1: \documentclass{emulateapj}

2: 

3: \lefthead{VAN DER WEL ET AL.}

4: \righthead{EVOLUTION OF CLUSTER AND FIELD MDR}

5: \slugcomment{To Appear in ApJ, 670, 206}

6: 

7: \begin{document}

8: 

9: \title{The Evolution of the Field and Cluster Morphology-Density

10: Relation for Mass-Selected Samples of Galaxies}

11: 

12: 

13: \author{A. van der Wel\altaffilmark{1}, B. P. Holden\altaffilmark{2},

14:   M. Franx\altaffilmark{3}, G.D. Illingworth\altaffilmark{2},

15:   M. P. Postman\altaffilmark{4}, D. D. Kelson\altaffilmark{5}, \\

16:   I. Labb\'e\altaffilmark{5}, Wuyts, S.{3},

17:   J. P. Blakeslee\altaffilmark{6}, H. C. Ford\altaffilmark{1}}

18: 

19: \altaffiltext{1}{Department of Physics and Astronomy,

20: Johns Hopkins University, 3400 North Charles Street, Baltimore, MD

21: 21218; wel@pha.jhu.edu}

22: \altaffiltext{2}{University of California Observatories/Lick Observatory, University of California, Santa Cruz, CA 95064}

23: \altaffiltext{3}{Leiden Observatory, Leiden University, P.O.Box 9513, NL-2300 AA Leiden, Netherlands}

24: \altaffiltext{4}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218}

25: \altaffiltext{5}{Carnegie Observatories, Carnegie Institution of Washington, 813 Santa Barbara Street, Pasadena, CA 91101}

26: \altaffiltext{6}{Department of Physics and Astronomy, Washington State University, Pullman, WA 99163-2814}

27: \altaffiltext{7}{Based

28: on observations with the \textit{Hubble Space Telescope}, obtained at

29: the Space Telescope Science Institute, which is operated by AURA,

30: Inc., under NASA contract NAS5-26555, and observations made with the

31: \textit{Spitzer Space Telescope}, which is operated by the Jet

32: Propulsion Laboratory, California Institute of Technology, under NASA

33: contract 1407.}

34: 

35: \begin{abstract}

36:   The Sloan Digital Sky Survey (SDSS) and photometric/spectroscopic

37:   surveys in the GOODS-South field (the Chandra Deep Field-South,

38:   CDF-S) are used to construct volume-limited, stellar mass-selected

39:   samples of galaxies at redshifts $0<z<1$.  The CDF-S sample at

40:   $0.6<z<1.0$ contains 207 galaxies complete down to $M=4\times

41:   10^{10}~M_{\odot}$ (for a ``diet'' Salpeter initial mass function),

42:   corresponding to a luminosity limit for red galaxies of

43:   $M_{\rm{B}}=-20.1$.  The SDSS sample at $0.020<z<0.045$ contains

44:   2003 galaxies down to the same mass limit, which corresponds to

45:   $M_{\rm{B}}=-19.3$ for red galaxies.  Morphologies are determined

46:   with an automated method, using the S\'ersic parameter $n$ and a

47:   measure of the residual from the model fits, called ``bumpiness,''

48:   to distinguish different morphologies.  These classifications are

49:   verified with visual classifications.  In agreement with previous

50:   studies, $65\%-70\%$ of the galaxies are located on the red

51:   sequence, both at $z\sim 0.03$ and at $z\sim 0.8$. Similarly,

52:   $65\%-70\%$ of the galaxies have $n>2.5$. The fraction of E+S0

53:   galaxies is $43\%\pm 3\%$ at $z\sim 0.03$ and $48\%\pm 7\%$ at

54:   $z\sim 0.8$; i.e., it has not changed significantly since $z\sim

55:   0.8$.  When combined with recent results for cluster galaxies in the

56:   same redshift range, we find that the morphology-density relation

57:   for galaxies more massive than $0.5M^*$ has remained constant since

58:   at least $z\sim 0.8$. This implies that galaxies evolve in mass,

59:   morphology, and density such that the morphology-density relation

60:   does not change.  In particular, the decline of star formation

61:   activity and the accompanying increase in the stellar mass density

62:   of red galaxies since $z\sim 1$ must happen without large changes in

63:   the early-type galaxy fraction in a given environment.

64: \end{abstract}

65: 

66: \keywords{galaxies: clusters: general---galaxies: elliptical and lenticular, cD---galaxies: evolution---galaxies: formation---galaxies: fundamental parameters---galaxies: general---galaxies: photometry}

67: 

68: \section{INTRODUCTION}

69: 

70: The morphology-density relation \citep[MDR;][]{dressler80}, i.e., the

71: observation that high-density environments contain a higher fraction

72: of early-type galaxies than low-density environments, provides a clue

73: about the formation and evolution of galaxies, and suggests that

74: environment plays an important role in shaping the galaxy population.

75: The interpretation of the MDR, however, is not straightforward because

76: of additional correlations between morphology, galaxy mass, color,

77: star formation history, and metallicity; all quantities that depend on

78: environment

79: \citep[e.g.,][]{hogg03,kauffmann03a,kauffmann04,blanton05a,baldry06}.

80: There are strong indications that the stellar mass of a galaxy,

81: independently of the environment, determines the color

82: \citep{baldry06} and concentration \citep{kauffmann04} of a galaxy. In

83: addition, the environment plays a role in star formation activity

84: \citep{kauffmann04}.

85: 

86: It is a matter of debate whether the fate of a galaxy is ultimately

87: determined by its mass or its environment. Studies of galaxy

88: properties such as those mentioned above at higher redshifts help to

89: answer those questions, and there is abundant evidence for both

90: scenarios. The massive galaxy populations of $z\sim 1$ clusters appear

91: to be fully assembled and passively evolving

92: \citep[e.g.,][]{depropris07}, whereas the galaxy population as a whole

93: undergoes significant evolution between $z=1$ and the present, both in

94: terms of mass and color \citep[e.g.,][]{bell04b} and in terms of star

95: formation \citep[e.g.,][]{lefloch05}. This points to a strong relation

96: between formation epoch and environment. On the other hand, high-mass

97: galaxies have had lower specific star formation rates than low-mass

98: galaxies since at least $z\sim 2$ \citep[e.g.,][]{juneau05,noeske07},

99: and there is only a very modest age difference between field and

100: cluster early-type galaxies with masses $M>10^{11}~M_{\odot}$ at

101: $z\sim 1$ \citep{vandokkum07}, with possibly a stronger age dependence

102: on mass than environment \citep{vanderwel05,treu05b}. All these

103: results, at first sight contradictory, will have to be reconciled with

104: each other.

105: 

106: The MDR, and its evolution with redshift, plays a crucial role in our

107: understanding of the effect of the environment on galaxy evolution,

108: and significant evolution of the early-type galaxy fraction has been

109: found between $z\sim 1$ and the present

110: \citep[e.g.,][]{dressler97,smith05,postman05}.  These studies are

111: based on samples that are selected by luminosity, and to ensure that

112: the $z\sim 1$ samples contain the progenitors of the galaxies in the

113: local samples, the luminosity limit for samples at different redshifts

114: is corrected for evolution. It is assumed that the evolution is the

115: same for all galaxies, which might not be the case. In addition,

116: luminosity is very sensitive to bursts of star formation that are

117: likely far more prevalent at $z\sim 1$ than in the local

118: universe. Therefore, it is worthwhile to adopt an alternative

119: selection method that is less sensitive to potentially rapidly

120: changing galaxy properties such as the star formation rate.  Stellar

121: mass is a quantity that changes less rapidly than luminosity, such

122: that it might be a more suitable tracer of the $z\sim 1$ progenitors

123: of local galaxies.  In addition, it is more straightforward to compare

124: stellar masses to model predictions.  Obviously, mass is not an ideal

125: tracer of the evolving galaxy population, just like luminosity, as it

126: increases through star formation and mergers. Still, in \citet[][,

127: hereafter Paper I]{holden07} we show that there is no discernible

128: evolution between $z=0.83$ and the present in the morphological mix of

129: galaxies that are more massive than $4\times 10^{10}~M_{\odot}$

130: ($\sim$25\% of the mass of a typical cluster galaxy, $\sim$50\% of the

131: mass of a typical field galaxy).  The evolution in the early-type

132: fraction seen in luminosity-selected samples of cluster galaxies is

133: due to the relative increase in the number of blue, star-forming,

134: low-mass galaxies.

135: 

136: Given the abundant evidence for significant evolution since $z\sim 1$

137: in the galaxy population outside clusters, it is important to measure

138: the early-type galaxy fraction at lower densities.  It is not

139: inconceivable that the early-type galaxy fraction does evolve strongly

140: in those environments because of the strong evolution of the galaxy

141: population in the field and in groups. On the other hand, if

142: morphological evolution is primarily driven by environmental effects,

143: the morphological mix of the galaxy population in a given density

144: might not evolve. Several authors have studied the morphological mix

145: of galaxies at $z\sim 1$

146: \citep[e.g.,][]{bell04a,bundy05,capak07,abraham07}. A range of

147: techniques and sample selection methods is used in those

148: investigations, and it is not always clear how the high-$z$ results

149: compare to local studies and how sample selection affects the

150: interpretation. In particular, mass estimates and morphologies are

151: generally not obtained with the same method for local and distant

152: galaxy samples, such that systematic errors may contribute to the

153: observed evolution in, for example, the early-type fraction.

154: 

155: In this paper we construct mass-selected samples of low- and

156: high-redshift galaxies with morphologies classified in an internally

157: consistent manner, such that the morphological mix of the galaxy

158: population at low and high redshift can be directly compared both with

159: each other and with our recent results on samples of cluster galaxies

160: over the same redshift range (Paper I). In \S~2 we describe our

161: galaxy samples extracted from the Sloan Digital Sky Survey (SDSS) at

162: redshifts $0.020<z<0.045$ and the Chandra Deep Field-South (CDF-S) at

163: redshifts $0.6<z<1.0$ and determine galaxy masses and morphologies.

164: In \S~3 we present and discuss our results and their implications

165: for the evolution of the early-type galaxy fraction in our field

166: samples and, through comparison with cluster galaxies, how the MDR

167: evolves with redshift. Finally, in \S~4 we put our results in the

168: context of the evolving red sequence and the decline in the average

169: cosmic star formation rate density.  We adopt the standard

170: cosmological parameters, $(\Omega_{\rm{M}},~\Omega_{\Lambda},~h) =

171: (0.3,~0.7,~0.7)$.

172: 

173: 

174: \section{DATA}

175: 

176: First, we construct a complete, stellar-mass-selected, volume-limited

177: sample of galaxies at redshifts $0.02<z<0.045$ from the Sloan Digital

178: Sky Survey \citep[SDSS;][]{york00}, Data Release 5

179: \citep[DR5;][]{adelman07}.  Similarly, we construct a stellar

180: mass-selected, volume-limited sample at $0.6<z<1.0$ in the Chandra

181: Deep Field-South (CDF-S).  Then we determine the morphologies of the

182: galaxies in both samples in a consistent manner and quantify the

183: environment (parameterized by the local surface galaxy density) in

184: which the galaxies are situated.

185: 

186: 

187: \subsection{A Mass-Selected, Volume-Limited SDSS Galaxy Sample}\label{secsdss}

188: 

189: The sample consists of two redshift bins: galaxies at redshifts

190: $0.02<z<0.03$ with stellar masses $4\times

191: 10^{10}<M/M_{\odot}<1.6\times 10^{11}$, and galaxies at redshifts

192: $0.035<z<0.045$ with stellar masses $M > 1.25\times

193: 10^{11}~M_{\odot}$.  The reason for this split is to ensure

194: spectroscopic completeness on both the faint and the bright end and to

195: minimize systematic problems with the photometry of bright galaxies

196: (see below). The $g$-band magnitudes of the galaxies in these samples

197: range from $g_{\rm{mod}}=13.7$ to 16.5. Here $g_{\rm{mod}}$ is the

198: model magnitude from the SDSS pipeline (see below for the

199: definition). The upper limit $g_{\rm{mod}}=13.7$ ensures that this

200: sample is not incomplete due to the cutoff at bright magnitudes in the

201: SDSS spectroscopic survey \citep{strauss02}.  This source of

202: incompleteness is essentially a surface brightness limit imposed to

203: avoid saturation.  The lower magnitude limit $g_{\rm{mod}}=16.5$

204: ensures spectroscopic completeness at the faint end \citep{blanton05b}

205: and the feasibility of determining morphologies without being limited

206: by background noise.  Finally, only those galaxies located at least

207: 130 pixels from the edge of the image tiles are included in the

208: sample, in order to avoid problems with the morphological analysis

209: described below. This does not introduce systematic effects other than

210: decreasing the volume of the sample by 28\%.  The lower mass limit of

211: $M=4\times 10^{10}~M_{\odot}$ is chosen to match the $z\sim 0.8$

212: sample that we construct below (see \S~\ref{secdatacdfs}).

213: Approximately two-thirds of the stars are located in galaxies with

214: $M>4\times 10^{10}~M_{\odot}$ \citep{bell03}, such that in terms of

215: stellar mass our samples contain a representative fraction of the

216: galaxy population.

217: 

218: \input{tab1.tex}

219: 

220: 

221: 

222: Total magnitudes can be underestimated due to overestimation of the

223: sky brightness by the SDSS reduction pipeline \citep{lauer07}. The

224: magnitude of this effect depends on both galaxy size and surface

225: brightness, i.e., to first order on apparent magnitude.  Since we are

226: working with a large sample we are primarily interested in a

227: statistical, magnitude-dependent correction for the sample as a whole.

228: For a dozen galaxies with magnitudes across the range of our sample we

229: determine total magnitudes by hand.  The true background is determined

230: by SExtractor \citep{bertin96}, using the global background mode, and

231: the total flux is measured with the {\tt ellipse} task in IRAF.  We

232: find that $g_{\rm{tot}}=g_{\rm{mod}}-0.2(15-g_{\rm{mod}})$ for

233: $g_{\rm{mod}}<15$ and $g_{\rm{tot}}=g_{\rm{mod}}$ for

234: $g_{\rm{mod}}>15$.  This correction is of similar magnitude to the

235: corrections applied by \citet{vonderlinden07}, who designed a method

236: to obtain accurate, corrected magnitudes for individual galaxies.  Our

237: correction is not perfect for individual galaxies but does provide a

238: sufficiently accurate correction for the sample as a whole. The impact

239: of the correction on the stellar masses that we derive below is $\sim$

240: 0.1 dex for the brightest galaxies, such that even if the correction

241: we apply is uncertain on the 50\% level, the effect on the stellar

242: masses is only 0.05 dex, which is negligible with respect to the other

243: uncertainties on the mass estimates.  The systematic error will be

244: still lower.

245: 

246: We adopt $u-g=u_{\rm{mod,c}}-g_{\rm{mod,c}}$ and

247: $g-r=g_{\rm{mod,c}}-r_{\rm{mod,c}}$ as the galaxy colors, where

248: $u_{\rm{mod,c}}$, $g_{\rm{mod,c}}$, and $r_{\rm{mod,c}}$ are the model

249: magnitudes, corrected for Galactic extinction. Here $r_{\rm{mod,c}}$

250: is either the De Vaucouleurs model magnitude or the exponential model

251: magnitude, depending on which one fits best to the observed surface

252: brightness profile in the $r$ band.  The best-fit radius, ellipticity,

253: and orientation as determined in the $r$ band are subsequently used to

254: measure $u_{\rm{mod,c}}$ and $g_{\rm{mod,c}}$, with only the surface

255: brightness as a free parameter \citep[see][for a description of the

256: data reduction pipeline]{lupton01}.  The colors thus obtained are

257: therefore equivalent to the color within the $r$-band half-light

258: radius.

259: 

260: $K$-corrections are not negligible, even at such low redshifts

261: \citep{blanton03}.  We derive $g_{\rm{tot},0}$, $(u-g)_0$, and

262: $(g-r)_0$ ($g$-band magnitudes, and $u-g$ and $g-r$ colors

263: $K$-corrected to $z=0$) by following the technique applied to higher

264: redshifts galaxies by \citet{blakeslee06}, \citet{holden06} and Paper

265: I.  We derive linear relations between colors and magnitudes in the

266: observed and rest frames from synthetic spectra for stellar

267: populations from \citet{bruzual03} with a range of metallicities and

268: star formation histories (exponentially declining star formation rates

269: with different timescales).  In Table \ref{tab1} we show the

270: transformations for redshifts $z=0.02$, 0.03, and 0.04.

271: 

272: Subsequently, stellar masses are estimated using the relation between

273: $(g-r)_0$ and $M/L$ derived by \citet{bell03} for a ``diet'' Salpeter

274: initial stellar mass function (IMF):

275: \begin{eqnarray*}

276: \log \bigg ( \frac{M}{L_{\rm{g}}} \bigg ) = 1.519 (g-r)_0 - 0.499 .

277: \end{eqnarray*}

278: The diet Salpeter IMF \citep[e.g.,][]{bell03} gives values of $M/L$

279: that are 0.15 dex lower than in the standard Salpeter IMF, which is

280: due to the flat slope for stellar masses below $0.35~M_{\odot}$.  The

281: $K$-corrections and color-$M/L$ conversions can be combined to give a

282: unique relation between stellar mass, apparent magnitude, color, and

283: redshift:

284: \begin{eqnarray*}

285: \log\bigg (\frac{M}{M_{\odot}}\bigg ) = f(g_{\rm{tot}};(g-r);z) = \\

286: -0.399 \bigg [g_{\rm{tot}} - 3.65 (g-r) - 4.79 \log\bigg (

287:   \frac{z}{0.03}\bigg ) - 39.55 \bigg].

288: \end{eqnarray*}

289: In order to compare with previous work and the $z\sim 0.8$ sample

290: described below, we derive the standard $U-B$ and $B-V$ colors and

291: $B$-band magnitudes in the Vega system.

292: 

293: \begin{figure}

294: \epsscale{1.2}

295: \plotone{f1.eps}

296: \caption {Comparison between stellar masses (as used in this paper)

297:  and kinematic masses ($\propto R\sigma^2$) for galaxies at redshifts

298:  $0.030<z<0.045$. The dotted line indicates the one-to-one

299:  correspondence between $M_{\rm{dyn}}$ and $M_{\rm{phot}}$. The solid

300:  line is the best linear fit, which has slope 0.9 and shows that the

301:  offset between the two mass estimates is, on average, 0.10 dex.  The

302:  scatter is 0.17 dex.}

303: \label{masscheck_sdss}

304: \end{figure}

305: 

306: In Figure \ref{masscheck_sdss} we compare our stellar mass estimates

307: with kinematic mass estimates, where the latter is defined as

308: $M_{\rm{dyn}}/M_{\odot}=\log(R)+2\log(\sigma_c)+6.07$

309: \citep{jorgensen96}.  Here, $R$ is either the De Vaucouleurs or the

310: exponential disk scale radius in kiloparsecs as measured in the $g$

311: band, depending on which profile provides the better fit, and

312: $\sigma_c$ is the aperture-corrected velocity dispersion

313: \citep[see][]{bernardi03a} in kilometers per second. There is an

314: average, systematic offset of 0.10 dex, the stellar mass estimates

315: being lower than the kinematic mass estimates. This can be due to the

316: choice of the diet Salpeter IMF and/or the presence of nonbaryonic

317: matter, even though $\sigma$ is measured at the centers of the

318: galaxies, which are dominated by stars.  The offset varies slightly

319: with mass (by $\lesssim \pm 0.05$ dex) for galaxies with

320: $M_{\rm{phot}}>4\times 10^{10}~M_{\odot}$, the mass limit of the

321: sample used in this paper.  Such a mass dependence of the difference

322: between kinematic and stellar mass estimates is also seen at higher

323: redshifts (see \S~\ref{secdatacdfs}). The scatter in Figure

324: \ref{masscheck_sdss} is 0.17 dex, which implies that the errors in the

325: photometric mass estimates for individual galaxies are less than

326: $\sim$50\%. Note that this is an upper limit on the real uncertainty

327: since kinematic mass estimates have their own uncertainties. We

328: conclude that the color measurements and the resulting stellar mass

329: estimates are sufficiently accurate for our purposes.

330: 

331: The final sample consists of 2003 galaxies that are more massive than

332: $M=4\times 10^{10}~M_{\odot}$. Of those, 545 are in the

333: $0.035<z<0.045$ subsample of galaxies with masses $M>1.25\times

334: 10^{11}~M_{\odot}$. When we compute numbers and fractions of numbers

335: of galaxies we give those high-mass galaxies lower weight (0.395),

336: which is the ratio of the volumes occupied by the $0.035<z<0.045$

337: sample and the $0.02<z<0.03$ sample.

338: 

339: \begin{figure}

340: \epsscale{1.2}

341: \plotone{f2.eps}

342: \caption{ \textit{Small circles:} Galaxies brighter than $z_{850}=24$

343:   in the CDF-S. \textit{Large circles:} Galaxies with spectroscopic

344:   redshifts $0.9<z<1.0$.  A mass-selected sample will be biased most

345:   strongly at the high-redshift end ($z=1$) and against those galaxies

346:   with the highest $M/L$, i.e., those with the reddest colors.  The

347:   area indicated by the rectangle indicates the colors of the reddest

348:   galaxies at $z\sim 1$.  In Fig. \ref{compl} we address the redshift

349:   completeness of galaxies with such colors.}

350: \label{vi_iz}

351: \end{figure}

352: 

353: 

354: \subsection{A Mass-Selected, Volume-Limited Galaxy Sample in the CDF-S}\label{secdatacdfs}

355: 

356: We combine photometric and spectroscopic redshift samples to construct

357: a stellar mass-limited sample of galaxies in the CDF-S in the redshift

358: range $0.6<z<1.0$. We use the photometric redshift catalog by

359: \citet{wuyts07}, which is based on a $K$-band-selected sample.  We

360: supplement these with spectroscopic redshifts from the literature

361: \citep{lefevre04,mignoli05,vanderwel05,vanzella06}. Only those

362: galaxies for which the spectroscopic redshifts have been measured with

363: reasonable confidence (best or second-best quality flags) are

364: considered. Of the final sample, 142 out of 207 or 69\% have

365: spectroscopic redshifts. We refer to the respective publications for

366: explanations of the confidence indicators.

367: 

368: Extensive comparisons by Wuyts et al. show that in the redshift range

369: of our sample, there are no systematic differences between

370: spectroscopic and photometric redshifts, regardless of color and

371: magnitude cuts imposed on the sample.  There are occasional

372: catastrophic failures, but their number is small ($1\%-2\%$), and

373: therefore this does not introduce significant systematic effects.

374: 

375: The $K$-band limit of the photometric redshift catalog determines the

376: mass completeness limit of our sample. Since we use $v_{606}$-,

377: $i_{775}$-, and $z_{850}$-band ACS (\textit{Hubble Space Telescope}

378: Advanced Camera for Surveys) photometry to calculate masses,

379: luminosities, and rest-frame colors we quantify the completeness limit

380: in terms of $z_{850}$-band total magnitudes.  The $z_{850}$-band total

381: magnitudes are the SExtractor Best magnitude, but we add flux (0.2

382: mag) to correct for light outside the aperture

383: \citep{benitez04,blakeslee06}.  We measure $v_{606}-i_{775}$ and

384: $i_{775}-z_{850}$ within a $0.5''$ diameter aperture, and we correct

385: for differential point-spread function (PSF) effects. The latter

386: correction is negligible in $v_{606}-i_{775}$ but ranges up to 0.07

387: mag (for point sources) in $i_{775}-z_{850}$. The aperture is chosen

388: such that we measure the color within the effective radius or

389: half-light radius of a typical $z\sim 0.8$ galaxy, which corresponds

390: closely to the model colors we use to define the colors of the

391: galaxies in the $z\sim 0.03$ sample (see \S~\ref{secsdss}).

392: 

393: \begin{figure}

394: \epsscale{1.2}

395: \plotone{f3.eps}

396: \caption{ \textit{Dashed line with crosses:} Fraction of galaxies in

397:   the CDF-S with known spectroscopic and/or photometric redshifts as a

398:   function of $z_{850}$ magnitude.  The completeness begins to

399:   decrease at $z_{850}\sim 23-23.5$.  \textit{Dotted line with

400:   circles:} Completeness of galaxies within the rectangle indicated in

401:   Fig. \ref{vi_iz}, i.e., those galaxies with colors similar to the

402:   reddest galaxies at $z\sim 1$.  This sample is essentially complete

403:   down to $z_{850}=24$, which implies that we can use $z_{850}=24$ to

404:   define the mass limit of the sample.}

405: \label{compl}

406: \end{figure}

407: 

408: 

409: 

410: Selection effects in a mass-limited sample will be strongest at the

411: high end of a redshift bin (here $0.6<z<1.0$) and for the galaxies

412: with the highest mass-to-light ratio ($M/L$), i.e., the reddest

413: galaxies.  In Figure \ref{vi_iz} we show the $v_{606}-i_{775}$ and

414: $i_{775}-z_{850}$ colors of the galaxies in the CDF-S with

415: spectroscopic redshifts $0.9<z_{\rm{spec}}<1.0$.  The rectangle, which

416: outlines the region with the reddest galaxies at that redshift,

417: indicates for which colors incompleteness first begins to affect our

418: mass-limited sample.  In Figure \ref{compl} we show that for those

419: colors the $K$-band-selected photometric redshift catalog is virtually

420: complete down to $z_{850}=24$, fainter than the completeness limit of

421: galaxies with a typical color. In the following, we adopt $z_{850}=24$

422: as the magnitude limit which we use to define the mass completeness

423: limit of our sample.  Bluer galaxies near this mass limit are much

424: brighter, and incompleteness does not play a role for those objects.

425: The reason for using a $z_{850}$-band magnitude limit to define the

426: mass completeness limit instead of the $K$-band magnitude limit of the

427: photometric redshift catalog is that we derive stellar masses by

428: estimating mass-to-light ratios in the rest-frame $B$-band, which

429: corresponds to the observed $z$ band at $0.6<z<1.0$.

430: 

431: Stellar masses are derived with the same method as used for the $z\sim

432: 0.03$ galaxy sample. First, observed magnitudes and colors ($z_{850}$,

433: $v_{606}-i_{775}$, and $i_{775}-z_{850}$) are used to derive

434: rest-frame $B$-band luminosities and rest-frame $U-B$ and $B-V$

435: colors.  For $z<0.85$ we use $v_{606}-i_{775}$ to derive $(U-B)_0$ and

436: $v_{606}-z_{850}$ to derive $(B-V)_0$ and $M_{\rm{B},0}$; for $z>0.85$

437: we use $v_{606}-z_{850}$ to derive $(U-B)_0$ and $i_{775}-z_{850}$ to

438: derive $(B-V)_0$ and $M_{\rm{B},0}$.  These choices optimally match

439: the observed and rest-frame bands, and the derived rest-frame colors

440: are well behaved as a function of redshift. Examples for various

441: redshifts are given in Table \ref{tab1}.  The rest-frame $U$-, $B$-,

442: and $V$-band magnitudes and colors are in the Vega system, as opposed

443: to the other magnitudes and colors, which are in the AB system. This

444: is necessary to facilitate the comparison with earlier results.

445: Finally, the empirical relation between $M/L$ and $B-V$ from

446: \citet{bell03} is used to estimate the stellar masses:

447: \begin{eqnarray*}

448: \log\bigg ( \frac{M}{L_{\rm{B}}}\bigg )) = 1.737 (B-V)_0 - 0.994.

449: \end{eqnarray*}

450: 

451: For 14 early-type galaxies in the $z\sim 0.8$ sample kinematic

452: measurements are available from \citet{vanderwel05}. There is a

453: systematic offset between the kinematic and stellar mass estimates of

454: 0.1 dex (the kinematic masses being larger) and a scatter of 0.3 dex.

455: These results are consistent with earlier determinations of the

456: robustness of stellar mass estimates at $z\sim 0.8-1$

457: \citep{vanderwel06b,holden06}.  Most importantly, the systematic

458: difference between kinematic and stellar mass estimates is the same

459: (0.1 dex) for the $z\sim 0.03$ sample and the $z\sim 0.8$ sample.

460: Therefore, we conclude that our stellar mass estimates do not

461: introduce systematic effects into our samples that exceed 0.1 dex in

462: mass as was found earlier by both \citet{vanderwel06b} and

463: \citet{holden06}. Uncertainties at that level do not significantly

464: affect our results.  The difference between the kinematic and stellar

465: mass estimates is mass dependent \citep[see, e.g.,][and Paper

466: I]{holden06}, but since this is also the case for our $z\sim 0.03$

467: sample (see \S~\ref{secsdss}), and at the same level, this does not

468: affect our results significantly.

469: 

470: \begin{figure}

471: \epsscale{1.2}

472: \plotone{f4.eps}

473: \caption{ Rest-frame $B$-band luminosity vs. stellar mass for galaxies

474:   in the CDF-S with redshifts $0.6<z<1.0$, using spectroscopic

475:   redshifts if available, and, to ensure completeness, including

476:   galaxies with only photometric redshifts as well.  Our magnitude

477:   completeness limit $z_{850}=24$ corresponds to the luminosity limit

478:   at $z=1$ (\textit{dotted line}).  In addition, we adopt

479:   $\log(M/L_B)=0.8$ (\textit{dashed line}) as the maximum value that

480:   the $M/L$ of any galaxy can take.  This is a reasonable assumption

481:   as this is the highest $M/L$ for galaxies at the high-mass,

482:   high-luminosity end of our sample.  The magnitude and $M/L$ limits

483:   cross at $M>4\times 10^{10}~M_{\odot}$.  This is the mass

484:   completeness limit of our sample. At lower masses, our sample is

485:   incomplete for faint, red (high $M/L$) galaxies (see Fig.

486:   \ref{compl}).  The mass-selected sample contains 207 galaxies, 142

487:   of which have spectroscopic redshifts.}

488: \label{L_M}

489: \end{figure}

490: 

491: \begin{figure*}

492: \epsscale{1.1}

493: \plottwo{f5a.eps}{f5b.eps}

494: \caption{ S\'ersic $n$ vs. bumpiness $B$ for all 207 galaxies in the

495:   mass-selected $z\sim 0.8$ sample (\textit{left}) and the 2003

496:   galaxies in the $z\sim 0.03$ sample (\textit{right}). The 200

497:   galaxies with visual classifications are separately indicated, and

498:   the remainder of the sample is indicated with small dots.  The

499:   S\'ersic index $n$ is restricted to be between $n=1$ and 4, but to

500:   enhance the readability of the figure we have randomly distributed

501:   the values of $n$ between 0.5 and 1 for galaxies with $n=1$ and

502:   between $4<n<5$ for galaxies with $n=4$. Visual classifications are

503:   indicated by different symbols: early types are indicated by open

504:   circles, late types by crosses.  The curved line, $B=0.065(n+0.85)$,

505:   indicates the separation between early types and late types

506:   according to the $B-n$ classification.  The correspondence between

507:   the visual and automated classifications is very good. See text for

508:   more details and Figure \ref{panels} for illustration.}

509: \label{morphcheck}

510: \end{figure*}

511: 

512: For late-type galaxies a comparison between kinematic and stellar mass

513: estimates is less straightforward. Recently, \citet{kassin07} showed

514: that the relation between stellar mass and kinematic mass as defined

515: by a combination of the rotation speed and the velocity dispersion of

516: the gas does not vary by more than $\sim$0.1 dex over the redshift

517: range $0.1<z<1.2$.  However, the stellar masses from Kassin et al.

518: are determined by fitting stellar population models to broadband

519: spectral energy distributions, which is not directly comparable with

520: our method.  If Kassin et al. had used the color-$M/L$ relations from

521: \citet{bell03}, the stellar masses would have changed such that at

522: $z\sim 1$ they would be $\sim$0.2 dex higher than those of local

523: galaxies with the same kinematic mass (Paper I). We comment on the

524: effects of possible biases where relevant.

525: 

526: 

527: 

528: In Figure \ref{L_M} we show the rest-frame $B$-band luminosities and

529: stellar masses of the galaxies in our $z\sim 0.8$ sample.  The highest

530: $M/L$ that occurs (for luminous, red galaxies) is

531: $M/L_{\rm{B}}=6.3$. We adopt this as the $M/L$ upper boundary, which,

532: combined with the magnitude limit derived above ($z_{850}=24$), gives

533: us the mass completeness limit of our sample: $M=4\times

534: 10^{10}~M_{\odot}$ (see Fig. \ref{L_M}).  The mass-limited sample

535: contains 207 galaxies.

536: 

537: The average $M/L$ is $M/L_{\rm{B}}=2.4$ for the galaxies with

538: $M>4\times 10^{10}~M_{\odot}$. For the $z\sim 0.03$ sample this is

539: $M/L_{\rm{B}}=4.9$. This implies 0.75 mag evolution between the two

540: samples in the $B$-band, or 1 mag per unit redshift. This is

541: consistent with the evolution of $L^*$ \citep[e.g.,][]{brown07}, but

542: somewhat less than the $M/L_{\rm{B}}$ evolution inferred via

543: fundamental plane analyses for $z\sim 1$ field early-type galaxies

544: with similar masses \citep[0.55

545: dex;][]{vanderwel05,treu05b,vandokkum07}.  This difference may contain

546: contributions from a systematic uncertainty in the high-$z$ mass

547: estimates as described above, and an increased number of very red,

548: dusty galaxies at $z\sim 1$ with high $M/L$. 

549: 

550: \subsection{Morphological Classifications}\label{secmorph}

551: 

552: Traditionally, visual morphological classifications are used for

553: morphological studies of high-redshift galaxy populations, for example

554: the evolution of the MDR \citep[e.g.,][]{postman05}.  In recent years,

555: however, increased sample sizes have initiated the emergence of

556: several alternative, automated classification schemes

557: \citep{conselice00,abraham03,lotz04,blakeslee06}. In addition to the

558: traditional, visual classifications we therefore also deploy a

559: quantitative measure of the galaxy morphologies. This enables us to

560: analyze the large sample of $z\sim 0.03$ galaxies and to examine, in a

561: quantitative manner, the effect of cosmological surface brightness

562: dimming on the detectability of disks, spiral arms, and

563: irregularities. The visual classifications are used to verify the

564: robustness of this automated classification method.  In this paper we

565: simply distinguish between early- and late-type galaxies, or, in the

566: $T$-type classification scheme \citep[see][]{postman05}, between

567: galaxies with type $T\leq 0$ (E, E/S0, and S0) and galaxies with type

568: $T>0$ (Sa and later).  The separation into subclasses, for example E

569: and S0, will be investigated in a future paper.

570: 

571: Visual morphologies were determined by A. v. d. W.  by examining the

572: $g$-band images of 200 randomly chosen galaxies in the mass-selected

573: $z\sim 0.03$ sample and the $z_{850}$-band images of all 207 galaxies

574: in the $z\sim 0.8$ sample.  For 35 galaxies in the $z\sim 0.8$ sample

575: with $T\sim 0$ or peculiar morphologies M. P. also examined the images

576: and assigned morphologies.  This provides an estimate of the random

577: uncertainty and ensures that the same threshold ($T=0$) is adopted as

578: in \citet{postman05}.  Based on these results, we conservatively

579: assume a 5.5\% uncertainty in the early-type galaxy fraction due to

580: classification errors.  We note that the $z_{850}$-band imaging with

581: an integration time of 12,000 s and a $10\sigma$ point-source

582: detection limit of $z_{850}=27$ is very deep with respect to the

583: faintest galaxies in our sample with $z_{850}=24$.  In Figure

584: \ref{morphcheck} we indicate early-type galaxies with open circles and

585: late-type galaxies with crosses.

586: 

587: Quantitative morphologies are determined for all galaxies with the

588: technique outlined by \citet{blakeslee06}, who showed that the

589: S\'ersic parameter $n$ and bumpiness parameter $B$ can effectively

590: distinguish between early- and late-type galaxies.  GALFIT

591: \citep{peng02} is used to fit S\'ersic profiles to the $g$-band and

592: $z_{850}$-band images; $n$ is constrained to values between $n=1$ and

593: 4. The only reason for not allowing larger values of $n$ is that the

594: effective radii for high values of $n$ become large and uncertain.

595: This would strongly affect the measurement of the bumpiness $B$, which

596: is the rms of the residual as measured within two effective radii.  In

597: order to reduce the effect of shot noise the residual is slightly

598: smoothed. In addition, other objects and a circular region around the

599: center are masked. This central region is masked out because of slight

600: PSF mismatches and because of central, small-scale deviations from a

601: S\'ersic profile. Because of the latter effect the adopted radius is

602: distance dependent. On visual inspection we choose a 4 pixel radius

603: for galaxies at $z=0.03$ in the low-$z$ sample and a 2 pixel radius

604: for galaxies at $z=0.80$ in the high-$z$ sample, both scaled inversely

605: proportional with angular diameter distance for galaxies at different

606: redshifts.

607: 

608: 

609: The advantage of the $B-n$ classification over other automated

610: classifiers is that PSF smearing is taken into account. This is

611: essential, in particular when redshift-dependent trends are

612: investigated in data sets with very different photometric

613: properties. In Figure \ref{panels} we show image cutouts and give the

614: $n$ and $B$ parameters for 20 randomly selected galaxies from our

615: $z\sim 0.0$ and 0.8 samples, in order to illustrate the correspondence

616: between the automated classification method and their morphological

617: appearance.  The parameters $n$ and $B$ are shown for all galaxies

618: with visual morphologies in Figure \ref{morphcheck}. A simple linear

619: relation separates early- and late-type galaxies: $B=0.065(n+0.85)$

620: (Fig. \ref{morphcheck}, \textit{solid line}).  Note that these

621: coefficients are taken from \citet{blakeslee06}, and are not chosen to

622: optimize the correspondence between the visual and $B-n$

623: classifications for these particular samples.  The correspondence

624: between our results, in particular for the $z\sim 0.03$ sample, and

625: the results by Blakeslee et al. illustrates the power of the $B-n$

626: method.  Once PSF smearing effects and noise properties are properly

627: taken into account, data sets with roughly similar spatial resolution

628: provide a unique measurement of galaxy morphology through the $B-n$

629: classification.  This extends the conclusions from Blakeslee et

630: al. who could not claim universality of their selection criteria

631: because of their homogeneous data sets and the small redshift range

632: ($z=0.83-0.84$) of their samples.

633: 

634: \begin{figure*}

635: \epsscale{2.4}

636: \plottwo{f6a.eps}{f6b.eps}

637: \caption{\textit{Top two rows:} $3.4'' \times 3.4''$ $z_{850}$-band

638:  cutouts for 10 randomly chosen galaxies from the sample at

639:  $0.6<z<1.0$. \textit{Bottom two rows:} $51'' \times 51''$ $g$-band

640:  cutouts for 10 randomly chosen galaxies from the sample at $z\sim

641:  0.03$. The physical scale of both sets of images is comparable:

642:  $\sim$30 kpc. Visual classifications (late or early) and $n$ and $B$

643:  parameters are listed for each galaxy to illustrate the

644:  correspondence between the visual morphologies and the automated

645:  classification method (see also Fig. \ref{morphcheck}).}

646: \label{panels}

647: \end{figure*}

648: 

649: 

650: The visual and $B-n$ classifications agree for 93\% of the galaxies in

651: the $z\sim 0.8$ sample (193 of 207) and for 89\% of the galaxies in

652: the $z\sim 0.03$ sample (179 of 200; see also Fig. \ref{morphcheck}).

653: Half of the mis-classifications have $n$- and $B$-values that put them

654: close to the criterion that separates the early types from the late

655: types (Fig. \ref{morphcheck}, \textit{solid line}). The disagreement

656: can be explained by the limited signal-to-noise ratio ($S/N$) of the

657: images, which shows that the true random uncertainty in the visual and

658: $B-n$ methods is $\sim$4\%. The other half of the disagreements are

659: mostly due to strong, central deviations from a S\'ersic profile

660: (probably point sources) or, for several galaxies in the $z\sim 0.03$

661: sample, large-scale deviations from a smooth profile that our fitting

662: method does not place within two effective radii.  Despite the

663: nonnegligible numbers of erroneous classifications ($\sim$10\%), the

664: net difference between the two classification methods in the ratio of

665: the numbers of early-type and late-type galaxies is less than 0.5\%.

666: This is true for both the $z\sim 0.03$ sample and the $z\sim 0.8$

667: sample.  Hence, there is virtually no systematic difference between

668: the visual and $B-n$ classifications for the samples as a whole.

669: 

670: Even though the $z\sim 0.03$ and 0.8 samples have internally

671: consistent morphological classifications, there may be a systematic

672: difference between the two samples, which are located at very

673: different cosmological distances. Our morphological classifications

674: could depend on redshift, mainly because of the lower $S/N$ of the

675: high-$z$ galaxy images.  Because the scales of the PSF and pixels of

676: the SDSS and ACS data sets are very similar in terms of physical size

677: at the sample redshifts, we can test the redshift dependence of our

678: morphological classification.  We take 50 galaxies in our sample

679: spanning the full range in magnitude and with redshifts

680: $0.025<z<0.030$, such that the SDSS pixel scale corresponds to the

681: pixel scale of an ACS image of a galaxy at $z=0.80$.  We add noise to

682: the $g$-band SDSS images, taking into account $K$-corrections and the

683: relative depths of the SDSS and ACS images, and reapply our two

684: classification methods, which gives $n_{0.80}$ and $B_{0.80}$, the

685: values for $n$ and $B$ as inferred from the simulated images.  In

686: Figure \ref{morphz_sdss} we compare those with the original values of

687: $n$ and $B$, and we find that neither quantity shows a systematic

688: difference.  In addition, the visual morphologies of the simulated

689: images are not different from the original visual morphologies: all

690: features that decide on the morphology of the $z\sim 0.03$ galaxies

691: are still visible in the simulated $z=0.80$ images.  We conclude that

692: our classification methods do not suffer from systematic effects that

693: introduce significant differences between the morphologies of the

694: $z\sim 0.03$ and $z\sim 0.8$ galaxies.

695: 

696: \begin{figure}

697: \epsscale{1.2} 

698: \plotone{f7.eps} 

699: \caption{ S\'ersic parameter $n$ and bumpiness parameter $B$ for 50

700:  galaxies at redshifts $0.025<z<0.030$ as determined from the original

701:  SDSS $g$-band images vs. the difference between $n$ and $B$ as

702:  determined from simulated $z=0.80~$ $z$-band images of the same

703:  galaxies and the original values. There are no systematic differences

704:  between the original low-$z$ and simulated high-$z$ morphological

705:  indicators, which indicates that the morphologies as determined for

706:  our $z\sim 0.03$ and $\sim 0.8$ samples are not systematically

707:  different.}

708: \label{morphz_sdss}

709: \end{figure}

710: 

711: \subsection{Local Density Estimates}\label{secdens}

712: 

713: Traditionally, local projected surface densities are computed by

714: measuring the distance to the $n$-th nearest neighbor brighter than,

715: for example, $M_{\rm{V}}+0.8z<-19.78$ \citep[e.g.,][]{postman05}.

716: However, since we work with mass-selected samples in this paper, it is

717: more consistent to measure the distance to the $n$th nearest neighbor

718: that is more massive than, in this case, $M>4\times

719: 10^{10}~M_{\odot}$.  Below we discuss the difference between the two

720: approaches.

721: 

722: We derive the local galaxy densities for the 142 galaxies in our

723: $z\sim 0.8$ sample with spectroscopic redshifts.  We compute the

724: distance to the $n$-th nearest neighbor with a difference in

725: rest-frame radial velocity of $\Delta v<1000~\rm{km~s}^{-1}$ and more

726: massive than $M=4\times 10^{10}~M_{\odot}$ (or, alternatively, more

727: luminous than $M_{\rm{V}}+0.8z<-19.78$).  Here $n$ is the number of

728: galaxies with such properties in the CDF-S, with a maximum of 7.  The

729: reason for allowing $n$ to be smaller than 7 is the limited number of

730: galaxies at a given, spectroscopic redshift.  We note that for

731: galaxies with fewer than 7 neighbors (12 of 142 for the

732: luminosity-limited sample, 36 of 142 for the mass-limited sample) the

733: density estimate does not systematically change for any value $3<n<7$.

734: The inferred surface densities are corrected for edge effects and for

735: incompleteness of the spectroscopic redshift catalog, as inferred from

736: the spectroscopic + photometric redshift catalog.

737: 

738: The distributions of $\Sigma_{\rm{M}}$ and $\Sigma_{\rm{L}}$ are shown

739: in Figure \ref{Dhist}; $\Sigma_{\rm{M}}$ is typically half of

740: $\Sigma_{\rm{L}}$, which is only a small difference considering the

741: large spread (3 orders of magnitude) of densities in the samples.  For

742: galaxies with only few neighbors, the computed values are likely upper

743: limits due to small area of the CDF-S.  Obviously, for galaxies

744: without close neighbors, the computed densities are certainly upper

745: limits.  The densities as computed here are, in principle, not

746: directly comparable with the computations for galaxies in cluster

747: environments \citep{postman05} as those use a wider redshift range to

748: search for neighbors to compensate for the large velocity dispersion

749: of the cluster members. A systematic difference of up to a factor of 2

750: may be expected due to the difference in bin width.  However, since

751: the environments under consideration have a range in density of

752: several orders of magnitude this does not compromise our inferences.

753: 

754: In a method analogous to that used for the $z\sim 0.8$ sample, we also

755: compute local galaxy densities by finding the distance to the seventh

756: nearest neighbor with $M=4\times 10^{10}~M_{\odot}$ (or

757: $M_{\rm{V}}+0.8z<-19.78$) and $\Delta v<1000~\rm{km~s}^{-1}$ for each

758: galaxy in the $z\sim 0.03$ sample.  The spectroscopic survey is 100\%

759: complete at these magnitudes, therefore corrections are not needed.

760: The distribution of $\Sigma$ is shown in Figure \ref{Dhist}.  As is

761: the case for the $z\sim 0.8$ sample, $\Sigma_{\rm{M}}$ and

762: $\Sigma_{\rm{L}}$ typically differ by a factor of 2 for the $z\sim

763: 0.03$ sample (see Table \ref{tab2}).

764: 

765: 

766: There is a difference of $\sim$0.7 dex between the median values of

767: $\Sigma$ for the $z\sim 0.03$ and 0.8 samples.  This is mainly because

768: we use the angular diameter distance $D_{\rm{A}}$ to calculate the

769: physical distances between the galaxies, and not the transverse

770: co-moving distance $D_{\rm{M}}=D_{\rm{A}}(1+z)$. This provides, to

771: first order, redshift-independent values of the local density for

772: bound systems such as clusters, but for expansion-dominated regions

773: the density will be redshift dependent.  Using $D_{\rm{M}}$ would

774: provide a redshift-independent measurement of the density and would

775: lower the $z\sim 0.8$ densities by 0.5 dex with respect to the $z\sim

776: 0.03$ densities. This demonstrates that the galaxies in both samples

777: are located in similar (co-moving) environments, besides an average,

778: apparent over-density of a factor of $1.5-2$ in the CDF-S which has

779: been noted before \citep[e.g.,][]{bundy05}.  This difference is not of

780: great interest for this study since we consider a range in density

781: that spans several orders of magnitude (see \S~\ref{secmdr}). Since we

782: are interested in the properties of the galaxy population as a

783: function of physical density, and to facilitate the comparison with

784: work on cluster galaxies, we use $\Sigma$ as calculated with

785: $D_{\rm{A}}$ in the remainder of this paper.

786: 

787: 

788: \section{The Evolution of the Early-Type Galaxy Fraction and the Morphology-Density Relation}

789: 

790: In \S~\ref{secetf} we calculate the evolution of the early-type galaxy

791: fraction in our field samples. In \S~\ref{secmdr} we compare this with

792: samples of cluster galaxies over the same redshift range and determine

793: the evolution of the MDR.

794: 

795: \begin{figure}

796: \epsscale{1.2}

797: \plotone{f8.eps}

798: \caption{ Distribution of local surface densities of the galaxies in

799:   our samples. (\textit{a}) Distribution of the $0.6<z<1.0$ sample,

800:   where the vertical dashed line indicates the value of $\Sigma$

801:   corresponding to one galaxy within the area covered by the survey

802:   data: objects with values of $\Sigma$ close to this line should be

803:   regarded as upper limits. (\textit{b}) Same as (\textit{a}), but for

804:   the $0.015<z<0.045$ sample. The solid histograms show the

805:   distributions of $\Sigma_{\rm{M}}$, the surface density as

806:   determined by regarding neighbors that are more massive than

807:   $M=4\times 10^{10}~M_{\odot}$. The dotted histograms show the

808:   distributions of $\Sigma_{\rm{L}}$, the surface density as

809:   determined by regarding neighbors that are brighter than

810:   $M_{\rm{V}}+0.8z=-19.78$.}

811: \label{Dhist}

812: \end{figure}

813: 

814: 

815: \subsection{The Field Early-Type Galaxy Fractions at $z\sim 0.8$ and 0}\label{secetf}

816: 

817: Our $z\sim 0.03$ sample consists of 2003 galaxies, the $z\sim 0.8$

818: sample of 207 galaxies. These volume-limited samples are complete down

819: to a stellar mass limit of $M=4\times 10^{10}~M_{\odot}$, about half

820: of the typical mass of a local field galaxy, or $M_{\rm{B}}\sim -20.1$

821: for red galaxies in the $z\sim 0.8$ sample. This luminosity limit for

822: red galaxies is similar to the $z=0.8-1$ luminosity limits of the

823: COMBO-17 \citep{bell04b}, DEEP2 \citep{faber07} and NDWFS

824: \citep{brown07} data sets used for recent red-sequence studies.

825: 

826: The early-type galaxy fraction at $z\sim 0.8$, for galaxies that are

827: more massive than $M=4\times 10^{10}~M_{\odot}$, is

828: $f_{\rm{ET}}=0.48\pm 0.07$.  This is based on the quantitative

829: morphological classification method described above, but we find the

830: same if we use the visual classifications (see also

831: \S~\ref{secmorph}).  If only galaxies with spectroscopic redshifts are

832: included (142 of 207 have spectroscopic redshifts), we find

833: $f_{\rm{ET}}=0.52$. The reason that the spectroscopic sample has a

834: slightly, but not significantly, higher early-type galaxy fraction is

835: that several of the spectroscopic campaigns specifically targeted red

836: objects.

837: 

838: \input{tab2.tex}

839: 

840: The main contributor to the error of 0.07 is the uncertainty in the

841: morphological classifications (0.055).  The second main source of

842: uncertainty is Poisson noise due to the limited sample size (0.036).

843: Smaller contributions include errors in photometric redshifts, color

844: transformations, and $M/L$ estimation.  All these individual

845: contributions are added in quadrature to obtain 0.07.

846: 

847: The early-type fraction at $z\sim 0.03$, for galaxies that are more

848: massive than $M=4\times 10^{10}~M_{\odot}$, is $f_{\rm{ET}}=0.43\pm

849: 0.03$.  Essentially, the only significant contributor to the error is

850: the uncertainty in the morphological classifications, which is

851: slightly smaller than for the $z\sim 0.8$ sample because of the higher

852: $S/N$ of the images.

853: 

854: The early-type galaxy fractions in the mass-selected $z\sim 0.03$ and

855: 0.8 samples are not significantly different (see Table \ref{tab2} and

856: Figure \ref{z_etf}).  This implies that if only two morphological

857: classes are considered, then the morphological composition of the

858: galaxy population did not change significantly between $z\sim 0.8$ and

859: the present, at least above our mass limit of $M=4\times

860: 10^{10}~M_{\odot}$. If anything, the early-type fraction is higher at

861: $z\sim 0.8$ than at $z\sim 0.03$.

862: 

863: \subsection{Comparison with Cluster Galaxies: The Morphology-Density Relation}\label{secmdr}

864: 

865: As we described above, the early-type galaxy fraction in our field

866: samples is $40\%-50\%$ at redshifts $0<z<1$.  In Figure \ref{z_etf} we

867: compare this with the results from Paper I, where we measure the

868: evolution of the cluster early-type fraction for galaxies down to the

869: same mass limit ($M=4\times 10^{10}~M_{\odot}$) and over the same

870: redshift range.  For galaxies more massive than $M=4\times

871: 10^{10}~M_{\odot}$ the early-type galaxy fraction is constant with

872: redshift, both in the field and in massive clusters (see

873: Fig. \ref{z_etf}). The same MDR exists at $z=0$ and 0.8, with

874: early-type fractions of $40\%-50\%$ in low-density environments and

875: $>$80\% in high-density environments.

876: 

877: The cluster samples from Paper I consist of galaxies with local

878: densities $> 50~\rm{Mpc}^{-2}$, our field samples consist of galaxies

879: with densities $< 50~\rm{Mpc}^{-2}$. In Figure \ref{morphdens} we show

880: the MDR at redshifts $z\sim 0.03$ and $\sim0.8$.  In this figure we

881: split our $z\sim 0.03$ field sample of 2003 galaxies into five density

882: bins, ranging from 0.01 to 10 Mpc$^{-2}$.  Our $z\sim 0.8$ field

883: sample is represented by a single data point, because of the limited

884: sample size. The average logarithm of the density of the 142 galaxies

885: with spectroscopic redshifts is used to estimate the density for the

886: sample as a whole.  The MDR has not evolved significantly since $z\sim

887: 0.8$ for galaxies more massive than $M=4\times 10^{10}~M_{\odot}$. The

888: early-type galaxy fraction in regions with a given local density has

889: remained constant over the past $\sim$7 Gyr.

890: 

891: \begin{figure}[b]

892: \epsscale{1.2}

893: \plotone{f9.eps}

894: \caption{ Early-type galaxy fraction as a function of redshift.  The

895:  circles indicate our samples of field galaxies ($\Sigma\lesssim

896:  50~\rm{Mpc}^{-2}$) at $z\sim 0.03$ (\textit{open blue circle}) and

897:  $z\sim 0.8$ (\textit{filled red circle}).  The squares indicate the

898:  samples of cluster galaxies ($\Sigma>50 ~\rm{Mpc}^{-2}$) from Paper I

899:  at $z=0.02$ (\textit{open blue square}, representing the Coma

900:  Cluster) and $z=0.83$ (\textit{filled red square}, representing the

901:  CL 0152 and MS 1054 clusters).  The fraction of early-type galaxies

902:  has not changed significantly since $z\sim 0.8$ for galaxy

903:  populations in similar environments.}

904: \label{z_etf}

905: \end{figure}

906: 

907: At first sight, this result seems to be at odds with previous studies

908: that reported significant evolution of the early-type galaxy fraction

909: \citep{smith05,postman05}. However, those analyses are based on

910: luminosity-selected samples, and, as is shown in Paper I, the observed

911: evolution is driven by blue, star-forming, low-mass galaxies. Samples

912: that are complete in terms of stellar mass do not show evidence for

913: evolution in the early-type fraction.  The evolution in the early-type

914: fraction seen in luminosity-selected samples is driven by the

915: evolution of the fraction of S0 galaxies, whereas the fraction of E

916: galaxies is roughly constant

917: \citep{dressler97,treu03,postman05,desai07}. It remains to be seen how

918: the relative fractions of E and S0 galaxies evolve in mass-selected

919: samples.

920: 

921: \begin{figure*}

922: \epsscale{1.1}

923: \plottwo{f10a.eps}{f10b.eps}

924: \caption{ Morphology-density relation for mass-selected galaxies

925:  ($M>4\times 10^{10}~M_{\odot}$) at $z\sim 0$ and $z\sim 0.8$. The

926:  symbols are the same as in Figure \ref{z_etf}.  Our $z\sim 0.03$

927:  field sample (\textit{open blue circles connected by the dotted

928:  line}) have been split into five density bins.  Over almost 3 orders

929:  of magnitude in density the morphology-density relation has been in

930:  place since at least $z=0.8$ and has not evolved within the

931:  measurement errors.  The difference between the two panels is the use

932:  of $\Sigma_{\rm{M}}$ (the surface density of galaxies more massive

933:  than $M=4\times 10^{10}~M_{\odot}$, or $\sim0.5M_*$) in (\textit{a}),

934:  and the use of $\Sigma_{\rm{L}}$ (the surface density of galaxies

935:  more luminous than $M_{\rm{V}}+0.8z<-19.78$, or $\sim0.5L_*$) in

936:  (\textit{b}). Other than a shift of 0.3 dex in density the use of

937:  $\Sigma_{\rm{M}}$ and $\Sigma_{\rm{L}}$ is interchangeable.}

938: \label{morphdens}

939: \end{figure*}

940: 

941: 

942: 

943: The MDR in the $z\sim 0.03$ data set shown in Figure \ref{morphdens}

944: suggests that the early-type fraction is low ($\sim$40\%) for all

945: local surface densities $\Sigma_{\rm{M}}\lesssim 1~\rm{Mpc}^{-2}$.

946: The transition to very high early-type fractions ($\sim$100\%) happens

947: at $1~\rm{Mpc}^{-2}<\Sigma_{\rm{M}}<100~\rm{Mpc}^{-2}$.  This suggests

948: that there is a critical density above which galaxies are

949: progressively more likely to undergo morphological transformations, as

950: was earlier noted by \citet{goto03}.  Despite the small area and the

951: small dynamic range in densities, there is evidence for this

952: transition in our $z\sim 0.8$ sample: if we take the galaxies with

953: spectroscopic redshifts in the highest 20th percentile of the density

954: distribution (typically, $\Sigma_{\rm{M}}\sim 30~\rm{Mpc}^{-1}$), we

955: find an early-type fraction of $f_{\rm{ET}}=0.8\pm 0.1$; for galaxies

956: in the lowest 20th percentile (typically, $\Sigma_{\rm{M}}\sim

957: 0.2~\rm{Mpc}^{-1}$), we find $f_{\rm{ET}}=0.4\pm 0.1$. Without

958: over-emphasizing these numbers because the sub-sample with

959: spectroscopic redshifts is not necessarily representative of the whole

960: $z\sim 0.8$ sample, we do note that these values follow the same trend

961: as seen in the $z\sim 0.03$ sample (see Fig. \ref{morphdens}).

962: 

963: 

964: \section{Discussion}

965: 

966: \subsection{The Evolution of the Star Formation Rate}\label{secsf}

967: 

968: Our conclusion that for galaxies with masses $M>4\times

969: 10^{10}~M_{\odot}$ the early-type galaxy fraction has not

970: significantly changed in any environment over the past 7 Gyr has to be

971: reconciled with the well-constrained decrease of the star formation

972: rate (SFR) density by an order of magnitude since $z\sim 1$

973: \citep[see, e.g.,][]{madau96,wolf03,lefloch05,bell05,perez05,zheng06}.

974: It may seem counter-intuitive to have the SFR decrease rapidly without

975: a change in the morphological mix of the galaxy population. On the

976: other hand, a non-changing early-type galaxy fraction does not

977: necessarily imply that the galaxy population itself does not

978: evolve. Below we investigate the star formation activity in our

979: samples.

980: 

981: We use the very deep, publicly available GOODS\footnote{The Great

982: Observatories Origins Survey (GOODS) data are available at

983: http://data.spitzer.caltech.edu/popular/goods/Documents/goods\_dr3.html}

984: $24~\mu\rm{m}$ imaging from the Multiband Imaging Photometer for

985: \textit{Spitzer} \citep[MIPS;][]{rieke04} to verify the star formation

986: activity in our $z\sim 0.8$ sample of mass-selected galaxies. Total

987: $24~\mu\rm{m}$ fluxes are measured by fitting the MIPS PSF to the

988: positions of objects detected in the $3.6~\mu\rm{m}$ band of the

989: Infrared Array Camera \citep[IRAC;][]{fazio04} on board

990: \textit{Spitzer}.  This technique reduces the effect of blending and

991: allows for more accurate photometric measurements \citep[for more

992: details, see][]{wuyts07}.  The $24~\mu\rm{m}$ flux is used to estimate

993: the bolometric infrared luminosity $L_{\rm{IR}}$ via the conversions

994: from \citet{chary01}. Subsequently, we use $L_{\rm{IR}}$ to estimate

995: the SFR \citep{kennicutt98}, where the SFR is scaled down by 0.15 dex

996: to account for the difference between the standard Salpeter IMF and

997: the diet Salpeter IMF that we use. For more details concerning the

998: $K$-correction and uncertainties (roughly a factor of 2 in the SFR) we

999: refer to \citet{vanderwel07}.  We conservatively use $50~\mu{\rm{Jy}}$

1000: as the flux limit, as at lower flux levels the poor constraint on the

1001: positions of $24~\mu\rm{m}$ sources requires verification of the

1002: optical counterpart for each individual object, which is beyond the

1003: scope of this paper.

1004: 

1005: \begin{figure*}

1006: \epsscale{1.1}

1007: \plottwo{f11a.eps}{f11b.eps}

1008: \caption{ \textit{Top left:} SFR (derived from $24~\mu\rm{m}$

1009:  photometry) as a function of rest-frame color for the $z\sim 0.8$

1010:  sample.  The red circles are early-type galaxies, the blue circles

1011:  are late-type galaxies. Filled circles are objects with significant

1012:  ($>5\sigma$) $24~\mu\rm{m}$ detections, open circles are objects

1013:  without significant detections (upper limits are shown).

1014:  \textit{Bottom left:} For the same sample as in the top left panel,

1015:  the specific SFR as a function of color.  \textit{Top right:} SFR

1016:  (derived from modeling optical emission lines; Brinchmann et

1017:  al. 2004) as a function of color for the $z\sim 0.03$ sample. Open

1018:  red circles are early-type galaxies, and filled blue circles are

1019:  late-type galaxies.  \textit{Bottom right:} For the same sample as in

1020:  the top right panel, the specific SFR as a function of color. For

1021:  both the low- and high-redshift samples, the star formation activity

1022:  is low in early-type galaxies.  SFRs are higher for late types, with

1023:  a marked decrease in SFR for red late-type galaxies between $z=0.8$

1024:  and the present. The dotted lines in the two bottom panels indicate

1025:  the expected specific SFRs associated with galaxies that will

1026:  double/have doubled their stellar masses between $z=1$ and the

1027:  present under the assumption that their SFRs decline (exponentially

1028:  as a function of redshift) by an order of magnitude between $z=1$ and

1029:  the present. The decreased star formation activity in our $z\sim

1030:  0.03$ sample with respect to our $z\sim 0.8$ sample is consistent

1031:  with the decline of the cosmic SFR.}

1032: \label{UB_SFR}

1033: \end{figure*}

1034: 

1035: 

1036: 

1037: In Figure \ref{UB_SFR} we show the SFR and the SFR per unit stellar

1038: mass, the specific SFR, as a function of rest-frame $U-B$ color for

1039: the 187 (of 207) $z\sim 0.8$ galaxies with MIPS coverage. Of the

1040: early-type galaxies, 11\% are detected at $50~\mu{\rm{Jy}}$, at least

1041: half of which are likely active galactic nuclei \citep{vanderwel07}.

1042: Of the late-type galaxies, 76\% have fluxes $>50~\mu{\rm{Jy}}$, the

1043: majority of which are most likely due to star formation. The typical

1044: SFR of a late-type galaxy is $\sim 20~M_{\odot}~\rm{yr}^{-1}$, whereas

1045: that of an early type is $\lesssim 5~M_{\odot}~\rm{yr}^{-1}$.  The

1046: specific SFRs are $\lesssim 10^{-10}~\rm{yr}^{-1}$ for early types and

1047: $\lesssim 10^{-9}~\rm{yr}^{-1}$ for late types.  These values imply an

1048: increase in stellar mass of less than 10\% per Gyr for the early-type

1049: galaxies and up to 100\% for the late types (see Fig. \ref{UB_SFR}).

1050: These findings are consistent with the work by \citet{bell05}, who

1051: show that the star formation activity at $z\sim 0.7$ is mainly due to

1052: star formation in normal spiral galaxies with masses $>2\times

1053: 10^{10}~M_{\odot}$.

1054: 

1055: \citet{brinchmann04} derived SFRs for galaxies in the SDSS by modeling

1056: their optical emission lines. The most robust determination of the SFR

1057: for nearby galaxies such as those in our $z\sim 0.03$ sample is that

1058: measured within the spectroscopic fiber of the SDSS spectrograph. We

1059: use the stellar mass derived within the same fiber aperture

1060: \citep{kauffmann03a},\footnote{The stellar masses and SFRs are

1061: publicly available at http://www.mpa-garching.mpg.de/SDSS/DR4/} and

1062: scale the SFR, comparing the fiber stellar mass with our total stellar

1063: mass (\S~\ref{secsdss}). We show the SFR and specific SFR in Figure

1064: \ref{UB_SFR}.  The global trend is similar to that observed at $z\sim

1065: 0.8$, with early-type galaxies showing lower star formation activity

1066: than late-type galaxies.

1067: 

1068: It should be kept in mind that SFRs of the $z\sim 0.03$ galaxies are

1069: determined from optical emission lines as measured with a $3''$ fiber,

1070: such that extended star formation can be missed, and a systematic

1071: difference between the total SFRs of the $z\sim 0.8$ galaxies is

1072: created. In addition, it is well known that IR-derived and emission-

1073: line-derived SFRs can be intrinsically different and depend on

1074: metallicity and the amount of extinction.  Therefore, we have to

1075: restrict ourselves to the description and interpretation of

1076: evolutionary trends of at least an order of magnitude.

1077: 

1078: Despite the large uncertainties and the high upper limit of

1079: $5~M_{\odot}~\rm{yr}^{-1}$ for the $z\sim 0.8$ sample, Figure

1080: \ref{UB_SFR} shows that the overall star formation activity decreases

1081: between $z\sim 0.8$ and the present (see also Table \ref{tab2}).  One

1082: prominent difference between the $z\sim 0.03$ and $\sim 0.8$ samples

1083: is the marked decrease of star formation activity by 2 orders of

1084: magnitude in late-type galaxies that are located on the red sequence.

1085: In the local universe, their SFRs are as low as those of early-type

1086: galaxies, whereas at $z\sim 0.8$ their SFRs are at least an order of

1087: magnitude higher than those of the early types.

1088: 

1089: In Figure \ref{UB_SFR} we indicate the specific SFR expected for

1090: galaxies that double their stellar mass between $z=1$ and the present,

1091: assuming an exponential decline in SFR by a factor of 10 over that

1092: period. Such a level of star formation, both at $z\sim 0.03$ and at

1093: $z\sim 0.8$, coincides with the typical specific SFR of late-type

1094: galaxies.  The decrease in star formation activity as observed in our

1095: samples follows the cosmic average trend of decline by an order of

1096: magnitude since $z\sim 1$. At the same time, the morphological mix has

1097: remained unchanged, at least, if only two morphological classes are

1098: considered. These results are therefore not mutually exclusive and

1099: have to be reconciled with each other.

1100: 

1101: 

1102: \begin{figure*}

1103: \epsscale{1.1} 

1104: \plottwo{f12a.eps}{f12b.eps}

1105: \caption{ Stellar mass vs. rest-frame $U-B$ color for the $z\sim 0.8$

1106:  sample (\textit{left}) the $z\sim 0.03$ sample (\textit{right}).

1107:  Early-type galaxies are indicated by filled red circles symbols,

1108:  late-type galaxies by open blue circles. Color evolution is readily

1109:  visible, with the red sequence being significantly bluer at $z\sim

1110:  0.8$ than at $z\sim 0.03$. The solid lines indicate the best fits to

1111:  the red sequence as defined by the early-type galaxies, and the

1112:  dashed lines separate blue and red galaxies (see text for an

1113:  explanation). The relative numbers of red galaxies in the $z\sim

1114:  0.03$ and $\sim 0.8$ samples are not significantly different (see

1115:  also Table \ref{tab2}). In addition, the fraction of red galaxies

1116:  with late-type morphologies is similar in both samples.}

1117: \label{MCD}

1118: \end{figure*}

1119: 

1120: \subsection{The Red Sequence and the Evolution of the Stellar Mass Density}\label{seccolormass}

1121: 

1122: In Figs. \ref{MCD} we show the rest-frame $U-B$ color distribution as

1123: a function of stellar mass for the low- and high-redshift samples.

1124: The $U-B$ colors of the $z\sim 0.8$ sample are corrected for evolution

1125: within the sample redshift range $0.6<z<1.0$, assuming

1126: $\Delta(U-B)=0.22(z-0.8)$ (see below).  This is a small effect of

1127: typically 0.02 mag. We define the early-type galaxy red sequence for

1128: the $z\sim 0.03$ sample by a linear fit, with the slope and the zero

1129: point as free parameters, and iteratively rejecting $3\sigma$ outliers

1130: (8\% are rejected, virtually all blueward of the red sequence). In

1131: order to minimize systematic effects, we define the red sequence for

1132: the $z\sim 0.8$ sample with the same slope (0.08 mag dex$^{-1}$) and

1133: scatter (0.05 mag), with only the zero point as a free parameter.  The

1134: red sequence for both samples is shown in Figure \ref{MCD}.  Color

1135: evolution between $z\sim 0.8$ and $\sim 0.03$ of 0.15 mag, or 0.20 mag

1136: per unit redshift, is readily visible. We also fit the slope and

1137: scatter of the $z\sim 0.8$ red sequence independently, resulting in a

1138: higher scatter (0.08 mag) and a steeper slope (0.11 mag dex$^{-1}$).

1139: 

1140: For both samples we define blue galaxies as galaxies with colors more

1141: than 2 $\sigma$ below the red sequence (with $\sigma=0.05$ as defined

1142: above; see Fig. \ref{MCD}). With this definition, we find similar

1143: red-galaxy fractions of 64\% and 68\% for the $z\sim 0.8$ and $\sim

1144: 0.03$ samples, respectively (see Table \ref{tab2}).  Like the

1145: early-type fraction, the red-galaxy fraction shows no strong evolution

1146: with redshift. The red-galaxy fraction is significantly higher than

1147: the early-type fraction in both samples \citep[see

1148: also][]{bundy06}. This is due to the presence of many late-type

1149: galaxies on the red sequence: $38\%\pm 8\%$ for the $z\sim 0.8$ sample

1150: and $45\%\pm 3\%$ for the $z\sim 0.03$ sample of the red galaxies are

1151: late types (see also Fig. \ref{UB_N}). Conversely, in either sample

1152: less than 20\% of the blue galaxies have early-type morphologies.

1153: 

1154: Given the fact that the majority of the galaxies in both our samples

1155: are located on the red sequence, our earlier conclusion that the

1156: early-type galaxy fraction at a given density does not evolve with

1157: redshift seems, at first sight, to be at odds with the observation

1158: that the stellar mass density of red-sequence galaxies has doubled

1159: since $z=1$ \citep[e.g.,][]{bell04b,brown07}.  The luminosity limit

1160: for red galaxies at $z=1$ in our sample ($M_{\rm{B}}\sim -20$) is

1161: similar as that of, e.g., \citet{bell04b} and \citet{brown07};

1162: therefore, the increase in mass density found by those authors has to

1163: take place above the mass threshold of our samples. In fact, even in

1164: our samples, which cover only small volumes, we observe an increase in

1165: stellar mass density from $z=0.8$ to the present.

1166: 

1167: If the mass function of galaxies evolves, it is not entirely

1168: appropriate to compare galaxy samples at different redshifts down to

1169: the same mass limit. If galaxies evolve by a factor of 2 in stellar

1170: mass between $z=0.08$ and the present, the mass limit at $z=0.8$

1171: should be chosen 0.3 dex lower than at $z=0$, otherwise many of the

1172: progenitors in the $z=0$ sample will not be included in the $=0.8$

1173: sample. We test this effect by choosing a mass limit for our $z\sim

1174: 0.03$ sample that is 0.3 dex higher ($M=8\times 10^{10}~M_{\odot}$)

1175: than for the $z\sim 0.8$ sample. The early-type fraction increases

1176: slightly, from $43\%\pm3\%$ to $48\%\pm3\%$. Due to the errors it is

1177: not clear whether this agrees better with the $z\sim 0.8$ early-type

1178: galaxy fraction ($48\%\pm7\%$), but it is striking that the values are

1179: the same.

1180: 

1181: We conclude that the growth of the red-sequence galaxy population must

1182: happen in such a way that the MDR is conserved. The accretion of red

1183: galaxies onto the red sequence and/or the growth of galaxies on the

1184: red sequence must preserve the relative number of late- and early-type

1185: galaxies with masses $M>4\times 10^{10}~M_{\odot}$. The galaxy

1186: population evolves along the MDR, whereas the MDR itself is a fixed

1187: relationship between environment and morphological appearance.

1188: 

1189: \begin{figure*}

1190: \epsscale{1.1}

1191: \plottwo{f13a.eps}{f13b.eps}

1192: \caption{ Rest-frame $U-B$ color distribution of the mass selected

1193:   $z\sim 0.8$ (\textit{left}) and $z\sim 0.03$ (\textit{right})

1194:   samples. The solid lines indicate the early-type galaxies, and the

1195:   dashed lines indicate the late-type galaxies. The relative numbers

1196:   of red galaxies and the relative numbers of red galaxies with

1197:   late-type morphologies are not significantly different at $z\sim

1198:   0.03 $ and $\sim 0.8$. }

1199: \label{UB_N}

1200: \end{figure*}

1201: 

1202: Since many star-forming, late-type galaxies are located on the red

1203: sequence, the stellar mass density of red-sequence galaxies will

1204: increase through \textit{in situ} star formation.  The question is how

1205: much this contributes to the observed increase in stellar mass density

1206: of red galaxies.  If we make the assumption that the SFR of the

1207: galaxies in our $z\sim 0.8$ sample decreases exponentially at the same

1208: rate as the cosmic average, i.e., by an order of magnitude between

1209: $z=1$ and the present, the average late-type galaxy on the red

1210: sequence will increase its stellar mass by $20\%-80\%$. The large

1211: uncertainty is caused by the absolute uncertainty in the IR-derived

1212: SFR of a factor of 2 (see \S~\ref{secsf}).  Given such an increase in

1213: stellar mass, there will be galaxies that are below our mass cutoff at

1214: $z\sim 0.8$ but would not be during the present epoch. We can estimate

1215: this effect by calculating the total stellar mass in red, late-type

1216: galaxies in our $z\sim 0.03$ sample with masses less than $1.2-1.8$

1217: times the threshold. Their contribution to the total stellar mass in

1218: red galaxies is $2\%-11\%$, such that the total increase in the

1219: stellar mass density through \textit{in situ} star formation in red

1220: late types is a factor of $1.2-2$. This is close the observed

1221: evolution of a factor of $\sim$2. One caveat is the possible evolution

1222: with redshift of the color-$M/L$ relation, which we use to estimate

1223: the stellar masses (see \S~\ref{secdatacdfs}). This may cancel out the

1224: growth in stellar mass as estimated above.

1225: 

1226: With the hypothesis that merging does not play a role in shaping the

1227: red sequence galaxy population, the above described process of star

1228: formation in late types, but not in early types, will change the

1229: morphological mix of a mass-selected sample of red galaxies. The

1230: early-type fraction will then decrease with time or increase with

1231: redshift. As mentioned in \S~\ref{seccolormass}, the fraction of

1232: early-type galaxies on the red sequence is indeed consistent with an

1233: increase from $z\sim 0.03$ to $\sim 0.8$ (from 55\% to 62\%). Even

1234: though this difference is not significant, at least it is consistent

1235: with a growth in stellar mass of late-type galaxies by a factor of

1236: 1.5.

1237: 

1238: These speculations depend on the assumption that a large reservoir of

1239: lower mass, red galaxies is available to provide the necessary growth

1240: of stellar mass through star formation on the red sequence. Given the

1241: evidence for the opposite, i.e., the lack of faint, red galaxies at

1242: significant look-back times \citep[e.g.,][]{tanaka05,delucia07},

1243: truncation of star formation in blue galaxies is possibly required to

1244: explain the growth of the red sequence. The above estimates of growth

1245: through \textit{in situ} star formation merely serve as a cautionary

1246: statement to indicate that a factor of 2 evolution in the red-galaxy

1247: mass density is rather modest and can be accounted for in various

1248: ways.

1249: 

1250: Of course, if, for some reason, current measurements overestimate the

1251: evolution in the red-galaxy population, our results can be regarded as

1252: evidence for passive evolution of red galaxies, in terms of both star

1253: formation and mass assembly. We stress, however, that this is most

1254: likely not the case, given the mounting evidence for considerable

1255: evolution in the stellar mass density of red galaxies since $z=1$.

1256: 

1257: 

1258: \subsection{Comparison with Other Studies}

1259: Numerous studies have analyzed the morphological and structural

1260: properties of galaxies in the local universe and at high

1261: redshifts. Even though it is not feasible to discuss all previous

1262: work, we put our work on the local galaxy population in the context of

1263: several other studies and compare our results with those from several

1264: recent efforts with the similar objective to measure the evolution

1265: with redshift of the early-type fraction outside massive clusters.

1266: 

1267: \citet{goto03} derive the MDR for luminosity-selected galaxies at

1268: $0.05<z<0.10$. Their morphological classifications are simply based on

1269: the concentration of the light profile, and they find early-type

1270: galaxy fractions of $\sim40\%-60\%$ at low densities, depending on the

1271: strictness of the classification criterion for early types. This is

1272: consistent with our results. \citet{goto03} note that at densities

1273: below $\Sigma\sim 1~\rm{Mpc}^{-2}$ the early-type fraction changes

1274: less rapidly with density than at higher densities. This trend is even

1275: more pronounced in our mass-selected sample: at densities

1276: $\Sigma<1~\rm{Mpc}^{-2}$ the early-type fraction is constant at $\sim

1277: 40\%$ and only increases at higher densities (see Fig.

1278: \ref{morphdens} and \S~\ref{secmdr}).

1279: 

1280: The contributions of galaxies with different morphologies to the total

1281: stellar mass density have been measured by \citet{bell03}, who find

1282: that nowadays 76\% of the stars in galaxies more massive than

1283: $M=4\times 10^{10}~M_{\odot}$ reside in early-type galaxies. In our

1284: $z\sim 0.03$ sample, we find a much lower value of 47\%. The cause of

1285: this difference is that Bell et al. use a single parameter (the

1286: concentration index) as a proxy for morphology, which corresponds with

1287: visual classifications 70\% of the time \citep[see

1288: also][]{strateva01}. Our method uses the bumpiness parameter $B$ in

1289: addition to the PSF-corrected concentration of the light profile,

1290: quantified by the S\'ersic parameter $n$.  Thus, early-type galaxies

1291: are required to be not only highly concentrated but also smooth.  This

1292: extension increases the agreement with visual classifications to

1293: $\sim$90\% (see \S~\ref{secmorph}).  This improvement is not due to

1294: smaller random errors: the contribution to the total stellar mass

1295: density of highly concentrated galaxies in our $z\sim 0.03$ sample

1296: (those with $n>2.5$) is 70\%, in good agreement with the result by

1297: \citet{bell03}, and higher than that of early-type galaxies as defined

1298: by both visual classifications and by our $B-n$ method. Many of the

1299: galaxies with $n>2.5$ have late-type morphologies, which is only

1300: revealed through visual classifications or the use of an additional

1301: parameter such as bumpiness.  As was demonstrated by

1302: \citet{kauffmann04}, the concentration distribution of the galaxy

1303: population provides crucial insight into the formation and evolution

1304: of galaxies. However, this structural property should be distinguished

1305: from morphology. The latter is related to the smoothness of the light

1306: profile in addition to its concentration.  It may well be that

1307: structure and morphology are correlated but distinct physical

1308: parameters that are affected by, e.g., galaxy mass and environment in

1309: different ways.

1310: 

1311: Recently, several studies have addressed the evolution with redshift

1312: of the morphological mix of field galaxies.  \citet{abraham07} find

1313: that for galaxies more massive than $M=4\times 10^{10}~M_{\odot}$ the

1314: fraction of the stellar mass density residing in early-type galaxies

1315: is $\sim$80\% at $z\sim 1$, similar to the numbers from \citet{bell03}

1316: and our $n$-selected sample at low redshift.  Abraham et al.  use two

1317: parameters to distinguish early types from late types: in addition to

1318: the Gini coefficient (which, for early types and spiral galaxies,

1319: closely corresponds to the concentration index) they use asymmetry as

1320: a morphological indicator.  However, the addition of asymmetry does

1321: not substantially improve the agreement with visual classifications

1322: for galaxies with type Sb or later \citep[see, e.g.,][]{conselice03}.

1323: The method used by \citet{abraham07} therefore effectively classifies

1324: galaxies by concentration.  In our $z\sim 0.8$ sample, highly

1325: concentrated galaxies (with $n>2.5$) contribute 72\% to the total

1326: stellar mass density, similar to what Abraham et al. find

1327: ($\sim$80\%).  The stellar mass in early-type galaxies as defined by

1328: our $B-n$ method or visual classifications is lower (57\%), consistent

1329: with the results by \citet{bundy05}, which are also based on visual

1330: classifications.  We conclude that the difference between our numbers

1331: and those from \citet{abraham07} are due to the difference between the

1332: classifying methods.  Note that the relative early-type galaxy mass

1333: density does not evolve between $z\sim 1$ and the present, independent

1334: of the applied morphological classification method.

1335: 

1336: \citet{capak07}, who use the Gini coefficient to determine

1337: morphologies, find a significant decrease with redshift (in the range

1338: $0.4<z<1.2$) of the early-type fraction in high-density regions (with

1339: $\Sigma>100~\rm{Mpc^{-2}}$).  At lower densities (with

1340: $\Sigma<50~\rm{Mpc^{-2}}$) they find a constant early-type fraction

1341: between $z=0.4$ and 1.2 that is consistent with the early-type

1342: fraction that we find.  However, when they compare their results with

1343: the local early-type fraction, measured with different classification

1344: methods, they do find significant evolution, also at the lowest

1345: densities. The sample constructed by \citet{capak07} is selected by

1346: luminosity, and is therefore not directly comparable with our

1347: sample. As was demonstrated above, selecting early-type galaxies by

1348: Gini coefficient or concentration index alone overestimates the number

1349: of early-type galaxies as compared to visual classifications or our

1350: $B-n$ method. This, combined with the difference between luminosity-

1351: and mass-selected samples, may conspire to yield an early-type

1352: fraction similar to that we find for visually classified galaxies in a

1353: mass-selected sample.

1354: 

1355: We conclude that our results are either consistent with previous work

1356: or can be explained by the differences in approach. We note that the

1357: GOODS/ACS imaging that we use to determine galaxy morphologies at

1358: $z\sim 0.8$ is much deeper than then imaging used by \citet{capak07}

1359: and effectively as deep as the data used by \citet{abraham07} in terms

1360: of $S/N$ for red galaxies at $z\sim 1$. Moreover, our study is the

1361: first of its kind to quantify galaxy morphologies in the local

1362: universe and at higher redshifts in an internally consistent manner.

1363: 

1364: 

1365: \section{SUMMARY}

1366: 

1367: In this paper we examine the early-type galaxy (E+S0) fraction and its

1368: evolution for stellar-mass-selected, volume-limited samples of

1369: galaxies at redshifts $0<z<1$.  At low redshift we estimate stellar

1370: masses and determine morphologies for 2003 galaxies at redshifts

1371: $0.02<z<0.045$ in the SDSS, complete down to a mass of $M=4\times

1372: 10^{10}~M_{\odot}$, 60\% of the mass of a typical ($L_*$) galaxy.  In

1373: addition, we construct a similar sample of 207 galaxies in the CDF-S

1374: at redshifts $0.6<z<1.0$, complete down to the same mass limit.  The

1375: stellar mass estimates are based on the relation between color and

1376: $M/L$, which has been proved to work without significant systematic

1377: effects with respect to kinematic mass measurements at low and high

1378: redshift.  Morphologies are determined with an automated method, based

1379: on the S\'ersic parameter and the bumpiness of the residual.  We

1380: compare this method with the traditional visual classifications and

1381: find no systematic differences when we split the sample into two

1382: morphological classes (Sp+Irr and E+S0).  We estimate the local

1383: surface density with an $n$th nearest neighbor method. The density for

1384: the galaxies in our samples is typically $\Sigma \sim 1~\rm{Mpc^{-2}}$

1385: and ranges from $\sim$0.01 to $100~\rm{Mpc^{-2}}$. The sample

1386: characteristics are summarized in Table \ref{tab2}.

1387: 

1388: We find that for galaxies with masses higher than $M=4\times

1389: 10^{10}~M_{\odot}$ the early-type fraction in the field and group

1390: environment probed by our samples has not changed significantly

1391: between $z=0.8$ and the present. The early-type fraction is

1392: $43\%\pm3\%$ for the $z\sim 0.03$ sample, and $48\%\pm7\%$ for the

1393: $z\sim 0.8$ sample. When we combine this with the unchanging

1394: early-type fraction of $>80\%$ in dense ($>100~\rm{Mpc}^{-2}$) cluster

1395: environments over the same redshift range and down to the same mass

1396: limit (Paper I), we find that the MDR has not evolved significantly

1397: between $z\sim 0.8$ and the present over at least 3 orders of

1398: magnitude in density.

1399: 

1400: In both our $z\sim 0.03$ sample and our $z\sim 0.8$ sample,

1401: $65\%-70\%$ of the galaxies (see also Table \ref{tab2}) are located on

1402: the red sequence, in agreement with previous work. In addition,

1403: $65\%-70\%$ of the galaxies in both samples have S\'ersic indices

1404: $n>2.5$.  Our samples follow the well-constrained trends that the

1405: stellar mass density of red galaxies has increased and the global star

1406: formation activity has decreased between $z\sim 1$ and the present.

1407: In fact, star formation in red galaxies could contribute significantly

1408: to the growth in the stellar mass density of red galaxies without

1409: strong evolution in the early-type galaxy fraction.

1410: 

1411: Future work will include a more detailed analysis of the morphological

1412: mix, e.g., the separation of the late types into irregular and spiral

1413: galaxies and the separation of the early types into E and S0 galaxies.

1414: Furthermore, studies in larger fields with high-resolution imaging

1415: will be used to measure the shape of the MDR at low densities in more

1416: detail.  In particular, it is important to study intermediately dense

1417: regions, i.e., the infalling regions around clusters. At those

1418: densities the morphological mix of the galaxy population appears to

1419: undergo the most rapid change as a function of density.  A crucial

1420: step forward in understanding the role of the environment and the role

1421: of the internal properties of galaxies in shaping their stellar

1422: populations and morphologies will be to include the evolution of the

1423: density itself in analyses such as those carried out in this paper.

1424: Another challenge will be to establish at what redshift the MDR

1425: emerges, whether this coincides with the emergence of the red

1426: sequence, and how this relates to the buildup of the stellar mass

1427: function.

1428: 

1429: \acknowledgements { We thank the referee, Bob Abraham, for his

1430:  thorough reading of the manuscript and positive feedback. We would

1431:  like to thank Pieter van Dokkum for helpful suggestions and Sandy

1432:  Faber for stimulating discussion.  A. v. d. W. acknowledges support

1433:  from NASA grant NAG5-7697.}

1434: 

1435: \bibliographystyle{apj}

1436: \begin{thebibliography}{69}

1437: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1438: 

1439: \bibitem[{{Abraham} {et~al.}(2007){Abraham}, {McCarthy}, {Glazebrook},

1440:   {Mentuch}, {Nair}, {Yan}, {Savaglio}, {Crampton}, {Murowinski}, {Juneau}, {Le

1441:   Borgne}, {Carlberg}, {Jorgensen}, {Roth}, {Chen}, \& {Marzke}}]{abraham07}

1442: {Abraham}, R.~G. et al. 2007, astro-ph/0704.1475

1443: 

1444: \bibitem[{{Abraham} {et~al.}(2003){Abraham}, {van den Bergh}, \&

1445:   {Nair}}]{abraham03}

1446: {Abraham}, R.~G., {van den Bergh}, S., \& {Nair}, P. 2003, \apj, 588, 218

1447: 

1448: \bibitem[{{Adelman-McCarthy} {et~al.}(2007)}]{adelman07}

1449: Adelman-McCarthy, J. et al. 2007, \apjs, 172, 634 

1450: 

1451: \bibitem[{{Baldry} {et~al.}(2006){Baldry}, {Balogh}, {Bower}, {Glazebrook},

1452:   {Nichol}, {Bamford}, \& {Budavari}}]{baldry06}

1453: {Baldry}, I.~K., {Balogh}, M.~L., {Bower}, R.~G., {Glazebrook}, K., {Nichol},

1454:   R.~C., {Bamford}, S.~P., \& {Budavari}, T. 2006, \mnras, 373, 469

1455: 

1456: \bibitem[{{Bell} {et~al.}(2004{\natexlab{a}}){Bell}, {McIntosh}, {Barden},

1457:   {Wolf}, {Caldwell}, {Rix}, {Beckwith}, {Borch}, {H{\" a}ussler}, {Jahnke},

1458:   {Jogee}, {Meisenheimer}, {Peng}, {Sanchez}, {Somerville}, \&

1459:   {Wisotzki}}]{bell04a}

1460: {Bell}, E.~F. et al. 2004{\natexlab{a}}, \apjl, 600, L11

1461: 

1462: \bibitem[{{Bell} {et~al.}(2003){Bell}, {McIntosh}, {Katz}, \&

1463:   {Weinberg}}]{bell03}

1464: {Bell}, E.~F., {McIntosh}, D.~H., {Katz}, N., \& {Weinberg}, M.~D. 2003, \apjs,

1465:   149, 289

1466: 

1467: \bibitem[{{Bell} {et~al.}(2005){Bell}, {Papovich}, {Wolf}, {Le Floc'h},

1468:   {Caldwell}, {Barden}, {Egami}, {McIntosh}, {Meisenheimer},

1469:   {P{\'e}rez-Gonz{\'a}lez}, {Rieke}, {Rieke}, {Rigby}, \& {Rix}}]{bell05}

1470: {Bell}, E.~F. et al. 2005, \apj, 625, 23

1471: 

1472: \bibitem[{{Bell} {et~al.}(2004{\natexlab{b}}){Bell}, {Wolf}, {Meisenheimer},

1473:   {Rix}, {Borch}, {Dye}, {Kleinheinrich}, {Wisotzki}, \& {McIntosh}}]{bell04b}

1474: {Bell}, E.~F. et al.  2004{\natexlab{b}}, \apj, 608, 752

1475: 

1476: \bibitem[{{Ben{\'{\i}}tez} {et~al.}(2004){Ben{\'{\i}}tez}, {Ford}, {Bouwens},

1477:   {Menanteau}, {Blakeslee}, {Gronwall}, {Illingworth}, {Meurer}, {Broadhurst},

1478:   {Clampin}, {Franx}, {Hartig}, {Magee}, {Sirianni}, {Ardila}, {Bartko},

1479:   {Brown}, {Burrows}, {Cheng}, {Cross}, {Feldman}, {Golimowski}, {Infante},

1480:   {Kimble}, {Krist}, {Lesser}, {Levay}, {Martel}, {Miley}, {Postman}, {Rosati},

1481:   {Sparks}, {Tran}, {Tsvetanov}, {White}, \& {Zheng}}]{benitez04}

1482: {Ben{\'{\i}}tez}, N. et al. 2004, \apjs, 150, 1

1483: 

1484: \bibitem[{{Bernardi} {et~al.}(2003){Bernardi}, {Sheth}, {Annis}, {Burles},

1485:   {Eisenstein}, {Finkbeiner}, {Hogg}, {Lupton}, {Schlegel}, {SubbaRao},

1486:   {Bahcall}, {Blakeslee}, {Brinkmann}, {Castander}, {Connolly}, {Csabai},

1487:   {Doi}, {Fukugita}, {Frieman}, {Heckman}, {Hennessy}, {Ivezi{\'c}}, {Knapp},

1488:   {Lamb}, {McKay}, {Munn}, {Nichol}, {Okamura}, {Schneider}, {Thakar}, \&

1489:   {York}}]{bernardi03a}

1490: {Bernardi}, M. et al. 2003,

1491:   \aj, 125, 1817

1492: 

1493: \bibitem[{{Bertin} \& {Arnouts}(1996)}]{bertin96}

1494: {Bertin}, E., \& {Arnouts}, S. 1996, \aaps, 117, 393

1495: 

1496: \bibitem[{{Blakeslee} {et~al.}(2006){Blakeslee}, {Holden}, {Franx}, {Rosati},

1497:   {Bouwens}, {Demarco}, {Ford}, {Homeier}, {Illingworth}, {Jee}, {Mei},

1498:   {Menanteau}, {Meurer}, {Postman}, \& {Tran}}]{blakeslee06}

1499: {Blakeslee}, J.~P. et al. 2006, \apj, 644, 30

1500: 

1501: \bibitem[{{Blanton} {et~al.}(2003){Blanton}, {Brinkmann}, {Csabai}, {Doi},

1502:   {Eisenstein}, {Fukugita}, {Gunn}, {Hogg}, \& {Schlegel}}]{blanton03}

1503: {Blanton}, M.~R. et al. 2003, \aj,

1504:   125, 2348

1505: 

1506: \bibitem[{{Blanton} {et~al.}(2005{\natexlab{a}}){Blanton}, {Eisenstein},

1507:   {Hogg}, {Schlegel}, \& {Brinkmann}}]{blanton05a}

1508: {Blanton}, M.~R., {Eisenstein}, D., {Hogg}, D.~W., {Schlegel}, D.~J., \&

1509:   {Brinkmann}, J. 2005{\natexlab{a}}, \apj, 629, 143

1510: 

1511: \bibitem[{{Blanton} {et~al.}(2005{\natexlab{b}}){Blanton}, {Lupton},

1512:   {Schlegel}, {Strauss}, {Brinkmann}, {Fukugita}, \& {Loveday}}]{blanton05b}

1513: {Blanton}, M.~R., {Lupton}, R.~H., {Schlegel}, D.~J., {Strauss}, M.~A.,

1514:   {Brinkmann}, J., {Fukugita}, M., \& {Loveday}, J. 2005{\natexlab{b}}, \apj,

1515:   631, 208

1516: 

1517: \bibitem[{{Brinchmann} {et~al.}(2004){Brinchmann}, {Charlot}, {White},

1518:   {Tremonti}, {Kauffmann}, {Heckman}, \& {Brinkmann}}]{brinchmann04}

1519: {Brinchmann}, J., {Charlot}, S., {White}, S.~D.~M., {Tremonti}, C.,

1520:   {Kauffmann}, G., {Heckman}, T., \& {Brinkmann}, J. 2004, \mnras, 351, 1151

1521: 

1522: \bibitem[{{Brown} {et~al.}(2007){Brown}, {Dey}, {Jannuzi}, {Brand}, {Benson},

1523:   {Brodwin}, {Croton}, \& {Eisenhardt}}]{brown07}

1524: {Brown}, M.~J.~I. et al. 2007, \apj, 654, 858

1525: 

1526: \bibitem[{{Bruzual} \& {Charlot}(2003)}]{bruzual03}

1527: {Bruzual}, G., \& {Charlot}, S. 2003, \mnras, 344, 1000

1528: 

1529: \bibitem[{{Bundy} {et~al.}(2005){Bundy}, {Ellis}, \& {Conselice}}]{bundy05}

1530: {Bundy}, K., {Ellis}, R.~S., \& {Conselice}, C.~J. 2005, \apj, 625, 621

1531: 

1532: \bibitem[{{Bundy} {et~al.}(2006){Bundy}, {Ellis}, {Conselice}, {Taylor},

1533:   {Cooper}, {Willmer}, {Weiner}, {Coil}, {Noeske}, \& {Eisenhardt}}]{bundy06}

1534: {Bundy}, K. et al. 2006, \apj, 651, 120

1535: 

1536: \bibitem[{{Capak} {et~al.}(2007){Capak}, {Abraham}, {Ellis}, {Mobasher},

1537:   {Scoville}, {Sheth}, \& {Koekemoer}}]{capak07}

1538: {Capak}, P.~L., {Abraham}, R.~G., {Ellis}, R.~S., {Mobasher}, B., {Scoville},

1539:   N.~Z., {Sheth}, K., \& {Koekemoer}, A. 2007, astro-ph/0703668

1540: 

1541: \bibitem[{{Chary} \& {Elbaz}(2001)}]{chary01}

1542: {Chary}, R., \& {Elbaz}, D. 2001, \apj, 556, 562

1543: 

1544: \bibitem[{{Conselice}(2003)}]{conselice03}

1545: {Conselice}, C.~J. 2003, \apjs, 147, 1

1546: 

1547: \bibitem[{{Conselice} {et~al.}(2000){Conselice}, {Bershady}, \&

1548:   {Jangren}}]{conselice00}

1549: {Conselice}, C.~J., {Bershady}, M.~A., \& {Jangren}, A. 2000, \apj, 529, 886

1550: 

1551: \bibitem[{{De Lucia} {et~al.}(2007){De Lucia}, {Poggianti},

1552:   {Arag{\'o}n-Salamanca}, {White}, {Zaritsky}, {Clowe}, {Halliday}, {Jablonka},

1553:   {von der Linden}, {Milvang-Jensen}, {Pell{\'o}}, {Rudnick}, {Saglia}, \&

1554:   {Simard}}]{delucia07}

1555: {De Lucia}, G. et al. 2007, \mnras, 374, 809

1556: 

1557: \bibitem[{{De Propris} {et~al.}(2007){De Propris}, {Stanford}, {Eisenhardt},

1558:   {Holden}, \& {Rosati}}]{depropris07}

1559: {De Propris}, R., {Stanford}, S.~A., {Eisenhardt}, P.~R., {Holden}, B., \&

1560:   {Rosati}, P. 2007, \aj, 133, 2209

1561: 

1562: \bibitem[{{Desai} {et al.}(2007){Desai}, {Dalcanton}, {Arag{\'o}n-Salamanca},  

1563: 	{Jablonka}, {Poggianti}, {Gogarten}, {Simard}, {Milvang-Jensen}, 

1564: 	{Rudnick}, {Zaritsky}, {Clowe}, {Halliday}, {Pell{\'o}}, {Saglia}, 

1565: 	{White}}]{desai07}

1566: {Desai}. V. et al. 2007, \apj, 660, 1151

1567: 

1568: \bibitem[{{Dressler}(1980)}]{dressler80}

1569: {Dressler}, A. 1980, \apj, 236, 351

1570: 

1571: \bibitem[{{Dressler} {et~al.}(1997){Dressler}, {Oemler}, {Couch}, {Smail},

1572:   {Ellis}, {Barger}, {Butcher}, {Poggianti}, \& {Sharples}}]{dressler97}

1573: {Dressler}, A. et al. 1997,

1574:   \apj, 490, 577

1575: 

1576: \bibitem[{{Faber et al.}(2007)}]{faber07}

1577: Faber, S.~M. et al. 2007, \apj, 665, 265

1578: 

1579: \bibitem[{{Fazio} {et~al.}(2004){Fazio}, {Hora}, {Allen}, {Ashby}, {Barmby},

1580:   {Deutsch}, {Huang}, {Kleiner}, {Marengo}, {Megeath}, {Melnick}, {Pahre},

1581:   {Patten}, {Polizotti}, {Smith}, {Taylor}, {Wang}, {Willner}, {Hoffmann},

1582:   {Pipher}, {Forrest}, {McMurty}, {McCreight}, {McKelvey}, {McMurray}, {Koch},

1583:   {Moseley}, {Arendt}, {Mentzell}, {Marx}, {Losch}, {Mayman}, {Eichhorn},

1584:   {Krebs}, {Jhabvala}, {Gezari}, {Fixsen}, {Flores}, {Shakoorzadeh}, {Jungo},

1585:   {Hakun}, {Workman}, {Karpati}, {Kichak}, {Whitley}, {Mann}, {Tollestrup},

1586:   {Eisenhardt}, {Stern}, {Gorjian}, {Bhattacharya}, {Carey}, {Nelson},

1587:   {Glaccum}, {Lacy}, {Lowrance}, {Laine}, {Reach}, {Stauffer}, {Surace},

1588:   {Wilson}, {Wright}, {Hoffman}, {Domingo}, \& {Cohen}}]{fazio04}

1589: {Fazio}, G.~G. et al.  2004, \apjs, 154, 10

1590: 

1591: \bibitem[{{Goto} {et~al.}(2003){Goto}, {Yamauchi}, {Fujita}, {Okamura},

1592:   {Sekiguchi}, {Smail}, {Bernardi}, \& {Gomez}}]{goto03}

1593: {Goto}, T. et al. 2003, \mnras, 346, 601

1594: 

1595: \bibitem[{{Hogg} {et~al.}(2003){Hogg}, {Blanton}, {Eisenstein}, {Gunn},

1596:   {Schlegel}, {Zehavi}, {Bahcall}, {Brinkmann}, {Csabai}, {Schneider},

1597:   {Weinberg}, \& {York}}]{hogg03}

1598: {Hogg}, D.~W. et al. 2003, \apjl, 585, L5

1599: 

1600: \bibitem[{{Holden} {et~al.}(2006){Holden}, {Franx}, {Illingworth}, {Postman},

1601:   {Blakeslee}, {Homeier}, {Demarco}, {Ford}, {Rosati}, {Kelson}, \&

1602:   {Tran}}]{holden06}

1603: {Holden}, B.~P. et al. 2006, \apjl, 642, L123

1604: 

1605: \bibitem[{{Holden} {et~al.}(2007){Holden}, {Illingworth}, {Franx}, {Blakeslee}, 

1606:   {Postman}, {Kelson}, {van der Wel}, {Demarco}, {Magee}, {Tran}, {Zirm}, 

1607:   {Ford}, {Rosati}, \& {Homeier}}]{holden07}

1608: {Holden}, B.~P. et al. 2007, \apj, 670, 195

1609: 

1610: \bibitem[{{J\o rgensen} {et~al.}(1996){J\o rgensen}, {Franx}, \&

1611:   {Kjaergaard}}]{jorgensen96}

1612: {J\o rgensen}, I., {Franx}, M., \& {Kjaergaard}, P. 1996, \mnras, 280, 167

1613: 

1614: \bibitem[{{Juneau et al.}(2005)}]{juneau05}

1615: Juneau, S. et al. 2005, \apjl, 619, L135

1616: 

1617: \bibitem[{{Kassin et al.}(2007)}]{kassin07}

1618: {Kassin, S. A. et al.} 2007, \apj, 660, 35

1619: 

1620: \bibitem[{{Kauffmann} {et~al.}(2003){Kauffmann}, {Heckman}, {White}, {Charlot},

1621:   {Tremonti}, {Brinchmann}, {Bruzual}, {Peng}, {Seibert}, {Bernardi},

1622:   {Blanton}, {Brinkmann}, {Castander}, {Cs{\'a}bai}, {Fukugita}, {Ivezic},

1623:   {Munn}, {Nichol}, {Padmanabhan}, {Thakar}, {Weinberg}, \&

1624:   {York}}]{kauffmann03a}

1625: {Kauffmann}, G. et al. 2003, \mnras, 341, 33

1626: 

1627: \bibitem[{{Kauffmann} {et~al.}(2004){Kauffmann}, {White}, {Heckman},

1628:   {M{\'e}nard}, {Brinchmann}, {Charlot}, {Tremonti}, \&

1629:   {Brinkmann}}]{kauffmann04}

1630: {Kauffmann}, G. et al. 2004, \mnras, 353, 713

1631: 

1632: \bibitem[{{Kennicutt}(1998)}]{kennicutt98}

1633: {Kennicutt}, R.~C. 1998, \araa, 36, 189

1634: 

1635: \bibitem[{{Lauer} {et~al.}(2007){Lauer}, {Gebhardt}, {Faber}, {Richstone},

1636:   {Tremaine}, {Kormendy}, {Aller}, {Bender}, {Dressler}, {Filippenko}, {Green},

1637:   \& {Ho}}]{lauer07}

1638: {Lauer}, T.~R. et al. 2007, \apj, 662, 808

1639: 

1640: \bibitem[{{Le F{\`e}vre} {et~al.}(2004){Le F{\`e}vre}, {Vettolani}, {Paltani},

1641:   {Tresse}, {Zamorani}, {Le Brun}, {Moreau}, {Bottini}, {Maccagni}, {Picat},

1642:   {Scaramella}, {Scodeggio}, {Zanichelli}, {Adami}, {Arnouts}, {Bardelli},

1643:   {Bolzonella}, {Cappi}, {Charlot}, {Contini}, {Foucaud}, {Franzetti},

1644:   {Garilli}, {Gavignaud}, {Guzzo}, {Ilbert}, {Iovino}, {McCracken}, {Mancini},

1645:   {Marano}, {Marinoni}, {Mathez}, {Mazure}, {Meneux}, {Merighi}, {Pell{\`o}},

1646:   {Pollo}, {Pozzetti}, {Radovich}, {Zucca}, {Arnaboldi}, {Bondi}, {Bongiorno},

1647:   {Busarello}, {Ciliegi}, {Gregorini}, {Mellier}, {Merluzzi}, {Ripepi}, \&

1648:   {Rizzo}}]{lefevre04}

1649: {Le F{\`e}vre}, O. et al. 2004, \aap, 428, 1043

1650: 

1651: \bibitem[{{Le Floc'h} {et~al.}(2005){Le Floc'h}, {Papovich}, {Dole}, {Bell},

1652:   {Lagache}, {Rieke}, {Egami}, {P{\'e}rez-Gonz{\'a}lez}, {Alonso-Herrero},

1653:   {Rieke}, {Blaylock}, {Engelbracht}, {Gordon}, {Hines}, {Misselt}, {Morrison},

1654:   \& {Mould}}]{lefloch05}

1655: {Le Floc'h}, E. et al. 2005, \apj, 632, 169

1656: 

1657: \bibitem[{{Lotz} {et~al.}(2004){Lotz}, {Primack}, \& {Madau}}]{lotz04}

1658: {Lotz}, J.~M., {Primack}, J., \& {Madau}, P. 2004, \aj, 128, 163

1659: 

1660: \bibitem[{{Lupton} {et~al.}(2001){Lupton}, {Gunn}, {Ivezi{\'c}}, {Knapp}, \&

1661:   {Kent}}]{lupton01}

1662: {Lupton}, R., {Gunn}, J.~E., {Ivezi{\'c}}, Z., {Knapp}, G.~R., \& {Kent}, S.

1663:   2001, in ASP Conf. Ser. 238: Astronomical Data Analysis Software and Systems

1664:   X, ed. F.~R. {Harnden}, Jr., F.~A. {Primini}, \& H.~E. {Payne}, 269--+

1665: 

1666: \bibitem[{{Madau} {et~al.}(1996){Madau}, {Ferguson}, {Dickinson}, {Giavalisco},

1667:   {Steidel}, \& {Fruchter}}]{madau96}

1668: {Madau}, P., {Ferguson}, H.~C., {Dickinson}, M.~E., {Giavalisco}, M.,

1669:   {Steidel}, C.~C., \& {Fruchter}, A. 1996, \mnras, 283, 1388

1670: 

1671: \bibitem[{{Mignoli} {et~al.}(2005){Mignoli}, {Cimatti}, {Zamorani}, {Pozzetti},

1672:   {Daddi}, {Renzini}, {Broadhurst}, {Cristiani}, {D'Odorico}, {Fontana},

1673:   {Giallongo}, {Gilmozzi}, {Menci}, \& {Saracco}}]{mignoli05}

1674: {Mignoli}, M. e al. 2005, \aap, 437, 883

1675: 

1676: \bibitem[{{Noeske} {et~al.}(2007){Noeske}, {Weiner}, {Faber}, {Papovich},

1677:   {Koo}, {Somerville}, {Bundy}, {Conselice}, {Newman}, {Schiminovich}, {Le

1678:   Floc'h}, {Coil}, {Rieke}, {Lotz}, {Primack}, {Barmby}, {Cooper}, {Davis},

1679:   {Ellis}, {Fazio}, {Guhathakurta}, {Huang}, {Kassin}, {Martin}, {Phillips},

1680:   {Rich}, {Small}, {Willmer}, \& {Wilson}}]{noeske07}

1681: {Noeske}, K.~G. et al. 2007, \apjl, 660, L43

1682: 

1683: \bibitem[{{Peng} {et~al.}(2002){Peng}, {Ho}, {Impey}, \& {Rix}}]{peng02}

1684: {Peng}, C.~Y., {Ho}, L.~C., {Impey}, C.~D., \& {Rix}, H.-W. 2002, \aj, 124, 266

1685: 

1686: \bibitem[{{P{\'e}rez-Gonz{\'a}lez} {et~al.}(2005){P{\'e}rez-Gonz{\'a}lez},

1687:   {Rieke}, {Egami}, {Alonso-Herrero}, {Dole}, {Papovich}, {Blaylock}, {Jones},

1688:   {Rieke}, {Rigby}, {Barmby}, {Fazio}, {Huang}, \& {Martin}}]{perez05}

1689: {P{\'e}rez-Gonz{\'a}lez}, P.~G. et al. 2005, \apj, 630, 82

1690: 

1691: \bibitem[{{Postman} {et~al.}(2005){Postman}, {Franx}, {Cross}, {Holden},

1692:   {Ford}, {Illingworth}, {Goto}, {Demarco}, {Rosati}, {Blakeslee}, {Tran},

1693:   {Ben{\'{\i}}tez}, {Clampin}, {Hartig}, {Homeier}, {Ardila}, {Bartko},

1694:   {Bouwens}, {Bradley}, {Broadhurst}, {Brown}, {Burrows}, {Cheng}, {Feldman},

1695:   {Golimowski}, {Gronwall}, {Infante}, {Kimble}, {Krist}, {Lesser}, {Martel},

1696:   {Mei}, {Menanteau}, {Meurer}, {Miley}, {Motta}, {Sirianni}, {Sparks}, {Tran},

1697:   {Tsvetanov}, {White}, \& {Zheng}}]{postman05}

1698: {Postman}, M. et al. 2005, \apj, 623, 721

1699: 

1700: \bibitem[{{Rieke} {et~al.}(2004){Rieke}, {Young}, {Engelbracht}, {Kelly},

1701:   {Low}, {Haller}, {Beeman}, {Gordon}, {Stansberry}, {Misselt}, {Cadien},

1702:   {Morrison}, {Rivlis}, {Latter}, {Noriega-Crespo}, {Padgett}, {Stapelfeldt},

1703:   {Hines}, {Egami}, {Muzerolle}, {Alonso-Herrero}, {Blaylock}, {Dole}, {Hinz},

1704:   {Le Floc'h}, {Papovich}, {P{\'e}rez-Gonz{\'a}lez}, {Smith}, {Su}, {Bennett},

1705:   {Frayer}, {Henderson}, {Lu}, {Masci}, {Pesenson}, {Rebull}, {Rho}, {Keene},

1706:   {Stolovy}, {Wachter}, {Wheaton}, {Werner}, \& {Richards}}]{rieke04}

1707: {Rieke}, G.~H. et al. 2004, \apjs, 154, 25

1708: 

1709: \bibitem[{{Smith} {et~al.}(2005){Smith}, {Treu}, {Ellis}, {Moran}, \&

1710:   {Dressler}}]{smith05}

1711: {Smith}, G.~P., {Treu}, T., {Ellis}, R.~S., {Moran}, S.~M., \& {Dressler}, A.

1712:   2005, \apj, 620, 78

1713: 

1714: \bibitem[{{Strateva} {et~al.}(2001){Strateva}, {Ivezi{\'c}}, {Knapp},

1715:   {Narayanan}, {Strauss}, {Gunn}, {Lupton}, {Schlegel}, {Bahcall}, {Brinkmann},

1716:   {Brunner}, {Budav{\'a}ri}, {Csabai}, {Castander}, {Doi}, {Fukugita}, {Gy{\H

1717:   o}ry}, {Hamabe}, {Hennessy}, {Ichikawa}, {Kunszt}, {Lamb}, {McKay},

1718:   {Okamura}, {Racusin}, {Sekiguchi}, {Schneider}, {Shimasaku}, \&

1719:   {York}}]{strateva01}

1720: {Strateva}, I. et al. 2001, \aj, 122, 1861

1721: 

1722: \bibitem[{{Strauss} {et~al.}(2002){Strauss}, {Weinberg}, {Lupton}, {Narayanan},

1723:   {Annis}, {Bernardi}, {Blanton}, {Burles}, {Connolly}, {Dalcanton}, {Doi},

1724:   {Eisenstein}, {Frieman}, {Fukugita}, {Gunn}, {Ivezi{\'c}}, {Kent}, {Kim},

1725:   {Knapp}, {Kron}, {Munn}, {Newberg}, {Nichol}, {Okamura}, {Quinn}, {Richmond},

1726:   {Schlegel}, {Shimasaku}, {SubbaRao}, {Szalay}, {Vanden Berk}, {Vogeley},

1727:   {Yanny}, {Yasuda}, {York}, \& {Zehavi}}]{strauss02}

1728: {Strauss}, M.~A. et al. 2002, \aj, 124, 1810

1729: 

1730: \bibitem[{{Tanaka} {et~al.}(2005){Tanaka}, {Kodama}, {Arimoto}, {Okamura},

1731:   {Umetsu}, {Shimasaku}, {Tanaka}, \& {Yamada}}]{tanaka05}

1732: {Tanaka}, M. et al. 2005, \mnras, 362, 268

1733: 

1734: \bibitem[{{Treu} {et~al.}(2005){Treu}, {Ellis}, {Liao}, {van Dokkum}, {Tozzi},

1735:   {Coil}, {Newman}, {Cooper}, \& {Davis}}]{treu05b}

1736: {Treu}, T. etal. 2005, \apj, 633,

1737:   174

1738: 

1739: \bibitem[{{Treu} {et~al.}(2003){Treu}, {Ellis}, {Kneib}, {Dressler}, {Smail}, 

1740:  {Czoske}, {Oemler}, {Natarajan}}]{treu03}

1741: {Treu}, T. et al. 2003, \apj, 591, 53 

1742: 

1743: \bibitem[{{van der Wel} {et~al.}(2005){van der Wel}, {Franx}, {van Dokkum},

1744:   {Rix}, {Illingworth}, \& {Rosati}}]{vanderwel05}

1745: {van der Wel}, A., {Franx}, M., {van Dokkum}, P.~G., {Rix}, H.-W.,

1746:   {Illingworth}, G.~D., \& {Rosati}, P. 2005, \apj, 631, 145

1747: 

1748: \bibitem[{{van der Wel} {et~al.}(2006){van der Wel}, {Franx}, {Wuyts}, {van

1749:   Dokkum}, {Huang}, {Rix}, \& {Illingworth}}]{vanderwel06b}

1750: {van der Wel}, A., {Franx}, M., {Wuyts}, S., {van Dokkum}, P.~G., {Huang}, J.,

1751:   {Rix}, H.-W., \& {Illingworth}, G. 2006, \apj, 652, 97

1752: 

1753: \bibitem[{{van der Wel et al.}(2007)}]{vanderwel07}

1754: {van der Wel et al.} 2007, astro-ph/0705.3394

1755: 

1756: \bibitem[{{van Dokkum} \& {van der Marel}(2007)}]{vandokkum07}

1757: {van Dokkum}, P.~G., \& {van der Marel}, R.~P. 2007, \apj, 655, 30

1758: 

1759: \bibitem[{{Vanzella} {et~al.}(2006){Vanzella}, {Cristiani}, {Dickinson},

1760:   {Kuntschner}, {Nonino}, {Rettura}, {Rosati}, {Vernet}, {Cesarsky},

1761:   {Ferguson}, {Fosbury}, {Giavalisco}, {Grazian}, {Haase}, {Moustakas},

1762:   {Popesso}, {Renzini}, {Stern}, \& {The Goods Team}}]{vanzella06}

1763: {Vanzella}, E. et al. 2006, \aap, 454, 423

1764: 

1765: \bibitem[{{von der Linden} {et~al.}(2007){von der Linden}, {Best}, {Kauffmann},

1766:   \& {White}}]{vonderlinden07}

1767: {von der Linden}, A., {Best}, P.~N., {Kauffmann}, G., \& {White}, S.~D.~M.

1768:   \mnras, 379, 867

1769: 

1770: \bibitem[{{Wolf} {et~al.}(2003){Wolf}, {Meisenheimer}, {Rix}, {Borch}, {Dye},

1771:   \& {Kleinheinrich}}]{wolf03}

1772: {Wolf}, C., {Meisenheimer}, K., {Rix}, H.-W., {Borch}, A., {Dye}, S., \&

1773:   {Kleinheinrich}, M. 2003, \aap, 401, 73

1774: 

1775: \bibitem[{{Wuyts} {et~al.}(2007)}]{wuyts07}

1776: {Wuyts}, S. et al. 2007, \apj, 655, 51

1777: 

1778: \bibitem[{{York} {et~al.}(2000){York}, {Adelman}, {Anderson}, {Anderson},

1779:   {Annis}, {Bahcall}, {Bakken}, {Barkhouser}, {Bastian}, {Berman}, {Boroski},

1780:   {Bracker}, {Briegel}, {Briggs}, {Brinkmann}, {Brunner}, {Burles}, {Carey},

1781:   {Carr}, {Castander}, {Chen}, {Colestock}, {Connolly}, {Crocker}, {Csabai},

1782:   {Czarapata}, {Davis}, {Doi}, {Dombeck}, {Eisenstein}, {Ellman}, {Elms},

1783:   {Evans}, {Fan}, {Federwitz}, {Fiscelli}, {Friedman}, {Frieman}, {Fukugita},

1784:   {Gillespie}, {Gunn}, {Gurbani}, {de Haas}, {Haldeman}, {Harris}, {Hayes},

1785:   {Heckman}, {Hennessy}, {Hindsley}, {Holm}, {Holmgren}, {Huang}, {Hull},

1786:   {Husby}, {Ichikawa}, {Ichikawa}, {Ivezi{\'c}}, {Kent}, {Kim}, {Kinney},

1787:   {Klaene}, {Kleinman}, {Kleinman}, {Knapp}, {Korienek}, {Kron}, {Kunszt},

1788:   {Lamb}, {Lee}, {Leger}, {Limmongkol}, {Lindenmeyer}, {Long}, {Loomis},

1789:   {Loveday}, {Lucinio}, {Lupton}, {MacKinnon}, {Mannery}, {Mantsch}, {Margon},

1790:   {McGehee}, {McKay}, {Meiksin}, {Merelli}, {Monet}, {Munn}, {Narayanan},

1791:   {Nash}, {Neilsen}, {Neswold}, {Newberg}, {Nichol}, {Nicinski}, {Nonino},

1792:   {Okada}, {Okamura}, {Ostriker}, {Owen}, {Pauls}, {Peoples}, {Peterson},

1793:   {Petravick}, {Pier}, {Pope}, {Pordes}, {Prosapio}, {Rechenmacher}, {Quinn},

1794:   {Richards}, {Richmond}, {Rivetta}, {Rockosi}, {Ruthmansdorfer}, {Sandford},

1795:   {Schlegel}, {Schneider}, {Sekiguchi}, {Sergey}, {Shimasaku}, {Siegmund},

1796:   {Smee}, {Smith}, {Snedden}, {Stone}, {Stoughton}, {Strauss}, {Stubbs},

1797:   {SubbaRao}, {Szalay}, {Szapudi}, {Szokoly}, {Thakar}, {Tremonti}, {Tucker},

1798:   {Uomoto}, {Vanden Berk}, {Vogeley}, {Waddell}, {Wang}, {Watanabe},

1799:   {Weinberg}, {Yanny}, \& {Yasuda}}]{york00}

1800: {York}, D.~G. et al. 2000, \aj, 120, 1579

1801: 

1802: \bibitem[{{Zheng} {et~al.}(2006){Zheng}, {Bell}, {Rix}, {Papovich}, {Le

1803:   Floc'h}, {Rieke}, \& {P{\'e}rez-Gonz{\'a}lez}}]{zheng06}

1804: {Zheng}, X.~Z., {Bell}, E.~F., {Rix}, H.-W., {Papovich}, C., {Le Floc'h}, E.,

1805:   {Rieke}, G.~H., \& {P{\'e}rez-Gonz{\'a}lez}, P.~G. 2006, \apj, 640, 784

1806: 

1807: \end{thebibliography}

1808: 

1809: \end{document}

1810: