Mon.Not.R.Astron.Soc.000,1–5(2007) Printed2February2008 (MNLATEXstylefilev2.2) The luminosity – halo-mass relation for brightest cluster galaxies S. Brough1⋆, W. J. Couch1, C. A. Collins2, T. Jarrett3, D. J. Burke4, R. G. Mann5 1Centre for Astrophysics and Supercomputing, Swinburne University of Technology, Hawthorn, VIC 3122, Australia 8 2Astrophysics Research Institute, Liverpool John Moores University,Egerton Wharf, Birkenhead, CH41 1LD, UK 0 3Infrared Processing and Analysis Center, California Institute of Technology, 770 South Wilson Avenue, Pasadena, CA 91125, USA 0 4Harvard-Smithsonian Centerfor Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA 2 5SUPA, Institute for Astronomy, Universityof Edinburgh, Royal Observatory, Blackford Hill, Edinburgh, EH9 3NJ, UK n a J Accepted...Received...;inoriginalform2007 8 ] h ABSTRACT p We examine the central-galaxy luminosity – host-halo mass relation for 54 Brightest - Group Galaxies (BGGs) and 92 Brightest Cluster Galaxies (BCGs) at z < 0.1 and o present the first measurement of this relation for a sample of known BCGs at 0.1 < tr z < 0.8 (z¯∼ 0.3). At z < 0.1 we find LK ∝ M200.204±0.08 for the BCGs and the early- as type BGGs in groups with extended X-ray emission and LK ∝ M200.101±0.10 for the [ BCGs alone. At 0.1<z <0.8 we find LK ∝M200.208±0.11. We conclude that there is no evidence for evolution in this relationship between z <0.1 and z <0.8: BCG growth 1 appearstostillbelimitedbythetimescalefordynamicalfrictionattheseearliertimes, v 0 not proceeding according to the predictions of current semi-analytic models. 7 Key words: Galaxies: clusters: general – galaxies: elliptical and lenticular, cD – 1 galaxies: evolution – galaxies:formation 1 . 1 0 8 0 1 INTRODUCTION sive clusters, as shown by their surface brightness profiles : (Brough et al. 2005). However, that evolution must have v Brightest cluster galaxies (BCGs) are the most mas- predominantly occurred before z ∼ 1 as shown by the uni- i sive galaxies known in the Universe and are observed X formity in their absolute magnitudes (Brough et al. 2002), to have remarkably small dispersion in their absolute r and their steep metallicity gradients (Brough et al. 2007). a magnitudes compared to other early-type cluster galaxies An important clue in this context lies in the central (e.g. Arag´on-Salamanca et al. 1998; Collins & Mann 1998). Theirposition atthecentreofclusterssuggests theirevolu- galaxy luminosity–host-halo mass (LcMh) relation, which shows a scaling between BCG mass and that of its host tion and their unique properties are linked to their special environment (e.g. Edge 1991; von derLinden et al. 2007). halo. Lin & Mohr (2004) found that LK ∝ M200.206±0.04 for However, it is still not clear exactly which processes drive 93 BCGs at z < 0.1, Popesso et al. (2007) found Lr ∝ M0.33±0.04 for 217 BCGs at 0.01 < z < 0.25 in the Sloan their evolution. 200 Digital SkySurvey(SDSS),andHansen et al. (2007)found Ofcentralimportanceinthiscontextisthemassassem- blyhistoryofthesemostmassivegalaxies andhowthatde- hLii ∝ M200.300±0.01 for the SDSS MaxBCG cluster sample pendsontheirenvironment.Semi-analyticmodelsbasedon (0.1<z <0.3). Studiesof the LcMh-relation havenot ven- tured beyond z ∼ 0.3, so it is not clear whether this rela- mergertreesfromtheMillenniumN-bodysimulationpredict tionship evolveswith cosmic time. that,whilethestarsinBCGsformathigh-redshift,themass ofthegalaxyonlyassemblesrecently(doublinginmasssince The LcMh-relation is also a natural output from anal- z ∼ 1; de Lucia & Blaizot 2007), with the most massive yses of cosmological simulations: Halo Occupation Dis- galaxies in the most massive clusters having undergone the tribution functions (HOD; e.g. vanden Bosch et al. 2005; mostmassassembly.ObservationsofBCGsagreethattheir Zhenget al. 2007) and Conditional Luminosity Functions properties are consistent with having undergone mergers (CLF;e.g.Cooray&Milosavljevi´c2005;henceforthCM05). (e.g. Boylan-Kolchin et al. 2006; Bernardi et al. 2007) and These statistically assign numbers of galaxies or the lumi- BCGs in the most massive clusters do appear to have un- nosity distribution of galaxies in a given dark matter halo, dergonemorestellar massevolution than thoseinless mas- distinguishingbetweencentralandsatellitegalaxiesanden- ablinga direct comparison with observations. Arelated ap- proach is the application of a direct relationship between ⋆ E-mail:[email protected] the mass of a dark matter halo or sub-halo and the ob- 2 Brough et al. served galaxy luminosity (e.g. Vale & Ostriker 2006, 2007; ClusterX-rayluminosityisdirectlyproportionaltothe henceforthVO06,VO07).CM05predicthLi∝M<0.3 above square of the electron density of the intra-cluster medium 200 halo masses of 4×1013h−1M⊙ as the timescale on which and provides a measure of the host system mass. X-ray dynamical friction causes satellite galaxies to fall in to the luminosities for the low-redshift clusters were measured centre exceeds the age of the host system halo. VO06 pre- from the ROSAT All Sky Survey (RASS), paired with co- dict hLKi∝M100.208 as a result of the hierarchical formation ordinates from Abell et al. (1989) and Lynam 1999. The ofstructureintheUniverse,withaprescriptionforsub-halo X-ray luminosities for the high-redshift clusters were ob- massloss.WeinterprettheLcMh-relationinthisparadigm: tainedfromtheEinsteinMediumSensitivitySurvey(EMSS; that it results from the BCG growing in step with its host Gioia & Luppino 1994) and Serendipitous High-Redshift halo. Archival Cluster ROSAT catalogues (Burkeet al. 2003). In Yang et al. (2005) analysed the HOD of galaxy groups Brough et al.(2002)theRASSpassband(0.1−2.4kev)data in the 2-degree Field Galaxy Redshift Survey (0.01 < z < were transformed to the EMSS passband (0.3−3.5 keV). 0.2)andfoundhLcen,bJi∝Mh0.25.Morerecently,Yanget al. Hereweusethesametransformationtoplaceallthesevalues (2007) determined the CLF of groups (0.01 < z < 0.2) in intheRASSpassband,i.e.LX(0.1−2.4keV)=LX(0.3−3.5 the SDSS and found hLcen,ri ∝ Mh0.17. At higher redshifts, keV)/1.08.ThehostclusterscoverawiderangeinX-raylu- Zhenget al. (2007) analysed the HOD using the two-point minosity:5×1042 <LX <5×1044(h−1ergs−1)atredshifts correlation function of the DEEP2 galaxy redshift survey z<0.1 and6×1042 <LX <1×1045 (h−1 ergs−1)at red- (z¯∼1) in comparison to the SDSS(z¯∼0) and determined shifts 0.1<z<0.8. the LcMh-relation and its evolution. They found the VO06 TheGroupEvolutionMultiwavelengthSurvey(GEMS; modeltobeagood fittotheirdataatz∼0andz ∼1sug- Osmond & Ponman 2004; Forbes et al. 2006) is a heteroge- gesting that this relationship has not evolved. White et al. neous sample of 60 groups selected from optically-selected (2007) found hLcen,Bi ∝ Mh0.36 at z ∼ 0.5 from their HOD groupcatalogues observedbytheROSAT X-raysatellite to analysis of ∼ 2L⋆ red galaxies in the NOAO Deep Wide enableX-rayclassification.Removingthe6groupsforwhich Field Survey. These studies, however, do not directly iden- Osmond & Ponman (2004) were unable to find at least 4 tify central galaxies. galaxies associated with theoriginal optical position, leaves In this Letter, we examine the central-galaxy luminos- a sample of 54 groups with 2×1040 <LX <4×1043 (h−1 ity – host halo mass relation for a sample of known BCGs erg s−1). Of these 54 groups, 35 have extended intra-group in groups and clusters at z < 0.1 and, for the first time, X-rayemission,13haveX-rayemissiononlyassociatedwith the relationship for known BCGs in clusters at redshifts thecentral galaxy and 6are undetected (at 3σ aboveback- 0.1 < z < 0.8. This will enable us to directly determine ground)inX-rays.Osmond & Ponman(2004)alsoidentified the evolution in this relationship for known BCGs and to theBGG and published thegalaxies’ morphologies. comparetotheHODanalysesattheseredshifts.Itwillalso All the GEMS groups are covered by the 2 Mi- enable ustoexamine thede Lucia & Blaizot (2007) predic- cron All Sky Survey Extended Source Catalogue (2MASS; tion of these galaxies having doubled in mass since z ∼1. Jarrett et al. 2000). Our photometric reductions were per- In Section 2 we introduceour sample and then present formed to be as consistent as possible with Brough et al. our comparison of BGG and BCG luminosities with host (2002). All 54 groups are at redshifts z < 0.03 (z¯ ∼ 0.01) systemmassinSection3.Wediscusstheimplicationsofour andforalargeproportionofthesamplethecircularaperture results in Section 4 and draw our conclusions in Section 5. photometryin2MASSdoesnotextendtolargeenoughradii Throughout this paperwe assume H0=70 km s−1 Mpc−1, tomeasureanapertureofradius12.5h−1kpc(themaximum Ωm =0.3 and ΩΛ =0.7. possibleradius70′′ =14h−1 kpcatz∼0.01). Theaperture magnitudes for this sample were therefore measured manu- allyfromthe2MASSscansandcorrectedtoabsolutemagni- tudesfollowingBrough et al.(2002).Theyhavephotometric 2 DATA uncertainties of ∼0.04 mag. TheBCGsample isfrom Brough et al. (2002) andcontains ThegroupX-rayluminositiesaretakenfromtheGEMS 92 BCGs at z 6 0.1 and 63 at 0.1 < z < 0.8 (z¯ ∼ 0.3). groupcatalogue. Thesearebolometric luminosities, extrap- These are selected as the brightest galaxies closest to the olated to the radius corresponding to an overdensity of X-ray centroid of their host cluster and are all early-type 500 times the critical density (r500), with uncertainties of galaxies. logLX ∼0.05. Brough et al. (2007) showed that the GEMS InBrough et al.(2002),K-bandmagnitudesweremea- groupX-rayluminositiesareconsistentwiththosemeasured sured in an aperture of radius 12.5h−1 kpc. The apparent bytheROSAT-ESOFluxLimitedX-raygalaxyclustersur- magnitudeswerecorrectedtoabsolutemagnitudesusingK- vey (REFLEX; B¨ohringer et al. 2004) extrapolated to 12 andpassiveevolutioncorrections (Yoshii & Takahara1988) times the core radius of the cluster, with an uncertainty producedbytheGISSEL96stellarpopulationsynthesiscode of logLX = 0.05. We repeated our low-redshift group and (Bruzual & Charlot1993).Thegalaxieswereassumedtobe cluster analysis using REFLEX X-ray luminosities for the 10Gyrsold,tohaveformedinaninstantaneousburst,andto 40 clusters that are in common with this work and found have evolved passively since z ∼2. The assumptions of age that our results are unchanged. We therefore present this andredshiftofformationhavenegligibleeffectonthecorrec- work using thefull low-redshift sample of 92 clusters. tion value. The magnitudes were also corrected for galactic We convert the group X-ray luminosities to the mass absorption (typical values AK ∼0.02 mag) using the maps withinanoverdensityof200timesthecriticaldensity,M200, of Schlegel et al. (1998). The uncertainties in these magni- using the LX(r500,Bol)−M120.103±0.27 relation measured for tudesare ∼0.04 mag. 15 GEMS groups from Brough et al. (2006). We then use Luminosity – Mass Relation for BCGs 3 theLX(0.1−2.4keV)−M120.507±0.08 relation forclustersfrom Reiprich & B¨ohringer(2002)forourclusters. Weestimated the uncertainties in our mass measurements by perform- ing a Monte-Carlo analysis of how the uncertainties in the LX measurementspropagatethroughtoourmassestimates given theuncertainty in theslope of theLX−M200 relation used for our conversion. These were found to be of order 0.08 in log M200 for the groups and 0.02 for the clusters. Ettori et al. (2004) show that the Reiprich & B¨ohringer (2002) relationship does not evolve significantly over the redshift range analysed here, we therefore use the cluster LX(0.1−2.4keV)−M200 relationshipforourhigherredshift sample. 2.1 Aperture Magnitudes IntheirBCGstudy,Lin & Mohr(2004)makethecasethat aperturemagnitudesarenotsensitivetothemassgrowthof BCGsandinsteaduse2MASSK-bandisophotalmagnitudes (measured within the 20th magnitude isophote). We test whether thisaffects ourdata in several different ways: Wecomparedtheapparentmagnitudesforthe24galax- iesin common with Lin & Mohr(2004)and foundan offset of∆mK =0.35±0.03magresultingfromtheuseofthedif- ferentmeasurements(Figure1).Thisoffsetisnotcorrelated with themagnitudes or with cluster mass. We re-measure the magnitudes in a larger aperture Figure 1.Upper Panel:Comparingourapparentaperturemag- of 25h−1kpc for 4 BCGs at the same mean luminosity in nitudeswiththeapparentisophotalmagnitudesfromLin&Mohr high-mass (M¯ ∼4×1014h−1M⊙),and in low-mass clusters (2004). We find an offset of ∆mK =0.35±0.03 mag for the 24 (M¯ ∼1×1014h−1M⊙).Ifweweresubjecttoapertureeffects galaxies in common. Lower Panel: Comparing model aperture andisophotalluminositieswithhostsystemmass. thenthemagnitudesofthegalaxiesinthelow-massclusters would increase less than those in the high-mass clusters. However, the increases are consistent: ∆MK(low mass) = 0.39±0.02, ∆MK(high mass)=0.44±0.04. a bound system is not detected in the less-massive groups, Finally, we use the R-band surface brightness profiles it is more reliable to determine this relationship from the from Graham et al. (1996) of 24 galaxies in common with early-type BGGs in groups with extended X-ray emission thiswork.Wethenpredicted(assumingR−K =2.6)aper- and BCGs (eBGG+BCG sample). For these we find a rela- ture luminosities within r = 12.5h−1 kpc and isophotal lu- tionship LK ∝ M200.204±0.08. This is consistent with values minosities within K =20 magarcsec−2. The lower panel of from the literature (e.g. Lin & Mohr 2004; Popesso et al. Figure 1 shows that aperture luminosities follow the same 2007; Hansen et al. 2007). We note that the uncertainties relationship as isophotal luminosities with cluster mass. in these data have insignificant effect on the fitted rela- Wethereforeconcludethatourresultsarerobusttoany tions.AnalysingtheBCGsaloneyieldsashallowerrelation: possible aperture effects. LK ∝ M200.101±0.10. This relationship is consistent with no increase of BCG luminosity with increasing cluster mass. However, dueto thescatter in thedata it is also consistent with therelationship for theeBGG+BCG sample and with 3 RESULTS those relationships in the literature. IntheupperpanelofFigure2weplottheBCG(z <0.1)and WhethertheLcMh-relationevolveswithredshiftisalso BGGapertureK-bandluminositiesversustheirhostsystem interesting. We examine the 63 BCGs at higher redshift mass.Thegalaxiesinthelower-masssystemsarefainterand, (0.1 < z < 0.8) in the lower panel of Figure 2. The best- morphologically, more likely to be late-type galaxies (Sa to fitting relationship to these data is LK ∝ M200.208±0.11. This Irr) in systems without extended intra-group X-rays. The is entirely consistent with thelow-redshift relations. dispersioninthecentralgalaxyluminositiesdecreasesasthe In Figure 3 we show how our results compare to those host halo mass increases, as also observed by Zheng et al. measured in the literature, and to the predictions of CM05 (2007).Theysuggestthatthisisaresultofabroaderdistri- and VO06. To add another high-redshift (z ∼ 1) point, we bution of major star formation epochs in lower mass halos. assume that because the Zheng et al. (2007) data fit the This is consistent with our observation that many of the VO06model, they can berepresented byL∝M0.28. There galaxies in thelow-mass halos are late-types. is no appreciable offset between the relationship measured Fitting a relationship to these data we find LK ∝ from HOD analyses or directly from BCG samples. These M0.46±0.06 for all BGGs and BCGs. This is steeper than suggest that there is no evolution in this relationship be- 200 has previously been observed. However, given that the ex- tween z ∼ 0.01 and z ∼ 1 . Our data are also consistent tendedX-rayemission characteristicofthepotentialwellof with thepredictions of both CM05 and VO06. 4 Brough et al. 4 DISCUSSION BCGs are often thought to be the most conspicuous ex- amples of central cluster galaxies undergoing rapid growth throughmerging.Thesedatasuggestapictureinwhichthe near-infraredluminosity(∼stellarmass)ofearly-typeBCGs evolvessteadilywiththegrowthoftheirhostcluster.Inob- serving the LcMh-relation to have a slope ∼ 0.3 at higher redshift, BCG growth appears to still be in the ‘dynamical friction time-scale-limited’ regime (as per CM05) at these earliertimes.Hence,BCG-halomassgrowthiscontrolledby this restriction over the cosmic time interval studied here, and is not consistent with BCGs doubling in mass from z ∼ 1 to z ∼ 0 as suggested by deLucia & Blaizot (2007). These observations are more consistent with BCGs accret- ing 10−20 per cent in stellar mass as their cluster doubles in mass. This conclusion is consistent with the observation that the position of BCGs on scaling relations such as the fun- damental plane is a result of having undergone at least 1 major merger event (e.g. Boylan-Kolchin et al. 2006; Bernardi et al.2007).Thesemajormergereventshavebeen directly observed: Yamada et al. (2002) observed a nearly Figure 2.Therelationshipof BGG andBCGaperture K-band equal-mass BCG merger in a ∼ 2×1014h−1M⊙ cluster at luminosities with host system mass. The upper panel shows the z∼1.26. Rines et al. (2007) haveobserved a plumearound relationship for the BGGs and the BCGs in low-redshift clus- a BCG in a ∼ 1×1014h−1M⊙ cluster at z ∼ 0.4. Both ters (z 60.1). In this panel the points indicate the BCGs (filled of these systems lie within the scatter of our data on the squares) and the properties of the BGGs (circles denote early- LcMh-relation. type galaxies, triangles denote late-types) and their host groups (filled points indicate extended X-ray emission, open points in- Observations of the mass assembly of other massive dicate galaxy halo emission and arrows indicate upper limits early-typegalaxypopulationsarecontradictory.Someclaim for the X-ray undetected groups). The solid line indicates the LK ∝ M200.204±0.08 relation for the (eBGG+BCG sample), while evidence for no evolution beyond passive for these galaxies the dot-dashed line indicates the LK ∝ M200.101±0.10 relation for (cela.gim. Csitmelalattrimetaassl. e2v0o0l6u;tiBonunodfyaetfaacl.to2r00o6f)2, wsihniclsetzoth∼er1s the BCGs alone. The lower panel shows the relationship for the BCGs in high-redshift clusters (0.1 < z < 0.8). The solid (e.g. van Dokkum 2005; Bell et al. 2006). In between these line indicates the best-fit LK ∝ M200.208±0.11 relation for these extremes are studies more consistent with this work, find- high-redshiftdata and the dashed lineindicates the low-redshift ing evidence for growth of 6 50 per cent since z ∼ 1 eBGG+BCGsamplefordirectcomparison. (e.g. Glazebrook et al. 2004;Wakeet al. 2006;Brown et al. 2007;White et al.2007;Masjedi et al.2007;McIntosh et al. 2007). From a theoretical perspective, we have demonstrated that our data are consistent with the predictions of two HOD models at z ∼ 0.1. However, these models do not make predictions for higher redshifts. Examining the evo- lution of BCGs in a Λ-Cold Dark Matter Universe, the semi-analyticalmodelofde Lucia & Blaizot(2007)predicts that ∼50 per cent of the mass of BCGs is assembled since z ∼1, with an equivalent growth in the host halo over this timescale. This is inconsistent with our observations. How- ever, Almeida et al. (2007) have analysed the stellar mass evolutionoflargeredgalaxiesbetweenz ∼0.24andz∼0.5 intheDurhamsemi-analyticmodels,andfindthat,onaver- Figure 3. Comparing our results with gradients from LcMh- age, large redgalaxies assemble ∼25percentof theirmass relationstudiesintheliteraturewithredshift.Starsindicatethe since z∼1. predictions from CM05 and VO06, filled squares indicate previ- ous BCG studies, open squares indicate previous HOD studies Theresultspresentedinthisletterareclearlyconsistent andopencirclesindicatetheresultspresentedhere. with the picture emerging from studies of the evolution of massiveearly-typegalaxies andsemi-analyticmodelsofthe evolution of these massive galaxies, that massive galaxies haveincreased in stellar mass by650 percent sincez∼1. Luminosity – Mass Relation for BCGs 5 5 CONCLUSIONS Cooray A., Milosavljevi´c M., 2005, ApJ, 627, L85, CM05 de Lucia G., Blaizot J., 2007, MNRAS,375, 2 We have examined the BGG and BCG luminosity – host Edge A. C., 1991, MNRAS,250, 103 halo mass relation at: z <0.1 and, at higher redshifts than EttoriS.,Tozzi P.,Borgani S.,RosatiP.,2004, A&A,417, previously measured: 0.1<z <0.8. Weconclude that: 13 • BCGs follow a relationship with themass of theirhost Forbes D. A., et al. 2006, PASA,23, 38 cluster. This relationship does not evolve from z ∼ 1 to Gioia I. M., LuppinoG. A., 1994, ApJS, 94, 583 z∼0. Glazebrook K., et al., 2004, Nature, 430, 181 • Our data are consistent with the predictions of two il- Graham A., Lauer T. R., Colless M., Postman M., 1996, lustrativeHODmodelsfortherelationshipofcentralgalaxy ApJ,465, 534 luminosity with host halo mass. HansenS.M.,SheldonE.S.,WechslerR.H.,KoesterB.P., • TheBCG–halo-massrelationappearstobecontrolled 2007, arXiv:0710.3780 bythetimescalefordynamicalfrictionevenat0.1<z<0.8. Jarrett T. H., Chester T., Cutri R., Schneider S., Skrut- skieM., HuchraJ. P., 2000, AJ, 119, 2498 Lin Y-T., Mohr J. J., 2004, ApJ, 617, 879 LynamP.D.,1999,PhDthesis,LiverpoolJohnMooresUni- ACKNOWLEDGEMENTS versity We thank the referee for instructive comments. We also Masjedi M., Hogg D. W., Blanton M. R., 2007, thankM.J.I.Brown,C.Blake,A.Arag´on-SalamancaandA. arXiv:0708.3240 Edge for helpfuldiscussions. SBand WJC acknowledge the McIntosh D.H.,GuoY.,HertzbergJ., KatzN.,Mo H.J., support of the Australian Research Council. DJB acknowl- vanden Bosch F. C., YangX., 2007, arXiv:0710.2157 edgesthesupportofNASAcontractNAS8-39073.Thispub- Osmond J. P.F.,Ponman T.J., 2004, MNRAS,350, 1511 lication makes use of data products from the Two Micron PopessoP.,BivianoA.,B¨ohringerH.,RomanielloM.,2007, AllSkySurvey(2MASS)whichisajointprojectoftheUni- A&A,464, 451 versity of Massachusetts and the Infrared Processing and Reiprich T. H., B¨ohringer H., 2002, ApJ,567, 716 Analysis Center/California Institute of Technology, funded Rines K., Finn R., Vikhlinin A., 2007, ApJ, 665, 9 bytheNationalAeronauticsandSpaceAdministration and SchlegelD.J.,FinkbeinerD.P.,DavisM.,1998,ApJ,500, theNational Science Foundation. 525 Vale A., OstrikerJ. P., 2006, MNRAS,371, 1173 Vale A., OstrikerJ. P., 2007, astro-ph/0701096 vandenBoschF.C.,WeinmannS.M.,YangX.,MoH.J., REFERENCES LiC., Jing Y.P., 2005, MNRAS,361, 1203 Abell G., Corwin H.G., Olowin R.P., 1989, ApJS,70, 1 van DokkumP., 2005, AJ, 130, 2647 Almeida C., Baugh C. M., Wake D. A., Lacey C. G., von der Linden A., Best P. N., Kauffmann G., Benson A. J., Bower R. G., Pimbblet K., 2007, WhiteS. D.M., 2007, 379, 867 arXiv:0710.3557 WakeD. A., et al., 2006, MNRAS,372, 537 Arag´on-Salamanca A.,BaughC.M.,KauffmannG.,1998, WhiteM.,ZhengZ.,BrownM.J.I.,DeyA.,JannuziB.T., MNRAS,297, 427 2007, ApJ, 655, 69 Bell E. F., et al., 2006, ApJ,640, 241 Yamada T., Koyama Y., Nakata F., Kajisawa M., Bernardi M., Hyde J. B., Sheth R. K., Miller C J., TanakaI.,KodamaT.,OkamuraS.,DeProprisR.,2002, Nichol R.C., 2007, AJ, 133, 1741 ApJL,577, 89 B¨ohringer H., et al., 2004, A&A,425, 367 YangX.,MoH.J.,YingY.P.,vandenBosch F.C.,2005, Boylan-Kolchin M., Ma C., Quataert E., 2006, MNRAS, MNRAS,358, 217 369, 1081 Yang X., Mo H. J., van den Bosch F. C., 2007, BroughS.,CollinsC.A.,BurkeD.J.,MannR.G.,Lynam arXiv:0710.5096 P. D.,2002, MNRAS,329, 533 Yoshii Y., Takahara F., 1988, ApJ, 326, 1 BroughS.,CollinsC.A.,BurkeD.J.,LynamP.D.,Mann Zheng Z., Coil A.L., ZehaviI., 2007, ApJ,667, 760 R.G., 2005, MNRAS,364, 1354 Brough S., Forbes D. A., Kilborn V. A., Couch W., 2006, ThispaperhasbeentypesetfromaTEX/LATEXfileprepared MNRAS,370, 1223 bythe author. Brough S., Proctor R., Forbes D. A., Couch W. J., Collins C. A., Burke D. J., Mann R. G., 2007, MNRAS, 378, 1507 Brown M. J. I., Dey A., Jannuzi B. T., Brand K., Ben- son A. J., Brodwin M., Croton D. J., Eisenhardt P. R., 2007, ApJ, 654, 858 Bruzual A.G., Charlot S., 1993, ApJ, 405, 538 Bundy K., et al., 2006, ApJ, 651, 120 Burke D. J., Collins C. A., Sharples R. M., Romer A. K., Nichol R.C., 2003, MNRAS,341, 1093 Cimatti A., DaddiE., RenziniA., 2006, A&A,453, 29 Collins, C. A. & Mann, R. G. 1998, MNRAS,297, 128