1: %%

2: %% Beginning of file 'sample.tex'

3: %%

4: %% Modified 2005 December 5

5: %%

6: %% This is a sample manuscript marked up using the

7: %% AASTeX v5.x LaTeX 2e macros.

8: 

9: %% The first piece of markup in an AASTeX v5.x document

10: %% is the \documentclass command. LaTeX will ignore

11: %% any data that comes before this command.

12: 

13: %% The command below calls the preprint style

14: %% which will produce a one-column, single-spaced document.

15: %% Examples of commands for other substyles follow. Use

16: %% whichever is most appropriate for your purposes.

17: %%

18: \documentclass[12pt,preprint]{aastex}

19: %  \usepackage{vmargin}

20: %  \setpapersize{A4}

21: %  \setmargins{1.5cm}{1.5cm}% % linker & oberer Rand

22: %             {18cm}{25cm}%   % Textbreite und -hoehe

23: %             {12pt}{25pt}%   % Kopfzeilenhoehe und -abstand

24: %             {0pt}{30pt}%    % \footheight (egal) und Fusszeilenabstand

25: 

26: %% manuscript produces a one-column, double-spaced document:

27: 

28: %\documentclass[manuscript]{aastex}

29: 

30: %% preprint2 produces a double-column, single-spaced document:

31: 

32: %% \documentclass[preprint2]{aastex}

33: 

34: %% Sometimes a paper's abstract is too long to fit on the

35: %% title page in preprint2 mode. When that is the case,

36: %% use the longabstract style option.

37: 

38: %% \documentclass[preprint2,longabstract]{aastex}

39: 

40: %% If you want to create your own macros, you can do so

41: %% using \newcommand. Your macros should appear before

42: %% the \begin{document} command.

43: %%

44: %% If you are submitting to a journal that translates manuscripts

45: %% into SGML, you need to follow certain guidelines when preparing

46: %% your macros. See the AASTeX v5.x Author Guide

47: %% for information.

48: 

49: \newcommand{\vdag}{(v)^\dagger}

50: \newcommand{\myemail}{enno.middelberg@csiro.au}

51: \newcommand{\uJy}{\ensuremath{\mu{\rm Jy}}}

52: \newcommand{\um}{\ensuremath{\mu{\rm m}}}

53: 

54: % For an easier bibliography

55: %\usepackage{natbib}

56: \bibliographystyle{aa}

57: 

58: % enables to put figures and tables sideways

59: % ``figuresright'' fixes the relative positions of table and caption

60: % independent of page number

61: %\usepackage{amsmath}

62: 

63: %% You can insert a short comment on the title page using the command below.

64: 

65: %\slugcomment{Not to appear in Nonlearned J., 45.}

66: 

67: %% If you wish, you may supply running head information, although

68: %% this information may be modified by the editorial offices.

69: %% The left head contains a list of authors,

70: %% usually a maximum of three (otherwise use et al.).  The right

71: %% head is a modified title of up to roughly 44 characters.

72: %% Running heads will not print in the manuscript style.

73: 

74: \shorttitle{ATLAS Radio Observations of ELAIS-S1}

75: \shortauthors{Middelberg et al.}

76: 

77: %% This is the end of the preamble.  Indicate the beginning of the

78: %% paper itself with \begin{document}.

79: 

80: \begin{document}

81: 

82: %% LaTeX will automatically break titles if they run longer than

83: %% one line. However, you may use \\ to force a line break if

84: %% you desire.

85: 

86: \title{Deep ATLAS Radio Observations of the ELAIS-S1/{\it Spitzer}

87: Wide-Area Infrared Extragalctic field}

88: 

89: %% Use \author, \affil, and the \and command to format

90: %% author and affiliation information.

91: %% Note that \email has replaced the old \authoremail command

92: %% from AASTeX v4.0. You can use \email to mark an email address

93: %% anywhere in the paper, not just in the front matter.

94: %% As in the title, use \\ to force line breaks.

95: 

96: \author{Enno Middelberg}

97: \affil{Astronomischies Institut der Universit\"at Bochum, Universit\"atsstr. 150, 44801 Bochum, Germany; middelberg@astro.rub.de}

98: 

99: \author{Ray P. Norris}

100: \affil{Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia}

101: 

102: \author{Tim J. Cornwell}

103: \affil{Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia}

104: 

105: \author{Maxim A. Voronkov}

106: \affil{Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia}

107: 

108: \author{Brian D. Siana}

109: \affil{Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125}

110: 

111: \author{Brian J. Boyle}

112: \affil{Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia}

113: 

114: \author{Paolo Ciliegi}

115: \affil{INAF, Osservatorio Astronomico di Bologna, Via Ranzani 1, I-40127 Bologna, Italy}

116: 

117: \author{Carole A. Jackson}

118: \affil{Australia Telescope National Facility, PO Box 76, Epping NSW 1710, Australia}

119: 

120: \author{Minh T. Huynh}

121: \affil{Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125}

122: 

123: \author{Stefano Berta}

124: \affil{Dipartimento di Astronomia, Universit\`a di Padova, Vicolo dell'Osservatorio 2, 35122 Padova, Italy}

125: 

126: \author{Stefano Rubele}

127: \affil{Dipartimento di Astronomia, Universit\`a di Padova, Vicolo dell'Osservatorio 2, 35122 Padova, Italy}

128: 

129: \author{Carol J. Lonsdale}

130: \affil{Center for Astrophysics and Space Sciences, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0424}

131: 

132: \author{Rob J. Ivison}

133: \affil{UK Astronomy Technology Centre, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ}

134: 

135: \author{Ian Smail}

136: \affil{Institute for Computational Cosmology, Durham University, South Road, Durham DH1 3LE, UK}

137: 

138: \author{Seb J. Oliver}

139: \affil{Astronomy Centre, CPES, University of Sussex, Falmer, Brighton BN1 9QJ, UK}

140: 

141: 

142: 

143: %\author{C. D. Biemesderfer\altaffilmark{4,5}}

144: %\affil{National Optical Astronomy Observatories, Tucson, AZ 85719}

145: %\email{aastex-help@aas.org}

146: %

147: %\and

148: %

149: %\author{R. J. Hanisch\altaffilmark{5}}

150: %\affil{Space Telescope Science Institute, Baltimore, MD 21218}

151: %

152: %% Notice that each of these authors has alternate affiliations, which

153: %% are identified by the \altaffilmark after each name.  Specify alternate

154: %% affiliation information with \altaffiltext, with one command per each

155: %% affiliation.

156: 

157: %\altaffiltext{1}{Visiting Astronomer, Cerro Tololo Inter-American Observatory.

158: %CTIO is operated by AURA, Inc.\ under contract to the National Science

159: %Foundation.}

160: %\altaffiltext{2}{Society of Fellows, Harvard University.}

161: %\altaffiltext{3}{present address: Center for Astrophysics,

162: %    60 Garden Street, Cambridge, MA 02138}

163: %\altaffiltext{4}{Visiting Programmer, Space Telescope Science Institute}

164: %\altaffiltext{5}{Patron, Alonso's Bar and Grill}

165: 

166: %% Mark off your abstract in the ``abstract'' environment. In the manuscript

167: %% style, abstract will output a Received/Accepted line after the

168: %% title and affiliation information. No date will appear since the author

169: %% does not have this information. The dates will be filled in by the

170: %% editorial office after submission.

171: 

172: \begin{abstract}

173: 

174: We have conducted sensitive ($1\,\sigma<30\,\uJy$) 1.4\,GHz radio

175: observations with the Australia Telescope Compact Array of a field

176: largely coincident with infrared observations of the {\it Spitzer}

177: Wide-Area Extragalactic Survey. The field is centred on the European

178: Large Area ISO Survey S1 region and has a total area of

179: $3.9^\circ$. We describe the observations and calibration, source

180: extraction, and cross-matching to infrared sources. Two catalogues are

181: presented; one of the radio components found in the image and one of

182: radio sources with counterparts in the infrared and extracted from the

183: literature. 1366 radio components were grouped into 1276 sources, 1183

184: of which were matched to infrared sources. We discover 31 radio

185: sources with no infrared counterpart at all, adding to the class of

186: Infrared-Faint Radio Sources.

187: 

188: \end{abstract}

189: 

190: %% Keywords should appear after the \end{abstract} command. The uncommented

191: %% example has been keyed in ApJ style. See the instructions to authors

192: %% for the journal to which you are submitting your paper to determine

193: %% what keyword punctuation is appropriate.

194: 

195: \keywords{Catalogs, Galaxies: Active, Galaxies: Evolution, Radio

196: Continuum: Galaxies, Surveys}

197: 

198: 

199: \section{Introduction}

200: 

201: In this paper describing early results from the Australia Telescope

202: Large Area Survey (ATLAS), we present a 1.4\,GHz survey of the

203: European Large Area ISO Survey (ELAIS) S1 field

204: (\citealt{Oliver2000}). This is the second survey paper describing

205: results from ATLAS, and is complementary to the paper by

206: \cite{Norris2006a}, which describes observations of the Chandra Deep

207: Field South.

208: 

209: ATLAS is an ambitious study of galaxies and their evolution since

210: z$\gtrsim$3, using predominantly observations in the radio regime. Two

211: large areas (about 3.5\,deg$^2$ each) have been surveyed with high

212: sensitivity, using the Australia Telescope Compact Array (ATCA) at

213: 1.4\,GHz, to complement multi-wavelength observations in the infrared

214: with the {\it Spitzer} Space Telescope. The immediate goals of the

215: observations are, in brief, to determine whether the radio-infrared

216: relation holds at high redshifts; to search for overdensities of

217: high-z ULIRGs which mark the positions of protoclusters in the early

218: universe; to trace the radio luminosity function to high redshifts; to

219: determine the relative contribution of starbursts and AGN to the

220: overall energy density of the universe; and to open a new parameter

221: space to allow for serendipitous discovery of rare sources. However,

222: surveys such as this have proven in the past to have a substantial

223: impact on longer time-scales, when they are used as a resource in a

224: broad variety of studies.

225: 

226: It was decided early on in this project to observe two separate sky

227: regions, rather than one larger area, to exclude cosmic variance which

228: might affect the results. Both fields extend beyond $2^\circ$ in one

229: dimension, which is sufficient to sample structures at any one

230: redshift which have evolved to more than 150\,Mpc at the present

231: epoch. Nevertheless, such surveys are still prone to cosmic

232: variance. The CDFS is known to contain some large-scale structures

233: (\citealt{Vanzella2005} and references therein), predominantly at

234: redshifts of 0.73 and 1.1. These structures are not obvious clusters,

235: but ``sheets'' in the original CDFS. This finding demonstrates the

236: need to sample large areas for an unbiased view of galaxy formation.

237: 

238: It should be noted that the ELAIS-S1 field has been also observed by

239: \cite{Gruppioni1999} using the ATCA in June 1997, reaching a fairly

240: uniform rms of 80\,\uJy\ across the observed area, a factor of three

241: higher than the average rms in our current observations. Also, the

242: area they observed is slightly larger than the field described

243: here. We have cross-matched their sources to ours and briefly discuss

244: the results.

245: 

246: This paper is organised as follows. Section~\ref{sec:obs} describes

247: the observations and Section~\ref{sec:images} details the source

248: extraction process and the cross-identification of radio sources with

249: the SWIRE catalogue. Section~\ref{sec:cats} provides a description of

250: the catalogues and the literature search for counterparts in other

251: surveys. Section~\ref{sec:class} gives a description of the

252: classification and a few individual sources are described in more

253: detail in Section~\ref{sec:individual}. We provide a short analysis of

254: the radio-infrared relation in Section~\ref{sec:r-ir} and our

255: conclusions in Section~\ref{sec:conc}.

256: 

257: \section{Observations}

258: \label{sec:obs}

259: 

260: As of December 2006 we have completed about 50\,\% of the planned

261: observations of the ELAIS-S1 field. The sensitivity here is slightly

262: higher than in the CDFS, partly because of slightly longer integration

263: times (between 10.5\,h per pointing and 13.5\,h per pointing, compared

264: to 8.2\,h per pointing over most of the CDFS), and partly because

265: there is a strong interfering source in the CDFS field. However, a

266: 3.8\,Jy source (PKS\,0033-44) limits the dynamic range of the ELAIS-S1

267: observations even though it is well outside our pointings. An overview

268: of the observed area is reproduced in Figure~\ref{fig:overview}.

269: 

270: 

271: \subsection{Radio observations}

272: 

273: The radio observations were carried out on 27 separate days between 9

274: January 2004 and 24 June 2005 with the Australia Telescope Compact

275: Array (ATCA), with a total net integration time on the pointings of

276: 231\,h, in a variety of configurations to maximise the $(u,v)$

277: coverage (Table~\ref{tab:obs}). However, $(u,v)$ coverage is probably

278: not a crucial factor in aperture synthesis when the field is dominated

279: by point sources as in our case. 3.89\,deg$^2$ in the ELAIS-S1 field

280: were analysed (this is the total area in the mosaic where the primary

281: beam response is $>10$\,\%) in a mosaic consisting of 20 overlapping

282: pointings (Table~\ref{tab:coords}). The full width at half maximum

283: (FWHM) of the primary beams at 1.4\,GHz is $35^\prime$. The pointings

284: were observed for one minute each, and the calibrator 0022-423 was

285: observed after each cycle of 20 pointings for two minutes. Amplitude

286: calibration was done using PKS\,1934-638 as a primary calibrator,

287: which was observed for 10\,min before or after each observing run. It

288: was assumed to have a flux density of 15.012\,Jy at 1.34\,GHz and

289: 14.838\,Jy at 1.43\,GHz, corresponding to the centres of the two ATCA

290: frequency bands. Each band had a bandwidth of 128\,MHz over 33

291: channels, so the total observing bandwidth was 256\,MHz. In the

292: observation in early 2004, the higher band was only slightly affected

293: by terrestrial radio-frequency interference (RFI), but this

294: deteriorated in 2004, requiring considerable effort to edit the data

295: properly to avoid losing a large fraction of good data. The lower band

296: was mostly free of RFI and required little editing. In the early

297: stages of the project in 2004, only pointings 1-12 were observed (the

298: upper three rows of circles in Figure~\ref{fig:overview}), but the

299: surveyed area was extended in 2005 by adding pointings 13-20 to the

300: field. The new pointings were initially observed for a longer time to

301: catch up with the older pointings, resulting in a different $(u,v)$

302: coverage and a little more integration time. Pointings 1-12 have net

303: integration times of 10.5\,h per pointing, whereas pointings 13-20

304: have net integration times of 13.5\,h per pointing. After editing, the

305: predicted noise level is 22\,\uJy\ in the centre of the

306: mosaic. Towards the image edges, the noise level increases due to

307: primary beam attenuation.

308: 

309: 

310: 

311: \subsubsection{Calibration}

312: 

313: The data were calibrated using Miriad (\citealt{Sault1995}) standard

314: procedures, following recommendations for high dynamic range

315: imaging. The raw data come in RPFITS format, and are converted into

316: the native Miriad format using ATLOD. ATLOD discarded every other

317: frequency channel (which are not independent from one another, hence

318: no information is lost) and flagged one channel in the

319: higher-frequency band which contained a multiple of 128\,MHz, and thus

320: was affected by self-interference at the ATCA. We also did not use the

321: channels at either end of the band where the sensitivity dropped

322: significantly. The resulting data set contained two frequency bands,

323: with 13 channels and 12 channels respectively, all of which are 8\,MHz

324: wide, and so the total net bandwidth in the data was

325: 25$\times$8\,MHz=200\,MHz.

326: 

327: The data were bandpass-calibrated to prepare for RFI removal with

328: Pieflag (\citealt{Middelberg2006a}). Pieflag derives baseline-based

329: statistics from a channel which is free of, or only very slightly

330: affected by, RFI, and searches the other channels for outliers and

331: sections of high noise. It is therefore important to

332: bandpass-calibrate the data before using it. Pieflag eliminated all

333: RFI-affected data which would have been flagged in a visual

334: inspection, while minimising the amount of erroneously flagged good

335: data. On average, approximately 3\,\% and 15\,\% of the data were

336: flagged in the lower and higher band, respectively.

337: 

338: After flagging, the bandpass calibration was removed as it may have

339: been affected by RFI in the calibrator observations, and

340: repeated. Phase and amplitude fluctuations throughout each observing

341: run were corrected using the interleaved calibrator scans, and the

342: amplitudes were scaled by correction factors derived from the

343: observations of the primary calibrator. The data were then split by

344: pointing and imaged.

345: 

346: \subsubsection{Imaging}

347: 

348: The data for each of the 20 pointings were imaged separately using

349: uniform weighting and a pixel size of $2.0''$. The 25 frequency

350: channels were gridded separately to increase the $(u,v)$ coverage. The

351: relatively high fractional bandwidth of the observations (15\,\%)

352: required the use of Miriad's implementation of multi-frequency clean,

353: MFCLEAN, for deconvolution, to account for spectral indices across the

354: observed bandwidth and to reduce sidelobes. After a first iteration,

355: model components with a flux density of more than $1\,{\rm

356: mJy\,beam^{-1}}$ were used in phase self-calibration, to correct

357: residual phase errors. The data were then re-imaged and cleaned with

358: 5000 iterations, at which point the sidelobes of strong sources were

359: found to be well below the thermal noise. The models were convolved

360: with a Gaussian of $10.26^{\prime\prime}\times7.17^{\prime\prime}$

361: diameter at position angle $0^\circ$, and the residuals were

362: added. The restored images of the 20 pointings were merged in a linear

363: mosaic using the Miriad task LINMOS, which divides each image by a

364: model of the primary beam to account for the attenuation towards the

365: edges of the image, and then uses a weighted average for pixels which

366: are covered by more than one pointing. As a result, pixels at the

367: mosaic edges have a higher noise level. Regions beyond a perimeter

368: where the primary beam response drops below 3\,\% (this occurs at a

369: radius of $35.06^\prime$ from the centre of a pointing) were blanked.

370: 

371: Imaging of the data turned out to be challenging, but the sensitivity

372: of the image presented here is mostly within 25\,\% of the predicted

373: sensitivity. In the south-eastern corner of the mosaic, mild artefacts

374: remain due to the presence of the 3.8\,Jy radio source PKS\,0033-44,

375: which is located about $1^\circ$ away from the centre of pointing

376: 13. The noise level of the present image could only be reached by

377: including this source in the CLEANed area. Because of a combination of

378: the high resolution of the image, the distance of the source from the

379: pointing centre, and the requirement of multi-frequency clean to

380: provide images which are three times larger than the area to clean, we

381: had to generate very large images with 16384 pixels on a side, plus an

382: additional layer of the same extent for the spectral index. These

383: images cannot be handled by 32-bit computers because the required

384: memory exceeds their address space, and we had to employ a 64-bit

385: machine to image the data.

386: 

387: The cause of the residual sidelobes is still the subject of

388: investigation. At present, we suspect that non-circularities in the

389: sidelobe pattern of the primary beams are the culprit. The interfering

390: source sits on the maximum of the first antenna sidelobe and, in the

391: course of the observations, rotates through the sidelobe pattern due

392: to the azimuthal mounting of the antennas. We have measured the

393: primary beam response of two ATCA antennas in great detail using a

394: geostationary satellite at 1.557\,GHz, and derived a model of their

395: far-field reception patterns. Unfortunately, we were unable to

396: reproduce the sidelobe pattern arising from PKS\,0033-44, and no

397: correction from this exercise has been applied to our data.

398: 

399: \subsubsection{Image properties}

400: 

401: The sensitivity is not uniform across the image due to primary beam

402: attenuation, however, it is quite homogeneous in the central

403: 1\,deg$^2$ of the image. A cumulative histogram of an image of the

404: noise in this area, made with SExtractor (\citealt{Bertin1996}),

405: revealed that only 2\,\% of the image has a noise of 22\,\uJy\ or

406: less, consistent with the theoretical expectations. However, 75\,\% of

407: pixels have a noise of 27.5\,\uJy\ or less, which is 25\,\% higher

408: than the expected noise.  We conclude that in the regions which are

409: not affected by sidelobes from PKS\,0033-44 the sensitivity of the

410: image is close to the theoretical expectations.

411: 

412: \subsubsection{Clean bias}

413: 

414: Clean bias is an effect in deconvolution which redistributes flux from

415: point sources to noise peaks in the image, thereby reducing the flux

416: density of the real sources. As the amount of flux which is taken away

417: from real sources is independent of the sources' flux densities, the

418: fractional error this causes is largest for weak sources. The effect

419: of clean bias in our calibration procedure has been analysed as

420: follows. We have added to the data of one pointing (rms=30\,\uJy)

421: 132 point sources at random positions, with flux densities between

422: 150\,\uJy\ and 3\,mJy. The number of sources added with a particular

423: SNR were N=40 (5\,$\sigma$), 15 (6\,$\sigma$), 15 (7\,$\sigma$), 15

424: (8\,$\sigma$), 15 (9\,$\sigma$), 10 (10\,$\sigma$), 10 (12\,$\sigma$),

425: 5 (16\,$\sigma$), 3 (20\,$\sigma$), 2 (30\,$\sigma$), 1

426: (50\,$\sigma$), and 1 (100\,$\sigma$). 

427: 

428: The data have then been used to form an image in the same way as the

429: final image was made, and each source's flux density was extracted

430: using a Gaussian fit, and then divided by the injected flux

431: density. This test was repeated 30 times to build up significant

432: statistics, in particular for the sources with high SNR. We found that

433: using 5000 iterations in cleaning did not cause a significant clean

434: bias ($<2.5$\,\%), whereas using 50000 iterations did cause the

435: extracted fluxes to be reduced by up to 5\,\%

436: (Figure~\ref{fig:clean_bias}). We conclude that the flux densities in

437: our catalogue are only marginally affected by clean bias.

438: 

439: 

440: \subsubsection{Comparison to earlier observations}

441: 

442: We have compared the flux densities and positions of components in our

443: image to those of \cite{Gruppioni1999} (G99). We have obtained their

444: image and selected 83 isolated components with $S>0.5$\,mJy in regions

445: where our noise level was below 30\,\uJy. All sources were detected

446: with an SNR$>$6 by G99. These sources were grouped into bins with

447: $2^n$\,mJy to $2^{n+1}$\,mJy (n=-1,0,1,2,3,4), the flux densities were

448: extracted from G99's image using the same methods as for our image,

449: and the ratios $S/S_{\rm G99}$ were computed. The median ratios were

450: 1.36 (0.5\,mJy-1\,mJy), 1.43 (1\,mJy-2\,mJy), 1.19 (2\,mJy-4\,mJy),

451: and 1.16 (4\,mJy-8\,mJy). The two highest bins with 8\,mJy-16\,mJy and

452: 16\,mJy-32\,mJy had ratios very close to one, but only two

453: measurements each, hence the statistics are not reliable.

454: 

455: Our analysis suggests that our flux densities are systematically

456: higher than G99's, although $S/S_{\rm G99}$ appears to approach unity

457: towards higher flux densities. We have found that our flux extraction

458: procedure reproduces the catalogued fluxes of G99 to within 3\,\%,

459: hence we conclude that our procedure is working and the effect is

460: real. The cause of this discrepancy is not known, but possible

461: explanations are (i) calibration differences: G99 used amplitude

462: self-calibration with a relatively sparse array and very short

463: solution intervals, which may have affected the flux densities. We did

464: not use amplitude self-calibration at all because it was not found to

465: improve our image significantly; (ii) $(u,v)$ coverage: G99 had only

466: one configuration at the ATCA whereas we had six, yielding more

467: constraints in deconvolution. Also G99 imaged the data from both IF

468: bands separately and averaged the images later, thus using only one

469: half of their data in the deconvolution stage.

470: 

471: We also tested for a systematic position offset between the components

472: of G99 and ours. We found a mean offset of

473: $0.112^{\prime\prime}\pm0.016^{\prime\prime}$ in right ascension and

474: of $0.017^{\prime\prime}\pm0.022^{\prime\prime}$ in declination, and

475: conclude that systematic position offsets are negligible.

476: 

477: 

478: \subsection{{\it Spitzer} observations}

479: 

480: The {\it Spitzer} observations of the ELAIS-S1 field were carried out as

481: part of the {\it Spitzer} Wide-Area Infra-Red Extragalactic Survey (SWIRE)

482: program, as described by \cite{Lonsdale2003}. Approximately

483: 6.9\,deg$^2$ were observed in the ELAIS-S1 region at 3.6\,\um,

484: 4.5\,\um, 5.8\,\um, and 8.0\,\um\ with the Infrared Array Camera (IRAC)

485: and at 24\,\um\ with the Multiband Imaging Photometer (MIPS). The

486: sensitivities in the five bands are 4.1\,\uJy, 8.5\,\uJy, 48.2\,\uJy,

487: 53.0\,\uJy, and 252\,\uJy. Here we use the fourth data release,

488: containing more than 400.000 sources (Surace et al. 2007, in prep.).

489: 

490: \subsection{Optical observations}

491: 

492: The optical follow-up observations of the ELAIS-S1 field are called

493: the ESO-Spitzer Imaging Extragalactic Survey (ESIS). The observations

494: were carried out with the Wide Field Imager (WFI) of the 2.2\,m

495: La~Silla ESO-MPI telescope and with the VIsible Multi Object

496: Spectrograph (VIMOS) on the VLT, to cover 5\,deg$^2$ in BVRIz. Only

497: approximately 1.5\,deg$^2$ have yet been covered (\citealt{Berta2006})

498: with WFI and these data are included in our catalogue. The filters

499: used are WFI B/99 (later replaced by B/123), V/89 and Rc/162, and the

500: catalogue is 95\,\% complete at 25$^{\rm m}$ in the B and V bands, and

501: at 24.5$^{\rm m}$ in the R band (all in Vega units).

502: 

503: 

504: 

505: \section{Image analysis}

506: \label{sec:images}

507: 

508: \subsection{Component extraction}

509: 

510: This section describes the procedure we used to extract radio sources

511: from the image and to subsequently match these radio sources to

512: infrared sources. In our terminology, a radio component is a region of

513: radio emission which is best described by a Gaussian. Close radio

514: doubles are very likely to be best represented by two Gaussians and

515: are therefore deemed to consist of two components. Single or multiple

516: components are called a radio source if they are deemed to belong to

517: the same object.

518: 

519: The rms of the image varies from 22\,\uJy\ in the best regions to

520: 1\,mJy towards the edges of the image, caused by primary beam

521: attenuation. It is therefore not possible to use the same cutoff, in

522: terms of flux density per pixel, above which a pixel is deemed a

523: detection of a source and below which pixels are deemed

524: noise. Furthermore, flux densities measured towards the image edges

525: are increasingly affected by uncertainties in the primary beam model,

526: and we therefore restricted our image analysis to those sources which

527: lie in regions where the theoretical sensitivity is below 250\,\uJy.

528: 

529: We used SExtractor to create an image of the noise, by which we

530: divided the radio image to obtain an image of signal-to-noise (called

531: the SNR map). The SNR map has unity noise everywhere, and can be

532: analysed using a single criterion. We used the Miriad task IMSAD to

533: look for islands of SNR$>$5, and then used this catalogue as input for

534: a visual inspection of the total intensity image at the locations

535: where SNR$>$5. Sources were re-fitted using the total intensity image,

536: and were subsequently cross-identified with IR sources and

537: classified. If either of the two axes of a fitted Gaussian was smaller

538: than the restoring beam's corresponding axis, the fit was repeated

539: using a Gaussian with the major and minor axis fixed to the restoring

540: beam and the position angle set to zero. Also very weak sources were

541: in general found to be better represented with fixed-size Gaussians.

542: 

543: The integrated flux densities of extended sources were obtained by

544: integrating over the source area, rather than summing the flux

545: densities of their constituents. This is because even multiple

546: Gaussians are seldom a proper representation of extended sources, and,

547: using this technique, even very faint emission between components is

548: included.

549: 

550: We estimated the error of the integrated flux densities using Eq.~(1)

551: in \cite{Schinnerer2004}, which is based on \cite{Condon1997},

552: assuming a relative error of the flux calibration of 5\,\% whereas

553: \cite{Schinnerer2004} assumed 1\,\%. In the case of extended sources,

554: where the integrated flux density was measured by integrating over a

555: polygon in the image, we assumed a 5\,\% scaling error and added to

556: that in quadrature an empirical error arising from the shape and size

557: of the area over which was integrated:

558: 

559: \begin{equation}

560: \Delta S = \sqrt{(0.05S)^2 + (10^{-7}/S)^2}

561: \end{equation}

562: 

563: where $S$ is the flux density in Jy. For extended sources with 10\,mJy,

564: 1\,mJy and 0.5\,mJy, the total errors are thus 0.5\,mJy (5\,\%),

565: 0.11\,mJy (11\,\%), and 0.2\,mJy (40\,\%), respectively, which

566: describe the errors found empirically reasonably well.

567: 

568: The uncertainties in the peak flux densities were estimated using

569: Eq.~(21) in \cite{Condon1997}. Errors in right ascension and

570: declination are the formal errors from Gaussian fits plus a

571: $0.1^{\prime\prime}$ uncertainty from the calibrator position added in

572: quadrature.

573: 

574: 

575: \subsubsection{Deconvolution of components from the restoring beam}

576: 

577: All radio components were deconvolved from the restoring beam. If a

578: deconvolution was not possible, or the deconvolution yielded a point

579: source, the component was deemed to be unresolved and the deconvolved

580: size has been left blank in Table~\ref{tab:radio1}.

581: 

582: 

583: \subsection{The cross-identification process}

584: 

585: The cross-matching process was as follows. The region used for the fit

586: and the ellipse indicating the FWHM were inspected, along with the

587: corresponding parts of the following images: the SNR map, a naturally

588: weighted radio image with lower resolution (and slightly higher

589: sensitivity), a superuniformly weighted radio image with higher

590: resolution (but lower sensitivity), and the 3.6\,\um\ SWIRE image with

591: superimposed SNR map contours. Furthermore, the locations of

592: catalogued SWIRE sources within $30^{\prime\prime}$ of the fitted

593: coordinates were shown on the SWIRE images.

594: 

595: It was then decided (i) whether each radio component was a genuine

596: detection or likely to be a sidelobe, (ii) how it could be matched to

597: catalogued or uncatalogued SWIRE sources, (iii) whether multiple radio

598: components constituted radio emission from a single object, and (iv)

599: whether extended components needed to be divided into

600: sub-components. Emission deemed to be sidelobes was found

601: predominantly towards the edges of the image and associated with, and

602: directly adjacent to, strong sources.

603: 

604: Most sidelobes were discovered because the naturally-weighted image,

605: which has a different sidelobe pattern and higher sensitivity but

606: lower resolution, showed no evidence of a source at the position of a

607: possible source in the uniformly weighted image. Our catalogue of

608: radio components contains 1366 components; 15 were deemed to be

609: sidelobes and have been marked as such (all with SNR$<$6), leaving 1351

610: genuine radio components.

611: 

612: The separation between a radio component and a SWIRE source cannot

613: easily be used as a parameter in the cross-identification process. In

614: some cases, despite a relatively large separation, the

615: cross-identification is relatively clear because the SWIRE source is

616: extended towards the SWIRE source, such as in the examples shown in

617: Figure~\ref{fig:large_sep}.

618: 

619: 1134 radio components (88.9\,\%) could be characterized properly by a

620: single Gaussian and were judged to be the only radio counterpart of a

621: catalogued SWIRE source. A fraction of these displayed the morphology

622: of doubles in a superuniformly weighted image. 15 sources (1.2\,\%)

623: had uncatalogued SWIRE counterparts.

624: 

625: 32 sources (2.5\,\%) were deemed to be radio doubles, consisting of

626: two radio components; and 26 sources (2.0\,\%) consisted of two or

627: more components, displaying more complex morphologies like triplets or

628: core-jet morphologies.

629: 

630: We have tested for systematic radio-IR position offsets by calculating

631: the average offsets of 533 sources which consist of a single radio

632: component and a catalogued SWIRE counterpart, and have SNR$>$10. The

633: offsets have a mean of $(0.08\pm0.03)''$ in right ascension and

634: $(0.06\pm0.03)''$ in declination. Although the offset is formally

635: significantly different from zero, we note that it is less than a

636: tenth of a pixel in the radio image.

637: 

638: All sources classified as radio doubles have been reviewed using the

639: criteria developed by \cite{Magliocchetti1998} based on an analysis of

640: the FIRST survey (\citealt{Becker1995}): Two radio components are

641: likely to be part of a double when (a) their separation measured in

642: arcsec is less than $100(S/100)^{0.5}$, where S is the total flux of

643: the two constituents, and (b) their flux densities do not differ by

644: more than a factor of four. We give the results of this test in the

645: source table. It should be noted that the test has been derived from a

646: large sample of galaxies (236000) and is purely

647: empirical. Furthermore, the FIRST survey is shallower (rms=0.14\,mJy)

648: than ours, and so statistically may contain different objects from the

649: survey presented here. It is therefore no surprise that some of our

650: radio sources which are clearly radio doubles fail the test. For

651: example, S923 (Figure~\ref{fig:examples}) fails on criterion (a), but

652: satisfies criterion (b).

653: 

654: \subsection{The false cross-identification rate}

655: 

656: Because the SWIRE field has a high IR source density (58700 sources

657: per deg$^2$), there is some chance that a radio component falls within

658: a few arcseconds of an infrared source, although it is not physically

659: connected to it. The two sources would be wrongly cross-matched, and

660: hence there is a fraction of erroneous cross-identifications in our

661: source catalogue, an upper limit of which we estimate as follows.

662: 

663: From the source density, one can calculate that on average 0.01423

664: SWIRE sources fall within $1''$ of any one point in the field. The

665: number of SWIRE sources within $1''-2''$ of any one point is 0.0427,

666: and within $2''-3''$ is 0.0711. We have confirmed these numbers

667: experimentally by searching near several hundred random positions in

668: the SWIRE catalogue.

669: 

670: In our catalogue, 1134 sources consist of a single component and have

671: a good SWIRE cross-identification. Of these, 656 have a separation of

672: less than $1''$, 350 have a separation of $1''-2''$, 86 have a

673: separation of $2''-3''$, and 45 have a separation of more than $3''$.

674: 

675: Of the original 1134 cross-identifications, a fraction of 0.01423, or 16

676: sources, are expected to be purely coincidental, and are found among

677: the 656 sources with sub-arcsec cross-identifications. Thus, a

678: fraction of $16/656=0.024$ is likely to be coincidental (and wrong).

679: 

680: With the sub-arcsec cross-identifications now accounted for,

681: $(1134-656)=478$ sources remain. Of these, a fraction of 0.0427, or 20

682: sources, will fall within $1''-2''$ of an infrared source by

683: coincidence. Thus, a fraction of $20/481=0.042$ is coincidental.

684: 

685: Repeating the steps above leaves $(1134-656-350)=128$ sources which have

686: not yet been cross-identified. Putting 128 sources randomly on the

687: SWIRE image yields a coincidental counterpart within $2''-3''$ for a

688: fraction of 0.0711, or 9 sources. Thus, a fraction of $9/86=0.105$ is

689: coincidental. The statistics of the remaining 45 sources with

690: separations $>3''$ are not meaningful because the separations are

691: dominated by extended radio objects which are not expected to coincide

692: with infrared sources. A summary of this estimate is shown in

693: Table~\ref{tab:fxids}.

694: 

695: We stress that the rates of false cross-identifications given here are

696: upper limits. A false cross-identification does not only require a

697: false counterpart within a few arcseconds of the radio position, but

698: it also requires that the true counterpart is much fainter than the

699: false one. The second requirement reduces the rate of false

700: cross-identifications well below our estimate.

701: 

702: 

703: \section{The component and source catalogues}

704: \label{sec:cats}

705: 

706: Following \cite{Norris2006a} we publish two catalogues, one containing

707: the component data (Table~\ref{tab:radio1}), and one containing radio

708: sources and their infrared counterparts (Table~\ref{tab:source1}).

709: 

710: \subsection{The components catalogue}

711: 

712: The component catalogue contains information about Gaussian components

713: fitted to the radio image. It does not contain information about the

714: grouping of components to sources, which is exclusively left to the

715: source catalogue in the next section. 

716: 

717: 

718: 

719: \subsection{The source catalogue}

720: 

721: The distribution of integrated flux densities for the 1276 catalogued

722: sources is shown in Figure~\ref{fig:fluxes}. We have carried out a

723: Kolmogorov-Smirnov-test using the ELAIS-S1 and CDFS integrated flux

724: densities, to test the likelihood that the two samples are drawn from

725: the same parent distribution. Because the two fields have different

726: sensitivities, the catalogues cannot be compared in full, but a flux

727: cutoff has to be used. Furthermore, we restricted the test to sources

728: within $48'$ of the field centres and required an rms of between

729: 30\,\uJy\ to 40\,\uJy, to exclude regions with elevated noise levels

730: towards the image edges. We find that when only sources with flux

731: densities of more than 0.5\,mJy are compared (ELAIS-S1: 137 sources,

732: CDFS: 130 sources), the probability that the two samples are drawn

733: from the same parent distribution is 73.7\,\%. When the minimum

734: required flux density is lowered to 0.4\,mJy (ELAIS-S1: 179 sources,

735: CDFS: 151 sources) or 0.3\,mJy (ELAIS-S1: 222 sources, CDFS: 186

736: sources), the probabilities are 18.3\,\% and 25.8\,\%,

737: respectively. We conclude that in regions with similar sensitivities

738: the distribution of radio sources in the ELAIS-S1 and CDFS fields is

739: identical at a flux density level of more than 0.3\,mJy.

740: 

741: In the source catalogue, comments on the cross-match and the radio

742: morphology are recorded as follows. If no comment is given, we had no

743: doubt about the identification; "uncatalogued counterpart" means that

744: we had no doubt that the radio source is associated with a clearly

745: visible IR source at either 3.6\,\um\ or 24\,\um\ which is not listed

746: in the SWIRE catalogue (data release 4); "IFRS" means that a radio

747: source could not be reasonably matched to any IR counterpart at all

748: and did not appear to be associated with another radio source;

749: "confused XID" means that the radio source is likely to be associated

750: with the SWIRE source we give, but that other sources cannot be ruled

751: out; "unclear XID" means that the identification was too ambiguous to

752: make a reasonable choice. We also comment on the radio morphology if

753: the source is anything but a single Gaussian. In the case of

754: multiple-component sources we give the component numbers which were

755: deemed to be associated with the source, and we comment on extension

756: or blending with other radio sources. The coordinates of sources are

757: generally those of the radio observations, but in the case of sources

758: with more than one component and with a clear IR counterpart, the

759: SWIRE coordinates have been adopted as the source position. In the

760: case of more than one component without a clear IR component the

761: flux-weighted mean of the radio components has been used.

762: 

763: 

764: 

765: \subsection{Identification of sources with other catalogues and literature data}

766: 

767: The ELAIS-S1 region has already been surveyed with the ATCA at

768: 1.4\,GHz by \cite{Gruppioni1999} with a $1\,\sigma$ sensitivity of

769: 80\,\uJy, and we have cross-matched their catalogue to ours, resulting

770: in 366 matches. We have also searched the NASA Extragalactic

771: Database\footnote{\url{http://nedwww.ipac.caltech.edu/index.html}} for

772: objects within $2^{\prime\prime}$ of the sources in our catalogue, and

773: found matches to 105 sources, sometimes with multiple names. We mostly

774: give the designations from the ELAIS 15\,\um\ catalogue

775: (\citealt{Oliver2000}), the APMUKS catalogue (\citealt{Maddox1990}),

776: and the 2MASS catalogue (\citealt{Skrutskie2006}). These

777: cross-identifications have been included in Table~\ref{tab:source1}.

778: 

779: We searched for available redshifts and found that 59 objects within

780: $2^{\prime\prime}$ of our sources had catalogued redshifts, mostly

781: from \cite{LaFranca2004}. A histogram of the redshifts is shown in

782: Figure~\ref{fig:redshifts}. Unlike in the CDFS, there is no indication

783: of cosmic large-scale structure in this histogram. However, the number

784: of redshifts is small and may not be sufficient to show inconspicuous

785: large-scale structure.

786: 

787: We have cross-matched our source catalogue to sources from the Sydney

788: University Molonglo Sky Survey (SUMSS, \citealt{Bock1999}), which is a

789: survey of the southern sky at 843\,MHz, using the Molonglo Observatory

790: Synthesis Telescope (MOST). The sensitivity of SUMSS is of the order

791: of $\sim1\,{\rm mJy\,beam^{-1}}$, so that the faintest sources have a

792: flux density of the order of $\sim5\,{\rm mJy}$. Assuming a spectral

793: index of $\alpha=-0.7$ ($S\propto\alpha$), typical for radio emission

794: from AGN, this corresponds to $S_{\rm 1.4GHz}=3.5\,{\rm mJy}$, so only

795: the brightest ATLAS sources will be present in SUMSS. We found 73

796: matches to sources catalogued in the 1 June 2006 data

797: release\footnote{\url{http://www.astrop.physics.usyd.edu.au/sumsscat/}}

798: and give the results in Table~\ref{tab:sumss}. There were no SUMSS

799: sources without 1.4\,GHz counterpart in the ATCA image.

800: 

801: We have also searched for counterparts in the AT20G survey

802: (\citealt{Ricci2004}), which is a survey of the southern sky with the

803: ATCA at 18\,GHz, but found no match.

804: 

805: 

806: 

807: \section{Classification}

808: \label{sec:class}

809: 

810: \subsection{AGN}

811: 

812: Here we discuss the classification of sources as AGN based on their

813: morphology, their ratio of 24\,\um\ to radio flux, and using

814: literature information.

815: 

816: Radio sources exhibiting a double-lobed, triple, or more complex

817: structures, e.g. with jets, were generally classified as AGN. Examples

818: for classification based on morphology are S829, S923, S926, S930.1,

819: S1192 and S1189, all of which are described in more detail in

820: Section~\ref{sec:individual}.

821: 

822: From {\it Spitzer} 24\,\um\ and VLA 20\,cm detections in the First

823: Look Survey (\citealt{Condon2003}), \cite{Appleton2004} derive

824: $q_{24}={\rm log}(S_{24\um}/S_{20{\rm cm}})=0.84$. Here, sources with

825: ${\rm log}(S_{24\um}/S_{20{\rm cm}})<-0.16$, i.e. more than 10 times

826: the radio flux density as predicted by the radio-infrared relation,

827: were classified as AGN. 

828: 

829: In total 75 sources were classified as AGN based on their radio

830: morphology, 128 sources based on their radio excess compared to the

831: radio-infrared relation at 24\,\um, and 9 sources had been classified

832: as AGN by \cite{LaFranca2004}, based on optical spectroscopy. 14

833: sources were classified as AGN using more than one criterion, and thus

834: 198 sources were classified as AGN. We note that, with the exception

835: of the three sources S606, S717, and S813, all sources which were

836: classified as AGN based on their morphology {\it and} which had

837: catalogued 24\,\um\ flux densities were also classified as AGN based

838: on their departure from the radio-infrared relation as given by

839: \cite{Appleton2004}. We plot the 20\,cm flux densities as a function

840: of 24\,\um\ flux densities of all sources in

841: Figure~\ref{fig:radio-ir}. AGN are plotted separately according to how

842: they have been classified.

843: 

844: 

845: 

846: 

847: \subsection{Infrared-Faint Radio Sources (IFRS)}

848: 

849: We find 31 sources with no detectable infrared counterpart. These

850: sources have been dubbed ``Infrared-Faint Radio Sources'', or IFRS, by

851: \cite{Norris2006a}, and may be more extreme cases of the ``Optically

852: Invisible Radio Sources'' found by \cite{Higdon2005}. As they are

853: invisible in the optical and infrared, there is only very limited

854: information available. A Kolmogorov-Smirnov test reveals a 80.1\,\%

855: probability that the distribution of flux densities of the IFRS is

856: drawn from the same parent distribution as all flux densities, though

857: the IFRS sources tend to have lower radio flux densities than the

858: entire sample. We have carried out VLBI observations of three IFRS in

859: our sample (S427, $S_{\rm 1.4GHz}=21.4\,{\rm mJy}$; S509, $S_{\rm

860: 1.4GHz}=22.2\,{\rm mJy}$; and S775, $S_{\rm 1.4GHz}=3.6\,{\rm mJy}$)

861: to determine whether they are AGN hosts and the contribution to the

862: arcsec-scale flux density from an AGN, but the results are not yet

863: available. However, \cite{Norris2007b} have successfully detected an

864: IFRS in the CDF-S field.

865: 

866: 

867: 

868: \section{Notes on individual sources}

869: \label{sec:individual}

870: 

871: We comment on a few examples to illustrate the cross-identification

872: process. The sources discussed here are shown in

873: Figure~\ref{fig:examples}.

874: 

875: 

876: \begin{itemize}

877: \item{\it Sources S829 and S829.2} S829 is an example of a mildly

878: extended object, which is best represented with two Gaussians (C829

879: and C829.1). However, at higher resolution it begins to resemble a

880: double-lobed or core-jet morphology, and it is centred on the IR

881: source SWIRE4\_J003251.97-433037.2 in between the two radio

882: components, and hence was classified as an AGN. The nearby source

883: S829.2 is a relatively weak radio source (0.30\,mJy) which coincides

884: ($\theta \sim 1.5^{\prime\prime}$) with SWIRE4\_J003251.87-433016.7.

885: 

886: \item{\it Sources S923, S930, S930.1 and S926} These four sources lie within

887: less than $2^\prime$ of each other and form a striking quartet at

888: first sight. Source S923 is without doubt a classical double-lobed

889: radio galaxy with an integrated flux density of 5.9\,mJy. The SWIRE

890: source SWIRE4\_J003042.10-432335.4 is located on the line connecting

891: the peaks of the two constituent radio components C923 and C931 and is

892: therefore identified as the infrared counterpart. Source S930 is an

893: otherwise inconspicuous radio source with an infrared counterpart,

894: SWIRE4\_J003038.21-432305.9, within $0.28^{\prime\prime}$. The

895: naturally weighted image indicates a faint bridge of emission between

896: components C930 and C941, hence both components have been grouped into

897: S930.  It blends with source S930.1, which consists of the two faint

898: radio components C930.1 and C930.2 with integrated flux densities of

899: 0.47\,mJy and 0.55\,mJy, respectively. In between components C930.1

900: and C930.2 is a very faint, uncatalogued infrared source, and thus the

901: radio morphology together with the location of the IR source indicates

902: that this is a double-lobed radio galaxy. Source S655 has a relatively

903: bright IR counterpart (SWIRE4\_J003035.03-432341.6) centred on the

904: brighter one of its two radio components C926 and C926.1, with

905: 2.7\,mJy and 0.93\,mJy, respectively. Unlike in sources S923 and

906: S930.1, the IR source is centred on one of its constituents, but it

907: was deemed more likely that both C926 and C926.1 are associated with

908: this source rather than to postulate that C926 is the radio

909: counterpart to SWIRE4\_J003035.03-432341.6 and that C926.1 is a

910: separate source with no IR counterpart.

911: 

912: \item{\it Sources S1189 and S1197} Source S1189 is a beautifully extended,

913: large radio source. The number of constituent radio components is

914: somewhat arbitrary, but there exists a low-SNR bridge of emission

915: which connects the main part of the source and component C1212,

916: $2^\prime$ north, as well as many more low-SNR patches in between. The

917: brightest part of S1189 is centred on SWIRE4\_J003427.54-430222.5

918: (separation $0.75''$), which we therefore identify as IR counterpart,

919: and which has the morphology of an elliptical galaxy in optical

920: images. Source S1197, $0.60^{\prime\prime}$ from

921: SWIRE4\_J003419.55-430151.7, is unlikely to be connected to

922: S1189. S1189 has the shape and extent of Wide-Angle Tail galaxies

923: (WAT, \citealt{Miley1980}) such as NGC\,1265

924: (\citealt{Owen1978}). Their characteristic C-shape is believed to be

925: caused by ram pressure against the jets while the galaxy is moving

926: through the intracluster medium. The jets in S1189 are strongly bent

927: backwards and almost touch another at the far ends. This is

928: illustrated in Figure~\ref{fig:1189}, where we have drawn contours

929: beginning at $2\,\sigma$ to emphasize the effect. WAT radio sources

930: can be used as cluster signposts (e.g.,

931: \citealt{Blanton2003}), but there is no known cluster at the position

932: of S1189, although there is a little overdensity of galaxies at

933: $115''$ to its south-west, centred on a bright galaxy with elliptical

934: morphology. The source is in the SUMSS catalogue with a flux density

935: of 36.2\,mJy, compared to a 1.4\,GHz flux density of

936: 45.0\,mJy. However, it is clearly extended in the SUMSS image and not

937: well represented by a Gaussian. Integrating over the source area in

938: the SUMSS image yielded a total flux density of 51\,mJy, and hence a

939: spectral index of $\alpha=-0.25$.

940: 

941: \item{\it Source S1192} This source is an example of a triple radio

942: galaxy. It consists of the three components C1192, C1192.1, and

943: C1192.2 with mJy flux densities. The brightest component, C1192.1, is

944: centred on the IR source SWIRE4\_J003320.68-430203.6. The other two

945: components are several arcsec away from the nearest IR sources, and

946: the overall morphology thus indicates that this is a bent triple radio

947: galaxy. It therefore also could be a WAT.

948: 

949: \item{\it Source S773} Source S773 is a rather faint radio

950: source with $S_{\rm 20cm}=0.37\,{\rm mJy}$, but it has a very bright

951: infrared counterpart with $S_{24\um}=28\,{\rm mJy}$ within $0.72''$,

952: and is one of the few objects clearly visible in the SWIRE 70\,\um\

953: image. Its unusual ratio of ${\rm log}_{10}(S_{24\um}/S_{\rm

954: 20cm})=1.89$ lets it clearly stand out in Figure~\ref{fig:radio-ir} as

955: a separated dot in the bottom right corner. It has been classified by

956: \cite{LaFranca2004} as a type 1 AGN (based on optical line widths in

957: excess of 1200\,km/s) at redshift 0.143. Furthermore, it is one of the

958: brightest X-ray sources found by

959: \cite{Alexander2001} in a {\it BeppoSAX} survey of the ELAIS-S1

960: region.

961: 

962: \item{\it Source S1081} Source S1081 is a very extended ($B_{\rm

963: maj}=93''$), low-surface brightness source. Nevertheless, its

964: integrated flux density is 2.4\,mJy, and there is no obvious

965: association with any one of the many nearby infrared sources. It is

966: unlikely to be a sidelobe, as this region of the image is very good

967: and free of artefacts. Furthermore, its extent indicates that it is

968: not a noise spike, which would have a similar size as the restoring

969: beam. We have convolved the radio image of \cite{Gruppioni1999} with a

970: $1'$ restoring beam, to increase its sensitivity to extended

971: structures, but their image was not sensitive enough to confirm or

972: refute the reality of S1081. The nature of this object is unclear:

973: given its size and the lack of a strong component it could be a

974: cluster radio relic. Such objects are interpreted as leftovers of

975: cluster mergers.

976: 

977: 

978: \end{itemize}

979: 

980: 

981: \section{The radio-infrared relation}

982: \label{sec:r-ir}

983: 

984: One of the goals of the ATLAS project is to trace the radio-infrared

985: relation to very faint flux levels, to determine whether the relation

986: exists in the early universe. Using {\it Spitzer} and VLA observations

987: of the First Look Survey (\citealt{Condon2003}), \cite{Appleton2004}

988: have determined a value of $q_{24}={\rm log}(S_{24\um}/S_{\rm

989: 20cm})=0.84\pm0.23$. Here, we note that \cite{Boyle2007} have employed

990: a statistical analysis of the ELAIS-S1 radio image at the known

991: positions of SWIRE sources. They find $q_{24}=1.46$ using the

992: observations presented here, and exactly the same value of $q_{24}$

993: using the CDFS observations of \cite{Norris2006a}. \cite{Boyle2007}

994: present an extensive description of the analysis and of simulations,

995: and we refer the reader to their paper for details. We note, however,

996: that the discrepancy of the value of $q_{24}$ found by

997: \cite{Appleton2004} and \cite{Boyle2007} remains unresolved.

998: 

999: We plot in Figure~\ref{fig:q24} a histogram of all individual values

1000: of $q_{24}$ where 24\,\um\ fluxes were available, without

1001: k-correction. We also indicate on the diagram the distribution (also

1002: without k-correction) found by \cite{Appleton2004}. We note that these

1003: authors also presented a k-correction for their data, but it was too

1004: small to reconcile their result with the result by

1005: \cite{Boyle2007}. The tail towards lower values of $q_{24}$ can be

1006: explained as arising from AGN, which have a radio excess and so do not

1007: obey the radio-infrared relation. Conversely, the sharp cutoff of the

1008: histogram at $q_{24}$ is caused by a lack of objects with an infrared

1009: excess. This is expected when one interprets the infrared emission as

1010: arising from star formation, which in turn generates radio emission

1011: according to the radio-infrared relation. \cite{Appleton2004} excluded

1012: AGN based on spectroscopic observations and thus their sample is not

1013: contaminated by AGN, and they do not see the tail towards low values

1014: of $q_{24}$.

1015: 

1016: We note that the distribution of $q_{24}$ found by \cite{Norris2006a}

1017: has a different shape than ours. It is rather constant between

1018: $q_{24}=-0.5$ and $q_{24}=1.5$, and indicates a double-peaked

1019: distribution. However, the differences in sensitivity make it

1020: impossible to construct similar samples from the CDFS data presented

1021: by \cite{Norris2006a} and the data presented here. We therefore

1022: postpone a detailed analysis of the distribution of $q_{24}$ to the

1023: time when the ATLAS survey is complete.

1024: 

1025: \section{Conclusions}

1026: \label{sec:conc}

1027: 

1028: We have presented the first data from the ATLAS observations of the

1029: ELAIS-S1 region, and a list of 1276 radio sources extracted from the

1030: image. Radio sources have been matched to infrared SWIRE sources and

1031: been classified as AGN if the morphology, radio-to-infrared ratio, or

1032: the literature indicated so. We discover another 31 Infrared-Faint

1033: Radio Sources, bringing the total number of these objects found with

1034: the ATLAS survey to 55, and find no significant difference between the

1035: distribution of source flux densities between the ELAIS-S1 and the CDFS

1036: at $S_{\rm 20\um}>0.3\,{\rm mJy}$. We find a distribution of

1037: $q_{24}={\rm log}(S_{\rm 20\um}/S_{\rm 20cm})$ which is in broad

1038: agreement with the distribution found by

1039: \cite{Appleton2004}. No further interpretation of our data is

1040: presented, partly because other essential information such as

1041: redshifts is not yet available and partly because the observations are

1042: not yet complete.

1043: 

1044: \acknowledgments{The Australia Telescope Compact Array is operated by the CSIRO

1045: Australia Telescope National Facility. IRS acknowledges support from

1046: the Royal Society. This research has made extensive use of the

1047: NASA/IPAC Extragalactic Database (NED) which is operated by the Jet

1048: Propulsion Laboratory, California Institute of Technology, under

1049: contract with the National Aeronautics and Space Administration.}

1050: 

1051: %\bibliography{refs}

1052: \begin{thebibliography}{33}

1053: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1054: 

1055: \bibitem[{{Alexander} {et~al.}(2001){Alexander}, {La Franca}, {Fiore},

1056:   {Barcons}, {Ciliegi}, {Danese}, {Della Ceca}, {Franceschini}, {Gruppioni},

1057:   {Matt}, {Matute}, {Oliver}, {Pompilio}, {Wolter}, {Efstathiou},

1058:   {H{\'e}raudeau}, {Perola}, {Perri}, {Rigopoulou}, {Rowan-Robinson}, \&

1059:   {Serjeant}}]{Alexander2001}

1060: {Alexander}, D.~M., {La Franca}, F., {Fiore}, F., {et~al.} 2001, \apj, 554, 18

1061: 

1062: \bibitem[{{Appleton} {et~al.}(2004){Appleton}, {Fadda}, {Marleau}, {Frayer},

1063:   {Helou}, {Condon}, {Choi}, {Yan}, {Lacy}, {Wilson}, {Armus}, {Chapman},

1064:   {Fang}, {Heinrichson}, {Im}, {Jannuzi}, {Storrie-Lombardi}, {Shupe},

1065:   {Soifer}, {Squires}, \& {Teplitz}}]{Appleton2004}

1066: {Appleton}, P.~N., {Fadda}, D.~T., {Marleau}, F.~R., {et~al.} 2004, \apjs, 154,

1067:   147

1068: 

1069: \bibitem[{{Becker} {et~al.}(1995){Becker}, {White}, \& {Helfand}}]{Becker1995}

1070: {Becker}, R.~H., {White}, R.~L., \& {Helfand}, D.~J. 1995, \apj, 450, 559

1071: 

1072: \bibitem[{{Berta} {et~al.}(2006){Berta}, {Rubele}, {Franceschini}, {Held},

1073:   {Rizzi}, {Lonsdale}, {Jarrett}, {Rodighiero}, {Oliver}, {Dias}, {Buttery},

1074:   {Fiore}, {La Franca}, {Puccetti}, {Fang}, {Shupe}, {Surace}, \&

1075:   {Gruppioni}}]{Berta2006}

1076: {Berta}, S., {Rubele}, S., {Franceschini}, A., {et~al.} 2006, \aap, 451, 881

1077: 

1078: \bibitem[{{Bertin} \& {Arnouts}(1996)}]{Bertin1996}

1079: {Bertin}, E. \& {Arnouts}, S. 1996, \aaps, 117, 393

1080: 

1081: \bibitem[{{Blanton} {et~al.}(2003){Blanton}, {Gregg}, {Helfand}, {Becker}, \&

1082:   {White}}]{Blanton2003}

1083: {Blanton}, E.~L., {Gregg}, M.~D., {Helfand}, D.~J., {Becker}, R.~H., \&

1084:   {White}, R.~L. 2003, \aj, 125, 1635

1085: 

1086: \bibitem[{{Bock} {et~al.}(1999){Bock}, {Large}, \& {Sadler}}]{Bock1999}

1087: {Bock}, D.~C.-J., {Large}, M.~I., \& {Sadler}, E.~M. 1999, \aj, 117, 1578

1088: 

1089: \bibitem[{{Boyle} {et~al.}(2007){Boyle}, {Cornwell}, {Middelberg}, {Norris},

1090:   {Appleton}, \& {Smail}}]{Boyle2007}

1091: {Boyle}, B.~J., {Cornwell}, T.~J., {Middelberg}, E., {et~al.} 2007, \mnras,

1092:   accepted

1093: 

1094: \bibitem[{{Colless} {et~al.}(2001){Colless}, {Dalton}, {Maddox}, {Sutherland},

1095:   {Norberg}, {Cole}, {Bland-Hawthorn}, {Bridges}, {Cannon}, {Collins}, {Couch},

1096:   {Cross}, {Deeley}, {De Propris}, {Driver}, {Efstathiou}, {Ellis}, {Frenk},

1097:   {Glazebrook}, {Jackson}, {Lahav}, {Lewis}, {Lumsden}, {Madgwick}, {Peacock},

1098:   {Peterson}, {Price}, {Seaborne}, \& {Taylor}}]{Colless2001}

1099: {Colless}, M., {Dalton}, G., {Maddox}, S., {et~al.} 2001, \mnras, 328, 1039

1100: 

1101: \bibitem[{{Condon}(1997)}]{Condon1997}

1102: {Condon}, J.~J. 1997, \pasp, 109, 166

1103: 

1104: \bibitem[{{Condon} {et~al.}(2003){Condon}, {Cotton}, {Yin}, {Shupe},

1105:   {Storrie-Lombardi}, {Helou}, {Soifer}, \& {Werner}}]{Condon2003}

1106: {Condon}, J.~J., {Cotton}, W.~D., {Yin}, Q.~F., {et~al.} 2003, \aj, 125, 2411

1107: 

1108: \bibitem[{{Gruppioni} {et~al.}(1999){Gruppioni}, {Ciliegi}, {Rowan-Robinson},

1109:   {Cram}, {Hopkins}, {Cesarsky}, {Danese}, {Franceschini}, {Genzel},

1110:   {Lawrence}, {Lemke}, {McMahon}, {Miley}, {Oliver}, {Puget}, \&

1111:   {Rocca-Volmerange}}]{Gruppioni1999}

1112: {Gruppioni}, C., {Ciliegi}, P., {Rowan-Robinson}, M., {et~al.} 1999, \mnras,

1113:   305, 297

1114: 

1115: \bibitem[{{Higdon} {et~al.}(2005){Higdon}, {Higdon}, {Weedman}, {Houck}, {Le

1116:   Floc'h}, {Brown}, {Dey}, {Jannuzi}, {Soifer}, \& {Rieke}}]{Higdon2005}

1117: {Higdon}, J.~L., {Higdon}, S.~J.~U., {Weedman}, D.~W., {et~al.} 2005, \apj,

1118:   626, 58

1119: 

1120: \bibitem[{{Jones} {et~al.}(2004){Jones}, {Saunders}, {Colless}, {Read},

1121:   {Parker}, {Watson}, {Campbell}, {Burkey}, {Mauch}, {Moore}, {Hartley},

1122:   {Cass}, {James}, {Russell}, {Fiegert}, {Dawe}, {Huchra}, {Jarrett}, {Lahav},

1123:   {Lucey}, {Mamon}, {Proust}, {Sadler}, \& {Wakamatsu}}]{Jones2004}

1124: {Jones}, D.~H., {Saunders}, W., {Colless}, M., {et~al.} 2004, \mnras, 355, 747

1125: 

1126: \bibitem[{{La Franca} {et~al.}(2004){La Franca}, {Gruppioni}, {Matute},

1127:   {Pozzi}, {Lari}, {Mignoli}, {Zamorani}, {Alexander}, {Cocchia}, {Danese},

1128:   {Franceschini}, {H{\'e}raudeau}, {Kotilainen}, {Linden-V{\o}rnle}, {Oliver},

1129:   {Rowan-Robinson}, {Serjeant}, {Spinoglio}, \& {Verma}}]{LaFranca2004}

1130: {La Franca}, F., {Gruppioni}, C., {Matute}, I., {et~al.} 2004, \aj, 127, 3075

1131: 

1132: \bibitem[{{Lonsdale} {et~al.}(2003){Lonsdale}, {Smith}, {Rowan-Robinson},

1133:   {Surace}, {Shupe}, {Xu}, {Oliver}, {Padgett}, {Fang}, {Conrow},

1134:   {Franceschini}, {Gautier}, {Griffin}, {Hacking}, {Masci}, {Morrison},

1135:   {O'Linger}, {Owen}, {P{\' e}rez-Fournon}, {Pierre}, {Puetter}, {Stacey},

1136:   {Castro}, {Del Carmen Polletta}, {Farrah}, {Jarrett}, {Frayer}, {Siana},

1137:   {Babbedge}, {Dye}, {Fox}, {Gonzalez-Solares}, {Salaman}, {Berta}, {Condon},

1138:   {Dole}, \& {Serjeant}}]{Lonsdale2003}

1139: {Lonsdale}, C.~J., {Smith}, H.~E., {Rowan-Robinson}, M., {et~al.} 2003, \pasp,

1140:   115, 897

1141: 

1142: \bibitem[{{Maddox} {et~al.}(1990){Maddox}, {Efstathiou}, {Sutherland}, \&

1143:   {Loveday}}]{Maddox1990}

1144: {Maddox}, S.~J., {Efstathiou}, G., {Sutherland}, W.~J., \& {Loveday}, J. 1990,

1145:   \mnras, 243, 692

1146: 

1147: \bibitem[{{Magliocchetti} {et~al.}(1998){Magliocchetti}, {Maddox}, {Lahav}, \&

1148:   {Wall}}]{Magliocchetti1998}

1149: {Magliocchetti}, M., {Maddox}, S.~J., {Lahav}, O., \& {Wall}, J.~V. 1998,

1150:   \mnras, 300, 257

1151: 

1152: \bibitem[{{Middelberg}(2006)}]{Middelberg2006a}

1153: {Middelberg}, E. 2006, Publications of the Astronomical Society of Australia,

1154:   23, 64

1155: 

1156: \bibitem[{{Miley}(1980)}]{Miley1980}

1157: {Miley}, G. 1980, \araa, 18, 165

1158: 

1159: \bibitem[{{Norris} {et~al.}(2006){Norris}, {Afonso}, {Appleton}, {Boyle},

1160:   {Ciliegi}, {Croom}, {Huynh}, {Jackson}, {Koekemoer}, {Lonsdale},

1161:   {Middelberg}, {Mobasher}, {Oliver}, {Polletta}, {Siana}, {Smail}, \&

1162:   {Voronkov}}]{Norris2006a}

1163: {Norris}, R.~P., {Afonso}, J., {Appleton}, P.~N., {et~al.} 2006, \aj, 132, 2409

1164: 

1165: \bibitem[{{Norris} {et~al.}(2007){Norris}, {Tingay}, {Phillips}, {Middelberg},

1166:   {Deller}, \& {Appleton}}]{Norris2007b}

1167: {Norris}, R.~P., {Tingay}, S.~J., {Phillips}, C., {et~al.} 2007, \mnras

1168: 

1169: \bibitem[{{Oliver} {et~al.}(2000){Oliver}, {Rowan-Robinson}, {Alexander},

1170:   {Almaini}, {Balcells}, {Baker}, {Barcons}, {Barden}, {Bellas-Velidis},

1171:   {Cabrera-Guerra}, {Carballo}, {Cesarsky}, {Ciliegi}, {Clements}, {Crockett},

1172:   {Danese}, {Dapergolas}, {Drolias}, {Eaton}, {Efstathiou}, {Egami}, {Elbaz},

1173:   {Fadda}, {Fox}, {Franceschini}, {Genzel}, {Goldschmidt}, {Graham},

1174:   {Gonzalez-Serrano}, {Gonzalez-Solares}, {Granato}, {Gruppioni},

1175:   {Herbstmeier}, {H{\'e}raudeau}, {Joshi}, {Kontizas}, {Kontizas},

1176:   {Kotilainen}, {Kunze}, {La Franca}, {Lari}, {Lawrence}, {Lemke},

1177:   {Linden-V{\o}rnle}, {Mann}, {M{\'a}rquez}, {Masegosa}, {Mattila}, {McMahon},

1178:   {Miley}, {Missoulis}, {Mobasher}, {Morel}, {N{\o}rgaard-Nielsen}, {Omont},

1179:   {Papadopoulos}, {Perez-Fournon}, {Puget}, {Rigopoulou}, {Rocca-Volmerange},

1180:   {Serjeant}, {Silva}, {Sumner}, {Surace}, {Vaisanen}, {van der Werf}, {Verma},

1181:   {Vigroux}, {Villar-Martin}, \& {Willott}}]{Oliver2000}

1182: {Oliver}, S., {Rowan-Robinson}, M., {Alexander}, D.~M., {et~al.} 2000, \mnras,

1183:   316, 749

1184: 

1185: \bibitem[{{Owen} {et~al.}(1978){Owen}, {Burns}, \& {Rudnick}}]{Owen1978}

1186: {Owen}, F.~N., {Burns}, J.~O., \& {Rudnick}, L. 1978, \apjl, 226, L119

1187: 

1188: \bibitem[{{Puccetti} {et~al.}(2006){Puccetti}, {Fiore}, {D'Elia}, {Pillitteri},

1189:   {Feruglio}, {Grazian}, {Brusa}, {Ciliegi}, {Comastri}, {Gruppioni},

1190:   {Mignoli}, {Vignali}, {Zamorani}, {La Franca}, {Sacchi}, {Franceschini},

1191:   {Berta}, {Buttery}, \& {Dias}}]{Puccetti2006}

1192: {Puccetti}, S., {Fiore}, F., {D'Elia}, V., {et~al.} 2006, \aap, 457, 501

1193: 

1194: \bibitem[{{Ricci} {et~al.}(2004){Ricci}, {Sadler}, {Ekers}, {Staveley-Smith},

1195:   {Wilson}, {Kesteven}, {Subrahmanyan}, {Walker}, {Jackson}, \& {De

1196:   Zotti}}]{Ricci2004}

1197: {Ricci}, R., {Sadler}, E.~M., {Ekers}, R.~D., {et~al.} 2004, \mnras, 354, 305

1198: 

1199: \bibitem[{{Sault} {et~al.}(1995){Sault}, {Teuben}, \& {Wright}}]{Sault1995}

1200: {Sault}, R.~J., {Teuben}, P.~J., \& {Wright}, M.~C.~H. 1995, in ASP Conf. Ser.

1201:   77: Astronomical Data Analysis Software and Systems IV, 433

1202: 

1203: \bibitem[{{Schinnerer} {et~al.}(2004){Schinnerer}, {Carilli}, {Scoville},

1204:   {Bondi}, {Ciliegi}, {Vettolani}, {Le F{\`e}vre}, {Koekemoer}, {Bertoldi}, \&

1205:   {Impey}}]{Schinnerer2004}

1206: {Schinnerer}, E., {Carilli}, C.~L., {Scoville}, N.~Z., {et~al.} 2004, \aj, 128,

1207:   1974

1208: 

1209: \bibitem[{{Serjeant} {et~al.}(2001){Serjeant}, {Efstathiou}, {Oliver},

1210:   {Surace}, {H{\'e}raudeau}, {Linden-V{\o}rnle}, {Gruppioni}, {La Franca},

1211:   {Rigopoulou}, {Morel}, {Crockett}, {Sumner}, {Rowan-Robinson}, \&

1212:   {Graham}}]{Serjeant2001}

1213: {Serjeant}, S., {Efstathiou}, A., {Oliver}, S., {et~al.} 2001, \mnras, 322, 262

1214: 

1215: \bibitem[{{Shectman} {et~al.}(1996){Shectman}, {Landy}, {Oemler}, {Tucker},

1216:   {Lin}, {Kirshner}, \& {Schechter}}]{Shectman1996}

1217: {Shectman}, S.~A., {Landy}, S.~D., {Oemler}, A., {et~al.} 1996, \apj, 470, 172

1218: 

1219: \bibitem[{{Skrutskie} {et~al.}(2006){Skrutskie}, {Cutri}, {Stiening},

1220:   {Weinberg}, {Schneider}, {Carpenter}, {Beichman}, {Capps}, {Chester},

1221:   {Elias}, {Huchra}, {Liebert}, {Lonsdale}, {Monet}, {Price}, {Seitzer},

1222:   {Jarrett}, {Kirkpatrick}, {Gizis}, {Howard}, {Evans}, {Fowler}, {Fullmer},

1223:   {Hurt}, {Light}, {Kopan}, {Marsh}, {McCallon}, {Tam}, {Van Dyk}, \&

1224:   {Wheelock}}]{Skrutskie2006}

1225: {Skrutskie}, M.~F., {Cutri}, R.~M., {Stiening}, R., {et~al.} 2006, \aj, 131,

1226:   1163

1227: 

1228: \bibitem[{{Vanzella} {et~al.}(2005){Vanzella}, {Cristiani}, {Dickinson},

1229:   {Kuntschner}, {Moustakas}, {Nonino}, {Rosati}, {Stern}, {Cesarsky}, {Ettori},

1230:   {Ferguson}, {Fosbury}, {Giavalisco}, {Haase}, {Renzini}, {Rettura}, {Serra},

1231:   \& {The Goods Team}}]{Vanzella2005}

1232: {Vanzella}, E., {Cristiani}, S., {Dickinson}, M., {et~al.} 2005, \aap, 434, 53

1233: 

1234: \bibitem[{{Wegner} {et~al.}(2003){Wegner}, {Bernardi}, {Willmer}, {da Costa},

1235:   {Alonso}, {Pellegrini}, {Maia}, {Chaves}, \& {Rit{\'e}}}]{Wegner2003}

1236: {Wegner}, G., {Bernardi}, M., {Willmer}, C.~N.~A., {et~al.} 2003, \aj, 126,

1237:   2268

1238: 

1239: \end{thebibliography}

1240: 

1241: \clearpage

1242: 

1243: \begin{figure}

1244: \includegraphics[width=\linewidth]{f1.eps}

1245: \caption{An overview of the observed area. The circles indicate the

1246: 20 antenna pointings and the FWHM of the primary beams. The thin

1247: contours show noise levels of 25\,\uJy, 35\,\uJy, and 45\,\uJy, as

1248: calculated by SExtractor (\citealt{Bertin1996}). The thick contour

1249: indicates where the predicted sensitivity is 250\,\uJy\ and marks the

1250: area which we have analysed. The image has been clipped where the

1251: response of the antenna primary beams has dropped to below 3\,\% of

1252: its peak value.}

1253: \label{fig:overview}

1254: \end{figure}

1255: 

1256: \begin{figure}

1257: \includegraphics[width=\linewidth]{f2.eps}

1258: \caption{Cumulative histogram of the pixel values of the rms map in the

1259: central 1\,deg$^2$ of the observed area.}

1260: \label{fig:cumhist}

1261: \end{figure}

1262: 

1263: \begin{figure}

1264: \includegraphics[width=\linewidth]{f3.eps}

1265: \caption{Results from our tests for clean bias. Shown is the median

1266: normalized flux of sources extracted from simulated images as a

1267: function of SNR. Using 5000 iterations in cleaning does not produce a

1268: significant clean bias, but using 50000 iteration does, although the

1269: bias is comparatively small.}

1270: \label{fig:clean_bias}

1271: \end{figure}

1272: 

1273: 

1274: 

1275: \begin{figure}

1276: \includegraphics[width=\linewidth]{f4.eps}

1277: \caption{Examples of sources with relatively large ($>3''$)

1278: separations between the fitted radio position and the catalogued SWIRE

1279: position. Shown are the radio SNR contours from SNR=4 and increasing

1280: by factors of $\sqrt{2}$, superimposed on the SWIRE 3.6\,\um\ image as

1281: greyscale. {\it Left:} S517 is strong and clearly extended towards

1282: SWIRE4\_J003815.62-435142.0, which was deemed to be associated despite

1283: a separation of $4.4''$. {\it Middle:} Source S655 is separated by

1284: $5.3''$ from its SWIRE counterpart. Shown here is a portion of the

1285: radio image made with super-uniform weighting, and hence higher

1286: resolution, which shows the extension clearly. {\it Right:} Source

1287: S1034 is similarly extended towards a SWIRE source, with a separation

1288: of $3.4''$.}

1289: \label{fig:large_sep}

1290: \end{figure}

1291: 

1292: 

1293: 

1294: \begin{figure}

1295: \includegraphics[width=\linewidth]{f5.eps}

1296: \caption{A histogram of the integrated flux densities of the sources in

1297: our survey. A histogram of the IFRS flux densities is drawn with

1298: dashed lines.}

1299: \label{fig:fluxes}

1300: \end{figure}

1301: 

1302: \begin{figure}

1303: \includegraphics[width=\linewidth]{f6.eps}

1304: \caption{A histogram of the 59 redshifts available for objects in our

1305: catalogue, taken from the literature. There is no hint of large-scale

1306: structure, but this may be hidden by too few data points.}

1307: \label{fig:redshifts}

1308: \end{figure}

1309: 

1310: 

1311: \begin{figure}

1312: \includegraphics[width=\linewidth]{f7.eps}

1313: \caption{20\,cm vs. 24\,\um\ flux densities of all sources, with AGN

1314: plotted separately. Symbols indicate the type of AGN classification:

1315: pluses show non-AGN; crosses indicate AGN classified based on an

1316: 10-fold excess of radio emission compared to the infrared-radio

1317: emission derived by \cite{Appleton2004}; filled squares indicate AGN

1318: classified based on their radio morphology; and filled circles

1319: indicate sources classified as AGN in the literature.  The line

1320: indicates $q_{24}={\rm log}(S_{24\um}/S_{\rm 20cm})=0.84$ found by

1321: \cite{Appleton2004}. The flattening of the distribution towards lower

1322: values of $S_{24\um}$ is caused by the limited sensitivity of the

1323: radio observations, which at low 24\,\um\ flux densities are only able

1324: to pick up objects with comparatively high radio flux densities. For

1325: a detailed analysis of the radio-infrared relation derived from the

1326: ELAIS-S1 radio and 24\,\um\ data, see Boyle et al. (2007).}

1327: \label{fig:radio-ir}

1328: \end{figure}

1329: 

1330: \begin{figure}

1331: \includegraphics[width=0.9\linewidth]{f8.eps}

1332: \caption{\small Six sample extracts from the radio image (contours),

1333: superimposed on the 3.6\,\um\ {\it Spitzer} image

1334: (greyscale). Contours start at SNR=4 and increase by factors of

1335: two. The rms in the images is 27\,\uJy\ (top left), 46\,\uJy\ (top

1336: right), 49\,\uJy\ (middle left), 43\,\uJy\ (middle right), 29\,\uJy\

1337: (lower left), and 28\,\uJy\ (lower right). See the text for a

1338: detailed description of these sources.}

1339: \label{fig:examples}

1340: \end{figure}

1341: 

1342: \begin{figure}

1343: \center

1344: \includegraphics[width=14cm]{f9.eps}\\

1345: % \includegraphics[width=14cm]{q_elais_subsample.eps}

1346: \caption{Histogram of $q_{24}$ from all ELAIS-S1

1347: data. The solid Gaussian indicates $q_{24}=0.84\pm0.23$ as found by

1348: \cite{Appleton2004}, and the dashed Gaussian $q_{24}=1.39\pm0.02$ as

1349: found by \cite{Boyle2007}. The histogram peak is in broad agreement

1350: with the \cite{Appleton2004} results, and the tail towards low values

1351: of $q_{24}$ is caused by AGN, which are included in our data but were

1352: discarded by \cite{Appleton2004}. Why the \cite{Boyle2007} peak does

1353: not agree with the histogram and the \cite{Appleton2004} distribution

1354: is not understood.}

1355: \label{fig:q24}

1356: \end{figure}

1357: 

1358: \begin{figure}

1359: \includegraphics[width=\linewidth]{f10.eps}

1360: \caption{Source S1189, drawn with contours starting at

1361: $2\,\sigma=70\,\uJy$ and increasing by factors of two. The two jets

1362: are strongly bent backwards, and their far ends almost touch each

1363: other. The morphology suggests that the source is moving through a

1364: relatively dense medium, indicating the presence of a yet unknown

1365: galaxy cluster.}

1366: \label{fig:1189}

1367: \end{figure}

1368: 

1369: \clearpage

1370: 

1371: \begin{table}

1372: \scriptsize

1373: \begin{tabular}{lll}

1374: \hline

1375: \hline

1376: Date    &   Config    & Int. time  \\

1377:         &             & hours      \\

1378: (1)     &  (2)        & (3)        \\

1379: \hline

1380: 09,10,11 Jan 04   			& 6A      & 8.91, 8.77, 6.99  \\ 

1381: 30 Jan 04, 01 Feb 04 			& 6B      & 9.11, 9.47  \\

1382: 19, 27 Dec 04; 01, 02, 03 Jan 05 	& 1.5D    & 3.82, 9.09, 9.89, 8.41, 8.97  \\

1383: 09, 10, 11, 20, 21, 22 Jan 05 		& 750B    & 9.69, 9.51, 10.59, 4.15, 8.74, 7.29  \\

1384: 25 Mar 05; 08, 11 Apr 05		& 6A      & 8.9, 9.23, 9.02	  \\

1385: 24, 26, 30 Apr 05; 01 May 05		& 750A    & 8.16, 8.9, 8.74, 8.53  \\

1386: 08, 09 Jun 05 				& EW367   & 9.28, 9.05  \\

1387: 19, 24 Jun 05 				& 6B      & 9.17, 9.3  \\

1388: \hline

1389: \end{tabular}

1390: \caption{Observing dates, array configurations and net integration

1391: times on ELAIS-S1 pointings.}

1392: \label{tab:obs}

1393: \end{table}

1394: 

1395: \begin{table}

1396: \scriptsize

1397: \begin{tabular}{lll}

1398: \hline

1399: \hline

1400: Source / Pointing    &   RA        & Dec        \\

1401:                      &             &            \\

1402: (1)                  &  (2)        & (3)        \\

1403: \hline

1404: 1934-638             & 19:39:25.02 & -63:42:45.62 \\

1405: 0022-423             & 0:24:42.99  &  -42:02:03.95 \\

1406: 1                    & 0:32:03.55  &  -43:44:51.24 \\

1407: 2                    & 0:31:10.95  &  -43:27:59.64 \\

1408: 3                    & 0:32:05.04  &  -43:11:18.84 \\

1409: 4                    & 0:33:51.29  &  -43:11:24.96 \\

1410: 5                    & 0:32:57.67  &  -43:28:09.00 \\

1411: 6                    & 0:33:50.79  &  -43:44:57.36 \\

1412: 7                    & 0:35:38.02  &  -43:44:57.36 \\

1413: 8                    & 0:34:44.40  &  -43:28:11.88 \\

1414: 9                    & 0:35:37.51  &  -43:11:24.96 \\                        

1415: 10                   & 0:37:23.76  &  -43:11:18.84 \\

1416: 11                   & 0:36:31.13  &  -43:28:09.00 \\

1417: 12                   & 0:37:25.25  &  -43:44:51.24 \\

1418: 13                   & 0:36:31.13  &  -44:01:42.84 \\

1419: 14                   & 0:37:25.25  &  -44:18:34.44 \\

1420: 15                   & 0:35:38.02  &  -44:18:34.44 \\

1421: 16                   & 0:34:44.40  &  -44:01:42.84 \\

1422: 17                   & 0:32:57.67  &  -44:01:42.84 \\

1423: 18                   & 0:33:50.79  &  -44:18:34.44 \\

1424: 19                   & 0:32:03.55  &  -44:18:34.44 \\

1425: 20                   & 0:31:10.95  &  -44:01:42.84 \\

1426: \hline

1427: \end{tabular}

1428: \caption{Coordinates of the calibrators and the pointings depicted in

1429: Figure~\ref{fig:overview}.}

1430: \label{tab:coords}

1431: \end{table}

1432: 

1433: 

1434: \begin{table}

1435: \begin{tabular}{lrrr}

1436: \hline

1437: \hline

1438: Separation & N & m & \%\\

1439: (1)     &  (2)        & (3)        \\

1440: \hline

1441: $<1''   $  & 656   & 16    & 2.4\\

1442: $1''-2''$  & 350   & 20    & 4.2\\

1443: $2''-3''$  & 86    & 9     & 10.5\\

1444: $>3''   $  & 45    & -     & -\\

1445: \hline

1446: \end{tabular}

1447: \caption{Summary of the upper limits on the number of false

1448: cross-identifications. Column 1 gives the separation, column 2 the

1449: number of sources within this range, column 3 the number of radio

1450: sources likely to be wrongly cross-identified with infrared sources,

1451: and column 4 gives this number as a percentage.}

1452: \label{tab:fxids}

1453: \end{table}

1454: 

1455: 

1456: \begin{deluxetable}{llllrrrrrrrrrr}

1457: \tabletypesize{\scriptsize}

1458: \tablewidth{0pt}

1459: \rotate

1460: \tablecaption{Radio component data\label{tab:radio1}}

1461: \tablehead{

1462:                 &\colhead{Name} 	      & \colhead{RA}  	& \colhead{Dec}    & \colhead{RA err} & \colhead{Dec err} & \colhead{Peak} & \colhead{err} & \colhead{Int} & \colhead{err}    & \colhead{rms}     & \colhead{$B_{\rm maj}$} & \colhead{$B_{\rm min}$} & \colhead{PA}\\

1463:                 &               	      &               	&                  & \colhead{arcsec} & \colhead{arcsec}  & \colhead{mJy}  & \colhead{mJy} & \colhead{mJy} & \colhead{mJy}    & \colhead{\uJy\ } & \colhead{arcsec}        & \colhead{arcsec}        & \colhead{$^\circ$}\\

1464: \colhead{(1)}   & \colhead{(2)}               & \colhead{(3)} 	& \colhead{(4)}    & \colhead{(5)}   & \colhead{(6)}      & \colhead{(7)}    & \colhead{(8)}      & \colhead{(9)}    & \colhead{(10)}   &  \colhead{(11)}   & \colhead{(12)}          & \colhead{(13)}          & \colhead{(14)}}

1465: \startdata

1466: C75 		& ATELAIS J003419.30-442647.2 & 00:34:19.308302 & -44:26:47.213520 & 0.33 & 0.23 & 0.25 & 0.03 & 0.25 & 0.01 & 31 & 10.26 & 7.17  & 0   \\

1467: C76 		& ATELAIS J003247.08-442628.8 & 00:32:47.088391 & -44:26:28.830840 & 0.20 & 0.16 & 0.39 & 0.04 & 1.02 & 0.03 & 42 & 16.73 & 11.48 & 151 \\

1468: C77 		& ATELAIS J003138.76-442620.6 & 00:31:38.765112 & -44:26:20.670360 & 0.16 & 0.12 & 0.35 & 0.04 & 0.35 & 0.01 & 39 & 10.26 & 7.17  & 0   \\

1469: C78 		& ATELAIS J003152.54-442620.6 & 00:31:52.548125 & -44:26:20.666040 & 0.14 & 0.10 & 0.40 & 0.04 & 0.40 & 0.01 & 40 & 10.26 & 7.17  & 0   \\

1470: C79 		& ATELAIS J003248.60-442625.7 & 00:32:48.606058 & -44:26:25.750680 & 0.38 & 0.26 & 0.31 & 0.04 & 0.31 & 0.02 & 42 & 10.26 & 7.17  & 0   \\

1471: C80 		& ATELAIS J003659.30-442622.2 & 00:36:59.305858 & -44:26:22.295400 & 0.26 & 0.29 & 0.17 & 0.04 & 0.37 & 0.02 & 40 & 14.09 & 10.97 & 126 \\

1472: C81 		& ATELAIS J003320.05-442617.8 & 00:33:20.053469 & -44:26:17.850480 & 0.09 & 0.07 & 0.42 & 0.04 & 0.47 & 0.01 & 39 & 10.79 & 7.63  & 21  \\

1473: C82 		& ATELAIS J003832.11-442540.6 & 00:38:32.113102 & -44:25:40.639080 & 0.12 & 0.08 & 1.26 & 0.06 & 1.26 & 0.02 & 62 & 10.26 & 7.17  & 0   \\

1474: C83 		& ATELAIS J003052.17-442537.3 & 00:30:52.170276 & -44:25:37.398360 & 0.28 & 0.20 & 0.29 & 0.05 & 0.48 & 0.02 & 55 & 15.14 & 7.89  & 30  \\

1475: C84 		& ATELAIS J003253.48-442543.5 & 00:32:53.487876 & -44:25:43.583880 & 0.09 & 0.06 & 0.39 & 0.03 & 0.56 & 0.01 & 36 & 12.66 & 8.35  & 2   \\

1476: C85 		& ATELAIS J003836.66-442513.5 & 00:38:36.662047 & -44:25:13.595520 & 0.06 & 0.05 & 1.66 & 0.06 & 2.57 & 0.03 & 67 & 12.04 & 9.46  & 165 \\

1477: C86 		& ATELAIS J003602.72-442539.8 & 00:36:02.721341 & -44:25:39.837720 & 0.06 & 0.05 & 1.31 & 0.04 & 1.50 & 0.02 & 40 & 10.56 & 7.99  & 177 \\

1478: C87 		& ATELAIS J003757.04-442516.6 & 00:37:57.045794 & -44:25:16.619160 & 0.41 & 0.29 & 0.28 & 0.04 & 0.28 & 0.02 & 44 & 10.26 & 7.17  & 0   \\

1479: C88 		& ATELAIS J003543.38-442534.9 & 00:35:43.389367 & -44:25:34.921200 & 0.24 & 0.15 & 0.20 & 0.04 & 0.20 & 0.01 & 38 & 10.26 & 7.17  & 0   \\

1480: C89 		& ATELAIS J003215.03-442521.8 & 00:32:15.038647 & -44:25:21.858960 & 0.24 & 0.22 & 0.21 & 0.03 & 0.29 & 0.02 & 37 & 10.97 & 9.24  & 152 \\

1481: \enddata

1482: \tablecomments{A section of the table with component data. Table~\ref{tab:radio1} is

1483: published in its entirety in the electronic edition of the

1484: Astronomical Journal.  A portion is shown here for guidance regarding

1485: its form and content.  {\it Column 1:} a component number we use in

1486: this paper. In some cases, sources were split up into sub-components,

1487: resulting in component numbers such as ``C5'' and ``C5.1''. However,

1488: this is no anticipation of the grouping of components to sources,

1489: which was carried out independently; {\it column 2:} designation for

1490: this component. In the case of single-component sources, this is

1491: identical to the source name used in table 5. This is the formal IAU

1492: designation and should be used in the literature when referring to

1493: this component; {\it columns 3/4:} right ascension and declination

1494: (J2000.0); {\it columns 5/6:} uncertainties in Right Ascension and

1495: Declination. These include the formal uncertainties derived from the

1496: Gaussian fit together with a potential systematic error in the

1497: position of the calibrator source of 0.1 arcsec; {\it columns 7/8:}

1498: peak flux density at 20\,cm (in mJy) of the fitted Gaussian component,

1499: and the associated error as described in the text; {\it columns 9/10:}

1500: integrated flux density at 20\,cm (in mJy) of the fitted Gaussian

1501: component, and the associated error; {\it column 11:} the value (in

1502: \uJy) of the rms map generated by SExtractor at the position of the

1503: component; {\it columns 12/13/14:} the FWHM of the major and minor

1504: axes of the Gaussian in arcsec, and its position angle in degrees.}

1505: \end{deluxetable}

1506: 

1507: \begin{deluxetable}{lrrrrc}

1508: \tabletypesize{\scriptsize}

1509: \tablewidth{0pt}

1510: \tablenum{\ref{tab:radio1}}

1511: \rotate

1512: \tablecaption{Radio component data (continued)}

1513: \tablehead{

1514:                & \colhead{Dec. Peak} & \colhead{Dec. $B_{\rm maj}$} & \colhead{Dec. $B_{\rm min}$} &\colhead{Dec. PA}   & \colhead{sidelobe?}\\

1515:                & \colhead{mJy}       & \colhead{arcsec}             & \colhead{arcsec}             & \colhead{$^\circ$} \\

1516: \colhead{(1)}  & \colhead{(15)}      & \colhead{(16)}               & \colhead{(17)}               & \colhead{(18)}     & \colhead{(19)}}

1517: \startdata

1518: C75 		\\

1519: C76 		& 0.68 & 13.82 & 7.99 & 141 &   \\

1520: C77 		\\

1521: C78 		\\

1522: C79 		\\

1523: C80 		& 0.44 & 11.61 & 5.25 & 112 & * \\

1524: C81 		\\

1525: C82 		\\

1526: C83 		\\

1527: C84 		& 1.3 & 7.43 & 4.27 & 4 &  	  \\ 

1528: C85 		& 5.21 & 7.27 & 4.99 & 129 &    \\

1529: C86 		& 13.12 & 3.64 & 2.30 & 108 &   \\

1530: C87 		\\

1531: C88 		\\

1532: C89 		& 2.15 & 6.85 & 1.45 & 110 &    \\

1533: \enddata

1534: \tablecomments{Radio component data (continued). {\it Column 15:} the

1535: deconvolved peak flux density of the component in mJy. If the

1536: undeconvolved fitted major or minor axis size was within one formal

1537: standard error of the restoring beam size, no value is given; {\it

1538: columns 16/17/18:} the deconvolved FWHM major and minor axes of the

1539: Gaussian in arcsec, and its position angle in degrees. If the

1540: undeconvolved fitted major or minor axis size was within one formal

1541: standard error of the restoring beam size, no value is given; {\it

1542: column 19:} an asterisk in this column indicates that this component

1543: was deemed to be a sidelobe.}

1544: \end{deluxetable}

1545: 

1546: \begin{deluxetable}{llp{7mm}lllrr}

1547: \tabletypesize{\scriptsize}

1548: \tablewidth{0pt}

1549: \rotate

1550: \tablecaption{Radio source data \label{tab:source1}}

1551: \tablehead{

1552:                    & \colhead{Name}  & \colhead{Comp.} & \colhead{RA}  & \colhead{Dec} &\colhead{SWIRE source} & \colhead{$S_{\rm 20cm}$}& \colhead{$\Delta S_{\rm 20cm}$}\\

1553:                    &                 &                 &               &               &                       & \colhead{mJy}           & \colhead{mJy}                  \\

1554:   \colhead{(1)}    & \colhead{(2)}   & \colhead{(3)}   & \colhead{(4)} & \colhead{(5)} & \colhead{(6)}         & \colhead{(7)}           & \colhead{(8)}                  }

1555: \startdata

1556: S693	& ATELAIS J003320.72-434030.1	& C693		& 00:33:20.72	& -43:40:30.11	& SWIRE4\_J003320.74-434030.1 	& 0.38	& 0.05	\\

1557: S694	& ATELAIS J003020.95-433942.8	& C694		& 00:30:20.95	& -43:39:42.89	& SWIRE4\_J003020.97-433942.7 	& 49.58	& 2.48	\\

1558: S695	& ATELAIS J003414.72-434030.7	& C695		& 00:34:14.72	& -43:40:30.74	& 	 			& 0.15	& 0.03	\\

1559: S696	& ATELAIS J003402.27-434008.6	& C696		& 00:34:02.27	& -43:40:08.60	& SWIRE4\_J003402.20-434014.8 	& 0.49	& 0.04	\\

1560: S697	& ATELAIS J003841.55-433925.0	& C697, C697.1	& 00:38:41.55	& -43:39:25.06	& SWIRE4\_J003841.54-433925.0 	& 13.32	& 0.67	\\

1561: S698	& ATELAIS J003412.39-434005.8	& C698		& 00:34:12.39	& -43:40:05.84	& SWIRE4\_J003412.32-434005.2 	& 0.16	& 0.03	\\

1562: S699	& ATELAIS J003513.86-433959.1	& C699		& 00:35:13.86	& -43:39:59.19	& SWIRE4\_J003513.86-433959.0 	& 0.33	& 0.04	\\

1563: S700	& ATELAIS J003703.48-433935.5	& C700		& 00:37:03.48	& -43:39:35.56	& SWIRE4\_J003703.00-433935.3 	& 0.41	& 0.06	\\

1564: S701	& ATELAIS J003141.08-433917.2	& C701		& 00:31:41.08	& -43:39:17.22	& SWIRE4\_J003141.18-433916.8 	& 0.50	& 0.07	\\

1565: S702	& ATELAIS J003038.12-433903.8	& C702, C710	& 00:30:38.12	& -43:39:03.89	& SWIRE4\_J003038.11-433903.8 	& 1.48	& 0.10	\\  	

1566: S703	& ATELAIS J003616.52-433917.5	& C703		& 00:36:16.52	& -43:39:17.55	& SWIRE4\_J003616.54-433918.3 	& 13.83	& 0.69	\\

1567: S704	& ATELAIS J003544.33-433930.2	& C704		& 00:35:44.33	& -43:39:30.25	& SWIRE4\_J003544.38-433930.4 	& 0.22	& 0.04	\\

1568: S705	& ATELAIS J003517.65-433931.9	& C705		& 00:35:17.65	& -43:39:31.97	& SWIRE4\_J003517.66-433931.0 	& 0.18	& 0.03	\\

1569: S706	& ATELAIS J003815.05-433906.5	& C706		& 00:38:15.05	& -43:39:06.53	& SWIRE4\_J003814.95-433907.5 	& 0.34	& 0.07	\\

1570: S707	& ATELAIS J003828.03-433847.2	& C707, C713	& 00:38:28.03	& -43:38:47.26	& SWIRE4\_J003828.02-433847.2 	& 6.00	& 2.44	\\

1571: 

1572: \enddata

1573: \tablecomments{A section of the table with radio source data. Table~\ref{tab:source1} is

1574: published in its entirety in the electronic edition of the

1575: Astronomical Journal.  A portion is shown here for guidance regarding

1576: its form and content.  {\it Column 1:} source number we use in this

1577: paper; {\it column 2:} designation for this source. In the case of

1578: single-component sources, this is identical to the component name used

1579: in Table 3. This is the formal IAU designation and should be used in

1580: the literature when referring to this source; {\it column 3:}

1581: components which are deemed to belong to this source; {\it columns

1582: 4/5:} right ascension and declination (J2000.0). In the case of

1583: single-component sources, this is the radio position of the

1584: component. In the case of multi-component sources with good infrared

1585: identification, the SWIRE position is used. In the case of

1586: multi-component sources without infrared identification, the

1587: coordinates are a flux-weighted mean of the components' coordinates;

1588: {\it column 6:} name of the SWIRE source; {\it columns 7/8:}

1589: integrated radio flux density of the source in mJy and the associated

1590: error. In the case of extended or multiple-component sources, the flux

1591: density has been integrated over the source region, rather than taking

1592: the sum of its constituent components}

1593: \end{deluxetable}

1594: 

1595: \begin{deluxetable}{lrrrrrrrrccrcp{6cm}}

1596: \tabletypesize{\scriptsize}

1597: \tablewidth{0pt}

1598: \tablenum{\ref{tab:source1}}

1599: \rotate

1600: \tablecaption{Radio source data (continued)}

1601: \tablehead{

1602:        	       & \colhead{$S_{\rm 3.6\um}$}  & \colhead{$S_{\rm 4.5\um}$}    & \colhead{$S_{\rm 5.8\um}$}  & \colhead{$S_{\rm 8.0\um}$}  & \colhead{$S_{\rm 24\um}$}& \colhead{B}   & \colhead{V}   & \colhead{R}   & \colhead{AGN} & \colhead{M}    & \colhead{z}    & \colhead{ref}  & \colhead{comment}\\

1603:        	       & \colhead{\uJy}              & \colhead{\uJy}                & \colhead{\uJy}              & \colhead{\uJy}              & \colhead{\uJy}           & \colhead{mag} & \colhead{mag} & \colhead{mag}\\

1604: \colhead{(1)}  & \colhead{(9)}               & \colhead{(10)}                & \colhead{(11)}              & \colhead{(12)}              & \colhead{(13)}           & \colhead{(14)}& \colhead{(15)}& \colhead{(16)}& \colhead{(17)}& \colhead{(18)} & \colhead{(19)} & \colhead{(20)} & \colhead{(21)} }

1605: \startdata

1606: S693    & 253.94        & 275.32        & 281.19        & 561.27        & 2892.44         & 21.16   & 19.95   & 19.08   &        &          &         &         &                                                                                                                       \\

1607: S694    & 435.13        & 312.91        & 158.69        & 86.96         &                 &         &         &         &        &          &         &         &                                                                                                                       \\                                        

1608: S695    &               &               &               &               &                 &         &         &         &        &          &         &         & unclear XID                                                                                                           \\

1609: S696    & 7.22          & 13.64         &               &               &                 &         &         &         &        &          &         &         & IFRS                                                                                                                  \\                                        

1610: S697    & 155.47        & 103.93        & 94.40         & 49.86         & 233.13          & 22.39   & 21.53   & 20.51   & mf     &  x/-     &         &         & clearly a radio double, M-test fails due flux ratio of constituents                                                   \\  

1611: S698    & 20.21         & 27.24         &               & 186.36        &                 & 24.68   & 24.78   & 24.04   &        &          &         &         &                                                                                                                       \\

1612: S699    & 2038.05       & 1476.46       & 844.10        & 639.00        & 156.85          & 18.02   & 16.65   & 15.99   & f      &          & 0.11    & 6dF     &                                                                                                                       \\                                                        

1613: S700    & 39.99         & 26.94         &               & 247.30        &                 & 22.69   & 22.04   & 21.23   & f      &          &         &         &                                                                                                                       \\

1614: S701    & 145.32        & 83.23         & 58.80         &               &                 &         &         &         &        &          &         &         &                                                                                               			\\

1615: S702    & 64.81         & 57.32         &               &               &                 &         &         &         & m      &          &         &         & extended, low surface brightness object, bridge of emission towards C710, which has no XID, hence core-jet morphology \\

1616: S703    & 63.27         & 68.60         & 63.05         &               &                 & 25.14   &         &         &        &          &         &         & confused XID                                                                                                          \\                                

1617: S704    & 14.14         & 15.81         &               &               &                 & 24.22   & 24.11   & 23.73   &        &          &         &         &                                                                                                                       \\

1618: S705    & 45.86         & 48.11         & 54.40         & 655.28        &                 & 24.03   & 23.90   &         &        &          &         &         & confused XID                                                                                                          \\

1619: S706    & 33.36         & 38.41         & 64.01         & 757.86        &                 & 24.85   & 24.39   & 24.07   &        &          &         &         & confused XID                                                                                                          \\

1620: S707    & 6277.45       & 4263.60       & 9520.63       & 40203.53      & 27526.39        & 15.81   & 15.22   & 14.71   &        &          &         &         & C713 probably associated                                                                                              \\  

1621: 

1622: 

1623: \enddata

1624: \tablecomments{Radio source data (continued). {\it Column 9-13:} flux density

1625: of the infrared counterpart in the four IRAC bands at

1626: 3.6\,\um--8.0\,\um\ and in the MIPS band at 24\,\um, in

1627: \uJy. Aperture-corrected flux densities have been used unless the

1628: source was clearly extended, in which case the flux in a Kron aperture

1629: has been used; {\it column 14-16:} optical magnitude of the

1630: counterpart; {\it column 17:} flag indicating whether a source has

1631: been classified as AGN or not, and based on what criteria. An ``f''

1632: indicates AGN classification based on the far-infrared-radio relation,

1633: an ``m'' based on morphology, and an ``l'' based on classification

1634: taken from the literature; {\it column 18:} result of the test

1635: developed by

1636: \cite{Magliocchetti1998} as described in the text, performed for

1637: double radio sources. A ``-'' indicates failure, a ``x'' success of

1638: the two parts of the test (separation and flux density ratio of the

1639: constituents); {\it column 19:} source redshift; {\it column 20:}

1640: reference for the redshift. The codes indicate the following

1641: publications: 2df - \cite{Colless2001}, 6dF -

1642: \cite{Jones2004}, A01 - \cite{Alexander2001}, L04 - \cite{LaFranca2004}, 

1643: P06 - \cite{Puccetti2006}, S01 - \cite{Serjeant2001}, S96 - \cite{Shectman1996},

1644: W03 - \cite{Wegner2003}; {\it column 21:} comment}

1645: \end{deluxetable}

1646: 

1647: \begin{deluxetable}{lll}

1648: \tabletypesize{\scriptsize}

1649: \tablewidth{0pt}

1650: \tablenum{\ref{tab:source1}}

1651: \rotate

1652: \tablecaption{Radio source data (continued)}

1653: \tablehead{

1654:        	       & \colhead{G99 name} & \colhead{other name}\\

1655: \colhead{(1)}  & \colhead{(22)}     & \colhead{(23)} }

1656: \startdata

1657: S693    & ELAIS20R\_J003021-433943 &                          \\                     

1658: S694    &                          &                          \\                     

1659: S695    & ELAIS20R\_J003402-434011 &                          \\                     

1660: S696    & ELAIS20R\_J003842-433924 &                          \\                     

1661: S697    &                          &                          \\                     

1662: S698    &                          & 2MASX J00351384-4339588  \\                     

1663: S699    &                          &                          \\                     

1664: S700    &                          &                          \\                     

1665: S701    &                          &                          \\                     

1666: S702    & ELAIS20R\_J003617-433918 &                          \\                     

1667: S703    &                          &                          \\                     

1668: S704    &                          &                          \\                     

1669: S705    &                          &                          \\                     

1670: S706    & ELAIS20R\_J003828-433849 & ESO 242- G 021           \\                     

1671: S707    & 			   &                          \\ 

1672: \enddata

1673: \tablecomments{Radio source data (continued). {\it Column 22:} The designation by \cite{Gruppioni1999}; {\it column 23:} Other names obtained from NED}

1674: \end{deluxetable}

1675: 

1676: \begin{table}[htpb]

1677: \scriptsize

1678: \begin{tabular}{lllrrrp{5cm}}

1679: \hline

1680: \hline

1681: Source 	& SUMSS RA        & SUMMS Dec       & $S$     &  Sep.  	& $\alpha$ & Comment \\

1682:        	&                 &                 & mJy     & arcsec 	&          & \\

1683: (1)    	& (2)             & (3)             & (4)     & (5)    	& (6)      & (7)\\

1684: \hline

1685: S207    & 00:30:48.60     & -44:14:33.10    & 34.6    & 14.40   & -1.06   & S207, S207.2 and S212 all blend together in this source \\

1686: S220    & 00:37:09.57     & -44:14:08.40    & 11.6    & 2.18    & -1.48   &							    \\

1687: S258    & 00:32:04.54     & -44:11:32.70    & 59.0    & 2.36    & -1.00   &							    \\

1688: S272    & 00:36:44.04     & -44:10:54.90    & 13.3    & 2.53    & -0.96   & blends with S278					    \\

1689: S293    & 00:29:25.72     & -44:08:24.80    & 12.4    & 2.08    & -1.11   & blends with S304					    \\

1690: S296    & 00:36:50.07     & -44:08:59.70    & 14.7    & 3.00    & -0.80   &							    \\

1691: S311    & 00:37:20.40     & -44:07:31.20    & 74.3    & 4.34    & -0.78   &							    \\

1692: S313    & 00:31:10.76     & -44:07:41.70    & 14.4    & 1.13    & -1.01   &							    \\

1693: S325    & 00:34:52.73     & -44:07:26.30    & 10.8    & 1.31    & -1.13   &							    \\

1694: S347    & 00:35:38.62     & -44:06:03.60    & 15.9    & 15.98   & -0.20   &							    \\

1695: S355    & 00:30:19.00     & -44:04:33.40    & 14.9    & 5.67    & -0.70   &							    \\

1696: S360    & 00:34:58.74     & -44:04:59.30    & 26.2    & 1.95    & -0.53   &							    \\

1697: S371    & 00:38:54.63     & -44:03:29.20    & 12.1    & 0.99    & -0.37   &							    \\

1698: S381    & 00:39:40.19     & -44:02:10.20    & 23.1    & 2.20    &  0.19   & 							    \\

1699: S390.1  & 00:37:19.70     & -44:01:49.80    & 11.0    & 3.64    & -3.13   & blends with 390                                         \\

1700: \hline

1701: \end{tabular}

1702: \caption{A section of the table with SUMSS counterparts to 1.4\,GHz

1703: radio sources. Table~\ref{tab:sumss} is published in its entirety in

1704: the electronic edition of the Astronomical Journal.  A portion is

1705: shown here for guidance regarding its form and content. {\it Column

1706: 1:} The source names we use in this paper; {\it Columns 2/3:} SUMMS

1707: right ascension and declination; {\it Column 4:} SUMMS flux density in

1708: mJy; {\it Column 5:} separation of the SUMMS source to the source

1709: coordinates in Table~\ref{tab:source1}; {\it Column 6:} the spectral

1710: index; {\it Column 7:}  comment}

1711: \label{tab:sumss}

1712: \end{table}

1713: 

1714: 

1715: 

1716: \end{document}

1717: 

1718: