Previous explanations for the allograft reaction have been based on the concept that antigen causes immunocyte activation, following engagement of the immunocyte's specific receptor. This notion lead to the concept of immune surveillance, the idea that the evolutionary pressure responsible for the development of the vertebrate immune system involved in allograft rejection was a need to recognize and destroy tumor cells that carried novel antigens. Allografts were rejected because they were, in effect, mistaken for tumor cells. At the practical level, these ideas suggested that a solution to the allograft problem required treatment of the recipient in a way that would reduce or eliminate the recipient's immune response to the grafted tissue. We have rejected these ideas on the grounds that the basic premise, the notion that antigen alone drives T cell differentiation, is invalid. To explain the origin of the allograft response, we have developed a theory of allogeneic reactivity based on the concept that a stimulator cell is required for the activation of blood cells involved in both nonspecific inflammatory reactions and specific cellular immunity. This theory provides a conceptual link between invertebrate and vertebrate alloreactivity and explains why the MHC and factors controlling the expression of T cell activity map in the same region of the genome. According to this theory, it is blood cells carried within the transplanted tissue and not transplantation antigen on the surface of graft parenchymal cells, that constitute the major barrier to allotransplantation. Experimentally we have presented evidence for a two-signal mechanism for T cell activation. Both antigen and an inductive stimulus are required for T cell activation, and neither factor alone induces detectable T cell activation. Organ culture of thyroid tissue for 4 weeks renders it non-immunogenic without altering its antigenic composition. Furthermore, cyclophosphamide pretreatment of the thyroid donor, a procedure that does not destroy the vascular bed of the donor tissue, also reduces its immunogenicity. These findings are of both theoretical and practical importance. They show that transplantation antigen carried on the parenchymal cells of a transplant do not constitute the major barrier to allotransplantation and, at least in the case of thyroid and parathyroid transplantation, indefinite allograft survival can be achieved by treatments directed at the transplanted tissue and not the recipient.