Passerine birds comprise over half of avian diversity, but have proved difficult to classify. Despite a long history of work on this group, no comprehensive hypothesis of passerine family–level relationships was available until recent analyses of DNA–DNA hybridization data. Unfortunately, given the value of such a hypothesis in comparative studies of passerine ecology and behaviour, the DNA–hybridization results have not been well tested using independent data and analytical approaches. Therefore, we analysed nucleotide sequence variation at the nuclear RAG–1 and c–mos genes from 69 passerine taxa, including representatives of most currently recognized families. In contradiction to previous DNA–hybridization studies, our analyses suggest paraphyly of suboscine passerines because the suboscine New Zealand wren Acanthisitta was found to be sister to all other passerines. Additionally, we reconstructed the parvorder Corvida as a basal paraphyletic grade within the oscine passerines. Finally, we found strong evidence that several family–level taxa are misplaced in the hybridization results, including the Alaudidae, Irenidae, and Melanocharitidae. The hypothesis of relationships we present here suggests that the oscine passerines arose on the Australian continental plate while it was isolated by oceanic barriers and that a major northern radiation of oscines (i.e. the parvorder Passerida) originated subsequent to dispersal from the south.