Many events and requirements of the developmental program of human hematopoietic stem cells have not yet been discovered. A major impediment has been the lack of an appropriate experimental system. At present the conditions for maintaining human stem cells in vitro are not fully known. As a result within a short period the small stem cell pool is lost due to differentiation, making it difficult to examine the correlation between these cells and their function in vivo. Most of our knowledge of hematopoietic stem cells is from animal models in which purified stem cell canididates are assayed based on their functional ability to rescue lethally conditioned recipients. The permanent correction of many genetic disorders of the hematopoietic system requires efficient methods for introducing genes into stem cells in vitro. However, progress has been hindered by the absence of preclinical models that assay the repopulating capacity of primitive human cells. In addition, the development of therapy for malignant diseases also requires assays to identify the target leukemic stem cells based on their ability to initiate the disease. The recent development of methods to transplant or implant both normal and leukemic cells into immune-deficient mice provides the foundation for human stem cell assays. These models assay the repopulating capacity of primitive human cells and provide an important approach to identify and characterize human stem cells, both normal and leukemic. This review focuses on the development of functional assays for normal and leukemic human stem cells and on the new insights that these models are beginning to provide on the organization of the human stem cell hierarchy.