Whether or not allorestriction should be distinguished from alloreactivity depends on one's model of the TCR-ligand interaction. If the ligand is viewed as a determinant formed by a meld between peptide and the MHC-encoded restricting element, then the TCR, like the BCR, has a single combining site specific for the composite epitope (the Centric Model). If, however, one views the recognition of peptide and the MHC-encoded restricting element as independent, then interactions at two sites of the TCR must be integrated to signal the T cell (the Tritope Model). As TCR recognition of the MHC-encoded restricting element is, by definition, restricted (allele-specific), then under the Centric Model, all TCR signaling interactions with the composite epitope are due to allorestriction, which is peptidespecific. In contrast, under the Tritope Model, there are two classes of signaling interaction, allorestriction and alloreactivity. Alloreactivity is peptide-unspecific and is triggered by recognition of the allo-MHC-encoded restricting element allele. Alloreactivity is incompatible with the Centric Model, under which one would predict that it does not exist. Selected data are analyzed to illustrate the importance of this distinction.