Long-term maintenance of normal hematopoiesis in vitro is possible when very primitive progenitors are cocultivated with certain non-hematopoietic stromal cells that may co-exist in (or be derived from cells that co-exist in) hematopoietic tissues. Such long-term cultures (LTC) have been used to develop quantitative assays for the most primitive populations of hematopoietic cells currently detectable in adult marrow. In addition they provide a unique model for analysis of the complex molecular mechanisms that may regulate primitive hematopoietic cell population dynamics in vivo. Similar studies with LTC of cells from patients with chronic myeloid leukemia (CML) have made it possible to detect and characterize very primitive neoplastic cell populations in this disease. These latter studies have revealed differences in the properties of primitive CML cells that both reflect and explain their increased turnover and are thus presumably part of the mechanism that enables the neoplastic clone to expand in vivo. In addition, the most primitive neoplastic cells in CML patients are abnormally distributed between the marrow and blood and their ability to maintain their numbers in LTC has also been found to be defective. Assessment of the number and behaviour of primitive cells in LTC of CML marrow has been used to identify those patients most likely to benefit from intensive therapy supported by transplantation of cultured autologous marrow. Twenty-two such CML patients have now been treated with this experimental protocol. The results to date have clearly established the feasibility of this novel treatment strategy and, together with more recent laboratory findings, suggest future avenues for significantly improving the management of CML patients.