T lymphocytes are activated by the interaction between the T-cell antigen receptor (TCR) and peptides presented by major histocompatibility complex (MHC) molecules. The avidity of this TCR-pMHC interaction is very low. Therefore, several hypotheses have been put forward to explain how T cells become specifically activated despite this handicap: conformational change model, aggregation model, kinetic segregation model, sequential interaction model and permissive geometry model. In the present paper, we conducted experiments to distinguish between the TCR-aggregation model and the TCR-conformational change model. The results obtained using a TCR capture ELISA with Brij 98-solubilized TCR molecules from normal or activated T cells showed that the ligand-TCR interaction causes structural changes in the CD3 epsilon cytoplasmic tail as well as in the extracellular TCR beta FG loop region. Size-fractionation experiments with Brij 98-solubilized TCR/CD3/co-receptor complexes from naïve or activated CD4(+) or CD8(+) T cells demonstrated that such complexes are found as either dimers or tetramers. No monomers or multimers were detected. We propose that: (1) ligand-TCR interaction results in conformational changes in the CD3 epsilon cytoplasmic tail leading to T-cell activation; (2) CD3 epsilon cytoplasmic tail interaction with intracellular proteins may dissociate pMHC and co-receptors (CD4 or CD8) from TCR/CD3 complexes, thus leading to the arrest of T-cell activation; and (3) T-cell activation appears to occur among dimers or tetramers of TCR/CD3/co-receptor complexes interacting with self and non-self (foreign) peptide-MHC complexes.