Protoclusters associated with z > 2 radio galaxies. I. Characteristics of high redshift protoclusters

1.9 MB·English

by  B. P. Venemans

Astronomy&Astrophysicsmanuscriptno.3941 February5,2008
(DOI:willbeinsertedbyhandlater)
Protoclusters associated with z > 2 radio galaxies ⋆
I. Characteristics of high redshift protoclusters
B.P.Venemans1,2,H.J.A.Ro¨ttgering1,G.K.Miley1,W.J.M.vanBreugel3,8,C.DeBreuck4,J.D.Kurk5,L.
Pentericci6,S.A.Stanford3,9,R.A.Overzier1,S.Croft3,8,andH.Ford7
7
0 1 SterrewachtLeiden,P.O.Box9513,2300RA,Leiden,TheNetherlands
0 2 InstituteofAstronomy,UniversityofCambridge,MadingleyRoad,Cambridge,CB30HA,UnitedKingdom
2 3 LawrenceLivermoreNationalLaboratory,P.O.Box808,LivermoreCA,94550,USA
n 4 EuropeanSouthernObservatory,KarlSchwarzschildStraße2,D-85748Garching,Germany
a 5 INAF,OsservatorioAstrofisicodiArcetri,LargoEnricoFermi5,50125,Firenze,Italy
J 6 OsservatorioAstronomicodiRoma,ViadiFrascati33,00040MontePorzioCatone,Italy

3 7 Dept.ofPhysics&Astronomy,TheJohnsHopkinsUniversity,3400NorthCharlesStreet,BaltimoreMD,21218-2686,USA
8 UniversityofCalifornia,Merced,P.O.Box2039,Merced,CA95344,USA

2 9 UniversityofCalifornia,Davis,1ShieldsAve,Davis,CA95616,USA
v
7
Received29July2005/accepted5October2006
6
5
0 Abstract.WepresenttheresultsofalargeprogramconductedwiththeVeryLargeTelescopeandaugmentedbyobservations
1 withtheKecktelescopetosearchforformingclustersofgalaxiesnearpowerfulradiogalaxiesat2.0<z<5.2.BesidesMRC
6 1138–262 atz = 2.16,theradiogalaxyobservedinourpilotprogram,weobtainednarrow-andbroad-bandimagesofeight
0 radiogalaxiesandtheirsurroundings.TheimagingwasusedtoselectcandidateLyαemittinggalaxiesin∼3×3Mpc2 areas
/ neartheradiogalaxies.Atotalof300candidateemitterswerefoundwitharest-frameLyαequivalentwidthofEW > 15Å
h 0
andsignificanceΣ≡EW /∆EW >3.Follow-upspectroscopywasperformedon152candidatesinsevenoftheradiogalaxy
p 0 0
- fields.Ofthese,139wereconfirmedtobeLyαemitters,fourwerelowredshiftinterlopersandninewerenon-detections.With
o theadoptedcriteriathesuccessrateis139/152=91%.Inaddition,14objectswithEW <15and/orΣ<3wereconfirmedto
0
r beLyαemitters.Combinedwiththe15LyαemittersnearMRC1138–262,wehavedeterminedLyαredshiftsfor168objects
t
s neareightradiogalaxies.
a
AtleastsixofoureightfieldsareoverdenseinLyαemittersbyafactor3–5ascomparedtothefielddensityofLyαemitters
:
v atsimilarredshifts,althoughthestatisticsinourhighestredshiftfield(z = 5.2)arepoor.Also,theemittersshowsignificant
i clusteringinvelocityspace.Intheoverdensefields,thewidthofthevelocitydistributionsoftheemittersisafactor2–5smaller
X
thanthewidthofthenarrow-bandfilters.Takentogether,weconcludethatwehavediscoveredsixformingclustersofgalaxies
ar (protoclusters).Weestimatethatroughly75%ofpowerful(L2.7GHz >1033ergs−1Hz−1sr−1)highredshiftradiogalaxiesreside
inaprotocluster.Theprotoclustershavesizesofatleast1.75Mpc,whichisconsistentwiththestructuresizesfoundbyother
groups.Byusingthevolumeoccupiedbytheoverdensitiesandassumingabiasparameterofb = 3−6,weestimatethatthe
protoclustershavemassesintherange2−9×1014M⊙.Theseprotoclustersarelikelytobeprogenitorsofpresent-day(massive)
clustersofgalaxies.Forthefirsttime,wehavebeenabletoestimatethevelocitydispersionofclusterprogenitorsfromz∼5to
∼2.Thevelocitydispersionoftheemittersincreaseswithcosmictime,inagreementwiththedarkmattervelocitydispersion
innumericalsimulationsofformingmassiveclusters.
Keywords.galaxies:active—galaxies:clusters:general—cosmology:observations—cosmology:earlyUniverse—cos-
mology:largescalestructureofUniverse
1. Introduction
Send offprint requests to: B. P. Venemans, e-mail:
[email protected] Clusters of galaxiesare the largestand most massive gravita-
⋆ Based on observations carried out at the European Southern tionallyboundstructuresintheUniverse.Theyareinteresting
Observatory, Paranal, Chile, programs 66.A-0597, LP167.A-0409, objectstostudyformanyreasons.
68.B-0295and70.A-0589.AlsobasedondataobtainedattheW.M.
KeckObservatory,whichisoperatedasascientificpartnershipamong
theCaliforniaInstituteofTechnology,theUniversityofCaliforniaand wasmadepossiblebythegenerousfinancialsupportoftheW.M.Keck
theNationalAeronauticsandSpaceAdministration.TheObservatory Foundation.
2 B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies
First,clusterscontainlargenumbersofgalaxiesatspecific powerful high redshift radio galaxies might be such tracers,
redshifts,makingthemexcellentlaboratorieswithwhichtoin- as they are forming massive galaxies in dense environments
vestigate the formation and evolution of galaxies. For exam- (e.g., Carillietal., 1997; Deyetal., 1997; Athreyaetal.,
ple, the analysis of galaxies in z ∼ 1 clusters showed that 1998; Pentericcietal., 1998; Papadopoulosetal., 2000;
the stars in massive, early-type galaxies formed at z > 2 Pentericcietal., 2000b; Archibaldetal., 2001; Jarvisetal.,
(e.g., Ellisetal., 1997; Stanfordetal., 1998; Blakesleeetal., 2001; DeBreucketal., 2002, 2003a,b; Stevensetal., 2003;
2003; vanDokkum&Stanford, 2003; Holdenetal., 2005). Zirmetal., 2003; Reulandetal., 2004, see e.g. Carilli et al.
Investigatingthegalaxypopulationof(forming)clustersatz> 2001fora review).Targetedsearchesforcompaniongalaxies
2 could provide knowledge of the formation process of such nearpowerfulradiosourcesatz>1havelongyieldedpromis-
massivegalaxies(e.g.,Eggenetal.,1962;Larson,1974).Also ing results (e.g., LeFe`vreetal., 1996; Pascarelleetal., 1996;
becauseclustersarethemostextremeoverdenseregionsinthe Keeletal., 1999; Sa´nchez&Gonza´lez-Serrano, 1999, 2002;
Universe, they allow an efficient investigation of the interac- Best, 2000; Halletal., 2001; Nakataetal., 2001; Bestetal.,
tionbetweengalaxiesandtheirenvironment(e.g.,Milesetal., 2003;Woldetal.,2003;Barretal.,2004).
2004; Tanakaetal., 2004; vanZeeetal., 2004; Goto, 2005; We therefore started a large program with the Very Large
Nakataetal.,2005;Tranetal.,2005). Telescope(VLT) tosystematically searchforgalaxyoverden-
Asecondreasontostudyclustersisthattheycanplacecon- sities near radio galaxies in the redshift range 2 < z < 5.2.
straints on cosmology. The number density of massive clus- This program was initiated after a successful pilot project in
ters is a strong function of the fundamentalcosmological pa- whichtheenvironmentoftheradiogalaxyMRC1138–262at
rametersΩ and σ , and the evolutionof cluster abundances z=2.16wasinvestigated.Deepnarrow-bandimagesofthera-
M 8
withredshiftdependsprimarilyonΩ (e.g.,Ekeetal.,1996). dio galaxy and the surrounding field were obtained to search
M
The number density of rich clusters at z > 0.5 has already foranexcessofLyαemittinggalaxies(Kurketal.,2000).The
been successfully used to constrain the values of cosmologi- imagingandfollow-upspectroscopyresultedin∼40candidate
calparameters(e.g.,Bahcalletal.,1997;Bahcall&Fan,1998; Lyαemitters(Kurketal.,2000,2004b)ofwhich15werecon-
Ettorietal.,2003). firmedatz = 2.16±0.02(Pentericcietal.,2000a;Croftetal.,
Several studies of massive clusters with redshifts up to 2005).ThepresenceofaformingclusterassociatedwithMRC
z=1.4havefoundlittleevolutionintheclusterproperties(e.g., 1138–262 was firmly established by the subsequent discov-
Tozzietal., 2003; Hashimotoetal., 2004; Maughanetal., eryofsignificantpopulationsof(spectroscopically-confirmed)
2004;Rosatietal.,2004;Mullisetal.,2005).Despitethelarge Hα emitters (Kurketal., 2004a,b), QSOs (Pentericcietal.,
lookbacktimes, clusters at z ∼ 1 appear to be verysimilar to 2002;Croftetal.,2005)andextremelyredobjects(Kurketal.,
localclusters.Forexample,thez = 1.3clusterRDCS1252.9- 2004b). The VLT large program was augmented by observa-
2927 has thermodynamicalpropertiesand metallicity that are tionswith the Kecktelescope, to take advantageof the excel-
very similar to those of lower redshift clusters (Rosatietal., lent UV throughputof LRIS-B for spectroscopy and imaging
2004). To study when and how clusters and their galaxies belowtheLymanbreakofsomeofthez∼3objects.
formed,asampleofclustersatz≫1isneeded. In this paper, we present the results of our program to
Unfortunately, conventional methods for finding distant search for forming clusters (protoclusters) near radio galax-
clustersbecomeimpracticalatz>1.SearchesforextendedX- ies up to z = 5.2. Early results of the program, the discov-
ray sources are difficult because the surface brightness of the ery ofgalaxyoverdensitiesat z = 4.1and z = 5.2,havebeen
X-ray emission fades as (1+z)4. Although large optical sur- presented in Venemansetal. (2002) and in Venemansetal.
veyshave been successful in finding galaxyclusters at z . 1 (2004).Adetailedanalysisofthefieldtowardstheradiogalaxy
bysearchingforconcentrationsofredgalaxies(e.g.,Gladders, MRC0316–257atz = 3.13isgiveninVenemansetal.(2005,
2002), the detection of z > 1 clusters with the same method hereafterV05).InV05thedatareductionsteps,selectionpro-
requiressensitive, wide field near-infraredcameras which are cedureofcandidateLyαemittersandtheassessmentofthedata
notyetavailable.Inthefuture,surveysexploitingtheSunyaev- quality are described in detail. Here we will follow the same
Zeldovich(SZ)effect(e.g.,Carlstrometal.,2002)willbeable steps as V05 for the data reduction and analysis of the other
todetectclustersofgalaxiesatz≫1.However,atthismoment fields.Whileinthispaperwewillfocusontheenvironmentof
thesensitivityofSZsurveysisnotsufficienttodetectanyofthe radio galaxies, in the second paper of this series we will de-
known clusters at z > 1 (Carlstrometal., 2002; Rosatietal., scribe the properties of the individual Lyα emitting galaxies
2004). In recent years, several distant forming clusters (pro- discoveredinourprogram.
toclusters) have been serendipitously discovered in field sur- The structure of this paper is as follows: in Sect. 2.1 we
veys (e.g., Steideletal., 1998, 2005; Shimasakuetal., 2003; presentthetargetsofourprogram,andinSect.2.2–2.4wegive
Ouchietal., 2005). For example, Steideletal. (1998) found thedetailsoftheimagingobservations,theselectionofcandi-
a large scale structure of Lyman Break galaxies (LBGs) at date Lyα emitters and follow-up spectroscopy.In Sect. 3, the
z ∼ 3.1 in one of their fields. This single discovery demon- resultsoftheimagingandspectroscopyaredescribedforeach
stratedthepowerofsuchstructuresasitcouldbederivedthat individualfield.Asummaryoftheresultsandtheevidencefor
LBGsmustbeverybiasedtracersofmass. thepresenceofprotoclustersneartheradiogalaxiesaregiven
A different approach to find (forming) clusters is to inSect.4.Thepropertiesoftheprotoclustersarepresentedin
search for a galaxy concentration near a presumed tracer Sect.5,followedbyadiscussionofourresultsinSect.6anda
of high density regions. There is considerable evidence that summaryinSect.7.
B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies 3
Table1. Detailsoftheradiogalaxiesobservedinourprogram. 2.2.Imagingobservations
To search for structures of Lyα emitting galaxiesnear the ra-
Name α δ z La diogalaxies,thefieldssurroundingtheradiogalaxieswereob-
J2000 J2000 2.7GHz
BRL1602–174 160501.7 −173418.4 2.04 2.0×1034 servedin a narrow-bandfilter andat least onebroad-bandfil-
MRC2048–272 205103.5 −270304.1 2.06 6.3×1033 ter.Thenarrow-bandfilterswerechosentoencompasstheLyα
MRC1138–262 114048.2 −262909.5 2.16 1.3×1034 lineattheredshiftoftheradiogalaxy,andthebroad-bandfil-
MRC0052–241 005429.8 −235131.1 2.86 8.6×1033 terswereselectedtomeasuretheUVcontinuumredwardofthe
MRC0943–242 094532.7 −242849.7 2.92 7.2×1033 Lyαline.
MRC0316–257 031812.0 −253510.8 3.13 1.4×1034
All the narrow-bandimaging and most of the broad-band
TNJ2009–3040 200948.1 −304007.4 3.16 2.8×1033
imaging obtained in the large program were performed with
TNJ1338–1942 133826.1 −194230.8 4.11 9.6×1033
the FOcal Reducer/ low dispersion Spectrograph 2 (FORS2;
TNJ0924–2201 092419.9 −220142.0 5.20 1.5×1034
Appenzeller&Rupprecht, 1992) in imaging mode. Before
a Radio luminosity at a rest-frame frequency of 2.7 GHz in 2002 April, the detector in FORS2 was a SiTE CCD with
ergs−1Hz−1sr−1 2048×2048 pixel2 and a pixel scale of 0′.′2pixel−1. In 2002
April, the SiTE CCD was replaced by two MIT CCDs each
with2048×2048pixel2withascaleof0′.′125pixel−1.Thepix-
elswerebinnedby2×2,whichdecreasesthereadouttimebya
factorof2andgivesapixelsscaleof0′.′25pixel−1.Thefieldof
view, which is restricted by the geometryof the Multi-Object
In this article, we adopt a Λ-dominated cosmology with Spectroscopyunit,is6.′8×6.′8.
H = 70kms−1Mpc−1,Ω = 0.3,andΩ = 0.7.Magnitudes
0 M Λ Additional (broad-band) imaging was obtained using the
aregivenintheABsystem(Oke,1974).
Low Resolution Imaging Spectrometer (LRIS, Okeetal.,
1995) on the Keck I telescope. LRIS has two arms: a blue
channel which is optimised for observations in the blue part
of the optical spectrum (LRIS-B, see McCarthyetal., 1998;
2. Observations
Steideletal., 2004, for more information)and a red channel
for observing in the red (LRIS-R, Oke et al. 1995). The red
2.1.Sampleselection
armisequippedwithaTektronixCCDwith2048×2048pixel2.
The pixel scale is 0′.′21pixel−1, resulting in a field of view of
Thetargetsforourprogramwereselectedfromalistofapprox-
7.′3×7.′3.Thebluechannelhastwo2048×4096MarconiCCDs
imately150radiogalaxieswithknownredshiftsofz > 2.Our
with a pixel scale of 0′.′135pixel−1. There is a small gap be-
targetswerechosentohavelargeradioluminosities(L >
1033 ergs−1Hz−1sr−1). Because our goal was to sea2.r7cGhHzfor tweentheCCDsofroughly13′.′5.Thefieldofviewis∼8′×8′1.
Observationsweresplitintoseparateexposuresoftypically
companion Lyα emitting galaxies near the radio sources, the
1200–1800s in the narrow-bandand 240–800s in the broad-
radiogalaxieshadtohaveredshiftsoptimumforimagingwith
band.Individualexposureswereshifted10′′–15′′ withrespect
thenarrow-bandfiltersthatwereavailableattheVLT.Also,the
radiosourcesneedto lie in thesouthernhemisphere(δ . 0◦) toeachothertofacilitatetheidentificationofcosmicraysand
theremovalofresidualflat-fielderrors.Thedatawerereduced
to allow deep imaging of the source with the VLT. Applying
these criteria to the list of∼ 150z > 2 radiosourcesreduced usingstandardroutineswithinthereductionsoftwarepackage
IRAF2. These routines included bias subtraction using either
thenumberofpossibletargetsforourprogramto16ofwhich
nineareatz ∼2.1(thisincludesthetargetofourpilotproject, bias frames or the overscan region of the CCD, flat fielding
MRC1138–262atz = 2.16),threeareatz ∼ 2.9andfourare withtwilightskyflatsandilluminationcorrectionusingtheun-
atz ∼ 3.1.Giventherelativelysmallnumberofradiosources registeredscienceframes.
AllscienceimageswereregisteredontheICRFastromet-
atz> 2.9thatcouldbeobservedwiththeVLT,wedidnotap-
ricframeofreference(Ma&Feissel,1998),usingtheUSNO-
ply other criteria. To extend the redshift range, we purchased
A2.0 catalogue (Monetetal., 1998; Monet, 1998). The rela-
narrow-bandfiltersthatwerecentredonthewavelengthofthe
tive positionsof objectsin the fields are accurate to 0′.′1–0′.′2.
Lyα line at a redshift of z = 4.1 and z = 5.2. The following
The absolute accuracy is dominated by the uncertainty in the
nineradiosourceswerechosenastargetsforourprogram:BRL
USNO-A2.0catalogueof0′.′25(Deutsch,1999).
1602–174(hereafter1602),MRC2048–272(2048)andMRC
The photometric calibration was performed using several
1138–262(1138,allthreeatz = 2.1),MRC0052–241(0052)
photometric and spectrophotometric standard stars from the
andMRC 0943–242(0943,bothat z = 2.9),MRC 0316–257
catalogues of Stone & Baldwin (1983), Baldwin & Stone
(0316)andTNJ2009–3040(2009,bothatz=3.1),TNJ1338–
1942 (1338, z = 4.1) and TN J0924–2201 (0924, z = 5.2).
1 Thespecificationsforthebluechanneldetectorsgivenhereapply
Because these targetswere selected only on the basis of their
fordatatakenafter2002June.
radio luminosity and position on the sky, we regardour sam- 2 IRAF is distributed by the National Optical Astronomy
ple as representativeof luminousradio sources. The position, Observatories,whichareoperatedbytheAssociationofUniversities
redshift and radio power of the targets are given in Table 1. for Research in Astronomy, Inc., under cooperative agreement with
IndividualradiogalaxiesarebrieflydescribedinSect.3. theNationalScienceFoundation.
4 B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies
Table2. Overviewoftheimagingobservationsoftheradiogalaxyfields.
Date Telescope Instrument Field Filter λa(Å) ∆λb(Å) Seeing tc Depthd
c exp
2001Mar24&25 VLTUT2 FORS2 1338-1e FILT 621 5 6199 59 0′.′6 33300 28.2
2001Mar24&25 VLTUT2 FORS2 1338-1e SpecialR 6550 1650 0′.′6 6300 28.9
2001Mar24–26 VLTUT2 FORS2 0943 HeII/6500 4781 68 0′.′7 22500 28.6
2001Mar25&26 VLTUT2 FORS2 0943 BesselB 4290 880 0′.′9 4500 28.7
2001Mar25&26 VLTUT2 FORS2 1602 OII 3717 73 0′.′8 15000 26.6
2001Mar26 VLTUT2 FORS2 1602 BesselB 4290 880 0′.′65 2700 27.6
2001May21–23 VLTUT2 FORS2 2048 OII 3717 73 0′.′95 25200 –f
2001May22&23 VLTUT2 FORS2 2048 BesselB 4290 880 1′.′05 3600 –f
2001Sep20–22 VLTUT4 FORS2 2048 OII 3717 73 0′.′95 25200 27.9
2001Sep20&21 VLTUT4 FORS2 2048 BesselB 4290 880 1′.′05 3000 28.8
2001Sep20&21 VLTUT4 FORS2 0316 OIII/3000 5045 59 0′.′7 23400 28.4
2001Sep20&21 VLTUT4 FORS2 0316 BesselV 5540 1115 0′.′7 4860 28.9
2001Sep22 VLTUT4 FORS2 0052 HeII 4684 66 0′.′75 5400 –f
2001Oct20 VLTUT4 FORS2 0052 HeII 4684 66 0′.′75 18000 28.3
2001Oct20 VLTUT4 FORS2 0052 BesselB 4290 880 0′.′8 4800 29.0
2002Mar8 VLTUT4 FORS2 0924 FILT 753 8 7528 89 0′.′8 14400 –f
2002Mar8 VLTUT4 FORS2 0924 BesselI 7680 1380 0′.′8 2700 –f
2002Apr17–19 VLTUT4 FORS2 0924 FILT 753 8 7528 89 0′.′8 28800 28.1
2002Apr17–19 VLTUT4 FORS2 0924 BesselI 7680 1380 0′.′8 8640 28.5
2002Apr17–19 VLTUT4 FORS2 1338-2e FILT 621 5 6199 59 0′.′75 25200 28.2
2002Apr17–19 VLTUT4 FORS2 1338-2e SpecialR 6550 1650 0′.′75 4500 28.9
2002Apr17–19 VLTUT4 FORS2 2009 OIII/3000 5045 59 0′.′9 21600 –f
2002Apr17–19 VLTUT4 FORS2 2009 BesselV 5540 1115 0′.′85 4800 –f
2002Apr19 VLTUT4 FORS2 0924 BesselV 5540 1115 1′.′05 3600 28.6
2002Sep6–8 VLTUT4 FORS2 0316 BesselI 7680 1380 0′.′7 4680 28.7
2002Sep8 VLTUT4 FORS2 2009 OIII/3000 5045 59 0′.′9 7200 28.0
2002Sep8 VLTUT4 FORS2 2009 BesselV 5540 1115 0′.′85 2400 28.6
2002Sep8 VLTUT4 FORS2 0052 BesselV 5540 1115 0′.′75 5400 29.2
2002Sep8 VLTUT4 FORS2 0052 BesselI 7680 1380 0′.′55 4800 28.5
2003Jan31 KeckI LRIS-B 0316 u′ 3550 600 1′.′25 4050 –f
2003Feb1&4 KeckI LRIS-B 0316 u′ 3550 600 1′.′25 9000 29.8
2003Feb4 KeckI LRIS-B 0943 u′ 3550 600 1′.′2 9600 –f
2003Feb4 KeckI LRIS-R 0943 V 5473 948 1′.′25 5600 29.4
2004Jan19 KeckI LRIS-B 0943 u′ 3550 600 1′.′2 7000 29.8
2004Jan19 KeckI LRIS-R 0943 I 8331 3131 0′.′90 6000 28.4
aCentralwavelengthofthefilterinÅ.
bFullwidthathalfmaximum(FWHM)ofthefilterinÅ.
cTotalexposuretimeinseconds.
d1σdepthoftheresultingimageinmagper(cid:3)′′.
eThefieldtowardsTNJ1338–1942wasobservedattwodifferentpointings,seeSect.3.8.
f Forimagesobtainedduringtwodifferentobservingsessions,onlythetotaldepthislistedinthefinalentry.
(1984),Oke (1990)andLandolt(1992).The magnitudezero- 2.3.Candidateselection
points derived from these standard stars are consistent with
Objects in the images were detected using the program
each other within 2–3%. Zero-points in the Vega system are
SExtractor(Bertin&Arnouts,1996).Thenarrow-bandimages
converted to the AB system using the transformations of
weretakentodetectobjects,andaperturephotometrywassub-
Bessell (1979) and Smithetal. (2002). The zero-points were
sequently performed on both the narrow-bandand the broad-
correctedforgalacticextinctionasestimatedbySchlegeletal.
bandimages.
(1998).
To assess the completeness of the source detection, artifi-
cialandrealpointsourceswereaddedtothenarrow-bandim-
ageandrecovered.Thecompletenesslimitwasdefinedasthe
Table2summarizestheimagingobservationsofourLarge narrow-bandmagnitudeatwhich50%oftheaddedsourceswas
Program targets and gives the properties of the narrow- and recovered(seeV05fordetails).
broad-bandfiltersthatwereused.Also,theimagequality(see- Detectedobjectswererequiredtohaveasignal-to-noiseof
ing)anddepth(in1σlimitingmagnitudespersquarearcsec- > 5 in thenarrow-bandimage.Thecolorsofthedetectedob-
ond)aregiveninTable2. jectsweremeasuredincircularapertures,whilethe“total”flux
B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies 5
Table 3. Depth, sensitivity and size of the imaging observa- 2.4.Spectroscopicobservations
tionsoftheradiogalaxiesfields,andthenumberofcandidates
Toconfirmwhetherthecandidateemissionlineobjectsarelo-
selectedineachfield.
cated at the redshift of the radio galaxy, spectra were taken
Field ma Fb Lc Nd Areae ofcandidateemitters.Prioritywasgiventothemostluminous
lim 5σ 5σ
ergs−1cm−2 ergs−1 arcmin2 candidates.InTable4asummaryisgivenofthespectroscopic
1602 24.4 9.7×10−17 3.4×1042 2 42.3 observationsofcandidateemittersintheradiogalaxyfields.
2048 25.4 3.3×10−17 1.2×1042 10 43.6 Mostofthespectroscopicobservationswereperformedus-
1138 25.2f 3.5×10−17f 1.4×1042f 37f 46.6f inguser-definedmasks(themulti-objectspectroscopy(MXU)
0052 26.1 1.1×10−17 8.4×1041 57 44.9 mode of FORS2 and multi-slit spectroscopy(MSS) modesof
0943 26.1 7.4×10−18 6.2×1041 65 46.5 LRIS in Table 4). This mode allowed us to observe between
0316 26.3 6.9×10−18 6.9×1041 77 45.8
20and40objectsperslitmask.Theslitstypicallyhadalength
2009 25.7 1.3×10−17 1.3×1042 21 46.7 of10′′–12′′andawidthof1′.′0–1′.′4.Thegrismsusedwerese-
1338-1 26.0 4.7×10−18 8.9×1041 31g 40.1
lectedtohavethehighestthroughputatthewavelengthofthe
1338-2 26.2 5.8×10−18 1.1×1042 33g 48.9
0924 25.5 7.0×10−18 2.3×1042 14 46.8 Lyαlineoftheradiogalaxyanda resolutionthatmatchesthe
width of the Lyα lines (∼ 300 kms−1, V05) to maximize the
a Magnitude at which 50% of artificial and real point sources that confirmationrate.
wereaddedtothenarrow-bandimage,wererecovered. Individual exposures were typically 1800–2700 s, which
bLinefluxofanemitterwithnocontinuumthatisdetectedatthe5σ ensuredthatthespectrawerelimitedbytheskynoise.Between
levelinanaperturewithadiametertwicethatoftheseeingdisc. theexposuresthepointingofthetelescopewasshifted2′′–5′′
c Lineluminosityofanemitterwithnocontinuumthatisdetectedat
alongtheslitstoenablemoreaccurateskysubtraction.Forthe
the 5σ level in an aperture with a diameter twice that of the seeing
fluxcalibrationlongslitexposuresof(atleast) oneofthefol-
disc.
lowing spectroscopic standard stars EG 274, Feige 67, Feige
d NumberofcandidateLyαemittersthatfulfilstheselectioncriteria
110,GD 108,LTT 377,LTT 1020,LTT 1788,LTT 6248and
EW >15ÅandEW /∆EW >3.
0 0 0
eImagingareausefulforselectionofLyαemitters. LTT 7987 (Stone&Baldwin, 1983; Baldwin&Stone, 1984;
f ValuestakenfromKurketal.(2004b). Oke,1990)wereused.Thefluxcalibrationisaccuratetoabout
g The 1338-1 and 1338-2 fields overlap and have an area of 9.3 ∼ 5%. This does not take into account the uncertainties due
arcmin2incommon.Thetotalnumberofuniquecandidateemittersin toslitlosses.Becausewecalculatedthetotalfluxoftheemis-
thetwofieldsis54inanareaof79.7arcmin2. sionlinesusingtheimagingphotometry,wedidnotattemptto
correctthespectraforfluxfallingoutsidetheslit.
The details of the reduction of the spectroscopic data is
givenin V05.In the nextsection,we will describethe results
wasmeasuredinanellipticalaperture.Acorrectionwasmade intheindividualfields.
to the “total” fluxto accountforthe flux outsidethe elliptical
aperture. More details on the object detection, completeness
3. Results
assessment, aperturesizes forthe photometryand “total”flux
correctioncanbefoundinV05. Inthissectionwedescribetheresultsoftheimaging,candidate
FollowingVenemansetal.(2002)andV05,weselectedob- selections and follow-up spectroscopy of the Lyα emitters in
jects with a rest-frame equivalent width EW > 15 Å and a eachofthenineradiogalaxyfields,followedbyadescription
0
significanceΣ≡EW /∆EW > 3asgoodcandidateLyαemit- ofthediffuseLyαhalosoftheradiogalaxies.Forthedetailsof
0 0
ters. In V05 a detailed description is presented on how EW theresultsinthe1138,0316and0924fieldswerefertoKurk
0
and ∆EW are computed from the available photometry. For et al. (2000),Pentericciet al. (2000),Venemansetal. (2004),
0
two fields that are imaged in at least two broad-band filters, Kurketal.(2004b),V05andCroftetal.(2005).Belowabrief
the0316and0052fields,theUVcontinuumslopeβ(f ∝ λβ) summaryoftheresultsinthesefieldsisgiven.
λ
ofcandidateemitterswas also computed.For theotherfields, RedshiftandLyαline propertiesweremeasuredbyfitting
a“flat”continuumslopeβ = −2wasusedtoselectthecandi- aGaussianfunctiontotheemissionline.Ifabsorptionfeatures
dates,whichisclosetothemedianβofconfirmedLyαemitters werepresent,acombinationofaGaussianandaVoigtabsorp-
in the 0316 field (β = −1.76, V05). In each field the candi- tion profile was fitted. The properties of the confirmed Lyα
date emission line galaxies were visually inspected and spu- emitters in the various fields are summarized in Tables A.1–
rious sources (like spikes of bright, saturated stars) were re- A.5. In the Tables, the objectsare orderedon increasingright
moved from the catalogues. The resulting lists should have a ascension.
very low fraction of contaminants, such as low redshift inter-
lopers(∼5%asesimatedbyV05).
3.1.BRL1602–174,z = 2.04
In Table 3 the 50% completeness limit, number of candi-
dateemittersandareaofeachobservedfieldislisted.Wealso TheradiosourceBRL1602–174wasopticallyidentifiedwith
calculated the 5σ limiting line flux and line luminosity of an a galaxywith m = 21.4objectalongthe radio axis. A spec-
R
emitterwithanegligiblecontinuum.Thenumberofcandidates trum of this galaxy yielded a redshift of 2.043±0.002 based
inthe1138fieldistakenfromKurketal.(2004b). onfouremissionlines(Bestetal., 1999).ThisplacestheLyα
6 B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies
Table4. Overviewofthespectroscopicobservationsoftheradiogalaxyfields.
Date Telescope Instrument Field Modea Grismname Dispersionb Resolutionc td
exp
2001May20&22 VLTUT2 FORS2 1338-1 MXU,maskA GRIS 600RI 1.32 260×1′.′0 31500
2001May21&22 VLTUT2 FORS2 1338-1 MXU,maskB GRIS 600RI 1.32 265×1′.′0 35100
2001Sep22 VLTUT4 FORS2 0316 MOS GRIS 1400V 0.50 130×1′.′5 12600
2001Oct18&19 VLTUT4 FORS2 2048 MXU GRIS 600B 2.40 410×1′.′0 16200
2001Oct18 VLTUT4 FORS2 0316 MXU,maskA GRIS 1400V 1.00 140×1′.′0 10800
2001Oct18–20 VLTUT4 FORS2 0316 MXU,maskB GRIS 1400V 1.00 140×1′.′0 29100
2001Nov15&16 VLTUT3 FORS1 0316 PMOS GRIS 300V 2.64 630×0′.′8 19800
2002Jan14 KeckI LRIS 0943 MSS,maskA 600/4000 1.01 310×0′.′85 10800
2002Jan15 KeckI LRIS 0943 MSS,maskB 600/4000 1.01 325×0′.′9 9000
2002Sep6&7 VLTUT4 FORS2 2009 MXU GRIS 1400V 0.62 110×0′.′75 19800
2002Sep6 VLTUT4 FORS2 0052 MXU,maskA GRIS 1400V 0.62 100×0′.′65 8400
2002Sep6&7 VLTUT4 FORS2 0052 MXU,maskB GRIS 1400V 0.62 100×0′.′65 17550
2002Sep7 VLTUT4 FORS2 0052 MXU,maskC GRIS 1400V 0.62 95×0′.′6 10800
2002Sep8 VLTUT4 FORS2 0052 MOS GRIS 1400V 0.62 95×0′.′6 9600
2003Jan31 KeckI LRIS 0943 MSS,maskC 600/4000 0.61 205×0′.′7 7200
2003Feb1 KeckI LRIS 0943 MSS,maskD 600/4000 0.61 215×0′.′75 7200
2003Feb1&4 KeckI LRIS 1338-2 MSS 400/8500 1.85 320×0′.′75 9000
2003Mar3&4 VLTUT4 FORS2 0924 MXU GRIS 600RI 1.66 225×0′.′85 20676
2003Mar4 VLTUT4 FORS2 1338-2 MXU GRIS 1200R 0.76 125×0′.′85 18600
aExplanationofthedifferentobservingmodes:
MOS:Multi-objectspectroscopymodeofFORS2,performedwith19movableslitletswithlengthsof20′′–22′′.
MXU:Multi-objectspectroscopymodeofFORS2withauser-preparedmask.
PMOS:SpectropolarimetrymodeofFORS1using9movableslitletsof20′′.
MSS:Multi-slitspectroscopymodeofLRIS,whichusesacustomlaser-cutmask.
bDispersioninÅpixel−1.
cTheresolutionisgivenforboththedispersionandspatialaxis.Theunitsare[kms−1]×[′′].
dTotalexposuretimeinseconds.
lineoftheradiogalaxyinthelowestwavelengthnarrow-band imaged again in 2001 September. The combined 14 hours of
filteravailableforFORS2,theOfilter.Thefieldwasimaged narrow-bandobservationsofthisfieldhaveadepthcomparable
for250minutesin thenarrow-bandandfor45minutesin the to that of the field surrounding MRC 1138–262 (Kurketal.,
B-band.Duetothehighextinctiontowardsthisfield(A ≈ 1, 2000, Table3).Intotal10candidateLyαemitterswerefound
B
Schlegeletal.,1998)andthelowquantumefficiencyatwave- inthisfield.Thenumberofcontaminantsisexpectedtobelow
lengths< 4000Å of the FORS2 detector,only two candidate inthisfield,becausetheonlystronglinethatfallsinthefilter
Lyαemitterswerefound.Nospectraweretakeninthisfield. is[O]λ3727ataredshiftofz<0.007.
Spectroscopicobservations
3.2.MRC2048–272,z = 2.06
MRC 2048–272 was listed in the 408 MHz Molonglo Due to geometrical constraints only three of the candidate
Reference Catalogue (Largeetal., 1981) and identified with emitterscouldbeobservedatthesametime.Additionaltargets
a z = 2.06 object by McCarthyetal. (1996). High resolution were included on the slitmask, including objects with a low
imaging in the infrared with the HST revealed three separate equivalentwidthEW <15Å.Twoofthethreecandidatesand
0
components within 3′′ (Pentericcietal., 2001), of which the theradiogalaxyshowalineina4hrspectrum(seeTableA.1
central object was identified as the radio galaxy. The surface andFig.1).Thethirdemitterhasmostlikelyanemissionline
brightness of the central object could be well fit by a de thatistoofaint(<3×10−17ergs−1cm−2)tobeconfirmedina4
Vaucouleursprofile,indicatingthatadynamicallyrelaxedstel- hrspectroscopicobservationwithFORS2atbluewavelengths
larpopulationisinplaceinthisradiogalaxy(Pentericcietal., (λ ∼ 3750 Å). The two confirmed emitters have a redshift
2001). very close to that of the radio galaxy with relative velocities
of 100 and 10 kms−1. A third Lyα emitter was found among
Imagingobservations the candidate emitters with a lower equivalent width. This
We observed the field in 2001 May under moderate seeing galaxy is a bright Lyα emitter located ∼ 4600 kms−1 away
conditions(∼1′′)for7hoursintheOnarrow-bandfilterand fromtheradiogalaxy,withtheemissionlineattheedgeofthe
for 1 hour in the B-band. Because of the low efficiency of narrow-bandfilter(Fig.1).
the detector at λ < 4000 Å and the high extinction towards
the field of A ≈ 0.4 (Schlegeletal., 1998), the field was Volumedensity
B
B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies 7
Fig.1. Spectra of the radio galaxy MRC 2048–272(left) and the three confirmed Lyα emitters near the radio galaxy (right).
Inserted in the left plot is a close-up of the Lyα line of the radio galaxy. The dotted curve represents the transmission of the
narrow-bandfilterthatwasusedtoselectthecandidateLyαemitters.Theuncertaintyinthefluxdensityisindicatedbytheerror
barintherightcornerofeachspectrum.
Kurketal. (2004b) found that the 1138 field is overdense in
Lyαemittersbyafactorof4±2comparedtothefield(volume)
density of Lyα emitters as derived by Stiavellietal. (2001).
Comparing the volume density of Lyα emitters near MRC
2048–272tothatofemittersnearMRC1138–262(Kurketal.,
2004b),thedensityinthe2048fieldisafactor3.4+1.8smaller.
−1.2
TheerrorsarebasedonPoissonstatisticsinthesmallnumbers
regime(Gehrels,1986).Usingthisfactor,thevolumedensityof
emittersnear2048is1.2+0.8timesthefielddensityofemitters.
−0.7
AdirectcomparisonwithStiavellietal.(2001)givesasimilar
density (n /n = 0.7+1.8). The small difference between
2048 field −0.6
the two density estimates (at the 0.3σ level) is due the small
numbers of galaxies involved in the comparison. The density
in the 2048field is consistentwith no overdensityof emitters
neartheradiogalaxy.
Fig.2. Spectra of low redshift emission line galaxies which
could contaminate our sample. These spectra were observed
3.3.MRC1138–262,z = 2.16
through the 1400V grism that was used for the spectroscopy
observationsofthe0052,0316and2009fields.Duetotheres-
This radio galaxy was the targetof our pilot project(see also olutionofR = 2100,interlopinggalaxiesatlow redshiftscan
Sect. 1). Narrow- and broad-band imaging with FORS1 on easilybedistinguishedfromLyαemittersatz∼3.
the VLT resulted in the detection of 37 candidate Lyα emit-
ters(Kurketal., 2000, 2004b). Subsequentspectroscopycon-
firmed15Lyαemitterstobeneartheradiogalaxyatz = 2.16 atedwiththeradiogalaxy(Kurketal.,2004a,b),increasingthe
(Pentericcietal., 2000a). The density of Lyα emitters in this numberofconfirmedprotoclustermembersto>25.
field is roughly a factor 4 higher as compared to field stud-
ies (Pentericcietal., 2000a; Kurketal., 2004b). This field is
3.4.MRC0052–241,z= 2.86
also overdense in X-ray sources (Pentericcietal., 2002) and
extremelyredobjects(EROs,Kurketal.,2004b).Intotal,four The optical counterpart of the radio source 0052–241 from
oftheX-raysourcesareconfirmedtobemembersoftheproto- the Molonglo Reference Catalogue (Largeetal., 1981) was
cluster (Pentericcietal., 2002; Croftetal., 2005). Also, nine found by McCarthyetal. (1996) to be a m = 23.2 object.
R
Hα emitters were spectroscopically confirmed to be associ- A spectrum of this object showed strong Lyα emission at a
8 B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies
Fig.3. SpectraoftheconfirmedemittersneartheradiogalaxyMRC0052–241.Thespectrumoftheradiogalaxyisshowninthe
top-rightpaneloftheFigure.Theerrorbarintherightcornerofeachspectrumindicatestheaverageerrorinthefluxdensityof
eachpixel.AllspectraareboxcaraveragedoverthreepixelsandnormalizedtothepeakoftheLyαlineatarelativefluxdensity
of0.7.Theoffsetbetweenthespectrais1.0.Thearrowindicatesthecenteroftheemissionlineofeachemitter.Theredshift,flux
andwidthoftheemissionlinescanbefoundinTableA.2.
B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies 9
Fig.4. Left: Velocity distribution of the confirmed emitters near MRC 0052–241. The median redshift (z = 2.8600) of the
emittersistakenaszero-point.Thebinsizeis150kms−1.Therelativevelocityoftheradiogalaxy(70kms−1,seeTableA.2)is
indicatedbyanarrow.Thesolidlinerepresentstheselectionfunctionofthenarrow-bandfilter,normalizedtothetotalnumber
ofconfirmedemitters. Right:Spatialdistributionofthe confirmed(circlesanddiamonds)andremainingcandidate(stars)Lyα
emitters.Thedashedquadranglerepresentstheoutlineoftheimagingarea.Theradiogalaxyisdenotedbyafilledellipse,that
depictstheapproximateshapeandpositionangleoftheradiogalaxyLyαhalo(seeTable5).Thecirclesrepresenttheemitters
witharedshiftsmallerthanthemedianredshift(z < 2.8600)andthediamondsthosewithz > 2.8600.Thesizeofthesymbols
are scaled accordingto the velocity offset fromthe median, with smaller symbolsstanding for emitters with a redshiftfarther
fromthemedian.
redshiftofz=2.86(McCarthyetal.,1996). candidates, four objects from a list with 20 candidate line
emitterswith8Å<EW < 15Åwereobserved.Twoofthem
0
Imagingandspectroscopicobservations were confirmedto be Lyα emitters atz ∼ 2.86.One of these,
Atz=2.86theLyαlineisshiftedintotheFORS2narrow-band emitter 0052.LAE36,has a Lyα line that falls at the red edge
He filter. The field was imaged for 390 min in this narrow- ofthefilter.Usingthemeasuredredshift(z = 2.876insteadof
band and for 80, 90 and 80 min in the B-band, V-band and z = 2.86)thecalculatedequivalentwidthisEW0 ∼ 24Å.The
I-band respectively (Table 2). Analysis of these data resulted spectraofthe37confirmedLyαemittersandthatoftheradio
in a list of 57 candidate Lyα emitters with EW > 15 Å and galaxyareshowninFig.3.
0
EW /∆EW >3(Sect.2.3).
0 0
Volumedensity
Follow-up spectroscopy of candidate emitters was carried
out in 2002 September at the VLT. 36 candidates were ob- The density of Lyα emitters can be compared directly to the
served in four masks with individualexposuretimes between field density of emitters. Fynboetal. (2003) used the same
140 and 292.5 min under good seeing conditions (0′.′6–0′.′65, narrow-band filter at the VLT to observe a field which con-
see Table 4). The resolution of the grism that we used (the tainsadampedLyαabsorber.Thedepththeyreachintheirim-
1400Vgrism)isR=2100,whichishighenoughtoresolvethe agesisapproximately0.7magdeepercomparedtoournarrow-
[O] λλ3726,3729and C λλ1548,1551doublets (see Fig. bandimage,makingtheirobservationssuitableforcomparison.
2 for a few examples). Objects with these emission lines in Fynboetal. (2003) find 14 emitters(of which 13are spectro-
theirspectrumarethemaincontaminantsinsearchesforz∼3 scopicallyconfirmed)withaLyαflux>8×10−18ergs−1cm−2
Lyαemitters(e.g.,Fynboetal., 2003).Thehighresolutionof and a rest-frame equivalent width > 15 Å. To the same limit
the 1400V grism allows us to discriminate high redshift Lyα andinthesamefieldofview,wefind48(candidate)emitters.
emittersfromlowredshiftinterlopers. Thisimpliesthattheoverdensityinthe0052fieldis3.4+1.5.
−1.0
Duetocosmicvariance,theuncertaintyinthenumberden-
Results sity of Lyα emitters in a small field of view like that of the
Of the 36 objects observed 35 were confirmed to be Lyα VLT images is generally higher than the uncertainty derived
emitters at a redshift z ∼ 2.86. The 36th candidate was most assuming Poisson statistics (e.g., Somervilleetal., 2004). To
likelytoofainttobedetected.Thiscandidatehadthesmallest overcome this, a comparison can be made with the density
linefluxoftheobjectsinthemask.Inadditiontothe36good of Lyα emitters at z ∼ 3.1 in a 0.13 deg2 field as found by
10 B.P.Venemansetal.:Protoclustersassociatedwithz>2radiogalaxies
Fig.5. LimitsontheobservedfluxratiosforthesingleemissionlinesourcesnearMRC0943–242(squares).Alllimitsare2σ
upperlimits.Lineratiosoflowredshift[O]emittersfromTerlevichetal.(1991)areplottedaspluses.Themeasuredlimitson
thelineratiosareconsistentwiththeidentificationoftheemissionlineswithLyα(seetext).
Ciardulloetal. (2002). This comparison gives a volume den- Instead,themajorityoftheemitters(31outofthe37)appearto
sityofLyαemittersinthe0052fieldof2.5+1.8 timesthefield beclusteredintwogroupswithvelocitydispersionsof180and
−1.1
density. The large (Poisson) errors are due to the small num- 230 kms−1. Interestingly, a similar redshift distribution was
berstatistics.Morerecently,Hayashinoetal.(2004)measured foundneartheradiogalaxyMRC1138–262(Pentericcietal.,
the blankfield space densityof Lyαemittersat z ∼ 3.1 using 2000a).Thevelocityoftheradiogalaxylies∼70kms−1 from
theSuprime-CamontheSubarutelescope.Ina0.17deg2field, themedianredshiftoftheemittersandfallsinsidethelargerof
they find 55 candidateLyαemitters with an observedequiva- the two groups.The combinationof the observedoverdensity
lentwidth>154Ådowntoanarrow-bandmagnitudeof25.3. ofLyαemitterswiththeclumpyredshiftdistributionprovides
Applyingthesameequivalentwidthandmagnitudecriteriato evidencethattheLyαemittersnearMRC0052–241arepartof
our field and taking into account the difference in luminosity aformingclusterofgalaxiesatz=2.86.Thepropertiesofthis
distancegives11sources.Theresultingdensityofemittersnear structurewillbediscussedinSect.5.
MRC0052–241isafactor3.1+1.4higherthanthefielddensity. Incontrastto the velocitydistribution,the emittersdo not
−1.0
Theweightedaverageofthethreeestimatesoftheoverdensity haveapreferredlocationonthesky.Thestructureofemitters
is3.0+0.9. appears not to be bounded by the image, indicating that the
−0.6
It is interesting to compare the density of Lyα emitters overdensityseenin thisfield extendsbeyondtheedgesofour
in the 0052 field with the density of emitters near the radio fieldofview.
galaxy MRC 0316–257 at z = 3.13 (V05). The depth and
sensitivityoftheimagingofthetwofieldsareverysimilar(see
3.5.MRC0943–242,z= 2.92
Tables 2 and 3). There are 52 candidate emitters in the 0052
fieldwithanarrow-bandmagnitudebrighterthan26.1,against This powerful radio source with a flux density of 1.1 Jy at
59in the 0316field. Takinginto accountthe differencein the 408MHz hasa redshiftof z = 2.923(Ro¨ttgeringetal., 1995;
volume probed by the narrow-band images, the ratio of the Ro¨ttgeringetal., 1997). The radio galaxy is surrounded by
numberdensityisn0052/n0316 =0.8+−00..22.Becausethe0316field a metal-enriched, low surface brightness gas halo extending
is overdense in Lyα emitters by a factor 3.3+0.5 (V05), this for at least 8′′ (67 kpc, Villar-Mart´ınetal., 2003). The esti-
−0.4
means that the 0052 field has an overdensityof Lyα emitting mated dynamical mass of the gas halo is 7 − 44 × 1011 M⊙
galaxies of 2.7+0.8, which is consistent within the errors with (Villar-Mart´ınetal., 2003; Jarvisetal., 2003). This could be
−0.6
theotherestimates. anindicationthatwearewitnessingtheformationofamassive
galaxy.
Velocityandspatialdistributions
Thevelocityhistogramandthespatialdistributionofthecan- Imagingandspectroscopicobservations
didate and confirmed emitters are shown in Fig. 4. The red- The field surroundingthe radiogalaxywas observedwith the
shifts of the confirmed emitters are not uniformly distributed VLTin2001March.Imagesweretakeninthenarrow-bandfor
throughout the narrow-band filter (solid curve in Fig. 4). 375minand75minintheB-band(Table2).Apartofthefield