In this study we have reproduced the prolongation of graft survival by ITI in a rat heart transplant model in which an ITI of an optimal number of donor bone-marrow cells (BMC) was given together with 1ml ALS IP 14 days before transplant. The efficacy of this protocol was critically dependent on the donor-recipient haplotype and influenced by antigenic strength and MHC disparity but not by non-MHC background genes. In strain disparities where ITI was unsuccessful, this was caused by alloreactive recent thymic emigrant cells. In a high responder strain combination the effect was highly dependent on the dose of BMC in the intrathymic injection. Moreover it was readily reproduced with injection of antigen by the intravenous route, even at a lower dose than that required via the intrathymic route. This was in contrast to the other strain combinations tested in which the beneficial effect of donor antigen injection was specific to the intrathymic route, and it suggested that the effect in this group might be at least partly dependent on peripheral mechanisms. The polyclonal ALS can easily be demonstrated to be non-specific in its depletion of peripheral lymphocytes at the dose used in these studies, and we have shown that it also penetrates the thymus. Therefore treatment with ALS may have more effects that mere disablement of peripheral alloreactive T cells, thus complicating the interpretation of the experiments. We have therefore refined our model by recruiting in place of ALS a more specific agent, the partially depleting anti-CD4 monoclonal antibody MRC-OX38, which is equally effective as an adjunct to ITI. We have shown using flow cytometry and immunohistochemistry, that this and certain other monoclonal antibodies, at a therapeutic dose, do not cross the "blood-thymus barrier", and therefore do not complicate the model as ALS potentially does. We would recommend this approach in further studies.