The treatment of type I diabetes by islet cell transplantation, while promising, remains restricted due to the incomplete efficacy and toxicity associated with current immunosuppression, and by limited organ availability. Given reports suggesting bone marrow derived stem cell plasticity, we sought to determine whether such cells could give rise to pancreatic islet cells in vivo. In the context of autoimmune diabetes, we transplanted unfractionated bone marrow from beta-gal trangenic donor mice into NOD mice prior to, at, and two weeks beyond the onset of disease. Successful bone marrow engraftment before diabetes onset prevented disease in all mice and for 1 year after transplant. However, despite obtaining full hematopoietic engraftment in over 50 transplanted mice, only one mouse became insulin independent, and no beta-Gal positive islets were detected in any of the mice. To test whether tolerance to islets was achieved, we injected islets obtained from the same allogeneic donor strain as the hematopoietic cells into 4 transplant recipients, and 2 had a reversion of their diabetes. Thus allogeneic bone marrow transplantation prevents autoimmune diabetes and tolerizes the recipient to donor islet grants, even in diabetic animals, yet the capacity of bone marrow derived cells to differentiate into functional islet cells, at least without additional manipulation, is limited in our model.