Mon.Not.R.Astron.Soc.000,1–??() Printed25January2010 (MNLATEXstylefilev2.2) The CoNFIG Catalogue - II. Comparison of Space Densities in the FR Dichotomy. 0 M. A. Gendre1⋆, P. N. Best2 and J. V. Wall1 1 0 1Department of Physics and Astronomy, The University of British Colombia, 6224 Agricultural Rd, Vancouver, BC, V6T 1Z1, Canada 2 2Institute for Astronomy, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, United Kingdom n a J Accepted .Received;inoriginalform 5 2 ABSTRACT ] This paper focuses on a comparisonof the spacedensities ofFRI andFRII sourcesat O different epochs, with a particular focus on FRI sources. C First, we present the concluding steps in constructing the Combined NVSS-FIRST . Galaxy catalogue (CoNFIG), including new VLA observations,optical identifications h and redshift estimates. The final catalogue consists of 859 sources over 4 samples p - (CoNFIG-1, 2, 3 and 4 with flux density limits of S1.4GHz =1.3, 0.8, 0.2 and 0.05 Jy o respectively).Itis95.7%completeinradiomorphologyclassificationand74.3%ofthe r sources have redshift data. t s Combining CoNFIG with complementary samples, the distribution and evolution of a FRI and FRII sources are investigated. We find that FRI sources undergo mild evo- [ lution and that, at the same radio luminosity, FRI and FRII sources show similar 1 space density enhancements in various redshift ranges, possibly implying a common v evolution. 4 Key words: Surveys - Radio Continuum: Galaxies - Galaxies: Active - Galaxies: 1 5 Statistics - Galaxies: Luminosity Function 4 . 1 0 0 1 INTRODUCTION theother havebeen observed (Capetti et al. 1995). 1 : Radio AGN are classified in various ways such as luminos- v InaninitialmodelingofthespacedensityofradioAGN, ity, spectral type or morphology. The Fanaroff-Riley (FR) i Wall & Jackson(1997)andJackson & Wall(1999)assumed X classification (Fanaroff & Riley 1974) provides a classifica- that the cosmic evolution of radio loud AGN was based tion of extended radio sources. The FRI objects have the r on a division of the radio sources into a high-luminosity a highest surface brightness along the jets and core, reside component (P178−MHz > 1025WHz−1sr−1) corresponding in moderately rich cluster environments (Hill & Lilly 1991) to FRIIs and a low-luminosity component showing no cos- and include sources with irregular structure (Parma et al. mic evolution, corresponding to FRIs. With the advent of 1992). In contrast, FRII sources show the highest surface large-scale redshift surveys for nearby galaxies, many au- brightness at the lobe extremities, as well as more colli- thors,includingSnellen & Best(2001),Willott et al.(2001), mated jets, are found in more isolated environments and Sadler et al.(2002)andRigby,Best & Snellen(2008),found generally display stronger emission lines (Rawlings et al. significantevolutionforlowpowersources–butmildevolu- 1989;Baum & Heckman 1989). tionincomparisonwiththatofthehigh-luminositysources. The FRI/FRII dichotomy is based purely on the ap- Rigby,Best & Snellen(2008)arguedthatifFRIsandFRIIs pearanceoftheradioobjectsand,althoughsomehypotheses have similar evolution, the dual-population scheme could exist(e.g.Bicknell1995),themechanismsdifferentiatingthe be reduced to a single-population model. Their sample was two populations are still unknown.If sources with different however confined to a small number of low flux density FR classes undergo different evolution, this might imply sources. that their fundamental characteristics, such as the black A dedicated study and comparison of FRI and FRII holespinorjetcomposition, aredifferenttoo. However,the sources and their evolution, using large samples of sources cut between FRI and FRII is somewhat ambiguous: hybrid of each type, is the key to understandingthese populations sources showing jets FRI-like on one side and FRII-like on anddeterminingiftheFRclassificationisvalidorifadiffer- ent classification, such as whether they display high or low excitation emission lines, is physically more relevant. ⋆ E-mail:[email protected] Afurthermotivationforstudyingthecosmicbehaviour 2 Melanie A. Gendre, P. N. Best and J. V. Wall of radio AGNs is to assess their contribution to feedback themorphology-dependentluminositydistributionsandthe processes.Thecurrentparadigmforgalaxyformation, hier- FRI/FRII source counts, as well as cosmic evolution of the archical build-up in a Cold Dark Matter (CDM) universe, radio luminosity function. Finally, §7 summarizes the re- impliesthatthemost massivegalaxies inthelocal Universe sults. oughttobethelargest andbluestandhavethehigheststar Throughout this paper, we assume a standard ΛCDM formingrateofallgalaxies.Yet,observationsshowthatthey cosmology with H0=70 km s−1 Mpc−1, ΩM=0.3 and are red, old galaxies, and the bulk of star formation is ob- ΩΛ=0.7. served at earlier epochs. This is known as downsizing, first described byCowie et al. (1996). AGN negative feedback, in which the ignition of the nucleus in a star-forming galaxy ejects the gas into the 2 THE CONFIG CATALOGUE inter-galactic medium, is a possible way to understand this phenomenon. AGNs jets could indeed be responsi- 2.1 Catalogue definition ble for reducing or even stopping star formation, breaking The CoNFIG Catalogue consists of 4 samples, CoNFIG-1, the hierarchical buildup (Silk & Rees 1998; Granato et al. 2, 3 and 4, which include all sources selected from the 2001; Quilis et al. 2001). AGN can also have a posi- NVSS catalogue with S1.4GHz >1.3, 0.8, 0.2 and 0.05 Jy tive feedback effect, whereby pressure from the jets com- respectively in defined areas (see Fig. 1 and Table 1). presses the inter-stellar medium and induces star forma- The NRAO-VLA Sky Survey (NVSS; Condon et al. tion(vanBreugel et al.2004;Klamer et al.2004).However, 1998) is a 1.4-GHz continuum survey covering the entire modelingofjetpoweranditsrelationtostarformationhave skynorthofδ =−40◦ (correspondingtoanareaof10.3sr). shown that theoverall effect is adecrease in star formation The completeness limit is ∼2.5 mJy/beam with an rms of rate (Antonuccio-Delogu & Silk 2008). ∼0.45 mJy/beam. The catalogue from the survey contains If AGN feedback from the heating and ejection of gas over 1.8 million sources, implying a surface density of ∼50 in the ISM possibly suppresses star formation, it is reason- sources per square degree. It was carried out with theVery abletothinkthat it shouldberelated totheenergyoutput Large Array (VLA) in D and DnC configuration (the D from the jets. Best et al. (2006) studied the output energy configuration being the most compact VLA configuration from AGNs and concluded that heating dissipated in the with a maximum antenna separation of ∼1 km), providing host galaxy is dominated by low-luminosity radio sources, an angular resolution of about 45 arcsec FWHM. which tend to be confined predominantly to the size of the Since the median angular size of faint extra-galactic galaxy and its halo. Such sources also stay ‘on’ for a longer sources at the CoNFIG flux density levels is .10 arcsec period of time than high luminosity sources, allowing heat (Condon et al.1998),mostsourcesinNVSSareunresolved, tobesuppliedpseudo-continuously.Schawinskiet al.(2009) and the flux density measurements are quite accurate. investigated the relation between the amount of molecular However, the large beam size does not reveal precise gas and AGN activity in galaxies and concluded that a low structureofsourcesordeterminepositions accurateenough luminosity AGN episode was sufficient to suppress residual to establish unambiguous optical counterparts. star formation in early typegalaxies. Very large sources resolved in NVSS within Establishing the potential space-density behaviour of the initial samples, such as a few 3CRR sources radio AGN is thus important in studying the precise role (Laing, Riley & Longair 1983), need to be considered. of the feedback mechanisms. Could feedback be linked to In some of these cases, two or more NVSS ‘sources’ with source morphology as it is to luminosity? Do FRI sources S1.4GHz> Slim are actually components of a much larger have a higher impact on star formation rate than typically resolved source. Multi-component sources in which each more powerful FRII sources? component has S1.4GHz<Slim but with a total flux density S1.4GHz>Slim,alsoneedtobeconsidered.Forthispurpose, The lack of a large comprehensive catalogue of NVSSsources with S1.4GHz<Slim were selected and, if any morphologically-classified radio sources has been a limiting other source in the catalogue was located within 4 arcmin factor in all these studies. This is the goal of the CoNFIG of the listed source, the combination was set aside as a catalogue, which we propose to use in modeling the radio candidate extended source. The final decision on whether luminosity function of AGN. or not the sources were actually components of a resolved This is the second paper in a series studying extended source was made by visual inspection of the NVSS contour radio galaxies and their role in AGN feedback. Paper I plots. (Gendre& Wall 2008) outlined the initial construction of AsummaryofeachsampleisgiveninTable2.Because the Combined NVSS-FIRST Galaxy (CoNFIG) sample. the area of the CoNFIG-2, 3 and 4 samples overlap with This paper describes the complete catalogue, including op- CoNFIG-1, all statistics estimated from CoNFIG 2, 3 and tical identifications and redshift estimates, as well as a pre- 4 useonly sources with Slim < S1.4GHz< 1.3 Jy. liminary study of FRI and FRIIspace densities. The structure of this paper is as follows. The con- struction of the catalogue is explained in §2 while §3 2.2 Spectral indices describes how the morphologies were determined. Opti- cal identifications and redshift information are discussed Inordertocomputetheradioluminosity,thespectralindex in §4 and an overall summary of the catalogue is given α (defined as Sν ∝ να) of each source needs to be deter- in §5, along with the introduction of complementary mined. To achieve this, flux densities at different frequen- datasets that will be used in the modelling. §6 describes cies for each source were compiled and the spectral index The CoNFIG Catalogue II. 3 Figure 2. FIRST contour plot of a characteristic example of an FRII source, 3C 223. The hot spots are located at the endsofthealignedlobes. Figure 1.Mapofthesampleregions.Eachsampleislocated in the North field of FIRST (grey area). CoNFIG-1 (red contour), CoNFIG-2 (green hatched), CoNFIG-3 (blue diagonally cross- hatched)andCoNFIG-4(pinkverticallycross-hatched)haveflux densitylimitsof1.3,0.8,0.2and0.05Jyrespectively. Definition oftheregionsanddetailsofthesamplescanbefoundinTables1 and2. Table 1. Region corners for the CoNFIG samples ( RA; DEC { } in HRS;deg )asshowninFig.1. { } Figure 3. FIRST contour plot of a characteristic example C-1 C-2 C-3 C-4 of an FRI source, 3C 272.1 (M84). The regions of highest 17.7;64.0 9.30;60.0 14.7;30.0 14.1; 3.0 surfacebrightnessarelocatedalongthejets. { } { } { } { } 7.0;64.0 13.35; 60.0 16.0;30.0 14.7; 3.0 { } { } { } { } 7.3;30.0 13.35; 5.0 16.0;10.0 14.7; 3.5 { } { − } { } { − } 17.3;24.8 9.30; 5.0 14.7;10.0 14.1; 3.5 by Gendre & Wall (2008), primarily by looking at FIRST { } { − } { } { − } 15.5; 8.0 { − } and NVSScontour plots. 9.1; 8.0 { − } TheFaintImagesoftheRadioSkyatTwenty-cmsurvey (FIRST; White et al. 1997) is another 1.4-GHz continuum survey with the VLA, covering an area of ∼9030 deg2 in- computed following the relation: cluding the North Galactic Pole. The completeness limit is ∆log(S) α= (1) ∼1 mJy/beam with a typical rms of 0.15 mJy/beam. The ∆log(ν) surveyyielded∼811,000 sources,implyingasurfacedensity A summary of the different frequencies and corresponding of ∼90 sources per square degree. It was carried out in B surveys used to retrieve the flux density information can configuration (the B configuration having a maximum an- be found in Gendre& Wall (2008). We were able to com- tenna separation of ∼10 km), which provides an angular putethelow-frequencyspectralindex(with178MHz6ν 6 resolution of about 5 arcsec FWHM. This survey comple- 1.4GHz) for99.6%, 97.7%, 89.3% and52.7% ofthesources ments the NVSS survey well, providing a beam size small in CoNFIG-1, 2, 3 and 4 respectively. enoughtoresolvethestructureofmostnearbyextendedra- dio sources and source positions to better than 1 arcsec to enable cross-waveband identification. If the FIRST/NVSS contour plot displays distinct hot spots at the edge of the lobes (as in Figure 2), and the 3 MORPHOLOGY lobesarealigned, thesourcewas classified asFRII.Sources 3.1 Initial classification withcollimatedjetsshowinghotspotsnearthecoreandjets wereclassifiedasFRI(seeFigure3).Wideangletailsources Theinitialmorphologiesweredeterminedeitherfromprevi- (WAT;Leahy1993)aswellasmostirregular-lookingsources ously referenced work or following the procedure described (Parma et al. 1992) were also classified as FRI. Sources of size smaller than 1 arcsec or previously classified as QSOs Table2.CharacteristicsoftheCoNFIGsamples,asdescribedin wereclassifiedas‘compact’whileextendedsourcesforwhich section2. the FRI/FRII classification was impossible to determine were classified as ‘uncertain’. Slim Area #of #not (Jy) (deg2 ) sources inC1 C-1 1.30 4924 273 - 3.2 VLA observations C-2 0.80 2915 243 132(54.3%) C-3 0.20 370 286 270(94.4%) InadditiontotheobservationsdescribedbyGendre& Wall C-4 0.05 52 185 184(99.4%) (2008),radioobservationsof213extendedCoNFIGsources with previously uncertain morphological classification were 4 Melanie A. Gendre, P. N. Best and J. V. Wall made at 1.4-GHz using the VLA in A configuration. These Table3.MorphologyofthesourcesintheCoNFIGsamples.The observations included polarization measurements for 31 morphologyofeachsourcewasdeterminedbylookingatFIRST sources as preliminary work for a possible study of the andNVSS contour plots or fromVLA observations as described morphology-dependent polarized source count. in 3.1and 3.2.Sourcesofsizesmallerthan3arcsecorpreviously § § The A-configuration, the most extended VLA configu- classifiedasQSOswereclassifiedas‘compact’(C)whileextended rationwithamaximumantennaseparationof∼36km,pro- sources for which the FRI/FRII classification was impossible to vides a synthesized beam of 1.4 arcsec FWHM at 1.4-GHz. determine were classified as ‘uncertain’ (U). In each cases, the correspondingpercentage ofsampleisgiveninitalic. Threefrequencybandswereused:(1)twoIFsof1464.9 and 1385.1-MHz, with a bandwidth of 50-MHz (2) two IFs of 1372.5 and 1422.5-MHz, with a bandwidth of 25-MHz and (3)twoIFsof1425.5 and1397.5-MHz, withabandwidthof C-1 C-2 C-3 C-4 Tot. % of sample 25-MHz. Frequency bands 2 and 3 were used in the polar- ization measurements. FRI 25 7 22 17 71 The exposure time was computed for each source such 9.2 5.3 8.1 9.2 8.3 as to provide a signal-to-noise ratio of at least 5, and FRII 149 75 152 90 466 the exposures were split into two or three separate inte- 54.6 56.8 56.3 48.9 54.2 grations to improve uv coverage. The primary calibrator C 86 47 88 64 285 31.5 35.6 32.6 34.8 33.2 3C286 (1331+305) was observed several times during the U 13 3 8 13 37 run. Nearby secondary calibrators were observed approxi- 4.8 2.3 3.0 7.1 4.3 mately every 30 min to provide phase calibration. All data werereducedusingstandardproceduresincorporatedwithin theAIPS software provided by NRAO. 4 OPTICAL IDENTIFICATIONS AND REDSHIFTS A preliminary search for counterparts was performed using the unified catalogue of radio objects of Kimball & Ivezi´c (2008)1, and optical identifications were obtained, princi- pallyfromtheSloanDigitalSkySurvey(SDSS; York et al. 2000). TheSDSS,withthe2.5metertelescopeatApachePoint 3.3 Final classification Observatory,NewMexico,hasimagedonequarteroftheen- tire sky in ugriz magnitudes2, as well as performing a spec- 62.5% of sources in the CoNFIG sample were classified troscopic redshift survey. The seventh data release (DR7; eitherasFRI(I)orFRII(II).Followingtheunifiedmodelof Abazajian et al. 2009) imaging survey contains a total of AGN(Jackson & Wall1999)core-jet sourceswereclassified 357millionobjectsover11,663deg2 whilethespectroscopic as FRII. Hybrid sources, showing jets FRI-like on one side surveycontains 1.6 million objects over 9380 deg2 . and FRII-like on the other (Capetti et al. 1995), were clas- sifiedaccordingtothecharacteristicsofthemostprominent Ks band photometric information was obtained from 2MASS. The Two Micron All Sky Survey (2MASS; jet. Extended sources for which FRI/FRII identification Skrutskieet al. 2006) is a near-infrared survey using 1.3 m was ambiguous were classified as uncertain (U). Sources telescopesatMountHopkinsinArizonaandCTIOinChile. with size smaller than 3 arcsec were classified as compact (C) or (C*), depending on whether or not the source was It aimed at imaging the entire sky in J, H and KS mag- nitudes. The now-complete catalogue, divided into a point confirmed compact from the VLBA calibrator list (see source and an extended source (semi-major axis >10” in Beasley et al. 2002; Fomalont et al. 2003; Petrov et al. size) catalogue, contains 472 million sources over 99.998% 2006; Kovalev et al. 2007) or the Pearson-Readhead survey of thesky. (Pearson & Readhead 1988). Finally, sources of type (S*) Asummaryofthenumberofidentifiedopticalcounter- correspond to confirmed compact sources which show a steep (α 6 −0.6) spectral index. These are probably parts is given in Table 4. compact steep-spectrum (CSS) sources. The final classification for each source is shown in 4.1 Spectroscopic and photometric redshifts Appendix A and the distribution of morphological types is presented in Table 3. Contour plots of extended sources, Spectroscopic redshifts were obtained for 45.5% of the cat- including the VLA observation presented in §3.2, are alogue(seeTable5)usingeithertheSIMBAD3 databaseor presented in Appendix B. In order to study the evolution theSDSSDR7catalogue. of the space density of FRI and FRII sources accurately, each extended source was assigned a sub-classification - confirmed (c) or possible (p) - depending on how clearly 1 http://www.astro.washington.edu/akimball/radiocat/ 2 The limiting magnitudes at the detection limit given in thesource showed either FRI or FRII characteristics. Abazajianetal. (2009) correspond to a 95% detection repeata- The complete catalogue consists of 859 sources, with bility for point sources. However, for galaxies, these are typ- 71 (8.3%) FRIs (50 confirmed, 21 possible), 466 (54.2%) ically between half a magnitude and a magnitude brighter FRIIs (390 confirmed, 76 possible), 285 (33.2%) compact at the same signal to noise ratio (from SDSS project book: sources and 37 (4.3%) uncertain sources. http://www.astro.princeton.edu/PBOOK/camera/camera.htm) 3 http://simbad.u-strasbg.fr/simbad The CoNFIG Catalogue II. 5 Figure 4. The SDSS magnitude-redshift relations were computed by finding the best fit (solid lines) to data from CoNFIG non-QSO sources having both spectroscopic redshift and SDSS magnitude information (dots). The relations (Equ.2-4) were used to estimate photometricredshiftsforsourcesnotinthephotoz2 catalogue, butwithanSDSScounterpart. Table4.NumbersofSDSSand2MASSopticalidentificationsfor shifts and KS-band information from the2MASS extended theCoNFIGsamples.Ineachcases,thecorrespondingpercentage and point source catalogues. ofsampleisgiveninitalic. log(z)=−3.515+0.204KS 2MASS extendedsources (5) log(z)=−4.800+0.279KS 2MASS point sources (6) SDSS 2MASS The relations, shown in Fig. 5, provide good estimates of All FRI FRII C U All % of sample redshifts up to KS = 15.5. They were used to estimate photometric redshifts for 6 sources which had no SDSS C-1 233 25 125 73 10 117 spectroscopic or photometric redshifts available but had 85.3 100 83.9 84.9 76.9 42.9 2MASS counterparts (KS 615.5). C-2 108 6 62 37 3 44 81.8 85.7 82.7 78.7 100 33.3 Overall, 74.3% of the sources in the CoNFIG cata- C-3 190 20 111 53 6 47 loguehavespectroscopicorphotometricredshiftinformation 70.4 90.9 73.0 60.2 75.0 17.4 C-4 110 17 52 37 4 22 available,withmeanandmedianredshiftsofzmean =0.714 59.8 100 57.8 57.8 30.8 12.0 and zmed = 0.588. The redshift distributions, by samples and morphological types, is shown in Fig. 6. Tot. 641 68 350 200 23 230 74.6 95.8 75.1 70.2 62.2 26.8 4.2 Sources with no redshift information A total of 221 sources in the CoNFIG catalogue, mostly in Because redshift information is essential to computing CoNFIG-3 and 4, have no redshift information available. spacedensitiesandexaminingtheirevolution,weestimated 104 of these sources are of morphological type I, II or redshifts for sources with no spectroscopic data available. U (we will ignore sources of type C for the time being, For a number of sources with an SDSS counterpart being only interested in the study of extended radio identified but with no spectroscopic information available, sources). One way to include these sources in the space photometricredshiftswereretrievedfromtheSDSSphotoz2 density modelling is to assign an estimated redshift to catalogue (Oyaizu et al. 2008), which covers SDSS galaxies each of them (by the procedures described below), com- with r622.0. For other galaxies (excluding the 285 sources putetheRLF,repeattheprocedureandaveragetheresults. identified as ‘compact’, which are most likely QSOs), red- shifts were estimated using a magnitude-redshift relation- BasedonSDSSnon-detectionwecandeterminealower ship computed from SDSS-identifiedCoNFIG non-compact redshiftlimit forthesesources. Thei-bandbeingeffectively (i.e. non-QSO)sources with spectroscopic redshifts: the deepest SDSS band for objects with the typical colours log(z)=−3.599+0.170i (2) ofhigh-redshiftradiogalaxies,Equ.2wasusedtodetermine log(z)=−3.609+0.175z (3) the lower limit, yielding a value of zlim ≃ 1.0. To account for the spread in the i-z relation, the estimate of the limit log(z)=−3.660+0.169r (4) was drawn randomly from a Gaussian of variance 0.1, TherelationsareshowninFig.4andwereusedtoestimate centered on zlim=1.0. photometric redshifts for 73 sources. Our next step was to use the (admittedly naive) KS-zrelationswerealsoobtainedusingdatafromCoN- assumption that the redshift of the radio source could be FIG non-compact sources having both spectroscopic red- estimated from the distribution of measured or estimated 6 Melanie A. Gendre, P. N. Best and J. V. Wall Table 5. Distribution and ranges of redshifts for the CoNFIG samples. Spectroscopic redshifts are retrieved either from the SIMBADdatabaseorfromtheSDSScatalogue.Photometricred- shiftsareeitherobtainedfromtheSDSSphotoz2catalogueores- timatedusingeithertheSDSSmag-zrelationdefinedbyEqu.2-4 orthe KS-zrelationdefined byEqu. 5-6. Ineach cases, the cor- respondingpercentage ofsampleisgiveninitalic. C-1 C-2 C-3 C-4 Total Number of Sources in Sample 273 132 270 184 Redshift types % of sample Figure 5. The KS-zrelation was computed by finding the best Spectro. 226 67 54 44 fit(solidpinkanddot-dashedredlinesrespectively)todatafrom 82.8 58.8 20.0 23.9 CoNFIGnon-QSOsourceshavingbothspectroscopicredshiftand Photo. photoz2 29 33 71 35 KS-bandinformationfromthe2MASSextended(bluetriangles) 10.6 25.0 26.3 19.0 andpointsource(orangedots)catalogues.Therelations(Equ.5- sdssmag-z 5 13 38 17 6) were used to estimate photometric redshifts for sources with 1.8 5.3 13.3 9.2 a magnitude KS 6 15.5, which corresponds to an upper esti- KS-z 3 1 2 0 mated redshift limit of z=0.43 (dotted lines) from the extended 1.1 0.8 0.7 0.0 source relation. For comparison, the K-z relations from CEN- SORS (Brookes etal. 2006) and (Willottetal. 2003) are shown Total 263 114 165 96 inlightanddarkgreydashedlinesrespectively. 96.3 86.4 61.1 52.2 FRI 25 7 21 17 100 100 95.4 100 distribution for thesource. FRII 145 65 112 52 To complete the catalogue redshift distributions, we 97.3 86.7 73.7 57.8 determinedthateachsourcewithnoredshiftwillcontribute C 80 39 26 23 a fraction to each redshift bin, following its assigned 93.0 83.0 29.5 35.9 probability distribution. For space densities computation U 13 3 6 4 100 100 75.0 30.8 (see §6), approximated redshifts were assigned to each source by making random realizations following the proba- Redshift ranges C-1 C-2 C-3 C-4 bility distribution, repeating the process in a Monte-Carlo manner. min. 0.003 0.011 0.018 0.006 All max. 3.530 2.707 2.408 2.677 Because most of the approximate redshifts are greater mean 0.711 0.760 0.623 0.828 than z=0.3, the redshift upper-limit used to define the med. 0.555 0.599 0.564 0.695 local universe, the results of the local radio luminosity min. 0.003 0.011 0.032 0.006 function (LRLF) are completely unaffected by redshift FRI max. 0.269 0.309 1.847 1.531 uncertainties. As the redshift lower-limits used in the mean 0.071 0.128 0.264 0.261 computationoftheapproximateredshiftsaremostlyz>1.0, med. 0.049 0.099 0.116 0.150 resultsouttoz∼1.0arealso notsignificantlyaffected.Over min. 0.036 0.098 0.062 0.138 the range 1.06z62.0, the results are likely to be impacted. FRII max. 2.183 1.711 2.408 2.677 Nevertheless, the fact that the redshift distribution is mean 0.637 0.660 0.674 0.938 well determined over that range implies that the impact med. 0.523 0.566 0.604 0.800 is perhaps not severe. Beyond z=2.0, results would be unreliableastheredshiftdistributionisnotwelldetermined min. 0.034 0.160 0.018 0.133 C max. 3.530 2.707 1.764 2.235 and the use of approximate redshifts may have introduced mean 1.024 1.050 0.665 1.026 significant biases. med. 0.880 0.795 0.580 0.725 redshifts for sources of similar flux density. For each of 5 CATALOGUE SUMMARY AND the 113 sources, we derived the sample of sources with COMPLEMENTARY SAMPLES redshift information available and flux densities within the range of a tenth to ten times the flux density of the The CoNFIG catalogue (Appendix A1) consists of 859 source with no redshift. The redshift distribution of this sources over 4 samples, CoNFIG-1, 2, 3 and 4 with flux sample wascomputed andfitwith apolynomial; theregion densitylimitsS1.4GHz=1.3,0.8,0.2and0.05Jyrespectively. of this polynomial above the calculated redshift limit was Spectral indices were computed for 86.0% of the sources then normalized to determine the redshift probability using flux densities at different frequencies for each source. The CoNFIG Catalogue II. 7 The catalogue is 95.7% complete for radio morphologies of the sample members have morphological classification, and 74.3% complete for redshift information. including57 FRI,18 FRIIand 6 uncertain sources. Sources were morphologically classified into 6 cat- The final list, including the complementary samples, egories, using NVSS, FIRST and VLA 1.4GHz A- contains 1114 sources and is 75.9% complete for redshift configurationobservationcontourplotsaswellaspreviously information and 94.2% complete for radio morphologies. referenced information. Sources of typeI and II correspond It includes a total of 136 FRI (78 confirmed, 58 possible) to Fanaroff & Riley (1974) morphologies; extended sources and 571 FRII (477 confirmed, 94 possible) sources, mak- for which FRI/FRII identification was uncertain were ing it one of the largest, most comprehensive databases of classified as typeU; sources with size smaller than 1 arcsec morphologically-classified radio sources and an important were classified asC typeorC*-type,dependingon whether tool in thestudy of AGN space-densities. or not the source was confirmed compact; sources of S* type correspond to confirmed compact sources which show a steep spectral index. OpticalcounterpartswereobtainedfromtheSDSSand 6 SOURCE STATISTICS AND EVOLUTION 2MASS catalogues for 74.6% and 26.8% of the sources re- The main goal of CoNFIG is to produce a comprehensive spectively. Spectroscopic redshift information was retrieved catalogue of morphologically-classified radio sources to be from SDSS and the SIMBAD database, while photometric used in the modeling of the radio luminosity function of redshifts (or redshift estimates) were compiled from the AGN,in orderto investigate theirevolution and therole of SDSS photoz2 catalogue, or using the KS-z or SDSS mag-z the different types in feedback processes. For this purpose, relations (Equ.2-6). wecomputedtheluminositydistributionsandsourcecounts basedonmorphologicalclassification,tobeusedintheRLF To improve the flux density coverage of the catalogue, modeling. three complementary samples were appended (Appendix A2, A3and A4): 6.1 Luminosity distribution and the P-z plane • The 3CRR (Third Cambridge Revised) cata- The luminosity distribution is computed for each morpho- logue (Laing, Riley & Longair 1983) is complete to logicaltype(FRI,FRII,CandU)forsourceswithavailable S178MHz=10 Jy and contains 173 sources over an area of redshift information, using the 1.4-GHz flux density and 4.2sr.TheconversionfromS178MHz toS1.4GHz withα=0.8 spectral index values of each source. When the latter was yield a flux density limit of S1.4GHz≈1.92 Jy. In order to unavailable, a value of α=−0.8 was used. This introduced maximize the completeness of the sample at 1.4-GHz, we a minimal bias in the results, since extended sources in the increased the flux density limit to S1.4GHz=3.5 Jy. The CoNFIG samples have a median spectral index of −0.75 compiled spectral indices were used in the conversion and less than 6% of them have α > −0.5. Finally, sources for each 3CRR source. After excluding sources already with no redshift information were included, with redshifts present in the CoNFIG samples, 38 sources were selected as estimated in §4.2, and the resulting distributions are to complement the CoNFIG catalogue. All sources were shown in Fig. 7. morphologically classified, either using the classification of Laing, Riley & Longair (1983) or following the method A wide coverage of the P-z plane is essential to any described in §3, and the sample includes 8 FRI, 24 FRII modeling of the radio luminosity function (Rawlings 2002). and 6 compact sources. The combination of CoNFIG, 3CRR, CENSORS and the • The CENSORS (Combined EIS-NVSS Survey Of Ra- Lynx&Herculessamples coversalargerangeofluminosity dio Sources) sample (Best et al. 2003) is complete to andredshift(Fig. 8and9),providingapowerfulbasisfrom S1.4GHz=7.2 mJy and contains 136 sources selected from which to study FRI and FRII sources. NVSS over the 6 deg2 of the ESO Imaging Survey (EIS) Patch D. The sample has spectroscopic redshifts for 68% 6.2 Source counts ofthesources,andopticalornear-IRidentifications(giving redshift estimates) for almost all of the remainder. The morphologically-dependent source counts (Fig. 10) LittleradiomorphologicalclassificationoftheCENSORS were compiled as described by Gendre & Wall (2008),from sources has been done as the image resolution is often not theCoNFIG and complementary samples. highenoughtoidentifythesourcemorphology.Forthisrea- AsseeninFig.7,uncertainsources(whichareextended son, the VLA observation program described in §3.2 also but uncertain to whether they are FRI or FRII) have a included 40 CENSORS sources, allowing us to morpholog- luminosity distribution closer to that of FRII sources than ically classify 84.5% of the CENSORS sources. The sam- FRI sources. Thus, we make the assumption to include ple includes13 FRI,64 FRII,38 compact and 21 uncertain uncertain sources into the FRII morphology group for the sources. source count. This inclusion does not make any significant • The Lynx & Hercules sample (Rigby,Snellen & Best change from thesource count of FRII sources only. 2007) is complete to a catalogue flux limit of S1.4GHz=0.5 mJy ,from radio images with inital flux The FRII sources dominate the total count, except at density limits of 0.07-0.09 mJy/bm It contains 81 sources lowfluxdensities(logS1.4GHz.−1.6),wheretheFRIsources withinanareaof0.6deg2 .Itiscompleteinredshiftestima- suddenlytakeover,constituting asignificant portion of the tion (49% spectroscopic and 51% photometric) and 95.6% mJyandsub-mJysourcesincontrasttoFRIIsources.Since 8 Melanie A. Gendre, P. N. Best and J. V. Wall Figure 6. Redshift distribution of the sources in the CoNFIG catalogue for each morphological type. Sources with spectroscopic, photometric photoz2, KS-z estimated and SDSS mag-z estimated redshifts are represented by the red solid, blue cross-hatched, green solid and purple diagonally hatched columns. The estimated contribution from sources with no redshift information available ( 4.2) is § showninblackverticallyhatched columns. Table 6.CoNFIG-1DataTable(example) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) 1 071338.15 +434917.20 B0710+439 2011.4 0.82 C* 0.5180 0.0010 S 071338.10 +434917.00o 2 071424.80 +353439.90 B0711+35 1467.1 0.41 C* 1.6260 S 071424.82 +353439.80o 3 071641.09 +532310.30 4C53.16 1501.4 0.63 II p 0.0643 0.0001 S 071641.21 +532309.60 10.2e − 4 073555.54 +330709.60 4C33.21 2473.1 0.56 C 0.7010 0.0939 P 073555.57 +330709.59 21.9 21.1 20.5 19.6 19.7 − 5 074110.70 +311200.40 J0741+3111 2284.3 0.38 C* 0.6300 0.0014 S 074110.71 +311200.22 17.0 16.5 16.6 16.6 16.7 14.0 6 074542.13 +314252.60 4C31.30 1357.8 0.55 II c 0.4608 0.0004 S 074541.67 +314256.70 15.7 15.5 15.6 15.4 15.3 12.9 7 074948.10 +555421.00 DA240 1660.4 −0.44 II c 0.0360 0.0001 S 074834.70 +554859.00o 8 075828.60 +374713.80 NGC2484 2717.9 −0.68 I c 0.0408 0.0002 S 075828.11 +374711.87 15.7 13.8 12.9 12.5 12.2 9.9e − 9 075947.26 +373850.20 4C37.21 1691.2 0.84 II c − 10 080135.32 +500943.00 TXS0757+503 1471.7 1.02 II p 0.4855 0.0419 P 080135.35 +500943.99 22.2 21.6 20.6 20.0 19.3 11 080531.31 +241021.30 3C192 5330.6 −0.67 II c 0.0600 S 080535.00 +240950.36 18.1 16.2 15.4 14.9 14.6 12.4e 12 081003.67 +422804.00 3C194 2056.6 −0.78 II c 1.1840 S 081003.60 +422804.00o − 13 081259.48 +324305.60 4C32.24 1522.5 0.68 II c 0.4306 0.0047 I 08130.27 +324243.71 22.0 20.7 19.8 19.0 18.4 − 14 081336.07 +481301.90 3C196 15010.0 0.75 II c 0.8710 S 081336.07 +481302.64 18.6 17.9 17.7 17.5 17.3 14.8 − 15 081947.55 +523229.50 4C52.18 2104.2 0.62 II c 0.1890 S 081947.51 +523227.13 20.7 19.0 17.9 17.4 17.0 16 082133.77 +470235.70 3C197.1 1787.1 −0.75 II c 0.1280 0.0012 S 082133.61 +470237.36 19.1 17.7 16.8 16.3 16.0 13.5e − 17 082144.02 +174820.50 4C17.44 1875.1 0.57 C 0.2960 0.0002 S 082144.02 +174820.30 21.0 19.6 18.1 17.6 17.2 14.9 − 18 082324.72 +222303.70 4C22.21 2272.4 0.34 C* 2.2103 0.0013 S 082324.76 +222303.30 20.3 19.8 19.3 18.9 18.5 15.4 − 19 082447.27 +555242.60 4C56.16A 1449.4 0.25 C* 1.4181 0.0016 S 082447.24 +555242.71 18.2 18.1 17.9 17.8 17.8 − 20 082455.43 +391641.80 4C39.23 1480.8 0.56 C* 1.2160 0.0010 S 082455.48 +391641.92 18.3 18.1 17.8 17.6 17.3 14.2 − 21 082725.40 +291844.80 3C200 2043.1 0.92 II c 0.4580 S 082725.38 +291845.04 21.7 20.4 19.1 18.5 18.0 15.3 − 22 083110.00 +374209.90 4C37.24 2259.6 0.65 C 0.9188 0.0014 S 083110.01 +374209.58 19.2 18.7 18.6 18.6 18.4 − 23 083318.80 +510307.80 4C51.25 1313.5 0.81 II c 0.5621 0.0419 P 083318.72 +510306.88 26.9 22.3 20.5 19.6 19.2 − 24 083448.37 +170046.10 3C202 1882.8 0.72 II c 0.6237 0.1740 P 083448.22 +170042.44 23.0 22.4 21.7 21.1 21.9 − 25 083454.91 +553421.00 4C55.16 8283.1 0.01 C 0.2412 0.0014 S 083454.90 +553421.11 19.6 17.9 16.7 16.1 15.8 − 26 083753.51 +445054.60 4C45.17 1528.9 0.60 II c 0.2072 0.0009 S 083752.76 +445025.95 20.3 18.4 17.1 16.6 16.2 14.0 − 27 083906.50 +575413.40 3C205 2257.7 0.86 II c 1.5360 S 083906.54 +575417.06 17.9 17.4 17.0 16.6 16.5 14.5 − 28 084047.70 +131223.90 3C207 2613.0 0.81 II c 0.6804 0.0010 S 084047.59 +131223.62 18.7 18.1 18.0 17.9 17.7 15.0 − 29 084331.63 +421529.70 B30840+424A 1409.7 0.41 C* 0.8393 0.1758 P 084331.64 +421529.38 25.6 22.6 21.3 21.3 20.2 30 084753.83 +535236.80 NGC2656 1542.3 −0.58 I p 0.0453 0.0002 S 084753.07 +535234.25 16.3 14.3 13.4 13.0 12.7 10.4e − 31 084757.00 +314840.50 4C31.32 1482.0 0.52 II c 0.0673 0.0003 S 084759.05 +314708.34 16.5 14.5 13.6 13.2 12.9 12.8 − 32 085308.83 +135255.30 3C208 2364.3 0.97 II c 1.1115 0.0014 S 085308.61 +135254.84 17.9 17.9 17.6 17.6 17.8 33 085439.35 +140552.10 3C208.1 2163.8 −0.71 IIc p 1.0200 S 085439.32 +140551.86 19.8 19.6 19.3 19.3 19.2 34 085448.87 +200630.70 PKS0851+202 1511.8 −0.21 C* 0.4190 0.0016 S 085448.87 +200630.71 16.4 15.8 15.4 15.0 14.7 11.8eT h 35 085740.64 +340406.40 3C211 1798.4 0.77 II c 0.4789 0.0618 P 085740.28 +340404.91 22.8 21.9 20.6 19.8 19.4 e − 36 085810.07 +275050.80 3C210 1807.8 0.83 II c 1.1690 S 085810.04 +275054.17 23.3 22.6 21.5 21.0 20.2 C 37 085841.51 +140943.80 3C212 2370.8 −0.87 II c 1.0430 S 085841.45 +140944.78 21.0 20.0 19.1 18.8 18.6 15.3o − N 38 090105.40 +290145.70 3C213.1 2003.4 0.58 II p 0.1940 0.0002 S 090105.26 +290146.92 20.1 18.5 17.6 17.1 16.9 15.2F − 39 090304.04 +465104.70 4C47.29 1754.9 0.39 C* 1.4710 0.0024 S 090304.01 +465104.21 19.3 19.3 18.9 18.7 18.7 I − G 40 090631.88 +164613.00 3C215 1586.2 0.95 II c 0.4115 0.0003 S 090631.80 +164612.00 25.1 22.3 22.6 20.8 21.0 15.5 41 090734.92 +413453.80 4C41.19 1394.5 −0.70 II c 0.4783 0.0379 P 090733.18 +413444.10 21.9 20.5 19.1 18.3 18.0 C − a 42 090850.56 +374820.20 3C217 2086.4 0.95 II c 0.8980 S 090850.67 +374819.69 22.3 22.2 21.2 20.3 19.8 t − a 43 090933.53 +425347.40 3C216 4233.8 0.77 S* 0.6700 0.0014 S 090933.50 +425346.50 19.9 19.3 18.7 18.3 18.0 14.6l − o 44 091204.00 +161829.70 4C16.27 1374.6 0.77 II c 0.9182 0.4087 P 091203.99 +161829.99 21.6 21.7 21.8 22.0 22.5 g − u 45 091404.83 +171552.40 4C17.48 1527.3 0.74 II c 0.5395 0.0291 P 09145.21 +171554.35 24.3 21.0 19.7 18.6 18.4 e 46 092107.54 +453845.70 3C219 8101.6 −0.79 II c 0.1744 0.0012 S 092108.62 +453857.39 19.2 17.9 16.7 16.3 16.0 13.1eI − I 47 092249.93 +530221.20 4C53.18 1597.8 0.77 II c 0.5974 0.1361 P 092249.90 +530221.00 23.5 25.3 21.9 20.9 20.2 . − 48 092703.04 +390220.70 4C39.25 2884.6 0.29 C* 0.6967 0.0019 S 092703.01 +390220.87 17.0 16.7 16.6 16.7 16.6 14.0 − 49 093033.45 +360123.60 3C220.2 1875.1 0.68 II c 1.1570 0.0013 S 093033.54 +360125.18 18.6 18.3 17.8 17.6 17.6 15.89 − 10 Melanie A. Gendre, P. N. Best and J. V. Wall Figure7.Luminositydistributionsforcompactandextended(FRI,FRIIanduncertain)sources.Thecross-hatchedcolumnsrepresent the estimated contribution to each luminosity bin of sources with no redshift informationavailable, following the method presented in 4.2 § Figure 8.P-zplanecoverage forthefourCoNFIGsamples,as wellas the3CRR,CENSORSandLynx&Herculessamples,byradio- morphological type (limited only to sources with estimated redshifts). The dot-dashed lines show the survey limits for each sample. Sources are identified by their radio morphological classification: FRIs, FRIIs, uncertain and compact sources are represented by blue stars,redcircles,greendotsandblackcrossesrespectively. most of the FRI count at low flux densities is composed 2005)andfoundevidencethatFRIsundergosignificantevo- of low-luminosity sources at low redshift, our results show lutionoverz <0.7.OurresultsalsoshowthatFRIsundergo thatFRIobjectsmustundergosomemildevolution.Thisis lessevolutionthanFRIIs,andtheydonotparticipatemuch consistentwiththeresultsofSadler et al.(2007),whostud- inthesource-count“evolutionbump”aroundS1.4GHz∼1Jy. iedlowpowersourcesinthe2SLAQsurvey(Richards et al.