All vertebrates respond to antigenic challenge by specific cellular reactions and by producing circulating antibodies, and they also contain recognition molecules that are evolutionary relics of primitive non-immune recognition. One such ancient recognition molecule is C-reactive protein (CRP), a member of the pentraxin family, that has homologs occurring in species as diverse as vertebrates, tunicates and the horseshoe crab (an ancient arachnoid). This molecule has lectin-like properties, can act as an opsonin and interact with complement and cells in a manner paralleling immunoglobulins (Igs). The horseshoe crab lectin and CRP, although unrelated to Igs, share functional idiotopes with classical antibodies. This finding reflects either an evolutionary convergence or a mechanism of “mini gene insertion” that allows molecules of distinct evolutionary histories to react to the same ligands. Serum antibodies of all vertebrates are polydisperse in charge and are composed of polypeptide chains comparable in mass to those of mammalian light and heavy chains. Molecular genetic studies of placoderm derived vertebrates are incomplete but are sufficient to allow the conclusion that Igs of these species are specified by variable (V), joining (J) and constant (C) gene segments and that rearrangement are an essential feature for the generation of antibody diversity. Here we present new evidence following from the use of antibodies directed against synthetic joining region peptides as probes in the study of rearranging Igs in evolution, the use of recombinant DNA technology to study T cell receptor Vβ genes in a goldfish genomic library and the isolation and characterization of a gene fragment specifying sandbar shark light chain C region. We reached the following conclusions; (1) J region segments are the most conserved in evolution, and this most probably reflects the essential requirements for these gene segments in the formation of intact Ig genes by rearrangement; (2) the framework segments of V regions are highly conserved in vertebrate evolution for both T cell receptors and classical Igs; and (3) although C region segments of light chains of lower vertebrates are homologous to their mammalian counterparts, the degree of conservation of C region structure in phylogeny is apparently less than that for V regions. Essentially, phylogenetic trees can be built using C regions but not V regions. We have identified a molecule of approximate mass of 26kDa in the hemolymph of the tunicate, Boltenia ovipera , that is serologically cross-reactive with shark heavy chain and with J region peptides. This molecule, however, differs from Igs of all vertebrates, including the lamprey, because it does not show the characteristic polydispersity in charge. The existence of gene segments resembling V, J and C in species more primitive than true vertebrates, such as tunicates, is a key issue, because it might provide information enabling us to determine the primordial state and arrangement of the elements of the classical immune system. The existing data are consistent with a “big bang”, in which a massive duplication event occurred very early in vertebrate evolution that generated the multigenic V, J and C system required to generate recognition diversity of antibodies and T cell receptors.