Conclusions

 

A large multiwavelength survey has been initiated, consisting of an extensive (two degree diameter), deep radio survey and optical observations. Additional observations of selected sources in the near-infrared have been made, and spectroscopy has established redshifts for about one quarter of the optical counterparts of radio sources. A catalogue of 938 radio sources brighter than 0.3 mJy and 232 brighter than 0.1 mJy has been compiled. Of the 1079 radio sources detected in total, 541 have been identified optically.

It has been established through 2dF spectroscopy that a significant proportion of the PDS radio sources are identified with emission line galaxies, which are most likely to be exhibiting starburst activity. The implication is that deep radio surveys are certainly useful tools for the detection of these galaxies at moderate redshifts. Until more redshifts for the sources with fainter optical counterparts are measured, it is difficult to comment on whether this provides a good cosmological probe of the starburst phenomenon. However if the redshift distributions predicted from the model bivariate luminosity function are accurate, it would be expected that this is indeed the case.

Evolution of these phenomena can be investigated through less direct means by calling upon models of optical and radio luminosity functions. It has been established that to reproduce the observed source counts to a limiting flux density of 0.1mJy in the radio, strong evolution of the starburst galaxy population is required, at a rate (1+z)^Q where $Q=3.3\pm0.8$, in addition to the already well established need for evolution of the AGN population. For the optical galaxies, however, to a limiting magnitude of R=22, no evolution of the luminosity function needs to be invoked to reproduce the observed source counts.

The model BLF constructed in Chapter 4 has been used to make predictions of the BSC distribution, and redshift distributions. These compare qualitatively well with the observations, although it is clear that more work is required to produce a more accurate model. Establishing these refinements to the BLF model will form part of the continuing research on the PDS.

Through an analysis of the spectral observations, and in the absence of spectroscopy for a complete sample, it can be said that it appears the sub-mJy radio population is comprised of an increasing fraction of starburst galaxies compared to the AGN sources. This agrees qualitatively with the distributions expected from the source count predictions of the luminosity evolution model of the radio luminosity functions. The model source counts predict the same surface density for the starburst and AGN populations at about 0.2mJy, with the AGN being dominant at higher flux densities, and starbursts at lower. A quantitative comparison of this figure with observations will require a complete sample of redshifts. Another interesting observation is the small number of galaxies which are difficult to distinguish as either starbursts or AGN from emission line diagnostics. A few of the emission line galaxies would be classified differently from different diagnostic diagrams, and several more often appear on the border separating the distinct regions. It is possible that these objects are exhibiting both AGN and starburst characteristics, and this claim should be investigated more closely.

The Phoenix Multiwavelength Deep Survey is an ongoing project, which has obtained additional observational data since the drafting of this thesis. At the time of writing, a further set of spectroscopic observations with the 2dF have been made, as have CASPIR observations of additional sub-mJy radio sources. An additional $\sim$150 hours of observations using the ATCA have been made to extend the radio survey to fainter flux densities than have ever before been observed at 1.4GHz, with an anticipated $5\sigma$ flux density limit of $60\,\mu$Jy over a one degree diameter region within the PDF. The results established through the work contained in this thesis are an exciting first step in the analysis of PDS observations, and further observations and research will continue to expand our understanding of galaxy evolution, and the sub-millijansky and microjansky radio source population.