1: \documentclass[12pt,preprint]{aastex}
2:
3: \usepackage{graphicx}
4: %\usepackage{epsfig}
5:
6: \newcommand{\HI}{{\sc H\,i}}
7: \newcommand{\HII}{{\sc H\,ii}}
8: \newcommand{\kms}{\rm km\ s^{-1}}
9:
10: \shorttitle{Radio Polarimetry of ELAIS N1}
11: \shortauthors{Taylor et.\ al.}
12:
13: \begin{document}
14:
15: \title{Radio Polarimetry of the ELAIS N1 Field: \\
16: Polarized Compact Sources}
17:
18: \author{A.\ R.\ Taylor, J.\ M.\ Stil, and J.\ K.\ Grant}
19: \affil{Department of Physics and Astronomy, University of Calgary}
20:
21: \author{T.\ L.\ Landecker, R.\ Kothes, R.\ I.\ Reid, A.\ D.\ Gray }
22: \affil{Dominion Radio Astrophysical Observatory, Herzberg Institute of Astrophysics \\
23: National Research Council of Canada}
24:
25: \author{Douglas Scott}
26: \affil{Department of Physics and Astronomy, University of British Columbia}
27:
28: \author{P.\ G.\ Martin, A.\ I.\ Boothroyd}
29: \affil{Department of Astronomy, University of Toronto}
30:
31: \author{G.\ Joncas}
32: \affil{D\'epartement de physique, g\'enie physique et d'optique, Universit\'e Laval}
33:
34: \author{Felix\ J.\ Lockman}
35: \affil{National Radio Astronomy Observatory}
36:
37: \author{J.\ English}
38: \affil{Department of Physics, University of Manitoba}
39:
40: \author{A.\ Sajina}
41: \affil{Spitzer Science Center, California Institute of Technology}
42:
43: \and
44:
45: \author{J.\ R.\ Bond}
46: \affil{Canadian Institute for Theoretical Astrophysics, University of Toronto}
47:
48: \begin{abstract}
49: We present deep polarimetric observations at 1420 MHz of the European
50: Large Area {\sl ISO} Survey North 1 region (ELAIS N1) as part of the
51: Dominion Radio Astrophysical Observatory {\sl Planck} Deep Fields
52: project. By combining closely spaced apertures synthesis fields, we
53: image a region of 7.43 square degrees to a maximum sensitivity in
54: Stokes $Q$ and $U$ of $78\,\mu$Jy beam$^{-1}$, and detect 786 compact
55: sources in Stokes $I$. Of these, 83 exhibit polarized emission. We
56: find that the differential source counts ($\log N - \log p$) for
57: polarized sources are nearly constant down to $p > 500\,\mu$Jy, and
58: that these faint polarized radio sources are more highly polarized
59: than the strong source population. The median fractional polarization
60: is $4.8 \pm 0.7$\% for polarized sources with Stokes $I$ flux density
61: between 1 and 30 mJy; approximately three times larger than sources
62: with $I > 100$ mJy. The majority of the polarized sources have been
63: identified with galaxies in the {\sl Spitzer} Wide Area Infrared
64: Extragalactic Survey (SWIRE) image of ELAIS N1. Most of the galaxies
65: occupy regions in the IRAC 5.8/3.6 $\mu$m vs.\ $8.0/4.5\,\mu$m
66: color-color diagram associated with dusty AGNs, or with ellipticals
67: with an aging stellar population. A few host galaxies have colors
68: that suggests significant PAH emission in the near-infrared. A small
69: fraction, 12\%, of the polarized sources are not detected in the SWIRE
70: data. None of the polarized sources in our sample appears to be
71: associated with an actively star-forming galaxy.
72: \end{abstract}
73:
74: \keywords{polarization --- techniques: polarimetric ---
75: radio continuum : galaxies --- galaxies : evolution --- individual (ELAIS N1)}
76:
77: \section{Introduction}
78:
79: Observation of polarized radiation at radio wavelengths is one of the
80: prime means to study the roles of magnetic fields in astrophysics,
81: through synchrotron emissivity which samples magnetic fields in
82: relativistic plasmas, and through Faraday Rotation, created by
83: radiation propagation through magnetized thermal plasmas.
84: Understanding the origin and evolution of magnetic fields is a key
85: science goal of the The Square Kilometre Array, a next-generation
86: radio telescope under development by the international community
87: \citep{schilizzi2004}. The primary observational tool to study the
88: magnetic universe will be an SKA all-sky Rotation Measure survey of
89: background radio sources and diffuse Galactic emission down to
90: polarized flux density levels of $\sim0.1\,\mu$Jy \citep{bg04}.
91: However, while modern source counts approach flux density
92: sensitivities of $\sim10\,\mu$Jy in total intensity
93: \citep{windhorst2003,hopkins2003}, very little is known about the
94: polarization properties of the faint radio source population. The
95: most extensive analyses of polarization of compact extragalactic
96: sources were carried out by \citet{mesaetal2002} and
97: \citet{tucci2004}, who used the NVSS data \citep{condon1998} to derive
98: statistical polarization properties for $\sim$30,000 sources with
99: $S_{\rm 1.4GHz} > 100$ mJy. Similarly, \citet{bg04} used NVSS sources
100: with total flux density greater than 80 mJy to extrapolate polarized
101: source counts to $\mu$Jy levels.
102:
103: \citet{mesaetal2002} found that the mean fractional polarization of
104: radio sources in the NVSS brighter than 80 mJy was anti-correlated
105: with flux density, especially for steep-spectrum radio sources
106: ($\alpha < -0.5$, for $S_\nu \sim \nu^\alpha$). \cite{tucci2004}
107: confirmed this result for the median of the fractional polarization
108: for steep-spectrum sources only (87\% of their sample), but found no
109: significant trend for flat spectrum sources (13\% of their sample).
110: \cite{tucci2004} also noted that the flat shape of the polarized
111: source counts indicates a dependence of the fractional polarization on
112: flux density.
113:
114: Radio sources with $S_{\rm 1.4GHz} > 100$ mJy are predominantly
115: associated with flat or steep spectrum Active Galactic Nuclei. Star
116: forming galaxies begin to be a significant fraction of the population
117: at flux densities less than a few mJy (see e.g. Hopkins 2000,
118: Windhorst 2003). However, there is still ongoing debate about the
119: fraction of the sub-mJy radio sources which is radio-quiet active
120: galactic nuclei (AGN) \citep{gruppioni1999,simpson2006}. It is thus
121: uncertain that polarization properties derived from the strong radio
122: source population are applicable to the sub-mJy radio sources.
123:
124:
125: We have begun a project called the DRAO Planck Deep Fields to explore
126: the high latitude sky at high sensitivity in polarized radio continuum
127: and in atomic hydrogen emission. The project uses the DRAO Synthesis
128: Telescope at 1.4 GHz to create deep images of two fields, one with a
129: very low column of foreground material, the ELAIS N1 region $(l,b) =
130: (84\degr, +45\degr)$, and a larger region of highly structured
131: infrared cirrus emission at $(l,b) = (135\degr, +40\degr)$. This paper
132: reports initial results from the first 30\% of observations of ELAIS
133: N1. The ELAIS N1 (European Large Area {\sl ISO} Survey North 1) field
134: \citep{oliver2000}, is an area of approximately 2 square degree chosen
135: for a mid-infrared survey of distant galaxies with {\sl ISO}. The
136: field was selected to minimize confusion with Galactic cirrus and the
137: zodiacal background, being one of the lines of sight to the
138: extragalactic sky with minimum {\sl IRAS} 100 $\mu$m emission. A
139: larger area that includes ELAIS N1 was later observed by the {\sl
140: Spitzer} Wide Area Infrared Extragalactic Survey (SWIRE)
141: \citep{lonsdale03}. The large amount of archival data from these
142: extragalactic surveys make this region ideal for studies of the faint
143: polarized radio source population.
144:
145:
146: \section{Observations and Data Processing}
147: \subsection{Synthesis Observations}
148:
149: The DRAO Synthesis Telescope (DRAO ST) is described in detail in
150: \citet{landecker2000}. The telescope is a seven-element east-west
151: array of 9-m diameter antennas. Three antennas are moved to
152: provide complete sampling of the UV plane from the shortest baseline
153: (12.86 m) to the longest baseline (617.18 m) after a full synthesis of
154: 12 times 12 hours. The array has a primary beam
155: size 107\farcm2 (FWHM) at $1420\,\rm MHz$, which makes it an effective
156: instrument for wide-field surveys. The first sidelobe of the
157: synthesized beam is at the 3\% level, and side-lobes farther from the main
158: lobe of the beam are less than 0.5\%. The first grating ring of the
159: synthesized beam appears at $2\fdg8$ from the main lobe at 1420 MHz,
160: which is outside the field of view defined by the 10\% sensitivity
161: level of the primary beam.
162:
163: The telescope observes simultaneously the \HI\ 21 cm line and
164: continuum at 408 MHz and full polarimetry in four 7.5 MHz wide
165: frequency bands centered around $1420\,\rm MHz$. The antennas have
166: prime focus feeds and at 1420 MHz receive both right-hand (R) and
167: left-hand (L) circularly polarized radiation. The observations and
168: data processing techniques used in this paper follow those that have
169: been used to obtain high-fidelity wide-field polarimetric images of
170: the Galactic plane with the DRAO ST \citep{taylor2003,landecker2007}.
171:
172: The system temperature of the telescope as described by
173: \citet{landecker2000} was $60\,\rm K$, leading to an rms noise in a
174: synthesized image of a single field of $280\,\mu$Jy~beam$^{-1}$ at
175: field center. However, starting in 2003, the sensitivity was enhanced
176: by a series of improvements, completed by the time the current
177: observations began in 2004. These improvements comprised installation
178: of new low-noise amplifiers, modifications to telescope structures to
179: reduce ground noise, and installation of shielding fences to further
180: block ground radiation from entering the aperture. The system
181: temperature of the seven individual antennas now spans the range
182: $35\,\rm K$ to $60\,\rm K$, and the overall system temperature is
183: $\sim45\,\rm K$, leading to a field-center rms noise of
184: $210\,\mu$Jy~beam$^{-1}$ ($53 \sin \delta\ \rm mK$) at 1420 MHz. A
185: higher sensitivity and larger field of view are obtained by creating a
186: mosaic of overlapping fields. The most uniform sensitivity across a
187: mosaic is obtained if the fields are centered on a hexagonal grid. The
188: field centre separations for the survey presented here is $22\arcmin$,
189: which is $20.5 \%$ of the FWHM diameter of the primary beam.
190:
191: Observations
192: for the DRAO ELAIS N1 survey began in August 2004 and will continue to
193: create a final mosaic of 30 fields. This paper presents 21-cm
194: continuum polarimetry of the first 10 fields.
195: Figure~\ref{surveyarea-fig} shows the location of the 10-field mosaic
196: in relation to the {\sl ISO} ELAIS N1 field, and the SWIRE survey of ELAIS~N1.
197: The theoretical maximum 1-$\sigma$ sensitivity in our mosaic of 10 fields is $\sim80\,\mu$Jy.
198:
199:
200: \subsection{Polarization Data Processing and Calibration}
201:
202: Complex gains for the center of each field were determined by
203: observing the unresolved and unpolarized sources
204: 3C\,147 and 3C\,295 between 12-hour observing runs. The absolute
205: polarization angle was calibrated by observing the highly
206: linearly polarized source 3C\,286 once every 4 days.
207:
208: The polarization images do not contain much flux, making
209: self-calibration ineffective. Therefore R and L gain solutions derived from
210: self-calibration of the Stokes $I$ images were applied to the polarization
211: data as well. The resulting visibility data sets for each field were
212: then corrected for the effects of instrumental polarization across the
213: field of view (which leads to leakage of Stokes $I$ power into $Q$ and
214: $U$). Instrumental polarization is a complex quantity, with phase and
215: amplitude terms. These were measured for the seven individual antennas
216: with a holographic technique, using the unpolarized source 3C~295, on a
217: $15'$ grid across the primary beam. After interpolation, these
218: measurements were used to predict conversion of $I$ into $Q$ and $U$ at
219: any point in the beam based on CLEAN component models from processing
220: the $I$ image for a field. Residual errors in instrumental polarization
221: for an individual field are estimated at 0.25\% for the field center,
222: growing to 1\% at a distance of $75'$. With a field separation in the
223: mosaic image of $22'$ the instrumental polarization in the central 3.6
224: square degrees of the mosaic remains less than 0.5\%. Towards the edge
225: of the mosaic instrumental terms may be as large as 1\%.
226:
227: After these initial procedures there are usually still confusing
228: arc-like structures left in the images, centered on bright sources
229: inside and outside the primary beam. These are the result of residual
230: complex gain errors at large distance from the field center. Effects
231: from these are removed from the visibilities using a procedure called
232: modcal, which is in principle a direction-dependent self-calibration
233: \citep{willis99}.
234:
235: The antenna sidelobes are highly polarized, and sources outside the
236: primary beam can produce strong spurious polarized signals. The Sun is
237: seen in the sidelobes whenever it is above the horizon, but its effects
238: are usually confined to the shorter baselines because of its large
239: extent. The effects of the Sun were removed by making images centered on
240: the Sun's position and removing the response from the visibilities.
241: Terrestrial interference, which is always polarized, is another source
242: of spurious polarization. Interference effects in images appear as a
243: spurious source concentrated around the North Celestial Pole, and can be
244: largely eliminated by making an image at the Pole and correcting the
245: visibilities. Radiation from the ground appears to be polarized, and
246: can appear as a correlated signal in visibilities corresponding to short
247: baselines; this effect is more difficult to remove. In some particularly
248: bad cases the data for affected interferometer spacings were simply
249: flagged and removed.
250:
251: \subsection{The Images}
252:
253: Figure~\ref{EN21-fig} shows the deep 21 cm continuum images of the
254: ELAIS N1 area in Stokes $I$, $Q$, and $U$. The images are centered on
255: $\alpha_{2000}$ = $16^{\rm h} 11^{\rm m} $, $\delta_{2000}$ =
256: $+55\degr$ and cover an area of 7.43 square degrees (the area within
257: the thick gray line in Figure \ref{surveyarea-fig}). The angular
258: resolution varies with the declination over the images and is given by
259: $b_\alpha \times b_\delta = 49'' \times 49'' {\rm cosec } \delta $.
260: At the mosaic center the resolution is $49'' \times 59''$. The noise
261: near the center of the mosaic (white square in Figure~\ref{EN21-fig})
262: is measured to be $78\,\mu$Jy in $Q$, and $U$ (see
263: \S~\ref{detect-sec}). The Stokes $I$ image is not limited by confusion,
264: but the noise is slightly higher at $85\,\mu$Jy near the center of the
265: mosaic, probably because of a contribution from faint sources.
266:
267: The final images are virtually free of artifacts, so the sensitivity is
268: limited by the noise. The rms brightness sensitivity is $17.4\,\rm mK$.
269: The dynamic range near the center of the mosaic is more than
270: 3000:1 in Stokes $I$. With very few exceptions, sources in
271: Figure~\ref{EN21-fig} appear as compact (nearly) unresolved sources,
272: that can be characterized by their peak intensity and a single
273: polarization angle. This is consistent with expectations from the
274: angular size -- flux density relation for extragalactic radio sources
275: \citep{windhorst2003}; the median angular size of a $1\,\rm Jy$ source is
276: $\sim 10''$.
277:
278: \section{Compact Polarized Sources}
279:
280: \subsection{Source Detection}
281: \label{detect-sec}
282:
283: Flux densities and positions of all sources in the pilot deep field
284: images were measured with a source extraction algorithm that fits a
285: two-dimensional Gaussian and a background level to each source. The
286: mosaic images were multiplied by the primary beam response function of
287: the mosaic to obtain an image with a uniform noise level, equal to the
288: noise level at the center of the image. This operation retains the
289: correct signal-to-noise ratio for each source, but the resulting
290: uniform noise level greatly facilitates automated source extraction.
291: The inverse primary beam correction is applied to the measured flux
292: densities of sources from the uniform noise map to transform back to
293: true flux density. The rms noise level in the uniform noise images
294: was measured by fitting a Gaussian function to the distribution of
295: amplitudes in the image. The result is shown in
296: Figure~\ref{fig:Q_Unoise}. The noise distribution in both the $Q$ and
297: $U$ images is well fitted by a Gaussian with dispersion $\sigma =
298: 78\,\mu$Jy.
299:
300: The polarized flux density image ($p = \sqrt {Q^2 + U^2}$) was searched
301: for polarized sources. For Gaussian noise in the $Q$ and $U$ images with
302: rms amplitude $\sigma$, the statistical distribution of the noise in
303: a measurement of $p$ for a source with an intrinsic polarized flux
304: density of $p_o$ is a Rice distribution
305: \citep{rice1945,vinokur1965,simmons1985}
306: \begin{equation}
307: f(p|p_{\rm o}) = \frac{p}{\sigma} e^{-\frac{(p^2 + p_{\rm o}^2)}{2 \sigma^2}}
308: I_{\rm o}(\frac{p p_{\rm o}}{\sigma^2}) \ ,
309: \end{equation}
310: Here $I_{\rm o}$ is
311: the modified Bessel function of the first kind. For $p_{\rm o} = 0$, the
312: distribution reduces to a Rayleigh distribution,
313: \begin{equation}
314: f(p|0) = \frac{p}{\sigma} e^{-\frac{p^2 }{2 \sigma^2}} \ ,
315: \end{equation}
316: which gives the probability distribution of pixel amplitudes in the
317: $p$ image in the absence of polarized emission. The noise in the $p$
318: image has a non-zero mean and has higher probability of positive peaks
319: above a given detection threshold than Gaussian noise. We searched
320: the $p$ map down to a level of $4.55\sigma$, which has an equivalent
321: probability for false positive signals to the $4\sigma$ level for a
322: Gaussian distribution. The measured polarized flux density $p$ was
323: corrected for noise bias to obtain an estimate of $p_{\rm o}$ through
324: the relation $p_{\rm o}^2 = p^2 - \sigma^2$, which is a good
325: approximation if the signal to noise ratio is larger than 4
326: \citep{simmons1985}.
327:
328: The 83 sources detected are listed in Table~\ref{table_polsources},
329: which gives the position of each source, the integrated total flux
330: density, noise bias-corrected peak polarized intensity, polarization
331: position angle, fractional polarization defined as the ratio of the
332: bias-corrected peak polarized intensity to the peak total intensity,
333: and the spectral index of the total flux density between $325\,\rm
334: MHz$ and $1420\,\rm MHz$ if the source appears in the WENSS catalogue
335: \citep{rengelink1997}. To ensure possible spurious polarized sources
336: due to instrumental polarization are not included, we conservatively
337: remove sources with observed fractional polarization, $\Pi$, less than
338: 1\%. Only two sources were removed from the sample for this reason.
339:
340: \subsection{Distribution of Fractional Polarization}
341: \label{maxlike-sec}
342:
343: The intrinsic fractional polarization $\Pi_{\rm o}$ of radio sources
344: provides astrophysical information about the nature of the polarized
345: sources. However, the observed fractional polarization $\Pi$ is
346: sensitive to the noise in $p$ and in $I$. In addition to $p$ being a
347: biased estimator of $p_{\rm o}$, the error distribution of the ratio
348: $p/I$ has strong non-Gaussian wings, so $\Pi$ is not a very accurate
349: estimate of $\Pi_{\rm o}$ even for relatively high signal-to-noise
350: sources. This is illustrated in Figure~\ref{polstat_demo}, which
351: shows the relation between $\Pi$ and Stokes $I$ flux density for an
352: artificial sample of sources, all with $\Pi_{\rm o} = 5$\% and
353: Gaussian noise added with equal amplitudes in Stokes $I$, $Q$, and
354: $U$. Only those sources with a $p$ flux density more than $5\sigma$
355: are shown. Error bars represent standard error propagation in $\Pi$
356: from the errors in $p$ and $I$. The high values of $\Pi$ at
357: signal-to-noise ratio less than $\sim$100 is a result of the detection
358: threshold in $p$ and the non-Gaussian error statistics of $\Pi$. The
359: effect is much larger than the polarization noise bias alone.
360:
361: These problems highlight the need for careful analysis of the effects
362: of noise and polarization detection thresholds in studies of the
363: fractional polarization of faint sources. Previous studies have
364: focused on the fractional polarization of bright radio sources in the
365: NVSS. \citet{mesaetal2002}, \citet{tucci2004}, and \citet{bg04}
366: considered polarized sources in the NVSS with Stokes $I$ flux density
367: stronger than 80 mJy, 100 mJy, and 80 mJy respectively. The high flux
368: density thresholds in these studies were adopted to achieve a good
369: level of completeness in $\Pi$ down to the limit set by residual
370: instrumental polarization ($\Pi \approx 1\%$). These studies found a
371: $\Pi$ distribution that decreases monotonically with increasing $\Pi$,
372: and with a median $\Pi \approx 1.8\%$. For these bright sources,
373: noise effects are small, and the $\Pi$ distribution should be close to
374: the intrinsic $\Pi_{\rm o}$ distribution down to limits set by
375: residual instrumental polarization.
376:
377: In this paper we investigate the shape of the $\Pi_{\rm o}$
378: distribution for much fainter Stokes $I$ flux densities than those
379: considered in previous work. The present data have angular resolution
380: similar to the NVSS, so differences in the measured degree of
381: polarization because of a difference in resolution are not
382: expected. Our results can be compared directly with results based on
383: the NVSS.
384:
385: Noise effects such as those illustrated in Figure~\ref{polstat_demo}
386: were taken into account by a Monte Carlo analysis. A set of simulated
387: catalogs was generated to accurately represent the effects of noise
388: and the polarized flux density detection threshold in the data.
389: Stokes $I$ flux densities of simulated sources were drawn from the fit
390: to observed source counts by \citet{windhorst1990}. We assume in
391: these simulations that the Stokes $I$ source counts of polarized
392: sources have the same shape as those for all radio sources. This is a
393: reasonable assumption because $\sim 80\%$ of radio sources in the NVSS
394: display significant polarization \citep{mesaetal2002,tucci2004,bg04}.
395: The intrinsic Stokes I flux density is multiplied by the degree of
396: polarization, $\Pi_{\rm o}$, drawn from an assumed $\Pi_{\rm o}$
397: distribution to obtain the intrinsic polarized intensity $p_{\rm o}$.
398:
399: The intrinsic Stokes $I_{\rm o}$ and $p_{\rm o}$ of a simulated source
400: are transformed to observed flux densities $I$ and $p$ by adding noise
401: with statistical properties identical to the properties of the noise
402: in the data. First, $p_{\rm o}$ is converted into intrinsic Stokes
403: $Q_{\rm o}$ and $U_{\rm o}$, assuming a random polarization angle.
404: The error in the flux density is assumed to consist of a part that is
405: proportional to the noise in the image at the location of the source,
406: and a part that is proportional to the flux density of the source,
407: added in quadrature. The error in the flux density $S$ is evaluated as
408: \begin{equation}
409: \sigma_S = S \sqrt { C_1^2 +C_2^2 \Bigl( {\sigma^2 \over S^2} \Bigr) } \ \ ,
410: \end{equation}
411: where $S$ represents the intrinsic flux density $I_{\rm o}$, $Q_{\rm
412: o}$, or $U_{\rm o}$, $\sigma$ is the rms noise in the image and $C_1$
413: and $C_2$ are constants. The value of $C_1$ was determined from the
414: rms variation of the flux density of bright sources in the ten fields
415: after field registration \citep{taylor2003}. Flux densities of
416: sources brighter than $100\,\rm mJy$ varied by $2.5\%$ (rms) over the
417: ten fields. From this we adopt $C_1 = 0.025$. The value of the
418: constant $C_2 = 1.3$ was taken to be that found by
419: \citet{rengelink1997} from Monte Carlo simulations for flux density
420: errors in the Westerbork Northern Sky Survey (WENSS). The value of
421: $\sigma$ is different for each source, to represent variation of the
422: noise with location in the mosaic. The distribution of polarized
423: intensities of the resulting simulated sources, have the same
424: statistical effects as the observed polarized intensities, including
425: the effects of noise bias, the detection threshold and the variation
426: of the noise with position in the mosaic.
427:
428: To derive the $\Pi_{\rm o}$ distribution directly from the data, we
429: compare the distribution of the data in a $\log (I) - \log (p)$
430: diagram with the probability distribution for simulated source
431: catalogues. Figure~\ref{beck-fig} shows the observed data points, and
432: the model probability distribution assuming the $\Pi_{\rm o}$
433: distribution function of \citet{bg04}. As expected, sources brighter
434: than $\sim 80\,\rm mJy$ are represented well by this model. However,
435: fainter sources in our sample are more highly polarized than predicted
436: by this distribution. This is clearly visible in
437: Figure~\ref{beck-fig} for sources with $10 < I < 30\,\rm mJy$, where
438: an offset exists between the distribution of observed sources and the
439: predicted ridge of maximum source density. A 2-dimensional
440: Kolmogorov-Smirnov test \citep{peacock1983} rejected the hypothesis
441: that the data were drawn from the simulated distribution at the 99.9\%
442: confidence level.
443:
444: The best fit $\Pi_{\rm o}$ distribution was derived by fitting source
445: probability density distributions to the data in the $\log(I) -
446: \log(p)$ plane for trial $\Pi_{\rm o}$ distributions. The trial
447: $\Pi_{\rm o}$ distributions were represented by a low-order
448: Gauss-Hermite series, also called a Gram-Charlier series, following
449: the description of \citet{vdmarel1993},
450: \begin{equation}
451: f(\Pi_{\rm o})=\exp \Bigl({-{\Pi_{\rm o}^2 \over 2\sigma_{\Pi_{\rm o}}^2}}\Bigr) \Bigl[1 + \sum_{i=3}^{N} h_i H_i(\Pi_{\rm o}/\sigma_{\Pi_{\rm o}}) \Bigr].
452: \label{GHseries-eq}
453: \end{equation}
454: We consider only modest deviations from a Gaussian, since previous
455: work on bright NVSS sources suggests that the shape of the
456: distribution is nearly Gaussian. Assuming that the $\Pi_{\rm o}$
457: distribution peaks at zero and declines monotonically with increasing
458: $\Pi_{\rm o}$, we use only the fourth term ($i = 4$) in
459: Equation~\ref{GHseries-eq}, which results in symmetric deviations. A
460: coefficient $h_4> 0$ means that the wings of the distribution are
461: stronger than the wings of a Gaussian distribution, as shown
462: graphically by \citet{vdmarel1993}. Higher order terms were not
463: considered because of the limited size of our data sample at this
464: time. The $\Pi_{\rm o}$ distribution defined by
465: Equation~\ref{GHseries-eq} thus has two free parameters, the Gaussian
466: standard deviation, $\sigma_{\Pi_{\rm o}}$, and the amplitude of the lowest
467: order symmetric deviation from a Gaussian, $h_4$. The fits maximize
468: the likelihood $L$ of the data as a function of these parameters,
469: \begin{equation}
470: L = \prod_{i=1}^{N_{\rm data}} P(I_i,p_i | \sigma_{\Pi_{\rm o}}, h_4),
471: \end{equation}
472: with the probability of an observed ($I_i, p_i$) for a given
473: $\sigma_{\Pi_{\rm o}}$ and $h_4$
474: \begin{equation}
475: P(I_i,p_i | \sigma_{\Pi_{\rm o}}, h_4) = {1 \over M_{\rm model}} \sum_{j=1}^{M_{\rm model}} e^{ {(I_i-I_j)^2 \over 2 \sigma_{I,i}^
476: 2 } + {(p_i-p_j)^2 \over 2 \sigma_{p,i}^2 } }.
477: \end{equation}
478: The product over $i$ is over all sources in the data, $N_{\rm
479: data}$. Whereas the sum over $j$ is over all simulated sources in the
480: catalog, $M_{\rm model}$, for a particular $\sigma_{\Pi_{\rm o}}$ and
481: $h_4$. Typically $M_{\rm model} \sim 10^5$. The values $\sigma_{I,i}$
482: and $\sigma_{p,i}$ are the observed errors in Stokes $I$ and polarized
483: flux density for the $i^{\rm th}$ source.
484:
485:
486: The maximum likelihood $\Pi_{\rm o}$ distribution was found through a
487: grid search over the parameter space. The best fitting model has
488: parameters $\sigma_{\Pi_{\rm o}} = 7\%$, and $h_4 = 0.05$. The
489: probability density function of this model is shown along with the
490: data in Figure~\ref{bestfit_Pi}. The uncertainty in the best-fitting
491: parameters was evaluated empirically. Three hundred randomly selected
492: samples, each containing on average the same number of sources as the
493: observed sample, were drawn from the best fitting simulated
494: catalog. Each of these samples was fitted with the maximum likelihood
495: fit to evaluate the spread of the best-fitting parameters. Two thirds
496: of these fits yielded a $\sigma_{\Pi_{\rm o}}$ within 1\% of the
497: maximum likelihood value $7\%$. We conclude that $\sigma_{\Pi_{\rm o}}
498: = (7.0 \pm 1.0) \%$. The fitted value of $h_4$ is not independent of
499: $\sigma_{\Pi_{\rm o}}$, because a larger $h_4$ can partially
500: compensate for a smaller $\sigma_{\Pi_{\rm o}}$. From the same 300
501: experimental fits, two thirds yielded a value $h_4 < 0.1$. Although
502: the data are consistent with $h_4 = 0$, those fits with $h_4$
503: contrained to be zero yield an average $\sigma_{\Pi_{\rm o}} = (8 \pm
504: 1)\%$. The data therefore suggest that the $\Pi_{\rm o}$ distribution
505: may be somewhat broader than a simple Gaussian with $\sigma_{\Pi_{\rm
506: o}} = 7.0\%$.
507:
508: The best-fitting model was also subjected to a 2-dimensional
509: Kolmogorov-Smirnov test. The hypothesis that the data were drawn from
510: the best-fit distribution in Figure~\ref{bestfit_Pi} was rejected at
511: the 98\% confidence level. This is much better than the result for the
512: \citet{bg04} distribution, but it is still suggestive that all the
513: data are not well represented by the maximum likelihood model
514: distribution. This is entirely the result of the fact that the best
515: fitting Gauss-Hermite distribution does not fit the bright sources in
516: the sample very well. We were able to produce a better fit by
517: creating a set of hybrid simulated catalogs that use the \citet{bg04}
518: distribution for brighter sources and our best fitting distribution
519: for faint sources. The transition between these regimes was made
520: smooth, with equal contributions from the two distributions at a flux
521: density of 30 mJy. These hybrid catalogs fitted the data
522: significantly better, with the best fitting model, using $\sigma_{\Pi_{\rm o}} =
523: 7.0\%$ and $h_4 = 0.05$ at low flux density, passing the 2-dimensional
524: Kolmogorov-Smirnov test. The median fractional polarization of the
525: best fit $\Pi_{\rm o}$ distribution for polarized sources with Stokes
526: $I$ below $30\,\rm mJy$ is $(4.8 \pm 0.7$)\%.
527:
528:
529: \subsection{Source Counts}
530: \label{counts-sec}
531:
532: To derive source counts of the polarized sources, the effect of the
533: varying noise level over the mosaic on source detection probability
534: (completeness correction) was measured as a function of flux density
535: by repeating the source detection on 1000 simulated images having the
536: same noise characteristics and source density as the data. Separate
537: simulations were performed for total intensity and for polarized
538: intensity images. Each simulated image contained random Gaussian
539: noise smoothed to the resolution of the DRAO images, and the same
540: uniform rms amplitude as the data. For each simulated polarization
541: image, two independent noise images were created to represent the $Q$
542: and $U$ images. Sources were placed at random positions, with flux
543: densities drawn from the source-count curve derived by
544: \citet{windhorst1990} between 0.1 and $500\,\rm mJy$. Sources below
545: the detection limit were included in the simulations to account for
546: crowding in the field, and the possibility that faint sources are
547: observed above the detection limit because of noise. The resulting
548: synthetic Stokes $I$ and polarized intensity images were then searched
549: for sources in an identical manner as for the observed images.
550:
551: Figure~\ref{dNdPmodel-fig} shows the derived differential source
552: counts for total intensity (log $N$ -- log $I$) and polarized flux
553: density (log $N$ -- log $p$) at $1420\,\rm MHz$, normalized to the
554: Euclidean expectation in the conventional way. The polarized source
555: counts are also listed in Table~\ref{polcounts-tab}. Counts were
556: derived in bins starting at $500\,\mu$Jy to avoid sources with
557: completeness correction greater than 10. This resulted in the removal
558: of the fifteen faintest sources from the counts. The Stokes $I$
559: source counts show good agreement with the Windhorst curve, although
560: we find somewhat higher numbers around $10\,\rm mJy$. The polarized
561: source counts are nearly flat in the flux density range observed,
562: consistent with the increased fractional polarization of the faint
563: radio sources.
564:
565: Figure~\ref{dNdPmodel-fig} shows predicted polarized source count
566: curves derived by convolving Windhorst's Stokes $I$ counts with the
567: strong-source $\Pi_{\rm o}$ distribution from \citet{bg04} (dashed
568: curve), as well as our hybrid $\Pi_{\rm o} $ distribution fit to the
569: $\log I - \log p$ distribution in \S~\ref{maxlike-sec} (solid
570: curve). The observed polarized source counts show an excess over the
571: \citet{bg04} extrapolation for $p < 3\,\rm mJy$, and are consistent
572: with the prediction based on the derived higher $\Pi_{\rm o}$ for
573: these sources. The data point at the lowest flux density lies
574: marginally above our predicted curve. This may suggest a continuing
575: trend toward even higher fractional polarization in the sub-mJy
576: population. The complete DRAO 30-field survey of the region, and
577: other deep polarization surveys, will test this.
578:
579:
580:
581:
582: \subsection{Identification with Spitzer objects}
583:
584: The positions of the polarized radio sources were examined in the {\sl
585: Spitzer} SWIRE images of ELAIS N1. Accurate positions for the sources
586: were obtained from the VLA FIRST survey images \citep{white1997},
587: which provides 1.4~GHz continuum Stokes $I$ images at $5\arcsec$
588: resolution with a 1-$\sigma$ sensitivity of $150\,\mu\rm
589: Jy~beam^{-1}$. The sensitivity of the FIRST images is sufficient to
590: detect a Stokes $I$ counterpart for every polarized source in the
591: ELAIS N1 deep field. Approximately 35\% of the polarized radio
592: sources showed resolved structure on scales of 2\arcsec\ to 30\arcsec\
593: in the FIRST images. A polarized source may be associated with a
594: radio lobe instead of the core of a radio galaxy. Visual inspection
595: avoided misidentification in such cases. As a comparison to the
596: polarized sources we also searched the SWIRE images for identification
597: of sources with Stokes $I$ flux density larger than $1\,\rm mJy$ but
598: no detectable polarization.
599:
600: In total, 54 polarized sources were unambiguously identified with {\sl
601: Spitzer} objects. Another 17 polarized sources had uncertain
602: identifications, i.e.\ there was more than one possible counterpart
603: within the errors of the radio position. Two sources were not covered
604: by SWIRE. The remaining 12\% of the sources (10 objects) have no
605: counterpart in the SWIRE images. Similar statistics resulted from the
606: search for counterparts of the Stokes $I$ sources with no detectable
607: polarized emission. A wide range in flux density and angular size was
608: found among the identified {\sl Spitzer} galaxies. Although some faint
609: galaxies appear unresolved in the {\sl Spitzer} images, the identified
610: polarized sources seem to be mainly associated with elliptical
611: galaxies. Three of the polarized sources have counterparts in the {\sl
612: Spitzer} images that are too faint to appear in the {\sl Spitzer}
613: ELAIS N1 source catalog. These sources were not included in the
614: subsequent analysis.
615:
616: Figure~\ref{spitzercol-fig} shows a near-infrared color-color diagram
617: of {\sl Spitzer} galaxies identified with radio sources that had
618: catalogued flux densities in all four IRAC bands at 3.6, 4.5, 5.8, and
619: $8.0\,\mu$m. This includes 41 of the polarized sources. Since many
620: of the sources are identified with extended galaxies, we used the
621: isophotal fluxes in each band. Our analysis of this color-color
622: diagram is based on the modeling of \citet{sajina2005}, which divided
623: the diagram into the four regions separated by dashed lines in
624: Figure~\ref{spitzercol-fig}. Boundaries between these regions were
625: defined so as to separate galaxies depending on the strength of
626: near-infrared PAH bands and the slope of the near-infrared
627: continuum. Region 1 is mainly populated by sources with a continuum
628: that rises with wavelength. The near-infrared spectrum of these
629: sources is usually dominated by non-equilibrium emission of
630: stochastically heated very small dust grains, interpreted as PAH
631: destruction by the hard ultraviolet spectrum of an AGN. Region 2 is
632: occupied mainly by dusty star-forming galaxies with strong PAH bands
633: at red-shift $z < 0.5$. This is because the strongest PAH features at
634: low redshift contribute to the flux in IRAC bands at 3.6 and
635: $8.0\,\mu$m. Region 3 is occupied by galaxies with fainter PAH
636: emission, or by dusty starforming galaxies at redshift $z = 0.5 -
637: 1.5$. A conspicuous concentration of galaxies in region 3 is the blue
638: clump near $\log(S_{5.8})/\log(S_{3.6}) = -0.6$,
639: $\log(S_{8.0})/\log(S_{4.5}) = -0.8$. Galaxies in the blue clump have
640: a near-infrared continuum that declines with wavelength, since the
641: spectrum is dominated by the starlight of an old stellar
642: population. These are elliptical galaxies, at a wide range of
643: redshift. Region 4 is populated by PAH-dominated sources at redshift
644: $z = 1.5 - 2$.
645:
646: The number of sources by region in Figure~\ref{spitzercol-fig} is
647: listed in Table~\ref{table_irac_colcol}. The host galaxies of the
648: polarized sources occupy mainly regions 1 and 3. The source in
649: region 4 cannot be considered a convincing high-redshift galaxy in
650: view of uncertainties in the photometry and its proximity to the
651: boundary with regions 1 and 3. The majority (71\%) of the host
652: galaxies of polarized sources in the {\sl Spitzer} color-color diagram
653: are either in region 1 or in the blue clump. In both cases, the radio
654: emission is interpreted as emission from an AGN.
655:
656: Eleven polarized sources (27\%) are found in the region of
657: PAH-dominated galaxies, with $\log(S_{8.0})/\log(S_{4.5}) > -0.5$, a
658: few dex above the blue clump. Their location in the
659: color-color diagram suggests either PAH-dominated galaxies at redshift
660: $0.5 - 1.5$, or galaxies with faint PAH bands at lower redshift
661: \citep{sajina2005}. Inspection of the FIRST and SWIRE images of these
662: sources shows that all appear to be elliptical galaxies with a smooth
663: morphology, and some have resolved symmetric radio sources, suggestive
664: of radio lobes. This suggests that the polarized radio emission in
665: these galaxies is associated with AGN activity, while the PAH
666: emission originates from dust at a substantial distance from the AGN,
667: where it is shielded from the hard ultraviolet radiation that would
668: destroy the PAHs. Dust in elliptical galaxies is a common phenomenon
669: \citep{goudfrooij1994}. It can be produced in the envelopes of cool
670: red giant stars or it can be acquired through a merger with a gas-rich
671: galaxy.
672:
673: The Stokes $I$ sources with no detectable polarized emission generally
674: occupy the same areas of the color-color diagram as the polarized
675: sources, with two exceptions. First, a significant fraction (15\%) of
676: the Stokes $I$ sources is located in region 2, the area where star
677: forming, PAH-dominated galaxies are expected. These objects are
678: generally the faintest Stokes I radio sources, and none would be
679: detectable in polarization at our sensitivity level. The Stokes $I$
680: sources in region 2 are likely members of a population of star-forming
681: galaxies that is believed to make up a large fraction of the radio
682: source population below $\sim$1 mJy. The second exception is that
683: Table~\ref{table_irac_colcol} indicates an excess of polarized sources
684: (27\%) relative to sources with no detected polarization (14\%) in
685: region 3b, associated with galaxies having PAH emission. This
686: difference is significant if Poisson errors are assumed. Confirmation
687: will require a more complete sample of polarized radio sources
688: identified with {\sl Spitzer} galaxies.
689:
690:
691: \subsection{Nature of the mJy polarized source population}
692:
693: The polarized sources found in the ELAIS N1 field have a median Stokes
694: $I$ flux density of 12 mJy. Models of radio source populations fitted
695: to the total radio source counts suggest that most radio sources with
696: $1420\,\rm MHz$ flux density $\gtrsim 1\,\rm mJy$ are steep-spectrum
697: radio galaxies, whose power is ultimately derived from accretion onto
698: a compact object. However, it is not clear a priori that a faint
699: polarization-selected sample of radio sources is representative of the
700: entire population. In principle, a highly polarized population of
701: faint radio sources may constitute a significant fraction of a sample
702: of faint polarized sources.
703:
704: In Section~\ref{maxlike-sec} we presented evidence that faint
705: extragalactic polarized radio sources are on average more highly
706: polarized than bright sources, with a median fractional polarization
707: approximately three times higher. From an analysis of sources with
708: Stokes $I > 100$ mJy in the NVSS, \citet{mesaetal2002} also noted that
709: the median fractional polarization of radio sources increases with
710: decreasing flux density, from 1.05\% for Stokes $ I > 800$ mJy to
711: 1.84\% between $100-200\,\rm mJy$. From a similar analysis of the
712: NVSS, \citet{tucci2004} showed that the anti-correlation between flux
713: density and percentage polarization occurs only for steep-spectrum
714: sources. They found that the median percentage polarization for
715: steep-spectrum sources increased from 1.14\% for flux densities
716: greater than $800\,\rm mJy$ to 1.77\% between $100 - 200\,\rm mJy$.
717: Our result extends this to much lower flux densities and indicates a
718: much stronger effect for faint sources, resulting in a median
719: polarization of $4.8\%$ at Stokes $I = 10 - 30\,\rm mJy$. Analysis
720: of Table~\ref{table_polsources} shows that these faint polarized
721: emitters are dominated by steep-spectrum sources; all but one of the
722: polarized sources for which a 325 --$1420\,\rm MHz$ spectral index
723: exists (75\%) has $\alpha < -0.4$.
724:
725: Polarized sources that can be identified with galaxies in the {\sl
726: Spitzer} ELAIS N1 deep field, are associated with elliptical
727: galaxies. Most of the host galaxies have near-infrared colors typical
728: for dust emission from the vicinity of an AGN, or an old stellar
729: population with no significant dust emission. The majority of the
730: polarized sources is associated with AGN activity for this reason.
731: The remaining 11 polarized sources associated with galaxies having PAH
732: emission are also likely to contain an AGN. Some are clearly resolved
733: double-lobed objects in the FIRST images, and all appear to be
734: elliptical galaxies with a smooth brightness distribution. We
735: conclude that there is no evidence for galaxies in our sample of
736: polarized sources in which the radio emission is powered by star
737: formation. The higher fractional polarization of faint radio sources
738: may be related to a population of radio-quiet AGN in which fainter
739: radio emission correlates with conditions that favor increased
740: polarization, for example more ordered magnetic fields or less
741: internal Faraday depolarization. High-resolution polarimetry of these
742: objects will provide more insight into their nature.
743:
744:
745: \section{Conclusions}
746:
747: We present sensitive observations of a complete sample of compact
748: polarized radio sources, as part of a deep integration of the ELAIS N1
749: region made with the Synthesis Telescope at the Dominion Radio
750: Astrophysical Observatory. A total of 83 polarized sources was
751: detected in the ten-field mosaic.
752:
753: The distribution of fractional polarization of faint polarized sources
754: was investigated with a Monte-Carlo analysis that generates synthetic
755: source lists with the same noise statistics and observational
756: selection criteria as the data. Maximum-likelihood fits of the
757: synthetic source lists to the data in the $\log(I)$ - $\log(p)$ plane
758: yielded a best fitting Gauss-Hermite function with $\sigma_{\Pi_{\rm o}} = (7.0\
759: \pm\ 1.0)\%$, $h_4 = 0.05\ \pm 0.05$ for the distribution of intrinsic
760: fractional polarization. The data demonstrate a trend of increasing
761: fractional polarization with decreasing flux density.
762:
763: Polarized source counts from the ELAIS N1 deep field are presented
764: down to $0.5\,\rm mJy$. We find that the Euclidean-normalized
765: polarized counts remain flat below $1\,\rm mJy$. The distribution of
766: fractional polarization derived from our Monte Carlo analysis is
767: convolved with the total-intensity source counts to produce a
768: prediction of the polarized source counts. The predicted
769: Euclidean-normalized polarized counts are nearly flat to $\sim 2\,\rm
770: mJy$, in good agreement with the data. However, the data at the
771: faintest polarized flux densities suggests a continuing trend of
772: increased polarization fraction with decreasing flux density.
773:
774: The near-infrared color-color diagram for host galaxies identified
775: with the polarized sources in the ELAIS N1 field shows that most of
776: the host galaxies are ellipticals, or galaxies for which the
777: near-infrared spectrum is dominated by stochastically heated very
778: small grains, presumably from the vicinity of an AGN. Some of the host
779: galaxies appear to have PAH bands in their near-infrared spectrum, but
780: the morphological resemblance with ellipticals, and the fact that some
781: of these polarized sources are resolved radio galaxies in the FIRST
782: survey, indicates that these objects also harbour AGN. We suggest that
783: the higher degree of polarization indicates a difference between AGN
784: observed at a flux density of hundreds of mJy, and fainter AGN.
785:
786: \section*{Acknowledgments}
787:
788: Observations and research on the DRAO Planck Deep Fields are supported
789: by the Natural Sciences and Engineering Research Council of Canada and
790: the National Research Council Canada. Ev Sheehan of DRAO was
791: instrumental in improving the sensitivity of the DRAO Synthesis
792: Telescope, with the outcome recorded in this paper. We are indebted to
793: him for his skill and dedication to this difficult task. The Dominion
794: Radio Astrophysical Observatory is operated as a National Facility by
795: the National Research Council of Canada.
796:
797: \begin{thebibliography}{}
798:
799: \bibitem[Beck \& Gaensler(2004)]{bg04}
800: Beck, R.\ and Gaensler, B.\ M.\ 2004, New Astronomy Reviews, 48, 1289
801:
802: \bibitem[Bernardi et~al.(2003)]{bernardi03} Bernardi, G., Carretti, E., Coriglioni, S., Sault, R.\ J., Kesteven, M.J. \& Poppi, S.\ 2003, \apjl, 594, L5.
803:
804: \bibitem[Bernardi et~al.(2006)]{bernardi06} Bernardi, G., Carretti, E., Sault, R.\ J., Cortiglioni, S.\ \& Poppi, S.\ 2006, \mnras, 370, 2064
805:
806: \bibitem[Carretti, et~al.(2005)]{carretti05} Carretti, E., Bernardi, G., Sault, R.J., Cortiglioni, S.\ \& Poppi, S.\ 2005, \mnras, 358, 1
807:
808: \bibitem[Carretti et~al.(2006)]{carretti06} Carretti, E., Poppi, S., Reich, W., Reich, P., F\"urst, E., Bernardi, G., Cortiglioni, S.\ \& Sbarra, C.\ 2006, \mnras, 367, 132
809:
810: \bibitem[Condon et~al.(1998)]{condon1998} Condon, J.\ J., Cotton, W.\ D., Greisen, E.\ W., Yin, Q.\ F., Perley, R.\ A., Taylor, G.\ B., \& Broderick, J. J. 1998, \aj, 115, 1693
811:
812: \bibitem[De Vries et~al.(2002)]{devries2002} De Vries, W. H., Morganti, R., R\"ottgering, H. J. A., Vermeulen, R., Van Breugel, W., Rengelink, R., \& Jarvis, M. J. 2002, \aj, 123, 1784
813:
814: \bibitem[Goudfrooij et~al.(1994)]{goudfrooij1994} Goudfrooij, P., Hansen, L., Jorgensen, H. E., \& Norgaard-Nielsen, H. U. 1994, \aaps, 105, 341
815:
816: \bibitem[Gruppioni et~al.(1999)]{gruppioni1999} Gruppioni, C., Mignoli, M., \& Zamorani, G. 1999, \mnras, 304, 199
817:
818: \bibitem[Hopkins et~al.(2000)]{hopkins2000} Hopkins, A., Windhorst, R., Cram, L., \& Ekers, R. 2000, Experimental Astronomy, 10, 419
819:
820: \bibitem[Hopkins et~al.(2003)]{hopkins2003} Hopkins, A. M., Afonso, J., Chan, B., Cram, L. E., Georgakakis, A., \& Mobasher, B. 2003, \aj, 125, 465
821:
822: \bibitem[Landecker et~al.(2000)]{landecker2000} Landecker, T. L., Dewdney, P. E., Burgess, T. A., Gray, A. D., Higgs, L. A., Hoffmann, A. P., Hovey, G. J., Karpa, D. R., Lacey, J. D., Prowse, N., Purton. C. R., Roger R. S., Willis, A. G., Wyslouzil, W., Routledge, D., \& Vaneldik, J. F. 2000, \aaps, 145, 509
823:
824: \bibitem[Landecker et~al.(2007)]{landecker2007} Landecker, T. L., et al. 2007, in prep.
825:
826: \bibitem[Londsdale et~al.(2003)]{lonsdale03} Lonsdale, C.J., et~al.\ 2003, \pasp, 115, 897
827:
828: \bibitem[Mesa et~al.(2002)]{mesaetal2002} Mesa, D., Baccigalupi, C., De Zotti, G., Gregorini, L., Mack, K.-H.,Vigotti, M.\ \& Klein, U.\ 2002, \aap, 396, 463
829:
830: \bibitem[Oliver et~al.(2000)]{oliver2000} Oliver, S., Rowan-Robinson, M., Alexander, D. M., Almaini, O., Balcells, M., Baker, A. C., et.\ al. 2000, \mnras, 316, 749
831:
832: \bibitem[Peacock(1983)]{peacock1983} Peacock, J. A. 1983, \mnras, 202, 615
833:
834: \bibitem[Rengelink et~al.(1997)]{rengelink1997} Rengelink, R. B., Tang, Y., De Bruyn, A. G., Miley, G. K., Bremer, M. N., R\"ottgering H. J. A., \& Bremer, M. A. R. 1997, A\&AS, 124, 259
835:
836: \bibitem[Rice(1945)]{rice1945} Rice, S. O. 1945, Bell Syst. Tech. J., 24, 46
837:
838: \bibitem[Sadler et~al.(2006)]{sadler2006} Sadler, E. M., Ricci, R., Ekers, R. D., Ekers, J. A., Hancock, P. J., Jackson, C. A., Kesteven, M. J., Murphy, T., Phillips, C., Reinfrank, R. F., Staveley-Smith, L., Subrahmanyan, R., Walker, M. A., Wilson, W. E., de Zotti, G. 2006, \mnras, 371, 898
839:
840: \bibitem[Sajina et~al.(2005)]{sajina2005} Sajina, A., Lacy, M., \& Scott, D. 2005, \apj, 621, 256
841:
842: \bibitem[Schilizzi(2004)]{schilizzi2004} Schilizzi, R.\ T.\ 2004, The Square Kilometre Array, in Proceedings of the SPIE, 5489, 62
843:
844: \bibitem[Simmons \& Stewart(1985)]{simmons1985} Simmons, J. F. L., \& Stewart, B. G. 1985, \aap, 142, 100
845:
846: \bibitem[Simpson et~al.(2006)]{simpson2006} Simpson, C., Mart\'inez-Sansigre, A., Rawlings, S., Ivison, R., Akiyama, M., Sekiguchi, K., Takata, T., Ueda, Y., \& Watson, M.\ 2006, \mnras, 372, 741
847:
848: \bibitem[Taylor et~al.(2003)]{taylor2003} Taylor, A. R., Gibson, S. J., Peracaula, M., Martin, P. G., Landecker, T. L., Brunt, C. M., Dewdney, P. E., Dougherty, S. M., Gray, A. D., Higgs, L. A., Kerton, C. R., Knee, L. B. G., Kothes, R., Purton, C. R., Uyaniker, B., Wallace, B. J., Willis, A. G., \& Durand, D. 2003, \aj, 125, 3145
849:
850: \bibitem[Tucci et~al.(2004)]{tucci2004} Tucci, M., Mart\'inez-Gonz\'alez, E., Toffolatti, L., Gonz\'alez-Nuevo, J., \& De Zotti, G. 2004, \mnras, 349, 1267
851:
852: \bibitem[Vaillancourt(2006)]{vaillancourt2006} Vaillancourt, J. E. 2006, \pasp, 118, 1340
853:
854: \bibitem[Van der Marel \& Franx(1993)]{vdmarel1993} Van der Marel, R. P., \& Franx, M. 1993, \apj, 407, 525
855:
856: \bibitem[Vinokur(1965)]{vinokur1965} Vinokur, M., 1965, Ann. d'Astrophys., 28, 412
857:
858: % Reference for FIRST:
859: \bibitem[White et~al.(1997)]{white1997} White, R. L., Becker, R. H., Helfand, D. J., \& Gregg, M. D. 1997, \apj, 475, 479
860:
861: \bibitem[Willis(1999)]{willis99} Willis A. G. 1999, A\&AS, 136, 603
862:
863: \bibitem[Willner(2006)]{willner2006} Willner, S. P., Coil, A. L., Goss, W. M., Ashby, M. L. N., Barmby, P., Huang, J.-S., Ivison, R., Koo, D. C., Egami, E., \& Miyazaki, S. 2006, \aj, 132, 2159
864:
865: \bibitem[Windhorst et~al.(1990)]{windhorst1990} Windhorst, R.\ A., Mathis, D., and Neuschaefer, L.\ 1990, in ASP Conference Series, 10 {\it Evolution of the Universe of Galaxies}, Eds. R.G. Kron, p. 389
866:
867: \bibitem[Windhorst(2003)]{windhorst2003} Windhorst, R.\ A.\ 2003, The micro-Jansky and nano-Jansky Population, in {\it New Astronomy Reviews}, 47, 357.
868:
869: \end{thebibliography}{}
870: \clearpage
871: \begin{figure}[h]
872: \begin{center}
873: \resizebox{14cm}{!}{\includegraphics[angle=0]{f1.eps}}
874: \caption{ Survey area of the first 10 fields of the DRAO mosaic in
875: relation to the sky coverage of the original ELAIS N1 field (dashed
876: lines) and the SWIRE survey (solid black lines). The ten field
877: centers are indicated by the $+$ symbols. Gray contours indicate the
878: sensitivity at 1.5, 2, and 3 times the theoretical noise level in the
879: center of the mosaic, and the edge of the field of view (thick
880: line).
881: \label{surveyarea-fig}
882: }
883: \end{center}
884: \end{figure}
885:
886:
887: \begin{figure}[h]
888: \begin{center}
889: %\resizebox{14cm}{!}{\includegraphics[angle=0]{f2.eps}}
890: \caption{ {\bf [This Figure is provided as a separate image f2.gif]}
891: Continuum images of the ELAIS N1 field at 21 cm wavelength. Top row:
892: Stokes I (gray scales linear from $-0.1$ to $+5$ mJy
893: beam$^{-1}$). Middle row: Stokes $Q$ (gray scales linear from $-1$ to
894: $+1$ mJy beam$^{-1}$). Bottom row: Stokes $U$ (gray scales linear from
895: $-1$ to $+1$ mJy beam$^{-1}$). Panels on the right show an enlargement
896: of the area indicated by the white frame in the Stokes $Q$ image.
897: \label{EN21-fig}
898: }
899: \end{center}
900: \end{figure}
901:
902: \begin{figure}
903: \begin{center}
904: %\resizebox{14cm}{!}{\includegraphics[angle=0]{f3.eps}}
905: \caption{ {\bf [This Figure is provided as a separate image f3.gif]}
906: The distribution of amplitudes in the $Q$ and $U$ images after
907: dividing by the mosaic weights to produce an image with uniform noise
908: over the map equal to the noise value at the map center. Gaussian
909: fits to the distributions (solid curves) were used to measure the map
910: center rms at $78\,\mu$Jy.
911: \label{fig:Q_Unoise}
912: }
913: \end{center}
914: \end{figure}
915:
916:
917: \begin{figure}
918: \begin{center}
919: \resizebox{14cm}{!}{\includegraphics[angle=0]{f4.eps}}
920: \caption{ Illustration of the effect of noise and the polarized flux density
921: detection threshold on $\Pi$ as a function of flux density. This figure shows
922: the variation of $\Pi$ with flux density for a simulated sample of
923: sources, all 5\% polarized, with random Gaussian noise added in $I$, $Q$,
924: and $U$. The error bars are derived from standard error propagation,
925: assuming the noise is known. The curve shows the effect of polarization
926: bias on $\Pi$, defined as $\sqrt{(p_{\rm o}^2+\sigma^2)}/I$ \citep{simmons1985}.
927: \label{polstat_demo}
928: }
929: \end{center}
930: \end{figure}
931:
932: \begin{figure}
933: \begin{center}
934: \resizebox{14cm}{!}{\includegraphics[angle=0]{f5.eps}}
935: \caption{ Distribution in $\log(I)$ - $\log(p)$ of observed sources
936: (open circles), compared with the simulated distribution that assumes
937: the $\Pi_{\rm o}$ distribution of \citet{bg04}. Gray scales and contours show
938: the 2-dimensional probability density function of sources in the
939: simulated catalog. The inner contours enclose 25\% and 50\% of the simulated
940: sources, while the outer contour encloses 90\%. Two lines mark
941: the loci of sources with $\Pi = 1\%$ (right) and $\Pi = 10\%$ (left).
942: \label{beck-fig}
943: }
944: \end{center}
945: \end{figure}
946:
947: \begin{figure}
948: \begin{center}
949: \resizebox{14cm}{!}{\includegraphics[angle=0]{f6.eps}}
950: \caption{ Same as Figure~\ref{beck-fig}, now for the best fitting Gauss-Hermite
951: $\Pi_{\rm o}$ distribution, with $\sigma_{\Pi_{\rm o}} = 7.0\%$ and $h_4 = 0.05$.
952: \label{bestfit_Pi}
953: }
954: \end{center}
955: \end{figure}
956:
957: \begin{figure}
958: \begin{center}
959: \resizebox{14cm}{!}{\includegraphics[angle=0]{f7.eps}}
960: \caption{ Euclidean-normalized source counts for total flux density
961: (triangles) and polarized flux density (circles). The upper solid curve is the
962: fit to observed source counts from \citet{windhorst1990}. The lower
963: solid curve shows polarized source counts predicted by convolving the
964: Stokes $I$ source counts with the $\Pi$ distribution derived from ELAIS
965: N1 data for faint sources and the \citet{bg04} distribution for bright
966: sources, as explained in the text. The dotted curve shows polarized source
967: counts derived by convolving only the \citet{bg04} distribution with the
968: \citet{windhorst1990} source counts curve.
969: \label{dNdPmodel-fig}
970: }
971: \end{center}
972: \end{figure}
973:
974:
975:
976:
977: \begin{figure}
978: \begin{center}
979: \resizebox{14cm}{!}{\includegraphics[angle=0]{f8.eps}}
980: \caption{ {\sl Spitzer} near-infrared color-color diagram of ELAIS N1 radio
981: source host galaxies. Stars: polarized sources; circles: sources with
982: no detectable polarized emission. The division of this diagram into four
983: regions and their interpretation follows \citet{sajina2005} and is
984: explained in the text.
985: \label{spitzercol-fig}
986: }
987: \end{center}
988: \end{figure}
989:
990: \begin{deluxetable}{cccrrrrrr}
991: \tabletypesize{\scriptsize}
992: \tablecaption{ ELAIS N1 Polarized Sources \label{table_polsources} }
993: \tablehead{
994: \colhead{Source} & \colhead{RA} & \colhead{DEC} & \colhead{$I$} & \colhead{$p_{o}$} &
995: \colhead{PA} & \colhead{$\Pi_0$} & \multicolumn{2}{c}{$\alpha_{325-1420}$} \\
996: \colhead{Number} & \multicolumn{2}{c}{(J2000)} &\colhead{(mJy)}&\colhead{(mJy beam$^{-1}$)} &
997: \colhead{($^{\circ}$)} & \colhead{(\%)} & & \\
998: }
999: \startdata
1000: 1 & 16 02 05.52 $\pm$ 0.29 & 54 54 43.2 $\pm$ 3.0 & 16.82 $\pm$ 0.56 & 1.74 $\pm$ 0.11 & 26 & 10.4 $\pm$ 0.7 & $-$0.78 $\pm$ 0.06 \\
1001: 2 & 16 02 34.37 $\pm$ 0.48 & 54 54 01.8 $\pm$ 4.2 & 7.95 $\pm$ 0.40 & 0.98 $\pm$ 0.16 & 37 & 11.8 $\pm$ 1.9 & $-$0.75 $\pm$ 0.11 \\
1002: 3 & 16 02 42.00 $\pm$ 0.44 & 55 10 01.2 $\pm$ 4.4 & 8.51 $\pm$ 0.78 & 0.78 $\pm$ 0.10 & 71 & 19.9 $\pm$ 2.7 & $-$0.23 $\pm$ 0.22 \\
1003: 4 & 16 03 11.52 $\pm$ 0.34 & 55 39 07.2 $\pm$ 3.8 & 4.65 $\pm$ 0.40 & 1.37 $\pm$ 0.11 & 27 & 28.2 $\pm$ 2.4 & $>-$0.64 \hfill \quad & \\
1004: 5 & 16 03 34.08 $\pm$ 0.44 & 54 29 20.4 $\pm$ 3.9 & 4.53 $\pm$ 0.37 & 0.79 $\pm$ 0.10 & 36 & 17.2 $\pm$ 2.3 & $-$1.46 $\pm$ 0.07 \\
1005: 6 & 16 04 42.24 $\pm$ 0.14 & 54 38 45.6 $\pm$ 1.4 & 19.42 $\pm$ 0.55 & 1.88 $\pm$ 0.11 & 33 & 10.5 $\pm$ 0.7 & $-$0.99 $\pm$ 0.04 \\
1006: 7 & 16 05 05.52 $\pm$ 0.15 & 55 00 46.8 $\pm$ 1.4 & 27.86 $\pm$ 0.74 & 1.25 $\pm$ 0.11 & 7 & 5.3 $\pm$ 0.5 & $-$1.33 $\pm$ 0.03 \\
1007: 8 & 16 05 23.04 $\pm$ 0.20 & 54 29 31.2 $\pm$ 1.8 & 6.39 $\pm$ 0.29 & 1.16 $\pm$ 0.11 & $-$12 & 17.0 $\pm$ 1.6 & $>-$0.43 \hfill \quad & \\
1008: 9 & 16 05 38.88 $\pm$ 0.04 & 54 39 28.8 $\pm$ 0.4 & 187.61 $\pm$ 4.74 & 8.42 $\pm$ 0.24 & 21 & 5.0 $\pm$ 0.2 & $-$1.00 $\pm$ 0.03 \\
1009: 10 & 16 05 38.88 $\pm$ 0.20 & 54 41 27.6 $\pm$ 1.9 & 2.00 $\pm$ 0.35 & 1.28 $\pm$ 0.11 & $-$37 & 47.3 $\pm$ 4.5 & $>-$1.22 \hfill \quad & \\
1010: 11 & 16 06 01.44 $\pm$ 0.04 & 54 54 10.8 $\pm$ 0.5 & 209.87 $\pm$ 5.29 & 5.24 $\pm$ 0.17 & 15 & 2.9 $\pm$ 0.1 & $-$0.91 $\pm$ 0.03 \\
1011: 12 & 16 06 08.40 $\pm$ 0.15 & 55 16 04.8 $\pm$ 1.6 & 5.06 $\pm$ 0.28 & 0.86 $\pm$ 0.10 & 16 & 15.8 $\pm$ 2.0 & $-$0.69 $\pm$ 0.19 \\
1012: 13 & 16 06 13.44 $\pm$ 0.23 & 55 01 58.8 $\pm$ 2.3 & 1.14 $\pm$ 0.18 & 0.71 $\pm$ 0.11 & $-$14 & 52.6 $\pm$ 9.3 & $>-$1.59 \hfill \quad & \\
1013: 14 & 16 06 35.52 $\pm$ 0.10 & 54 35 02.4 $\pm$ 1.0 & 15.96 $\pm$ 0.51 & 1.61 $\pm$ 0.11 & 44 & 12.1 $\pm$ 0.9 & $-$0.73 $\pm$ 0.06 \\
1014: 15 & 16 06 58.80 $\pm$ 0.32 & 54 43 12.0 $\pm$ 3.0 & 4.31 $\pm$ 0.21 & 0.41 $\pm$ 0.10 & $-$9 & 8.6 $\pm$ 2.2 & $>-$0.69 \hfill \quad & \\
1015: 16 & 16 07 22.56 $\pm$ 0.22 & 55 31 04.8 $\pm$ 2.8 & 12.16 $\pm$ 0.45 & 0.73 $\pm$ 0.10 & 1 & 8.0 $\pm$ 1.2 & $-$1.05 $\pm$ 0.05 \\
1016: 17 & 16 08 21.36 $\pm$ 0.11 & 56 13 51.6 $\pm$ 0.8 & 221.05 $\pm$ 5.59 & 3.65 $\pm$ 0.15 & 54 & 2.0 $\pm$ 0.1 & 0.30 $\pm$ 0.03 \\
1017: 18 & 16 08 28.56 $\pm$ 0.28 & 54 10 37.2 $\pm$ 3.7 & 19.94 $\pm$ 0.56 & 0.59 $\pm$ 0.10 & 26 & 3.0 $\pm$ 0.5 & $-$0.68 $\pm$ 0.06 \\
1018: 19 & 16 08 38.64 $\pm$ 0.29 & 54 14 34.8 $\pm$ 3.3 & 2.41 $\pm$ 0.21 & 0.51 $\pm$ 0.10 & $-$19 & 17.0 $\pm$ 3.5 & $>-$1.09 \hfill \quad & \\
1019: 20 & 16 08 47.76 $\pm$ 0.20 & 56 11 16.8 $\pm$ 1.6 & 27.09 $\pm$ 1.03 & 1.66 $\pm$ 0.11 & 8 & 7.8 $\pm$ 0.6 & $-$1.19 $\pm$ 0.03 \\
1020: 21 & 16 08 58.32 $\pm$ 0.36 & 55 56 27.6 $\pm$ 2.2 & 9.29 $\pm$ 0.48 & 0.69 $\pm$ 0.10 & 69 & 12.0 $\pm$ 1.8 & $-$0.82 $\pm$ 0.09 \\
1021: 22 & 16 09 04.32 $\pm$ 0.26 & 56 10 33.6 $\pm$ 2.2 & 12.68 $\pm$ 0.60 & 1.33 $\pm$ 0.11 & 23 & 11.9 $\pm$ 1.0 & $-$1.13 $\pm$ 0.05 \\
1022: 23 & 16 09 11.04 $\pm$ 0.18 & 55 26 31.2 $\pm$ 2.2 & 4.51 $\pm$ 0.22 & 0.78 $\pm$ 0.10 & 4 & 16.9 $\pm$ 2.3 & $-$0.86 $\pm$ 0.17 \\
1023: 24 & 16 09 22.80 $\pm$ 0.15 & 56 15 03.6 $\pm$ 1.4 & 32.46 $\pm$ 0.95 & 2.14 $\pm$ 0.12 & 21 & 8.0 $\pm$ 0.5 & $-$0.63 $\pm$ 0.04 \\
1024: 25 & 16 09 31.68 $\pm$ 0.12 & 55 25 04.8 $\pm$ 1.3 & 15.32 $\pm$ 0.47 & 1.41 $\pm$ 0.11 & 25 & 11.4 $\pm$ 0.9 & $-$1.05 $\pm$ 0.05 \\
1025: 26 & 16 09 36.24 $\pm$ 0.18 & 55 27 03.6 $\pm$ 1.5 & 10.95 $\pm$ 0.37 & 1.12 $\pm$ 0.10 & 8 & 11.7 $\pm$ 1.1 & $-$0.71 $\pm$ 0.09 \\
1026: 27 & 16 09 44.40 $\pm$ 0.38 & 54 37 51.6 $\pm$ 2.9 & 6.61 $\pm$ 0.24 & 0.35 $\pm$ 0.10 & 59 & 5.3 $\pm$ 1.6 & $>-$0.40 \hfill \quad & \\
1027: 28 & 16 09 52.56 $\pm$ 0.24 & 55 07 08.4 $\pm$ 2.5 & 0.68 $\pm$ 0.14 & 0.42 $\pm$ 0.10 & 1 & 44.1 $\pm$ 12.0 & $>-$1.95 \hfill \quad & \\
1028: 29 & 16 10 03.12 $\pm$ 0.10 & 55 52 37.2 $\pm$ 1.1 & 96.83 $\pm$ 2.47 & 1.59 $\pm$ 0.11 & 30 & 2.0 $\pm$ 0.2 & $-$0.79 $\pm$ 0.03 \\
1029: 30 & 16 10 27.12 $\pm$ 0.14 & 54 12 54.0 $\pm$ 1.6 & 8.51 $\pm$ 0.28 & 1.24 $\pm$ 0.11 & 30 & 14.4 $\pm$ 1.3 & $-$0.43 $\pm$ 0.17 \\
1030: 31 & 16 10 57.84 $\pm$ 0.08 & 55 35 24.0 $\pm$ 0.7 & 19.46 $\pm$ 0.53 & 1.61 $\pm$ 0.11 & 5 & 9.7 $\pm$ 0.7 & $-$0.63 $\pm$ 0.06 \\
1031: 32 & 16 11 00.48 $\pm$ 0.06 & 54 42 03.6 $\pm$ 0.6 & 29.41 $\pm$ 0.76 & 2.03 $\pm$ 0.11 & 19 & 7.8 $\pm$ 0.5 & $-$0.74 $\pm$ 0.04 \\
1032: 33 & 16 11 20.40 $\pm$ 0.08 & 55 28 44.4 $\pm$ 0.9 & 18.22 $\pm$ 0.50 & 1.39 $\pm$ 0.11 & 3 & 8.8 $\pm$ 0.7 & $-$0.88 $\pm$ 0.05 \\
1033: 34 & 16 11 21.12 $\pm$ 0.30 & 54 31 55.2 $\pm$ 3.3 & 3.62 $\pm$ 0.34 & 0.43 $\pm$ 0.10 & $-$5 & 9.9 $\pm$ 2.4 & $>-$0.81 \hfill \quad & \\
1034: 35 & 16 11 29.04 $\pm$ 0.29 & 55 51 36.0 $\pm$ 2.6 & 13.22 $\pm$ 0.41 & 0.44 $\pm$ 0.10 & 25 & 4.3 $\pm$ 1.0 & $-$0.87 $\pm$ 0.06 \\
1035: 36 & 16 11 37.92 $\pm$ 0.29 & 53 59 34.8 $\pm$ 4.2 & 13.82 $\pm$ 0.44 & 0.82 $\pm$ 0.10 & $-$18 & 5.9 $\pm$ 0.8 & $-$1.03 $\pm$ 0.05 \\
1036: 37 & 16 11 38.16 $\pm$ 0.26 & 55 59 52.8 $\pm$ 2.7 & 4.50 $\pm$ 0.23 & 0.48 $\pm$ 0.10 & 14 & 11.3 $\pm$ 2.4 & $-$1.19 $\pm$ 0.11 \\
1037: 38 & 16 11 50.88 $\pm$ 0.22 & 55 00 54.0 $\pm$ 1.0 & 9.14 $\pm$ 0.41 & 0.82 $\pm$ 0.10 & 14 & 16.3 $\pm$ 2.1 & $-$0.63 $\pm$ 0.12 \\
1038: 39 & 16 12 12.24 $\pm$ 0.02 & 55 22 48.0 $\pm$ 0.2 & 312.36 $\pm$ 7.86 & 20.37 $\pm$ 0.52 & 29 & 8.0 $\pm$ 0.3 & $-$1.11 $\pm$ 0.03 \\
1039: 40 & 16 12 24.00 $\pm$ 0.13 & 55 26 02.4 $\pm$ 1.5 & 9.57 $\pm$ 0.34 & 0.74 $\pm$ 0.10 & 32 & 8.7 $\pm$ 1.2 & $-$0.56 $\pm$ 0.12 \\
1040: 41 & 16 12 28.56 $\pm$ 0.34 & 55 06 46.8 $\pm$ 2.0 & 10.06 $\pm$ 0.30 & 0.37 $\pm$ 0.10 & 59 & 4.1 $\pm$ 1.1 & $-$1.13 $\pm$ 0.06 \\
1041: 42 & 16 12 31.68 $\pm$ 0.31 & 54 18 10.8 $\pm$ 2.5 & 9.81 $\pm$ 0.36 & 0.43 $\pm$ 0.10 & 21 & 4.5 $\pm$ 1.1 & $-$0.73 $\pm$ 0.10 \\
1042: 43 & 16 12 35.28 $\pm$ 0.04 & 56 28 19.2 $\pm$ 0.4 & 176.88 $\pm$ 4.51 & 11.47 $\pm$ 0.31 & 39 & 7.8 $\pm$ 0.3 & $-$0.96 $\pm$ 0.03 \\
1043: 44 & 16 12 47.52 $\pm$ 0.33 & 55 02 31.2 $\pm$ 2.9 & 7.68 $\pm$ 0.25 & 0.35 $\pm$ 0.10 & 51 & 4.9 $\pm$ 1.4 & $-$0.74 $\pm$ 0.12 \\
1044: 45 & 16 12 51.36 $\pm$ 0.20 & 56 03 50.4 $\pm$ 2.4 & 2.59 $\pm$ 0.19 & 0.61 $\pm$ 0.10 & 25 & 20.9 $\pm$ 3.7 & $>-$1.04 \hfill \quad & \\
1045: 46 & 16 13 02.40 $\pm$ 0.15 & 54 32 27.6 $\pm$ 1.3 & 6.48 $\pm$ 0.25 & 0.73 $\pm$ 0.10 & 56 & 11.5 $\pm$ 1.7 & $-$0.86 $\pm$ 0.12 \\
1046: 47 & 16 13 16.80 $\pm$ 0.36 & 56 08 13.2 $\pm$ 3.8 & 3.85 $\pm$ 0.36 & 0.56 $\pm$ 0.10 & $-$28 & 14.1 $\pm$ 2.6 & $>-$0.77 \hfill \quad & \\
1047: 48 & 16 13 19.20 $\pm$ 0.09 & 54 16 40.8 $\pm$ 1.0 & 5.78 $\pm$ 0.43 & 1.34 $\pm$ 0.11 & 5 & 31.5 $\pm$ 2.8 & $>-$0.50 \hfill \quad & \\
1048: 49 & 16 13 25.92 $\pm$ 0.24 & 55 39 39.6 $\pm$ 2.4 & 3.06 $\pm$ 0.18 & 0.35 $\pm$ 0.10 & 17 & 12.2 $\pm$ 3.6 & $>-$0.93 \hfill \quad & \\
1049: 50 & 16 13 26.64 $\pm$ 0.10 & 55 15 46.8 $\pm$ 1.1 & 14.60 $\pm$ 0.14 & 0.80 $\pm$ 0.10 & 14 & 6.5 $\pm$ 0.9 & $-$1.19 $\pm$ 0.04 \\
1050: 51 & 16 13 28.80 $\pm$ 0.27 & 56 17 49.2 $\pm$ 2.1 & 17.29 $\pm$ 0.57 & 0.87 $\pm$ 0.10 & 26 & 6.8 $\pm$ 0.8 & $-$0.53 $\pm$ 0.07 \\
1051: 52 & 16 13 30.72 $\pm$ 0.06 & 54 27 21.6 $\pm$ 0.6 & 80.29 $\pm$ 2.03 & 2.67 $\pm$ 0.12 & 15 & 3.7 $\pm$ 0.2 & $-$0.85 $\pm$ 0.03 \\
1052: 53 & 16 13 36.72 $\pm$ 0.36 & 54 11 16.8 $\pm$ 2.7 & 1.70 $\pm$ 0.36 & 0.59 $\pm$ 0.10 & 1 & 22.9 $\pm$ 4.2 & $>-$1.33 \hfill \quad & \\
1053: 54 & 16 13 41.76 $\pm$ 0.19 & 56 11 49.2 $\pm$ 1.4 & 101.05 $\pm$ 7.62 & 2.80 $\pm$ 0.12 & 22 & 10.9 $\pm$ 0.6 & 0.09 $\pm$ 0.04 \\
1054: 55 & 16 13 48.48 $\pm$ 0.27 & 54 14 13.2 $\pm$ 2.5 & 8.12 $\pm$ 0.40 & 0.64 $\pm$ 0.10 & 18 & 7.3 $\pm$ 1.2 & $-$1.10 $\pm$ 0.07 \\
1055: 56 & 16 13 56.16 $\pm$ 0.25 & 54 57 28.8 $\pm$ 2.5 & 0.45 $\pm$ 0.09 & 0.28 $\pm$ 0.10 & $-$1 & 50.6 $\pm$ 19.2 & $-$3.97 $\pm$ 0.03 \\
1056: 57 & 16 13 56.84 $\pm$ 0.22 & 55 02 08.2 $\pm$ 2.3 & 0.81 $\pm$ 0.13 & 0.50 $\pm$ 0.02 & $-$31 & 49.8 $\pm$ 5.7 & $>-$1.82 \hfill \quad & \\
1057: 58 & 16 14 00.96 $\pm$ 0.20 & 53 57 21.6 $\pm$ 1.8 & 14.75 $\pm$ 0.49 & 1.08 $\pm$ 0.10 & 32 & 7.5 $\pm$ 0.8 & $-$0.41 $\pm$ 0.10 \\
1058: 59 & 16 14 16.80 $\pm$ 0.29 & 55 42 57.6 $\pm$ 3.5 & 1.14 $\pm$ 0.36 & 0.46 $\pm$ 0.10 & 56 & 64.0 $\pm$ 17.2 & $>-$1.60 \hfill \quad & \\
1059: 60 & 16 14 21.12 $\pm$ 0.06 & 55 36 39.6 $\pm$ 0.6 & 35.41 $\pm$ 0.93 & 3.00 $\pm$ 0.13 & 18 & 9.9 $\pm$ 0.5 & $-$0.74 $\pm$ 0.04 \\
1060: 61 & 16 14 32.64 $\pm$ 0.20 & 55 38 31.2 $\pm$ 2.5 & 1.27 $\pm$ 0.39 & 0.47 $\pm$ 0.10 & $-$79 & 90.8 $\pm$ 27.5 & $>-$1.52 \hfill \quad & \\
1061: 62 & 16 15 27.36 $\pm$ 0.13 & 54 27 10.8 $\pm$ 1.2 & 4.51 $\pm$ 0.20 & 0.97 $\pm$ 0.10 & 5 & 20.1 $\pm$ 2.3 & $>-$0.66 \hfill \quad & \\
1062: 63 & 16 15 30.96 $\pm$ 0.16 & 54 52 30.0 $\pm$ 2.0 & 5.43 $\pm$ 0.22 & 0.68 $\pm$ 0.10 & 20 & 13.1 $\pm$ 2.0 & $-$1.16 $\pm$ 0.09 \\
1063: 64 & 16 15 36.72 $\pm$ 0.19 & 53 46 37.2 $\pm$ 2.4 & 54.20 $\pm$ 1.46 & 2.64 $\pm$ 0.12 & 19 & 5.6 $\pm$ 0.3 & $-$0.85 $\pm$ 0.03 \\
1064: 65 & 16 15 49.68 $\pm$ 0.06 & 55 16 40.8 $\pm$ 0.7 & 27.80 $\pm$ 0.72 & 1.75 $\pm$ 0.11 & 35 & 7.4 $\pm$ 0.5 & $-$0.68 $\pm$ 0.04 \\
1065: 66 & 16 16 23.04 $\pm$ 0.14 & 55 27 00.0 $\pm$ 1.3 & 13.10 $\pm$ 0.40 & 0.90 $\pm$ 0.10 & 27 & 8.8 $\pm$ 1.0 & $-$0.89 $\pm$ 0.06 \\
1066: 67 & 16 16 23.52 $\pm$ 0.17 & 54 57 43.2 $\pm$ 1.6 & 10.35 $\pm$ 0.48 & 0.60 $\pm$ 0.10 & 10 & 6.0 $\pm$ 1.0 & $-$0.97 $\pm$ 0.07 \\
1067: 68 & 16 16 37.92 $\pm$ 0.18 & 55 45 14.4 $\pm$ 2.1 & 74.66 $\pm$ 1.90 & 0.85 $\pm$ 0.10 & 51 & 1.3 $\pm$ 0.2 & $-$0.61 $\pm$ 0.03 \\
1068: 69 & 16 16 39.36 $\pm$ 0.11 & 53 58 12.0 $\pm$ 0.9 & 351.23 $\pm$ 9.13 & 5.14 $\pm$ 0.19 & 23 & 1.7 $\pm$ 0.1 & $-$0.85 $\pm$ 0.03 \\
1069: 70 & 16 16 40.08 $\pm$ 0.18 & 56 20 38.4 $\pm$ 1.5 & 18.64 $\pm$ 0.70 & 1.37 $\pm$ 0.11 & 24 & 9.3 $\pm$ 0.8 & $-$0.52 $\pm$ 0.07 \\
1070: 71 & 16 17 57.60 $\pm$ 0.22 & 54 51 36.0 $\pm$ 3.0 & 14.65 $\pm$ 0.46 & 0.68 $\pm$ 0.10 & $-$17 & 5.9 $\pm$ 0.9 & $-$1.11 $\pm$ 0.04 \\
1071: 72 & 16 18 06.72 $\pm$ 0.36 & 54 42 46.8 $\pm$ 2.4 & 5.31 $\pm$ 0.26 & 0.65 $\pm$ 0.10 & 24 & 12.3 $\pm$ 2.0 & $>-$0.55 \hfill \quad & \\
1072: 73 & 16 18 32.64 $\pm$ 0.05 & 54 31 44.4 $\pm$ 0.5 & 48.68 $\pm$ 1.27 & 4.56 $\pm$ 0.15 & $-$1 & 10.3 $\pm$ 0.4 & $-$0.78 $\pm$ 0.03 \\
1073: 74 & 16 18 57.57 $\pm$ 0.12 & 54 29 26.2 $\pm$ 1.3 & 132.06 $\pm$ 5.88 & 2.55 $\pm$ 0.14 & 21 & 2.3 $\pm$ 0.2 & $-$1.39 $\pm$ 0.03 \\
1074: 75 & 16 18 59.28 $\pm$ 0.18 & 54 52 40.8 $\pm$ 1.7 & 40.08 $\pm$ 1.07 & 1.86 $\pm$ 0.11 & 21 & 6.0 $\pm$ 0.4 & $-$0.50 $\pm$ 0.04 \\
1075: 76 & 16 19 15.36 $\pm$ 0.20 & 55 05 13.2 $\pm$ 1.4 & 16.65 $\pm$ 0.50 & 0.69 $\pm$ 0.10 & $-$2 & 5.7 $\pm$ 0.9 & $-$0.30 $\pm$ 0.11 \\
1076: 77 & 16 19 19.20 $\pm$ 0.08 & 55 35 56.4 $\pm$ 0.8 & 53.48 $\pm$ 1.36 & 1.98 $\pm$ 0.11 & 15 & 4.1 $\pm$ 0.3 & $-$0.44 $\pm$ 0.04 \\
1077: 78 & 16 19 19.44 $\pm$ 0.06 & 54 48 25.2 $\pm$ 0.6 & 57.59 $\pm$ 1.46 & 3.32 $\pm$ 0.13 & 11 & 6.1 $\pm$ 0.3 & $-$0.67 $\pm$ 0.03 \\
1078: 79 & 16 19 24.24 $\pm$ 0.18 & 55 50 52.8 $\pm$ 1.6 & 2.18 $\pm$ 0.21 & 0.79 $\pm$ 0.10 & 19 & 26.8 $\pm$ 3.7 & $>-$1.16 \hfill \quad & \\
1079: 80 & 16 21 13.68 $\pm$ 0.10 & 55 23 42.0 $\pm$ 0.9 & 58.11 $\pm$ 1.49 & 2.62 $\pm$ 0.12 & 50 & 4.8 $\pm$ 0.3 & $-$0.94 $\pm$ 0.03 \\
1080: 81 & 16 21 18.72 $\pm$ 0.28 & 55 38 27.6 $\pm$ 2.3 & 7.97 $\pm$ 0.40 & 0.74 $\pm$ 0.10 & 7 & 8.5 $\pm$ 1.2 & $-$0.98 $\pm$ 0.08 \\
1081: 82 & 16 21 45.36 $\pm$ 0.10 & 55 49 37.2 $\pm$ 1.0 & 31.10 $\pm$ 1.01 & 2.05 $\pm$ 0.11 & 60 & 6.9 $\pm$ 0.4 & $-$0.62 $\pm$ 0.04 \\
1082: 83 & 16 22 08.64 $\pm$ 0.14 & 55 24 28.8 $\pm$ 1.6 & 14.23 $\pm$ 2.15 & 4.55 $\pm$ 0.15 & 6 & 27.7 $\pm$ 1.2 & $-$1.57 $\pm$ 0.03 \\
1083: \enddata
1084: \end{deluxetable}
1085:
1086: \begin{deluxetable}{lcccc}
1087: \tablecolumns{3}
1088: \tablewidth{0pc}
1089: %\tabletypesize{\scriptsize}
1090: \tablecaption{ Polarized source counts \label{polcounts-tab} }
1091: \tablehead{
1092: \colhead{$p$} & \colhead{\ \ \ \ \ \ \ \ \ } & \colhead{$N$} & \colhead{\ \ \ \ \ } & \colhead{$p^{2.5} dN/dp$}\\
1093: \colhead{(mJy)} & \colhead{\ } & \colhead{\ } & \colhead{\ } & \colhead{(Jy$^{1.5}$ sr$^{-1}$)} \\
1094: }
1095: \startdata
1096: \phantom{0}0.71 & & 29 & & 1.38 $\pm$ 0.26 \\
1097: \phantom{0}1.42 & & 20 & & 1.26 $\pm$ 0.28 \\
1098: \phantom{0}2.50 & & 11 & & 1.10 $\pm$ 0.33 \\
1099: \phantom{0}4.61 & & \phantom{0}5 & & 1.01 $\pm$ 0.45 \\
1100: \phantom{0}9.94 & & \phantom{0}2 & & 1.46 $\pm$ 1.03 \\
1101: 20.4 & & \phantom{0}1 & & 2.32 $\pm$ 2.32 \\
1102: \enddata
1103: \end{deluxetable}
1104:
1105:
1106:
1107:
1108: \begin{deluxetable}{ccccc}
1109: \tablecolumns{5}
1110: \tablewidth{0pc}
1111: %\tabletypesize{\scriptsize}
1112: \tablecaption{ Radio sources in the IRAC color-color diagram \label{table_irac_colcol} }
1113: \tablehead{
1114: \colhead{Region} & \multicolumn{2}{c}{Polarization detected} & \multicolumn{2}{c}{Other sources}\\
1115: \colhead{} & \colhead{Number} & \colhead{\%} & \colhead{Number} & \colhead{\%}
1116: \\
1117: }
1118: \startdata
1119: 1 &\phantom{0}8 & 20$\pm$\phantom{0}7 & 32 & 19$\pm$3 \\
1120: 2 &\phantom{0}0 & 0 & 25 & 15$\pm$3 \\
1121: 3a\tablenotemark{a} & 21 & 51$\pm$11 & 80 & 47$\pm$5 \\
1122: 3b\tablenotemark{b} & 11 & 27$\pm$\phantom{0}8 & 24 & 14$\pm$3 \\
1123: 4 &\phantom{0}1 &\phantom{0}2$\pm$\phantom{0}2 & \phantom{0}9 & \phantom{0}5$\pm$2 \\
1124: \enddata
1125: \tablenotetext{a}{Region 3 blue clump, selected by $\log(S_{8.0})/\log(S_{4.5})\leqq -0.5$}
1126: \tablenotetext{b}{Region 3 PAH, selected by $\log(S_{8.0})/\log(S_{4.5}) > -0.5$}
1127: \end{deluxetable}
1128:
1129:
1130: \end{document}
1131:
1132:
1133: