Most of what is known about murine (and human) hematopoietic stem cells (HSCs) comes from studies of fetal liver and adult bone marrow. More recently, there has been a burst of research activity aimed at defining how the hematopoietic system is established during early embryogenesis. Defining the ontogeny of HSCs should improve our ability to isolate and manipulate them for therapeutic purposes, and it has important implications for the study of basic developmental biology. Unraveling the origins of vertebrate hematopoiesis has been a thorny problem, largely because the first committed progenitors to arise in the embryo are rare, transient, and difficult to access. Little is known about the cytokines, stromal interactions, and cell-intrinsic genetic programs that regulate the onset of hematopoiesis and the genesis of HSCs. Solving these problems requires new approaches to complement studies of whole mouse embryos. An ideal model system would permit examination and manipulation of hematopoietic induction outside the uterus, facilitate isolation of the earliest committed hematopoietic cells, and reproduce the embryonic microenvironment. Murine embryonic stem (ES) cells potentially fulfill these requirements. In vitro differentiation of ES cells recapitulates early hematopoietic ontogeny in a tissue culture dish. Moreover, ES cells form the basis of an experimental strategy to create targeted ablations, or “knockouts,” of genes in mice. These loss-of-function experiments have identified several genes that act at specific stages of HSC development.