The idea that there may be a cut-off in the distribution of quasars at high redshifts (z∼4) has been of some recent interest through the work of Osmer (1982). The observation of such an epoch of quasar creation is potentially of great importance in relation to theories of galaxy formation, but the evidence from optically-selected quasar samples remains uncertain: quite apart from the notorious problems in achieving quantifiable completeness in objective-prism surveys, any observed lack of high-redshift quasars may always be attributed to absorption either by a neutral IGM or by dust in intervening galaxies. Radio-selected samples, however, do not suffer from these problems, and this paper aims to review what studies of extragalactic radio sources can tell us about the numbers of objects at the highest redshifts.