Spectral variability in faint high frequency peakers
aa r X i v : . [ a s t r o - ph . C O ] J un Mon. Not. R. Astron. Soc. , 1– ?? (2002) Printed 20 November 2018 (MN LaTEX style file v2.2) Spectral variability in faint high frequency peakers
M. Orienti , ⋆ , D. Dallacasa , , C. Stanghellini Dipartimento di Astronomia, Universit`a di Bologna, via Ranzani 1, I-40127, Bologna, Italy Istituto di Radioastronomia - INAF, Via P. Gobetti 101, I-40129 Bologna, Italy
Received 20 November 2018; accepted ?
ABSTRACT
We present the analysis of simultaneous multi-frequency Very Large Array (VLA)observations of 57 out of 61 sources from the “faint” high frequency peaker (HFP)sample carried out in various epochs. Sloan Digital Sky Survey SDSS data have beenused to identify the optical counterpart of each radio source. From the analysis ofthe multi-epoch spectra we find that 24 sources do not show evidence of spectralvariability, while 12 objects do not possess a peaked spectrum anymore at least inone of the observing epochs. Among the remaining 21 sources showing some degree ofvariability, we find that in 8 objects the spectral properties change consistently withthe expectation for a radio source undergoing adiabatic expansion. The comparisonbetween the variability and the optical identification suggests that the majority ofradio sources hosted in galaxies likely represent the young radio source population,whereas the majority of those associated with quasars are part of a different populationsimilar to flat-spectrum objects, which possess peaked spectra during short intervalsof their life, as found in other samples of high-frequency peaking objects. The analysisof the optical images from the SDSS points out the presence of companions around 6HFP hosted in galaxies, suggesting that young radio sources resides in groups.
Key words: galaxies: active - galaxies: evolution - radio continuum: general - quasars:general
Our knowledge of the first stages of the evolution of pow-erful radio sources is based on the study of the populationof high frequency peaking radio sources. In the frameworkof models explaining the evolution of individual radiosources, the spectral peak of young radio sources occurs athigh frequencies. Given their small size, in these sourcesthe synchrotron self-absorption (SSA) is a very effectivemechanism. As the source grows, the peak frequency isexpected to shift towards lower frequencies as a conse-quence of adiabatic expansion. An alternative explanationsuggests that the spectral peak is due to free-free absorptionfrom a ionized medium enshrouding the radio emission(Bicknell et al. 1997). Both scenarios are supported by theempirical anti-correlation found by O’Dea & Baum (1997)from the study of samples of compact steep spectrum (CSS)and gigahertz-peaked spectrum (GPS) radio sources. Theformer have peak frequencies around a few hundred MHz,typical sizes of a few kpc and ages of 10 - 10 years, whereasthe latter have spectra peaking around 1 GHz, typical sizesof about 1 kpc or less and ages of 10 - 10 years. However, ⋆ E-mail: [email protected] it is worth noting that the consistency between the sourcesize and the spectral peak often found in the most compactsources strongly support the synchrotron self-absorptionscenario (Orienti & Dallacasa 2008a; Tingay & de Kool2003).High frequency peakers (HFP; Dallacasa et al. 2000), witha spectral peak occurring at frequencies above a few GHz,are thus the best candidates to be newly born radio sources,with ages between 10 - 10 years.The radio properties of HFPs have been derived by theanalysis of the “bright” HFP sample (Dallacasa et al.2000; Tinti et al. 2005; Orienti et al. 2006a, 2007;Orienti & Dallacasa 2008b). In particular, from themulti-epoch analysis of their radio spectra it has been foundthat the sample is composed of two different populations.One population consists of radio sources that maintain theconvex spectrum without showing variability, whereas theother comprises radio sources that change their spectralshape, becoming also flat-spectrum objects, and possessingsubstantial flux density variability. The different spectralproperties shown by the two populations suggest that theformer represent young radio sources still in an early stageof their evolution, while the latter are beamed objects.The analysis of the multi-epoch spectral behaviour has c (cid:13) M. Orienti et al.
Table 1.
VLA observations and configurationsDate Conf Proj.ID CodeSep. 2003 AnB AD488 aJan.2004 BnC AD494 bNov.2006 C AO210 cApr.2007 D AO210 d proved to be a powerful tool to discriminate between thetwo populations (Torniainen et al. 2005; Tinti et al. 2005;Orienti et al. 2007). In fact, beamed radio sources, althoughusually characterized by a flat and variable spectrum,may be selected in samples of high-frequency peakingobjects when their emission is dominated by a flaringknot in a jet. On the other hand, young radio sources areknown to be the least variable class of extragalactic radiosources (O’Dea 1998). However, it must be mentionedthat in the youngest objects, substantial variability inthe optically-thick part of the spectrum is expected as aconsequence of either the source growth/evolution (e.g.J1459+3337, Orienti & Dallacasa 2008c), or changes inthe possible absorber screen, or a combination of both(Tingay & de Kool 2003).In this paper we present a multi-epoch analysis basedon simultaneous multi-frequency VLA data of the radiospectra of 57 high frequency peakers from the “faint”HFP sample (Stanghellini et al. 2009). This sample wasselected as the “bright” HFP sample (Dallacasa et al. 2000)by cross-correlating the 87GB survey at 4.9 GHz withthe NVSS at 1.4 GHz, and including only sources fainterthan 300 mJy at 4.9 GHz within a restricted area aroundthe northern galactic cap (for details on the sample seeStanghellini et al. 2009). The study of the radio propertiesof a sample of faint HFPs is the first step in understandingthe first stages of the radio source evolution. So far, spectralstudies have been carried out for the bright HFP sourcesonly, and an extension to fainter objects is necessary. Theevolution models developed so far (e.g. Fanti et al. 1995;Snellen et al. 2000; Kaiser & Alexander 1997) predict thatin the earliest stage the radio luminosity progressivelyincreases, implying that the youngest objects are likelyto be found among faint sources. Furthermore, in faintHFPs, boosting effects should be less relevant, making thecontamination from blazars less severe than what found insamples of brighter objects.Throughout this paper we assume the following cosmol-ogy: H = 71km s − Mpc − , Ω M = 0 .
27, and Ω Λ = 0 .
73 ina flat Universe. The spectral index is defined as S ( ν ) ∝ ν − α . Simultaneous multi-frequency VLA observations of 57out of the 61 sources from the “faint” HFP sample (Stanghellini et al. 2009) were carried out during differentruns between September 2003 and April 2007 (Table 1).Observations were performed in L band (with the two IFscentered at 1.415 and 1.665 GHz), C band (with the twoIFs centered at 4.565 and 4.935 GHz), X band (with thetwo IFs centered at 8.085 and 8.465 GHz), U band (14.940GHz), K band (22.460 GHz), and in Q band (43.340 GHz).At each frequency, the target sources were observed forabout 1 minute, cycling through frequencies. During eachrun, the primary flux density calibrator either 3C 286 or3C 48 was observed for about 3 minutes at each frequency.Secondary calibrators were chosen on the basis of theirdistance from the targets in order to minimize the telescopeslewing time, and they were observed for 1.5 min at eachfrequency, every 20 min.The data reduction was carried out following the standardprocedures for the VLA implemented in the NRAO AIPSpackage. The flux density at each frequency was mea-sured on the final image produced after a few phase-onlyself-calibration iterations. In the L band it was generallynecessary to image a few confusing sources falling within theprimary beam. All the target sources appeared unresolvedat any frequency. During the observations of a few sources,strong RFI at 1.420 and 1.665 GHz was present, and inthose cases the measurements of the flux density were notpossible.Uncertainties on the determination of the absolute fluxdensity scale are dominated by amplitude errors. Based onthe variations of antenna complex gains during the variousobservations, we can conservatively estimate an uncertaintyof ∼
3% in L, C, and X bands, ∼
5% in U band, and ∼ ∼
20 mJy. Results are presented in Section 4.
To determine the optical properties of the sources inthe faint HFP sample, we complemented the informationavailable in the literature with that provided by the SDSSDR7 (Abazajian et al. 2009). The optical properties of eachobject (like source extension and magnitude) have beencarefully inspected beyond the automated procedures inthe SDSS, in order to unambiguously identify the host (i.e.quasar, galaxy, or empty field) of each radio source.Of the 57 sources considered in this paper, 12 are identifiedwith galaxies with redshift between 0.03 and 0.6; 33 arequasars with a higher redshift, typically in the range from0.6 to 3.0, while 12 sources still lack an optical counterpart(labelled as empty field (EF) in Table 3).Images and optical information have been retrieved bymeans of the SDSS DR7 Finding Chart Tool. In Table 3 wereport the R magnitude, converted in the Johnson-Kron-Cousins BVRI system, and the spectroscopic redshift whenavailable. An “f” indicates a photometric redshift.In Fig. 1 we present the optical images from the SDSS DR7(Abazajian et al. 2009) of the sources hosted in galaxies. c (cid:13) , 1– ?? pectral variability in faint high frequency peakers J0804+5431 G z = 0.22 fz=0.26f
35 kpc z=0.26 fz=0.27 fz=0.25 fz=0.22 f z=0.26 f z=0.258
55 kpc
J0943+5113 G z = 0.42 f
43 kpc
J0951+3451 G z = 0.29 f
J1025+2541 G z = 0.46 f
58 kpc 40 kpc
J1058+3353 G z = 0.265z=0.26 fz=0.23 f
J1109+3831 G z=0.2+/−0.1 f
42 kpc z=0.119 z=0.152
J1218+2828 G z = 0.18 f
30 kpc z=0.19 f
52 kpc
J1240+2323 G z = 0.38 f
J1330+5202 G
66 kpc z = 0.58 f
Figure 1.
Optical images from the SDSS DR7 of the 12 HFP radio sources identified with a galaxy. On the images we report the sourcename, the redshift (an “f” indicates a photometric redshift). If the source forms a group, we report the redshift of the companion galaxies,when available. The field width has been chosen to show a region of about 250 kpc around the galaxy hosting the HFP radio source.
For the 3 galaxies with available spectroscopic data we showthe spectrum in Fig. 2 and we summarize the informationon the main lines in Table 2.
A characteristic arising from Fig. 1 is the presence ofcompanions within a projected distance of about 150 -200 kpc around 6 HFP galaxies. Although in J0804+5431,J1058+3353, J1109+3831, and J1218+2828 this evidencecomes from photometric information only, in the case of c (cid:13) , 1– ?? M. Orienti et al.
41 kpc
J1352+3603 G z = 0.27 f
78 kpc
J1530+2705 G z = 0.033z=0.031 z=0.031 z=0.032z=0.033
J1602+2646z=0.372 z=0.32 fz=0.371z=0.32 fz=0.31 f
46 kpc
Figure 1.
Continued.
J1530+2705 and J1602+2646 the association is confirmedby spectroscopic redshifts, supporting the idea that youngradio sources reside in groups, as found in other works(Orienti et al. 2006b; Snellen et al. 2002; Stanghellini et al.1993). The relatively small redshift of J1530+2705 allowsus to identify the spiral morphology of the brightestcompanions where also a bar is clearly visible.The galaxies hosting the HFPs are usually the brightestelliptical at the group centre. In the case of J0804+5431the radio source is hosted by an elliptical galaxy that isat a projected distance of about 160 kpc to the north-eastfrom the brightest galaxy at the centre of the group and itis at a projected distance of about 20 kpc to the north ofanother elliptical.An intriguing case is represented by J1109+3831 whosehosting galaxy is a spiral that is located at a projecteddistance of about 20 kpc from an elliptical. A possibleidentification error between optical and radio images hasbeen excluded by the analysis of the optical spectrum ofthe companion, which lacks the typical lines displayed byactive galaxies.Among the HFPs identified with galaxies, 3 objects(J1058+3353, J1530+2705, and J1602+2646) have anoptical spectrum in the SDSS DR7. For J1602+2646,the optical spectrum seems to be well fitted by a QSOtemplate, since a large fraction of the light comes from thenuclear region. However, both Fig. 1 and the analysis of thediagnostics O[II]/H β - O[III]/H β clearly indicate that thissource is hosted in a galaxy.The emission lines detected in these objects (Table 2)are those typical for radio galaxies, showing [O II] λ λλ α /N II, H β lines, and [O I] λ β emission lines, as expected in non-activeobjects.To check whether the galaxies of the faint HFP samplefollow the Hubble relation found by Snellen et al. (1996),we added the 12 galaxies with spectroscopic and photo-metric redshift in the GPS R-band Hubble diagram ofSnellen et al. (2002). The Hubble diagram, where the HFPgalaxies ( squares ) have been added to the GPS galaxies( crosses ) from Snellen et al. (2002), is presented in Fig. 3,and it indicates that HFP galaxies have a tight distributionin the apparent magnitude-redshift relation, as also foundin GPS galaxies (Snellen et al. 1996, 2002). Simultaneous multi-frequency observations carried out atdifferent epochs are necessary to monitor the spectral be-haviour and variability of high-frequency peaking radiosources. Variations in the spectral properties, like peak fre-quency, spectral shape and flux density, are strong indica-tors of the true nature of the source (Orienti et al. 2007;Tingay & de Kool 2003). In young radio sources the spec-tral properties are not expected to change, while spectralvariability is a typical characteristic of beamed radio sources.To determine a possible variation of the spectral peak, foreach source we fitted the simultaneous radio spectrum ateach epoch with a pure analytical function:Log S = a + Log ν · ( b + c Log ν )where S is the flux density, ν the frequency, and a, b and c are numeric parameters without any direct physicalmeaning. We prefer to adopt this function instead ofthat used by Stanghellini et al. (2009) because it better c (cid:13) , 1– ?? pectral variability in faint high frequency peakers Table 2.
Spectral lines of the HFP galaxies with available optical spectrum from the SDSS DR7. Column 1: source name (J2000); Col.2: line; Col. 3: line frequency in the observer’s frame; Col. 4: line flux density in the observer’s frame; Col. 5: equivalent width in the restframe. Source Line λ oss S line , obs EW rest ˚A 10 − erg s − cm − ˚AJ1058+3353 [O I] 7915.6 ± ± ± ± ± ± ± ± ± ± ± ± β ± ± ± α ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± α ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± β ± ± ± α ± ± ± represents the data, providing more accurate values for thepeak parameters. The best fits to the spectra are shownin Fig. 4, and the derived peak frequencies at the variousepochs are reported in Table 4. Statistical errors derivedfrom the fit are not representative of the real uncertaintyon the estimate of the peak frequency. For this reason weprefer to assume a conservative uncertainty on the peakfrequency of 10%. The position of the spectral peak iswell constrained when the peak occurs at a frequency wellsampled by the observations, becoming less accurate whenthe frequency coverage is not as appropriate. For example,in the case a source has a spectrum peaking at the edgesof the frequency coverage (i.e. L or K/Q bands), the fitprovides parameters that are less constrained than in thecase of sources with the spectral peak occurring around5-10 GHz, where both the optically thin and thick emissionare properly sampled. Spectra poorly constrained are alsothose lacking observations at some frequencies, as in thecase of J1044+4328, J1052+3355, and J1547+3518 wherethe lack of data either at 1.4/1.7 or 15.3 GHz precluded areliable determination of their peak frequencies.By comparing the distribution of the peak frequency of allthe sources at the various epochs, we do not find remarkabledifferences; the changes are usually within the uncertainties.The median value of the peak frequency of the whole sampleat each epoch has not changed significantly: ν p = 5 . ν p = 6 . ν p = 5 . α b ) and above( α a ) the peak frequency, by fitting a straight line in theoptically-thick and -thin part of the spectrum, respectively,following the approach by Torniainen et al. (2005) andOrienti et al. (2007). We considered “flat” those sourceswith both α b > − . α a < .
5. In a few sources,depending on the peak frequency, we could fit either α b or α a only, in order to avoid the flattening near the peak. Wefind that 12 sources (2 galaxies, 9 quasars, and 1 emptyfield) show a flat spectrum during at least one of theobserving epochs, implying that they are part of the blazarpopulation. The fitted optically thick and thin spectralindices are reported in Table 3.A case that is worth discussion in detail is the quasarJ1008+2533. The radio spectrum shown by this sourceduring two of the observing runs presented here turnedout to be a composition of two different spectra: convexat frequencies below 8.4 GHz, and inverted at higherfrequencies (see Fig. 4). This shape is similar to thatshown by the bright HFP J0927+3902. In J0927+3902the two-component spectrum is explained by its core-jetstructure (see e.g. Orienti et al. 2006a): at frequencies below1 GHz the spectrum is dominated by the emission fromthe jet, while at higher frequencies the contribution fromthe self-absorbed core becomes more important, becomingthe dominant emission above ∼
10 GHz. Such a scenariois well supported by pc-scale morphological informationby multi-frequency VLBI data (Alberdi et al. 2000). Asimilar explanation may apply to the case of the faint HFPJ1008+2533. Another possibility is that in our new epochs c (cid:13) , 1– ?? M. Orienti et al.
Table 3.
Multi-frequency VLA flux density of 57 sources from the faint HFP sample. Column 1: source name (J2000); Col. 2: opticalidentification from the SDSS DR7: G=galaxy, Q=quasar; EF=empty field; Col. 3: R magnitude; Col. 4: redshift. An “f” indicates aphotometric redshift from the SDSS DR7; Col. 5: the observing code from Table 1; Cols. 6 - 14: flux density at 1.4, 1.7, 4.5, 5.0, 8.1, 8.4,15.3, 22.2, and 43.2 GHz respectively; Cols. 15 and 16: the spectral index computed below and above the spectral peak, respectively.Source ID mag z code S . S . S . S . S . S . S . S . S . α b α a R mJy mJy mJy mJy mJy mJy mJy mJy mJy(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)J0736+4744 Q 20.3 - b 39 45 60 60 49 47 29 18 9 -0.3 0.8c 30 - 53 54 46 48 - 19 - -0.5 0.7J0754+3033 Q 17.3 0.769 a 69 77 170 173 182 180 151 130 - -0.6 0.3b 65 70 161 165 174 173 150 116 76 -0.6 0.4c - - 150 154 166 166 - 95 - - 0.6J0804+5431 G 18.1 0.22f c 37 - 82 81 73 72 - 37 - -0.7 0.4J0819+3823 Q 21.6 - a 18 25 115 120 101 96 48 17 - -1.4 1.3b 16 23 113 116 97 93 45 25 10 -1.6 1.0c 14 - 120 127 115 112 - 26 - -1.7 1.1J0821+3107 Q 16.9 2.625 c 93 - 95 92 75 74 - 31 - - 0.7J0905+3742 EF - - a 54 64 102 100 72 68 29 11 - -0.5 1.3b 51 74 102 99 71 68 35 18 8 -0.6 1.1J0943+5113 G 20.8 0.42f a 77 90 160 147 69 64 16 9 - -0.6 1.8c 72 - 163 152 79 75 - 17 - -0.6 1.4J0951+3451 G 19.2 0.29f a 19 29 62 62 56 55 37 26 - -1.0 0.6c 23 - 63 64 59 58 - 24 - -0.8 0.6J0955+3335 Q 17.3 2.491 a 45 58 104 106 95 92 60 36 - -0.6 0.7c 49 - 94 94 79 76 - 29 - -0.5 0.8J1002+5701 EF - - a 28 39 130 129 72 65 15 - - -1.3 1.8J1004+4328 EF - - b 11 15 36 37 29 29 10 16 10 -1.0 0.6c 12 - 36 37 33 32 - 19 - -1.0 0.4J1008+2533 Q 18.3 1.960 b 49 66 113 112 100 99 130 161 135 - -c 50 - 116 116 107 108 - 150 - - -J1020+2910 EF - - b 24 27 16 16 15 14 13 10 6 - 0.4c 16 - 15 15 14 14 - 7 - - 0.3J1020+4320 Q 18.8 1.964 b 118 157 253 247 191 183 116 80 31 -0.6 0.7J1025+2541 G 19.7 0.46f b 24 35 47 47 30 29 13 9 5 -0.6 1.0J1035+4230 Q 19.1 2.440 b 28 28 90 95 98 96 68 44 14 -0.7 0.7J1037+3646 EF - - b 70 94 146 141 99 94 55 33 11 -0.6 0.9J1044+2959 Q 18.9 2.983 b - - 144 177 163 159 120 96 65 - 0.6J1046+2600 EF - b 13 - 38 38 31 30 15 7 4 -0.8 1.0J1047+3945 Q 20.0 - b 41 - 39 39 30 30 20 12 9 - 0.4J1052+3355 Q 16.9 1.407 b - - 38 36 22 20 11 5 5 - 1.2c 14 - 35 32 20 18 - 6 - -0.8 1.1J1053+4610 EF - - b 11 - 36* 38 42 42 54 59 42 -0.6 -c 22 - 39 40 60 62 - 75 - -0.4 -J1054+5058 Q 22.0 - c 12 - 20 21 31 32 - 40 - -0.4 -J1058+3353 G 18.5 0.265 a 20 26 39 40 41 42 51 51 - -0.4 -J1107+3421 EF - - a 21 32 73 72 52 50 21 6 - -1.1 1.6b 25 38 75 73 52 48 25 14 6 -1.0 1.1c 28 - 73 72 51 48 - 9 - -0.8 1.3J1109+3831 G 17.7 0.2f a 12 14 53 59 90 91 77 56 - -1.1 0.4b 13 15 50 55 88 89* 81 59 23 -1.0 0.4c 14 - 50 55 95 98 - 53 - -1.1 0.6J1135+3624 EF - - a 28 37 58 58 46 45 21 11 - -0.6 1.0c 28 - 59 59 50 49 - 13 - -0.6 1.0J1137+3441 Q 18.6 0.835 a 25 34 78 83 85 121 157 169 - -0.7 -J1203+4803 Q 16.2 0.817 a 187 - 529 562 721 734 788 761 - -0.6 -c 218 252 416 425 452 458 - 384 - -0.4 0.2J1218+2828 G 18.1 0.18f a 25 28 92 96 98 97 69 58 - -0.7 0.5J1239+3705 Q 21.4 - a - - 93 101 129 130 112 86 - -0.5 0.4c 13 17 96 107 145 146 - 100 - -1.3 0.4c (cid:13) , 1– ?? pectral variability in faint high frequency peakers Table 3.
ContinuedSource ID mag z code S . S . S . S . S . S . S . S . S . α b α a R mJy mJy mJy mJy mJy mJy mJy mJy mJy(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)J1240+2323 G 21.0 0.38f a 27 28 52 53 59 60 54 49 - -0.6 0.2c 24 - 58 60 65 65 - 56 - -0.6 -J1240+2425 Q 16.9 0.831 a 78 - 60 58 45 44 32 27 - - 0.4c 57 - 58 56 45 44 - 17 - - 0.5J1241+3844 Q 21.5 - c 20 21 23 25 20 20 - 13 - -0.2 0.4J1251+4317 Q 18.7 1.453 a - - 54 57 80 82 112 116 - -0.4 -c 25 - 72 80 138 142 - 105 - -1.0 0.3J1258+2820 Q 19.2 - a 25 32 51 52 51 50 43 38 - -0.6 0.2c 23 34 45 45 54 54 - 54 - -0.3 -J1300+4352 Q 19.6 - c 140 155 116 115 97 95 - 81 - - 0.2J1309+4047 Q 18.9 2.910 a 37 - 139 131 118 115 76 40 - -1.1 0.8c 34 51 132 133 113 110 - 33 - -1.1 0.9J1319+4951 Q 19.1 - a 25 - 49 49 39 39 25 21 - -0.5 0.5c 22 30 50 49 43 43 - 20 - -0.7 0.6J1321+4406 Q 21.2 - a 62 - 73 74 71 70 58 50 - -0.1 0.3c 65 73 78 77 75 74 - 38 - -0.1 0.4J1322+3912 Q 17.5 2.985 a 117 - 227 223 181 176 118 86 - -0.6 0.6c 116 135 200 196 150 146 - 56 - -0.5 0.8J1330+5202 G 20.7 0.58f c 91 103 176 180 180 177 - 156 - -0.4 0.1J1336+4735 Q 19.7 - a - - 61 61 52 51 26 22 - - 0.7d 26 - 67 67 57 56 - 30 - - 0.5J1352+3603 G 18.0 0.27f a 65 70 102 103 95 93 65 37 - -0.4 0.7d 59 - 97 97 84 82 - 34 - -0.4 0.7J1420+2704 Q 20.3 - a 14 - 55 57 55 53 34 25 - -1.1 0.6d 14 - 62 63 58 56 - 26 - -1.2 0.6J1436+4820 EF - - a 20 - 72 72 61 58 32 13 - -1.0 1.1J1459+3337 Q 16.6 0.645 a 22 30 195 221 403 415 470 435 - -1.3 -J1530+2705 G 14.1 0.033 a 13 14 43 46 45 43 17 13 - -1.0 0.8d 11 12 48 49 42 41 - 23 - -1.1 0.5J1547+3518 Q 21.2 - a - - 53 57 67 67 74 72 - -0.3 -d 12 - 49 51 57 58 - 74 - -0.6 -J1602+2647 G 18.4 0.372 a 39 - 131 148 238 244 259 224 - -0.8 -J1613+4223 Q 19.8 - b 42 - 206 201 119 110 39 14 - -1.4 1.7d 40 - 201 194 114 105 - 13 - -1.4 1.7J1616+4632 Q 19.3 0.950 d 88 - 141 144 148 147 - 129 - -0.3 0.1J1617+3801 Q 19.0 1.607 a 23 - 71 75 99 100 95 67 - -0.8 0.4J1624+2748 EF - - a 20 - 109 118 175 177 198 173 - -1.0 -J1651+3417 EF - - b - - 46 48 58 58 49 22 16 - 0.9J1702+2643 Q 17.2 - b 35 39 41 42 49 50 61 65 73 -0.2 -J1719+4804 Q 15.3 1.084 a 75 - 126 135 158 157 115 85 - -0.4 0.6b 66 - 122 127 150 148 113 73 30 -0.5 0.7d 60 - 112 114 100 97 - 43 - -0.5 0.6 a flare from a self-absorbed knot in the jet occurred, causingan increase of the flux density at high frequencies, as alsofound in blazar objects. The lack of pc-scale morphologyinformation does not allow us to unambiguously determinethe origin of this complex spectrum.
Following the approach by Tinti et al. (2005) andOrienti et al. (2007) we investigate the presence of flux den-sity variability, by computing the variability index V : V = 1 m m X i =1 ( S i − S i ) σ i (1)where S i is the flux density at the i -th frequency measuredat one epoch, S i is the mean value of the flux densitycomputed averaging the flux density at the i -th frequencymeasured at all the available epochs, σ i is the rms on S i − S i , and m is the number of sampled frequencies. Weprefer to compute the variability index for each new epoch(see Table 4) instead of considering all the epochs togetherin order to better detect the presence of a possible burst.As in Tinti et al. (2005), we consider variable those sourceswith a variability index V > c (cid:13) , 1– ?? M. Orienti et al. S ( m J y ) (GHz) ν J0736+4744
J0736+4744 Q V S ( m J y ) (GHz) ν J0754+3033V J0754+3033 Q S ( m J y ) (GHz) ν J0804+5431 H J0804+5431 G J0819+3823 S ( m J y ) (GHz) ν H J0819+3823 QJ0821+3107 S ( m J y ) (GHz) ν V J0821+3107 Q J0905+3742 S ( m J y ) ν (GHz) H J0905+3742 EF J0943+5113 S ( m J y ) (GHz) ν H J0943+5113 G J0951+3451 S ( m J y ) (GHz) ν H J0951+3451 GJ0955+3335 (GHz) ν S ( m J y ) V J0955+3335 Q J1002+5701 S ( m J y ) (GHz) ν H J1002+5701 EF J1004+4328 S ( m J y ) (GHz) ν V J1004+4328 EF J1008+2533 S ( m J y ) (GHz) ν F J1008+2533 QJ1020+2910 S ( m J y ) (GHz) ν F J1020+2910 EF J1020+4320 S ( m J y ) (GHz) ν H J1020+4320 Q J1025+2541 S ( m J y ) (GHz) ν V J1025+2541 G J1035+4230 S ( m J y ) ν (GHz) H J1035+4230 Q
Figure 4.
Radio spectra of the 57 candidate HFPs from the “faint” HFP sample observed with the VLA during the observing runspresented in this paper. Asterisks and a dash-dotted line refer to the first epoch observations (1998-2000, Stanghellini et al. 2009); crossesand a dashed line refer to epoch a (2003); diamonds and a solid line refer to epoch b (2004); squares and a dotted line refer to epochs c,d (2006-2007). c (cid:13) , 1– ?? pectral variability in faint high frequency peakers J1037+3646 S ( m J y ) (GHz) ν H J1037+3646 EF J1044+2959 S ( m J y ) (GHz) ν H J1044+4328 Q J1046+2600 S ( m J y ) (GHz) ν H J1046+2600 EF J1046+2600 S ( m J y ) (GHz) ν F J1047+3945 QJ1052+3355 S ( m J y ) (GHz) ν V J1052+3355 Q J1053+4610 S ( m J y ) (GHz) ν V J1053+4610 EF J1054+5058 S ( m J y ) (GHz) ν H J1054+5058 Q J1058+3353 S ( m J y ) (GHz) ν F J1058+3353 GJ1107+3421 S ( m J y ) (GHz) ν H J1107+3421 EF J1109+3831 S ( m J y ) (GHz) ν V J1109+3831 G J1135+3624 S ( m J y ) (GHz) ν H J1135+3624 EF J1137+3441 S ( m J y ) (GHz) ν V J1137+3441 Q (GHz) ν J1203+4803 S ( m J y ) F J1203+4803 Q J1218+2828 S ( m J y ) (GHz) ν V J1218+2828 G J1239+3705 S ( m J y ) (GHz) ν V J1239+3705 Q J1240+2323 S ( m J y ) (GHz) ν V J1240+2323 G
Figure 4.
Continued.c (cid:13) , 1– ?? M. Orienti et al.
J1240+2425 S ( m J y ) (GHz) ν F J1240+2425 Q J1241+3844 S ( m J y ) (GHz) ν H J1241+3844 Q J1251+4317 S ( m J y ) (GHz) ν V J1251+4317 Q J1258+2820 S ( m J y ) (GHz) ν F J1258+2820 QJ1300+4352 S ( m J y ) (GHz) ν F J1300+4352 Q J1309+4047 S ( m J y ) (GHz) ν H J1309+4047 Q J1319+4851 S ( m J y ) (GHz) ν H J1319+4851 Q J1321+4406 S ( m J y ) (GHz) ν F J1321+4406 QJ1322+3912 S ( m J y ) (GHz) ν V J1322+3912 Q J1330+5202 S ( m J y ) (GHz) ν F J1330+5202 G J1336+4735 S ( m J y ) (GHz) ν V J1336+4735 Q J1352+3603 S ( m J y ) (GHz) ν V J1352+3603 GJ1420+2704 S ( m J y ) (GHz) ν H J1420+2704 Q J1436+4820 S ( m J y ) (GHz) ν H J1436+4820 EF S ( m J y ) (GHz) ν J1459+3337V J1459+3337 Q J1530+2705 S ( m J y ) (GHz) ν V J1530+2705 G
Figure 4.
Continued. c (cid:13) , 1– ?? pectral variability in faint high frequency peakers J1547+3518 S ( m J y ) (GHz) ν V J1547+3518 Q J1602+2646 S ( m J y ) (GHz) ν H J1602+2646 G J1613+4223 S ( m J y ) (GHz) ν H J1613+4223 Q J1616+4632 S ( m J y ) (GHz) ν F J1616+4632 QJ1617+3801 S ( m J y ) (GHz) ν H J1617+3801 Q J1624+2748 S ( m J y ) (GHz) ν H J1624+2748 EF J1651+3417 S ( m J y ) (GHz) ν H J1651+3417 EF J1702+2643 S ( m J y ) (GHz) ν F J1702+2643 QJ1719+4804 S ( m J y ) (GHz) ν V J1719+4804 Q
Figure 4.
Continued.
From the comparison of the multi-epoch spectralproperties and variability we find:- 24 objects maintain the convex spectrum without showingsignificant flux density variability (
V <
V >
The optical identification of the radio sources in sam-ples of CSS and GPS objects (e.g. Fanti et al. 1990;Stanghellini et al. 1998; Snellen et al. 1998; Fanti et al.2001) showed that in “faint” samples there is a higherfraction of objects identified with galaxies with respectto “bright” samples, and the fraction of galaxies seemsto be anti-correlated with the peak frequency, i.e. inCSS samples a higher percentage of radio sources arehosted in galaxies than in GPS and HFP samples. Thisis easily seen in Fig. 5, where we have plotted the peakfrequency versus the peak flux density for the sourcesfrom the bright HFP sample (Dallacasa et al. 2000),the faint HFP sample (Stanghellini et al. 2009), theGPS sample (Stanghellini et al. 1998), the faint GPSsample (Snellen et al. 1998); the half-Jansky GPS sam- c (cid:13) , 1– ?? M. Orienti et al.
J1058+3353J1530+2705J1602+2646
Figure 2.
Optical spectra from the SDSS DR7 of the galaxiesJ1058+3353 ( top ), J1530+2705 ( center ), J1602+2642 ( bottom ). R ( m a g ) z Figure 3.
The Hubble diagram of HFP galaxies ( squares ) andGPS galaxies ( crosses , Snellen et al. 2002). The dotted linerepresents the Hubble relation for GPS galaxies as found bySnellen et al. (1996). ple (Snellen et al. 2002), and the B3-VLA CSS sample(Fanti et al. 2001). In the latter sample only objects withspectra peaking above 100 MHz have been considered, sincefor sources with a peak at lower frequency, the turnoverfrequency could not be reliably constrained. In this plot,radio sources identified with galaxies (squares) mainlyoccupy both the top left part of the panel, i.e. low peakfrequency and high peak flux density, and the bottom rightpanel, i.e. high peak frequency (usually below 10 GHz)but low peak flux density. On the other hand, quasars(circles) are mostly found in the left part of the plot, withpeak frequencies above a few GHz, and peak flux densitiesthat span almost three orders of magnitude. Radio sourceswithout an optical identification (triangles) are found inthe same regions as galaxies, suggesting that the majorityof these objects share the same properties of the galaxies,but they are fainter probably because they are at higherredshift. It is worth noting that the criteria that have beenused to date to select radio source catalogues cannot pickup objects with both low peak frequency and low peak fluxdensity (bottom left panel), indicating that we miss thepopulation of faint CSS objects.Statistical studies on the correlation between the radiocharacteristics and the optical counterpart indicate thatthose objects hosted in a galaxy have the typical propertiesof young radio sources (i.e. symmetric radio structureand no spectral variability), whereas those identified withquasars are more similar to flat-spectrum radio objects.This suggests that there is a dichotomy between the opticalidentification and the radio properties: quasars are morelikely part of the flat-spectrum blazar population, whilegalaxies are likely associated with genuinely young radiosources. In the case of the sources from the faint HFPsample studied in this paper, the comparison between theiroptical identification and the radio variability showed thatthe majority of quasars have a variable spectrum, whilea smaller fraction maintain the convex spectrum withoutvariability (33% H, 39% V, and 28% F). In the case ofgalaxies, we found that the fraction of variable objects is c (cid:13) , 1– ?? pectral variability in faint high frequency peakers still larger than those without spectral variability (34%H, 50% V, and 16% F). For the sources still lacking anoptical identification we found that the majority do notshow significant spectral variability (75% H, 17% V, 8% F).Among the 12 sources with a flat radio spectrum, wefound that two objects (J1058+3353 and J1330+5202) areidentified with a galaxy, and one (J1020+2910) lacks anoptical identification. When we compare the variabilityproperties of sources with different optical identification bymeans of the Student’s t-statistic we find that there is asignificant difference ( > The anti-correlation found between the projected lin-ear size and the peak frequency (O’Dea & Baum 1997;Bicknell et al. 1997; Snellen et al. 2000) implies that thesources with the spectral peak occurring above a fewGHz should represent the population of the smallestradio sources whose radio emission has recently turnedon. Samples of high-frequency peaking objects have beenselected by choosing sources with an inverted radio spec-trum up to 5 GHz (Dallacasa et al. 2000; Torniainen et al.2005; Stanghellini et al. 2009), i.e. the highest observingfrequency where a large area survey is presently available.However, due to the selection criteria these samples havebeen found to comprise both young radio sources andflaring flat-spectrum objects selected during particularphases of their spectral variability, for example when theirradio emission is dominated by a knot in the jet. Thestudy of flux density and spectral variability based onrepeated simultaneous multi-frequency observations hasproved to be an ideal tool in discriminating the differentnature of the sources. It was found that in samples of brightobjects, where there is a high incidence of sources opticallyidentified with quasars, flat-spectrum blazar objects repre-sent the dominant population (e.g. Torniainen et al. 2007;Orienti et al. 2007; Jauncey et al. 2003). A higher incidenceof genuinely young radio sources is expected in samples offaint objects where the majority of radio sources should behosted in galaxies and boosting effects are supposed to playa minor role.The optical identification of the sources studied in thispaper by means of the SDSS DR7 indicates that 21% ofobjects are hosted in galaxies, i.e. similar to the fraction ofgalaxies in the bright HFP sample. However, in the faintsample, another 21% of objects lack optical identificationand thus a reliable comparison between the two samplescannot be done. The analysis of the optical images ofthe galaxies hosting HFP pointed out the presence ofcompanions around 6 HFP candidates, indicating thatyoung radio galaxies, like powerful extended radio sources,are in groups, as previously suggested by Stanghellini et al.(1993), indicating a continuity between compact youngobjects and the population of classical radio galaxies(O’Dea et al. 1996).Although in 4 galaxies the presence of companions issuggested by photometric information only, in J1530+2705 and J1602+2646 the association is made by spectroscopicredshift. The companion galaxies are located within a pro-jected distance of about 150 - 200 kpc from the target whichusually is hosted in the brightest elliptical at the groupcentre. A peculiar case is represented by J1109+3831, whoseparent galaxy seems to be a spiral that is interacting withan elliptical. Young radio sources are normally associatedwith ellipticals. The case represented by J1109+3831 maybe explained by the possible interaction between the hostingspiral and the companion that may have triggered the radioemission. The small redshift of J1530+2705 enabled us toidentify the morphology of its brightest companions thatturned out to be barred spirals. This group resembles thatof the bright HFP J0655+4100 (Orienti et al. 2006b), andin both cases the HFP is hosted by the central ellipticalgalaxy at the group centre. The presence of companiongalaxies in the environment of galaxies hosting youngradio sources suggests that the onset of the radio emissionmay be triggered by merger or interaction events thatoccurred not long ago. This scenario is supported bythe proximity of the companions in J0804+5431 and inJ1109+3831, although observations to establish a physi-cal interaction are needed to unambiguously verify this idea.From the analysis of the multi-epoch radio spectra ofthe sources in the faint HFP sample, we find a high frac-tion of objects displaying some level of variability. Thisresult does not imply that all these sources are part ofthe blazar population. In fact, changes in the radio spec-trum may be a direct consequence of the source expan-sion (e.g. Tingay & de Kool 2003). In newly born radiosources, the evolution timescales can be of the order of afew tens of years. Changes in the radio spectrum of suchyoung objects can be appreciable after the short time (5-8years) elapsed between the first and last observing run. Aclear example is represented by the faint HFP J1459+3337(Orienti & Dallacasa 2008c). This HFP showed a steadilyincreasing flux density at 1.4 and 5 GHz, in the opticallythick regime, and its spectral peak shifted from 30 GHzdown to 12 GHz in about 10 years (Edge et al. 1996). Thisbehaviour is consistent with the flux density and spectralevolution of a young object, with an age of about 50 years,undergoing adiabatic expansion.In the presence of adiabatic expansion of a homogeneoussynchrotron source, the radio spectrum undergoes a shifttowards low frequencies. In the optically-thick regime thismeans that at a given frequency the flux density S increaseswith time (see Pacholczyk 1970, Orienti et al. 2007, and Ori-enti et al. 2008b for a detail analysis of the radio spectrumevolution): S = S (cid:16) t + ∆ tt (cid:17) (2)where S and S are the flux densities at the time t and t + ∆ t , respectively. On the other hand, the spectral peak ν p moves to lower frequencies: ν p, = ν p, (cid:16) t t + ∆ t (cid:17) (3)where ν p, and ν p, are the peak frequency at the time t and t + ∆ t respectively. c (cid:13) , 1– ?? M. Orienti et al.
Table 4.
Peak frequency and flux density variability. Column 1: source name (J2000); Col. 2: peak frequency in GHz of the first epoch(1998-2000, Stanghellini et al. 2009); Cols 3, 4, and 5: peak frequency in GHz of epoch a (2003), b (2004), and c, d (2006-2007); Cols. 6,7, and 8: variability index V computed for the epoch a , b , c,d respectively; Col. 9: the classification of the source spectrum (V=variable,H=genuine HFP, F=flat; see Section 3). An asterisk indicates that the peak frequency is not reliable due to poor frequency coverage. :for the source J1008+2533 we report the peak of the lowest part of the spectrum.Source ν ep1 ν ep2 ν ep3 ν ep4 V a V b V c,d Var.(1) (2) (3) (4) (5) (6) (7) (8) (9)J0736+4744 3.6 3.9 4.6 6.3 22.2 VJ0754+3033 8.8 7.9 8.1 7.3 27.5 22.4 19.9 VJ0804+5431 5.4 5.5 2.3 HJ0819+3823 5.8 5.6 6.2 6.1 6.7 3.0 4.0 HJ0821+3107 3.3 2.6 166.6 VJ0905+3742 3.9 3.8 3.7 13.6 1.5 HJ0943+5113 3.7 3.4 3.7 14.7 12.6 HJ0951+3451 6.0 6.1 5.6 1.1 4.3 HJ0955+3335 5.8 5.2 4.5 7.4 57.9 VJ1002+5701 4.6 4.1 14.4 HJ1004+4328 8.0 5.5 6.5 10.3 10.9 VJ1008+2533 ∗ ∗ >
22 110.0 352.4 VJ1054+5058 > >
22 5.3 HJ1058+3353 6.4 >
22 71.6 FJ1107+3421 4.6 4.5 5.2 4.2 12.8 0.9 4.0 HJ1109+3831 8.1 9.3 9.2 9.3 5.1 8.4 14.1 VJ1135+3624 4.1 4.2 4.4 0.9 5.5 HJ1137+3441 23.0 >
22 146.8 VJ1203+4803 >
22 15.2 9.0 600.0 1058.1 FJ1218+2828 7.1 7.5 50.7 VJ1239+3705 9.5 9.5 10.0 23.6 35.2 VJ1240+2323 7.8 9.0 9.8 16.0 12.2 VJ1240+2425 3.8 0.7 2.6 55.7 5.3 FJ1241+3844 3.7 3.6 1.0 HJ1251+4317 7.5 >
22 11.4 150.3 430.3 VJ1258+2820 4.8 7.0 14.6 22.2 40.2 FJ1300+4352 5.8 < > ∗ (cid:13) , 1– ?? pectral variability in faint high frequency peakers Table 4.
Continued. Source ν ep1 ν ep2 ν ep3 ν ep4 V a V b V c,d Var.(1) (2) (3) (4) (5) (6) (7) (8) (9)J1602+2646 12.8 13.0 7.5 HJ1613+4223 4.6 4.3 4.4 5.9 12.4 HJ1616+4632 >
22 8.5 119.1 FJ1617+3801 11.6 9.4 1.7 HJ1624+2748 12.0 13.3 9.5 HJ1651+3417 8.6 8.4 9.0 HJ1702+2643 21.5 >
22 115.5 FJ1719+4804 9.7 6.0 7.5 4.8 37.6 33.8 472.3 V
In the optically-thin regime, the flux density at a given fre-quency decreases with time: S = S (cid:16) t + ∆ tt (cid:17) − δ (4)where δ is the spectral index of the electron energy dis-tribution N ( E ) ∝ E − δ that originates the radio emission.It is clear from this relationship that substantial variationin the optically-thick regime can be revealed in case that∆ t is a non-negligible fraction of the total source age ( t ).In our case, ∆ t ∼ t ∼
100 years.Among the sources with some changes in the spectralproperties compatible with such a scenario, other 7 sources(J0754+3033, J0955+3335, J1004+4328, J1025+2541,J1052+3355, J1322+3912, and J1547+3518) in addition toJ1459+3337 show a spectral and flux density variability thatmay be explained in terms of source expansion. Althoughin J1459+3337 there are several indicators supporting thisinterpretation (Orienti & Dallacasa 2008c), in the caseof the other sources mentioned here this assumption isbased on the flux density and peak variation on a smalltime range only. Additional observations spanning a longertime interval, together with information on the pc-scalemorphology are necessary in order to reliably constrain thesource nature.
We presented simultaneous multi-frequency VLA obser-vations of 57 sources from the faint HFP sample, carriedout at various epochs. From the comparison of the spectralproperties we found that 24 objects (4 galaxies, 11 quasars,and 9 empty fields) preserve their convex spectrum withoutshowing any evidence of flux density variability. Of theremaining sources, 12 objects (2 galaxies, 9 quasars, and 1empty field), selected on the basis of their convex spectrumin the first epoch by Stanghellini et al. (2009), turned outto show a flat spectrum in one of the subsequent observingepochs. The remaining 21 sources (6 galaxies, 13 quasars,and 2 empty fields) possess high levels of variability,although still displaying a convex spectrum. However,among these variable sources we found that in 8 objects thechanges in their spectra are consistent with what expected
Figure 5.
The peak frequency versus the peak flux density forthe radio sources from the B3-VLA CSS sample (Fanti et al.2001), the faint GPS sample (Snellen et al. 1998), the half-Jansky GPS sample (Snellen et al. 2002), the GPS sample fromStanghellini et al. (1998), the bright (Dallacasa et al. 2000) andthe faint HFP samples (Stanghellini et al. 2009). Circles, squaresand triangles refer to quasars, galaxies and empty field, respec-tively. if the source is undergoing adiabatically expansion. Thisimplies that out of the 57 sources studied in this paper, 32objects (56%) can still be considered young radio sourcecandidates. The remaining 25 sources (44%) are part ofthe flat-spectrum blazar population, indicating that alsoin samples of faint radio sources, where boosted effects arethought to play a minor role, a large fraction of sources arerepresented by flaring objects.The analysis of the optical images of the HFPs hosted bygalaxies pointed out the presence of companion galaxiesin the target environment, supporting the idea that youngradio sources reside in groups. The parent galaxy is usuallythe brightest elliptical at the group centre with the ex-ception of two sources. In J0804+5431 the galaxy hostingthe HFP is at the periphery of the group, and it seemsinteracting with a close elliptical. A surprising result isrepresented by the HFP J1109+3831 that is hosted in a c (cid:13) , 1– ?? M. Orienti et al. spiral that seems to be interacting with a close elliptical.The fact that young radio sources reside in groups supportthe idea that the interactions occurring between the galaxiesare at the origin of the radio emission.
ACKNOWLEDGMENTS
We thank S. Bardelli for his help on the analysis of the opti-cal spectra. The VLA and the VLBA are operated by the USNational Radio Astronomy Observatory which is a facilityof the National Science Foundation operated under cooper-ative agreement by Associated Universities, Inc. This workhas made use of the NASA/IPAC Extragalactic DatabaseNED which is operated by the JPL, Californian Institute ofTechnology, under contract with the National Aeronauticsand Space Administration. Funding for the SDSS and SDSS-II has been provided by the Alfred P. Sloan Foundation, theparticipating Institutions, the National Science Foundation,the U.S. Department of Energy, the National Aeronauticsand Space Administration, the Japanese Monbukagakusho,the Max Planck Society, and the Higher Education FundingCouncil for England. The SDSS was managed by the Astro-physical Research Consortium for the Participating Institu-tions.
REFERENCES
Abazajian, K.N., Adelman-McCarthy, J.K., Ag¨ueros,M.A., et al., 2009, ApJS, 182, 543Alberdi, A., Gomez, J.L., Marcaide, J.M., Marsher, A.P.,Perez-Torres M.A., 2000, A&A, 361, 529Bicknell, G.V., Dopita, M.A., O’Dea, C.P., 1997, ApJ, 485,112Dallacasa, D., Stanghellini, C., Centonza, M., Fanti, R.,2000, A&A, 363, 887Edge, A.C., Jones, M., Saunders, R., Pooley, G., Grainge,K., 1996, in Proceedings of the Second Workshop on GPSand CSS Radio Sources, ed. I.A.G. Snellen (Leiden: Lei-den Observatory), 208Fanti, R., Fanti, C., Schilizzi, R.T., Spencer, R.E., NanRendong, Parma, P., van Breugel, W.J.M., Venturi, T.,1990, A&A, 231, 333Fanti, C., Fanti, R.,Dallacasa, D., Schilizzi, R.T., Spencer,R.E., Stanghellini, C., 1995, A&A, 302, 31Fanti, C., Pozzi, F., Dallacasa, D., Fanti, R., Gregorini, L.,Stanghellini, C., Vigotti, M., 2001, A&A, 369, 380Jauncey, D.L., King, E.A., Bignall, H.E., et al., 2003,PASA, 20, 151Kaiser, C.R., Alexander, P., 1997, MNRAS, 286, 215O’Dea, C.P., Stanghellini, C., Baum, S.A., Charlot, S.,1996, ApJ, 470, 806O’Dea, C.P., Baum, S.A., 1997, AJ, 113, 148O’Dea, C.P., 1998, PASP, 110, 493Orienti, M., Dallacasa, D., Tinti, S., Stanghellini, C.,2006a, A&A, 450, 959Orienti, M., Morganti, R., Dallacasa, D., 2006b, A&A, 457,531Orienti, M., Dallacasa, D., Stanghellini, C., 2007, A&A,475, 813 Orienti, M., Dallacasa, D., 2008, A&A, 487, 885Orienti, M., Dallacasa, D., 2008b, A&A, 479, 409Orienti, M., Dallacasa, D., 2008c, A&A, 477, 807Pacholczyk, A.G., 1970, Radio Astrophysics (San Fran-cisco: Freeman & Co.)Snellen, I.A.G., Bremer, M.N., Schilizzi, R.T., Miley, G.K.,van Ojik, R., 1996, MNRAS, 279, 1294Snellen, I.A.G., Schilizzi, R.T., de bruyn, A.G., Miley,G.K., Rengelink, R.B., R¨ottgering, H.J., Bremer, M.N.,1998, A&AS, 131, 435Snellen, I.A.G., Schilizzi, R.T., Bremer, M.N., Miley, G.K.,de Bruyn, A.G., R¨ottgering, H.J.A., 1999, MNRAS, 307,149Snellen, I.A.G., Schilizzi, R.T., Miley, G.K., de Bruyn,A.G., Bremer, M.N., R¨ottgering, H.J.A., 2000, MNRAS,319, 445Snellen, I.A.G., Lehnert, M.D., Bremer, M.N., Schilizzi,R.T., 2002, MNRAS, 337, 981Stanghellini, C., O’Dea, C.P., Baum, S.A., Laurikainen, E.,1993, ApJS, 88, 1Stanghellini, C., O’Dea, C.P., Dallacasa, D., Baum, S.A.,Fanti, R., Fanti, C., 1998, A&AS, 131, 303Stanghellini, C., Dallacasa, D., Orienti, M., 2009, AN, 330,223Tingay, S.J., de Kool, M., 2003, AJ, 126, 723Tinti, S., Dallacasa, D., De Zotti, G., Celotti, A.,Stanghellini, C., 2005, A&A, 432, 31Torniainen, I., Tornikosky, M., Ter¨asranta, H., Aller, M.F.,Aller, H.D., 2005, A&A, 435, 839Torniainen, I., Tornikoski, M., L¨ahteenm¨aki, A., Aller,M.F., Aller, H.D., Mingaliev, M.G., 2007, A&A, 469, 451 c (cid:13) , 1–, 1–