A&A 382, 60{83 (2002) Astronomy DOI: 10.1051/0004-6361:20011754 & (cid:13)c ESO 2002 Astrophysics Hidden star-formation in the cluster of galaxies Abell 1689? P.-A. Duc1;2, B. M. Poggianti3, D. Fadda4;2, D. Elbaz2;5, H. Flores6;2, P. Chanial2, A. Franceschini7, A. Moorwood8, and C. Cesarsky8 1 CNRS URA2052 2 CEA, DSM, DAPNIA,Serviced’astrophysique, 91191 Gif{sur{Yvette Cedex, France 3 Osservatorio Astronomico di Padova, vicolo dell’Osservatorio 5, 35122 Padova, Italy 4 InstitutodeAstro(cid:12)sica deCanarias, Via Lactea s/n, 38200 La Laguna { Tenerife, Spain 5 Department of Astronomy & Astrophysics,University of California, Santa Cruz, CA 95064,USA 6 DAEC/LUL,Observatoire de Paris{Meudon, 5 place Jules Janssen, 92195 Meudon, France 7 Dipartimento diAstronomia, Universita(cid:18) diPadova, Vicolo dell’Osservatorio, 5, 35122 Padova, Italy 8 European SouthernObservatory,Karl-Schwarzchild-Strasse, 2 85748 Garching beiMu¨nchen, Germany Received 5 October 2001 / Accepted 28 November2001 Abstract. At a redshift of 0.18, Abell 1689 is so far the most distant cluster of galaxies for which substantial mid{infrared (MIR) data havebeen published.Itsmappingwith theISOCAMcamera onboard theISOsatellite allowedthedetectionof30clustermembersat6.75(cid:22)m(LW2(cid:12)lter)and16clustermembersat15(cid:22)m(LW3(cid:12)lter) within a clustercentric radius of 0.5 Mpc (Fadda et al. 2000a, Paper I). We present here the follow{up optical photometric and spectroscopic observations which were used to study the individual properties of the galaxy membersofA1689.Wecon(cid:12)rmthehighfraction ofbluegalaxies initially reportedinthisrichclusterbyButcher & Oemler (1984), that was challenged by some subsequent studies. We discuss the spectral and morphological properties of all cluster members in our spectroscopic sample, and of the MIR{detected galaxies in particular. Sources with a low [15 (cid:22)m]/[6.75 (cid:22)m] flux ratio typically consist of luminous passive early{type galaxies while thosewithahighMIRcolorindexaremainlyluminous,blue,emission{line,morphologicallydisturbedspirals,i.e. thestar{forminggalaxiesusuallyassociatedwiththe\Butcher{Oemler"e(cid:11)ect.Ontheotherhand,atleast30%of the15 (cid:22)m sources have optical counterpartsshowing no evidence for current star{formation activity,while their 15 (cid:22)m emission is most likely due to obscured star formation. We argue that the LW3 luminosity measured in the cluster members is a reliable tracer of the total infrared luminosity which in A1689 galaxies peaks at LIR = 6:2(cid:2)1010 L(cid:12). We derive from LIR a star{formation rate free of dust extinction, SFR(IR), which we comparewiththatdeterminedintheopticalfromthefluxofthe[OII]emissionline,SFR(opt).Thehighesttotal star formation rates (11 M(cid:12) yr−1) and dust extinction are measured in those galaxies exhibiting in their optical spectrum a signature of a dusty starburst. In contrast, none of the galaxies with post-starburst optical spectra has been detected by ISOCAM down to a 15 (cid:22)m flux limit corresponding to 1.4 M(cid:12) yr−1. We (cid:12)nd a median SFR(IR) of the LW3{detected galaxies of 2 M(cid:12) yr−1that is ten times higher than the median SFR(opt) of the [OII]{ detected galaxies. The ratio SFR(IR)/SFR(opt)is in fact very high, ranging between 10 and 100 for LW3{detected galaxies with [OII] emission. We conclude that a major part, at least 90%, of the star formation activity taking place in Abell 1689 is hidden. Whether the high extinction measured in the star-forming cluster members results from the cluster environmentitself or reflects a comparable extinction in the coeval (cid:12)eld is still unclear. Key words.galaxies: clusters: individual: Abell 1689 { galaxies: evolution { infrared: galaxies 1. Introduction universe, they highlighted the strong rise of the star for- mation comoving density up to z = 1 and its possible The history of star formation (SF) as a function of red- decline at higher redshift. Further multiwavelength stud- shift has recently received extensive attention since the ies have added points on the curve and addressed various reference work by Madau et al. (1996). Combining opti- biases and uncertainties in the slopes of the SF rise and cal observations on (cid:12)eld galaxies in the nearby (z = 0), decline. In contrast, the global star formation history in intermediate redshift (z < 1) and distant (1 < z < 4) clusters is practically unknown (See though Kodama & Bower 2001). A number of impediments make its study Send o(cid:11)print requests to: P.-A.Duc, e-mail: [email protected] particularly di(cid:14)cult. First of all, data for distant clus- ? Based onobservationscollected at theEuropeanSouthern ters (z > 1) are still missing. Furthermore, variations in Observatory,La Silla, Chile (ESO No. 61.A-0619). Article published by EDP Sciences and available at http://www.aanda.org or http://dx.doi.org/10.1051/0004-6361:20011754 P.-A.Ducet al.: Hidden star-formation in Abell 1689 61 cluster properties, such as the X{rayluminosity, the level emission lines and the presence of strong Balmer lines of sub-clustering and, generally, the dynamical status of in absorption. Classically, such objects are considered as the cluster, introduce a large dispersion in any trend. A post{starburstorpost-starforminggalaxies(e.g.Couch& \global" history { even just limited to rich clusters { can Sharples1987;Poggiantietal.1999).Couldtheyhowever only be obtained using large samples. Also, the star for- beextremee(a)galaxiesinwhichstarformationistotally mation history of clusters is largely related to the star obscured at optical wavelengths, as suggested by Smail formation history in the coeval (cid:12)eld via the infall rate et al. (1999) from radio-continuum observations? In ad- of (cid:12)eld galaxies which is itself redshift-dependent as well. dition, the post{starburst interpretation of the strongest Any uncertainty in the latter has a strong impact on the k+a’s requires that a starburst episode occurred in these former. Finally, problems with dust obscuration are crit- galaxies, possibly when they fell in the cluster. Optical ical when computing the (cid:12)eld star formation rate (SFR) surveyshaveyetfailedtopinpointtheputativestarburst- density,andtheycouldbeatleastasannoyinginclusters. ing progenitors, perhaps for time scale reasons (Couch The long{terme(cid:11)ect of the cluster environmentis un- etal.2001),and\e(a)"galaxieshavebeensuggestedtobe doubtedly to quench star{formationin its member galax- the most likely progenitors of the post-starburst galaxies ies(e.g.Abrahametal.1996;Poggiantietal.1999;Balogh (Poggiantiet al. 1999). et al. 1999). Indeed, nearby rich clusters contain more Undoubtedly deep mid{infrared surveys can address passive, non-starforminggalaxiesthan moredistant ones. someofthepreviousquestions.Fluxesat10{15(cid:22)m,along This is one aspect of the so{called Butcher{Oemler ef- with radio centimetric fluxes, provide the best estimates fect which is the observedincrease in the number of blue, of the dust obscured star formation rates (e.g. Elbaz presumably star{forming galaxies in clusters, as a func- et al. 2001)when far{infrareddata arenot available.The tionofredshift(Butcher&Oemler1984,hereafterBO84). latest generation of infrared cameras, such as ISOCAM Furthermore, at intermediate redshifts the star formation (Cesarsky et al. 1996) on board the ISO, have provided rate per cluster galaxy appears to be lower than in simi- for the (cid:12)rst time images in the mid{infrared regime with lartypesofgalaxiesinthesurrounding(cid:12)eld(e.g.Dressler a (cid:12)eld of view, spatial resolution and sensitivity (Genzel etal.1999;Baloghetal.2000).Unfortunately,estimatesof & Cesarsky 2000) particularly well suited to the study of theSFRrelyonmeasurementsofopticallinefluxes,often distant clusters. [OII], obtained with aperture limited slits. Some Hii re- WemadeuseoftheuniquecapabilitiesofISOCAMto gionsmighthavebeenmissedandhencetheSFRunderes- mapat6.75(cid:22)m(LW2(cid:12)lter,bandwidth:5.0{8.3(cid:22)m)and timated.Datafromnarrow{bandH imageshaverecently (cid:11) 15 (cid:22)m (LW3 (cid:12)lter, bandwidth: 11.6{18.0 (cid:22)m) a sample becomeavailableforafewz =0:2−0:3clusters(Balogh& of ten clusters in the redshift range 0.2{0.9 (Fadda et al. Morris2000).This kindof data hasthe advantageof cov- 2000a).The mid{infrareddata of themost nearbycluster eringthewholegalaxy,butitislesssensitivetolowlevels of our sample, Abell 1689, were presented in Fadda et al. of star formation activity. Moreover, the SFR derived in (2000b,hereafterPaperI).Towardsthisrichcluster,situ- the optical might su(cid:11)er from strong dust extinction. A atedataredshiftof0.181,wedetected45infraredsources constant dust absorption of 1 mag at H , typical of local (cid:11) (41 LW2 sources and 25 LW3 sources). Among them, spirals(Kennicutt1992),isusuallyassumedtocorrectthe 6 (13%) sources were identi(cid:12)ed as stars and 8 (18%) turn optically-derived SFR. However, dust obscuration varies out to be foreground or background objects. A1689 was dramatically from one galaxy to another, and it is likely foundtocontainapopulationofgalaxieswithhigh15(cid:22)m todependalsoontheenvironmentalconditions,aswillbe fluxesandhighMIRtoopticalfluxratios,whilesuchpop- discussed in this paper. Direct optical signatures of dust ulation is absent towards the central region of the Virgo enshrouded starbursts have already been found in some and Coma clusters. This was interpreted as revealing an clustergalaxiesthathaveaspectrumclassi(cid:12)edas,accord- \infrared Butcher-Oemler e(cid:11)ect". Moreover, the compar- ing to the authors, \e(a)" (Dressler et al. 1999), \a+em" ison of optical and MIR counts in the (cid:12)eld and towards (Balogh & Morris 2000) or \S+A" (in the (cid:12)eld, Flores A1689 seemed to suggest a possible excess in the number et al. 1999).All these spectra show a moderate[OII] (cid:21)3727 of MIR emitters in the cluster as compared to the num- line in emissionand strongearly Balmer absorptionlines. ber expected given the di(cid:11)erent spiral fraction in the two Such features are well reproduced by spectrophotomet- environmentsandthefactthatmostoftheLW3emitters ric models with a selective dust extinction which a(cid:11)ects arespiralgalaxies.Suchanexcesswouldindicate thatSF di(cid:11)erently young and older stars (Poggianti et al. 2001). is triggered in some galaxies by the cluster environment. Thoughthis spectralcombinationis abletorevealtheoc- The 15 (cid:22)m luminosities and MIR-to-optical flux ratios, currenceofsomelevelofhiddenstarformation,themodels haveshownthatitisimpossibletoquantifythetotalSFR 1 We have taken H =75 km s−1 Mpc−1 and q = 0:5. on the basis of optical observations alone. 0 0 Adopting this cosmology, the distance of Abell 1689 is Furthermore, several spectroscopic surveys of distant 767 Mpc, which is thevalueusedthroughout thepaper. With clusters (e.g. Couch & Sharples 1987; Fabricant et al. a more typical cosmology, such as Ω = 0:3;Ω = 0:7, and m (cid:21) 1991;Bargeret al.1998;Fisher et al.1998;Dressler et al. H =70 kms−1 Mpc−1,thedistancewouldbecome877Mpc. 0 1999)havedisclosedasigni(cid:12)cantpopulationofk+agalax- At the adopted distance of Abell 1689, 1 arcmin corresponds ies whose spectrum is characterized by the absence of to159 kpc. 62 P.-A.Ducet al.: Hidden star-formation in Abell1689 instead, werenot found to be signi(cid:12)cantlydi(cid:11)erentin the 2.2. Imaging data cluster and in the coeval (cid:12)eld. B2, V and R images of the cluster have been obtained In the present paper, we focus on the properties of during the spectroscopic run which was carried out in the galaxies members of A1689 that were detected with May 1998 at the New-Technology Telescope at La Silla ISOCAM, comparing the optical and the IR view of the observatory. The seeing was poor (about 1:500). Landolt Butcher-Oemler e(cid:11)ect. A major goal of this work is to (cid:12)elds of photometric standard stars (Landolt 1992) were establishtheamountofSF hiddenbydustinthegalaxies observedforfluxcalibration.Duetotheatmosphericcon- of A1689, and whether a population of starburst galaxies ditions, the accuracy of the absolute photometric zero- ispresentintheclusterbutwouldbeunrecognizedonthe pointsisnobetter than0.05mag.Theastrometryofeach basis of the optical data. frame was computed using several tens of reference stars from the USNO A1.0 catalog (Monet 1996) queried via Our study relies on an extensive spectroscopic survey the ESO SKYCAT browser. The images were corrected oftheclusterthathasbeencarriedoutwiththeESONew fordistortionsduringthisprocess.Sourceswereextracted Technology Telescope (NTT) and on a collection of new from the reduced BVR images using SExtractor (Bertin and previous photometric data. The literature on Abell & Arnouts 1996).The photometriccatalogis complete to 1689 is indeed very rich. The cluster was included in the B = 23 mag, V = 22:7 mag and R = 22:7 mag. The original paper highlighting what was later known as the corrected isophotal (\mag best") BVR magnitudes from Butcher{Oemler e(cid:11)ect (BO84). Several authors have ob- SExtractor are listed in Tables 1{4. tainedadditionalphotometricdataanddiscussedwhether Ourphotometric databasealsoincludes the Gunn g,r the rather large blue fraction found in A1689 by BO84 and i magnitudes measured by Molinari et al. (1996) was real or not (e.g. Gudehus & Hegyi 1991; Pickles & with the ESO 3.6 m telescope and independently by van der Kruit 1991; Molinari et al. 1996; Margoniner & Margoniner & de Carvalho (2000) with the CTIO 0.9 m de Carvalho 2000). The lensing properties of the cluster, telescope,theB andI magnitudesmeasuredbyDyeetal. studied through deep optical images, were used to deter- (2001) at the Calar{Alto 3.5 m telescope and the near- mine its mass pro(cid:12)le (Tyson & Fischer 1995; Dye et al. infrared K0{band magnitudes measured by de Propris 2001). The latter was also estimated from X{ray maps et al. (1999) at the CTIO 1.5 m. All these di(cid:11)erent cat- (Miralda-Escude & Babul 1995). Evidence for dynamical alogs were cross-identi(cid:12)ed. Unfortunately because these substructures wasfound byTeague et al. (1990)based on data have various origins and were extracted with di(cid:11)er- a set of spectroscopic data. ent methods, they cannot be easily compared. In partic- Thepaperisorganizedasfollows:inSect.2wepresent ular they turn out to be of little use to derive a reliable our spectroscopicandphotometricobservationsandcom- spectralenergydistribution.Inthefollowing,wewillthus pare them with data from the literature. In Sect. 3, we mostly rely on our own BVR photometric data. describethepropertiesoftheoptically-selected(Sect.3.2) Finally,archiveHST/WCPC2images(PI:Tyson)cov- and MIR-selected cluster members (Sect. 3.3). Finally, in eringthewholeMOS(cid:12)eldofviewhavebeenprocessedby Sect. 4, we discuss the star formation activity and the the MORPHS group and were kindly given to us. They dust extinction of the galaxies in the cluster and provide consist of recombined F555W and F814W images. The some clues on environmental e(cid:11)ects that a(cid:11)ected their HST astrometry was lost during the combination process evolution. and it was re-computed using the astrometric solution of our ground-base images as reference. 2.3. Spectroscopic data 2. Observations and data-reduction The spectroscopic observations were carried out with EMMI installed on the NTT. Using (cid:12)ve masks punched 2.1. MIR data before the observations, we obtained spectra of 111 dif- ferent objects towards the inner Mpc2 (50 (cid:2)70) of Abell TheobservationsanddatareductionoftheISOCAMmid- 1689. Each slitlet was 1:300 wide (3.4 kpc at the distance infrareddatahavebeendescribedindetailinPaperI.We of A1689) and at least 800 long. The slit length was ac- recallherethebasiccharacteristicsofthesurvey.Thetotal (cid:12)eldofviewwas60(cid:2)60 (correspondingto0:95(cid:2)0:95Mpc) tually adjusted to cover the whole length of the target with apixelsizeof600.The spatialresolutionandastrom- object keeping some blank sky at each side. Note that, due to geometrical constraints, the slit orientation was etry were good enough to allow an unambiguous opti- not necessarily along the main axis. The targets were se- cal identi(cid:12)cation of all the mid{infrared sources but the lected from a deep V-band image of the cluster which ones near the very crowded cluster center. The 90% com- wasavailableinthe NTTarchive.Wechosepreferentially pleteness limits were 0.2 and 0.4 mJy at 6.75 (cid:22)m and 15 (cid:22)m, respectively. The sensitivity limits of the LW2 2 The EMMI red channelblue{likeBb (cid:12)lter which was used and LW3 surveys were 0.15 mJy and 0.3 mJy, respec- has a transmission curve slightly di(cid:11)erent from the standard tively. Johnson’s (cid:12)lter. P.-A.Ducet al.: Hidden star-formation in Abell 1689 63 Table 1. Spectroscopically-con(cid:12)rmed cluster members in theMOS sample. ID ID(cid:3) RA DEC z B V R morph (MOS) (ISO) J2000 J2000 (MOS) mag mag mag 1 { 13:11:24.00 −1:19:07.6 0.1829 21.53 19.98 19.24 { 2 { 13:11:24.25 −1:21:14.8 0.1766 21.41 20.82 20.27 Scd 3 { 13:11:24.27 −1:18:38.1 0.1803 21.87 20.38 19.65 Sb 4 { 13:11:24.40 −1:21:11.5 0.1870 20.72 19.11 18.42 S0 5 { 13:11:24.41 −1:22:15.8 0.1785 20.48 18.81 18.08 { 6 { 13:11:25.07 −1:19:31.7 0.1797 21.44 19.79 19.07 S0 7 3 13:11:25.31 −1:20:37.6 0.1924 19.75 18.13 17.38 S0 8 .. 13:11:25.64 −1:17:24.8 0.1814 20.91 19.31 18.59 E 9 { 13:11:25.94 −1:19:36.0 0.1983 21.81 20.48 19.77 Sab 10 { 13:11:26.51 −1:22:01.4 0.1953 21.21 19.66 18.98 Sc 11 .. 13:11:27.03 −1:16:10.5 0.1945 20.08 18.86 18.13 Sbc 12 { 13:11:27.05 −1:18:49.2 0.1852 22.36 20.88 20.09 Sa 13 { 13:11:27.06 −1:21:43.4 0.1955 22.52 21.28 20.68 S0 14 4 13:11:27.07 −1:20:58.9 0.2153 20.81 19.68 19.10 Scd 15 { 13:11:27.37 −1:22:48.2 0.1832 21.13 19.43 18.73 { 16 6 13:11:27.69 −1:21:06.7 0.2165 20.59 19.50 19.06 Sd 17 { 13:11:27.80 −1:21:13.4 0.1977 21.55 19.95 19.19 { 18 { 13:11:27.86 −1:21:37.2 0.1752 21.13 19.66 18.94 S0 19 { 13:11:27.98 −1:18:44.5 0.1890 21.26 19.67 18.89 S0 20 { 13:11:28.15 −1:18:12.7 0.1902 22.16 20.44 19.85 S0 21 8 13:11:28.16 −1:20:44.0 0.1712 20.34 18.75 18.04 S0 22 { 13:11:28.29 −1:18:45.5 0.1999 21.04 19.60 18.92 { 23 10 13:11:28.30 −1:19:58.9 0.1780 20.29 18.95 18.34 Sab 24 { 13:11:28.42 −1:22:42.1 0.1842 20.48 18.82 17.96 { 25 { 13:11:28.56 −1:20:27.1 0.1909 21.34 19.77 19.08 S0 26 { 13:11:28.69 −1:19:03.3 0.1859 21.91 20.21 19.44 { 27 { 13:11:29.26 −1:19:17.3 0.1932 21.30 19.57 18.89 S0 28 12 13:11:29.27 −1:17:50.4 0.1726 20.34 19.24 18.79 Sab 29 14 13:11:29.44 −1:20:28.4 0.1842 18.42 16.87 16.02 Scd 30 .. 13:11:29.71 −1:17:22.1 0.1996 20.87 19.20 18.43 S0 31 { 13:11:29.74 −1:17:43.2 0.1830 22.45 21.11 20.44 Sbc 32 { 13:11:29.83 −1:20:15.8 0.2016 21.82 20.11 19.27 Sb 33 16 13:11:29.94 −1:20:40.8 0.2002 19.83 18.72 18.08 Sp 34 { 13:11:29.95 −1:22:07.9 0.1985 20.48 18.96 18.21 S0 35 { 13:11:29.96 −1:20:17.9 0.1801 21.62 19.96 19.28 S0/a 36 .. 13:11:29.98 −1:16:25.6 0.1785 21.17 19.56 18.84 { 37 18 13:11:30.01 −1:20:43.3 0.2038 19.53 18.10 17.34 { 38 { 13:11:30.11 −1:22:30.8 0.1910 20.81 19.34 18.67 S0 39 20 13:11:30.23 −1:20:29.8 0.1745 19.53 17.87 17.17 E 40 22 13:11:30.53 −1:20:44.3 0.1918 20.40 18.78 17.91 E 41 { 13:11:31.03 −1:21:28.3 0.1878 20.28 18.50 17.87 E 42 24 13:11:31.09 −1:21:25.7 0.1972 20.93 19.42 18.32 E 43 26 13:11:31.36 −1:19:33.4 0.1878 19.54 17.81 17.03 E 44 { 13:11:31.47 −1:19:25.5 0.1745 21.27 19.74 19.05 Sbc 45 .. 13:11:31.54 −1:17:28.4 0.1913 20.93 19.26 18.55 S0 46 30 13:11:32.07 −1:19:47.5 0.1801 20.37 18.43 17.72 Sa 47 { 13:11:32.17 −1:22:11.3 0.1855 20.57 18.90 17.93 { 48 { 13:11:32.43 −1:22:18.6 0.1794 21.35 20.46 19.98 ? 49 31 13:11:32.60 −1:18:42.1 0.1757 19.44 18.17 17.65 Scd 50 { 13:11:32.61 −1:18:37.9 0.1823 { 23.96 20.84 S0 51 32 13:11:32.62 −1:19:59.3 0.2022 19.13 17.39 16.63 { 52 33 13:11:32.67 −1:19:32.4 0.2009 19.55 17.66 16.83 S0 53 .. 13:11:33.08 −1:17:02.4 0.1891 19.94 18.29 17.58 { 54 35 13:11:34.02 −1:21:02.5 0.1813 19.73 18.49 17.94 Sa 55 37 13:11:34.47 −1:18:11.7 0.1989 19.66 18.34 17.73 Sc 56 .. 13:11:34.65 −1:17:43.7 0.1926 20.22 18.54 17.77 E 57 { 13:11:34.72 −1:20:59.8 0.1895 20.90 19.19 18.49 E 58 .. 13:11:35.10 −1:23:18.3 0.1925 21.42 19.90 19.37 { 59 40 13:11:35.31 −1:21:33.8 0.1870 20.39 18.69 17.95 { 60 .. 13:11:35.47 −1:17:42.9 0.1758 20.99 19.29 18.58 Sa 61 41 13:11:35.55 −1:20:13.0 0.2000 19.86 18.76 18.26 Sc 62 { 13:11:36.52 −1:18:47.3 0.1769 22.18 20.56 19.85 E 63 .. 13:11:36.58 −1:22:54.6 0.1758 20.15 19.08 18.45 { 64 { 13:11:37.01 −1:22:32.4 0.1878 21.73 20.31 19.59 { 65 .. 13:11:37.12 −1:17:07.7 0.1851 20.88 19.21 18.42 S0 66 { 13:11:37.80 −1:19:21.2 0.1839 19.92 18.32 17.61 E 67 .. 13:11:37.88 −1:22:38.0 0.1887 21.19 19.69 18.94 { 68 .. 13:11:37.91 −1:18:09.0 0.1804 19.91 18.17 17.45 E 69 43 13:11:38.23 −1:21:05.7 0.1964 20.17 18.82 18.21 Scd 70 .. 13:11:39.31 −1:16:49.8 0.1845 20.13 18.48 17.86 E 71 .. 13:11:39.56 −1:17:50.1 0.1977 20.86 19.28 18.51 S0 72 45 13:11:40.09 −1:19:52.3 0.1880 19.87 19.13 18.70 Sp(late) 73 .. 13:11:40.27 −1:18:01.1 0.1807 20.46 18.90 18.16 { 74 .. 13:11:43.38 −1:19:20.4 0.1837 19.83 18.01 17.29 { (cid:3) Galaxies detected at 15 (cid:22)m are highlighted. Those outside theISOCAM(cid:12)eld of view are indicated with \..". Undetected ones in both LW2 and LW3 (cid:12)lters are indicated with \{". 64 P.-A.Ducet al.: Hidden star-formation in Abell1689 Table 2. Spectroscopically-con(cid:12)rmed cluster members from Teague et al. (1990). RA DEC ID z B V R morph J2000 J2000 (ISO) mag mag mag 13:11:18.37 −1:18:40.3 { 0.2095 18.92 17.69 17.05 S0 13:11:20.74 −1:20:02.9 { 0.1835 20.47 18.83 18.07 Sa 13:11:21.53 −1:19:44.2 { 0.2171 20.76 19.13 18.41 { 13:11:26.85 −1:19:37.5 { 0.1754 20.34 19.02 18.38 E 13:11:27.08 −1:22:09.5 { 0.1840 21.17 19.41 18.63 SB0 13:11:27.89 −1:23:09.1 { 0.1842 20.48 18.77 18.02 { 13:11:28.96 −1:21:17.3 { 0.1947 20.53 18.86 18.16 E 13:11:29.04 −1:21:37.7 { 0.1858 21.21 19.57 18.81 S0 13:11:29.08 −1:21:55.8 { 0.1908 21.02 19.35 18.57 S0 13:11:29.35 −1:18:35.4 { 0.1751 20.81 19.16 18.44 { 13:11:30.00 −1:20:43.0 { 0.1987 { { { { 13:11:30.20 −1:20:28.0 { 0.1750 { { { { 13:11:30.50 −1:20:46.0 { 0.1987 { { { { 13:11:30.90 −1:20:31.0 { 0.1739 { { { { 13:11:31.00 −1:21:28.0 { 0.1865 { { { { 13:11:31.04 −1:20:53.1 { 0.1885 21.32 19.55 18.80 E 13:11:31.30 −1:19:33.0 { 0.1864 { { { { 13:11:32.06 −1:21:38.7 29 0.1770 20.57 18.98 18.33 S0 13:11:32.57 −1:23:52.1 { 0.1865 20.30 18.62 17.87 { 13:11:33.74 −1:18:44.7 { 0.1849 20.79 19.08 18.37 E 13:11:35.34 −1:20:43.4 { 0.1835 21.21 19.54 18.80 E 13:11:35.98 −1:23:41.4 { 0.1861 20.29 18.54 17.76 { 13:11:39.43 −1:19:07.2 { 0.1813 20.39 19.02 18.40 Sa Table 3. ISOCAMcluster members with photometric redshift. ID RA DEC z B V R morph (ISO) J2000 J2000 (phot) mag mag mag 7 13:11:27.78 −1:20:08.3 0.193 20.31 18.50 17.78 E 11 13:11:28.40 −1:20:25.7 0.202 21.13 19.19 18.53 E 13 13:11:29.36 −1:20:43.9 0.180 20.98 19.43 18.60 Sb 17 13:11:29.99 −1:20:29.1 0.191 20.42 18.67 17.77 E 19 13:11:30.18 −1:20:52.3 0.191 20.44 18.59 17.69 E 21 13:11:30.34 −1:20:46.0 0.191 20.92 19.28 18.19 E 39 13:11:35.21 −1:18:54.9 0.190 22.61 21.08 20.42 Sbc the optical counterparts to the MIR sources detected by range depended on each slitlet position on the mask. We ISOCAM (Paper I) and completed the gaps between the made sure that the redshifted 4000 (cid:23)A break was present slitlets with bright cluster candidates. The completeness in all spectra. The wavelength of H is available only for (cid:11) oftheMOSsampleisanalyzedinAppendixA.Itisshown a few galaxies whereas the [OII] position is reached (cid:21)3727 that the initial selection criterion results only in a slight in most cases. bias towards blue galaxies. Whenever possible, objects in Data reduction and extraction were performed using the same magnitude range were arranged in each mask. a set of IRAF procedures written by PAD. A normalized Thesourcelists ofclustermembers andfore-/background dome flat was used as a flat-(cid:12)eld. The wavelength cali- galaxiesaregiveninTables1and4,respectively.Thetotal bration, based on HeAr lamp spectra, was carried out on integration time was about 2 hours per mask, divided in the 2D spectra that were initially roughly re-positioned 4 exposuresof30 mineach.Thedisperser,ESOgrism#3, to a common wavelength using the info of the slitlet po- hada resolutionofabout700at6000(cid:23)A .Thewavelength sitions available in the headers. The spectra of several P.-A.Ducet al.: Hidden star-formation in Abell 1689 65 Table 4. Foreground and background galaxies in theMOS sample. RA DEC ID z B V R Morph J2000 J2000 (ISO) (MOS) mag mag mag 13:11:23.95 −1:21:46.3 { 0.1330 { 21.55 21.00 ? 13:11:24.58 −1:20:04.1 { 0.4816 { 22.65 21.51 Sd 13:11:27.19 −1:20:10.5 5 0.0862a 19.41 18.24 17.60 Sab 13:11:27.81 −1:18:53.3 { 0.3840 { { 22.70 Sc 13:11:28.25 −1:18:28.1 { 0.7220 22.94 21.67 20.91 Sd 13:11:28.31 −1:18:32.7 { 0.0130 21.90 21.34 20.88 E 13:11:28.73 −1:21:43.9 { 0.7900 21.05 20.97 20.34 Irr 13:11:29.67 −1:17:47.4 15 0.3972 20.99 19.72 18.97 Scd 13:11:30.73 −1:21:39.2 23 0.6919 21.89 21.46 20.60 Irr 13:11:33.00 −1:21:25.1 { 0.1430 23.16 22.06 21.43 S0/a 13:11:33.69 −1:19:39.5 { 0.3100 22.85 21.16 20.26 { 13:11:34.07 −1:22:35.8 { 0.4340 { 23.01 21.86 { 13:11:35.18 −1:20:30.9 { 0.5867 { 22.59 21.33 Sd 13:11:35.95 −1:22:29.0 { 0.2420 21.71 20.59 20.12 { 13:11:36.40 −1:22:06.1 { 0.9443 { 22.23 21.59 { 13:11:37.37 −1:18:37.0 { 0.0825b 19.30 18.56 18.10 Sc 13:11:38.76 −1:19:08.2 { 0.3695 22.05 20.61 19.97 Sd 13:11:41.83 −1:19:48.1 { 0.1030 21.19 19.94 19.35 { Notes:a [TCG90] 217;theredshift of0.2153 asgiveninTeagueetal.(1990), ispresumablywrong. b [TCG90] 006; theredshift of 0.1826, as given in Teague et al. (1990), is presumably wrong. spectrophotometricstandardstarswereobtainedwiththe equivalent widths by 15{20%. This is due to systematic samegrismand500 widelong-slits.Afewobjectswereob- di(cid:11)erencesintheadoptedlevelofthecontinuum.Inorder served through di(cid:11)erent masks with their corresponding tocompareourresultswiththoseoftheMORPHSgroup, slitlets putatvariouslocations.Wecouldhencecheckthe wedecidedto base ourspectralclassi(cid:12)cationandanalysis relative accuracyof the flux calibration in the MOS (cid:12)eld, on the measurements obtained with the semi-automatic which turns out to be better than 20%. method. For galaxies observed through di(cid:11)erent masks, Redshifts were determined from the average value of we retained the spectra with the highest signal to noise. individual emission and absorption lines with a stronger The spectrophotometric data of our MOS run are listed weight given to emission lines. For spectra with very low in Table 5. signal to noise, no obvious lines could be identi(cid:12)ed; in- stead the wavelength of the 4000 (cid:23)A decrement was used. 3. Results The redshift of thirty galaxies could be compared with that measuredbyTeagueetal.(1990).Theyareidentical 3.1. Cluster membership and velocity distribution (within 1%) for all of them, but two3. Our redshifts are The histograms shown in Fig. 1 collect all redshift in- listed in Tables 1 and 4. formation available for galaxies lying in the 50 (cid:2)70 (cid:12)eld Line measurements were performed with two tech- towards Abell 1689. Spectroscopic redshifts mainly come niques: manually, with the Gaussian (cid:12)tting provided by from our MOS survey and from the spectroscopic survey the \splot" procedure in IRAF, and semi-automatically by Teagueet al. (1990).In addition, Dyeet al. (2001)de- with a purposely written program (MORPHS collabora- termined the photometric redshifts of severalhundreds of tion, Dressler et al. 1999). They agree well with each galaxies observed through an optimized set of narrow{ other, although the manual technique underestimates the band and broad{band (cid:12)lters. Not surprisingly, all his- 3 [TCG90] 006 and [TCG90] 217 have a redshift of resp. togramsshowastrongconcentrationofgalaxiesinthered- 0.0825 and 0.0862 instead of 0.1826 and 0.2153 as reported shiftrange0.17{0.22,presumablymembersofAbell1689. inTeagueetal.(1990).Giventhenumberofspectralfeatures, The peak is at zspectro = 0:184. While the large disper- we believe our redshifts are thecorrect ones. sionin the redshift distributionbased onthe photometric 66 P.-A.Ducet al.: Hidden star-formation in Abell1689 Table 5. Spectrophotometric data of cluster members from theMOS sample. ID ID ID flux([OII](cid:21)3727) −EQW([OII](cid:21)3727) EQW(H(cid:14)) Class Comments (MOS) (ISO) (mask) ergcm−2s−1 (cid:23)A (cid:23)A 1 { MOS5:17 { { { k: poor 2 { MOS2:9 4.67(cid:6)0.19 32.8(cid:6)1.7 4.8:(cid:6)1.0 e(c) H(cid:12)=−8:2,OIII=−14:0 3 { MOS3:20 { { { ? poor 4 { MOS1:12 { { 3.8:(cid:6)2.1 k+a: poor 5 { MOS1:6 { { { k: OIII=−1:5? 6 { MOS3:17 { { { k: poor 7 3 MOS5:11 0.60(cid:6)0.32 3.9(cid:6)1.6 { k(e) 8 .. MOS1:31 { { { k: poor 9 { MOS4:17 { { 3.8(cid:6)1.4 k+a 10 { MOS1:7 { { { k: 11 .. MOS1:36 2.73(cid:6)0.28 8.6(cid:6)1.0 2.3(cid:6)0.7 e(c:) H(cid:12)=−2:9,OIII=−9:4,OIII2=−2:7 12 { MOS3:19 { { { ? poor 13 { MOS2:6 { { { ? poor 14 4 MOS4:10 1.85(cid:6)0.27 18.8(cid:6)3.3 { e(c) H(cid:12)=−3:6 15 { MOS5:3 { { { k: 16 6 MOS5:9 2.36(cid:6)0.11 60.0(cid:6)3.0 { e(b) H(cid:12)=−12:8,OIII=−44:4 17 { MOS5:8 { { { k 18 { MOS5:7 { { { k: OIII/sky? 19 { MOS5:18 { { { k 20 { MOS2:22 { { { k: poor 21 8 MOS5:10 { { { k 22 { MOS4:21 { { { k: poor 23 10 MOS2:15 3.05(cid:6)0.23 10.4(cid:6)1.0 2.1:(cid:6)0.8 e(a)+ <H(cid:18)+H(cid:17)+H(cid:16)>=3=5:1,H(cid:14) prob.higher,H(cid:12)=−0:8 24 { MOS1:4 { <3.3 { k:(e:) H(cid:12)=2:0 25 { MOS3:13 { { { k: 26 { MOS4:19 { { { ? 27 { MOS4:18 { { 3.0:(cid:6)1.3 k+a: 28 12 MOS1:29 { { 5.5(cid:6)0.7 k+a 29 14 MOS2:13 { { { k 30 .. MOS2:25 { { 2.1(cid:6)0.9 k 31 { MOS5:21 { { { k: poor 32 { MOS4:14 { { 6.7:(cid:6)2.4 k+a: poor 33 16 MOS4:11 { { { k: poor 34 { MOS5:5 { { { k 35 { MOS5:12 0.23:(cid:6)0.09 7.4:(cid:6)2.3 { k(e:) 36 .. MOS1:35 0.53:(cid:6)0.18 2.1:(cid:6)1.1 4.2:(cid:6)1.7 k+a:(e) NIIinem? 37 18 MOS4:12 { <5.4 1.8(cid:6)0.7 k(e::) 38 { MOS4:3 { { 2.8(cid:6)1.1 k+a: 39 20 MOS1:15 { { 2.0:(cid:6)0.6 k 40 22 MOS3:12 { { { k: poor 41 { MOS2:7 { { { k 42 24 MOS1:11 { { { k 43 26 MOS1:20 { { { k 44 { MOS5:16 { { { k: 45 .. MOS3:22 { { { k: poor 46 30 MOS4:16 { { { k: 47 { MOS2:4 { { { k 48 { MOS3:5 1.50(cid:6)0.17 46.2(cid:6)7.0 { e(b) H(cid:12)=−6:7,OIII=−11:4,H(cid:11)=−43:1 49 31 MOS1:24 2.48(cid:6)0.26 7.1(cid:6)0.9 5.5(cid:6)0.8 e(a) <H(cid:18)+H(cid:17)+H(cid:16)>=3=5:2 50 { MOS1:25 { { { ? verypoor 51 32 MOS1:18 { { 1.1(cid:6)0.8 k 52 33 MOS2:16 { { 2.0(cid:6)0.5 k 53 .. MOS1:32 { { { k 54 35 MOS1:13 1.90(cid:6)0.17 8.1(cid:6)0.8 3.5(cid:6)0.7 e(a)+ <H(cid:18)+H(cid:17)+H(cid:16)>=3=5:5,Ha+NII=−3:1 55 37 MOS1:27 1.00(cid:6)0.30 3.9(cid:6)1.0 2.8(cid:6)1.4 e(a)+ <H(cid:18)+H(cid:17)+H(cid:16)>=3=5:7,Ha+NII=−13 56 .. MOS1:30 { { 2.3:(cid:6)1.6 k(e:) H(cid:11)=−2:3? 57 { MOS3:11 { { { k: poor 58 .. MOS1:2 0.18:(cid:6)0.07 2.5:(cid:6)0.9 3.4(cid:6)1.2 k+a(e:) 59 40 MOS3:9 { { { ? verypoor 60 .. MOS4:24 { { { k+a: earlBalsv.str(tobemeas) 61 41 MOS2:14 { { 3.5(cid:6)0.8 sey1: OIII=−5:1,broadHbem=−3:7,broadHa+NII(cid:3)2=−28:4 62 { MOS2:19 { { { ? poor 63 .. MOS3:3 .. .. { e(c) skyonOII,noH(cid:14),OIII=−16:8,H(cid:12)=−6:5,H(cid:11)=−9:1 64 { MOS1:5 { { { k: poor 65 .. MOS4:25 { { { k no 66 { MOS2:17 { { 2.5(cid:6)1.0 k 67 .. MOS3:4 { { { k(e) NII2=−4:8 68 .. MOS3:21 { { 4.1:(cid:6)2.2 k+a: 69 43 MOS2:10 .. .. 2.9(cid:6)1.0 e(c) H(cid:12)=1:4,Ha>−4:2,NII*2 70 .. MOS2:26 .. .. { k 71 .. MOS5:20 .. .. { k H(cid:12)=1:9 72 45 MOS3:16 .. .. { e(b) OIII=−11:7,H(cid:12)=−15:4,Ha>−42:7,NII2 73 .. MOS4:23 .. .. { k: 74 .. MOS1:21 .. .. { ? poor Notes: \.." means that the [OII]or H line is outside theMOS wavelength range. The equivalentwidth of a numberof some (cid:14) other lines are indicated in thecomments. P.-A.Ducet al.: Hidden star-formation in Abell 1689 67 Fig.1. Redshift histogram towards Abell 1689 in the redshift range z =0−1. Several samples are included: optically selected galaxies with photometric redshifts from the catalog by Dye et al. (2001) (light shaded bars) or spectroscopic redshifts from our MOS survey and from the literature (shaded bars); ISOCAM{selected galaxies with either spectroscopic or photometric redshifts (dark shaded bars). The inset shows a close up on the redshift range z = 0:17−0:22, that considered for cluster membership. OnlyISOCAM{selected galaxies with well determined spectroscopic redshifts are includedin thelatter plot. The x{axis is scaled in kms−1. technique results fromthe uncertaintyof the method, the based criteria to assess the cluster membership. In the substantial velocityspreadindicated bythe spectroscopic followinganalysis,weconsiderasclustermembersat large observations for galaxies related to A1689 is real (see in- allgalaxiesintheabove{mentionnedredshiftrange.They set in Fig. 1). It reflects both the richness of the cluster might be long standing members, galaxies just falling in and its complex velocity structure. Applying a multiscale the cluster or belonging to merging sub-clusters. Their analysis, Girardiet al. (1997) distinguished three distinct excess of infrared activity as detected by ISOCAM might groups in A1689 that overlap spatially but are well sepa- trace the broad cluster environment, or the cluster itself. ratedinvelocity.Hence,whatweseeinprojectiontowards A detailed analysis of the precise role of the environment the central region of this cluster (see Fig. 2) is probably goes beyond the limits of this study for lack of statistics composedofdi(cid:11)erentsubclumpsmergingalongthelineof and(cid:12)eld coverage.Our surveyonly extends out to a clus- sight. tercentric radius of 0.5 Mpc, therefore inferring a radial Figure1indicatesthattheISOCAM{detectedgalaxies dependence of the properties of the cluster members is also show a well-de(cid:12)ned peak in correspondence with the not possible. clusterbaricentricvelocity.Actually,morethan80%ofthe ISOCAM sources have redshifts between 0.17{0.22. This 3.2. Properties of the optically selected cluster prominent excess of mid{infrared emitters related to the members cluster is realand does not result fromselection e(cid:11)ects in the optical follow-up (see Paper I). However, the velocity In this section we present the photometric, spectroscopic distributionofISOCAMsourcesrelatedtoA1689isflatter and morphological properties of galaxies in our spectro- thanthatofthebulkoftheclusterpopulation(seeinsetin scopic sample, comparing them when possible with the Fig.1).Thisresultisactuallyexpected.Wewillshowlater properties of lower and higher redshift clusters. that most MIR emitters are associated with spiral and emission{line galaxies.Studies of nearby clusters indicate a larger velocity dispersion for this population than for 3.2.1. The optical subsamples the early-type galaxies (e.g. Biviano et al. 1997). The observation of a well de(cid:12)ned peak in the redshift The 74 cluster members in our MOS survey are listed histogramat z =0:17−0:22 both in the optically selected in Table 1. In the following analysis we include an- and mid-infrared selected samples provide a redshift{ other 17 galaxies for which redshifts are available in the 68 P.-A.Ducet al.: Hidden star-formation in Abell1689 blue fractionshave alsobeen reportedin other clusters at z =0:2(e.g.inAbell115,Metevieretal.2000),butthere is clearly a large spread in f among rich clusters at this B redshift (Smail et al. 1998). 3.2.3. Spectral and morphological classi(cid:12)cation Weclassi(cid:12)edallspectrausingthesetofrulesproposedby Dressler et al. (1999), which are based on the rest-frame equivalent widths of the [OII] emission line and the (cid:21)3727 H absorptionline4.Nospectraltypecouldbeassignedto (cid:14) the spectra of 8 cluster members due to their low signal to noise. Representative spectra for each class are presented in Fig. 3 and the histogram of the spectral types is shown in Fig. 4. At least 50% of the spectroscopically con(cid:12)rmed members have spectra of type \k" { they show no emis- sionlinesandonlyweakBalmerabsorptionlines{andare hence typical of passive early type galaxies. The \k+a" type, characterized by the presence of strong Balmer ab- sorption lines typical of post{starburst/post-starforming Fig.2. Spatial and velocity distribution of cluster members. galaxies,accountsfor10to15%ofourspectroscopicsam- Photometrically con(cid:12)rmedmembersareshown with thesmall ple.Thisfractionissigni(cid:12)cantlyhigherthanthattypically stars. Spectroscopically con(cid:12)rmed members are shown with circles the size of which is proportional to their velocity in estimated in nearby clusters ((cid:24)1% according to Dressler the cluster. Finally, the (cid:12)lled (red) circles correspond to the 1987). A direct comparison with higher redshift clusters ISOCAM 15 (cid:22)m sources. is hindered by the di(cid:11)erent completeness as a function of galaxy magnitude of the various spectroscopic samples. In clusters at z (cid:24) 0:5, the k+a fraction was found to literature(Teagueetal.1990).ThesearelistedinTable2. be about 20% (MORPHS collaboration, Dressler et al. Wealsomakeuseofthe167galaxieswithanRbandmag- 1999;Poggiantiet al. 1999). Using a sample with a mean nitudebrighterthan22,whichhaveaphotometricredshift redshift lower than the MORPHS, the CNOC1 group determined by Dye et al. (2001) compatible with a clus- (Baloghetal.1999)estimatedamuchlowerproportionof ter membership. Those detected by ISOCAM are listed post-starburstgalaxies, less than 5%. On the other hand, in Table 3. For reference, we indicate in Table 4 the fore- Abraham et al. (1996) counted in Abell 2390 a propor- groundandbackgroundgalaxiesfoundinourMOSsurvey. tion of H {strong galaxies as high as 23%, in good agree- (cid:14) mentwithAbell1689wheretheH {strongobjects(k+a’s (cid:14) and e(a)’s) amount to about 20%. The two clusters are 3.2.2. Blue fraction situated at the same redshift. Intotalabout20%oftheclustermembersinoursam- The value of the blue fraction, f , in Abell 1689, i.e. the B ple are emission line galaxies showing some level of star proportionofgalaxiesbluer than the color{magnitudere- formation in the optical. Conversely, most of the galaxies lation, is controversial. In their original article, Butcher in the cluster classi(cid:12)ed as \blue" { those responsible for & Oemler (1984) computed a value of f = 0:09(cid:6)0:03. B the photometric Butcher-Oemlere(cid:11)ect { exhibit emission However, later on Gudehus & Hegyi (1991) estimated a lines that are indicative of a current star formation. lower value for f (0:05(cid:6)0:06) consistent with the blue B The spectroscopic sample includes only one fraction measured in local clusters. Recently, f was re- vised again and raised to: 0:093(cid:6)0:019 (MargBoniner & Seyfert 1 galaxy, identi(cid:12)ed by the broad H(cid:11) emission de Carvalho 2000) and even 0:191(cid:6) 0:015 (Margoniner line. Usual diagnostics to distinguish Seyfert 2/LINERs from HII regions (Veilleux & Osterbrock 1987) could et al. 2001). Such a discrepancy can be due to the many not be applied because the required H hydrogen, [NII] observational biases, the fuzziness of the de(cid:12)nition of the (cid:11) nitrogen and [SII] sulfur lines were outside the spectral blue fraction(i.e.the choiceofthecolorindex,luminosity range,k-correctionmodel,locationofthecolor-magnitude 4 Wehaveintroducedanadditionalspectralclass,the\k(e)" diagram etc.) and especially the background subtraction. type, which represents spectra similar to the k-type but with Weestimatedf usingourdatasetofcon(cid:12)rmedcluster B signs of at least one very weak emission line. The de(cid:12)nition members(seedetailsinAppendixB).Ourcomputedvalue ofthe\e(a)"class was alsoslightly modi(cid:12)ed,asindicated and is1.5{2timesthatoriginallyderivedbyButcher&Oemler justi(cid:12)ed in Sect. 3.3.4. Finally, seven galaxies for which the (1984)forthisclusterandappearstobemuchhigherthan [OII] line was outside the MOS wavelength range were (cid:21)3727 in the local rich clusters studied by these authors. Large classi(cid:12)ed based on their Hydrogen Balmer lines. P.-A.Ducet al.: Hidden star-formation in Abell 1689 69 Fig.3.Representativeopticalspectraofclustermemberswithdi(cid:11)erentspectraltypes.Allspectraexceptthebottomonebelong to LW3-detected galaxies. From bottom to top, spectra are ordered by increasing 15 (cid:22)m flux.The MOS ID and spectral type are indicated to the right. Fig.4.SpectraltypehistogramofallMOS(lightshaded),6.75 (cid:22)mdetected(shaded,LW2)and15(cid:22)mdetected(darkshaded, LW3)clustermembersintheISOCAM(cid:12)eldofview.NotethatalmostallLW3sourceshavealsobeendetectedintheLW2(cid:12)lter. The unknown class \?" either includes galaxies for which the signal to noise was too low to determine a redshift or ISOCAM sources missed by our MOS survey. range of most galaxies. Instead, we have used the recent Finally, we note that all spectral types are evenly dis- diagnostics of Rola et al. (1997), based on lines in the tributedthroughoutour(cid:12)eldofview.Star{forminggalax- blue only. We found that no AGN activity was required ies are present even at low projected clustercentric radii. to explaintheopticalspectraofthe emissionlinegalaxies The (cid:12)eld-of-view and the sub-clustering and related pro- but the Seyfert 1. jection e(cid:11)ects (see Sect. 3.1) are likely to be the reasons