Classical and more recent studies have provided a description of the in vivo behavior of the totipotent hematopoietic stem cell and its clonal progeny. These reconstitution experiments, employing clonotypic markers have shown that single or few engrafted lymphoid-myeloid stem cells are both necessary and sufficient for long-term, stable hematopoiesis in a reconstituted mouse. This underscores the remarkable developmental capacity of individual stem cell clones. Furthermore, the long-term and retransplantation studies have provided an indication of stem cell self-renewal ability. Taken together, the long-term analyses have also shed light on the dynamic behavior of engrafted stem cell clones and of the entire reconstituted hematopoietic system. A model is presented where the developmental and proliferative behavior of totipotent stem cells is a function of time. In this model commitment versus self-renewal decisions may be governed by stochastic mechanisms. However, the actual contribution by stem cells to particular mature cell populations may be more a function of lineage specific demands as they change over post-engraftment time.