We present the catalogue of a wide-field CFHT/WHT spectroscopic survey of the lensing cluster Cl0024+1654 at z=0.395. This catalogue contains 618 new spectra, of which 581 have identified redshifts. Adding redshifts available from the literature, the final

O.Czoskeetal.:SpectroscopicsurveyofCl0024:I.Thecatalogue5

essentialinordertoderivespectraltypesforthegalaxiesandanalyseindetailthespectralpropertiesoftheclustermembers.Forthesereasonswedidnotuseband-limiting lters andtheusablewavelengthrangewastypically4500–8500A,dependingonsignal-to-noiseratioandthequalityoftheskysubtractionattheredendofthespectralrange,wheretheskyemissionisdominatedbymolecularbands.TheobservationsatWHTweremadewiththeLowDispersionSurveySpectrograph(LDSS-2,Allington-Smithetal.1994),themed/bluegrismandtheLoralLOR1detector.Theusablewavelengthrangewas4000–7500 A,somewhatbluerthanfortheMOSobservations.AlltheLDSS-2maskswerecoveredbytwoexposureseach.Weused1′′wideslitsthroughout,resultinginares-olutionof～13 A,exceptforrun1wheretheslitwidth

usedwas1′′

.5withacorrespondinglyworseresolutionof～20 A.

3.Datareductionandanalysis3.1.Reductionofspectroscopicdata

Thespectrawerereducedusingthesemiautomaticpack-agemultired(LeF`evreetal.1995)whichinturnusesstandardIraftasksandtreatseachslitseparately.Thespectralimageswerede-biasedand at- eldcorrectedinthestandardway.Alow-order(mostlylinear)polynomial ttotheskyemissioninthespatialdirectionwasthensubtractedforeachindividualexposure.Duringthe tmostofthecosmicrayhitsintheskyregionofthespec-tralimagesweretakencareofbyasigma-rejectional-gorithm(pixelswithdeviationsofmorethan±2σfromthe twererejectedbeforere tting);neverthelessinsuf- cientlyrejectedcosmicsoccasionallysurvivedthe ttoproducefakeabsorptionfeaturesintheobjectspectrum.Geometricdistortionscausethetruespatial/dispersiondirectionstodeviatefromtherow/columndirectionsoftheCCDchip,especiallyattheedgesofthe eld.Sincesky ttingwasdonerow-wise(column-wisefortheWHTdata),skyemissionlinestendtobeimperfectlysub-tracted.Isolatedemissionlines(inparticular[Oi]λ5577,whichroughlycoincideswiththe4000 AbreakforgalaxiesattheredshiftofCl0024+1654)couldsimplybemaskedforthesubsequentanalysisofthespectrum,buttheus-ablewavelengthrangeintheredwase ectivelylimitedbythemolecularbandemissionfromthesky.

Theindividualexposureswerethenaveragedintothe naltwo-dimensionalspectrum.Inthosecaseswheremorethantwoexposurespermaskwereavailable,thehighestpixelvaluewasrejected,thusaccountingforcosmicrayhits.Thesmallnumberofexposurespermaskmadecos-micrayrejectiondi cultandimperfect.Acosmicrayhitontheobjectspectrumresultedinafakeemissionfeature,ahitintheskyareainafakeabsorptionfea-ture.However,asarguedinSection3.2,cosmicsdonotingeneralin uencetheredshiftdetermination.Finally,variance-weightedone-dimensionalspectrawereextractedfromthecombinedspectralimages,wavelengthcalibrated

usinglampspectra(He/AratCFHT,Cu/AratWHT)

andapproximately ux-calibratedwithlong-slitspectrafromspectro-photometricstandardstars(listedinTable1).

Thewavelengthcalibrationspectrawereextractedus-ingastraighttrace(followingthecolumn/rowdirectionoftheCCD),unlikethecorrespondingobjectspectra,wherethetracefollowedthe exureintroducedbygeometriccameradistortions.Weveri edthatthiscausesonlyneg-ligibleerrorsbyre-calibratingseveralspectrawithstrong exurewithcalibrationspectraextractedusingthesametraceasfortheobjectspectra.

We nallyperformedacheckonthewavelengthcali-brationbyextractingthreespectrapermaskwithoutskysubtractionandmeasuringthepositionofprominentskylinesusingthecalibrationfromthelamps.FortheWHTmasksthisrevealedsystematicshifts,whichareduetothefactthatthelampspectrawerenottakenimmediatelybe-foreorafterthescienceexposures.Thepositioningofthemasksinthemaskholderwasthereforenotidenticaldur-inglampandscienceexposures.Thesesystematicshiftstranslateto～10 3intermsofredshiftandacorrectionforeachmaskwasappliedtoalltheredshiftsdeterminedfromthismask.Nosuchsystematice ectwasfoundfortheCFHTmasks.

SevenexamplespectraandtheV-bandimagesforthecorrespondingobjectsareshowninFig.2.

3.2.Redshiftdetermination

Sincealargefractionofthespectrainoursamplehaveratherlowsignal-to-noiseratio,wedecidedtoidentifyred-shiftsbyeye,whichisbetterat ndingrealabsorptionandemissionlinesamongstnoisethanautomaticredshiftiden-ti cationtechniques.Also,fakeemissionandabsorptionfeaturesduetocosmicscouldbeidenti edthiswaybysim-plycheckingtheoriginaltwo-dimensionalspectralimages.Allredshiftswereidenti edbyoneofus(OCforruns2-4,GSforrun1)andcheckedbyatleastoneothermemberoftheteam.Theredshifts(runs2-4only)giveninTable4carrya agwhichindicatesa(somewhatsubjective)levelofcon dencethattheredshiftidenti cationiscor-rect.“Secure/A”redshiftsweredeterminedfromspectrawhereemissionlinesand/orseveralabsorptionlinesareclearlyseen,“uncertain/D”redshiftsarebasedonaten-tativeidenti cationofasinglelineorpossiblyseveralweakabsorptionlines,and“probable/B”and“possible/C”in-dicateintermediatelevelsofcon dence.

Inordertoobtainamoreobjectiveestimateofthemeasurement(asopposedtoidenti cation)errorofourredshifts,weusedthecross-correlationtechniqueimple-mentedinthetaskxcsaointheIrafpackagervsao(Kurtz&Mink1998).Inallcases,theby-eyeredshiftidenti cationwasfedtoxcsaoastheinitialredshiftes-timateandtheallowedredshiftrangewasrestrictedtoabout±5%aroundthisredshift.Inthissense,xcsaowas