It was reported that the dose of self-antigens can determine the consequence of deletional tolerance and donor T cells are critical for tolerance induction in mixed chimeras. This study aimed at assessing the effect of cell doses and marrow T cells on engraftment and tolerance induction after prenatal bone marrow transplantation. Intraperitoneal cell transplantation was performed in FVB/N (H-2Kq) mice at gestational day 14 with escalating doses of adult C57BL/6 (H-2Kb) marrows. Peripheral chimerism was examined postnatally by flow cytometry and tolerance was tested by skin transplantation. Transplantation of light-density marrow cells showed a dose response. High-level chimerism emerged with a threshold dose of 5.0 × 106 and host leukocytes could be nearly replaced at a dose of 7.5–10.0 × 106. High-dose transplants conferred a steady long-lasting donor-specific tolerance but were accompanied by >50% incidence of graft-versus-host disease. Depletion of marrow T cells lessened graft-versus-host disease to the detriment of engraftment. With low-level chimerism, tolerance was a graded phenomenon dependent upon the level of chimerism. Durable chimerism within 6 months required a threshold of = 2% chimerism at 1 month of age and predicted a 50% chance of long-term tolerance, whereas transient chimerism (<2%) only caused hyporesponsiveness to the donor. Tolerance induction did not succeed without peripheral chimerism even if a large amount of injected donor cells persisted in the peritoneum. Neither did an increase in cell doses or donor T-cell contents benefit skin graft survivals unless it had substantially improved peripheral chimerism. Thus, peripheral chimerism level can be a simple and straightforward test to predict the degree of prenatal immune tolerance.