Corneal grafts are believed to enjoy a degree of "immunological privilege" primarily due to the avascularity of the recipient bed. In this study orthotopic full-thickness corneal grafts were carried out in the inbred rat, using a technique that is a close model of corneal grafting in humans. The survival times of corneal grafts on nonvascularized beds of 28 fully allogeneic strain combinations were determined without the use of immunosuppression. Some combinations were rejected rapidly, e.g. DA (RT1a) into BN (RT1n) with a mean survival time (MST) +/- SD of 7.8 +/- 1.3 days, and some at a moderate rate, e.g. AO (RT1u) into LEW (RT1l) with an MST of 23.1 +/- 10.0 days, whereas in other cases survival was indefinite, e.g. WAG (RT1u) into PVG (RT1c), an MST of greater than 100 days. Orthotopic corneal grafts on nonvascularized beds between DA and AO parents and the F1, followed the basic rules of transplantation genetics. In addition, the rate of graft rejection was significantly faster (P less than 0.001) with corneal grafts from DA into AO placed onto a vascularized compared with a nonvascularized corneal bed (MST of 6.8 +/- 2.4 or 12.1 +/- 4.0 days respectively). The rate of rejection of corneas on a vascularized bed was at a similar rate to that of orthotopic skin or heterotopic auxiliary heart grafts. The results indicate that the fate of a corneal allograft on a nonvascularized bed is dependent upon the particular combination of donor and recipient strain. No consistent association was observed between any donor or recipient RT1 haplotype and survival; this suggests that non-RT1 background genes may play a role in the survival of corneal grafts.