Dendritic cells (DC) play key rolls in various aspects of immunity. The functions of DC depend on the subsets as well as their location or activation status. Understanding developmental lineages, precursors and inducing factors for various DC subsets would help their clinical application, but despite extensive efforts, the precise ontogeny of various DC, remain unclear and complex. Because of their many functional similarities to macrophages, DC were originally thought to be of myeloid-lineage, an idea supported by many in vitro studies where monocytes or GM-CSF (a key myeloid growth factor) has been extensively used for generating DC. However, there has been considerable evidence which suggests the existence of lymphoid-lineage DC. After the confusion of myeloid-/lymphoid-DC concept regarding DC surface markers, we have now reached a consensus that each DC subset can differentiate through both myeloid- and lymphoid-lineages. The identification of committed populations (such as common myeloid- and lymphoid progenitors) as precursors for every DC subsets and findings from various knockout (KO) mice that have selected lymphoid- or myeloid-lineage deficiency appear to indicate flexibility of DC development rather than their lineage restriction. Why is DC development so flexible unlike other hematopoitic cells? It might be because there is developmental redundancy to maintain such important populations in any occasions, or such developmental flexibility would be advantageous for DC to be able to differentiate from any "available" precursors in situ irrespective of their lineages. This review will cover ontogeny of conventional (CD8 +/- DC) DC, plasmacytoid DC and skin Langerhans cells, and recently-identified many Pre-DC (immediate DC precursor) populations, in addition to monocytes and plasmacytoid DC, will also be discussed.