Many studies of phylogeography, speciation, and species limits restrict their focus to a narrow issue: gene tree monophyly. However, reciprocal monophyly does not provide an ideal touchstone criterion of any aspect of evolutionary divergence. There is a continuum of divergence stages as isolated populations go from initial allele frequency differences to well-differentiated species. Studying intermediate stages of divergence will increase our understanding of geographical speciation, species limits, and conservation priorities. We develop a conceptual framework and terminology for thinking about the stages of 'intermediate polyphyly'. The Holarctic clade of common ravens (Corvus corax), found throughout much of Eurasia and North America, provides a case study of these stages of intermediate divergence. We used coalescent, phylogenetic, and population genetic methods to investigate the history and current status of this Old World-New World distribution using 107 mitochondrial control region sequences. Phylogenetically, New World and Old World samples are intermixed. However, most samples are grouped into small subclades that are restricted to either the New World or the Old World, and only one haplotype is shared between the hemispheres. Analysis of moleculalr variance (amova) results reflect this low haplotype sharing between hemispheres (Phi(ST) = 0.13, P < 0.01). Isolation with Migration (im) coalescent results suggest a sustained period of divergence between the hemispheres and low levels of maternal gene flow. Although there has not been sufficient time to evolve reciprocal monophyly and some gene flow may occur, New World and Old World ravens are genetically quite distinct. We use this example to demonstrate these early stages of divergence as populations go from sharing only internal haplotypes, to sharing no haplotypes, to having population specific subclades. Studies of phylogeography, speciation and systematics will benefit from increased attention to these stages of intermediate polyphyly.