Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.atnf.csiro.au/people/rnorris/papers/n197.pdf Äàòà èçìåíåíèÿ: Thu Jan 6 05:43:34 2005 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 10:42:45 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 43

This version 06 January 2005

Radio Observations of the Hubble Deep Field South region: I. Survey Description and Initial Results
Ray P. Norris CSIRO Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia Minh T. Huynh Research School of Astronomy & Astrophysics, Mount Stromlo Observatory, The Australian National University, Canberra ACT 2611, Australia Carole A. Jackson, Brian J. Boyle, Ronald. D. Ekers, Daniel A. Mitchell, Rob ert J. Sault, Mark H. Wieringa CSIRO Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia Rob ert E. Williams Space Telescope Science Institute 3700, San Martin Drive, Baltimore, MD 21218, USA Andrew M. Hopkins
1

Department of Physics & Astronomy, University of Pittsburgh, 3941 O'Hara Street, Pittsburgh, PA15260, USA and James Higdon Department of Astronomy, Cornel l University, 215, Space Sciences Building, Ithaca, NY 14853-6801, USA ABSTRACT This pap er is the first of a series describing the results of the Australia Telescop e Hubble Deep Field South (ATHDFS) radio survey. The survey was conducted at four wavelengths 20, 11, 6, and 3 cm, over a 4-year p eriod, and achieves an rms sensitivity of ab out 10 µJy at each wavelength. We describ e the observations and data reduction processes, and present data on radio sources close to the centre of the HDF-S. We
1

Hubble Fellow


­2­ discuss in detail the prop erties of a subset of these sources. The sources include b oth starburst galaxies and galaxies p owered by an active galactic nucleus, and range in redshift from 0.1 to 2.2. Some of them are characterised by unusually high radio-tooptical luminosities, presumably caused by dust extinction. Subject headings: surveys ­ radio continuum: radio galaxies, quasars

1.

Introduction

Following the success of the original Hubble Deep Field pro ject in the northern sky (HDFN; Williams et al. (1996)) the WFPC2, STIS and NICMOS instruments of the Hubble Space Telescop e (HST) were used to produce very deep images of a corresp onding southern sky region the `Hubble Deep Field South' (HDF-S; Williams et al. 2000). The HDF-S differs from the HDF-N in that a field was chosen near a bright, high redshift (z=2.24) quasar to enable a study of the intervening intergalactic medium (Outram et al. 1999). Both Hubble Deep Fields have provided a focus for a wide range of research, and extensive ground-based observations have complemented the optical and infrared images obtained by the HST, leading to significant advances in our understanding of the evolution of the Universe. However, most of the these observations were conducted at optical and infrared wavelengths. Radio observations occupy an imp ortant observational niche as they prob e physical conditions that are faint or unobservable at optical/infrared wavelengths, and they are also unaffected by dust extinction. The HDF-N has b een studied in detail at radio wavelengths using deep Very Large Array (VLA) observations (Richards et al. 1998; Richards 2000) resulting in the most sensitive radio images to date, reaching an rms sensitivity of 4 µJy at 20 cm, and 1.6 µJy at 3 cm. At high flux density levels, radio source surveys are generally dominated by radio-loud active galaxies and quasars. However, at µJy levels, the source counts are increasingly dominated by a p opulation of starburst galaxies (Hopkins, Afonso, Cram, & Mobasher 1999), which are generally b elieved to b e high-redshift (and in some cases higher-luminosity) counterparts to the ultraluminous infrared galaxies typified by Arp220. Muxlow et al (2000) have shown that, in their sample, 70% of the µJy sources are starburst-typ e sources lying at redshifts b etween z 0.4 and 1.0. Another 20% are low-luminosity AGNs in elliptical galaxies at z 1. The remaining 10% of µJy sources have optically-faint hosts, close to or b eyond the HDF-N limit. Further supp ort for b oth starburst and AGN processes, based on observations at radio, farinfrared (FIR) and submillimetre (submm) wavebands, have b een given by Barger, Cowie, & Richards (2000), Afonso, Mobasher, Chan, & Cram (2001), Brusa et al. (2002), Chapman et al. (2003), Bergstr¨ & Wiklind (2004), and Georgakakis et al. (2004). These observations sugom gest that there is a large p opulation of highly-obscured, but very active galaxies at z = 1­5 which is resp onsible for producing the bulk of the extragalactic background in the FIR/submm region.


­3­ These galaxies may host the ma jor sites of massive star formation and fuelling of active galactic nuclei (AGN) at high redshifts. Moreover, there is tentative evidence that these galaxies are strongly clustered, with a correlation-length exceeding any other known high-z p opulation (Blain, Chapman, Smail, & Ivison 2004; Stevens et al. 2003). These results hint at a direct link b etween these dusty, high-redshift galaxies and the growth of large scale structure in the early Universe, which may b e the origin of the basic environmental variations of galaxy prop erties (e.g. morphologies, stellar ages) seen in the modern Universe. Our understanding of these phenomena requires sensitive radio observations, extending to the limits of current instrumentation, together with equally sensitive observations at other wavelengths. The combination of deep radio observations, deep HST observations, and deep observations at other wavelengths, provides a remarkable opp ortunity in the Hubble Deep Fields to study the evolution of these galaxies. This is the first of a series of pap ers presenting results from our deep radio survey of the HDF-S made with the Australia Telescop e Compact Array (ATCA). The principal goals of this survey are to: · Use source counts together with redshifts to study the cosmological evolution of the star formation rate, the evolutionary history of galaxies, and the origin of current stellar p opulations. · Understand the evolutionary relationship b etween galaxies and AGN and trace the development of sup ermassive black holes. · Trace the changing balance b etween obscured and unobscured activity over the lifetime of the Universe. Obscured activity produces no more than 30% of the total b olometric luminosity in the local Universe, whereas it has b een suggested that high-redshift obscured activity might account for 90% of the detected luminosity. We also need to relate this transformation to the changing roles of star formation and AGN activity. Pap er 1 (this pap er) describ es the observations and data reduction, presents the radio data at all four wavelengths in the inner region of the HDF-S, and discusses a representative subset of sources in this region in some detail. Pap er 2 (Huynh et al. 2005a) presents the extended 20 cm observations and catalogue of detected sources, and analyses the differential radio source counts. Pap er 3 (Huynh et al. 2005b) presents the extended 13, 6, and 3 cm observations, and discusses radio sp ectral indices and individual radio sources in the field. Pap er 4 (Huynh et al. 2005c) presents optical identifications, classification, and redshifts, and discusses the implications for the radio luminosity function, and the cosmological evolution of the star formation rate. Throughout this pap er, we use the following cosmological parameters: H0 = 71kms m =0.27 and =0.73.
-1

Mpc-1 ,


­4­ 2. 2.1. Observations

Observations of HDF-S Candidate Fields

As part of the field selection process for the HDF-S, prior to the HST observations, we observed candidate fields with the ATCA to ensure that the selected field had no strong radio sources that would prevent deep radio imaging to complement the deep optical image. At the same time, the candidate fields were searched by other groups to find a high-redshift quasar, and were studied at other wavelengths to search for extinction or infrared cirrus. In May 1997, we made ATCA observations of eight candidate fields at 21 cm. All fields were observed in a single 12-hour run in a snapshot observing mode. Each image covered a 68-arcmin square field, with the central 34 arcmin square b eing the primary candidate area. Table 1 shows the field centres and the resulting source statistics. Our results showed that the only detected quasar (which we identify b elow as ATHDFS J223337.5-603329 ) was only 10 arcmin from a strong radio source (identified in Pap er 2 as ATHDFS J223355.6-604315), presenting the risk that dynamic range limitations might reduce the sensitivity available at radio wavelengths. However, it was judged that this source would not degrade the radio observations sufficiently to justify moving the HDF-S to another candidate field, provided the precise location of the HDF-S was chosen to avoid the radio source whilst keeping the quasar within the field. The p osition of the HDF-S was selected accordingly, and the results presented here confirm the wisdom of that decision.

2.2.

Survey Observations

We imaged the HDF-S using the ATCA at four wavelengths: 20, 11, 6 and 3 cm, obtaining resolutions of 6, 4, 2 and 1 arcsec resp ectively. Observations were taken over a p eriod of four years from mid 1998 to late 2001, in a variety of configurations to maximise uv coverage (i.e. sampling of the Fourier plane). Table 2 summarises the observations, and Figure 1 shows the p ointing centres. A single p ointing centre was used at each wavelength, rather than a mosaic, to maximise the sensitivity that might b e obtained in the available observing time. The p ointing centres differ at the four frequencies b ecause of the different primary b eam size at the four wavelengths, and thus the different strategies necessary to deal with the strong confusing source ATHDFS J223355.6-604315 with S20 cm = 155 mJy. At 6 and 3 cm the observations are centred on the WFPC2 field, so that the strong source is well outside the primary b eam. At 20 cm and 11 cm the p ointing centre (RA = 22h 33m 37.6s and Dec = -60 33' 29" (J2000) ) was chosen to b e approximately half-way b etween the WFPC2 field and ATHDFS J223355.6-604315. This ensures that the strong source is well inside the primary b eam to minimise calibration problems.


­5­ Throughout the observations, the correlator was used in continuum mode (2 â 128 MHz bandwidth), with each 128 MHz band divided into 32 â 4 MHz channels. This correlator configuration achieves the highest sensitivity, but at the price of higher resolution sp ectral information. In most observations, we allocated the two bands to two different observing frequencies. In Table 2 the total hours column is the total amount of time sp ent on each 128 MHz band, and should b e halved to give an equivalent observing time with two 128 MHz bands. The primary flux density calibrator used was PKS B1934-638, which is the standard calibrator for ATCA observations (S = 14.95 Jy at 1.380 GHz; Reynolds (1994) ). We calibrated the complex antenna gains by frequently observing the secondary calibrators PKS B2205-636 and PKS B2333-528.

3.

Data Reduction 3.1. Imaging

We used the Australia Telescop e National Facility (ATNF) release of the MIRIAD (Sault, Teub en, & Wright 1995) software to reduce our data, and based our imaging and source extraction process on the procedure outlined by Prandoni et al. (2000). Because of the large observing bandwidth (2 â 128 MHz), the multi-frequency synthesis (Sault & Wieringa 1994) technique was necessary to improve uv-coverage and reduce bandwidth smearing. This technique makes a single image from multi-frequency data by gridding each sp ectral channel in its correct place in the uv plane, instead of at a location determined by the average over all channels. The MIRIAD implementation of multi-frequency synthesis also solves for sp ectral index, so that the image is not degraded by differing sp ectral indices of sources in the field. Before imaging, the data from each observing session were insp ected and the MIRIAD tive tasks TVFLAG and BLFLAG were used to flag bad data resulting from interference, problems or correlator failures. The primary calibrator data were flagged b efore calibration plied. The secondary calibrator and target data had bandpass and p olarisation calibration b efore insp ection and flagging. interacreceiver was apapplied

After flagging, the data were split into the separate observing bands b efore imaging and cleaning, so that calibration and imaging could account for frequency-dep endent terms. When imaging, we explored a range of robust weights, which is a hybrid form of uniform and natural weighting (Briggs 1995). Robust values of 1 or more resulted in a degraded b eam shap e, which translated into a strong circular artefact in the final image. Decreasing robust values gave tighter main lob es, and hence higher resolution, b etter clean models and more effective self calibration, but at the cost of sensitivity. A robustness value of 0 was chosen for the final imaging, as it was found to b e a good compromise b etween sensitivity and resolution. This resulted in a synthesized b eam of 7.1 â 6.2 arcsec at 20 cm.


­6­ Both the 11 cm and 6 cm images contained sidelob es from two strong off-field sources. The sidelob es in the 6 cm image were removed by increasing the image size to image an area four times larger, and cleaning a small region, 2 arcmin square, around each source. These clean comp onents were removed from the visibility data b efore the final 6 cm imaging. The 11 cm off-field sources were removed by offsetting the image region to include the two sources. The procedure then followed that at 6 cm to remove the source sidelob es. We exp 20 cm data, bright (> 1 of the clean erimented with a variety of approaches to deconvolution, including, in the case of the cleaning a region around the bright central source only, and cleaning `b oxes' around all mJy) sources. The numb er of clean iterations was chosen by monitoring the residuals model. 11 cm data were self-calibrated to improve the image quality. The b est selfwas from a clean model derived over the whole image, with the numb er of itera( 30000 iterations). Two iterations of b oth phase and amplitude self calibration improve the image quality, and in particular remove phase error strip es.

The 20 and calibration result tions set as ab ove were sufficient to

After correction for the primary b eam attenuation, image analysis was confined to a circular region of radius 20, 12, 5.5, and 3.5 arcmin at 20, 11, 6, and 3 cm resp ectively, at which radius the sensitivity falls to 39% of that in the primary b eam centre.

3.2.

Source Extraction

To identify sources in the image at each wavelength, we first divided each map by the noise map generated by SExtractor (Bertin 2004) to obtain a `signal to noise' map. The MIRIAD task IMSAD was then used to derive a preliminary list of source `islands' ab ove a cut-off of 4 . Each source `island' found by IMSAD was examined and refitted with an elliptical Gaussian to derive source flux densities and sizes. All sources and fit parameters were insp ected to check for obvious failures and p oor fits that need further analysis. The p eak flux from this fit was then compared with a p eak value derived from a parab olic fit to the source (MIRIAD task MAXFIT). If these two p eak values differed by less than 20% and the fitted p osition was inside the 0.9Speak flux density contour, then these fluxes and p ositions were assigned a quality flag of 1. If they failed this test but insp ection showed that the derived p osition and p eak flux density of the p eak were consistent with the data, then they were assigned a quality flag of 2. Cases which failed this second test fell into one of the following categories: · Sources which were b etter describ ed by two Gaussians. In this case, the IMSAD islands were split into two Gaussian comp onents, which are catalogued as A and B in Table 3. · Sources with a shap e which could not b e fitted by a Gaussian. No such sources are amongst those discussed in this pap er.


­7­ · Obviously spurious sources that corresp ond to artefacts or noise p eaks. These very rare cases were deleted. Finally, to reduce the incompleteness of our catalogue and minimise the probability of spurious detections, we included only sources with signal to noise greater than 5 in the final catalogue and in Table 3.

3.3.

The limiting sensitivity of synthesis observations

The radio images describ ed here are amongst the deep est made with the ATCA. However, future observations will probably b e able to reach even deep er sensitivity levels, b ecause of the planned introduction of a wide-bandwidth correlator (Compact Array Broadband Backend, or CABB) at the ATCA in 2006, and through the p otential future availability of even larger amounts of observing time for a few key science pro jects. The sensitivity to compact sources achievable with a synthesis telescop e is limited by four factors. · Thermal Sensitivity For many observations, particularly those with short integration times, images are limited by the thermal noise temp erature of the telescop e, which is set by the receivers, optics, and atmosphere. In the case of a deep observation, this is still likely to b e the limiting factor at high frequencies, but will b ecome less imp ortant than other factors for deep integrations at low frequencies. For multi-day observations with the ATCA, this will continue to b e the limiting factor at 6 and 3 cm, b ecause of the low confusion levels together with the faintness and low surface brightness of sources at these frequencies. For example, the ATCA can b e exp ected to reach an rms noise level of 5 µJy after observing at 6 cm for 16 *12 h (using two 128-MHz bands, and natural weighting). This level dep ends primarily on the system gain and temp erature, and confusion does not play a significant role. This sensitivity limit is often referred to as the 'theoretical limit'. · Confusion by adjacent sources Long integrations at low frequencies (20 and 13 cm) are likely to reach the confusion limit at which each source of interest is confused by an adjacent background source. Assuming the log N/logS function taken from Jackson (2004), the confusion level, for a 6 arcsec b eam at 20 cm, is 0.05 µJy/b eam (i.e. on average, one background 0.05 µJy source will fall within each b eam). Thus this does not currently present a limit to synthesis observations for any realistic amount of observing time. However, confusing that most of the AT a commonly stated rule-of source p er 100 synthesised sources are separated by at CA, using the same source -thumb is to require that there b e no more than one b eams at the same level as the target source, implying least 10 synthesised b eams. For the 6-km configuration statistics, and assuming a sp ectral index of -0.7, this


­8­ gives a practical sensitivity limit of 21, 6, 0.7, and 0.03 µJy at 20, 13, 6, and 3 cm resp ectively. The 20 cm observations describ ed here already approach this limit. We note that this ruleof-thumb is probably too conservative for deep-field surveys such as those presented here, in which valuable astrophysical information can still b e gained even when, in some cases, there is some confusion b etween adjacent sources. · Dynamic range limited by incomplete u-v coverage Even with p erfect calibration, the dynamic range of an image is limited by incomplete u-v coverage, and hence missing information, which cannot b e recovered even by sophisticated deconvolution algorithms. For a carefully chosen field around a bright radio source, with no strong confusing sources nearby, this is likely to b e the factor that ultimately limits the sensitivity. The highest dynamic range (defined as p eak flux in the image divided by the rms in a source-free region of the image) ever achieved in a synthesis image is 106 , achieved using excellent uv coverage and redundant spacings on the Westerb ork Synthesis Radio Telescop e (de Bruyn & Brentjens 2005). The highest dynamic range reached with the VLA is 3.5 â 105 (Conway, Garrington, Perley, & Biretta 1993), and a dynamic range of 105 has b een reached by Geller et al. (2000) with the ATCA. We note that the ATCA is in principle capable of almost complete uv coverage, by observing at every p ossible baseline, using a sp ecial set of 12 configurations (Manchester et al. 1984). However, the standard range of configurations currently offered to users does not include this sp ecial set. No deep ATCA observation has yet b een demonstrably limited in dynamic range by incomplete uv coverage, but we note that future deep-field observations may require the use of such sp ecial configurations to reach the ultimate sensitivity. In principle, even with a standard set of configurations, the uv coverage may b e significantly improved by the use of the multi-frequency synthesis describ ed ab ove (Sault & Wieringa 1994). However, although the MIRIAD implementation of multi-frequency synthesis solves for sp ectral index, it still assumes a constant sp ectral index for each source over the observed frequency range. Data are not yet available to determine whether this will limit the dynamic range in deep field observations. · Dynamic range limited by calibration errors The dynamic range will b e limited by calibration errors. While simple antenna-based complex gain errors can b e corrected by the selfcal technique available in MIRIAD and other packages, these implementations are unable to correct for gain errors that vary across the primary b eam. Particularly troublesome are strong confusing sources close to the steep edge of the primary b eam resp onse (or its sidelob es), since small p ointing errors may cause large amplitude fluctuations in these sources, preventing them from b eing adequately deconvolved from the image. Furthermore, the primary b eam resp onse is assumed in current analysis packages to b e circularly symmetric. The magnitude of this effect is uncertain, but an indication of its severity may b e estimated as follows. Assuming that the critical area of the primary b eam at 20 cm in which p ointing errors will b e significant occupies an annulus of radius 15 arcmin and width 5 arcmin, then


­9­ we exp ect one 40 mJy source to fall by chance within that annulus. A typical dynamic range of 500 p er 12-hour synthesis for sources in this region will then limit the sensitivity to 40 µJy for each 12-hour image. When n observations are combined, then if the p ointing errors are purely random, this will b e reduced by sqrt(n), giving a limiting rms sensitivity of 10 µJy for the set of observations describ ed in this pap er. Corresp onding limits for 11 and 6 cm are 4 and 1 µJy resp ectively. This limiting sensitivity at 20 cm is similar to the sensitivity obtained in the observations describ ed here, suggesting that more sophisticated calibration algorithms may b e necessary to enable significantly deep er imaging. If future observations wish to prob e more deeply in the presence of such strong confusing sources, then it will b e necessary to develop calibration packages that account for varying gain errors across the primary b eam. We note that the aips++ package contains an algorithm that can in principle handle these errors.

4.

Radio Sources in the Inner HDF-S Region

The final 20 cm image is shown in Figure 1. Overlaid on this image are the primary b eam sizes and locations for the ATCA observations at the four frequencies, and the locations of the HST WFPC, STIS, and NICMOS fields. Of particular note are the bright confusing source ATHDFS J223355.6-604315 at 6 arcmin south-east of image centre, the clearly identifiable multiple radio source in the north-east corner, and a radio galaxy with a jet-lob e structure ab out 7 arcmin south of the image centre. Catalogues covering the complete region imaged by the AT at the two longest frequencies will b e presented in Pap ers 2-4. All the radio data from the survey are available on http://www.atnf.csiro.au/research/deep/hdfs/, and in the NASA/IPAC Extragalactic Database (NED) on http://nedwww.ipac.caltech.edu/. In this remainder of this pap er we focus on a circular region of radius 6.5 arcmin, centred on the WFPC field at (RA = 22h 32m 56.22s and Dec = -60 33' 02.7" (J2000) ). The size of this region has b een chosen to include the HST WFPC2, NICMOS and STIS fields, and also cover all of the area inside the 6 cm and 3 cm observations. Thus all the 6 and 3 cm data from this survey are presented in this pap er. The 20 cm image of this region is shown in Figure 2. In Table 3 we show the combined source catalogue for all sources detected at a level of 5 times the local rms noise at one or more frequencies within this region. Sources were matched across the 4 frequencies by visual insp ection and the result is a catalogue of 87 radio sources. For sources detected at more than one wavelength, the p osition is derived from the image with the smallest errors in the source p osition, as noted in column 15 of Table 3. The columns of Table 3 are as follows.


­ 10 ­ Column 1 -- Reference numb er. These are used for brevity only within this pap er, and, to avoid ambiguity, the full source name should b e used by any pap ers referring to the catalogue. Column 2 -- Source name. Multiple comp onents of a single source are lab elled A, B, etc. Columns 3 & 4 -- Right Ascension (J2000), 1 uncertainty in arcsec. Note that these uncertainties refer to the formal p ositional uncertainties derived from the fitting process. To these should b e added, in quadrature, a systematic uncertainty of ab out 0.2 arcsec, representing the uncertainty in the p osition of the AT phase calibrator sources. Columns 5 & 6 -- Declination (J2000), 1 uncertainty in arcsec. Columns (7­10) -- Flux densities at 20, 11, 6, and 3 cm in µJy. Where the column is blank, then the source is either undetected at that wavelength, i.e. b elow 5local, , or lies outside the catalogued area. The p eak flux is given for unresolved sources and the integrated flux is given for resolved sources. Column (11) -- Deconvolved ma jor axis (FWHM), column 15. Column (12) -- Deconvolved minor axis (FWHM), column 15.
maj

in arcsec, at the wavelength indicated in

min

, in arcsec, at the wavelength indicated in

Column (13) -- Deconvolved p osition angle, PA, in degrees east of north, at the wavelength indicated in column 15. Column (14) -- Local signal to noise ratio. Column (15) -- Band flag - indicates the wavelength at which the source p osition and Gaussian fit parameters are measured: L= 20 cm, S= 11 cm, C= 6 cm, and X= 3 cm. Column (16) -- Quality flag, as describ ed in Section 3.2

5.

Prop erties of Selected Radio Sources a subset of 19 representative sources chosen from the list b een selected in any statistically well-defined way, and so not b e true for the p opulation of sources as a whole. The Pap er 3.

In this section, we discuss in detail given in Table 3. These sources have not conclusions drawn from this subset may full sample will b e discussed in detail in

It is particularly imp ortant to determine the origin of the radio luminosity, which may b e generated either by star formation or AGN activity. A p otential discriminator b etween these two mechanisms is morphology. AGNs frequently have a classic double-lob ed structure, whilst starburst galaxies are typically unresolved or amorphous. However, the resolution of the observations presented here is in most cases insufficient to distinguish b etween these.


­ 11 ­ Another p otential discriminant is radio sp ectral index. Starburst galaxies typically have a sp ectral index of ab out -0.7, whilst AGNs typically have a sp ectral index ranging from ab out 0 to ab out -1.4. Thus, while little is learnt from a sp ectral index of ab out -0.7, a sp ectral index which differs significantly from this, in either direction, implies AGN activity. However, the sp ectral indices presented here have a typical uncertainty, due to image noise, of 0.1 ­ 0.2, so only extreme sp ectral indices are useful indicators. There is a further uncertainty caused by the mismatch of the synthesised b eam at different frequencies. Huynh et al. (2005b) reduce this uncertainty by convolving the higher-frequency image with the lower-frequency b eam, and we use their sp ectral indices here, which therefore differ slightly from those which would b e obtained from the figures in Table 3. One particular class of radio source, the "gigahertz p eaked sp ectrum" or GPS galaxy, has a sp ectrum which rises to a maximum at centimetre wavelengths, and is b elieved to represent an early stage of AGN activity (Snellen et al. 1999). This characteristic sp ectrum is seen in two of the ob jects discussed here, and is an unambiguous indicator of AGN activity. Ap contains sp ectral it would otential cause of a flat radio sp ectrum might also b e that the shorter wavelength emission a comp onent of free-free emission from HI I regions in a starburst galaxy. However, the indices discussed here are based on the 20 and 11 cm radio emission, at which wavelength b e unusual for free-free emission to b e significant (Condon 1992).

A further complication for any interpretation of these sources is that a starburst galaxy may contain an obscured AGN in its nucleus. Local examples of this phenomenon are NGC6240 (Gallimore & Beswick 2004) and Mrk231 (Yun, Reddy, & Condon 2001) in which a starburst galaxy harb ours a hidden low-luminosity AGN, with comparable contributions to the luminosity from the starburst and AGN comp onents. A list of the sources and their optical, infrared, and derived prop erties is given in Table 4. In Figure 3 we show the 20 cm radio contours overlaid on CTIO images (Palunas et al. 2000). The columns of Table 4 are as follows. Column 1 -- Reference numb er, as in Table 3. Column 2 -- Source name. Column 3 -- Flux density at 20 cm. Column 4 -- Measured redshift. Redshifts with four decimal digits are sp ectroscopic redshifts, while those with two decimal digits are photometric redshifts. References for the redshifts are given in the discussion of individual sources. For reasons discussed in Huynh et al. (2005b), in cases where b oth Fernandez-Soto et al (1999) and Teplitz et al. (2001) have measured photometric redshifts, but no sp ectroscopic redshifts are available, we adopt the Fernandez-Soto et al (1999) redshift. Columns 5­9 -- AB magnitudes measured at V, R, I, J, and K bands. Except where otherwise stated, measured values of V,R,I are taken from the CTIO catalogue (Palunas et al. 2000). Other


­ 12 ­ references are given in the discussion of individual sources. Column (10) -- I-K colour. Column (11) -- Radio luminosity Column (12) -- Star formation rate implied by the radio luminosity, if all radio emission is assumed to b e generated by star formation activity, using the scaling given by Condon (1992) with a Saltp eter IMF (Q=5.5). For comparison, this algorithm assigns Arp220 a star formation rate of 300 M yr -1 . Column (13) -- Logarithm of the radio to optical luminosity ratio, calculated as 0.4 â (I - where I is the I-band magnitude shown in Table 2, and S20 , is the radio magnitude derived the 20 cm flux density using the AB Radio magnitude scale defined by Ivezic et al. (2002). few cases, indicated by the notes to Table 2 and in the discussion of individual sources, I has estimated from R or J magnitudes. For comparison, Arp220 has log (S20 /I ) = 4.31 S20 ), from In a b een

Column (14) -- Classification of the radio source. The arguments for each classification are given in the discussion of individual sources. SB indicates starburst activity, Sy indicates a Seyfert galaxy, and GPS indicates a gigahertz-p eaked-sp ectrum AGN. The rest of this section discusses each of the sources in Table 4. In each case, the discussion is headed by the short reference numb er, the full source name, and the source name given by Norris et al. (1999, 2001), if appropriate. We use a numb er of abbreviations throughout this section: FS refers to Fernandez-Soto et al (1999), and SMO refers to Sawicki & Mall´ -Ornelas (2003). We en refer frequently to the CTIO images (Palunas et al. 2000), the WFPC images (Williams et al. 2000; Casertano et al. 2000), the WFPC FF (flanking field) observations (Lucas et al. 2003), the NICMOS images (Yahata et al 2000), and the EIS (ESO Infrared survey; Nonino et al. (1999)). Except where stated otherwise, all radio sp ectral indices are taken from Huynh et al. (2005b), and refer to the sp ectral index b etween 20 and 11 cm.

3. ATHDFS J223245.6-603857 (source b) This bright, marginally resolved, 843 µJy source has a radio sp ectral index of (-0.69 ± 0.06), which reveals little ab out the origin of its radio luminosity. The source is located ab out one arcsec east of a faint, p ossibly disturb ed or interacting, galaxy seen in WFPC FF and, faintly, in CTIO images. Here we assume they are coincident, b ecause astrometric accuracy on the flanking fields is relatively p oor. FS fit an irregular template and obtain a photometric redshift of 0.75. If the resulting high radio luminosity is attributed entirely to star formation activity, this implies a star formation rate of 515 M yr -1 . The radio-optical luminosity ratio is also very high, implying a high degree of extinction. One explanation is that it is a very obscured starburst galaxy, ab out twice as active as Arp220, similar to those suggested by Barger, Cowie, & Richards (2000) to account for the strong SCUBA sources in HDF-N, and an alternative is that the source contains a hidden


­ 13 ­ AGN. Given the irregular sp ectral template, and the disturb ed morphology, it is likely that this is a very obscured starburst galaxy, although we cannot rule out the presence of an obscured AGN.

7. ATHDFS J223254.5-603748 This faint 92 µJy source is detected only in our 20 cm observations. It is coincident with a bright disturb ed barred spiral galaxy seen in CTIO and WFPC FF images. Huynh et al. (2005c) measure a sp ectroscopic redshift of 0.1798, and see a Seyfert-typ e sp ectra. Thus, this source app ears to b e a relatively normal Seyfert galaxy. We note that SMO have measured a sp ectroscopic redshift of 0.2668, but the reason for this discrepancy is unclear. Here we adopt the Huynh et al. (2005c) redshift of 0.1798.

12. ATHDFS J223316.5-603627 (source f ) This bright, unresolved, 649 µJy source has a radio sp ectral index of (-0.67 ± 0.06). It is coincident with a galaxy in the CTIO image that app ears extended, and is p ossibly an edge-on spiral. Teplitz et al. (2001) measure a photometric redshift of 0.60, while FS measure a photometric redshift of 0.29, using a starburst template. Here we adopt the FS redshift. The radio luminosity then implies a star formation rate of 95 M yr -1 . This, together with the colours of this galaxy, indicates vigorous, but not unusual, starburst activity, in a relatively normal galaxy.

19. ATHDFS J223338.8-603523 ( source m) The radio sp ectrum of this 185 µJy unresolved source is relatively flat (-0.35± 0.21) suggesting an AGN rather than a starburst galaxy. The source is coincident with a bright elliptical galaxy in the WFPC FF image, which is also seen as a J=18.3 galaxy in the EIS survey, with J-K=0.54. Huynh et al. (2005c) have measured a sp ectroscopic redshift of 0.2250, and we note that Teplitz et al. (2001) have measured a photometric redshift of z=0.16. The prop erties of this source are consistent with those of an elliptical galaxy hosting an AGN.

26. ATHDFS J223243.3-603442 This faint radio source is visible in our 20 cm and 6 cm observations, but not in the 11 cm observations, which give an upp er limit S11 cm < 30 µJy. It has a radio sp ectral index derived directly from the 6 and 20 cm observations of -0.23, suggesting that it is an AGN, although this conclusion is weakened by the faintness of the source, its non-detection at 11cm, and the different b eam sizes at 6 and 20 cm. The source is coincident with an extended and apparently disturb ed


­ 14 ­ system, p ossibly a merger, visible in CTIO and WFPC images. Teplitz et al. (2001) measured a photometric redshift of 0.57, FS measured a photometric redshift of 0.46, and SMO measure a sp ectroscopic redshift of 0.4233, which we adopt here. Mann et al. (2002) have observed it at 7 and 15 µm with ISO, and show that its sp ectral energy distribution (SED) resembles a cirrusdominated galaxy like M51 rather than a starburst. Sub ject to the uncertainty discussed ab ove, the radio sp ectrum suggests that this otherwise normal spiral galaxy may harb our a weak radio AGN in its nucleus.

29. ATHDFS J223245.5-603419 (source a) This bright radio source has a sp ectral index of (-0.43 ± 0.21), which is marginally flatter than that exp ected of a starburst galaxy. It is identified with the western-most memb er of a line of three galaxies clearly visible in the CTIO, AAT and EIS images, but unfortunately just two arcsec outside the deep WFPC field. Its radio image is slightly extended in the direction of the line of galaxies, suggesting that these may contribute weakly to the radio emission. Franceschini et al. (2003) suggest that most of the infrared emission is from the central ob ject of the triplet, and yet the p osition given in their table coincides with the western ob ject, which is unambiguously the origin of the strong radio emission. We list their J and K magnitude in Table 2. Glazebrook et al. (1998) have measured a sp ectroscopic redshift of 0.4606, while Vanzella et al. (2002) measure a redshift of 0.4594, and SMO measure z=0.4608. We adopt the mean of these, which is 0.4603. Teplitz et al. (2001) and FS have measured photometric redshifts of 0.57 and 0.46 resp ectively, and FS fit an sb c template to this galaxy. Based on its infrared prop erties, Franceschini et al. (2003) derive a mass of 1011 M and a star formation rate of 50 M yr -1 , which is slightly lower than the rate of 81 M yr -1 that we derive from the radio observations. Mann et al. (2002) have observed it at 7 and 15 µm with ISO, and show that its sp ectral energy distribution (SED) resembles a starburst galaxy like Arp220. The data presented here are consistent with this interpretation.

31. ATHDFS J223327.6-603414 (source h) This sp ectrum of this bright unresolved radio source p eaks at 11 cm, and we therefore classify it as a GPS galaxy, which is b elieved to represent an early stage of AGN activity (Snellen et al. 1999). There is no optical counterpart, but it lies ab out two arcsec from a small galaxy (at 22 33 27.40, -60 34 12.37) which is visible within the STIS flanking fields observations. The absence of any galaxy coincident with the radio source in the HDF flanking field observations gives limiting magnitudes shown in Table 4, leading to an extreme value of the radio-optical luminosity ratio, indicating that the GPS source is strongly obscured by its host galaxy.


­ 15 ­ 33. ATHDFS J223243.4-603351 This unresolved radio source has a flux density of 98 µJy at 20 cm , rising to 114 µJy at 3 cm . This inverted sp ectrum suggests that it is an AGN. It is coincident with a p oint-like optical source on the CTIO image, and Huynh et al. (2005c) have measured a sp ectroscopic redshift of 1.566, and note that it has the broad lines of a quasar. Palunas et al. (2000) (who refer to it as QSO B) and Franceschini et al. (2003) also identify this source as a quasar at z=1.56. Mann et al. (2002) have observed it at 7 and 15 µm with ISO, and suggest that it is a broad-line AGN, but measure a redshift of 0.0918 based on a single broad line. As this differs from the redshifts measured by Huynh et al. (2005c), Palunas et al., and Francheschini et al., we suggest that the Mann et al. redshift was based on a misidentified line, and adopt a redshift of 1.566 for this quasar.

34. ATHDFS J223306.0-603350 (source d) This bright radio source has a sp ectral index of (-0.62 ± 0.12), and is identified with the core of a relatively bright face-on barred spiral galaxy (R=17.2), which app ears to b e interacting with a small galaxy a few arcsec to its north. The radio source is significantly extended to the south at 11 cm, p ossibly indicating some structure within the galaxy. Glazebrook et al. (1998) have measured a redshift of 0.1733, and show that it has the characteristic emission lines of a star-forming galaxy. It is the brightest 15-µm source seen in the ISO observations by Mann et al. (2002) who show that its SED is that of a normal spiral galaxy rather than a starburst. However, its optical and radio luminosities suggest that it is a luminous starburst galaxy, and Francheschini et al. show that its infrared emission shows an excess over that exp ected from a normal spiral. We derive a star formation rate of 27 M yr -1 , and thus classify it as a starburst rather than a normal spiral.

35. ATHDFS J223258.5-603346 (source c) This unusual source is the strongest radio source in the WFPC field, and yet is optically a very faint (V=27.05) red source which is invisible in CTIO images and barely visible in WFPC images. It has b een discussed extensively by Norris et al. (1999) and Norris et al. (2001), where it is lab elled as "source c". A numb er of authors have measured photometric redshifts for this galaxy: FS (1.69), Fontana et al. (2004) (1.7), Lanzetta et al (2002) (1.69), Rodighiero, Franceschini, & Fasano (2001) (1.69), and Rudnick et al. (2001) (1.34). We note the high degree of consistency b etween these measurements, and adopt a redshift of 1.69 for this ob ject. A 12-hour sp ectroscopic observation on the VLT failed to detect any emission lines (Vanzella et al. 2002). This source has an unusually high radio-optical ratio (the highest in this subset), several hundred times greater than Arp220. Vanzella et al. (2001) measure I-KAB = 3.45, classifying it as an extremely red ob ject (ERO) according to the division sp ecified by Pozzetti & Mannucci


­ 16 ­ (2000). EROs (McCarthy et al. 2004) are thought to b e a mix of passively evolving at 1 < z < 2 and heavily obscured star-forming galaxies, also at z > 1. However, as E b oth starburst and early-typ e galaxies, this classification does not yield any further on this ob ject. Based on the infrared colours, Vanzella et al. (2001) also classify it as galaxy. It app ears to b e marginally extended in b oth the 3 cm image and the WFPC red galaxies ROs include information an elliptical image.

The radio sp ectral index of this source is (-0.65 ± 0.05). It is difficult to determine whether this source is a extremely dusty starburst or a dust-enshrouded AGN. Its radio sp ectral index is consistent with either, and SCUBA observations of similar galaxies in the HDF-N (Barger, Cowie, & Richards 2000) suggest that their radio emission is produced by star formation. At a redshift of 1.69, this would imply this source is ab out five times more luminous in the radio than Arp220. On the other hand, unpublished SIMBA (Nyman et al. 2001) observations by Wiklind, Bergstrom, Huynh, Norris, and Jackson failed to detect any 1.3 mm continuum emission greater than 7.5 mJy, whereas if this source is p owered by starburst activity, the associated dust would b e exp ected to have a flux density of > 15 mJy. We therefore consider it likely that this source is a dust-enshrouded AGN.

37. ATHDFS J223247.6-603337 This faint radio source is seen only in our 20 cm observations, and is coincident with a large late-typ e spiral galaxy in the CTIO image. Glazebrook et al. (1998) measure a sp ectroscopic redshift of 0.5803, and SMO measure a sp ectroscopic redshift of 0.5807. Photometric redshifts of 0.67 and 0.56 have b een measured by Teplitz et al. (2001) and FS resp ectively. Mann et al. (2002) note that its SED is that of a normal galaxy like M51. Rigop oulou et al. (2002) and Franceschini et al. (2003) show that, based on sp ectroscopy and infrared photometry resp ectively, it is an unusually massive galaxy, with a dynamical mass of 4.5 - 10 â 1011 M and a star formation rate of 45 M yr -1 . This is in good agreement with the star formation rate of 32 M yr -1 that we derive from its radio luminosity.

39. ATHDFS J223337.5-603329 (source l) This source represents the bright, unresolved, radio emission from the well-studied STIS quasar (Outram et al. 1999) at z=2.238, whose location was partly resp onsible for the location of the HDFS, as discussed in Section 2.1 ab ove. This source has b een extensively discussed in the literature (e.g. Sealey, Drinkwater, & Webb (1998); Palunas et al. (2000)), and so we restrict the discussion here to noting that the associated radio source is typical of a p owerful radio-loud quasar, with a sp ectral index of (-0.69 ± 0.05).


­ 17 ­ 43. ATHDFS J223327.9-603304 (source i) This bright double radio source app ears in Table 3 with two entries (lab elled A and B) corresp onding to the two lob es of the source, indicating that it is a radio galaxy or quasar. There is no optical counterpart visible in the CTIO images, but a faint counterpart is visible at J and K in the EIS images. The limit on I from the CTIO catalogue gives a radio-infrared luminosity ratio of 5.20, indicating that the source is highly obscured.

48. ATHDFS J223308.6-603251(source e) This strong extended radio source is coincident with a R=21.77 CTIO spheroidal galaxy visible in the CTIO, AAO, and WFPC FF1 images. Its radio sp ectral index is (-0.86 ± 0.03), suggesting a radio galaxy rather than a starburst. The radio source app ears to b e extended in a NW-SE direction at 20, 11, and 6 cm by ab out 5 arcsec, corresp onding to 10 kp c at z=0.5. At 3 cm it app ears as a core-dominated source with weak jets extending to the NW and SE. Both Teplitz et al. (2001) and FS measure photometric redshifts of 0.64. All these data are consistent with a classical radio-loud AGN of moderate radio p ower.

49. ATHDFS J223323.2-603249 (source g) This slightly extended 457 µJy source is notable in having a GPS sp ectrum, suggesting that it is p owered by AGN activity. No optical counterpart is visible in the CTIO image of this field, but it is weakly detected at J and K by EIS. For the calculation of the radio-optical ratio, the I magnitude has b een estimated as I=J+0.12, where 0.12 is the mean value of I-J for those sources in this subset for which a measured value of I-J is available. This source has one of the largest radio-optical ratios in our subset.

55. ATHDFS J223331.6-603222 (source k) This slightly resolved 295 µJy source has a sp ectral index of (-0.68 ± 0.15), which gives no clue to the origin of its emission. It is coincident with the central galaxy of a group of three galaxies, p ossibly merging, seen in the CTIO, AAO, WFPC-FF, and EIS images, with J-K= 0.8. FS measure a photometric redshift of 0.42, and Huynh et al. (2005c) measure a sp ectroscopic redshift of 0.4652, with the sp ectrum characteristic of a star-forming galaxy. Its extension at 20 and 13 cm is consistent with some radio emission from the other two galaxies in the group. Its radio luminosity implies a star formation rate of 123 M yr -1 , and its radio-optical ratio is ab out ten times that of Arp220. These observations are consistent with a very obscured starburst galaxy.


­ 18 ­ 67. ATHDFS J223236.5-603000 This is the strongest (1.5 mJy) radio source in the subset discussed in this pap er. It has a sp ectral index of (-0.86 ± 0.05), and no optical counterpart. The EIS upp er limits give it an extreme radio-optical ratio of 6.02. It is marginally resolved at 20 cm, and also shows a small amount of emission to the south in the 13 cm image, which is coincident with another galaxy, and may b e unrelated. None of the observed characteristics gives any clue as to whether this radio emission is p owered by AGN or star formation activity, but the high radio-optical luminosity ratio indicates a high obscuration by dust, whatever the origin of the radio emission.

69. ATHDFS J223316.8-602934 This extended strong (1.0 mJy) source has a flat sp ectrum (sp ectral index (-0.16 ± 0.06)), indicating that it is an AGN. It is coincident with a faint galaxy in the CTIO image, in a dense cluster. Photometric redshifts of 0.12 and 0.89 have b een obtained by FS and Teplitz et al. (2001) resp ectively. Here we adopt the FS redshift. Some of the extended radio emission may b e attributable to the other galaxies in the cluster.

70. ATHDFS J223329.1-602933 This extended 261 µJy source app ears in our source catalogue only as a 20 cm detection, although there is a marginal detection in the 11 cm observations with a p eak flux density of 86 µJy, giving a sp ectral index based on this uncorrected flux of -1.9, indicating an AGN. We note that this would b e classified as an "ultra-steep sp ectrum" by De Breuck, van Breugel, R¨ ottgering, & Miley (2000), and consequently may b e a high redshift candidate. A faint extended source, or cluster of sources, is visible in the CTIO image, but no redshift is available. Although V and I magnitudes are given in the CTIO catalogue, they have very high uncertainties and so have b een set to zero in Table 4. For the calculation of the radio-optical ratio, an I magnitude has b een estimated as I=R-0.83, where 0.83 is the mean value of R-I for those sources in this subset for which a measured value of R-I is available. The extended radio emission to the north and east may represent emission from the other galaxies in the cluster.

6.

Discussion

Of the 19 sources in our subset, we classify eleven as AGN or quasars, six as starbursts or star-forming galaxies, and two as comp osite or unknown. However, the selection criteria for this subset are not well-defined, and the subset may b e biased towards brighter ob jects, and so these fractions are unlikely to b e representative of the p opulation as a whole. For example, Prandoni


­ 19 ­ et al. (2001) find that 40% of sources < 1 mJy are starburst galaxies. A careful analysis of the distribution of the entire sample of sources from this survey will b e given by Huynh et al. (2005c). Ab out half of the ob jects in our subset have high radio-optical luminosity ratios which app ear to b e quite unlike those seen in the local Universe. However, a detailed comparison requires a significant correction (the k-correction) to b e made to the flux densities of these ob jects, b oth at infrared and radio wavelengths, to comp ensate for the fact that the observed wavelength differs significantly from the emitted wavelength. This correction dep ends sensitively on the intrinsic emission sp ectrum of the ob ject, and the redshift. Even in those cases where redshifts are known, the intrinsic sp ectrum of the galaxy is unknown, and so attempts to produce plots of radio/optical luminosities dep end sensitively up on assuming a template derived from the local galaxy p opulation, which may b e intrinsically different from those at high redshifts. Small differences in these sp ectra result in significantly different conclusions, and so the resulting plots may b e misleading. Thus we are unable to make a detailed comparison of radio-optical ratios with those seen in the local Universe until these ob jects are much b etter characterised. We also note that our lack of knowledge of the intrinsic sp ectra of these ob jects is likely to make photometric redshifts unreliable. In the detailed discussion of Section 5, we have noted several instances where a photometric redshift differs significantly from a sp ectroscopic redshift. However, the reliability of photometric redshifts for the heavily obscured ob jects will increase as high-quality mid-infrared observations b ecome available from the Spitzer Space Telescop e, together with the detailed models and templates based on those observations. A significant fraction of the ob jects in this subset app ear to b e heavily reddened, and fall into the category of "Extremely Red Ob jects" (EROs). A similar result is found in other Deep Fields observations such as those in the Hubble Deep Field North (e.g. Garrett et al. (2000), Richards (2000) ). For example, Waddington et al. (1999) detected a faint radio galaxy in the HDF-N (VLA J123642+621331) at a redshift of 4.42, which in several resp ects may b e similar to some of the ob jects describ ed here. Numerous observations of the EROs at a variety of wavelengths have established that this class includes b oth early-typ e evolved galaxies at z > 1, heavily obscured high ­ z active galaxies and starburst galaxies, and in some cases a combination of these (e.g. Afonso, Mobasher, Chan, & Cram (2001), Brusa et al. (2002), Bergstrom & Wiklind (2004), Georgakakis ¨ et al. (2004) ).

7.

Conclusion

In this first pap er of the series, we have presented the observations and describ ed some of the radio ob jects detected in the Hubble Deep Field South. Subsequent pap ers will present the entire catalogue and discuss identifications, source prop erties, and the statistical prop erties of the sample. The small subset of galaxies describ ed here show that:


­ 20 ­ · The galaxies include a mixture of starburst-dominated and AGN-dominated, with some evidence that the luminosity of some galaxies may b e generated by a combination of these phenomena. · Several of the galaxies have unusually large radio-optical luminosity ratios, and app ear to b e heavily obscured by dust. Whilst the subset discussed in detail here has not b een chosen according to any statistically well-defined procedure, it is already clear from this small sample that some of these galaxies are quite unlike those seen in the local Universe, p erhaps representing an earlier evolutionary stage. This has already b een noted by other authors conducting deep radio surveys, but the small numb ers of such ob jects makes it difficult to characterise these ob jects. It is also likely that many of these ob jects will b e more clearly characterised by mid-infrared results from the Spitzer Space Telescop e.

8.

Acknowledgements

We thank Tommy Wiklind and Stefan Bergstrom for their collab oration in the unpublished SIMBA observations. The Australia Telescop e Compact Array is part of the Australia Telescop e which is funded by the Commonwealth of Australia for op eration as a National Facility managed by CSIRO.

REFERENCES Afonso, J., Mobasher, B., Chan, B., & Cram, L. 2001, ApJ, 559, L101 Barger, A. J., Cowie, L. L., & Richards, E. A. 2000, AJ, 119, 2092 Barger, A. J., Cowie, L. L., Steffen, A. T., Hornschemeier, A. E., Brandt, W. N., & Garmire, G. P. 2001, ApJ, 560, L23 Bergstrom, S. & Wiklind, T. 2004, A&A, 414, 95 ¨ Bertin, E., 2004, http://terapix.iap.fr/rubrique.php?id rubrique=91 Blain, A. W., Chapman, S. C., Smail, I., & Ivison, R. 2004, ApJ, 611, 725 Briggs, D., 1995, PhD thesis, NMIMT. See http://www.aoc.nrao.edu/dissertations/dbriggs/ de Bruyn, G., & Brentjens, M., 2005, in preparation. Brusa, M., Comastri, A., Daddi, E., Cimatti, A., Mignoli, M., & Pozzetti, L. 2002, ApJ, 581, L89 Casertano, S., et al. 2000, AJ,120, 2747


­ 21 ­ Chapman, S. C., et al. 2003, ApJ, 585, 57 Condon, J., J., 1992, ARA&A, 30, 575 Conway, R. G., Garrington, S. T., Perley, R. A., & Biretta, J. A. 1993, A&A, 267, 347 De Breuck, C., van Breugel, W., R¨ ottgering, H. J. A., & Miley, G. 2000, A&AS, 143, 303 Fernandez-Soto et al., 1999, (private communication). (FS) Fontana, A., et al. 2004, in preparation. Franceschini, A., et al. 2003, A&A, 403, 501 Gallimore, J. F. & Beswick, R. 2004, AJ, 127, 239 Garrett, M. A., de Bruyn, A. G., Giroletti, M., Baan, W. A., & Schilizzi, R. T. 2000, A&A, 361, L41 Geller, R. M., Sault, R. J., Antonucci, R., Killeen, N. E. B., Ekers, R., Desai, K., & Whysong, D. 2000, ApJ, 539, 73 Georgakakis, A., Hopkins, A. M., Afonso, J., Sullivan, M., Mobasher, B., & Cram, L. E. 2004, MNRAS, 354, 127 Glazebrook et al., 1998, on http://www.aao.gov.au/hdfs/Redshifts/ (G98) Hopkins, A., Afonso, J., Cram, L., & Mobasher, B. 1999, ApJ, 519, L59 Huynh, Jackson, & Norris, 2005a, submitted to AJ (Pap er 2) Huynh, Jackson, & Norris, 2005b, in preparation. (Pap er 3) Huynh, Jackson, & Norris, 2005c, in preparation. (Pap er 4) Ivezi´ Z., et al. 2002, AJ, 124, 2364 c, Jackson, C. A. 2004, New Astronomy Review, 48, 1187 Labb e, I., Franx, M., Rudnick, G., Forster Schreib er, N.M., van der Werf, P., Rottgering, H., van Starkenburg, L., van de Wel, A., Kuijken, K., Rix, H.-W., Moorwood, A., Daddi, E., van Dokkum, P.G. 2003, AJ, in press Lanzetta et al 2002 ApJ 570 492 (see also www.sess.unysb.edu/astro/hdfs) Lucas, R. A., et al. 2003, AJ, 125, 398 Manchester, R.N., and the ATNF Configuration Study Group, 1984, ATCA Technical Memo, 1984, AT/17.2/003


­ 22 ­ Mann, R. G., et al. 2002, MNRAS, 332, 549 McCarthy 2004, ARAA, 42, 477 Muxlow, T.W.B., Wilkinson, P.N., Richards, A.M.S., Kellermann, K.I., Richards, E.A., Garrett, M.A. 1999, New Astr Reviews, 43, 623 Nonino et al, 1999A&AS, 137, 51 (see also http://www.eso.org/science/eis/eis gallery/HDFS/) Norris et al, in "Persp ectives on Radio Astronomy - Science with Large Antenna Arrays", ed. M.P.van Haarlem, NFRA, 1999 Norris, R.P., Hopkins, A., Sault, R.J., Mitchell, D.A., Ekers, R.D., Badia, F., Higdon, J., Wieringa, M.H., Boyle, B.J., Williams, R.E., 2001, in Proceedings of the ESO/ECF/STScI workshop:Deep Fields, eds Cristiani, S., Renzini, A., Williams, R.E., Springer Nyman, L.-° et al. 2001, The Messenger, 106, 40 A., Oliver, S., Mann, R.G., Carballo, R., Franceschini, A., Rowan-Robinson, M., Kontizas, M., Dap ergolas, A., Kontizas, E., Verma, A., Elbaz, D., Granato, G. L., Silva, L., Rigop oulou, D., Gonzalez-Serrano, J. I., Serjeant, S., Efstathiou, A., van der Werf, P. P. 2001, MNRAS, 332, 536 Outram, P. J., Boyle, B. J., Carswell, R. F., Hewett, P. C., & Williams, R. E. 1999, MNRAS, 305, 685 Palunas et al., 2000, ApJ, 541, 61 Pozzetti, L. & Mannucci, F. 2000, MNRAS, 317, L17 Prandoni, I., Gregorini, L., Parma, P., de Ruiter, H. R., Vettolani, G., Wieringa, M. H., & Ekers, R. D. 2000, A&AS, 146, 31 Prandoni, I., Gregorini, L., Parma, P., de Ruiter, H. R., Vettolani, G., Zanichelli, A., Wieringa, M. H., & Ekers, R. D. 2001, A&A, 369, 787 Reynolds, J.E., 1994, ATNF internal memo AT/39.3/040 Richards, E.A., Kellermann, K.I., Fomalont, E.B., Windhorst, R.A., Partridge, R.B. 1998, AJ, 117, 1039 Richards, E. A. 2000, ApJ, 533, 611 Rigop oulou, D., Franceschini, A., Aussel, H., Genzel, R., Thatte, N., & Cesarsky, C. J. 2002, ApJ, 580, 789 Rodighiero, G., Franceschini, A., & Fasano, G. 2001, MNRAS, 324, 491


­ 23 ­ Rudnick, G., et al. 2001, AJ, 122, 2205 Saracco, P., Giallongo, E., Cristiani, S., D'Odorico, S., Fontana, A., Iovino, A., Poli, F., & Vanzella, E. 2001, A&A, 375, 1 (see also www.stecf.org/hstprogrammes/ISAAC/HDFSdata.html) Sault, R. J. & Wieringa, M. H. 1994, A&AS, 108, 585 Sault, R. J., Teub en, P. J., & Wright, M. C. H. 1995, ASP Conf. Ser. 77: Astronomical Data Analysis Software and Systems IV, 4, 433 Sawicki, M. & Mall´ -Ornelas, G. 2003, AJ, 126, 1208 en Sealey, K. M., Drinkwater, M. J., & Webb, J. K. 1998, ApJ, 499, L135 Sealey, K., Drinkwater, M., Webb, J., Boyle, B.J., 1999, in Looking Deep into the Southern Sky, eds Morganti, R., Couch, W., ESO Astrophysics Symp osia, Springer-Verlag, Berlin Hidelb erg 1999 Snellen, I. A. G., Schilizzi, R. T., Miley, G. K., Bremer, M. N., Rottgering, H. J. A., & van ¨ Langevelde, H. J. 1999, New Astronomy Review, 43, 675 Stevens, J. A., et al. 2003, Nature, 425, 264 Teplitz H.I., Hill, R.S., Malumuth, E.M., Collins, N.R., Gardner, P.P., Woodgate, B.E., 2001, ApJ, 548, 127 Vanzella, E., et al. 2001, AJ, 122, 2190 Vanzella, E., et al. 2002, A&A, 396, 847 Vanzella, E., et al. 2004, A&A, 423, 761 Waddington, I., Windhorst, R. A., Cohen, S. H., Partridge, R. B., Spinrad, H., & Stern, D. 1999, ApJ, 526, L77 Williams, R. E. et al. 1996 AJ, 112, 1335 Williams, R. E. et al. 2000 AJ, 120, 2735 Yahata, N., Lanzetta, K.M., Chen, H-W., Fernandez-Soto, A., Pascarelle, S.M., Yahil, A.M., Puetter, R.C. 2000 ApJ, 538, 493 Yun, M. S., Reddy, N. A., & Condon, J. J. 2001, ApJ, 554, 803

A This preprint was prepared with the AAS L TEX macros v5.2.


­ 24 ­

Table 1: The Fields Searched and the resulting p eak flux and rms in each field Field HST Candidate Pointing centre (J2000) Peak Flux r ms Name Fields RA Dec (mJy) (mJy) 1A/B 3A/QSO 4A 15A/B 12A/B 11A/B 16A 14A 1A & 1B 3A & QSO (see note) 4A 15A & 15B 12A & 12B 11A & 11B 16A 14A 22:50:13.92 22:33:18.63 22:13:33.47 23:30:44.28 00:37:44.31 02:51:22.97 02:49:28.97 04:12:58.98 -60:19:26.2 -60:36:52.0 -61:02:03.0 -60:28:17.0 -62:18:37.8 -60:34:36.0 -63:30:46.5 -59:02:15.8 53 155 69 42 179 59 21 20 0.097 0.116 0.090 0.087 0.193 0.102 0.087 0.082

Table 2. HDF-S Pointing Centres and Total Observing Time
Central frequency 1.4 2.5 5.1 8.7
a

Primary beam FWHP/arcmin 33 22 10 5

Pointing centre (J2000) RA Dec 22:33:25.96 22:33:25.96 22:32:56.22 22:32:56.22 - - - - 60:38:09.0 60:38:09.0 60:33:02.7 60:33:02.7

Total Hours per band 190 190 208 200

Central rms µJy 16.1 11.4 10.9 11.8

GHz GHz GHz GHz

Hours observed are given per 128-MHz band. These numbers should be halved to give the observing time when observing with two 128-MHz bands.


Table 3.

Catalogue of ATHDFS sources within 6.5 arcmin of J2000 22h 32m 56.22", -60 33' 02.7". The columns are describ ed in Section 4.
Name RA (J2000) 7- 7- 6- 1- 6- 2- 5- 8- 6- 0- 6- 5- 4- 8- 4- 0- 3- 0- 8- 7- 2- 2- 4- 1- 3- 3- 6- 6- 5- 9- 6- 7- 4- 0- 5- 603921 603903 603857 603846 603835 603805 603748 603737 603717 603716 603657 603627 603542B 603544 603542A 603539 603537 603537 603523 603520 603459B 603459A 603450 603448 603448 603442 603434 603423 603419 603417 603414 603351 603351 603350 603346 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 32: 32: 32: 33: 33: 32: 32: 32: 32: 33: 32: 33: 32: 32: 32: 32: 32: 32: 33: 32: 32: 32: 32: 33: 32: 32: 32: 32: 32: 33: 33: 33: 32: 33: 32: 53. 58. 45. 07. 04. 48. 54. 32. 25. 26. 36. 16. 32. 29. 32. 53. 45. 24. 38. 23. 30. 29. 47. 07. 12. 43. 16. 31. 45. 11. 27. 29. 43. 06. 58. 70 79 65 13 66 29 53 85 68 01 60 55 56 89 40 09 34 02 84 70 21 21 41 18 39 32 63 69 51 97 67 74 47 07 59 RA 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 69 36 03 30 52 33 38 06 36 49 18 05 12 35 12 31 28 03 13 50 13 14 56 55 54 23 55 26 13 41 05 19 06 09 02 Dec (J2000) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 39: 39: 38: 38: 38: 38: 37: 37: 37: 37: 36: 36: 35: 35: 35: 35: 35: 35: 35: 35: 35: 34: 34: 34: 34: 34: 34: 34: 34: 34: 34: 33: 33: 33: 33: 21. 03. 57. 46. 35. 05. 48. 37. 17. 16. 57. 27. 53. 44. 42. 39. 37. 37. 23. 20. 03. 59. 50. 48. 48. 42. 34. 23. 19. 17. 14. 51. 51. 50. 46. 2 9 3 9 8 6 0 1 4 7 7 5 1 9 5 7 8 7 8 5 5 5 0 8 7 5 9 1 0 1 3 9 0 3 7 Dec S20 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 62 47 04 46 52 44 48 08 84 95 44 05 18 45 16 31 53 04 19 91 29 15 39 08 49 30 84 39 16 52 07 26 20 13 03 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.
cm

Ref 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS

S11

cm

S6cm 0.216 0.204 0.141 0.115 0.396 0.076 0.047 0.06 0.262 0.073 0.066 0.123 0.404

S3cm 0.114 0.238



ma j



min

PA

SN

band L L L L L L L L L L S L L L L L L L L L C L S L L C L L C L S C X L L

qual 2 2 1 2 2 2 2 1 2 2 2 1 1 2 1 2 2 1 1 2 2 1 2 2 2 2 2 2 2 2 1 2 2 1 1

J223253. J223258. J223245. J223307. J223304. J223248. J223254. J223232. J223225. J223326. J223236. J223316. J223232. J223229. J223232. J223253. J223245. J223224. J223338. J223223. J223229. J223229. J223247. J223307. J223212. J223243. J223216. J223231. J223245. J223311. J223327. J223329. J223243. J223306. J223258.

052 058 843 137 067 076 092 645 057 061 649 466 077 529 090 051 259 185 142 113 192 103 093 063 052 070 265 059 456 152 098 452 010

0. 0.

0.

0. 0. 0. 0.

0. 0. 0. 0. 0.

0. 0. 0. 0. 0.

536 074 489 095 451 236 334 836 145 128 081 059 150 498 081 288 658

0 0 2.27 4.9 0 0 3.85 3.63 0 0 0 0 6.82 0 6.67 0 0 3.53 0 9.1 0 0 0 8.75 5.21 0 0 0 0 0 1.38 0 0 4.45 2.23

0. 2.

2.

4. 5.

2. 4.

1. 2.

0.

2. 1.

0 0 5.1 0 0 5.7 73 30.5 75.9 15 17 9 0 0 5.6 0 0 7.4 43 42.4 7.7 2 19.6 45.8 0 0 5.3 0 0 5 0 0 7.5 0 0 56.2 54 -8.9 27.2 0 0 7.9 03 -9.7 30 0 0 9.2 0 0 5.3 27 -32.3 97.1 0 0 18.5 98 -4.4 6.4 0 0 11.9 0 0 20.3 0 0 5.3 85 20.7 5.3 35 -72.7 6.9 0 0 5.1 0 0 5.5 0 0 7.2 0 0 6.9 0 0 5.1 75 -69.7 34.1 0 0 5.3 0 0 8 76 11.5 31.2 33 -16.3 103.9

­ 25 ­


Table 3--Continued
Ref 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS Name J223225. J223247. J223253. J223337. J223302. J223306. J223339. J223327. J223327. J223242. J223234. J223209. J223308. J223323. J223229. J223212. J223317. J223212. J223335. J223331. J223302. J223303. J223254. J223316. J223256. J223304. J223241. J223216. J223303. J223331. J223224. J223236. J223253. J223316. J223329. 0- 6- 1- 5- 8- 2- 4- 9- 9- 6- 2- 7- 6- 2- 5- 9- 7- 9- 3- 6- 1- 1- 4- 0- 4- 8- 4- 6- 9- 1- 7- 5- 7- 8- 1- 603338 603337 603329 603329 603323 603307 603306 603304A 603304B 603258 603257 603253 603251 603249 603243 603234B 603235 603234A 603234 603222 603213 603132 603131 603127 603058 603031 603025 603016 603013 603007 603005 603000 602946 602934 602933 RA (J2000) 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 32: 32: 32: 33: 33: 33: 33: 33: 33: 32: 32: 32: 33: 33: 32: 32: 33: 32: 33: 33: 33: 33: 32: 33: 32: 33: 32: 32: 33: 33: 32: 32: 32: 33: 33: 25. 47. 53. 37. 02. 06. 39. 27. 29. 42. 34. 09. 08. 23. 29. 12. 17. 12. 35. 31. 02. 03. 54. 16. 56. 04. 41. 16. 03. 31. 24. 36. 53. 16. 29. 05 65 15 57 83 26 41 95 28 66 27 71 60 25 55 95 75 90 31 64 18 17 41 09 42 89 42 67 96 14 77 56 78 80 16 RA 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 28 42 58 03 61 09 62 13 44 06 61 58 06 03 07 04 41 03 35 08 33 35 47 43 39 44 30 54 46 31 66 02 38 02 12 Dec (J2000) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 33: 33: 33: 33: 33: 33: 33: 33: 33: 32: 32: 32: 32: 32: 32: 32: 32: 32: 32: 32: 32: 31: 31: 31: 30: 30: 30: 30: 30: 30: 30: 30: 29: 29: 29: 38. 37. 29. 29. 23. 07. 06. 04. 02. 58. 57. 53. 51. 49. 43. 43. 35. 34. 34. 22. 13. 32. 31. 27. 58. 31. 25. 16. 13. 07. 05. 00. 46. 34. 33. 7 0 1 1 8 9 0 8 0 1 4 9 7 2 6 3 2 6 5 2 2 8 5 6 9 2 9 7 6 9 5 6 7 7 4 Dec S20 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 64 56 71 04 40 35 32 17 59 20 79 56 08 04 16 06 38 03 55 11 43 47 41 67 35 25 27 51 05 40 47 03 52 05 22 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 2. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 1. 0.
cm

S11

cm

S6cm 373 093 196 360 108 657 716 120 394 569 -

S3cm 071 077 092 223 -



ma j



min

PA

SN 0 7.5 0 6.8 4 6.3 3 88.1 0 5.2 0 5.6 0 5.6 0 19.8 0 5.4 0 5.3 0 6.1 0 5.1 6 55 0 32.9 0 10.5 5 77.9 0 7.3 7 147.3 0 5.5 4 32.4 0 6.4 0 5.2 1 9 0 5.3 0 8.5 0 8.2 0 9.4 0 5.2 0 5.1 0 9 0 5.1 0 129.7 0 5.1 5 37.4 5 20.3

band L L L L L X L L L X L L L C C L L L L L L L L L L L L L L L L L L C L

qual 2 2 2 1 2 2 2 1 1 2 2 2 1 1 1 1 2 1 2 1 2 2 2 2 2 2 2 2 2 2 2 1 2 1 1

080 075 113 126 051 058 221 059 057 061 821 457 237 466 070 816 054 395 063 052 139 052 074 084 092 052 052 094 055 507 045 003 261

0.718 0.187 0.394 0.140 0.903 1.574 0.286 0.877 0.811 -

0.

6. 2.

0.

0.

0.

0. 0. 0. 0. 0. 0.

0. 0.

6.

5. 7. 3.

6.

0. 0.

3. 1. 5.

0 0 36 58 0 0 0 0 0 0 0 0 59 0 0 84 0 07 0 31 0 0 91 0 0 0 0 0 0 0 0 71 0 45 36

0 0 4.96 1.22 0 0 0 0 0 0 0 0 1.82 0 0 4.24 0 2.77 0 2.46 0 0 3.17 0 0 0 0 0 0 0 0 1.33 0 0.44 1.3

-34. 9.

-32.

­ 26 ­

-5. -54. 2.

64.

-30. 1. 3.


Table 3--Continued
Ref 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT AT HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS HDFS Name J223303. J223303. J223317. J223236. J223307. J223326. J223330. J223307. J223255. J223240. J223311. J223241. J223244. J223317. J223312. J223257. J223259. 0- 0- 7- 2- 7- 9- 5- 0- 9- 7- 5- 5- 5- 1- 3- 4- 9- 602927B 602927A 602916 602855 602853 602850 602849 602827 602810 602755 602725 602719 602719 602714 602707 602657 602654 RA (J2000) 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 22: 33: 33: 33: 32: 33: 33: 33: 33: 32: 32: 33: 32: 32: 33: 33: 32: 32: 01. 03. 17. 36. 07. 26. 30. 07. 55. 40. 11. 41. 44. 17. 12. 57. 59. 82 00 72 20 73 94 55 07 99 71 53 50 57 11 30 44 93 RA 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 41 34 50 31 25 15 51 11 20 34 32 09 33 38 41 39 33 Dec (J2000) - - - - - - - - - - - - - - - - - 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 60: 29: 29: 29: 28: 28: 28: 28: 28: 28: 27: 27: 27: 27: 27: 27: 26: 26: 30. 27. 16. 55. 53. 50. 49. 27. 10. 55. 25. 19. 19. 14. 07. 57. 54. 2 1 0 6 8 0 7 8 1 3 1 8 2 1 9 4 8 Dec S20 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 72 67 72 62 55 22 90 15 38 38 40 13 60 73 51 49 69 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
cm

S11

cm

S6cm 0.101 -

S3

cm

-

ma j



min

PA

SN 7.1 8.2 5.2 7.5 6.5 16.3 5 26.6 10.3 9.2 13.2 28 7 7.6 7 7.3 5.5

band L L L L L L L L L L L L L L L L L

qual 1 1 2 2 2 1 2 1 2 2 2 1 2 2 2 2 2

071 116 059 079 068 182 047 354 108 139 344 351 088 122 078 083 061

0.192 0.244 -

0 7.22 0 0 0 2.93 0 4.58 0 3.77 9.06 0 0 8.19 0 0 0

0 0 3.83 -1.4 0 0 0 0 0 0 1.13 -5.7 0 0 2.6 32.4 0 0 2.56 68 6.09 33.7 0 0 0 0 0.86 -22.1 0 0 0 0 0 0

­ 27 ­


Table 4.
Ref Name 6- 5- 5- 8- 3- 5- 6- 4- 0- 5- 6- 5- 9- 6- 2- 6- 5- 8- 1- S20 (mJy) 603857 603748 603627 603523 603442 603419 603414 603351 603350 603346 603337 603329 603304 603251 603249 603222 603000 602934 602933 0.843 0.092 0.649 0.185 0.063 0.265 0.456 0.098 0.452 1.010 0.075 1.126 0.221 0.821 0.457 0.395 1.507 1.003 0.261 z

Optical and derived prop erties of the subset
V R I J K log (L20 ) WH z -1 23.63 21.69 22.89 22.17 22.15 22.83 23. 22. 24. 22. 24. 07 36 12 43 28 log (S20 /I ) SFR M0 yr - Classification
1

Notes

3 7 12 19 26 29 31 33 34 35 37 39 43 48 49 55 67 69 70
a b c

ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS ATHDFS

J223245. J223254. J223316. J223338. J223243. J223245. J223327. J223243. J223306. J223258. J223247. J223337. J223327. J223308. J223323. J223331. J223236. J223316. J223329.

0.75 0.1798 0.29 0.2250 0.4233 0.4603 0 1.5660 0.1733 1.69 0.5803 2.2380 0 0.64 0 0.4652 0 0.12 0

25.45 19.11 21.75 20.04 20.55 21.95 0 20.28 17.77 0 21.05 17.16 >25 23.07 0 21.24 >25 22.53 0

24.6 18.6 21.2 19.3 19.6 21.1 >28 19.9 17.2 >25 20.0 16.9 >25 21.8 >25 20.6 >25 22.7 23.9

23.39 17.91 20.68 18.61 18.87 20.33 >26 19.45 16.58 25.75 19.16 16.51 >23.5 20.45 0 19.9 >23.5 21.05 0

18.28 18.72 19.84 19. 16. 23. 18. 16. 23. 78 68 44 86 92 91

17.74 18.09 18.94 19. 16. 22. 18. 16. 22. 53 36 31 17 40 42

23.53 25.05 19.58 24.34 18.78 23.01 22.42

5.72 2.57 4.52 3.15 2.79 4.00 >6.50 3.21 2.73 6.74 2.98 3.10 >6.19 4.53 6.17 4.00 >6.02 4.86 5.09

515 6 95 19 17 81 142 27 1579 32 2280 405 123 35

SB? Sy spiral SB elliptical AGN spiral AGN? SB GPS QSO SB Dusty AGN? Massive spiral RL QSO obscured AGN AGN GPS obscured SB ? AGN AGN

b b a h a a d a b b

­ 28 ­

b,f b

g

J and K from Francheschini et al, 2003. J and K from EIS.

J has large uncertainty. I,J,K from Vanzella et al 2001.

d e f

For the calculation of the radio-optical ratio, the 20 cm flux density is assumed to be equal to the 3cm flux density.

For the calculation of the radio-optical ratio, I has been estimated as I=J+0.12, where 0.12 is the mean value of I-J for those sources in this subset for which a measured value of I-J is available.
g For the calculation of the radio-optical ratio, I has b een estimated as I=R-0.83, where 0.83 is the mean value of R-I for those sources in this subset for which a measured value of R-I is available. h

R and I limits based on non-detection in WFPC flanking field observations.


­ 29 ­

Fig. 1.-- Position of the 4 background image is the size field, the inner circles repres resp ectively, which are set to The three p olygons show the and STIS QSO field (left).

p ointings of the ATHDFS relative to the HST WFPC2 field. The of the entire 20 cm primary b eam. Working inwards to the WFPC2 ent the catalogued areas of the 20, 11, 6, and 3 cm primary b eams the level where the sensitivity falls to 39% of that at the b eam centre. p ositions of the HST WFPC field (top right), NICMOS field (lower),


­ 30 ­

Fig. 2.-- ATCA 20 cm image of the region containing the sample of 87 sources listed in Table 3. This sample includes all sources detected, at any wavelength, within the circle, which has a radius of 6.5 arcmin.


­ 31 ­


­ 32 ­


­ 33 ­


­ 34 ­

Fig. 3.-- Radio images (contours) overlaid on CTIO I-band images (greyscale) for each of the sources in our subset. In each case, the named source is at the centre of the image. The radio image of source 41 (ATHDFS 2J223306.2-603307) shows 3 cm data; all others show 20 cm data. Contours are at 3, 5, 10, 20, 50 and 100 â the local rms.