The Redshift Cut-Off in the Luminosity Function of Radio Galaxies and Quasars

We describe a new investigation of the evolution of the radio luminosity function for extragalactic radio sources, with particular emphasis on the behaviour at high redshifts. Our analysis contrasts free-form and simple parametric modelling with model-independent techniques. Using data from a new complete sample at 2.7 GHz (the S > 100mJy Parkes selected regions) in addition to previously studied data sets, we are able to reach several important conclusions, as follows. (i) We strengthen previous conclusions that a 'redshift cut-off' exists in the distribution of flat-spectrum quasars, in the sense of a reduction in comoving density by a factor ~5 over the redshift range 2-4. (ii) We present the first clear evidence that similar behaviour also appears to apply to steep-spectrum quasars and radio galaxies, making this decline a universal feature for all high-luminosity radio sources. However, the 'cut-off' does appear to be more abrupt for the flat-spectrum population. These statements still depend on the accuracy of our redshift estimates for faint galaxies, and we identify future observations which can remove the remaining uncertainty. (iii) By allowing for a slowly evolving population at low luminosities, we are able to fit all existing redshift and source-count data with a model whereby luminous sources undergo pure luminosity evolution (but this is only possible for {OMEGA}_0_ ~ 1). We have also successfully constructed a luminosity/density evolution model, with continuing positive luminosity evolution plus a depression in comoving density at high redshift. Further improvements in the database are required to distinguish between these two alternative forms of high-redshift evolution.