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

2: \documentclass[preprint2]{aastex}

3: \newcommand{\vdag}{(v)^\dagger}

4: \newcommand{\fuvcenter}{1530\AA}

5: \newcommand{\nuvcenter}{2310\AA}

6: \newcommand{\fuvband}{1350-1750\AA}

7: \newcommand{\nuvband}{1750-2750\AA}

8: \newcommand{\fuvwidth}{400\AA}

9: \newcommand{\nuvwidth}{1000\AA}

10: \newcommand{\FUV}{{\it FUV~}}

11: \newcommand{\NUV}{{\it NUV~}}

12: \newcommand{\FUVnb}{{\it FUV}}

13: \newcommand{\NUVnb}{{\it NUV}}

14: 

15: \newcommand{\galex}{\textit{GALEX~}}

16: 

17: 

18: \shortauthors{}

19: \usepackage{natbib}

20: \bibpunct{(}{)}{;}{a}{}{,}

21: 

22: 

23: \begin{document}

24: \title{Clustering Properties of restframe UV selected galaxies II:\\

25: Migration of star formation sites with cosmic time from GALEX and

26: CFHTLS}

27: 

28: 

29: \author{ 

30:   S\'ebastien Heinis\altaffilmark{1,2},

31:   Bruno Milliard\altaffilmark{1}, 

32:   St\'ephane Arnouts\altaffilmark{1},

33:   J\'er\'emy Blaizot\altaffilmark{1,3},\\ 

34:   David Schiminovich\altaffilmark{4}, 

35:   Tam\'as Budav\'ari\altaffilmark{2},

36:   Olivier Ilbert\altaffilmark{5},

37:   Jos\'e Donas\altaffilmark{1}, 

38:   Marie Treyer\altaffilmark{1},\\ 

39:   Ted K. Wyder\altaffilmark{6},

40:   Henry J. McCracken\altaffilmark{7},

41:   Tom A. Barlow\altaffilmark{6}, 

42:   Karl Forster\altaffilmark{6},

43:   Peter G. Friedman\altaffilmark{6},\\

44:   D. Christopher Martin\altaffilmark{6},

45:   Patrick Morrissey\altaffilmark{6},

46:   Susan G. Neff\altaffilmark{8},

47:   Mark Seibert\altaffilmark{6},

48:   Todd Small\altaffilmark{6},

49:   Luciana Bianchi\altaffilmark{9},

50:   Timothy M. Heckman\altaffilmark{2},

51:   Young-Wook Lee\altaffilmark{10},

52:   Barry F. Madore\altaffilmark{11},\\

53:   R. Michael Rich\altaffilmark{12},

54:   Alex S. Szalay\altaffilmark{2},

55:   Barry Y. Welsh\altaffilmark{13},

56:   Sukyoung K. Yi\altaffilmark{10} and

57:   C. K. Xu \altaffilmark{6}

58:  }

59: \altaffiltext{1}{Laboratoire d'Astrophysique de Marseille, BP 8, Traverse

60: du Siphon, 13376 Marseille Cedex 12, France}

61: 

62: \altaffiltext{2}{Department of Physics and Astronomy, The Johns Hopkins

63: University, Homewood Campus, Baltimore, MD 21218}

64: 

65: \altaffiltext{3}{Max Planck Institut f\"ur astrophysik, D-85748

66: Garching, Germany}

67: 

68: \altaffiltext{4}{Department of Astronomy, Columbia University, New

69: York, NY 10027}

70: 

71: \altaffiltext{5}{Institute for Astronomy, 2680 Woodlawn Drive,

72: Honolulu, HI 96822}

73: 

74: \altaffiltext{6}{California Institute of Technology, MC 405-47, 1200 East

75: California Boulevard, Pasadena, CA 91125}

76: 

77: \altaffiltext{7}{Institut d'Astrophysique de Paris, Universit\'e

78: Pierre et Marie Curie, UMR 7095, 98 bis Bvd Arago, 75014 Paris,

79: France}

80: 

81: \altaffiltext{8}{Laboratory for Astronomy and Solar Physics, NASA Goddard

82: Space Flight Center, Greenbelt, MD 20771}

83: 

84: \altaffiltext{9}{Center for Astrophysical Sciences, The Johns Hopkins

85: University, 3400 N. Charles St., Baltimore, MD 21218}

86: 

87: \altaffiltext{10}{Center for Space Astrophysics, Yonsei University, Seoul

88: 120-749, Korea}

89: 

90: \altaffiltext{11}{Observatories of the Carnegie Institution of Washington,

91: 813 Santa Barbara St., Pasadena, CA 91101}

92: 

93: \altaffiltext{12}{Department of Physics and Astronomy, University of

94: California, Los Angeles, CA 90095}

95: 

96: \altaffiltext{13}{Space Sciences Laboratory, University of California at

97: Berkeley, 601 Campbell Hall, Berkeley, CA 94720}

98: 

99: 

100: 

101: 

102: 

103: 

104: \begin{abstract}

105:   {We analyze the clustering properties of ultraviolet selected

106:     galaxies by using \galex-~SDSS data at $z<0.6$ and CFHTLS deep

107:     $u'$ imaging at $z\sim 1$. These datasets provide a unique basis

108:     at $z\le 1$ which can be directly compared with high redshift

109:     samples built with similar selection criteria. We discuss the

110:     dependence of the correlation function parameters ($r_0$,

111:     $\delta$) on the ultraviolet luminosity as well as the linear bias

112:     evolution. We find that the bias parameter shows a gradual decline

113:     from high ($b_8 \gtrsim 2$) to low redshift ($b_8 \simeq

114:     0.79^{+0.1}_{-0.08}$). When accounting for the fraction of the

115:     star formation activity enclosed in the different samples, our

116:     results suggest that the bulk of star formation migrated from high

117:     mass dark matter halos at $z>2$ ($10^{12} \le M_{min} \le 10^{13}

118:     M_{\odot}$, located in high density regions), to less massive

119:     halos at low redshift ($M_{min} \le 10^{12} M_{\odot}$, located in

120:     low density regions). This result extends the ``downsizing''

121:     picture (shift of the star formation activity from high stellar

122:     mass systems at high $z$ to low stellar mass at low $z$) to the

123:     dark matter distribution.}\vspace{1.cm}

124: \end{abstract}

125: \keywords{Galaxies: UV - Correlation Function Evolution - Star

126: Formation - Downsizing}

127: \shorttitle{Migration of star formation sites with cosmic time from \galex and

128: CFHTLS}

129: \slugcomment{Submitted for publication in the Special GALEX Ap.J.Suppl. Issue}

130: \maketitle

131: 

132: \section{Introduction}

133: 

134: 

135: Accumulated evidence shows that the cosmic Star Formation Rate (SFR)

136: has been decreasing from $z \sim 1$ by a dramatic factor of about $5$

137: \citep{Hopkins_2004, Lilly_1996, Madau_1996, Schiminovich_2005,

138: Sullivan_2000, Wilson_2002}. This is linked to the decrease of the

139: contribution of the faint galaxies that dominate the star formation

140: density, and to the strong decline of the most ultraviolet-luminous

141: galaxies with time, given the redshift evolution of the 1500 \AA~

142: luminosity function \citep{Arnouts_2005}. Another aspect of this

143: evolution, known as ``downsizing'' \citep{Cowie_1996}, is the

144: observation that star formation activity shifts with time from high to

145: low stellar mass systems \citep[][and references therein]{Bundy_2005,

146: Jimenez_2005, Juneau_2005, Heavens_2004}.

147: 

148: The star formation history results from the interplay between the

149: physical processes driving the star formation fueling (gas cooling)

150: and regulation (feedback), both closely related to galaxy

151: environment. Recent simulations show that about half of the galaxy gas

152: is accreted through a cold mode, which dominates at high redshift in

153: high density environments, and shifts to low density environments in

154: the local Universe \citep{Keres_2005}. The type of the dominant

155: feedback process is expected to depend on galaxy host halo mass:

156: supernovae explosions \citep[e.g.][]{Benson_2003} at low mass, and AGN

157: \citep[e.g.][]{Croton_2006} at high mass. \citet{Cattaneo_2006} show

158: that the introduction of a critical halo mass above which there is a

159: complete shutdown of cooling and star formation is efficient to

160: reproduce the bimodality in galaxy properties observed in the local

161: Universe \citep[e.g.][]{Baldry_2004}.

162: 

163: 

164: In this paper, we propose to set constraints on the roles of these

165: different processes through cosmic time by assessing the spatial

166: distribution of star formation in the Universe from high to low

167: redshifts. A convenient method is to study the clustering properties

168: of restframe ultraviolet (UV) selected galaxies. This has already been

169: performed at high redshifts using Lyman Break Galaxies (LBGs) samples

170: to show that, at these epochs, star formation is highly clustered and

171: concentrated in overdense regions \citep{Adelberger_2005, Allen_2005,

172: Arnouts_2002, Foucaud_2003, Giavalisco_2001}. The study of the

173: redshift evolution of the clustering properties of actively star

174: forming galaxies has now been made possible in a homogeneous way with

175: the combination of restframe UV data collected from $z\sim4$ to $z =

176: 0$. To extend high-$z$ studies, we use GALEX observations in the

177: recent Universe and CFHTLS deep imaging at $z=1$. We compute the

178: angular correlation function (ACF) of star forming galaxies and derive

179: their bias and its evolution.

180: 

181: 

182: 

183: In a companion paper, \citet[][hereafter Paper I]{Milliard_2007}, we

184: describe in detail the methodology and the first results of the

185: angular correlation function measurements of UV selected galaxies

186: using a \galex sample at $z\le 0.6$. Section \ref{sec_sample}

187: summarizes the sample properties and presents a new restframe

188: UV-selected sample from the $u'$ band deep CFHTLS imaging survey that

189: we use to extend the analysis to higher redshift ($z\sim 1$). We then

190: investigate the dependence on redshift and UV luminosity of the

191: clustering properties: $r_0$, $\delta$ in sect. \ref{sec_results_Muv},

192: bias in sect. \ref{sec_bias}. In the last section we discuss the

193: evolution of the preferred sites of star formation over the last 90\%

194: of the age of the Universe.  

195: 

196: All magnitudes have been corrected for Galactic extinction using the

197: E(B-V) value from the \citet{Schlegel_1998} maps and the

198: \citet{Cardelli_1989} extinction law.  Throughout the paper, we adopt

199: the following cosmological parameters: $\Omega_{m}=0.3$,

200: $\Omega_{\Lambda}=0.7$, $H_0=70$ km s$^{-1}$Mpc$^{-1}$.

201: 

202: 

203: 

204: 

205: 

206: \section{Samples description}\label{sec_sample}

207: \subsection{\galex}\label{sec_galex_sample}

208: 

209: In this work, we use the same subsample of \galex Release 2 (GR2)

210: Medium Imaging Survey (MIS) fields cross-matched with SDSS DR5

211: presented in Paper I, and we refer to this paper for a full

212: description. We recall here the main characteristics of the

213: selection. We only keep \galex objects with SDSS counterparts within a

214: search radius of 4\arcsec and use the closest SDSS match. We select

215: galaxies as objects with SDSS \texttt{type} equal to 3. We use the

216: half of the MIS fields from the GR2 dataset with the lowest Galactic

217: extinction $\left( \langle E(B-V)\rangle \leq

218: 0.04\right)$. Photometric redshifts are computed using an empirical

219: method \citep{Connolly_1995} trained on SDSS spectroscopic

220: counterparts. The standard deviation estimated from the SDSS

221: spectroscopic redshifts is $\sigma = 0.03$. We then use a template

222: fitting procedure \citep{Arnouts_2070} to derive UV luminosities. Our

223: starting samples include objects with \FUV $<22$ or \NUV $<22$.

224: 

225: 

226: The \NUV absolute magnitude vs photometric redshift relation is shown

227: in figure \ref{fig_Mabs_zp}. The colors code the galaxy type

228: determined using a SED template fitting procedure: red represent

229: elliptical types, green spiral and blue irregular. We restricted

230: hereafter the samples to $-21.5<NUV_{\rm{abs}}<-14.$ and

231: $0.<z_{phot}<0.6$ (dashed lines on fig \ref{fig_Mabs_zp}). The same

232: cuts have been applied to the \FUV sample.

233: 

234: 

235: \begin{figure}[!t]

236:   \plotone{newf1.ps}

237:   \caption{ \small \NUV absolute magnitude-photometric redshift

238:   relation in the \galex sample. The colors code the type according to

239:   the best fitting template: red represent elliptical types, green

240:   spirals, and blue irregulars. The dashed lines indicate the additional

241:   cuts adopted: $z_{phot}<0.6$ and $-21.5<NUV_{\rm{abs}}<-14.$ The same cuts

242:   hold for \FUVnb.}  \label{fig_Mabs_zp}

243: \end{figure}

244: 

245: 

246: In the following, we consider both \FUV and \NUV bands and we divide

247: the samples in two bins according to the mean $UV$ absolute

248: magnitude. The figure \ref{fig_Muv_nzs} shows the photometric redshift

249: distributions of the \galex samples; the table \ref{tab_Muv_galex}

250: summarizes the properties of the samples.

251: 

252: \begin{figure}[!t]

253:   %\includegraphics[width=\hsize]{nz_mag_abs_cuts_ter.ps}

254:   \plotone{f2.eps}

255:   \caption{\small Redshift distribution of the subsamples cut in

256:   absolute UV magnitude: $M<\langle M \rangle$, solid lines;

257:   $M>\langle M \rangle$, dashed lines, \FUV is shown as blue and \NUV

258:   as red.}  

259:   \label{fig_Muv_nzs}

260: \end{figure}

261: 

262: 

263: \subsection{CFHTLS}\label{sec_cfhtls_sample}

264: 

265: 

266: 

267:  The CFHTLS-Deep survey consists of deep multi-colour images collected

268: through the $u' g'r' i' z'$ filters over four independent areas of 1

269: $deg^2$ each and reaching the limiting magnitude of $i'_{AB}\sim 26$.

270: In this work, we use the official CFHTLS data release T0003. For a

271: full presentation of the CFHTLS-Deep survey, we refer to

272: \citet{Schultheis_2006}\footnote{see also

273: \url{http://www.cfht.hawaii.edu/Science/CFHLS/} and

274: \url{http://www.ast.obs-mip.fr/article204.html}}. We built specific

275: masks from the $u$-band images to mask out stars, chips edges' and

276: artifacts. The total solid angle of the four fields used after masking

277: is 3.1 $deg^2$. The star/galaxy separation is based on the same method

278: as \citet{McCracken_2003} with the half-light radius versus $u$

279: magnitude plot. This selection has been applied down to $u =

280: 23$. Beyond this limit, we combine the photometric criterion with the

281: star/galaxy classification derived from the photometric redshift code

282: {\it Le Phare}, \citep{Arnouts_2070}.  \\

283:  

284: To construct the sample of UV selected galaxies at $z\sim 1$, we adopt

285: a $u'$ magnitude limit of $u'=24$, which ensures a genuine

286: $UV$-selected sample as the $u'$ effective wavelength (3587 \AA)

287: corresponds to 1848 \AA~at our mean redshift $\langle z \rangle =

288: 0.94$.  The fraction of objects lost (without any redshift selection)

289: with a $i'=26$ cut is on average 0.07\% over the four fields at

290: $u'=24$. The redshift selection of the sample is based first on a

291: color-color selection and then on the photometric redshift

292: selection. We do not adopt a single selection based on the photometric

293: redshifts because of the variable accuracy of the method due to

294: inhomogeneous exposure times in the five bands for the different

295: fields.\\

296: %

297: First, we use a color-color selection to isolate galaxies with $z\ge

298: 0.7$, based on VVDS photometric redshifts estimation, relying

299:   on multi-color data \citep{Ilbert_2006}. As shown in

300: figure~\ref{fig_color_sel}, the $(g-r)$ versus $(r-i)$ selection

301: criterion is efficient to separate galaxies at $z\ge 0.7$ (big dots)

302: from the lower redshift population (small dots). The line shows our

303: separation criterion. 96\% of galaxies with $z_{phot}\ge 0.7$ are

304: located below the line while less than 10\% of low~$z$ objects

305: ($z_{phot}\le 0.7$) fall in the same region.  \\

306: %

307: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

308: \begin{figure}[t]

309:   %\includegraphics[width=\hsize]{cfhtls_vvds_color_t0003.ps}

310:   \plotone{f3.eps}

311:   \caption{\small Redshift selection based on the ($g'-r'$) vs

312:     ($r'-i'$) diagram for the CFHTLS-D1 field.  The small dots show

313:     galaxies with $z_{phot}<0.7$ and big dots galaxies with

314:     $z_{phot}>0.7$.  The line represents the adopted color-color

315:     selection criterion.}

316: 

317:   \label{fig_color_sel} 

318: \end{figure}

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

320: 

321: The photometric redshifts are computed by using {\it Le Phare} code

322: and by adopting the method described by \citet{Ilbert_2006}. The

323: comparison with the spectroscopic redshifts, obtained by the VVDS in

324: the best photometric field (CFHTLS-D1, \citet{Lefevre_2005}), for our

325: $u'$ selected sample shows an accuracy of $\sigma(\Delta

326: z/(1+z))=0.03$ with 4\% of outliers (defined as $\Delta z\ge

327: 0.15\times (1+z)$).

328: 

329: In figure~\ref{fig_nz} we show the photometric redshift distribution

330: of the galaxies selected with the color criterion (dashed

331: histogram). The final sample is obtained by further selecting objects

332: with $0.7<z_{phot}<1.3$ (solid histogram). The absolute magnitudes in

333: the \galex bands are derived from the best fitting SEDs whose NUV-rest

334: flux are well constrained by the $u'$, $g'$, and $r'$ bands in the

335: redshifts range ($0.7\le z\le 1.3$). Note that as the $u'$ filter

336: shifts to \FUV wavelengths at $z\sim 1$, absolute magnitudes depend

337: very weakly on k-correction and best-fit fitting SEDs. As for \galex

338: samples, we divide the CFHTLS sample in two bins according to the mean

339: \FUV absolute magnitude and the resulting redshift distributions are

340: shown in Fig~\ref{fig_nz}.\\

341: %

342: The global properties of the CFHTLS UV samples are given in table

343: \ref{tab_Muv_cfhtls}.

344: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

345: \begin{figure}[t]

346:   %\includegraphics[width=\hsize]{nzs_color_zp_bis.ps}

347:   \plotone{f4.eps}

348:   \caption{\small Redshift distribution of the CFHTLS sample: the

349:     dashed histogram shows the photometric redshift distribution of

350:     the galaxies selected with the color criterion alone, while the

351:     solid histogram shows the final redshift distribution after

352:     selecting objects with $0.7<z_{phot}<1.3$. The dotted and

353:     dot-dashed histograms show the redshift distributions of the

354:     $FUV_{\rm{abs}} < -19.41$ and $FUV_{\rm{abs}} > -19.41$

355:     respectively.}

356:   \label{fig_nz} 

357: \end{figure}

358: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

359: 

360: 

361: 

362: \section{Redshift evolution of the correlation function of UV-selected galaxies}\label{sec_results_Muv}

363: We compute the ACF using the \citet{Landy_1993} estimator. We assume

364: that the ACF is well approximated by a power-law: $w(\theta) =

365: A_{w}\theta^{-\delta}$; we use a variable Integral Constraint (IC)

366: with $\delta$ as free parameter during the power-law fitting process,

367: and estimate the IC with the same method used by

368: \citet{Roche_1999}. We derive correlation lengths ($r_0$) for each

369: sample from the Limber equation \citep{Peebles_1980}, using the

370: corresponding redshift distribution. These quantities, as well as the

371: bias parameter\footnote{See Paper I for details on the computations.},

372: are summarized in table~\ref{tab_Muv_galex} and

373: table~\ref{tab_Muv_cfhtls}. The effects of the dust internal to

374: galaxies have again been neglected.

375: 

376: 

377: In figures~\ref{fig_Muv_wtheta_galex} and~\ref{fig_Muv_wtheta_cfhtls},

378: we show the ACFs of the \galex and CFHTLS samples respectively. The

379: ACFs are derived for the global samples and for two sub-samples with

380: UV absolute luminosity brighter and fainter than the mean

381: $<UV_{\rm{abs}}>$ of each sample. The angular scales probed for the

382: \galex samples are $0.005^{\circ}$ to $0.4^{\circ}$ (corresponding

383: respectively to comoving distances $0.07$ Mpc and $5.7$ Mpc at $z =

384: 0.2$), while $0.002^{\circ}$ to $0.4^{\circ}$ for the CFHTLS samples

385: (resp. $0.11$ Mpc and $23$ Mpc at $z =1$). These ACFs are fairly well

386: fitted by power-laws, even if a small dip appears at small scales in

387: the FUV \galex samples and also in the CFHTLS bright one (see sec.

388: \ref{Halo_terms}). The higher surface density of UV-selected galaxies

389: at $z\sim 1$ allows a less noisy estimation of the ACF at these epochs

390: than at $z<0.4$. 

391: 

392: \begin{figure*}[!t]

393:   \includegraphics[width=\hsize]{f5.eps}

394:   %\plotone{f5.eps}

395:   \caption{\small Angular correlation function of the \galex

396:   subsamples cut in absolute UV magnitude: $M<\langle M \rangle$,

397:   squares; $M>\langle M \rangle$, triangles and for comparison the

398:   total sample (circles). Left panel, \FUVnb; right panel, \NUVnb. The

399:   curves show the best fit power-law not corrected for the Integral

400:   Constraint bias. The upper axis shows the comoving scales

401:   corresponding to angular scales at $z = 0.18$ (left) or $z = 0.24$

402:   (right).}

403:   \label{fig_Muv_wtheta_galex}

404: \end{figure*}

405: 

406: \begin{figure}[!t]

407:   %\includegraphics[width=\hsize]{wtheta_more_bins_bis.ps}

408:   \plotone{f6.eps}

409:   \caption{\small Angular correlation function of the CFHTLS

410:   subsamples cut in absolute UV magnitude: $M<\langle M \rangle$,

411:   squares; $M>\langle M \rangle$, triangles and for comparison the

412:   total sample (circles). The curves show the best fit power-laws with

413:   the Integral Constraint correction terms subtracted. The upper axis

414:   shows the comoving scales corresponding to angular scales at $z =

415:   0.9$.}

416:   \label{fig_Muv_wtheta_cfhtls}

417: \end{figure}

418: 

419: %Table 1

420: \input{tab1}

421: 

422: %Table 2

423: \input{tab2}

424: 

425: 

426: 

427: \subsection{Clustering segregation with \FUV luminosity }

428: \label{sec_ro_Muv}

429: 

430: 

431: 

432: \begin{figure}[!t]

433:   

434:   %\includegraphics[width=\hsize]{ro_mabs_paper_cfhtls_t03.ps}

435:   \plotone{f7.eps}

436:   \caption{\small Dependence on absolute \FUV magnitude of the

437:   correlation length $r_0$ for low and high redshift restframe

438:   $UV$-selected galaxies: open squares, \citet{Adelberger_2005}; open

439:   star, \citet{Arnouts_2002}; open triangles, \citet{Foucaud_2003};

440:   open circles, \citet{Giavalisco_2001}. Our results are presented as

441:   filled circles (\FUVnb) and filled squares (\NUVnb) for the \galex

442:   samples and as filled stars for the CFHTLS samples. The mean

443:   redshifts of the samples are color-coded. Note that the results from

444:   a given study (including ours) are not all obtained from independent

445:   samples. Hence we distinguish global samples by plotting them with a

446:   bigger symbol size, in this figure and in the following ones as

447:   well, unless otherwise stated.}

448:   \label{fig_ro_Muv}

449: \end{figure}

450: 

451: 

452: The dependence of $r_0$ on \FUV luminosity in \galex and CFHTLS

453: samples is shown in fig.~\ref{fig_ro_Muv}, along with results from higher

454: redshift studies ($z\ge 2$)\footnote{We choose the $FUV$ absolute

455: magnitude for the comparison as most of high redshift samples are

456: $FUV$ restframe selected, and \galex results are not strongly

457: dependent of the UV band.\\ The mean absolute magnitudes of the LBG

458: samples have been obtained by deriving an average apparent magnitude

459: from the galaxy counts, and assuming a k-correction of

460: $2.5\log(1+z)$. \citet{Ouchi_2005} do not provide their counts, so we

461: computed the expected mean absolute magnitude given their limiting

462: absolute magnitude and the luminosity function of

463: \citet{Sawicki_2006}.}. As the different surveys probe different parts

464: of the UV luminosity function with little overlap, it is difficult to

465: draw firm conclusions. Nevertheless, significant differences between

466: the samples are apparent:

467: 

468: 

469: \begin{itemize}    

470: \item We use as reference the correlation function results from

471: \citet{Adelberger_2005}, \citet{Allen_2005}, \citet{Foucaud_2003},

472: \citet{Giavalisco_2001}, \citet{Lee_2006} and \citet{Ouchi_2001} at

473: $z>2$. At these redshifts, all studies conclude a significant

474: segregation of $r_0$ with UV luminosity (the more luminous the more

475: clustered) in the range $-23 \le FUV_{\rm{abs}}\le -20$.

476: %

477: \item At $z\sim 1$, a positive correlation of $r_0$ with \FUV is still

478: observed for $-20\le FUV_{\rm{abs}} \le -19$. Notably, our value of

479: $r_0$ at $FUV_{\rm{abs}} \sim -20$ is very close to that of

480: \citet{Adelberger_2005} obtained from $z \sim 2$ samples.

481: %

482: \item At $z\lesssim0.3$, we probe a fainter luminosity range ($-19 \le

483:  FUV_{\rm{abs}} \le -17$), and a weak anti-correlation of $r_0$ with \FUV is

484:  apparent, though given the error bars, it is compatible with no \FUV

485:  luminosity segregation of $r_0$.

486: \end{itemize}

487: %

488: 

489: The values of $r_0$ as a function of \FUV luminosity for \FUV selected

490: samples at different redshifts follow a unique smooth curve, with a

491: significant slope at the bright end $FUV_{\rm{abs}} \le -19$ and a

492: flat or slightly negative slope at the faint end. A similar

493: segregation is observed with $B$ luminosity at low redshift, with

494: optical selection criteria \citep{Benoist_1996, Guzzo_2000,

495: Norberg_2002, Willmer_1998, Zehavi_2005}. In particular,

496: \citet{Norberg_2001} showed that for blue-selected galaxies $r_0$

497: increases only slowly for galaxies fainter than $L^{B}_{*}$, while it

498: varies strongly for galaxies brighter than $L^{B}_{*}$. Indeed, using

499: $N$-body simulations \citet{Benson_2001} showed that $L^{B}_*$ could

500: be a natural boundary in the distribution of the halos hosting

501: galaxies, galaxies fainter than $L^{B}_*$ being hosted by a

502: mix of low and high mass halos, while galaxies brighter than $L^{B}_*$

503: hosted by more and more massive haloes. To check if $FUV_*$ could

504: play a similar role in UV samples, we show in

505: figure~\ref{fig_ro_Mstar} $r_0$ as a function of $\langle

506: FUV_{\rm{abs}} \rangle - FUV_{\ast}$, where the evolution of

507: $FUV_{\ast}$ with $z$ is taken from \citet{Arnouts_2005} (for $z<1$)

508: and \citet{Sawicki_2006} (for $z>1$). The luminosity dependence of

509: $r_0$ changes noticeably when expressed as a function of

510: $FUV_{\rm{abs}}-FUV_{\ast}$, as two different trends are observed

511: according to the redshift range:

512: 

513: 

514: \begin{itemize}

515:  \item At $z\ge 1$, for the high $z$ samples and our CFHTLS sample,

516:  the behavior of $r_0$ with $FUV_{\rm{abs}}-FUV_{\ast}$ is

517:  qualitatively compatible with the monotonic trend described

518:  above, the brighter galaxies being more clustered.

519:  \item At $z\le0.5$ (\galex samples) a radically different behavior of

520:  $r_0$ vs $FUV_{\rm{abs}}-FUV_{\ast}$ is seen. An anti-correlation or

521:  no correlation (given the errorbars) is observed, with brighter

522:  samples showing slightly lower $r_0$ than fainter ones.

523: \end{itemize}

524: This suggests that the luminosity segregation mechanisms of the

525: clustering at low redshifts work in a different regime, or that

526: $FUV_{\ast}$ is not the relevant variable.

527: 

528: 

529: \begin{figure}[!t]

530:   %\includegraphics[width=\hsize]{ro_mabs_mstar_cfhtls_t03.ps}

531:   \plotone{f8.eps}

532:   \caption{\small Same as in figure \ref{fig_ro_Muv} but versus $FUV_{\rm{abs}} -

533:   FUV_{\ast}$. }

534:   \label{fig_ro_Mstar}

535: \end{figure}

536: 

537: 

538: \subsection{ACF slope segregation with \FUV luminosity } 

539: \label{sec_delta_Muv}

540: 

541: \begin{figure}[!t]

542:   %\includegraphics[width=\hsize]{delta_mabs_cfhtls_t03.ps}

543:   \plotone{f9.eps}

544:   \caption{\small Dependence on absolute \FUV magnitude of the slope

545:   $\delta$ of the correlation function for low and high redshift

546:   restframe $UV$-selected galaxies. Results shown here come from

547:   studies allowing $\delta$ as a free parameter: open squares,

548:   \citet{Adelberger_2005}; open circles, \citet{Giavalisco_2001};

549:   crosses, \citet{Ouchi_2005} (results from the fit with Integral

550:   Constraint). Our results are presented as filled circles (\FUVnb)

551:   and filled squares (\NUVnb) for the \galex samples and as filled

552:   stars for the CFHTLS samples; the error on $\delta$ comes from the

553:   fitting procedure and the horizontal bars for our samples reflect

554:   the standard deviations of $FUV_{\rm{abs}}$. The mean redshifts of the

555:   samples are color-coded.}

556:   \label{fig_delta_Muv}

557: \end{figure}

558: 

559: \subsubsection{ACF slope}

560: The slope of the ACF ($\delta$), which describes the balance between

561: small and large scale separations, is an important indicator on the

562: nature of the spatial distribution of a given population. In paper I,

563: we found that the estimates of the slope inferred from the global

564: \galex samples ($\delta \simeq 0.81\pm 0.07$) are steeper than those

565: derived from optically selected blue galaxies in the local Universe:

566: $\delta\sim 0.6$ \citep{Budavari_2003, Madgwick_2002, Zehavi_2002,

567: Zehavi_2005}. \\

568: %

569: In figure~\ref{fig_delta_Muv}, we now analyse the dependence of

570: $\delta$ on UV luminosity for the different samples (\galex samples:

571: filled circles for \FUV and filled squares for \NUVnb; CFHTLS samples:

572: filled triangles; high $z$ samples: empty squares, open circles and

573: crosses; all points color-coded with redshift).\\

574: %

575: 

576: At $z > 3$, the compilation of measurements showed here, and

577: especially those at at $z = 4$, from \citet{Ouchi_2005} indicate that

578: the ACF slope steepens at higher UV luminosities.

579: 

580: \citet{Ouchi_2005} claimed that this trend, well modeled in the Halo

581: Occupation Distribution (HOD) framework, is not an actual slope

582: variation but it is related to the halo occupancy. Based on HOD

583: models, they show that the contribution of satellite galaxies ('one

584: halo term', see e.g. \citet{Zehavi_2004}) increases when selecting

585: brighter galaxies, by enhancing the small scale signal of the ACF ($ r

586: \lesssim 0.35 $ Mpc). This effect produces an apparent steepening of

587: the observed slope $\delta$.\\ Our GALEX sample at $z<0.4$ seems to

588: produce a similar although less pronounced effect. Our current low $z$

589: GALEX data do not allow to perform detailed comparison between

590: observations and HOD models, but we have investigated if the observed

591: steepening with luminosity can be partially due to the small scale

592: component. We fitted the GALEX ACF only at scales $r>0.4$ Mpc (see

593: sec. \ref{Halo_terms}) or $r>0.7$ Mpc in order not to include the one

594: halo term component. We do not observe any significant change with

595: respect to our initial slopes. However doing so we face at large

596: scales the problems of lower signal-to-noise ratio and more important

597: contribution of the Integral Constraint bias that prevent us to make

598: firm statements. This test thus relies on the efficiency of our power

599: law fitting process in recovering the true ACF (see Paper I).\\In

600: other words, at low redshift, we do not see evidence that the one halo

601: term plays a major role in the slope of the ACF, as observed at high

602: redshift, which is expected from simulations

603: \citep{Kravtsov_2004}. Hence this indicates that our clustering

604: parameters ($r_0$, $\delta$ and bias, $b_8$) reflect the large scale

605: clustering of star forming galaxies, which enables us to make

606: comparisons with analytical predictions for the clustering of Dark

607: Matter Haloes.

608: 

609: 

610: \subsubsection{Dip in the ACF ?}\label{Halo_terms}

611: 

612: The ACFs derived for the various \galex and CFHTLS samples are

613: globally well described by a power-law, but some of our ACFs show a

614: little dip around 0.35 Mpc, the \galex \FUV ones for instance, and

615: also the brightest CFHTLS sample at $z\sim1$ (at a slightly larger

616: scale $\sim 0.5$ Mpc). This recalls the departure to the power-law

617: observed in other surveys and interpreted as the transition between

618: the one and two halo terms in the HOD framework. \citet{Zehavi_2004}

619: showed that this transition occurs at $\sim 1.5-3$ Mpc for $r$-band

620: selected galaxies. It is expected that this scale should be shorter

621: for bluer galaxies, residing in less massive halos, as showed by

622: \citet{Magliocchetti_2003} in observations and \citet[][see their

623: fig. 22]{Berlind_2003} in simulations, with a transition scale for

624: late-types galaxies at $\sim0.45$ Mpc, close to what we

625: observe. Finally and very interestingly \citet{Ouchi_2005} observe

626: this transition for LBGs at $z\sim4$ at 0.35 Mpc, the same comoving

627: scale as we get.

628: 

629: Comparing measurements with predictions from HOD models is a natural

630: perspective of this work, to probe the redshift evolution of the halo

631: occupancy of star-forming galaxies. This will be addressed in details

632: in a forthcoming paper with enlarged datasets.

633: 

634: 

635: \section{Bias of star-forming galaxies from $z = 4$ to $z = 0 $}

636: \label{sec_bias}

637: 

638: The link between the properties of the galaxy distribution and the

639: underlying Dark Matter density field can be accessed via the bias

640: formalism. The bias parameter is indicative of the masses of the dark

641: matter halos that preferentially host the observed galaxy population

642: \citep[e.g.][]{Giavalisco_2001, Mo_2002, Ouchi_2004}, {\it i.e.} in

643: our case, actively star forming galaxies. The DM halo bias is a direct

644: output of \citet{Mo_2002} models. For galaxies, we assume a linear

645: bias to convert $r_0$ in $\sigma_{8,g}$, a common, though

646: questionable, assumption \citep[see e.g.][]{Marinoni_2005}.

647: 

648: \begin{figure}[htbp]

649:   %\includegraphics[width=\hsize]{bias_z_sigma_cfhtls_t03.ps}

650:   \plotone{f10.eps}

651:   \caption{\small Evolution with redshift of the bias of restframe

652:    $UV$-selected galaxies (symbols as figure \ref{fig_ro_Muv}). The

653:    curves show the effective bias of halos more massive than $M_{min}$

654:    (color-coded) according to \citet{Mo_2002} (see also Paper I). The

655:    inset is an enlargement of the low redshift area.}

656:   \label{fig_bias_z}

657: \end{figure}

658: 

659: \subsection{Redshift evolution of the bias}

660: Figure~\ref{fig_bias_z} shows as symbols the redshift evolution of the

661: bias parameter measured at 8 $h^{-1}$ Mpc defined as $b_8 =

662: \sigma_{8,g} / \sigma_{8,m}$ \citep[see e.g.][]{Magliocchetti_2000}

663: for the different samples discussed above.\footnote{For

664: \citet{Adelberger_2005} subsamples, $\delta$ values are not available;

665: we assumed that the slopes are the same than those of their global

666: samples. The expected relative error on the inferred bias is lower

667: than 10\% if $0<\delta<1$.} The bias values for our \galex and CFHTLS

668: samples are reported in tables \ref{tab_Muv_galex} and

669: \ref{tab_Muv_cfhtls} respectively. \\

670: %

671: The observed bias of star-forming galaxies shows a gradual increase

672: with look back time:

673: %

674: at $z>2$, UV galaxies are strongly biased \citep{Giavalisco_2001,

675: Foucaud_2003}, with $b_8\gtrsim 2$, and at a given redshift the bias

676: increases with \FUV luminosity (\FUV luminosity segregation).

677: %

678: At $z\sim 1$, the mean bias is $\langle b_8 \rangle =1.26 \pm0 .06$,

679: indicating that star-forming galaxies are closer tracers of the

680: underlying mass distribution at that time.

681: %

682: At $z\le 0.4$, given the error bars, the mean bias is consistent with

683: 0.8 for all \galex samples ($\langle b_8 \rangle =

684: 0.79^{+0.1}_{-0.08}$), a slight anti-bias independent of the UV

685: luminosity.\\

686: %

687: 

688: 

689: In figure~\ref{fig_bias_z}, we also show the effective bias evolution

690: derived from the \citet{Mo_2002} formalism for different minimum Dark

691: Matter Halo (DMH) mass thresholds. A comparison can be made to the

692: bias of star forming galaxies, if one assumes that most haloes do not

693: host more than one star-forming galaxy. This coarse assumption is

694: likely inaccurate for star-forming galaxies selected at high redshifts

695: in \FUV with a well developed one-halo term \citep{Kashikawa_2006,

696: Lee_2006}, but is acceptable at low redshifts in the \FUV since the

697: one-halo term does not seem to play a major role as discussed

698: sect. \ref{sec_delta_Muv}. \\

699: 

700: 

701: 

702: The mimimum masses of the DMH that produce the bias derived for

703: galaxies are $10^{12} M_{\odot} \lesssim M \lesssim 10^{13} M_{\odot}$

704: at $z\ge 2$, $10^{11} M_{\odot} \lesssim M \lesssim 10^{12} M_{\odot}$

705: at $z\simeq 1$ and $M \le 10^{12} M_{\odot}$ at $z<0.4$. There is an

706: obvious degeneracy of the models at low redshifts, but the locally

707: observed bias is definitely in the region of low cutoff masses. This

708: is a hint that observed star-forming galaxies at low redshift reside

709: preferentially in less massive halos than high $z$ star-forming

710: galaxies.

711: 

712: 

713: 

714: 

715: \begin{figure}[htbp]

716:   %\includegraphics[width=\hsize]{bias_z_sigma_cfhtls_t03.ps}

717:   \plotone{f11.eps}

718:   \caption{\small Bias of restframe $UV$-selected galaxies as a

719:    function of galaxy number density. The curves show the expected

720:    relation for the effective bias of Dark Matter halos at redshifts

721:    $z=0,0.5,1,2,3$, from bottom to top. We plot only independent

722:    samples here. The legend is the same as in fig.~\ref{fig_bias_z}.}

723:   \label{fig_bias_ngal}

724: \end{figure}

725: 

726: \subsection{Bias and galaxy number density}

727: In fig.~\ref{fig_bias_ngal} we show the bias a function of the galaxy

728: number density $n_{gal}$ for UV-selected samples and the predicted

729: relation between the effective bias and the number density of DMHs at

730: $z=0,0.5,1,2,3$ (curves from bottom to top). At high redshift ($z>1$),

731: we observe the well known luminosity segregation effect, brighter

732: galaxies (less abundant) having a larger bias, in good agreement with

733: DMH models predictions (the less abundant, the more clustered). In

734: contrast at low $z$ ($z<1$), a significant departure to this relation

735: is observed.  At $z\sim 0.9$, the CFHTLS data show a bias slightly

736: lower than the expected one according to the observed density with

737: $n_{gal}$ approximately 3 times lower than expected. This seems even

738: worse for our brightest samples at $z\sim 0.3$, as these galaxies are

739: about 10 times less numerous than expected according to their bias

740: values. In the model discussed here, we implicitly assume that one DMH

741: hosts one galaxy, which provides a fairly reasonable description of

742: the observations at high $z$, to the level of precision allowed here

743: \citep[see e.g.][for more detailed discussions on this

744: point]{Adelberger_2005, Ouchi_2004}. At $z<1$, this assumption may be

745: not valid anymore and our results suggest that star forming galaxies

746: (especially the brightest) are not hosted by a significant fraction of

747: the DMHs with similar clustering properties. This implies that the DMH

748: occupation fraction, that is roughly $>0.5$ at high redshift ($z>2$),

749: drops to 0.3 and 0.1 at $z=1$ and $z=0.3$ respectively.

750: 

751: 

752: % 

753: 

754: %Here again, 

755: \subsection{Bias and FUV LD fraction}

756: The very limited overlap in \FUV luminosities of the data at different

757: redshifts does not allow a derivation of the bias evolution with

758: redshift at fixed \FUV luminosity. However, despite the fact that low

759: $z$ samples reach fainter luminosities than high $z$ ones, they happen

760: to probe the same fraction of total \FUV luminosity densities, owing

761: to the strong evolution of the \FUV luminosity function with $z$

762: \citep{Arnouts_2005}. In particular at all redshifts the samples are

763: able to probe the bulk of star formation, {\it i.e.}  they encompass a

764: fraction of the \FUV luminosity density (LD) greater than 0.5. This

765: can be seen in fig.~\ref{fig_bias_LD} where we show the bias as a

766: function of the fraction of the total \FUV LD enclosed by the

767: different samples. This favorable situation allows us to track the

768: evolution with redshift of the clustering at a fixed fraction of the

769: \FUV LD, an essentially constant fraction of the star formation rate.

770: 

771: 

772: 

773: 

774: %

775: \begin{figure}[htbp]

776:   %\includegraphics[width=\hsize]{bias_LD_all_cfhtls_t03.ps}

777:   \plotone{f12.eps}

778:   \caption{\small Bias of restframe $UV$-selected galaxies (symbols as

779:    in figure \ref{fig_bias_z}) as a function of the fraction of the

780:    FUV total luminosity density represented by each sample. The legend

781:    is the same as in fig.~\ref{fig_bias_z}.}

782:   \label{fig_bias_LD}

783: \end{figure}

784: %

785: 

786: %

787: The fraction of the LD for each sample is computed by comparing the

788: total LD at the relevant redshift (from the \FUV Luminosity function

789: parameters of \citet{Arnouts_2005} and \citet{Sawicki_2006}), to the

790: LD enclosed by each sample according to its flux limits converted to

791: \FUV luminosity cuts\footnote{Note that while LBG samples are by

792: construction volume-limited, we cannot adopt this approach for the

793: \galex samples due to limited statistics. This means that faintest

794: galaxies are underrepresented, especially in our higher redshift

795: \galex samples. However, as we do not observe a strong luminosity

796: dependence of the bias within the \galex samples, we expect this has

797: only a small impact.}.\\ Note that we do not attempt to correct for

798: galaxy internal dust attenuation. As brighter UV galaxies present

799: higher extinction in the local Universe \citep{Buat_2005}, the strong

800: brightening of $FUV_{\ast}$ with redshift \citep[e.g.][]{Arnouts_2005}

801: may introduce a small bias as a given LD fraction could not correspond

802: exactly to the same star formation rate fraction at the different

803: redshifts we explore here.

804: 

805: 

806: %

807: The plot confirms the result already apparent in fig.~\ref{fig_bias_z}

808: of a significant decrease of the bias of UV selected galaxies from

809: high to low redshifts, but now selected on the basis of a physically

810: defined parameter, the fraction of the \FUV luminosity density. Near

811: an LD fraction of $0.5$, the bias is divided by a factor 3, between

812: redshifts near 3, shifting from $2.5$ in the redshift range $2-3$ down

813: to $0.8$ in the local universe.

814: 

815: 

816: 

817: \section{Discussion}\label{sec_discussion}

818: 

819: In paper I we reported on the overall clustering properties of the

820: UV-selected galaxies in \galex samples, the largest ones available to

821: date at low redshift and at these wavelengths. These samples allow for

822: the first time an investigation of the clustering properties of

823: UV-selected galaxies as a function of different parameters at

824: $z\lesssim 1$, which can be compared to higher redshift samples also

825: selected in the rest frame FUV.

826: 

827: The measurements from the \galex samples confirm previous results for

828: rest-UV selected galaxies at low redshifts indicating that they are

829: weakly clustered \citep{Heinis_2004}, with an autocorrelation function

830: well approximated by a power law in the range $0.2-5$ Mpc.  

831: 

832: 

833: 

834: At $z \sim 1$, the correlation length of the rest-UV selected galaxies

835: from CFHTLS $u'$ data is found comparable to those of the

836: emission-line samples from \citet{Coil_2004} in the same redshift

837: range, but slightly higher than those obtained by \citet{Meneux_2006}

838: for late type and irregular galaxies (their types 3 and 4, selected in

839: the visible). At those redshifts, according to our CFHTLS sample, star

840: forming galaxies are modestly biased with $\langle b_8 \rangle =1.26

841: \pm 0.06$, which under the linear bias hypothesis implies they are

842: closer tracers of the mass distribution than their higher redshift

843: counterparts. As opposed to the dependence found at redshifts above

844: $2$ \citep{Giavalisco_2001, Foucaud_2003}, no strong positive

845: correlation between the bias and the \FUV luminosity is observed in

846: the local universe, but rather a slight anti-correlation or no

847: correlation. At $z\le 0.4$, given the error bars, the mean bias is

848: consistent with 0.8 for all \galex samples ($\langle b_8 \rangle =

849: 0.79^{+0.10}_{-0.08}$) independently of the UV luminosity.

850: 

851: 

852: 

853: \subsection{\it Migration of the bulk of star formation sites from $z = 3$ to the local universe}

854: 

855:  

856: 

857: In this study, we find a decrease by a factor 3.1 of the bias with

858: respect to mass, from redshifts near $3$ to the local universe in the

859: UV flux-limited samples, and more importantly in samples selected in

860: UV luminosity so that they encompass a constant fraction of the

861: luminosity density at all $z$. This decrease is slightly larger than

862: the factor 2.7 derived from the \citet{Mo_2002} model for the $M \ge

863: 10^{12} M_{\odot}$ haloes that host most star formation at redshift

864: $3$, an indication that star forming galaxies tend to be hosted by

865: haloes of lower mass in the local universe. This is the main

866: conclusion of the present study.

867: 

868: 

869: 

870: The ``downsizing'' scenario \citep{Cowie_1996, Juneau_2005,

871: Bundy_2005, Heavens_2004} states that the star formation shifts from

872: high stellar mass systems at high redshift to low ones at low

873: redshift. Our results extend this vision in the sense that the same

874: trend is observed for the mass of the dark matter halos that host

875: actively star forming galaxies. 

876: 

877: 

878: The DMH mass migration of the bulk of the star formation might be

879: associated with regions of different densities. At high redshifts, LBGs

880: studies show that active star formation traced by the UV light resides

881: preferentially in overdense regions \citep{Adelberger_1998,

882: Blaizot_2004, Giavalisco_2002, Steidel_1998, Tasker_2006}. At low

883: redshift, \citet{Abbas_2005} showed that the slope of the fitted power

884: law is steeper in underdense regions, and that the correlation length

885: is smaller. The observed steeper ACFs for the more UV-luminous

886: galaxies at low $z$ suggest that the most star-forming objects reside

887: preferentially in regions where the local galaxy density is lower than

888: for the fainter ones, a result in agreement with direct optical based

889: studies of star formation as a function of galaxy density in the local

890: universe \citep{Gomez_2003, Lewis_2002}.

891: 

892: 

893: \acknowledgments

894: We thank Christian Marinoni for stimulating discussions.

895: 

896: 

897: \galex (Galaxy Evolution Explorer) is a NASA Small Explorer, launched

898: in April 2003. We gratefully acknowledge NASA's support for

899: construction, operation, and science analysis for the \galex mission,

900: developed in cooperation with the Centre National d'Etudes Spatiales

901: of France and the Korean Ministry of Science and Technology.

902: 

903: This study is also based on observations obtained with

904: MegaPrime/MegaCam, a joint project of CFHT and CEA/DAPNIA, at the

905: Canada-France-Hawaii Telescope (CFHT) which is operated by the

906: National Research Council (NRC) of Canada, the Institut National des

907: Science de l'Univers of the Centre National de la Recherche

908: Scientifique (CNRS) of France, and the University of Hawaii. This work

909: is based in part on data products produced at TERAPIX and the Canadian

910: Astronomy Data Centre as part of the Canada-France-Hawaii Telescope

911: Legacy Survey, a collaborative project of NRC and CNRS.

912: 

913: 

914: 

915: 

916: \begin{thebibliography}{}

917: 

918: % #########

919: % ### A ###

920: % #########

921: 

922: \bibitem[Abbas \& Sheth(2005)]{Abbas_2005} Abbas, U., \& Sheth, 

923: R.~K.\ 2005, \mnras, 364, 1327 

924: 

925: \bibitem[Adelberger et al.(1998)]{Adelberger_1998} Adelberger, K.~L., 

926: Steidel, C.~C., Giavalisco, M., Dickinson, M., Pettini, M., \& Kellogg, M.\ 

927: 1998, \apj, 505, 18 

928: 

929: \bibitem[Adelberger et al.(2005)]{Adelberger_2005} Adelberger, K.~L.,

930: Steidel, C.~C., Pettini, M., Shapley, A.~E., Reddy, N.~A., \& Erb,

931: D.~K.\ 2005, \apj, 619, 697

932: 

933: \bibitem[Allen et al.(2005)]{Allen_2005} Allen, P.~D., Moustakas, 

934: L.~A., Dalton, G., MacDonald, E., Blake, C., Clewley, L., Heymans, C., \& 

935: Wegner, G.\ 2005, \mnras, 360, 1244

936: 

937: \bibitem[Arnouts et al.(2002)]{Arnouts_2002} Arnouts, S., et al.\ 

938: 2002, \mnras, 329, 355

939: 

940: \bibitem[Arnouts et al.(2005)]{Arnouts_2005} Arnouts, S., et al.\ 

941: 2005, \apjl, 619, L43

942: 

943: \bibitem[Arnouts \& Ilbert, (in preparation)]{Arnouts_2070} Arnouts, S., \&

944: Ilbert, 0.\ in preparation

945: 

946: % #########

947: % ### B ###

948: % #########

949: 

950: \bibitem[Baldry et al.(2004)]{Baldry_2004} Baldry, I.~K., 

951: Glazebrook, K., Brinkmann, J., Ivezi{\'c}, {\v Z}., Lupton, R.~H., Nichol, 

952: R.~C., \& Szalay, A.~S.\ 2004, \apj, 600, 681 

953: 

954: \bibitem[Benoist et al.(1996)]{Benoist_1996} Benoist, C., 

955: Maurogordato, S., da Costa, L.~N., Cappi, A., \& Schaeffer, R.\ 1996, \apj, 

956: 472, 452

957: 

958: \bibitem[Benson et al.(2001)]{Benson_2001} Benson, A.~J., Frenk, 

959: C.~S., Baugh, C.~M., Cole, S., \& Lacey, C.~G.\ 2001, \mnras, 327, 1041 

960: 

961: \bibitem[Benson et al.(2003)]{Benson_2003} Benson, A.~J., Bower, 

962: R.~G., Frenk, C.~S., Lacey, C.~G., Baugh, C.~M., \& Cole, S.\ 2003, \apj, 

963: 599, 38 

964: 

965: \bibitem[Berlind et al.(2003)]{Berlind_2003} Berlind, A.~A., et al.\ 

966: 2003, \apj, 593, 1 

967: 

968: \bibitem[Blaizot et al.(2004)]{Blaizot_2004} Blaizot, J., 

969: Guiderdoni, B., Devriendt, J.~E.~G., Bouchet, F.~R., Hatton, S.~J., \& 

970: Stoehr, F.\ 2004, \mnras, 352, 571 

971: 

972: 

973: \bibitem[Buat et al.(2005)]{Buat_2005} Buat, V., et al.\ 2005, 

974: \apjl, 619, L51 

975: 

976: \bibitem[Budav{\' a}ri et al.(2003)]{Budavari_2003} Budav{\' a}ri, 

977: T., et al.\ 2003, \apj, 595, 59

978: 

979: 

980: \bibitem[Bundy et al.(2005)]{Bundy_2005} Bundy, K., et al.\ 2005, 

981: ArXiv Astrophysics e-prints, arXiv:astro-ph/0512465

982: 

983: \bibitem[Burgarella et al.(2006)]{Burgarella_2006} Burgarella, D., et 

984: al.\ 2006, \aap, 450, 69 

985: 

986: 

987: % #########

988: % ### C ###

989: % #########

990: 

991: 

992: \bibitem[Cardelli et al.(1989)]{Cardelli_1989} Cardelli, J.~A.,

993: Clayton, G.~C., \& Mathis, J.~S.\ 1989, \apj, 345, 245

994: 

995: \bibitem[Cattaneo et al.(2006)]{Cattaneo_2006} Cattaneo, A., Dekel, 

996: A., Devriendt, J., Guiderdoni, B., \& Blaizot, J.\ 2006, \mnras, 370, 1651

997: 

998: \bibitem[Coil et al.(2004)]{Coil_2004} Coil, A.~L., et al.\ 2004,

999: \apj, 609, 525

1000: 

1001: \bibitem[Connolly et al.(1995)]{Connolly_1995} Connolly, A.~J., 

1002: Csabai, I., Szalay, A.~S., Koo, D.~C., Kron, R.~G., \& Munn, J.~A.\ 1995, 

1003: \aj, 110, 2655 

1004: 

1005: \bibitem[Cowie et al.(1996)]{Cowie_1996} Cowie, L.~L.,

1006: Songaila, A., Hu, E.~M., \& Cohen, J.~G.\ 1996, \aj, 112, 839

1007: 

1008: \bibitem[Cross et al.(2001)]{Cross_2001} Cross, N., et al.\ 2001,

1009: \mnras, 324, 825

1010: 

1011: \bibitem[Croton et al.(2006)]{Croton_2006} Croton, D.~J., et al.\ 

1012: 2006, \mnras, 365, 11 

1013: 

1014: % #########

1015: % ### F ###

1016: % #########

1017: 

1018: \bibitem[Foucaud et al.(2003)]{Foucaud_2003} Foucaud, S., McCracken,

1019: H.~J., Le F{\`e}vre, O., Arnouts, S., Brodwin, M., Lilly, S.~J.,

1020: Crampton, D., \& Mellier, Y.\ 2003, \aap, 409, 835

1021: 

1022: % #########

1023: % ### G ###

1024: % #########

1025: 

1026: \bibitem[Giavalisco \& Dickinson(2001)]{Giavalisco_2001} Giavalisco, 

1027: M., \& Dickinson, M.\ 2001, \apj, 550, 177

1028: 

1029: \bibitem[Giavalisco(2002)]{Giavalisco_2002} Giavalisco, M.\ 2002, 

1030: \araa, 40, 579 

1031: 

1032: \bibitem[G{\'o}mez et al.(2003)]{Gomez_2003} G{\'o}mez, P.~L., et al.\

1033: 2003, \apj, 584, 210

1034: 

1035: 

1036: \bibitem[Guzzo et al.(2000)]{Guzzo_2000} Guzzo, L., et al.\ 2000, 

1037: \aap, 355, 1

1038: 

1039: % #########

1040: % ### H ###

1041: % #########

1042: 

1043: \bibitem[Heavens et al.(2004)]{Heavens_2004} Heavens, A., Panter, 

1044: B., Jimenez, R., \& Dunlop, J.\ 2004, \nat, 428, 625 

1045: 

1046: \bibitem[Heinis et al.(2004)]{Heinis_2004} Heinis, S., Treyer, M., 

1047: Arnouts, S., Milliard, B., Donas, J., Gal, R., Martin, D.~C., \& Viton, M.\ 

1048: 2004, \aap, 424, L9

1049: 

1050: \bibitem[Hopkins(2004)]{Hopkins_2004} Hopkins, A.~M.\ 2004, \apj, 

1051: 615, 209 

1052: 

1053: 

1054: % #########

1055: % ### I ###

1056: % #########

1057: \bibitem[Ilbert et al.(2006)]{Ilbert_2006} Ilbert, O., et al.\ 2006,

1058: \aap, 457, 841

1059: 

1060: % #########

1061: % ### J ###

1062: % #########

1063: \bibitem[Jimenez et al.(2005)]{Jimenez_2005} Jimenez, R., Panter, 

1064: B., Heavens, A.~F., \& Verde, L.\ 2005, \mnras, 356, 495

1065: 

1066: \bibitem[Juneau et al.(2005)]{Juneau_2005} Juneau, S., et al.\ 

1067: 2005, \apjl, 619, L135 

1068: 

1069: % #########

1070: % ### K ###

1071: % #########

1072: 

1073: \bibitem[Kashikawa et al.(2006)]{Kashikawa_2006} Kashikawa, N., et 

1074: al.\ 2006, \apj, 637, 631 

1075: 

1076: \bibitem[Kere{\v s} et al.(2005)]{Keres_2005} Kere{\v s}, D., 

1077: Katz, N., Weinberg, D.~H., \& Dav{\'e}, R.\ 2005, \mnras, 363, 2

1078: 

1079: \bibitem[Kravtsov et al.(2004)]{Kravtsov_2004} Kravtsov, A.~V., 

1080: Berlind, A.~A., Wechsler, R.~H., Klypin, A.~A., Gottl{\" o}ber, S., 

1081: Allgood, B., \& Primack, J.~R.\ 2004, \apj, 609, 35

1082: 

1083: % #########

1084: % ### L ###

1085: % #########

1086: 

1087: \bibitem[Landy \& Szalay(1993)]{Landy_1993} Landy, S.~D., \& 

1088: Szalay, A.~S.\ 1993, \apj, 412, 64 

1089: 

1090: \bibitem[Lee et al.(2006)]{Lee_2006} Lee, K.-S., Giavalisco, M.,

1091: Gnedin, O.~Y., Somerville, R.~S., Ferguson, H.~C., Dickinson, M., \&

1092: Ouchi, M.\ 2006, \apj, 642, 63

1093: 

1094: 

1095: \bibitem[Le F{\` e}vre et al.(2005)]{Lefevre_2005} Le F{\` e}vre, 

1096: O., et al.\ 2005, \aap, 439, 845

1097: 

1098: \bibitem[Lewis et al.(2002)]{Lewis_2002} Lewis, I., et al.\ 2002, 

1099: \mnras, 334, 673 

1100: 

1101: \bibitem[Lilly et al.(1996)]{Lilly_1996} Lilly, S.~J., Le Fevre, 

1102: O., Hammer, F., \& Crampton, D.\ 1996, \apjl, 460, L1 

1103: % #########

1104: % ### M ###

1105: % #########

1106: 

1107: \bibitem[Madau et al.(1996)]{Madau_1996} Madau, P., Ferguson, 

1108: H.~C., Dickinson, M.~E., Giavalisco, M., Steidel, C.~C., \& Fruchter, A.\ 

1109: 1996, \mnras, 283, 1388

1110: 

1111: \bibitem[Madgwick et al.(2002)]{Madgwick_2002} Madgwick, D.~S., et 

1112: al.\ 2002, \mnras, 333, 133

1113: 

1114: \bibitem[Magliocchetti et al.(2000)]{Magliocchetti_2000} Magliocchetti, 

1115: M., Bagla, J.~S., Maddox, S.~J., \& Lahav, O.\ 2000, \mnras, 314, 546

1116: 

1117: \bibitem[Magliocchetti \& Porciani(2003)]{Magliocchetti_2003} 

1118: Magliocchetti, M., \& Porciani, C.\ 2003, \mnras, 346, 186

1119: 

1120: \bibitem[Marinoni et al.(2005)]{Marinoni_2005} Marinoni, C., et al.\ 

1121: 2005, \aap, 442, 801 

1122: 

1123: \bibitem[McCracken et al.(2003)]{McCracken_2003} McCracken, H.~J., et 

1124: al.\ 2003, \aap, 410, 17

1125: 

1126:  \bibitem[Meneux et al.(2006)]{Meneux_2006} Meneux, B., et al.\ 

1127: 2006, \aap, 452, 387 

1128: 

1129: \bibitem[Milliard et al.(2007)]{Milliard_2007} Milliard, B., Heinis,

1130: S., Arnouts, S., Blaizot, J., Schiminovich, D., Budav{\' a}ri, T.,

1131: Donas, J., Treyer, M., Wyder, T.~K., \& GALEX Team \ 2006, \apj,

1132: submitted

1133: 

1134: \bibitem[Mo \& White(2002)]{Mo_2002} Mo, H.~J., \& White, 

1135: S.~D.~M.\ 2002, \mnras, 336, 112

1136: 

1137: % #########

1138: % ### N ###

1139: % #########

1140: 

1141: 

1142: 

1143: \bibitem[Norberg et al.(2001)]{Norberg_2001} Norberg, P., et al.\ 

1144: 2001, \mnras, 328, 64 

1145: 

1146: \bibitem[Norberg et al.(2002)]{Norberg_2002} Norberg, P., et al.\ 

1147: 2002, \mnras, 332, 827

1148: 

1149: % #########

1150: % ### 0 ###

1151: % #########

1152: 

1153: \bibitem[Ouchi et al.(2001)]{Ouchi_2001} Ouchi, M., et al.\ 2001, 

1154: \apjl, 558, L83

1155: 

1156: \bibitem[Ouchi et al.(2004)]{Ouchi_2004} Ouchi, M., et al.\ 2004, 

1157: \apj, 611, 685 

1158: 

1159: \bibitem[Ouchi et al.(2005)]{Ouchi_2005} Ouchi, M., et al.\ 2005, 

1160: \apjl, 635, L117 

1161: 

1162: 

1163: % #########

1164: % ### P ###

1165: % #########

1166: \bibitem[Peebles(1980)]{Peebles_1980} Peebles, P.~J.~E.\ 1980, 

1167: Princeton University Press

1168: 

1169: 

1170: % #########

1171: % ### R ###

1172: % #########

1173: \bibitem[Reddy et al.(2006)]{Reddy_2006} Reddy, N.~A., Steidel, 

1174: C.~C., Fadda, D., Yan, L., Pettini, M., Shapley, A.~E., Erb, D.~K., \& 

1175: Adelberger, K.~L.\ 2006, \apj, 644, 792 

1176: 

1177: \bibitem[Roberts et al.(2004)]{Roberts_2004} Roberts, S., et al.\ 

1178: 2004, \mnras, 352, 478 

1179: 

1180: 

1181: \bibitem[Roche \& Eales(1999)]{Roche_1999} Roche, N., \& Eales, 

1182: S.~A.\ 1999, \mnras, 307, 703 

1183: % #########

1184: % ### S ###

1185: % #########

1186: 

1187: 

1188: 

1189: \bibitem[Sawicki \& Thompson(2006)]{Sawicki_2006} Sawicki, M., \& 

1190: Thompson, D.\ 2006, \apj, 642, 653

1191: 

1192: \bibitem[Schiminovich et al.(2005)]{Schiminovich_2005} Schiminovich, D., 

1193: et al.\ 2005, \apjl, 619, L47 

1194: 

1195: \bibitem[Schlegel et al.(1998)]{Schlegel_1998} Schlegel, D.~J., 

1196: Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525 

1197: 

1198: \bibitem[Schultheis et al.(2006)]{Schultheis_2006} Schultheis, M., 

1199: Robin, A.~C., Reyl{\'e}, C., McCracken, H.~J., Bertin, E., Mellier, Y., \& 

1200: Le F{\`e}vre, O.\ 2006, \aap, 447, 185 

1201: 

1202: \bibitem[Steidel et al.(1998)]{Steidel_1998} Steidel, C.~C.,

1203: Adelberger, K.~L., Dickinson, M., Giavalisco, M., Pettini, M., \&

1204: Kellogg, M.\ 1998, \apj, 492, 428

1205: 

1206: 

1207: 

1208: \bibitem[Sullivan et al.(2000)]{Sullivan_2000} Sullivan, M., Treyer, 

1209: M.~A., Ellis, R.~S., Bridges, T.~J., Milliard, B., \& Donas, J.\ 2000, 

1210: \mnras, 312, 442

1211: 

1212: % #########

1213: % ### T ###

1214: % #########

1215: 

1216: \bibitem[Tasker \& Bryan(2006)]{Tasker_2006} Tasker, E.~J., \& 

1217: Bryan, G.~L.\ 2006, ArXiv Astrophysics e-prints, arXiv:astro-ph/0603684

1218: 

1219: 

1220: % #########

1221: % ### W ###

1222: % #########

1223: 

1224: 

1225: 

1226: \bibitem[Willmer et al.(1998)]{Willmer_1998} Willmer, C.~N.~A., da 

1227: Costa, L.~N., \& Pellegrini, P.~S.\ 1998, \aj, 115, 869

1228: 

1229: \bibitem[Wilson et al.(2002)]{Wilson_2002} Wilson, G., Cowie, L.~L.,

1230: Barger, A.~J., \& Burke, D.~J.\ 2002, \aj, 124, 1258

1231: 

1232: 

1233: % #########

1234: % ### Z ###

1235: % #########

1236: \bibitem[Zehavi et al.(2002)]{Zehavi_2002} Zehavi, I., et al.\ 

1237: 2002, \apj, 571, 172

1238: 

1239: \bibitem[Zehavi et al.(2004)]{Zehavi_2004} Zehavi, I., et al.\ 

1240: 2004, \apj, 608, 16

1241: 

1242: \bibitem[Zehavi et al.(2005)]{Zehavi_2005} Zehavi, I., et al.\ 

1243: 2005, \apj, 630, 1 

1244: 

1245: 

1246: \end{thebibliography}

1247: 

1248: \end{document} 

1249: 

1250: 

1251: