We estimated numbers of transplantable primitive stem cells (PSCs) and found evidence that the same PSC continuously produced circulating erythrocytes and lymphocytes. These estimations used the binomial formula on data from recipients of identical portions of marrow mixtures containing two distinguishable cell types. Analysis of variance was used to compare repeated tests within each recipient. Values of pi s or pi c, probabilities that two independently sampled cells were descended from the same PSC, were also estimated, as this does not require the unverified condition that all PSCs contribute equally to the differentiated cell population. Several months after transplantation, erythrocytes were descended from only a single PSC per 1-2 X 10(5) marrow cells injected, several times rarer than previously reported. Percentages of erythrocyte and lymphocyte types in each recipient were closely correlated, with r values ranging from 0.86 to 0.94, in groups receiving 2-8 X 10(5) marrow cells; apparently the same precursors repopulated both myeloid and lymphoid lines in each recipient, as expected of true PSCs. Our data did not fit the clonal succession model, which predicts sequential activation of new PSCs and deactivation of old. Between 76 and 154 days, differentiated erythrocyte precursors were probably exhausted, with no evidence for new precursor activation or for further change between 154 and 250 days. The percentage of newly produced erythrocytes (reticulocytes) of each donor type varied little when individual recipients were followed between 165 and 295 days after transplantation, and variances within recipients were similar at marrow doses from 8 to 200 X 10(5) cells, further contradicting models of sequential activation and deactivation of PSC clones. Thus, transplanted PSCs were continually active during much of the recipient's lifespan.