There is much interest in early T-cell development, particularly in relation to the diversification of the T-cell receptor repertoire^1,2 and the elucidation of the lineage relationships between T-cell populations in the thymus and peripheral lymphoid organs (reviewed in refs 3, 4). However, the requirements for the growth of the earliest thymic T-cell precursor in 13–14-day mouse embryo thymus in isolation from the thymic environment are unknown. Proliferation and maturation of such cells are not sustained either in the presence of monolayers of thymic stromal cells (refs. 5, 6 and see below) or by the addition of interleukin-2 (IL-2), despite the expression of receptors for this growth factor on a proportion of thymocytes displaying the immature Thy 1⁺ Lyt-2⁻L3T4⁻ phenotype in the embryonic thymus^7,8. In contrast, when maintained within the intact thymic environment in organ cultures^9,10, 13–14-day thymic stem cells do show a pattern of surface marker and functional development similar to that seen in vivo, suggesting that short-range growth signals, perhaps necessitating direct contact with organized epithelial cells, are required. We have shown, by exploiting the selective toxicity of deoxyguanosine (dGuo) for early T cells, that this organ culture system can be manipulated to produce alymphoid lobes that can be recolonized from a source of precursors in a transfilter system¹¹. We now show that recolonization of alymphoid lobes can also be achieved by association with T-cell precursors in hanging drops, allowing recolonization by exposure to defined numbers of precursors, including a single micromanipulated stem cell. Analysis of T-cell marker expression in these cultures shows that a single thymic stem cell can produce progeny of distinct phenotypes, suggesting that these marker-defined populations are not derived from separate pre-thymic precursors, but arise within the thymus.