The nature of faint radio sources in low frequency samples
The behaviour of the radio source population at high redshift has been a major area of research in extragalactic radio astronomy during the past decade. In particular, the question of at what redshift the co-moving number density of radio sources reaches a peak, has been of considerable interest. Radio surveys are particularly good for yielding large numbers of distant radio galaxies and quasars but at the price of having the edshift of the individual sources very difficult to obtain. This thesis has examined the behaviour of faint radio sources at large redshift, in particular those objects showing the classical double structure as is commonly found in samples selected at low frequency. The first part of this thesis identifies the need for at least approximate redshift information for complete flux limited samples to properly address the behaviour of these sources at large redshift; known redshift indicators are shown to be inadequate for this purpose. A new method of obtaining the intrinsic radio luminosity and hence redshift for classical double radio sources from the bright 3C sample from the radio data alone, is introduced. The same relationship is shown to be applicable to similar sources found at larger redshifts when selected from fainter samples. The middle part of this work is concerned with the construction of two new faint radio samples selected at 151 MHz. In addition to VLA and 5km telescope mapping, a review of existing optical data together with new CCD imaging and some spectroscopy is given. Finally, by taking these two new samples together with the intermediate sample of Eales and the bright sample of LRL, the new method of deriving radio luminosity has been used to explore the shape of the evolution function and degree of linear size evolution required for double radio sources at a much larger redshift than has been previously possible.