The clonal expansion theory of the immune system was first proposed over 30 years ago (Jerne, 1955; Burnet, 1959). It still forms the basis of current thinking about the immune system. It was not obvious to the authors of this paper that the combinatorial possibilities of the 20 available amino acids in antigen peptides could result in a system that could distinguish successfully between self and foreign antigen, utilizing only the trial and error method proposed by the clonal expansion theory. In this paper it is demonstrated that the theory is credible on this issue, at least with respect to major histocompatibility complex (MHC) restricted antigen recognition, and that the very high proportion of immature lymphocytes that die in the thymus is consistent with the theory. © 1993 Academic Press. All rights reserved.