma_{({\rm O}/{\rm H})}$ are

806:   smaller than the data points, since each bin contains $> \! 500$

807:   galaxies.}

808: \label{mz_fig3}

809: \end{figure}

810: 

811: 

812: \section{Discussion}

813: \label{sec_disc}

814: 

815: 

816: \subsection{Potential Selection Effects}

817: 

818: While we utilize the relatively conservative line--diagnostic criteria of

819: \citet{kauffmann03b} for excluding AGN from our sample, any significant

820: amount of contamination from AGN emission in the integrated galaxy spectra

821: could potentially impact the oxygen abundance measurements, biasing them

822: towards high (or low) metallicity. If AGN are strongly correlated with a

823: given environment (e.g., if they are preferrentially found in high--density

824: regions), then the metallicity--density relation could be (at some level) a

825: product of this underlying AGN--environment correlation. Of particular

826: interest is the relationship between Low Ionization Nuclear Emission--line

827: Regions \citep[LINERs,][]{heckman80} and environment as low--level AGN such

828: as LINERs are more likely to contaminate the star--forming sample than

829: their more powerful Seyfert counterparts.

830: 

831: 

832: Using the SDSS data set, several studies of the relationship between AGN

833: activity and environment have uncovered no significant correlation between

834: low--level AGN activity and local galaxy density. For example,

835: \citet{miller03} found that the fraction of AGN shows no variation with

836: environment within the SDSS early data release \citep{stoughton02}, a

837: result supported by later work using the larger DR4 data set

838: \citep{sorrentino06}. While analysis by \citet{montero08} shows that the

839: fraction of LINERs and Seyferts on the red sequence is potentially lower in

840: high--density environments locally, this result may not be representative

841: of the environments of LINERs in the blue cloud (i.e., among the

842: star--forming population). In partial agreement with the results of

843: \citet{montero08}, \citet{kauffmann04} conclude that the fraction of

844: galaxies hosting a powerful ($L$[O {\small III}]\ $> 10^7 L_{\sun}$) AGN

845: decreases in high--density environments. However, these are not the AGN

846: that are likely to contaminate our star--forming sample. For low--level AGN

847: activity, there is no evidence for a correlation with environment and thus

848: it is unlikely that contamination from AGN would contribute to the observed

849: correlation between metallicity and environment in our sample.

850: 

851: 

852: 

853: As shown by several studies of star--forming galaxies in the local

854: Universe, there is a correlation between gas--phase oxygen abundance and

855: galaxy morphology, such that more bulge--dominated systems are typically

856: more metal--rich \citep[e.g.,][]{vila92, zaritsky93, zaritsky94}. An

857: analogous trend is found when studying stellar metallicities among a more

858: diverse galaxy population \citep{gallazzi08}. Any relationship between

859: metallicity and environment separate from that observed with stellar mass

860: could therefore be a derivative of the well--known morphology--density

861: relation \citep[e.g.,][]{davis76, dressler80}. 

862: 

863: Given the strong correlations between luminosity, color, and morphology on

864: the blue cloud, the existence of a significant correlation between residual

865: environment, $\Delta_3$, and morphology is unlikely when none is found with

866: luminosity or rest--frame color. However, we investigate this possibility

867: using the S\'{e}rsic indices of \citet{blanton03a, blanton05a}. While the

868: S\'{e}rsic index is a measure of morphology derived from the fit of only a

869: single component to the galaxy's radial profile (versus bulge--disk

870: decomposition, for example), we find no dependence of mean residual

871: environment on S\'{e}rsic among our sample. Furthermore, recent analysis of

872: star--forming galaxies in the SDSS found that the mass--metallicity

873: relation shows no dependence on bulge fraction \citep{ellison08a}. Plus, as

874: stated in \S \ref{sec_intro}, \citet{tremonti04} found no correlation

875: between the scatter in the mass--metallicity relation and galaxy

876: concentration. Thus, we conclude that the portion of the scatter in the

877: mass--metallicity relation correlated with environment is not attributable

878: to variations in galaxy morphology.

879: 

880: 

881: 

882: 

883: \subsection{Theoretical Interpretation}

884: \label{sec_disc_1}

885: As discussed in \S \ref{sec_intro}, gas--phase metallicity and its

886: relationship with stellar mass within the star--forming population is

887: directly connected to feedback associated with star formation, as metals

888: are added to the ISM via supernovae and as gas is ejected via outflows and

889: accreted from the surrounding intergalactic medium (IGM). The presence of

890: outflows in star--forming galaxies has been supported by a variety of

891: observations \citep[e.g.,][]{lehnert96, frye02, weiner08}, but the physics

892: of this feedback mechanism remains poorly understood.

893: 

894: In an attempt to explain the mass--metallicity relation, early feedback

895: models \citep[e.g.,][]{dekel86, cole91, dekel03} employed energy--driven

896: winds, powered by supernovae explosions \citep{larson74}, to expel metals

897: from low--mass galaxies. Such models, however, fail to include the role of

898: winds in more massive systems (${\rm M}_{*} \gtrsim 10^{10}\ {\rm

899:   M}_{\sun}$), while observational work has shown outflows to be common at

900: galaxy stellar masses of $\gtrsim 10^{11}\ {\rm M}_{\sun}$

901: \citep[e.g.,][]{shapley03, rupke05, weiner08}.  

902: 

903: In contrast, the models of \citet{springel03a} incorporate winds at all

904: mass scales, but their simple prescription relies on winds of a constant

905: velocity \citep[$484\ {\rm km}/{\rm s}$,][]{springel03b}, independent of

906: galaxy mass. In disagreement with this approach, recent observations by

907: \citet{martin05} show outflow velocities to scale approximately linearly

908: with circular velocity (i.e., increase with $M_{*}$). Furthermore, simple

909: wind approximations such as that of \citet{springel03b} fail to reproduce

910: some properties of the IGM at higher redshift \citep[e.g., underpredicting

911: metal enrichment,][]{aguirre05} and the mass--metallicity relation at $z

912: \sim 2$ \citep{finlator07}.

913: 

914: 

915: Recent work by \citet{finlator07} has ventured to take a more detailed

916: approach to modeling the feedback in star--forming galaxies \citep[see

917: also][]{oppenheimer06}. In their model, outflows are pushed by

918: momentum--driven winds \citep{mqt05}, where momentum is deposited into the

919: ISM by coupling with the radiation from star formation through dust

920: absorption and where the wind speed scales with the galaxy's circular

921: velocity. Rather than assuming a wind that is driven in all directions

922: \citep[such as that of][]{springel03a}, \citet{finlator07} model polar

923: outflows with constrained opening angles ($\sim \! 45^{\circ}$) such that

924: the resulting outflows much more closely imitate those observed locally

925: \citep[e.g.,][]{veilleux05}.

926: 

927: 

928: In addition to assuming a wind speed that scales linearly with rotational

929: speed, the \citet{finlator07} model assumes that the mass--loading factor

930: --- the rate of mass ejection divided by the star--formation rate --- is

931: inversely related to the circular velocity. These scaling relations evolve

932: naturally for momentum--driven winds \citep{mqt05} and are in rough

933: agreement with results from other detailed feedback models

934: \citep[e.g.,][]{kobayashi07, brooks07}. Within this theoretical framework,

935: the gas--phase metallicity at any epoch depends on (\emph{i}) the mean

936: metallicity of accreted gas and (\emph{ii}) the mass--loading factor

937: \citep[see Equation 20 of][]{finlator07}.

938: 

939: 

940: In this model, the observed trends between metallicity and environment

941: would require either higher enrichment of the gas flowing into galaxies in

942: overdense regions and/or lower mass--loading factors in high--density

943: environments. There are many environment--dependent physical mechanisms

944: that could yield the former; for instance, supernova feedback from evolved

945: stars associated with intragroup or intracluster light will directly dump

946: metals (in particular, oxygen) into the IGM about galaxies in the

947: highest--density environments. In addition, galaxy mergers, harassment, and

948: ram--pressure stripping in groups and clusters can strip enriched gas from

949: member and infalling galaxies, thereby inflating the metal content of the

950: local gas reservoir relative to the gas supply of roughly primordial

951: composition that feeds galaxies in the field \citep[e.g.,][]{gunn72,

952:   moore96, hester06, gnedin98}.

953: 

954: 

955: Stripping of gas from cluster members could also contribute to a higher

956: gas--phase metallicity in extreme environments in a secondary manner. That

957: is, ram--pressure stripping could remove the outer portion (and therefore

958: most metal--poor segment) of a galaxy's gas halo. Since the mixing time

959: (assumed to be the dynamical time) for a disk galaxy is on the order of the

960: cluster crossing time ($\sim 2$ Gyr), if not stripped this metal--poor gas

961: would become effectively mixed, thereby reducing the mean metallicity

962: within the central $\sim 5-10$ kpc (the region sampled by an SDSS fiber).

963: 

964: 

965: In the most extreme environments, pressure from the intercluster medium

966: (ICM) could potentially resist such stripping

967: \citep[e.g.,][]{babul92}. However, hydrodynamical simulations have found

968: that the net effect of thermal pressure and ram--pressure stripping on a

969: cluster member still results in gas being removed from the galaxy,

970: contributing to the ICM \citep{murakami99}. On the other hand, numerical

971: and analytical modeling of feedback in isolated galaxies shows that the

972: ejection of metals from a galaxy's ISM is more likely to occur in regions

973: of lower pressure \citep[e.g.,][]{silich01, maclow99}. Thus, thermal

974: pressure (and its impact on the ability to drive an outflow) could account

975: for the relative decrease in metallicity for galaxies in low--density

976: environs.

977: 

978: 

979: 

980: Alternatively, the metallicity--environment relations presented in this

981: work could also result from variations in the mass--loading factor with

982: local galaxy density. While the mass--loading factor is, in principal, a

983: quantity that can be directly observed \citep[e.g.,][]{morganti05},

984: detailed radio measurements of a galaxy's gas mass are required. Since we

985: lack the required observations within the SDSS data set, we instead utilize

986: the SDSS spectroscopic data to look for signatures of variation in outflow

987: velocity with environment at $z \sim 0.1$. Although, wind speed does not

988: necessarily provide any information about the amount of mass expelled from

989: a galaxy, a significant variation in outflow velocity with environment

990: could be an indication that the net accretion rate (relative to the SFR) is

991: driving the observed metallicity--environment relations. From co--adding

992: two sets of spectra including several hundred strongly star--forming (${\rm

993:   H}\alpha\ {\rm equivalent\ width} > 30$\AA), massive (${\rm M}_{*} >

994: 10^{10}\ {\rm M}_{\sun}$) galaxies, we find no significant variation in the

995: Na D absorption profile between extreme (low--density and high--density)

996: environments. Admittedly, our analysis is limited to the most highly

997: star--forming galaxies, given the low resolution $({\rm R} \sim 1800)$ of

998: the SDSS spectra.

999: 

1000: 

1001: Another point to consider when searching for physical sources of the strong

1002: relationship between environment and metallicity is that galaxies

1003: populating high--density regions today likely formed early in the first

1004: overdensities. Predictions of early galaxy enrichment

1005: \citep[e.g.,][]{schaye03, dave06} indicate that these overdensities of gas

1006: at high--$z$ would be the most enriched environments, naturally producing a

1007: metallicity--environment relation \citep[see also][]{oppenheimer06}.

1008: Within the model of \citet{finlator07}, however, the gas--phase metallicity

1009: in a galaxy at $z \sim 0.1$ is a product of the recent ($< \! 1$ Gyr)

1010: accretion and star--formation activity, rather than a result of the

1011: integrated star--formation history of the galaxy \citep[see

1012: also][]{dalcanton07}. So while galaxies in high--density environs in the

1013: local Universe generally formed early in cosmic time and in the early

1014: density peaks, metallicity--environment relations imprinted at $z \gtrsim

1015: 2$ would not necessarily persist to the present.

1016: 

1017: 

1018: 

1019: 

1020: \subsection{Comparison to Related Work}

1021: \label{sec_disc_2}

1022: 

1023: 

1024: As this paper was being completed, a parallel analysis of the relationship

1025: between metallicity and environment in the SDSS was presented by

1026: \citet{mouhcine07}. Using a very similar data set, drawn from SDSS DR4 and

1027: employing the metallicity measurements of \citet{tremonti04}, they find

1028: that the mass--metallicity relation depends weakly on local

1029: environment. When dividing our sample into discrete bins according to

1030: overdensity, we also find a relatively weak connection between metallicity

1031: and environment at fixed stellar mass (see Fig.\ \ref{mz_fig}); that is,

1032: when plotting the median mass--metallicity relation in discrete bins of

1033: overdensity, we find what appear to be only small variations among

1034: environment regimes, in close agreement with Figure 5 of

1035: \citet{mouhcine07}.

1036: 

1037: However, studying the relationships between galaxy properties and

1038: environment in this manner is far less sensitive than the techniques

1039: presented herein. Measurements of local galaxy density are inherently

1040: noisier than measures of other galaxy properties, including rest--frame

1041: color, luminosity, stellar mass, and metallicity. Thus, when dividing a

1042: sample by environment, any trends in the data set are smeared out by the

1043: significant correlation between neighboring bins. While \citet{mouhcine07}

1044: conclude that gas--phase oxygen abundance is only weakly dependent on

1045: environment, we have presented evidence to the contrary, showing that the

1046: metallicity--environment relation is roughly equal in strength to the

1047: color--density relation. Furthermore, we find metallicity has a

1048: relationship with environment that is separate from the color--density or

1049: stellar mass--density relations.

1050: 

1051: 

1052: In contrast to our work and that of \citet{mouhcine07}, which trace galaxy

1053: environments on $\sim 1$--$2 h^{-1}$ Mpc scales over the full SDSS galaxy

1054: population, the analyses of \citet{kewley06} and \citet{ellison08b} probe a

1055: far more limited range of environments, focusing on the metallicity of

1056: galaxy pairs in the local Universe. Focusing on smaller scales, they find

1057: that galaxy pairs at close (projected) separations ($\lesssim \! 30\

1058: h^{-1}$ kpc) are biased towards lower metallicities. This result is

1059: attributed to inflows of metal--poor gas during the merger or interaction

1060: process, an effect that is also found in simulations \citep{perez06}. The

1061: number of close (projected separations $< \! 100\ h^{-1}$ kpc) pairs,

1062: however, is $\lesssim \! 1\%$ in the SDSS sample \citep[see

1063: also][]{deng06}, and thus such systems cannot be a significant contribution

1064: to the scatter in the mass--metallicity relation. While metallicity may be

1065: lower in close pairs, the dominant metallicity--environment relation moves

1066: towards higher metal enrichment in high--density environments.

1067: 

1068: 

1069: 

1070: 

1071: 

1072: 

1073: 

1074: 

1075: 

1076: \section{Summary and Conclusions}

1077: \label{sec_sum}

1078: 

1079: Using the measurements of gas--phase oxygen abundance from

1080: \citet{tremonti04} and local galaxy environment from \citet{cooper08}, we

1081: study the relationship between metallicity and environment in a sample of

1082: star--forming galaxies drawn from the SDSS data set. Our principal results

1083: are as follows.

1084: 

1085: 

1086: \begin{enumerate}

1087: 

1088: \item We find a strong metallicity--density relation (see Fig.\

1089:   \ref{metal_fig}b) in the local Universe such that more metal--rich

1090:   galaxies favor regions of higher galaxy overdensity. This relationship

1091:   between metallicity and environment follows (with comparable or greater

1092:   strength) that seen between environment and other fundamental properties

1093:   such as color, luminosity, SFR, or stellar mass.

1094: 

1095: \item After removing the mean color--luminosity--environment relation from

1096:   the SDSS data set, we find a significant residual relationship between

1097:   environment and metallicity (see Fig.\ \ref{resid_fig2}), suggesting that

1098:   metallicity has a relationship with environment separate from that

1099:   observed with color and luminosity (or with stellar mass). 

1100: 

1101: \item The residual metallicity--environment trend is largely driven by

1102:   galaxies in high--density regions such as groups and clusters, where the

1103:   local environment may be responsible for impacting the feedback and/or

1104:   gas accretion relative to galaxies of like stellar mass in lower--density

1105:   regions.

1106: 

1107: \item A non--negligible portion (at least $15\%$) of the scatter in the

1108:   mass--metallicity relation is correlated with local environment. 

1109: 

1110: 

1111: 

1112: \end{enumerate}

1113: 

1114: 

1115: 

1116: 

1117: 

1118: %%%%%%%%%%%%%%%%%%%%%%%

1119: %%% Acknowledgments %%%

1120: %%%%%%%%%%%%%%%%%%%%%%%

1121: 

1122: 

1123: \acknowledgments Support for this work was provided by NASA through the

1124: Spitzer Space Telescope Fellowship Program. C.A.T.\ acknowledges support by

1125: NASA through Hubble Fellowship grant HST--HF--01192.01, awarded by the

1126: Space Telescope Science Institute, which is operated by AURA Inc.\ under

1127: NASA contract NAS 5--26555. M.C.C.\ would like to thank Michael Blanton,

1128: David Hogg, and Renbin Yan for their assistance in utilizing the NYU--VAGC

1129: data products. This work benefited greatly from conversations with Kristian

1130: Finlator, Romeel Dav\'{e}, Ben Weiner, and Dennis Zaritsky.

1131: 

1132: Funding for the SDSS has been provided by the Alfred P.\ Sloan Foundation,

1133: the Participating Institutions, the National Science Foundation, the

1134: U.S.\ Department of Energy, the National Aeronautics and Space

1135: Administration, the Japanese Monbukagakusho, the Max Planck Society, and

1136: the Higher Education Funding Council for England. The SDSS Web Site is

1137: http://www.sdss.org/.

1138: 

1139: The SDSS is managed by the Astrophysical Research Consortium for the

1140: Participating Institutions. The Participating Institutions are the American

1141: Museum of Natural History, Astrophysical Institute Potsdam, University of

1142: Basel, University of Cambridge, Case Western Reserve University, University

1143: of Chicago, Drexel University, Fermilab, the Institute for Advanced Study,

1144: the Japan Participation Group, Johns Hopkins University, the Joint

1145: Institute for Nuclear Astrophysics, the Kavli Institute for Particle

1146: Astrophysics and Cosmology, the Korean Scientist Group, the Chinese Academy

1147: of Sciences (LAMOST), Los Alamos National Laboratory, the

1148: Max--Planck--Institute for Astronomy (MPIA), the Max--Planck--Institute for

1149: Astrophysics (MPA), New Mexico State University, Ohio State University,

1150: University of Pittsburgh, University of Portsmouth, Princeton University,

1151: the United States Naval Observatory, and the University of Washington.

1152: 

1153: 

1154: %%%%%%%%%%%%%%%%%%%%

1155: %%% Bibliography %%%

1156: %%%%%%%%%%%%%%%%%%%%

1157: %\bibliographystyle{apj}

1158: %\bibliography{apj-jour,environ_metal_refs}

1159: 

1160: \begin{thebibliography}{92}

1161: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1162: 

1163: \bibitem[{{Adelberger} {et~al.}(2005){Adelberger}, {Shapley}, {Steidel},

1164:   {Pettini}, {Erb}, \& {Reddy}}]{adelberger05}

1165: {Adelberger}, K.~L., {Shapley}, A.~E., {Steidel}, C.~C., {Pettini}, M., {Erb},

1166:   D.~K., \& {Reddy}, N.~A. 2005, \apj, 629, 636

1167: 

1168: \bibitem[{{Adelman-McCarthy} {et~al.}(2006)}]{adelman06}

1169: {Adelman-McCarthy}, J.~K. {et~al.} 2006, \apjs, 162, 38

1170: 

1171: \bibitem[{{Aguirre} {et~al.}(2005){Aguirre}, {Schaye}, {Hernquist}, {Kay},

1172:   {Springel}, \& {Theuns}}]{aguirre05}

1173: {Aguirre}, A., {Schaye}, J., {Hernquist}, L., {Kay}, S., {Springel}, V., \&

1174:   {Theuns}, T. 2005, \apjl, 620, L13

1175: 

1176: \bibitem[{{Babul} \& {Rees}(1992)}]{babul92}

1177: {Babul}, A. \& {Rees}, M.~J. 1992, \mnras, 255, 346

1178: 

1179: \bibitem[{{Balogh} {et~al.}(2004{\natexlab{a}})}]{balogh04a}

1180: {Balogh}, M. {et~al.} 2004{\natexlab{a}}, \mnras, 348, 1355

1181: 

1182: \bibitem[{{Balogh} {et~al.}(2004{\natexlab{b}}){Balogh}, {Baldry}, {Nichol},

1183:   {Miller}, {Bower}, \& {Glazebrook}}]{balogh04b}

1184: {Balogh}, M.~L., {Baldry}, I.~K., {Nichol}, R., {Miller}, C., {Bower}, R., \&

1185:   {Glazebrook}, K. 2004{\natexlab{b}}, \apjl, 615, L101

1186: 

1187: \bibitem[{{Balogh} {et~al.}(1998){Balogh}, {Schade}, {Morris}, {Yee},

1188:   {Carlberg}, \& {Ellingson}}]{balogh98}

1189: {Balogh}, M.~L., {Schade}, D., {Morris}, S.~L., {Yee}, H.~K.~C., {Carlberg},

1190:   R.~G., \& {Ellingson}, E. 1998, \apjl, 504, L75+

1191: 

1192: \bibitem[{{Blanton} {et~al.}(2005{\natexlab{a}}){Blanton}, {Eisenstein},

1193:   {Hogg}, {Schlegel}, \& {Brinkmann}}]{blanton05a}

1194: {Blanton}, M.~R., {Eisenstein}, D., {Hogg}, D.~W., {Schlegel}, D.~J., \&

1195:   {Brinkmann}, J. 2005{\natexlab{a}}, \apj, 629, 143

1196: 

1197: \bibitem[{{Blanton} \& {Roweis}(2007)}]{blanton07}

1198: {Blanton}, M.~R. \& {Roweis}, S. 2007, \aj, 133, 734

1199: 

1200: \bibitem[{{Blanton} {et~al.}(2003{\natexlab{a}})}]{blanton03b}

1201: {Blanton}, M.~R. {et~al.} 2003{\natexlab{a}}, \aj, 125, 2348

1202: 

1203: \bibitem[{{Blanton} {et~al.}(2003{\natexlab{b}})}]{blanton03a}

1204: ---. 2003{\natexlab{b}}, \apj, 594, 186

1205: 

1206: \bibitem[{{Blanton} {et~al.}(2005{\natexlab{b}})}]{blanton05b}

1207: ---. 2005{\natexlab{b}}, \aj, 129, 2562

1208: 

1209: \bibitem[{{Brinchmann} {et~al.}(2004){Brinchmann}, {Charlot}, {White},

1210:   {Tremonti}, {Kauffmann}, {Heckman}, \& {Brinkmann}}]{brinchmann04}

1211: {Brinchmann}, J., {Charlot}, S., {White}, S.~D.~M., {Tremonti}, C.,

1212:   {Kauffmann}, G., {Heckman}, T., \& {Brinkmann}, J. 2004, \mnras, 351, 1151

1213: 

1214: \bibitem[{{Brodie} \& {Huchra}(1991)}]{brodie91}

1215: {Brodie}, J.~P. \& {Huchra}, J.~P. 1991, \apj, 379, 157

1216: 

1217: \bibitem[{{Brooks} {et~al.}(2007){Brooks}, {Governato}, {Booth}, {Willman},

1218:   {Gardner}, {Wadsley}, {Stinson}, \& {Quinn}}]{brooks07}

1219: {Brooks}, A.~M., {Governato}, F., {Booth}, C.~M., {Willman}, B., {Gardner},

1220:   J.~P., {Wadsley}, J., {Stinson}, G., \& {Quinn}, T. 2007, \apjl, 655, L17

1221: 

1222: \bibitem[{{Bruzual} \& {Charlot}(2003)}]{bc03}

1223: {Bruzual}, G. \& {Charlot}, S. 2003, \mnras, 344, 1000

1224: 

1225: \bibitem[{{Capak} {et~al.}(2007){Capak}, {Abraham}, {Ellis}, {Mobasher},

1226:   {Scoville}, {Sheth}, \& {Koekemoer}}]{capak07}

1227: {Capak}, P., {Abraham}, R.~G., {Ellis}, R.~S., {Mobasher}, B., {Scoville}, N.,

1228:   {Sheth}, K., \& {Koekemoer}, A. 2007, \apjs, 172, 284

1229: 

1230: \bibitem[{{Cavaliere} {et~al.}(1992){Cavaliere}, {Colafrancesco}, \&

1231:   {Menci}}]{cavaliere92}

1232: {Cavaliere}, A., {Colafrancesco}, S., \& {Menci}, N. 1992, \apj, 392, 41

1233: 

1234: \bibitem[{{Chabrier}(2003)}]{charbrier03}

1235: {Chabrier}, G. 2003, \pasp, 115, 763

1236: 

1237: \bibitem[{{Charlot} \& {Longhetti}(2001)}]{charlot01}

1238: {Charlot}, S. \& {Longhetti}, M. 2001, \mnras, 323, 887

1239: 

1240: \bibitem[{{Cole}(1991)}]{cole91}

1241: {Cole}, S. 1991, \apj, 367, 45

1242: 

1243: \bibitem[{{Contini} {et~al.}(2002){Contini}, {Treyer}, {Sullivan}, \&

1244:   {Ellis}}]{contini02}

1245: {Contini}, T., {Treyer}, M.~A., {Sullivan}, M., \& {Ellis}, R.~S. 2002, \mnras,

1246:   330, 75

1247: 

1248: \bibitem[{{Cooper} {et~al.}(2005){Cooper}, {Newman}, {Madgwick}, {Gerke},

1249:   {Yan}, \& {Davis}}]{cooper05}

1250: {Cooper}, M.~C., {Newman}, J.~A., {Madgwick}, D.~S., {Gerke}, B.~F., {Yan}, R.,

1251:   \& {Davis}, M. 2005, \apj, 634, 833

1252: 

1253: \bibitem[{{Cooper} {et~al.}(2006)}]{cooper06}

1254: {Cooper}, M.~C. {et~al.} 2006, \mnras, 370, 198

1255: 

1256: \bibitem[{{Cooper} {et~al.}(2007)}]{cooper07}

1257: ---. 2007, \mnras, 376, 1445

1258: 

1259: \bibitem[{{Cooper} {et~al.}(2008)}]{cooper08}

1260: ---. 2008, \mnras, 383, 1058

1261: 

1262: \bibitem[{{Cucciati} {et~al.}(2006)}]{cucciati06}

1263: {Cucciati}, O. {et~al.} 2006, \aap, 458, 39

1264: 

1265: \bibitem[{{Dalcanton}(2007)}]{dalcanton07}

1266: {Dalcanton}, J.~J. 2007, \apj, 658, 941

1267: 

1268: \bibitem[{{Dav{\'e}} {et~al.}(2006){Dav{\'e}}, {Finlator}, \&

1269:   {Oppenheimer}}]{dave06}

1270: {Dav{\'e}}, R., {Finlator}, K., \& {Oppenheimer}, B.~D. 2006, \mnras, 370, 273

1271: 

1272: \bibitem[{{Davis} \& {Geller}(1976)}]{davis76}

1273: {Davis}, M. \& {Geller}, M.~J. 1976, \apj, 208, 13

1274: 

1275: \bibitem[{{Dekel} \& {Silk}(1986)}]{dekel86}

1276: {Dekel}, A. \& {Silk}, J. 1986, \apj, 303, 39

1277: 

1278: \bibitem[{{Dekel} \& {Woo}(2003)}]{dekel03}

1279: {Dekel}, A. \& {Woo}, J. 2003, \mnras, 344, 1131

1280: 

1281: \bibitem[{{Deng} {et~al.}(2006){Deng}, {Chen}, {Wu}, {Luo}, \& {He}}]{deng06}

1282: {Deng}, X.-F., {Chen}, Y.-Q., {Wu}, P., {Luo}, C.-H., \& {He}, J.-Z. 2006,

1283:   Chinese Journal of Astronomy and Astrophysics, 6, 411

1284: 

1285: \bibitem[{{Domainko} {et~al.}(2004){Domainko}, {Gitti}, {Schindler}, \&

1286:   {Kapferer}}]{domainko04}

1287: {Domainko}, W., {Gitti}, M., {Schindler}, S., \& {Kapferer}, W. 2004, \aap,

1288:   425, L21

1289: 

1290: \bibitem[{{Dressler}(1980)}]{dressler80}

1291: {Dressler}, A. 1980, \apj, 236, 351

1292: 

1293: \bibitem[{{Ellison} {et~al.}(2008{\natexlab{a}}){Ellison}, {Patton}, {Simard},

1294:   \& {McConnachie}}]{ellison08a}

1295: {Ellison}, S.~L., {Patton}, D.~R., {Simard}, L., \& {McConnachie}, A.~W.

1296:   2008{\natexlab{a}}, \apjl, 672, L107

1297: 

1298: \bibitem[{{Ellison} {et~al.}(2008{\natexlab{b}}){Ellison}, {Patton}, {Simard},

1299:   \& {McConnachie}}]{ellison08b}

1300: ---. 2008{\natexlab{b}}, \apj, \apj, accepted

1301: 

1302: \bibitem[{{Erb} {et~al.}(2006){Erb}, {Shapley}, {Pettini}, {Steidel}, {Reddy},

1303:   \& {Adelberger}}]{erb06}

1304: {Erb}, D.~K., {Shapley}, A.~E., {Pettini}, M., {Steidel}, C.~C., {Reddy},

1305:   N.~A., \& {Adelberger}, K.~L. 2006, \apj, 644, 813

1306: 

1307: \bibitem[{{Finlator} \& {Dave}(2007)}]{finlator07}

1308: {Finlator}, K. \& {Dave}, R. 2007, ArXiv e-prints, 704

1309: 

1310: \bibitem[{{Frye} {et~al.}(2002){Frye}, {Broadhurst}, \&

1311:   {Ben{\'{\i}}tez}}]{frye02}

1312: {Frye}, B., {Broadhurst}, T., \& {Ben{\'{\i}}tez}, N. 2002, \apj, 568, 558

1313: 

1314: \bibitem[{{Gallazzi} {et~al.}(2008){Gallazzi}, {Brinchmann}, {Charlot}, \&

1315:   {White}}]{gallazzi08}

1316: {Gallazzi}, A., {Brinchmann}, J., {Charlot}, S., \& {White}, S.~D.~M. 2008,

1317:   \mnras, 383, 1439

1318: 

1319: \bibitem[{{Garnett}(2002)}]{garnett02}

1320: {Garnett}, D.~R. 2002, \apj, 581, 1019

1321: 

1322: \bibitem[{{Gnedin}(1998)}]{gnedin98}

1323: {Gnedin}, N.~Y. 1998, \mnras, 294, 407

1324: 

1325: \bibitem[{{Gunn} \& {Gott}(1972)}]{gunn72}

1326: {Gunn}, J.~E. \& {Gott}, J.~R.~I. 1972, \apj, 176, 1

1327: 

1328: \bibitem[{{Heckman}(1980)}]{heckman80}

1329: {Heckman}, T.~M. 1980, \aap, 87, 152

1330: 

1331: \bibitem[{{Hester}(2006)}]{hester06}

1332: {Hester}, J.~A. 2006, \apj, 647, 910

1333: 

1334: \bibitem[{{Hogg} {et~al.}(2003)}]{hogg03}

1335: {Hogg}, D.~W. {et~al.} 2003, \apjl, 585, L5

1336: 

1337: \bibitem[{{Kauffmann} {et~al.}(2004){Kauffmann}, {White}, {Heckman},

1338:   {M{\'e}nard}, {Brinchmann}, {Charlot}, {Tremonti}, \&

1339:   {Brinkmann}}]{kauffmann04}

1340: {Kauffmann}, G., {White}, S.~D.~M., {Heckman}, T.~M., {M{\'e}nard}, B.,

1341:   {Brinchmann}, J., {Charlot}, S., {Tremonti}, C., \& {Brinkmann}, J. 2004,

1342:   \mnras, 353, 713

1343: 

1344: \bibitem[{{Kauffmann} {et~al.}(2003{\natexlab{a}})}]{kauffmann03a}

1345: {Kauffmann}, G. {et~al.} 2003{\natexlab{a}}, \mnras, 341, 33

1346: 

1347: \bibitem[{{Kauffmann} {et~al.}(2003{\natexlab{b}})}]{kauffmann03b}

1348: ---. 2003{\natexlab{b}}, \mnras, 346, 1055

1349: 

1350: \bibitem[{{Kewley} {et~al.}(2006){Kewley}, {Geller}, \& {Barton}}]{kewley06}

1351: {Kewley}, L.~J., {Geller}, M.~J., \& {Barton}, E.~J. 2006, \aj, 131, 2004

1352: 

1353: \bibitem[{{Kobayashi} {et~al.}(2007){Kobayashi}, {Springel}, \&

1354:   {White}}]{kobayashi07}

1355: {Kobayashi}, C., {Springel}, V., \& {White}, S.~D.~M. 2007, \mnras, 376, 1465

1356: 

1357: \bibitem[{{Kobulnicky} {et~al.}(2003)}]{kobulnicky03}

1358: {Kobulnicky}, H.~A. {et~al.} 2003, \apj, 599, 1006

1359: 

1360: \bibitem[{{Lamareille} {et~al.}(2004){Lamareille}, {Mouhcine}, {Contini},

1361:   {Lewis}, \& {Maddox}}]{lamareille04}

1362: {Lamareille}, F., {Mouhcine}, M., {Contini}, T., {Lewis}, I., \& {Maddox}, S.

1363:   2004, \mnras, 350, 396

1364: 

1365: \bibitem[{{Larson}(1974)}]{larson74}

1366: {Larson}, R.~B. 1974, \mnras, 169, 229

1367: 

1368: \bibitem[{{Lee} {et~al.}(2003){Lee}, {McCall}, \& {Richer}}]{lee03}

1369: {Lee}, H., {McCall}, M.~L., \& {Richer}, M.~G. 2003, \aj, 125, 2975

1370: 

1371: \bibitem[{{Lee} {et~al.}(2006){Lee}, {Skillman}, {Cannon}, {Jackson}, {Gehrz},

1372:   {Polomski}, \& {Woodward}}]{lee06}

1373: {Lee}, H., {Skillman}, E.~D., {Cannon}, J.~M., {Jackson}, D.~C., {Gehrz},

1374:   R.~D., {Polomski}, E.~F., \& {Woodward}, C.~E. 2006, \apj, 647, 970

1375: 

1376: \bibitem[{{Lehnert} \& {Heckman}(1996)}]{lehnert96}

1377: {Lehnert}, M.~D. \& {Heckman}, T.~M. 1996, \apj, 472, 546

1378: 

1379: \bibitem[{{Lequeux} {et~al.}(1979){Lequeux}, {Peimbert}, {Rayo}, {Serrano}, \&

1380:   {Torres-Peimbert}}]{lequeux79}

1381: {Lequeux}, J., {Peimbert}, M., {Rayo}, J.~F., {Serrano}, A., \&

1382:   {Torres-Peimbert}, S. 1979, \aap, 80, 155

1383: 

1384: \bibitem[{{Mac Low} \& {Ferrara}(1999)}]{maclow99}

1385: {Mac Low}, M.-M. \& {Ferrara}, A. 1999, \apj, 513, 142

1386: 

1387: \bibitem[{{Martin}(2005)}]{martin05}

1388: {Martin}, C.~L. 2005, \apj, 621, 227

1389: 

1390: \bibitem[{{Miller} {et~al.}(2003){Miller}, {Nichol}, {G{\'o}mez}, {Hopkins}, \&

1391:   {Bernardi}}]{miller03}

1392: {Miller}, C.~J., {Nichol}, R.~C., {G{\'o}mez}, P.~L., {Hopkins}, A.~M., \&

1393:   {Bernardi}, M. 2003, \apj, 597, 142

1394: 

1395: \bibitem[{{Montero-Dorta} {et~al.}(2008)}]{montero08}

1396: {Montero-Dorta}, A.~D. {et~al.} 2008, ArXiv e-prints, 801

1397: 

1398: \bibitem[{{Moore} {et~al.}(1996){Moore}, {Katz}, {Lake}, {Dressler}, \&

1399:   {Oemler}}]{moore96}

1400: {Moore}, B., {Katz}, N., {Lake}, G., {Dressler}, A., \& {Oemler}, A. 1996,

1401:   \nat, 379, 613

1402: 

1403: \bibitem[{{Morganti} {et~al.}(2005){Morganti}, {Tadhunter}, \&

1404:   {Oosterloo}}]{morganti05}

1405: {Morganti}, R., {Tadhunter}, C.~N., \& {Oosterloo}, T.~A. 2005, \aap, 444, L9

1406: 

1407: \bibitem[{{Mouhcine} {et~al.}(2007){Mouhcine}, {Baldry}, \&

1408:   {Bamford}}]{mouhcine07}

1409: {Mouhcine}, M., {Baldry}, I.~K., \& {Bamford}, S.~P. 2007, \mnras, 382, 801

1410: 

1411: \bibitem[{{Murakami} \& {Babul}(1999)}]{murakami99}

1412: {Murakami}, I. \& {Babul}, A. 1999, \mnras, 309, 161

1413: 

1414: \bibitem[{{Murray} {et~al.}(2005){Murray}, {Quataert}, \& {Thompson}}]{mqt05}

1415: {Murray}, N., {Quataert}, E., \& {Thompson}, T.~A. 2005, \apj, 618, 569

1416: 

1417: \bibitem[{{Oke} \& {Gunn}(1983)}]{oke83}

1418: {Oke}, J.~B. \& {Gunn}, J.~E. 1983, \apj, 266, 713

1419: 

1420: \bibitem[{{Oppenheimer} \& {Dav{\'e}}(2006)}]{oppenheimer06}

1421: {Oppenheimer}, B.~D. \& {Dav{\'e}}, R. 2006, \mnras, 373, 1265

1422: 

1423: \bibitem[{{Peeples} {et~al.}(2008){Peeples}, {Pogge}, \& {Stanek}}]{peeples08}

1424: {Peeples}, M.~S., {Pogge}, R.~W., \& {Stanek}, K.~Z. 2008, ArXiv e-prints, 804

1425: 

1426: \bibitem[{{Perez} {et~al.}(2006){Perez}, {Tissera}, {Scannapieco}, {Lambas}, \&

1427:   {de Rossi}}]{perez06}

1428: {Perez}, M.~J., {Tissera}, P.~B., {Scannapieco}, C., {Lambas}, D.~G., \& {de

1429:   Rossi}, M.~E. 2006, \aap, 459, 361

1430: 

1431: \bibitem[{{Pilyugin} \& {Ferrini}(2000)}]{pilyugin00}

1432: {Pilyugin}, L.~S. \& {Ferrini}, F. 2000, \aap, 354, 874

1433: 

1434: \bibitem[{{Rupke} {et~al.}(2005){Rupke}, {Veilleux}, \& {Sanders}}]{rupke05}

1435: {Rupke}, D.~S., {Veilleux}, S., \& {Sanders}, D.~B. 2005, \apjs, 160, 115

1436: 

1437: \bibitem[{{Salim} {et~al.}(2007)}]{salim07}

1438: {Salim}, S. {et~al.} 2007, \apjs, 173, 267

1439: 

1440: \bibitem[{{Schaye} {et~al.}(2003){Schaye}, {Aguirre}, {Kim}, {Theuns}, {Rauch},

1441:   \& {Sargent}}]{schaye03}

1442: {Schaye}, J., {Aguirre}, A., {Kim}, T.-S., {Theuns}, T., {Rauch}, M., \&

1443:   {Sargent}, W.~L.~W. 2003, \apj, 596, 768

1444: 

1445: \bibitem[{{Shapley} {et~al.}(2005){Shapley}, {Coil}, {Ma}, \&

1446:   {Bundy}}]{shapley05}

1447: {Shapley}, A.~E., {Coil}, A.~L., {Ma}, C.-P., \& {Bundy}, K. 2005, \apj, 635,

1448:   1006

1449: 

1450: \bibitem[{{Shapley} {et~al.}(2003){Shapley}, {Steidel}, {Pettini}, \&

1451:   {Adelberger}}]{shapley03}

1452: {Shapley}, A.~E., {Steidel}, C.~C., {Pettini}, M., \& {Adelberger}, K.~L. 2003,

1453:   \apj, 588, 65

1454: 

1455: \bibitem[{{Silich} \& {Tenorio-Tagle}(2001)}]{silich01}

1456: {Silich}, S. \& {Tenorio-Tagle}, G. 2001, \apj, 552, 91

1457: 

1458: \bibitem[{{Skillman} {et~al.}(1989){Skillman}, {Kennicutt}, \&

1459:   {Hodge}}]{skillman89}

1460: {Skillman}, E.~D., {Kennicutt}, R.~C., \& {Hodge}, P.~W. 1989, \apj, 347, 875

1461: 

1462: \bibitem[{{Smith} {et~al.}(2005){Smith}, {Treu}, {Ellis}, {Moran}, \&

1463:   {Dressler}}]{smith05}

1464: {Smith}, G.~P., {Treu}, T., {Ellis}, R.~S., {Moran}, S.~M., \& {Dressler}, A.

1465:   2005, \apj, 620, 78

1466: 

1467: \bibitem[{{Sorrentino} {et~al.}(2006){Sorrentino}, {Radovich}, \&

1468:   {Rifatto}}]{sorrentino06}

1469: {Sorrentino}, G., {Radovich}, M., \& {Rifatto}, A. 2006, \aap, 451, 809

1470: 

1471: \bibitem[{{Springel} \& {Hernquist}(2003{\natexlab{a}})}]{springel03a}

1472: {Springel}, V. \& {Hernquist}, L. 2003{\natexlab{a}}, \mnras, 339, 289

1473: 

1474: \bibitem[{{Springel} \& {Hernquist}(2003{\natexlab{b}})}]{springel03b}

1475: ---. 2003{\natexlab{b}}, \mnras, 339, 312

1476: 

1477: \bibitem[{{Stoughton} {et~al.}(2002)}]{stoughton02}

1478: {Stoughton}, C. {et~al.} 2002, \aj, 123, 485

1479: 

1480: \bibitem[{{Tremonti} {et~al.}(2004)}]{tremonti04}

1481: {Tremonti}, C.~A. {et~al.} 2004, \apj, 613, 898

1482: 

1483: \bibitem[{{Veilleux} {et~al.}(2005){Veilleux}, {Cecil}, \&

1484:   {Bland-Hawthorn}}]{veilleux05}

1485: {Veilleux}, S., {Cecil}, G., \& {Bland-Hawthorn}, J. 2005, \araa, 43, 769

1486: 

1487: \bibitem[{{Vila-Costas} \& {Edmunds}(1992)}]{vila92}

1488: {Vila-Costas}, M.~B. \& {Edmunds}, M.~G. 1992, \mnras, 259, 121

1489: 

1490: \bibitem[{{Weiner} {et~al.}(2008)}]{weiner08}

1491: {Weiner}, B.~J. {et~al.} 2008, \apj, in prep

1492: 

1493: \bibitem[{{York} {et~al.}(2000)}]{york00}

1494: {York}, D.~G. {et~al.} 2000, \aj, 120, 1579

1495: 

1496: \bibitem[{{Zaritsky}(1993)}]{zaritsky93}

1497: {Zaritsky}, D. 1993, \pasp, 105, 1006

1498: 

1499: \bibitem[{{Zaritsky} {et~al.}(1994){Zaritsky}, {Kennicutt}, \&

1500:   {Huchra}}]{zaritsky94}

1501: {Zaritsky}, D., {Kennicutt}, Jr., R.~C., \& {Huchra}, J.~P. 1994, \apj, 420, 87

1502: 

1503: \end{thebibliography}

1504: 

1505: 

1506: \end{document}