We present the results of a study of a sample of 375 Extremely Red Galaxies (ERGs) in the Phoenix Deep Survey, 273 of which constitute a subsample which is 80% complete to K_s = 18.5 over an area of 1160 arcmin^2. The angular correlation function for ERGs is estimated, and the association of ERGs with faint radio sources explored. We find tentative evidence that ERGs and faint radio sources are associated at z > 0.5. A new overdensity-mapping algorithm has been used to characterize the ERG distribution, and identify a number of cluster candidates, including a likely cluster containing ERGs at 0.5 < z < 1. Our algorithm is also used in an attempt to probe the environments in which faint radio sources and ERGs are associated. We find limited evidence that the I - K_s > 4 criterion is more efficient than R - K_s > 5 at selecting dusty star-forming galaxies, rather than passively evolving ERGs.

We estimate the star-formation rates and the stellar masses of the Extremely Red objects (EROs) detected in a 180arcmin2 Ks-band survey (Ks~20mag). This sample is complemented by sensitive 1.4GHz radio observations (12micro-Jy; 1sigma rms) and multiwaveband photometric data (UBVRIJ) as part of the Phoenix Deep Survey. For bright K<19.5mag EROs in this sample (I-K>4mag; total of 177) we use photometric methods to discriminate dust-enshrouded active systems from early-type galaxies and to constrain their redshifts. Radio stacking is then employed to estimate mean radio flux densities of 8.6 (3sigma) and 6.4micro-Jy (2.4sigma) for the dusty and early-type subsamples respectively. Assuming that dust enshrouded active EROs are powered by star-formation the above radio flux density at the median redshift of z=1 translates to a radio luminosity of 4.5e22W/Hz and a star-formation rate of SFR=25Mo/yr. Combining this result with photometric redshift estimates we find a lower limit to the star-formation rate density of ~0.02Mo/yr/Mpc^3 for the K<19.5mag dusty EROs in the range z=0.85-1.35. Comparison with the SFR density estimated from previous ERO samples (with similar selection criteria) using optical emission lines, suffering dust attenuation, suggests a mean dust reddening of at least E(B-V)~0.5 for this population. We further use the Ks-band luminosity as proxy to stellar mass and argue that the dust enshrouded EROs in our sample are massive systems, M>5e10Mo. We also find that EROs represent a sizable fraction (~50%) of the number density of galaxies more massive than M=5e10Mo at z~1, with almost equal contributions from dusty and early types. Similarly, we find that EROs contribute about half of the mass density of the Universe at z~1 after taking into account incompleteness because of the limit K=19.5mag.

We present optical and X-ray identifications for the sixty-four radio sources in the GOODS-S ACS field revealed in the ATCA 1.4 GHz survey of the Chandra Deep Field South. Optical identifications are made using the ACS images and catalogs, while the X-ray view is provided by the Chandra X-ray Observatory 1 Ms observations. Redshifts for the identified sources are drawn from publicly available catalogs of spectroscopic observations and multi-band photometric-based estimates. Using this multiwavelength information we provide a first characterization of the faint radio source population in this region. The sample contains a mixture of star-forming galaxies and active galactic nuclei, as identified by their X-ray properties and optical spectroscopy. A large number of morphologically disturbed galaxies is found, possibly related to the star-formation phenomena. In spite of the very deep optical data available in this field, seven of the sixty-four radio sources have no optical identification to z(850)~28 mag. Only one of these is identified in the X-rays.

The Phoenix Deep Survey is a multi-wavelength survey based on deep 1.4 GHz radio imaging, reaching well into the sub-100 microJy level. One of the aims of this survey is to characterize the sub-mJy radio population, exploring its nature and evolution. In this paper we present the catalog and results of the spectroscopic observations aimed at characterizing the optically ``bright'' (R<~ 21.5 mag) counterparts of faint radio sources. Out of 371 sources with redshift determination, 21% have absorption lines only, 11% show AGN signatures, 32% are star-forming galaxies, 34% show narrow emission lines that do not allow detailed spectral classification (due to poor signal-to-noise ratio and/or lack of diagnostic emission lines) and the remaining 2% are identified with stars. For the star-forming galaxies with a Balmer decrement measurement we find a median extinction of A(Ha)=1.9 mag, higher than that of optically selected samples. This is a result of the radio selection, which is not biased against dusty systems. Using the available spectroscopic information, we estimate the radio luminosity function of star-forming galaxies in two independent redshift bins at z~0.1 and 0.3 respectively. We find direct evidence for strong luminosity evolution of these systems consistent with L(1.4 GHz) ~ (1+z)^(2.7).

In this paper we explore the clustering properties and the environment of the Extremely Red Objects (EROs; I-K>4mag) detected in a ~180arcmin^2 deep (Ks~20mag) Ks-band survey of a region within the Phoenix Deep Survey, an on-going multiwavelength program aiming to investigate the nature and the evolution of faint radio sources. Using our complete sample of 289 EROs brighter than Ks=20mag we estimate a statistically significant (~3.7sigma) angular correlation function signal with amplitude Aw=8.7^{+2.1}_{-1.7}x10^{-3} consistent with earlier work based on smaller samples. This amplitude suggests a clustering length in the range ro=12-17h^{-1}Mpc, implying that EROs trace regions of enhanced density. Using a novel method we further explore the association of EROs with galaxy overdensities by smoothing the K-band galaxy distribution using the matched filter algorithm of Postman et al. (1996) and then cross-correlating the resulting density maps with the ERO positions. Our analysis provides direct evidence that EROs are associated with overdensities at redshifts z>1. We also exploit the deep radio 1.4GHz data (limiting flux 60microJy) available to explore the association of EROs and faint radio sources and whether the two populations trace similar large scale structures. Cross-correlation of the two samples (after excluding 17EROs with radio counterparts) gives a 2sigma signal only for the sub-sample of high-z radio sources (z>0.6). Although the statistics are poor this suggests that it is the high-z radio sub-sample that traces similar structures with EROs.

The Phoenix Deep Survey is a multi-wavelength galaxy survey based on deep 1.4 GHz radio imaging (Hopkins et al., 2003). The primary goal of this survey is to investigate the properties of star formation in galaxies and to trace the evolution in those properties to a redshift z=1, covering a significant fraction of the age of the Universe. By compiling a sample of star-forming galaxies based on selection at radio wavelengths we eliminate possible biases due to dust obscuration, a significant issue when selecting objects at optical and ultraviolet wavelengths. In this paper, we present the catalogs and results of deep optical (UBVRI) and near-infrared (Ks) imaging of the deepest region of the existing decimetric radio imaging. The observations and data-processing are summarised and the construction of the optical source catalogs described, together with the details of the identification of candidate optical counterparts to the radio catalogs. Based on our UBVRIKs imaging, photometric redshift estimates for the optical counterparts to the radio detections are explored.

The radio properties of hard (2-8keV) X-ray selected sources are explored by combining a single 50ks XMM-Newton pointing with the ultra-deep and homogeneous Phoenix radio (1.4GHz) survey (Hopkins et al. 2003). A total of 43 sources are detected above the X-ray flux limit f_X(2-8keV)=7.7e-15cgs with 14 of them exhibiting radio emission above ~40muJy (3sigma). The X-ray/radio matched population lies in the borderline between radio loud and quiet AGNs and comprises sources with both soft and hard X-ray spectral properties suggesting both obscured and unobscured systems. The spectroscopically identified sub-sample (total of 6 X-ray/radio matches) comprises narrow emission line AGNs (4) with hard X-ray spectral properties and broad line sources (2) with soft X-ray spectra. We find evidence that the fraction of X-ray/radio matches increases from ~20% for sources with rest-frame column density N_H<1e22cm^-2 to \~50% for more absorbed systems. Poor statistics however, limit the significance of the above result to the ~2sigma level. Also, the X-ray/radio matched sources have flatter coadded X-ray spectrum (Gamma=1.78^{+0.05}_{-0.03}) compared to sources without radio emission (Gamma=2.00^{+0.03}_{-0.04}). A possible explanation for the higher fraction of absorbed sources with radio emission at the muJy level is the presence of circum-nuclear starburst activity that both feeds and obscures the central engine. For a small sub-sample of z~0.4 radio emitting AGNs with N_H>1e22cm^-2 their combined spectrum exhibits a soft X-ray component that may be associated with star-formation activity, although other possibilities cannot be excluded. We also find that radio emitting AGNs make up about 13-20% of the hard-band X-ray background depending on the adopted normalisation.

Insensitive to dust obscuration, radio wavelengths are ideal to study star-forming galaxies free of dust induced biases. Using data from the Phoenix Deep Survey, we have identified a sample of star-forming extremely red objects (EROs). Stacking of the radio images of the radio-undetected star-forming EROs revealed a significant radio detection. Using the expected median redshift, we estimate an average star-formation rate of 61 M_sun/yr for these galaxies.

The Phoenix Deep Survey (PDS) is a multiwavelength survey based on deep 1.4 GHz radio observations used to identify a large sample of star forming galaxies to z=1. Photometric redshifts are estimated for the optical counterparts to the radio-detected galaxies, and their uncertainties quantified by comparison with spectroscopic redshift measurements. The photometric redshift estimates and associated best-fitting spectral energy distributions are used in a stacking analysis exploring the mean radio properties of U-band selected galaxies. Average flux densities of a few microJy are measured.

The Australia Telescope Compact Array (ATCA) has been used to survey at 1.4GHz, a small region (< 3 sq deg) overlapping with the 2dF Galaxy Redshift Survey (Colless et al. 2001, MNRAS, 328, 1039). We surveyed with a varying radio sensitivity, ranging from 1mJy - 20microJy (1 sigma). There are 365 2dFGRS sources with z > 0.001 lying within the surveyed region, of which 316 have reliable spectral classification. Following Sadler et al. 2002 (MNRAS, 329, 227), we visually classified 176 as AGN or early-type galaxies, and 140 as star-forming galaxies. We derived radio flux density measurement or upperlimits for each of the 365 2dFGRS sources. The fraction of radio detected 2dFGRS star-forming galaxies increases from ~ 50% at ~ 0.7mJy up to ~ 60% at ~ 0.2mJy. The mean redshift for the fraction of radio detected star-forming galaxies increases with increasing radio detection sensitivity, while the mean redshift is fairly constant for the AGN/early-type fraction. We found very similar radio detection rates of 2dFGRS galaxies for both the AGN/early-type and star-forming components. The radio detection rate increases approximately linearly with respect to the rate of increase in radio detection sensitivity. We derived the radio luminosity function for our sample and it was found to be consistent with that of Sadler et al. 2002. We have also compared the total flux densities of NVSS sources common to our survey, and we discuss strategies for a large-scale radio survey of the 2dFGRS sample.

The Phoenix Deep Survey (PDS) is a multiwavelength survey based on deep 1.4 GHz radio observations used to identify a large sample of star forming galaxies to z=1. Here we present an exploration of the evolutionary constraints on the star-forming population imposed by the 1.4 GHz source counts, followed by an analysis of the average properties of extremely red galaxies in the PDS, by using the "stacking" technique.

In this paper we use a 50ks XMM-Newton pointing overlapping with the Phoenix Deep Survey, a homogeneous radio survey reaching muJy sensitivities, to explore the X-ray properties and the evolution of star-forming galaxies. UV, optical and NIR photometry is available and is used to estimate photometric redshifts and spectral types for radio sources brighter than R=21.5mag (total of 82). Sources with R<21.5mag and spiral galaxy SEDs (34) are grouped into two redshift bins with a median of z=0.240 and 0.455 respectively. Stacking analysis for both the 0.5-2 and 2-8keV bands is performed on these subsamples. A high confidence level signal (>3.5sigma) is detected in the 0.5-2keV band corresponding to a mean flux of ~3e-16cgs for both subsamples. This flux translates to mean luminosities of ~5e40 and 1.5e41cgs for the z=0.240 and 0.455 subsamples respectively. Only a marginally significant signal (2.6sigma) is detected in the 2-8keV band for the z=0.455 subsample. We argue that the stacked signal above is dominated by star-formation. The mean L_X/L_B ratio and the mean L_X of the two subsamples are found to be higher than optically selected spirals and similar to starbursts. We also find that the mean L_X and L_1.4 of the faint radio sources studied here are consistent with the L_X-L_1.4 correlation of local star-forming galaxies. Moreover, the X-ray emissivity of sub-mJy sources to z~0.3 is found to be elevated compared to local HII galaxies. The observed increase is consistent with L_X evolution of the form (1+z)^3. Assuming that our sample is indeed dominated by starbursts this is direct evidence for evolution of such systems at X-ray wavelengths. Using an empirical L_X to SFR conversion we estimate a global SFR density at z~0.3 of \~0.029M_o/yr/Mpc in agreement with previous studies.

Archetypal analysis represents each individual member of a set of data vectors as a mixture (a constrained linear combination) of the pure types or archetypes of the data set. The archetypes are themselves required to be mixtures of the data vectors. Archetypal analysis may be particularly useful in analysing data sets comprising galaxy spectra, since each spectrum is, presumably, a superposition of the emission from the various stellar populations, nebular emissions and nuclear activity making up that galaxy, and each of these emission sources corresponds to a potential archetype of the entire data set. We demonstrate archetypal analysis using sets of composite synthetic galaxy spectra, showing that the method promises to be an effective and efficient way to classify spectra. We show that archetypal analysis is robust in the presence of various types of noise.

The initial Phoenix Deep Survey (PDS) observations with the Australia Telescope Compact Array have been supplemented by additional 1.4 GHz observations over the past few years. Here we present details of the construction of a new mosaic image covering an area of 4.56 square degrees, an investigation of the reliability of the source measurements, and the 1.4 GHz source counts for the compiled radio catalogue. The mosaic achieves a 1-sigma rms noise of 12 microJy at its most sensitive, and a homogeneous radio-selected catalogue of over 2000 sources reaching flux densities as faint as 60 microJy has been compiled. The source parameter measurements are found to be consistent with the expected uncertainties from the image noise levels and the Gaussian source fitting procedure. A radio-selected sample avoids the complications of obscuration associated with optically-selected samples, and by utilising complementary PDS observations including multicolour optical, near-infrared and spectroscopic data, this radio catalogue will be used in a detailed investigation of the evolution in star-formation spanning the redshift range 0 < z < 1. The homogeneity of the catalogue ensures a consistent picture of galaxy evolution can be developed over the full cosmologically significant redshift range of interest. The 1.4 GHz mosaic image and the source catalogue are available on the web at http://www.atnf.csiro.au/~ahopkins/phoenix/ or from the authors by request.

In the course of the follow-up multiwavelength study of a deep radio survey we have discovered that the milli-Jansky radio source PDFJ011423 is a low-redshift (z = 0.65) extremely red galaxy (ERG) with K=15.3, R-K = 5.8 and J-K = 3.1. Optical, infrared and radio photometry, together with optical and near-infrared spectroscopy, reveal a heavily obscured galaxy (A_V=5-6, from the observed Balmer decrement) undergoing vigorous star formation and presenting an active galactic nucleus (AGN). PDFJ011423 is a representative member of the dusty ERG population, providing a local counterpart for studying more distant ERGs.

We present results from a decimetric radio survey undertaken with the Very Large Array as part of a longer term goal to inter-compare star formation and dust extinction diagnostics, on a galaxy by galaxy basis, for a representative sample of nearby galaxies. For our survey field, Selected Area 57, star formation rates derived from 1.4GHz luminosities are compared with earlier nebular emission line and ultraviolet (UV) continuum diagnostics. We find broad correlations, over several decades in luminosity, between H-alpha, the UV continuum and 1.4GHz diagnostics. However, the scatter in these relations is found to be larger than observational errors, with offsets between the observed relations and those expected assuming constant star-formation histories and luminosity-independent extinction models. We investigate the physical origin of the observed relations, and conclude the discrepancies between different star-formation diagnostics can only be partly explained by simple models of dust extinction in galaxies. These models cannot by themselves explain all the observed differences, introducing the need for temporally varying star-formation histories and/or more complex models of extinction, to explain the entire dataset.

Different wavelength regimes and methods for estimating the space density of the star-formation rate (SFR) result in discrepant values. While it is recognised that ultra-violet (UV) and H-alpha emission line data must be corrected for the effects of extinction, the magnitude of the required correction is uncertain. Even when these corrections are made there remains a significant discrepancy between SFRs derived from UV and H-alpha measurements compared with those derived from far-infrared (FIR) and radio luminosities. Since the FIR/radio derived SFRs are not affected by extinction, and simple corrections to reconcile the UV and H-alpha measurement with these do not fully account for the discrepancies, a more sophisticated correction may be required. Recent results suggest that at least part of the solution may be a form of extinction which increases with increasing SFR (or luminosity, given the common assumption that SFR is proportional to luminosity). We present an analysis of the effects of a dust reddening dependent on star formation rate applied to estimators of SFR. We show (1) that the discrepancies between H-alpha and FIR/radio SFR estimates may be explained by such an effect, and we present an iterative method for applying the correction; and (2) UV-based estimates of SFR are harder to reconcile with FIR/radio estimates using this method, although the extent of the remaining discrepancy is less than for a non-SFR-dependent correction. Particularly at high redshift, our understanding of extinction at UV wavelengths may require a still more complex explanation.

The Infrared Space Observatory (ISO) is used to carry out mid-IR (7 and 15 micron) and far-IR (90 micron) observations of a sample of star-forming sub-mJy radio sources. By selecting the sample at radio wavelengths, one avoids biases due to dust obscuration. It is found that the mid-IR luminosities, covering the PAH features, measure the star formation rate for galaxies with $P_{1.4 GHz} < 10^{23}$ W Hz$^{-1}$. This is further confirmed using the H$\alpha$ luminosities. The far-IR emission is also found to trace the SFR over the whole range of radio and H$\alpha$ luminosities. The implication of the mid-IR measurements in estimating the SFRs from the future infrared space missions (SIRTF and ASTRO-F) is discussed.

The cluster DC0107-46 (Abell 2877) lies within the Phoenix Deep Survey, made at 1.4 GHz with the Australia Telescope Compact Array. Of 89 known optical cluster members, 70 lie within the radio survey area. Of these 70 galaxies, 15 (21%) are detected, with luminosities as faint as 10^20 W/Hz. Spectroscopic observations are available for 14/15 of the radio-detected cluster galaxies. Six galaxies show only absorption features and are typical low-luminosity AGN radio sources. One galaxy hosts a Seyfert 2 nucleus, two are star-forming galaxies, and the remaining five may be star-forming galaxies, AGNs, or both.

The redshift dependence of the luminosity density in certain wavebands (e.g. UV and H-alpha) can be used to infer the history of star formation in the populations of galaxies producing this luminosity. This history is a useful datum in studies of galaxy evolution. It is therefore important to understand the errors that attend the inference of star formation rate densities from luminosity densities. This paper explores the self-consistency of star formation rate diagnostics by reproducing commonly used observational procedures in a model with known galaxy populations, evolutionary histories and spectral emission properties. The study reveals a number of potential sources of error in the diagnostic processes arising from the differential evolution of different galaxy types. We argue that multi-wavelength observations can help to reduce these errors.

An international project is underway to design and build a radio telescope with an effective collecting area two orders of magnitude greater than the largest existing instruments. One of the many scientific goals of this instrument will be the investigation of the extragalactic radio source population at flux densities two to three orders of magnitude fainter than the limits of existing observations. We present simulations of the radio sky at 1.4 GHz down to a flux density limit of 0.1 microJy using extrapolations of known radio luminosity functions for two different population scenarios. The resulting simulations confirm that a resolution of 0.1" is necessary to avoid formal confusion, but source blending may still dominate if the intrinsic size of such faint sources is larger than a few kiloparsecs.

Using the Phoenix radio survey, a homogeneous survey selected at 1.4GHz and covering an area of ~3deg^2, we analyse the clustering of the sub-mJy radio population using angular correlation function analysis. Extensive simulations are carried out to investigate the significance of the estimated angular correlation amplitudes. Our analysis show that for the S_1.4>0.5mJy sub-samples the radio source distribution is anisotropic at the 2$\sigma$ significance level. Additionally, we estimate upper limits for the angular correlation amplitudes that, despite the large uncertainties, are in good agreement with the amplitude estimates for sources brighter than 1mJy, detected in the FIRST radio survey (Cress {et al.} 1997). Adopting a radio luminosity function and assuming an evolving spatial correlation function of the form \xi(r)=(r/r_{o})^{-\gamma} (1+z)^{-(3+\epsilon)}, with the evolution parametrised by \epsilon, we find an upper limit for the angular correlation length r_{o}~9h^-1Mpc for S_1.4>0.5mJy and \gamma=2.1. This agrees well with the value r_{o}~6-8h^-1Mpc estimated from the FIRST radio survey for sources brighter than 1mJy. Additionally, we quantify the characteristics, in terms of areal coverage and limiting flux density, of future deep radio surveys to yield a significant correlation amplitude detection and to explore possible changes of the correlation amplitude with flux density.

The significance of tidal interactions in the evolution of the faint radio population (sub-mJy) is studied using a deep and homogeneous radio survey (1.4 GHz), covering an area of 3.14 deg$^2$ and complete to a flux density of 0.4 mJy. Optical photometric and spectroscopic data are also available for this sample. A statistical approach is employed to identify candidate physical associations between radio sources and optically selected `field' galaxies. We find an excess of close pairs around optically identified faint radio sources, albeit at a low significance level, implying that the pairing fraction of the sub-mJy radio sources is similar to that of `field' galaxies (at the same magnitude limit) but higher than that of local galaxies.

The study of the nature of faint radio sources is of great importance since a significant fraction of these objects is thought to be composed of actively star-forming galaxies. Due to the increased sensitivity of radio telescopes, we are now not only able to catalogue large numbers of these sources in the sub-millijansky regime, but also to start the study of the nature of increasingly fainter microjansky sources. This paper presents a new very deep 1.4 GHz radio survey made as a part of the Phoenix Deep Survey, a project aimed to study the nature of the faintest radio sources. With a limiting sensitivity of 45 microJy, this new survey has allowed us to assemble a large number of sources with 1.4 GHz flux densities below 100 microJy. The resulting source counts and the analysis of the optical properties of the faintest radio sources are presented.

We present a deep 1.4 GHz survey made with the Australia Telescope Compact Array (ATCA), having a background RMS of 9 microJy near the image phase centre, up to 25 microJy at the edge of a 50' field of view. Over 770 radio sources brighter than 45 microJy have been catalogued in the field. The differential source counts in the deep field provide tentative support for the growing evidence that the microjansky radio population exhibits significantly higher clustering than found at higher flux density cutoffs. The optical identification rate on CCD images is approximately 50% to R=22.5, and the optical counterparts of the faintest radio sources appear to be mainly single galaxies close to this optical magnitude limit.

Using a recently completed survey of faint (sub-mJy) radio sources, selected at 1.4 GHz, a dust-free estimate of the local star formation rate (SFR) is carried out. The sample is 50% complete to 0.2 mJy, with over 50% of the radio sources having optical counterparts brighter than R = 21.5. Spectroscopic observations of 249 optically identified radio sources have been made, using the 2-degree Field (2dF) facility at the Anglo-Australian Telescope (AAT). Redshifts and equivalent widths of several spectral features (e.g., H\alpha and [OII]3727) sensitive to star formation have been measured and used to identify the star-forming and absorption-line systems. The spectroscopic sample is corrected for incompleteness and used to estimate the 1.4 GHz and H\alpha luminosity functions (LFs) and luminosity density distributions. The 1.4 GHz LF of the star-forming population has a much steeper faint-end slope (1.85) than that for the ellipticals (1.35). This implies an increasing preponderance of star-forming galaxies among the optically identified (i.e., z < 1) radio sources at fainter flux densities. The H\alpha LF of the faint radio population agrees with published H\alpha LFs derived from local samples selected by H\alpha emission. This suggests that the star-forming faint radio population is coincident with the H\alpha selected galaxies. The 1.4 GHz and H\alpha luminosity densities have been used to estimate the SFRs. The two estimates agree, both giving a SFR density of $0.032 M_\odot yr^{-1} Mpc^{-3}$ in the range z < 1.

Using a deep Australia Telescope Compact Array (ATCA) radio survey covering an area of ~3deg^{2} to a 4\sigma sensitivity of \ge 100 \muJy at 1.4GHz, we study the nature of faint radio galaxies. About 50% of the detected radio sources are identified with an optical counterpart revealed by CCD photometry to m_{R}=22.5 mag. Near-infrared (K-band) data are also available for a selected sample of the radio sources, while spectroscopic observations have been carried out for about 40% of the optically identified sample. These provide redshifts and information on the stellar content. Emission-line ratios imply that most of the emission line sources are star-forming galaxies, with a small contribution (\approx 10%) from Sy1/Sy2 type objects. We also find a significant number of absorption line systems, likely to be ellipticals. These dominate at high flux densities (> 1 mJy) but are also found at sub-mJy levels. Using the Balmer decrement we find a visual extinction A_{V}=1.0 for the star-forming faint radio sources. This moderate reddening is consistent with the V-R and R-K colours of the optically identified sources. For emission line galaxies, there is a correlation between the radio power and the H\alpha luminosity, in agreement with the result of Benn et al. (1993). This suggests that the radio emission of starburst radio galaxies is a good indicator of star-formation activity.

A deep survey of the European Large Area ISO Survey (ELAIS) field in the southern celestial hemisphere (hereinafter S1) has been carried out with the Australia Telescope Compact Array (ATCA) at 1.4 GHz. The S1 region, covering about 4 square degrees, has been also surveyed in the mid- and far-infrared (5-200 micron) with the Infrared Space Observatory (ISO). The radio survey provides uniform coverage of the entire S1 region, with a sensitivity (5 sigma) of 0.4 mJy over the whole area and 0.2 mJy in the center. To this sensitivity, virtually all the radio counterparts of the far-infrared extragalactic ISO sources should be detected. This paper presents a radio sample - complete at the 5 sigma level - consisting of 581 sources detected at 1.4 GHz. Of these radio sources, 349 have peak flux density in the range 0.2-1 mJy, forming a new homogeneous sample of sub-mJy radio sources. Due to its size, depth and multi-waveband coverage, the sample will allow us to study in greater detail the sub-mJy radio source population. The full catalogue will be available from http://athena.ph.ac.uk/

The first epoch Molonglo Galactic Plane Survey (MGPS1) is a radio continuum survey made using the Molonglo Observatory Synthesis Telescope (MOST) at 843 MHz with a resolution of 43" X 43" cosec |delta|. The region surveyed is 245 deg < l < 355 deg, |b| < 1.5 deg. The thirteen 9 deg X 3 deg mosaic images presented here are the superposition of over 450 complete synthesis observations, each taking 12 h and covering 70' X 70' cosec |delta|. The root-mean-square sensitivity over much of the mosaiced survey is 1-2 mJy/beam (1 sigma), and the positional accuracy is approximately 1" X 1" cosec |delta| for sources brighter than 20 mJy. The dynamic range is no better than 250:1, and this also constrains the sensitivity in some parts of the images. The survey area of 330 sq deg contains well over 12,000 unresolved or barely resolved objects, almost all of which are extra-galactic sources lying in the Zone of Avoidance. In addition a significant fraction of this area is covered by extended, diffuse emission associated with thermal complexes, discrete H II regions, supernova remnants, and other structures in the Galactic interstellar medium.

The decimetric luminosity of many galaxies appears to be dominated by synchrotron emission excited by supernova explosions. Simple models suggest that the luminosity is directly proportional to the rate of supernova explosions of massive stars averaged over the past 30 Myr. The proportionality may be used together with models of the evolving 1.4 GHz luminosity function to estimate the global star formation rate density in the era z < 1. The local value is estimated to be 0.026 solar masses per year per cubic megaparsec, some 50% larger than the value inferred from the Halpha luminosity density. The value at z ~ 1 is found to be 0.30 solar masses per year per cubic megaparsec. The 10-fold increase in star formation rate density is consistent with the increase inferred from mm-wave, far-infrared, ultra-violet and Halpha observations.

The decimetric radio continuum luminosity of a star-forming galaxy appears to be directly proportional to the rate of formation of supernovae in the galaxy. Since decimetric radiation does not suffer significant extinction and is not directive, radio luminosities may thus provide a particularly straightforward way to determine the current rate of star formation. Using a sample of over 700 local galaxies we confirm the utility of the radio luminosity as a measure of star formation rate by showing concordance with the rates predicted by U-band, H-alpha, and far-infrared luminosites. We also show that there are systematic discrepancies between these various indicators, suggesting that the H-alpha luminosity may underestimate the star formation rate by approximately an order of magnitude when the star formation rate is more than 20 solar mass per year. We use this calibration and the measured radio luminosities of sub-mJy radio sources to infer the star formation rate in approximately 60 star-forming galaxies at moderate (z = 0.1) redshifts, both as the actual rate and as the fraction of the existing mass of stars in the galaxy. For some of these objects the inferred current rate of star formation could increase the stellar mass in the galaxy by approximately 10% over an interval of about 30 Myr.