197 4Ap J . . .194 10 The AstrophysicalJournal,194:1-19,1974November15 © 1974.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedinU.S.A. present characteristicsandpastevolutionofclusters the propertiesofalargegroupclusterswidely inadequate forthispurpose,astudyhasbeenmadeof properties fromthosethatarepeculiartoasingle number ofobjectstopermitseparatingtheirsystematic varying formandrichness. cluster. Sincetheobservationsinliteratureare of galaxies,weneedinformationonasufficientlylarge parts. Thisfirstpaperdescribesobservationsofthe luminosities, andmorphologicaltypesoftheirmem- bers—as obtainedfromwide-fieldphotographicplates. that thedifferencesbetweenthesegroupscanbeex- homogeneous groups.Argumentswillbepresented which permitthemtobedividedintoafewfairly between thevariousobservedpropertiesofclusters matically findingandmeasuringgalaxiesonplates. Included isthedescriptionofaprocedureforauto- static propertiesofclusters—namely,thepositions, plained bydifferentmeandensitiesintheoriginal protoclusters, whichresultedindifferentratesof evolution. galaxy formationanddifferentratesofdynamical ratios. clusters, andwiththeirinferredmass-luminosity (1958) catalogofrichclusters. Withinthiscatalog, If wewishtoobtainsomeunderstandingofthe The resultsofthisstudywillbereportedintwo It willbeshownthattherearestrongcorrelations Paper IIwilldealwiththedynamicalpropertiesof The objectstobestudiedwere takenfromAbell’s © American Astronomical Society • Provided by theNASA Astrophysics Data System THE SYSTEMATICPROPERTIESOFCLUSTERSGALAXIES.I.PHOTOMETRY Subject heading:galaxies,clustersof three groups.Spiral-richclustershavecompositionssimilartothatofthefield,irregularmassdistributions ing galaxiesonphotographicplates.Itwasfoundthattheirpropertiesaresuchtheymaybedividedinto morphological type.ClusterswithcDgalaxiesarerichinellipticals,havesmooth,sphericalmassdistributions clusters. tendency towardenergyequipartition.Spiral-poorclusters,whicharedominatedbySOgalaxies,alsoshow low densityandnocentralconcentration,showsignsofsegregationmembersaccordingtomassor segregation bymassandmorphologicaltype,butarenotassmooth,dense,orcentrallycondensedthecD of highdensityandcentralconcentration,haveastrongdeficiencyspiralsinthecore,showconsiderable local minimumintheradialdensitydistribution,effectbeingstrongestcDclusters.Themean tionary stateofspiral-richclustersbutthatcDcluste dispersion inabsolutemagnitude.Argumentsarepresentedthatspiral-poorclustersrepresentalaterevolu- of clustersalltypesisindependentmass.Theluminosityfunctionsaresimilar,withasmall Fifteen richclustersofgalaxieshavebeenstudiedusinganewautomaticmethodforfindingandphotometer- The outerenvelopesofallclustershaveaprojecteddensitygradientslopenear3.0.Allshow Hale Observatories,CaliforniaInstituteofTechnology,CarnegieInstitutionWashington a) ChoiceofObjects I. INTRODUCTION II. OBSERVATIONS Received 1974March15;revisedJune19 Augustus Oemler,Jr. OF 15CLUSTERS ABSTRACT while themostevolved,areintrinsicallydifferent. those objectswereselectedwhich,ofeachrichness class, werethenearestexamplesnotatexcessivelylow presence ofmanyotherclusters.Exceptforitsrichness galactic latitude,andnotinfieldsconfusedbythe plate qualityorthelimitationsoftime.Also,two considered inmakingthischoice. class, thecharacteristicsofclusteritselfwerenot the literatureonitsinternalvelocitydispersion,and clusters werelateradded:A2197becauseofdatain Zw Cl1545.1+2104becauseithadalreadybeen biased sample,thefinalgroupof15clustersistoo the firstfourcolumnsoftable1.Unlessotherwise class, redshift,andgalacticlatitudeofeachobject,in characteristics withinAbell’scatalog. small andtoofortuitousaselectiontopermitconclu- studied forotherreasons(Oemler,Gunn,andOke noted, theredshiftshavebeentakenfromNoonan sions aboutthefrequenciesofclusterswithvarious using eitherthePalomar200-inch(5-m)or48-inch (1973). made onbakedIllaJplatesbehindaWratten4filter. hypersensitized 098emulsions, allbehind2.2mmof 098, andIIIaFemulsions atthe200-inchonwater- 1972). Therefore,whiletheoriginallistwasanun- (1.2-m) Schmidttelescope.Allgreenexposureswere James Gunn,andtwoofthe 48-inchplatesweretaken Red exposuresonthe48-inchweremade103aF, by StevenShectman.The200-inch plateswere5x7 Schott RG-1filter.The200-inch platesweretakenby Some ofthisgroupwerenotstudiedbecausepoor The finallistispresented,alongwiththerichness Green andredplateswereobtainedofeachcluster, b) PhotometricObservations 197 4Ap J . . .194 10 2 48-inch plateswereeither5x7inches(2?0x2?1), inches insize,givingausablefieldof2Tx23'.The All 10xinchand14plateshad calibrated platefromthesameboxwasdevelopedwith them unsuitableforphotometry,manyplatescould veloped for9minutesinMWP-2developer,usinga each batchof5x7inchplates.Allplateswerede- Sensitometer spotsplacedononecorner,and 6?2), dependingontheangularextentofcluster. the platethatwasusedispresented,foreachcluster, to measureallclustersinonecolor.Theemulsionof rocker agitatortoensureevendevelopment. the 48-inchtelescope,exceptforthoseofA2670,which in thelastcolumnoftable1.Allplatesweretakenon not beused.Asaconsequenceofthis,itwasimpossible came frombothtelescopes,andthoseofZwCl 10 xinches(4?34?3),or14(6?2 large aperture,typically40"indiameterfornearby from photoelectricphotometryofseveralgalaxiesin each field.Eachgalaxywasphotometeredthroughone /, issomewhattotheblueofF-band;andred eter mountedonthePalomar60-inchtelescope.Filters channel spectrophotometeroratwo-tubeS20photom- / —F=0.0andJ 0.0 forthephotographic photographic bands.Filtersusedwere,inthegreen, closely aspossiblethespectralresponseoftwo for the60-inchphotometerwerechosentomatchas objects. Thiswasdonewitheitherthe200-inchmulti- Wratten 44;and,inthered,3-mmSchottRG1, band, denotedbyF,issimilartoSandageandSmith’s 3-mm SchottBG38,1-mmGG460anda 1545.1+2104. photometry ofaLyr(Oke and Schild1970)through broadband photometryare tiedtotheabsolute end. Magnitudesaredefined sothat,foranA0star, are presentedinfigure1.Thegreenband,denotedby Schott BG20,and1-mmBG38. bands. Boththe200-inchmultichannel and60-inch (1963) Rband,butismoresharplycutoffatthered Because ofseriousemulsiondefectswhichrendered Absolute calibrationoftheseplateswasobtained The photographicandphotoelectricresponsecurves © American Astronomical Society • Provided by theNASA Astrophysics Data System Abell 400. Abell 194 Abell 1228 Abell 539 Abell 1367 Abell 1314... Abell 665 Abell 1413 Abell 2197 Abell 1904 Coma cluster Zw Cl1545.1+2104 Abell 2670 Abell 2199 Hercules cluster.... * Sargent(1973).tOemler Cluster Richnessz(degrees)Emulsion AUGUSTUS OEMLER,JR. Clusters Studied 0 0.0181 0 0.0335 2 0.0205 2 0.0230 5 0.183* 2 0.0360 2 0.0719 1 0.0267 1 0.0231 2 0.0312 3 0.1427 3 0.0753t 1 0.0344 1 0.0303 1 0.263§ TABLE 1 the multichannelspectrophotometersecondarystand- were constructedfromthemultichannelscansusing photographic responsefunctions;butbecauseofthe ards, /andFmagnitudesofstandardstarsandgalaxies the errorsinphotoelectriccalibrationsduetothis § Oemleretal.(1972). fairly narrowcolorrangeofthecalibrationobjects, that, forgalaxieswithlate-typespectra,itisapproxi- in thisstudy.Thephotoelectricphotometryshowed are notsignificantforphotometryoftheaccuracyused the field.ComparisonofZwicky’sphotometryinother mately truethat A1413. Platesofthisclusterwerecalibratedbyusing and Zwicky etal.(1961)photometryofbrightgalaxiesin for m<15.0,themagnitudesoflate-spectral-type cluster fieldswiththatdoneforthisstudyshowedthat, galaxies wererelatedby p (solid curves)responsefunctions fortheJandFbands. There werenophotoelectricdataononecluster, Fig. 1.—Photoelectric{dashed curves) andphotographic -17 IHaJ -43 IHaJ + 35IHaJ -45 103aF + 64IHaJ + 70103aF + 77IIIaF + 74IHaJ + 87098 + 44103aF + 62103aF -69 098 + 43103aF + 43IHaJ + 49098 m =+1.5±0.2.(3) / =F+0.35(F-F)(1) pF /- F=1.1±0.10.(2) Vol. 194 197 4Ap J . . .194 10 photographic plateswereusedinthisstudy.Thefirst, images onanegativecopyoftheplate,hasbeen which involvesmicrophotometertracingsofgalaxy No. 1,1974 very largeamountofphotometrywhichhadtobe method wasusedonA1314,A539,A2670,andZwCl described indetailelsewhere(Oemler1973).This two-axis microphotometerattachedtoasmallcom- procedure forbothfindingandmeasuringgalaxies,is done. Thesecondmethod,acompletelyautomated based ontheCaltechplatescanner.Thismachineisa puter. Thecomputercontrolscarriagemotion,pro- the firstpart,areaofplatetobestudiedis writes dataonmagnetictape. cesses thedigitizedoutputofphotometer,and 1545.1 +2104,butprovedtootimeconsumingforthe 256 overlappingrows.Onalargecomputer,thedata boxes perrow.Byoverlappingineachdirection,we scanned withalong,narrowslitinrasterpatternof ensure thatnogalaxyimagewillbemissedbecauseit along eachrowareaddedtoform1024overlapping the fieldduetovariationsinplatesensitivityandfog. fell halfinoneboxandanother.Oneventhe best ofplates,theskybackgrounddensityvariesover procedure ofignoringthedenserthanaveragepoints The variationinskydensityisfoundbyaniterative dimensional least-squarespolynomialfit. and fittingtheremaindertoalow-order(35)two- higher opacitythantheleast-squaresfittosky.All sented byagroupofpointsinthe256x1024array considered asobjectstobephotometered,andtheir plate scanner.Thepeakopacityofeachchosenpoint positions areaddedtoalistonmagnetictapeforthe of thosegroupswhosepeakopacitydiffersfromthat of theskybymorethansomethresholdvalueare plate scannermovesinturntoeachpositiononthelist and theskyfitaresavedforlateruse. centered ontheobject.Theaperturesizeanddigitizing resolution elementontheplate.Amongobjects rate arechosensothateachdatapointcoversone and performsarasterscanofsmallsquarearea image areaforagroupofobjects,thereisclear two groups,wemakeuseofthefactthatimages four timesasmanystarsgalaxies.Toseparatethe tude whereitisbecomingdifficulttodistinguishthe galaxies. Whenthemeanopacityisplottedagainst which havebeenphotometered,thereare,typically, of starsarehighersurfacebrightnessthanthose two groupsevenbyeye.Byusingaconservativecriter- ion fortheboundarybetweentwogroups,virtually separation ofstarsfromgalaxiesdowntothatmagni- line areinspectedbyeyeto remove anystars. doubtedly somediscrimination againstverycompact galaxies. Asmallnumberof objectsnearthedividing all normalgalaxiesarefound,althoughthereisun- galaxy magnitudesisastraightforward procedure.All Two methodsforthephotometryofgalaxieson The galaxymeasurementprocessisintwoparts.In Each starandgalaxyinthemeasuredfieldisrepre- In thesecondpartofmeasurementprocess, All starshavingbeeneliminated, obtainingthe © American Astronomical Society • Provided by theNASA Astrophysics Data System c) PlateMeasurementMethods SYSTEMATIC PROPERTIESOFCLUSTERSGALAXIES _1- tained fromtheSensitometerspotsoneachplate,then the skybackgroundisremovedbyusingleast- points ineachrasterscanareconvertedtorelative surface brightnessesusingthecharacteristiccurveob- squares fitobtainedearlier.Therelativebrightnesses we simplyaddtheintensitiesofallpointswhosesurface Now, toobtaintheisophotalluminosityofanobject, galaxies forwhichwehavephotoelectricphotometry. are convertedtoabsolutevaluesbyusingthose brightnesses aregreaterthanthatofthelimiting that theselectioncriterionforobjectswaspeak isophote. ness ofthisgroupobjects.Firstthegalaxiesare interval, weplotahistogramofthenumberobjects divided intomagnitudeintervals.Itwillberecalled from thesearchroutine.Theresultisaseriesofnear- measured opacityoftheirimages.Foreachmagnitude more andtruncatedbytheselectionlimit.As- of eachpeakopacity,thisquantityhavingbeensaved calculate, foreachcase,thefractionofobjectswhich fall belowthethreshold,andthusfindincomplete- suming thedistributionstobesymmetrical,wecan Gaussian curveswhich,withincreasingmagnitude,are ness asafunctionofmagnitude. potential ofhighaccuracy.Anotheristhatitcom- galaxies hasseveralgreatvirtues.Oneisthat,by pletely automatic,requiringinterventiononlyto measuring allpartsofthegalaxyimage,ithas below. Its practicalefficiencyandaccuracywillbediscussed specify thecriterionforseparatingstarsfromgalaxies. corrections forfieldobjectslimitthedepthtowhicha given fieldcanbestudied.Inordertocoverasufficient have beenmeasuredintwoparts.First,alargefieldis area andasufficientmagnituderange,mostclusters to obtainthefaintendofclusterluminosityfunc- measured overarestrictedmagnitudeintervalto tion, asmallfieldintheclustercenterismeasuredto survey thetotalextentofaclusterandobtainsome measure ofthegalaxybackgroundinitsvicinity.Then, much fainterlimit. radius (Abell1958)definedby typical cluster,about80percentofthetotalmassis deceleration parameter,<7,of0.0,thisisequivalent, If weassume,asshallthroughoutthispaper,a for nearbyobjects,toalinearsizeof3.0Mpc.For Hubble constant,of50kmsMpcanda contained withintheprojectedAbellradius. within 1Abellradiushave beenstudied,although an areaoutto2or3Abellradii. Usually,faintgalaxies of galaxiesinthebrightest3-mag intervalhascovered occasionally asomewhatsmaller areawascovered. 0 The finalstepistodeterminethelimitofcomplete- This entireprocedureforfindingandmeasuring Both thelimitationsoftimeanduncertaintiesin A convenientmeasureofclusterscaleistheAbell For mostoftheclustersinthisstudy,photometry r =5.15x105/czarcmin.(4) d) MeasurementofClusters a 3 197 4Ap J . . .194 10 the photographicphotometry. the galaxies,withabrightnessrangewithineach in thered.Thisprovedtobefaintestsurface limiting isophoteof24.1magpersquaresecondinthe the entiregalaxyouttodesiredlimitingisophote, true, intherelevantmagnituderange,forallclusters. legitimate onlyifthereisnomasssegregationdueto 4 cluster fieldsweremeasuredphotoelectrically.When problems duetoplatenoiseandincorrectsubtraction two-body relaxation.Weshallshowlaterthatthisis The procedureofinferringthefaintgalaxycontent teristic curvesoftheplates.Therandomdifferences due tothecombinederrorsinphotoelectricand in thecalibrationofindividualobjectswas0.10mag, clusters inthenormalmanner,standarddeviation cluster of2mag,wereusedtocalibrateplatesthese been checkedinseveralways.Thirty-sixgalaxies10 brightness whichcouldbeusedwithoutencountering green and,itsequivalent,23.0magpersquaresecond for thefaintergalaxiesinmostdistantclusters. brighter galaxiesinthenearestclustersto10"square area. Scanareasusedvariedfrom2'squareforthe chance ofincludinganotherobjectinthemeasured raster scanwaschosentobelargeenoughcontain etry methodusesanegativeoftheplateonwhich with theplatescannerduetoinaccuratemeasurements us thattherearenosystematicerrorsinphotometry deviation of0.13mag,mostwhichcanprobablybe magnitude scales,dueprobablytoerrorsinthecharac- systematic differenceof0.03magpermagnitudeinthe of theskybrightness. but notsolargethattherewouldbeasignificant of theentireclusterfromastudycenteris limit ofcompleteness.The IllaJ plateofthecluster implies a=0.18forthefirstmethod. which, assuminga=0.09forthesecondmethod, part. Therandomerrorsbetweenthetwosetsof galaxy centersaretheleastdense,mosteasilymeasured below thelevelofrandomerrors.Thisreassures systematic differenceinthemagnitudescaleswas been studiedusingthefirstphotometrymethod.Any scanner, ontheplateofA1314whichhadpreviously attributed tothepoorplate. between thetwosetsofmeasurementshadastandard There wasnocalibrationdifference.a finding chartwasprepared of allgalaxiesabovethe calibration errorsa=0.14,random0.18; measurements hadastandarddeviationof0.21mag, of thedensecentersgalaxies,sincefirstphotom- other ofratherpoorquality,theComacluster. separately calibratedplates,oneofverygoodandthe was theninspected,using this charttoseeifany random errorsa=0.09. for thesecondmethod,calibrationerrorsa=0.07, methods canbesummarizedthus:forthefirstmethod, The accuracyofthetwomethodsphotometryhas All galaxyphotometryinthisstudywasdonetoa In measuringindividualgalaxies,thesizeof Twenty-seven objectsweremeasured,withtheplate After photometryofeachcluster wascompleted,a The averagereliabilityofphotometrybythetwo Forty-eight objectsweremeasuredoneachoftwo © American Astronomical Society • Provided by theNASA Astrophysics Data System AUGUSTUS OEMLER,JR. with thoseofthesameobjectsbyRoodandBaum properties ofthegalaxies,obtainedfromthin,large- tunately, theclassificationofobjectsatdistancesa ticals, SO’s,orspirals(includingirregulars).Unfor- the sameIllaJplate.Objectswereclassifiedasellip- the screeningprocesses.Theseomissionsandinclusions the entirefield. plates showsthat,whileEandSOgalaxiesweretyped classifications ofobjectsintheComaclusterandA194 distinguish SOandspiralgalaxies.Acomparisonof usual morphologicaltypesarebasedonthenuclear is classifyingtheouterfeaturesofgalaxies,whereas few hundredmegaparsecsondeep,small-scalephoto- not onecluster.Theplatestudiedwasoffairquality. Abell’s catalog.Althoughthereissometendencyto chain. Theplateusedwasofonlyfairquality. classify manyearly-typespiralsasSO’s. fairly reliably,therewasasystematictendencyto (1967) andZwickyHumason(1964)using200-inch scures manyfeatures,makingitespeciallydifficultto scale plates.Perhapsworse,thepoorresolutionob- graphic platesisnotverysatisfactory.Necessarily,one each fieldwereclassifiedbymorphologicaltype,using were corrected. ticals. Thereisnoonegalaxyofoutstandingbright- subclustering, thereisnostrongevidencethatthis not seemtobeobscured.A5x7inchplateofgood center. Thisclusterisatlowgalacticlatitude;there 400. Thereisashortchainofbrightgalaxiesinthe plate studiedwasofonlyfairquality. most ofthebrightgalaxiesareellipticalsandSO’s. brightest galaxies,mostofwhichareSO’s,formalong similar numberofstarsusuallymanagedtoslipthrough ness limithadbeenmissedforvariousreasons.A galaxies weremissedorstarsincludedinthefinallist cD galaxywasdescribedby MorganandLesh(1965) compact andregularinappearance,withseveralbright ness. Theplatestudiedwasofgoodquality. spiral galaxies.Theplatestudiedwasofgoodquality. quality wasstudiedtoonelimitingmagnitudeover much nearbyobscuration,buttheclusteritselfdoes sitating largeextrapolationsinthelateranalysis.The Only theinnerpartofclusterwasstudied,neces- (1964). Itisamoderatelyregularcluster,inwhichthe of thegalaxiesinfieldbrighterthancomplete- of objects.Ingeneral,itwasfoundthat1-2percent plate ofratherpoorquality. IIIaF plateusedwasdeepbut ofpoorquality. good quality. as possiblythelargestofall supergiantgalaxies.The E’s andSO’sinthecenter.Theplateusedwasofvery Abell 1228isanirregularclusterdominatedby Abell 194wasstudiedbyZwickyandHumason Abell 665istheonlyclusterofrichnessclass5in Abell 539issimilarinformandcontentto Abell 400isamoderatelyregularclusterinwhich Abell 1367hasahighspiralcontentbutisfairly Abell 1314isaregularclusterdominatedbyellip- Details ofindividualclustersaresummarizedbelow. For nearbyclusters(z<0.04),allofthegalaxiesin Abell 1413isarich,regularcluster.Thehugecentral The Comaclusterwasstudied onone14xinch Vol. 194 197 4Ap J . . .194 10 -1 two clustersshowsthatonlytheoutermostisopleths is noevidencethatthesetwoclustersarenot,atpresent, A2197 andA2199havethesameredshifta photometered toauniformdepthovertheentire for details. though notasrich,toA2670.SeeOemleretal.(1972) velocity ofapproach1000kms,thecollapsetime projected separationofonly4.5Mpc.Assuminga galaxies (Matthews,Morgan,andSchmidt1964). central galaxy,NGC6166,isatypeexampleofcD 2147 islocated2°tothesouth.Agood-qualityplate by spiralgalaxies.ThesmallcompanionclusterAbell field. cD galaxyofmoderatesize.Agood-qualityplatewas the galaxiesintheirvicinity,shouldbeconsidered in thelateranalysistheywillbetreatedasindependent density ofbrightgalaxiesintheneighborhoodthese independent systems.Acontourmapofthenumber of thissystemisonly4.5billionyears.However,there was used. the clusteringscaleofobjectsataparticulardistance there aresomeproblemswithdecidingwhich,evenof systems. clusters’ innerisoplethsduetointeraction.Therefore, fluctuations, pluspracticallimitationsonthearea number densityofbackgroundgalaxiesarestrongon galaxies whichsurroundsthem.Sincemanyclusters difficulties ofseparatingthemfromthegeneralfield (1973). quality. No. 1,1974SYSTEMATICPROPERTIES tics, thesetwodistributions could themselvesbecom- which canbestudied,makeitverydifficulttoaccur- cluster members.However,themostserioustroubles about thepropertiesofclustersgalaxiesisdueto supergiant galaxy.ItisdiscussedindetailOemler been combinedintotwomeandistributions.Itwas ately inferthecontributionofbackgroundobjectsto all scales,butparticularlyatthosecorrespondingto are duetobackgroundgalaxies.Fluctuationsinthe are probablystillundergoingcollapseand/orinfall, overlap andthatthereisnodistortionofthetwo found that,tothelimitsofaccuracysetbystatis- able, allcountsoffieldgalaxiesineachcolorhave cluster. a clusterofgalaxiesfromcountsinthevicinity (^apparent magnitude)(Shectman1973,1974).These number versusmagnituderelation. Alloftheavailable bined, withacolorshiftof /—F=0.90,intoone 1?5 northofAbell2199.Theplateusedwasgood data arepresentedinfigure 2. Whilemostofthedata Abell 1904isarich,regularclusterwithcentral Abell 2199isacompact,regularcluster.Thelarge The Herculesclusterisanirregulardominated Abell 2670isacompact,regularclusterwithcentral Abell 2197isasomewhatirregularclusterlocated Zw Cl1545.1+2104issimilarinappearance,al- Much oftheuncertainty,reflectedinliterature, In ordertomakethemostofinformationavail- © American Astronomical Society • Provided by theNASA Astrophysics Data System a) ClusterMembership III. DATAANALYSIS OF CLUSTERSGALAXIES uncertainty inthebackgroundlevelis±50percent. to fieldbutalsofrompointwithinonefield. the brightendwithcountsinafewfieldsbyZwicky to beconsistentwithaconstantspacedensity. with magnitudeofthebrightgalaxiesdoesnotseem log A=C+0.6mlawoveronlyalimitedmagnitude in figure2istheadoptedmeanrelationusedforall will beassumedthattheseproportionsarecorrectat range. Thefalloffatthefaintendisexpecteddueto In allsubsequentanalysis,wewillconsiderthatthe clusters. Thescatterisquitelarge,notonlyfromfield photometric systemsinequation(3).Thedottedline from photographicphotometry.Asemiempiricalpro- classified bymorphologicaltype,thereare64percent etered forthisstudy,theyhavebeensupplementedat come fromtheouterpartsofclusterfieldsphotom- assume thattheluminosity profilesofallelliptical been describedindetailelsewhere (Oemler1973).We cedure forobtainingtherequiredextrapolationshas extrapolate beyondtheisophotalluminosityobtained all magnitudes. spirals, 23percentSO’s,and13ellipticals.It omitted kcorrection;however,theincreaseofnumber et al.(1961-1968)usingtherelationbetweentwo characteristic scaleandsurface brightness.Then,it galaxies havethesameshape, differingonlyintheir dotted lineistheadoptedrelationforallfields. interval infieldsneartheindicatedclustersofgalaxies.The Among thosefieldgalaxiesbrightenoughtobe To findthetotalluminosityofagalaxy,onemust It seemsthatthisrelationcanbefittotheusual Fig. 2.—Thenumberdensityoffieldgalaxiesper0.4-mag b) TotalMagnitudes 197 4Ap J . . .194 10 in theearlierstudy.Mostof theverticaldispersioncan tion oftheA1314datawasuncertain,andaccuracy luminosity isuniquelydeterminedbythemeansurface for ellipticalgalaxiesina number ofclusters.The Abies and(1972).Thecalculationshavethere- brightness ofthegalaxywithinlimitingisophote. was shown,theextrapolationfromisophotaltototal distribution showsthesame characteristic shapefound brightness isplottedagainst theisophotalmagnitude now available.Infigure3a,themeanredsurface fore beenrepeatedusingsomeoftheabundantdata and Baum1968)availablethen.Theabsolutecalibra- amount ofdataonA1314andtheComacluster(Rood mean relation,(b)ThedashedlinesarethecalculatedextrapolationstototalmagnitudeforellipticalandSOgalaxies.points 6 of theComadatahasrecentlybeenquestionedby adopted relationforalltypes. are empiricallydeterminedextrapolationsforellipticals{filledcircles),SO’s{openandspirals(x’s).Thesolidlineis the The extrapolationswerecalculatedfromalimited Fig. 3.—-(û)Meansurfacebrightnessversusabsolutemagnitudeofellipticalgalaxiesinseveralclusters.Thecurveistheadopted © American Astronomical Society • Provided by theNASA Astrophysics Data System AUGUSTUS OEMLER,JR. the luminositywithinS =26.0magpersquare the skybrightness,andwas thenusedtodetermine be attributedtomeasuringerrors,sothemeanrelation these semiempiricaldeterminations,theextrapolations Also shownbyadashedlinearetheextrapolationsfor reduced toremoveplatenoise andproperlysubtract were obtaineddirectlyforasmallnumberofobjects. curve; theseareshownbythedashedlineinfigure3Z>. shown bythesolidlinewillbeconsideredasdescribing Isophotometry oftheseobjects wasverycarefully etry ofSOgalaxiesbyFreeman(1970).Asacheckon all ellipticalgalaxies. extrapolations forEgalaxieswerecalculatedfromthis SO galaxiescalculatedinthatpaperfromthephotom- F Using theproceduredescribedinOemler(1973) Vol. 194 197 4Ap J . . .194 10 two agreedoverthemagnitudeintervaltheyhadin redshifts intable1.The/^-corrections werecalculated converted toabsoluteisophotal magnitudesusingthe common. Innocasewasthereastatisticallysignificant function ofthesmalleranddeeperfieldsothat fainter magnitudesbyrenormalizingtheluminosity luminosity functionofthelargefieldwasextendedto where Aj=0.27.Apparentisophotal magnitudeswere difference betweenthetwocurves. field, thatquadrantofthefieldindirection cluster—where therewasaneighboringclusterinthe there wasonlyone,theluminosityfunction other clusterwasomittedatradiigreaterthan30'.The nosity functionscannotbeusedtodirectlyobtainthe larger orsmallerareawasused,sothatthefinallumi- improved signal-to-noiseratiomadeitdesirable,a than thatwithinoneAbellradius.However,whenan generally constructedfromanareasomewhatlarger larger ofthetwofields,andinsinglefieldwhen the incompletenessateachmagnitudecalculated cluster richnessorthetotalluminosity. fields ofdifferinglimitingmagnitude,separatelumi- background tothegalaxycountsreachedone-third. nosity functionswereconstructedforeachfield.Inthe earlier. Forthoseclusterswhichwerestudiedintwo reached 25percentorwherethecontributionof tude wheretheincompletenessofphotometry that theassumptionsinOemler(1973)aboutthemare lated extrapolationsforSOandspiralgalaxies,implying ticals, apossibilityraisedintheearlierstudy.However, tion withinourownGalaxyusingtheprescription Before thispoint,thegalaxycountswerecorrectedfor ences betweentheempiricalextrapolationsfordifferent there isnorelationbetweentheempiricalandcalcu- used toconstructluminosityfunctionsforallclusters. in figure3bwillbeusedforallgalaxies. system, becomes given bySandage(1973)which,forthisphotometric morphological types,therelationshownasasolidline not correct.Sincethereappeartobenogreatdiffer- what overestimatetheextrapolationsforfainterellip- galaxies brighterthan—21.0.Thecalculationssome- pirical andcalculatedextrapolationsforelliptical filled circles,SO’sbyopenandspiralsx’s. figure 3bfortheseobjects,withellipticalsdenotedby Each luminosityfunctionwasextendedtothatmagni- nosity. Thequantitym(23)—(26)isplottedin second isophote,whichisveryclosetothetotallumi- No. 1,1974 F In thethreecases—A2197,A2199,andHercules The galaxymagnitudeswerecorrectedforabsorp- The onlysignificantcorrection wasthatforA539, The backgroundcorrectionsdeterminedabovewere There israthergoodagreementbetweentheem- © American Astronomical Society • Provided by theNASA Astrophysics Data System Aj =A—0.0{b>50°),(5) Aj =0.11(escb-X)(b<50°),(6) A =0.07(cscb-\)(b<50°).(7) F f c) DataReduction SYSTEMATIC PROPERTIESOFCLUSTERSGALAXIES 10 2 units of20magpersquare arcminute,whichis zontal scaleisinmegaparsecs andtheverticalscalein A1367, A2197,andtheHerculescluster).Theformer equivalent to9.62x10L Mpc“. Theerrorbars in §III«.Thesurfacebrightnessesofdistantclusters relation fromthemeancolorforfieldgalaxiesfound relation istakenfromequation(2)andthelatter clusters or/—F=0.9forspiral-rich(A1228, using J—V=0.3and-F1.1forspiral-poor The luminositieswereconvertedtothevisualband are correctedfortheeffectofkfactorand function, extrapolatedtoinfinitemagnitudeby large fieldwasaddedusingtheclusterluminosity using therelationinfigure3b.Theadditionallumi- tions fromisophotaltototalmagnitudeweremade (redshift, angulardiameter)-relationforq=0.0. nosity ingalaxiesbeyondthecompletenesslimitof was removedusingequations(6)and(7).Extrapola- figure 2,wassubtractedfromthisprofileandaseries lated inaseriesofconcentricrings.Thenthesurface brightness duetobackgroundobjects,calculatedfrom largest fieldstudiedofeachcluster,whosedepths the derivedprofilesshouldbeminimal.Using to thecentralgradientingalaxydensity,itseffecton ber, thesurfacebrightnessduetogalaxieswascalcu- ranged from2.5to4magbeyondthebrightestmem- of correctionsmade.Theeffectgalacticabsorption position is0.25Mpcandformostclustersmuch the lineofsight,anyway,andnumberobjects smaller. Sincethisuncertaintyisinverselyproportional mated thatintheworstcasesuncertaintythis plot ofthepositionsbrightergalaxies.Itisesti- the distributioninplaneofsky. is toosmalltoadequatelyspecifyanyothershapefor ing theclustersurface-brightnessprofiles.Thisisnot integrated functionatM~—21foundbyAbell. from thisstudydonotshowtheverysharpbreakin the totalareasstudied,agreementinA2199is taken ofdifferencesinthephotometricsystemsand cluster alsomatchratherwell,exceptthattheresults pared withresultsbyAbell(1974).Whenaccountis o However, circularsymmetrymustbeassumedalong as A194wherethebrightestgalaxiesformalongchain. a verysatisfactoryapproximationinsomecases,such excellent. ThetwoluminosityfunctionsoftheComa filled circlesareintegratedcounts. tions forallclustersarepresentedinfigure4.Open tudes werecalculatedforthiscluster.Luminosityfunc- circles aregalaxycountsper0.4-maginterval,and close totheJband.Therefore,ratherthanFmagni- Q for themoredistantclustersusingascanofcentral observed intheFbandwasemittedatwavelengths galaxy inA2670obtainedbyJamesGunn.Theredshift of ZwCl1545.1+2104issogreatthatthelight F The finalresultsarepresented infigure5.Thehori- In eachcluster,thecenterwasdeterminedfroma Spherical symmetryhasbeenassumedindetermin- Several oftheseluminosityfunctionscanbecom- log N=C+0.15m.(8) 7 Fig 4b Fig. 4a - p 4 The cluster luminosity functions. Open circles are galaxy counts per 0.4-mag interval, and filled integrated counts. IG

8

© American Astronomical Society • Provided by the NASA Astrophysics Data System 197 4Ap J . . .194 10 © American Astronomical Society • Provided by theNASA Astrophysics Data System N ß°) 9 b 6 I a £ rt* ü « 197 4Ap J . . .194 10 © American Astronomical Society • Provided by theNASA Astrophysics Data System 197 4Ap J . . .194 10 © American Astronomical Society • Provided by theNASA Astrophysics Data System Ph ü £ d Ô £ ü K 197 4Ap J . . .194 10 30 12 to thesumofallprofiles.Usingthiscurveanda background. tics andtheassumed50percentuncertaintyin the entireclusterluminosity(Oemler1973),hasbeen dominant. Ratherthanrelyingonindividualpointsin three-dimensional massdistribution,thebinding the diffusebackgroundlight,amountingtoone-third fitted toapowerlaws~r~-,whichisthebestfit at largeradii,wherethebackgroundgalaxiesare analyzing theseprofiles,theouterpartshavebeen A2670 comparablewiththosefortheotherclusters, was calculatedforeachcluster.Tomaketheresults smooth fittotheinnerpoints,totalluminosity are thesumofuncertaintiesduetoPoissonstatis- energy canbecalculatedforagivenmass-luminosity ratio. Inthesecondandthirdcolumnsoftable2are omitted. units), andthegravitationalradiusR(inmegapar- listed, foreachcluster,thetotalluminosity(insolar secs), thelatterbeingdefinedby tion ofeachtypehasbeencalculatedforthosegalaxies within Rand3magofthebrightestmember. These percentagesarelistedinthenextthreecolumns of thegalaxieswereclassified,fractionalcomposi- ties, theycanbedividedintocleartypes.Clustersof clusters showanalmostcontinuousrangeofproper- of table2. made aboutthedatapresentedaboveisthat,while their galaxycontent.Inspection oftable2andthe G galaxies havebeenclassified inthepastbyanumber of systems,usingseveral their observedproperties; one ofthemostnaturaland objectivemethodsisby descriptions in§\\dshowsthat theydivideintothree g The accuracyoftheseprofilesdeterioratesrapidly Then, usingvonZeipel’sinversiontoobtainthe For thoseclustersinwhichthemorphologicaltypes One ofthemostimmediatepointswhichcanbe © American Astronomical Society • Provided by theNASA Astrophysics Data System V. ANALYSISOFTHERESULTS 2 ^Grav =“GM[R.(9) g Abell 539 Abell 400 Abell 194...... Abell 1904 Abell 1367.... Abell 1228 Abell 665... Zw Cl1545.1+2104. Abell 2670 Abell 2199..... Abell 2197.. Abell 1413.. Abell 1314.... Coma cluster. Hercules cluster Cluster AUGUSTUS OEMLER,JR. 12 (10L) G 17.0 2.6 4.2 2.6 2.9 Cluster Parameters 5.5 5.7 7.8 9.4 3.3 3.5 3.1 L 3.6 8.2 1.8 TABLE 2 (Mpc) groups. ThecDclustershaveone(ortwo,inthecase A1413, theComacluster,A1904,A2199,A2670,and the Herculescluster,andA2197.Theremaining E:S0:Sp ratiosofabout3:4:2;thisgroupincludes from thatofthefield,andincludeA1228,A1367, Zw Cl1545.1+2104.Thespiral-richclustershavean of Coma)uniqueanddominantmemberhave clusters, whichmaybecalledspiral-poor,haveE:SOSp E:S0:Sp ratioofabout1:2:3,nottoodissimilar these clusters.Thisclassificationschemeismostclosely A1314, andA665.TheclassificationofA665isun- ratios ofabout1:2:1andincludeA194,A400,A539, 4.0 2.7 tent andthegeneralappearanceofclusters.Spiral- lines betweentypesdonotpreciselycoincide. portions ofgalaxytypesresemblethosethecD certain becauseofitsdistance,butappearance 4.1 2.9 2.8 2.6 the center.Onotherhand,cDclusterstendto related tothatofMorgan(1962),althoughthedividing clusters, thereisnoambiguityinthetypingofany suggests thatitbelongstothethirdratherthan 2.6 3.7 4.0 5.6 3.3 rich clusterstendtobeirregularinshape,withno second class.WiththeexceptionofA1314,whosepro- 3.7 3.9 with awell-definedcentertowardwhichthemassis strong correlationbetweengroupingsbygalaxycon- 3.4 3.7 than thespiral-richclustersbutlessregular highly concentrated.Thespiral-poorclustersfall resemble globularstarclusters,beingrathercircular, symmetry andnoapparentgradientindensitytoward midway betweenthesetwotypes,beingmoreregular Ro cD clusters. tative fashion.Onemeasureofthecompactnessa luminosity, insolarunits, in figure6.Spiralrich weighted harmonicmeanofthedistancesbetween cluster isitsgravitationalradius,whichamass- other propertiescanbeexpressedinamorequanti- posed onthemeantrendof sizewithluminosityisa cluster, inmegaparsecs,is plotted againstitstotal cluster members.Thegravitationalradiusofeach clusters aredenotedbyx’s, spiral-poorclustersby open circles,andcDclusters byfilledcircles.Super- As hasbeenpointedoutbyAbell(1974),thereisa Some ofthesecorrelationsgalaxycontentwith 7oE 35 35 38 19 19 17 14 15 18 14 7 oS0 41 47 40 43 25 53 56 36 56 36 %Sp 45 43 28 29 26 24 63 50 18 19 Vol. 194 No. 1, 1974 SYSTEMATIC PROPERTIES OF CLUSTERS OF GALAXIES 13 of points is the composite of several clusters, namely, A1228, A1367, and the Hercules cluster for the spiral- rich profile; A400, A539, and A1314 for the spiral-poor profile; and A1904, the Coma cluster, A2199, and A2670 for the cD cluster profile. Each cluster was 1 scaled by RG~ to equalize their size. The errors are the sum of those due to Poisson statistics and the assumed 50 percent uncertainty in the background. Three points can be made about these profiles. The outer envelopes all have steep slopes consistent with 3.0, similar to that predicted by the numerical models of Peebles (1970) and others and much steeper than that of 1.0 given by a projected isothermal sphere. The logU spiral-rich clusters have a flat central density gradient and little central concentration. As one moves to the Fig. 6.—Gravitational radius, in megaparsecs, versus total spiral-poor and then to the cD clusters, the central luminosity, in solar units, for all clusters. Spiral-rich clusters are designated by x’s; spiral-poor clusters, by open circles; concentration and density gradient increase, although and cD clusters, by filled circles. The solid line is the fit to the the latter never becomes as high as that of the envelope. data. Midway in the profiles is a plateau or local minimum at about 0.4Rç which grows in importance as one moves systematic tendency for the spiral-rich clusters to be from spiral-rich to spiral-poor to cD clusters. It is larger and the cD clusters smaller than average, as more prominent in some of the individual cD cluster expected from their appearance. profiles, such as A1904, A2199, and A2670. Because of this variation of size with type, and The significance of this feature becomes even stronger because the various types are not distributed evenly in the three-dimensional mass profiles. The dotted with L, a single fit to all the data would be misleading. curves in figure 8 are typical three-dimensional mass Instead, one straight line with a slope of 0.30 was distributions obtained by inverting a smooth curve fitted by least squares to the cD clusters, and a separate through the points. These are only typical curves; one line of slope 0.37 was fitted to the rest of the data. The could easily draw a projected curve through the cD mean of these two lines, weighted by the number of profile which would make the space density zero in clusters represented by each, is shown as a solid line some shell. in figure 6. Its slope of 0.34 ± 0.05 is essentially equal This feature has been seen by other investigators to 1/3; in other words, in the mean all clusters have and cannot be easily explained away. Bahcall (1971) the same density, independent of mass. has shown that it is not due to measuring errors or to Using this mean relation, the systematic trend of the presence of subclustering. It is seen in both the size with type is illustrated in figure 7. Plotted for each luminosity distribution and number counts (Clark cluster is the deviation of log RG from the mean line 1968). Since both of these depend on the brightness versus its E:Sp ratio; A2670 has been plotted, using of the galaxies, one might wish to reconcile the observ- an E : Sp ratio typical of cD clusters. With the exception ed profile with a smooth mass distribution by invoking of A1367, which on plates is noticeably compact for a a discontinuity in the mass-to-light ratio. This could be spiral-rich cluster, there is a quite tight relation between accomplished either by giving the galaxies in the core compactness and galaxy content. a higher mass-luminosity ratio or by removing mass The detailed distribution of luminosity within from their outer envelopes into a diffuse sea of stars, clusters of each type is illustrated in figure 8. Each set where it would not be observed with the techniques used. It is known that such a component of diffuse light exists in some clusters (Oemler 1973), but it is probably not massive enough. Also, a comparison of galaxies in the Coma cluster inside the feature (but excluding the very center) with those outside shows no difference in either their surface brightness or mean magnitude greater than 0.02 mag. From the informa- tion now available, we must conclude that this is a real feature, in the mass, as well as the light, distribu- tion in clusters. The composite cluster profiles have been broken down by the morphological types of the galaxies in figure 9. These graphs are in the same format as figure 8 except that the vertical axis represents number density rather than surface brightness. Although the uncertainties are large, the different galaxy types seem to have almost the same distribution, with one major Fig. 7.—The deviation of log RG of a cluster from the mean line in fig. 6 versus its elliptical to spiral ratio. exception : in the spiral-poor and cD clusters, the

© American Astronomical Society • Provided by the NASA Astrophysics Data System 197 4Ap J . . .194 10 © American Astronomical Society Sßoi 14 Provided bythe NASA Astrophysics Data System 00 JD T3 1 ’tí Jtí H .tí ’tí ‘a 12 •a 3 Ig ¡3 5 o tí O tí tí § tí 9’tí 5 U tí fe S a o o a tí tí -I tí tí W) Si c/3 ¿I w o a ^ O 85 r h Q ü ° a o sa ■ I -S 3 tí 197 4Ap J . . .194 10 © American Astronomical Society • Provided by theNASA Astrophysics Data System 15 197 4Ap J . . .194 10 the statisticalsignificanceofthisisratherlowand the center;infact,decreaseissteepenoughtomake projected densityofspiralgalaxiesdecreasestoward the featureat0.5i?. to participatethemost,andspiralsleast,in the meanandthatofspiralslessthanmean,but the spacedensityofspiralszeroincore.Thereis poor (opencircles),andcD(filledcircles)clusters. interactions. Presentedinfigure10isthemeanradius to theapproachenergyequipartitionbytwo-body corrections. AlsointhecDprofile,ellipticalsseem could easilybeduetoslighterrorsinthebackground {stars), spiral-poor{opencircles),andcD{filledcircles)clusters. galaxies withintheprojectedgravitationalradiusinspiral-rich in asystemwhichhaspartiallyevolvedtowardenergy is marginallysignificantevidencefordifferential whether clustersofgalaxiesshowmasssegregationdue of theellipticaldensityinenvelopeisgreaterthan some suggestion,inthecDclusters,thatgradient large galaxieshavemovedinward.Thesmaller but nonemorethan2magbelowthebrightestmember. mass segregationinthefirstfewmagnitudeintervals different magnituderangesinspiral-rich(stars),spiral- have gainedenergyfromthelargeobjectsandmoved lose energytothesmallerobjectsisshort,andthese The shapeoftheselattertwocurvesisthatexpected segregation. Inthespiral-poorandcDclusters,there of galaxieswithintheprojectedgravitationalradiusin treated inthesamewaythatwehavetheir from photometryofthecoreonly. inferring thefaintgalaxycontentofanentirecluster equipartition. Thetimescaleforthelargegalaxiesto This latterobservationjustifiesourprocedureof selves andthereforestillhavesimilardistributions. have nothadtimetoexchangeenergyamongthem- the differenttypeclusters. Compositedifferential by theirgalaxycontent,one mightexpecttosee outward; however,becauseofthelongtimescale,they 16 AUGUSTUS substantial differencesinthe luminosityfunctionsof galaxy distributions.Sincethe clusterswereclassified G There hasbeensomecontroversyinthepastover The spiral-richclustersshownoevidenceofmass Fig. 10.—Themeanradiusasafunctionofmagnitude The luminosityfunctionsoftheclusterscanbe © American Astronomical Society • Provided by theNASA Astrophysics Data System OEMLER, JR.Vol.194 luminosity functionsforthesamegroupsofclusters rich functionappearstobeslightlydifferent,butthe first-ranked membertypicalofthelatter.Thespiral- within theerrorsexceptforoutstandinglybright poor andcDluminosityfunctionsareidenticalto between thesethreecurvesarequitesmall.Thespiral- the first-rankedgalaxyisusuallyanSOorelliptical, predicted theoreticallybyPressandSchechter(1974) difference isofonlymarginalsignificance.Theshape /- F=1.1)of0.24mag.Thisisnotmuchlarger than thoseofEandSOgalaxies,thisobservation luminosity functionoccursatthesameabsolutemag- Abell (1962)seemstobeanoversimplification. The approximationoftwostraightlinessuggestedby but differsmarkedlyfromthatpredictedbyIcke(1973). as infigure8arepresented11.Thedifferences two types. implies thatthemassfunctionofspiralsmustcutoff with theobservationthateveninspiral-richclusters nitude asinthespiral-poorandcDfunctions,coupled of theseluminosityfunctionsisquitesimilartothat mass-luminosity ratioofspiralsisconsiderablylower is asignificantpoint.If,ascommonlythought,the tions andfittingbyeyetotheindividualfunctions that thiseffectisduetophotometryorreduction this studymaybesummarizedasfollows.Although than wouldbeexpectedfromuncertaintyinthefitting produces adispersioninabsolutemagnitude(assuming cant differenceswiththeexceptionofthoseA665 with thoseoftheindividualclustersshowsnosignifi- at asignificantlylowermassthanthoseoftheother end oftheluminosityfunction. mean luminosityfunctionfromthecompositefunc- errors. Abell(1962)hasassertedthatthereisauniver- end. Althoughbotharedistantobjects,itisunlikely and A2670,bothofwhichrapidlyleveloffatthefaint dividing themintothreegroups.Onegroup,labeled clusters displayawiderangeofcharacteristics,most depends onthesmallnumberofgalaxiesatbright alone, sincethenormalizationinabsolutemagnitude absolute magnitudeforallclusters.Constructinga sal luminosityfunctioncharacterizedbythesame above asspiral-rich,hascompositionssimilartothat They arefairlyhomogeneousincomposition,withno density andnotendencytocentralconcentration. They areirregularinappearance,havealowmean of theirgrossfeaturesmaybeaccountedforby perceptible segregationbymorphologicaltype,and of thefield,withahighproportionspiralgalaxies. encounters. ticals thantheothercluster types.ThesecDclusters show noevidenceofrelaxationbytwo-body by centralsupergiantgalaxies. Theyaredeficientin spirals andcontainamuchhigher proportionofellip- are dense,centrallyconcentrated, andsphericalin That thecutoffatbrightendofspiral-rich A comparisonofthecompositeluminosityfunctions The propertiesofclustersgalaxiesfoundfrom At theotherextremearethose clustersdominated VI. CONCLUSIONS 197 4Ap J . . .194 10 © American Astronomical Society • Provided by theNASA Astrophysics Data System 197 4Ap J . . .194 10 profiles areingoodqualitativeagreementwiththe two clustertypesissuchthatthetransmutationof predictions ofnumericalmodelscollapsedclusters. change thespiral-richcluster contentintothatofa the necessarydynamicalevolutionshouldoccuron composition ofthespiral-poorandcDclusters.Will the relationbetweenthesetypesofclustersis condensed, andshowtheeffectsofaconsiderableperiod They havesmoothmassdistributions,arecentrally the characteristicsexpectedofsystemsinasteadystate. tion hasoccurred.Finally,theirhomogeneouscom- might bedonebytheinteraction ofthegascontent changing thegalaxycontent.Thecontentof collapse, andthereonlyremainstheproblemof become typicalspiral-poorclusters?Aswehaveseen, spiral-rich clustersliketheHerculesclustereventually position, resemblingthatofthefield,showstheir low densitiesofthespiral-richclustersimplyalong spiral-poor cluster.Gunnand Gottshowedhowthis some fractionofthespiral galaxiesintoSO’swill of two-bodyrelaxationprocesses.Theirluminosity affinity tothegeneralfieldofunboundgalaxies. segregation alsosuggeststhatlittledynamicalevolu- which occurduringcollapse,andthelackofmass not survivethephasemixingandviolentrelaxation tempting toregardthemorphologicalsequenceof the growthofadensecore.Withthisinmind,itis tative behaviorcanbecharacterizedbyaconstant of aspiralgalaxywithhot intergalacticmediumin collapse time.Theirirregularmassdistributionscould morphological evidenceforthisisquitestrong.The former hadnotcollapsed,whiletheothershad.The clusters andtheothertypescouldbeexplainedif in dynamicalevolution. increase inregularityandsphericalsymmetry,by differences betweenwhatareherecalledspiral-rich until itreachesastateofquasi-equilibrium,itsquali- particles, initiallyatrest,throughitscollapseand spiral-rich, spiral-poor,andcDclustersasasequence collapsed. reached apointofmaximumexpansionandthen behind thegeneralHubbleexpansionuntilthey expanding universe.Withtime,theseareaslagged tion ofthesefacts.Letusassume,followingthecon- regions ofhigherthanaveragedensityintheearly ventional idea,thatclustersofgalaxiesbeganas make sometentativeattemptsatatheoreticalexplana- condensed astheformer. two-body relaxationamongthebrightermembers. are notquiteasregular,compact,orcentrally segregation bymassandmorphologicaltype,butthey the spiral-poorclusters,whosecompositionisdomin- cores. Theyshowevidenceofconsiderabledifferential ated bySOgalaxies.LikethecDclusters,theyshow galaxies isduetoacompleteabsenceofspiralsintheir appearance. Theirlowoverallproportionofspiral 18 Most importantforGunnandGott’sargument Conversely, thespiral-poorandcDclustersshow If onefollowstheevolutionofaboundsystem Gunn andGott(1972)firstsuggestedthatthe Even withoutanydynamicalinformation,wecan Somewhat intermediatebetweenthesetwotypesare © American Astronomical Society • Provided by theNASA Astrophysics Data System AUGUSTUS OEMLER,JR. irregularity atalmostthe same radius.Also,itis understand whyallclusters should startwiththesame surprising thattheviolentrelaxation duringcollapse with timeastheclustercollapses, butitisdifficultto irregularity intheinitialmass distributionwillgrow it wouldseemthatthisfeaturemustbeduetothe initial distributionofmassinthecluster’svicinity, is thepeculiarfeatureinradialmassdistributions. poor. proportions ofallgalaxytypes,butthedataarevery any significantdynamicalevolution.Itistruethat since theseclustershavepresumablynotundergone If webelievethesmallbumpinspiral-richprofile, been important,suggestsdifferencesintheinitial transmutation ofspiralsintoSO’shas,presumably,not precisely determinedfromthepresentdata.Inspection figure 7.Therelativeimportanceofinitialrates the latter. rich precursorsofthecDclustershaveallevolvedinto galaxy creationandlaterevolutioncannotbevery between compactnessandgalaxycontentseenin increases withdensityrelativetotheothergalaxy they haveevolvedfromthespiral-poorclusters.These proportion ofellipticalsseemstoruleouttheideathat central profiles,wouldsuggestthattheyaredynamic- encounters wouldbesufficientlystrong. One observationwhichcannotbeexplainedatpresent of thestructureandevolutionclustersgalaxies. of theouterpartsprofilesinfigure9where but, becauseoftheirshortcollapsetimes,theelliptical- clusters aretheprecursorsofspiral-poorclusters; were originallyfourtypesofclusters.Thespiral-rich types. ticals isduetoabirthrateofellipticalgalaxieswhich the fastest,cDclustersmusthavebegunas collisions inaclusteristoolow(GunnandGott1972), tical inappearance.However,thefrequencyofdirect speculate, then,thattheirhigherabundanceofellip- densest fluctuationsintheearlyuniverse.Onecould are thedynamicallyoldest,andthereforehaveevolved cluster typesare,then,intrinsicallydifferent.Ifthey ally olderthanthespiral-poorclusters,theirgreater and itisdoubtfulwhetherthetidalinteractioninnear comparable withthecollisiontimetendstobeinelastic two stellarsystemsinwhichthedynamicaltimeis of thecDclusters,togetherwiththeirslightlysteeper disks oftwocollidingSOgalaxiesarelikelytobe poor andcD—isnotsoclear.Thedirectcollisionof sufficiently disruptedtomakethefinalsystemellip- between thetwotypesofcollapsedclusters—spiral- in figure9. toward thecenterofspiral-poorandcDclustersseen (Aarseth andHills1972;Toomre1973),thusthe muting ofSOintoellipticalgalaxies,andsotherelation explanation forthedecreaseindensityofspirals the mostnatural—althoughnotonlypossible— the coreofacollapsedcluster,andthistheoryprovides Thus, wecanbegintodevelopaconsistentpicture We thereforehavetworeasonsforthetightrelation At thepresenttime,then,weonlyseethreeofwhat Therefore, whilethegreaterregularityandsymmetry There isnoequivalentmechanismforthetrans- Vol. 194 197 4Ap J . . .194 10 irregularities intheformofsubclusteringaresmoothed Although thisparticularproblemhasnotbeen seems toamplifyratherthansmoothoutthisfeature. studied, AarsethandHills(1972)haveshownthat needs tobedone. out duringcollapse.Clearly,moretheoreticalwork No. 1,1974 .1974,inStarsandStellarSystems,Vol.9,ed.A. .1962,inProblemsofExtragalactieResearch,ed.G.C. Abell, G.O.1958,Ap.J.Suppl,3,211. Aarseth, S.J.,andHills,J.G.1972,Astr.Ap.,21,255. Icke, V.1973,Astr.andAp.,27,1. Bahcall, N.A.1971,A.J.,76,995. Abies, H.D.,andP.G.1972,AASPhoto-Bulletin, Noonan, T.W.1973,A.J.,78,26. Augustus Oemler,Jr.:KittPeakNationalObservatory,950N.CherryAve.,Tucson,AZ85719 Freeman, K.C.1970,Ap.J.,160,811. Clark, E.1968,A.J.,73,1011. Gunn, J.E.,andGott,R.1972,Ap.J.,176,1. Morgan, W.1962,Ap.J.,135,1. Morgan, W.W.,andLesh,J.R.1965,Ap.J.,142,1364. Matthews, T.A.,Morgan,W.W.,andSchmidt,M.1964, This studywassuggestedbyDr.JamesGunn,andI Press) (inpress). No. 2. Ap. J.,140,35. M. SandageandJ.Kristian(Chicago:UniversityofChicago McVittie (NewYork:Macmillan). © American Astronomical Society • Provided by theNASA Astrophysics Data System SYSTEMATIC PROPERTIESOFCLUSTERSGALAXIES REFERENCES tions. Thisworkwasmadepossiblebytheverygener- with additionalsupportunderNSFgrantGP-36687X. which greatlyhelpedthiswork,andDennisPalm would alsoliketothankDr.StevenShectman, .1968,ibid.,73,442. ous assistanceofaVirginiaSteeleScottFellowship, Ranny Adamsfortheirassistancewiththeobserva- am indebtedtohimforhisguidanceandsupport.I Toomre, A.1973,IAUSymposiumNo.58. .1974,ibid.,188,233. Press, W.H.,andSchechter,P.1974,Ap.J.,187,425. Peebles, P.J.E.1970,A.J.,75,13. Richard Gott,andPaulSchechterforthemanytalks Zwicky, F.,Herzog,E.,Wild,P.,Karpowicz,M.,andKowal, Sargent, W.L.1973,Pub.A.S.P.,85,281. Sandage, A.,andSmith,L.1963,Ap.J.,137,1057. Sandage, A.1973,Ap.J.,183,711. Rood, H.J.,andBaum,W.A.1967,A.J.,72,398. Zwicky, F.,andHumason,M.L.1964,Ap.J.,139,269. Shectman, S.A.1973,Ap.J.,179,681. Oke, J.B.,andSchild,R.E.1970,Ap.J.,161,1015. Oemler, A.,Gunn,J.E.,andOke,B.1972,Ap.{Letters), Oemler, A.1973,^(/7./.,180,11. 176, L47. C. T.1961-1968,CatalogueofGalaxiesandClusters Galaxies (Pasadena:CaliforniaInstituteofTechnology). 19 © American Astronomical Society • Provided by the NASA Astrophysics Data System