The hallmark of a hematopoietic stem cell (HSC) is its capacity to sustain production of daughter cells with the same multi-lineage differentiation potential for the lifetime of the organism. Definitive evidence of HSC self-renewal requires the demonstration that daughter HSCs have been generated from the division of a single cell. This was first demonstrated by the detection of clones with the same proviral inserts in multiple secondary recipients of retrovirally-marked HSCs initially expanded in primary mice. Later, this approach showed that HSC self-renewal occurs in longterm marrow cultures. More recently, HSC self-renewal divisions were demonstrated by in vivo assessment of the progeny of highly purified (lin- Rhodamine-123- SP) HSCs stimulated to divide in single-cell cultures containing recombinant growth factors. We now describe the remarkable self-renewal activity that a proportion of HSCs can display in vivo when subjected to serial transplantation. In 7/25 (28%) sublethally irradiated B6-W41/W41-Ly-5.2 mice engrafted with a single lin- Rho- SP Ly5.1 bone marrow cell, the majority of the circulating WBCs (all lineages) 6 months later were Ly-5.1. 1.4-5.0% of the Ly-5.1 cells present in the marrow of 3 of these mice was then transplanted into 1 or 2 sublethally irradiated secondary B6-W41/W41-Ly-5.2 recipients. Six months later, at least one of these secondary recipients of cells from each of the 3 primary mice contained 1-79% Ly5.1 WBCs including all lineages. Ly5.1 cells from only one of the 3 clones regenerated in these secondary recipients repopulated tertiary mice, reflecting the heterogeneity of HSC self-renewal activity. Ly-5.1 marrow cells obtained 6 months post-transplant from the repopulated tertiary mice (1–4% of all the Ly-5.1 cells) were transplanted into 4th and 5th generation recipients at further 6-month intervals. Three of 6 of the 4th generation recipients and 1 of 4 of the 5th generation recipients were significantly repopulated (8–20% Ly5.1 cells in the blood after 12–16 weeks ). However, the proportions of Ly5.1 myeloid cells were low in one of the 3 repopulated 4th generation recipients and in the single repopulated 5th generation recipient (0.5–1% of the Ly5.1 clone in the blood after 12–16 weeks, compared to >5% in all other repopulated mice). Assuming that these latter 2 mice received at least one HSC, and that HSCs are randomly distributed throughout the entire bone marrow mass, it could be calculated from the serial transplant data that the most prolific original HSC transplanted had the ability to produce more than 780,000 daughter HSCs through the execution of at least 19 symmetric self-renewal divisions (and more if not all self-renewal divisions were symmetric). If the 2 mice with low myeloid repopulation are not included, the same calculation indicates an output capacity of 28,000 daughter HSCs requiring at least 14 symmetric self-renewal divisions. In addition, at least 20 subsequent divisions would need to have occurred to account for the >1 million donor-derived WBCs produced in each 5th generation mouse. These results provide the first data quantifying the extraordinary self-renewal and proliferative capacity of a subset of the HSCs in the marrow of normal adult mice.