Studies on the structural basis of antibody affinity maturation have been carried out by measuring the affinity of secreted antibodies, and information on structures has often been obtained from nucleotide sequences of BCRs of memory B cells. We considered it important to establish whether the repertoire of secreted antibodies from plasma cells is really in accord with that of BCRs on memory B cells at the same time points post-immunization. We isolated plasma cells secreting antibodies specific to (4-hydroxy-3-nitrophenyl)acetyl (NP) hapten by affinity matrix technology using biotin–anti-CD138 and streptavidin–NP–allophycocyanin, to which anti-NP antibodies secreted by autologous plasma cells bound preferentially. We found that plasmablasts occupied >90% of the antibody-secreting cell compartment in the primary response and that they secreted antibodies whose VH regions were encoded by V186.2+Tyr95+ sequences, which provided an increase in the medium level of affinity by somatic hypermutation (SHM) of heavy chains at position 33. After secondary immunization, a further increase in antibody affinity was observed, which was explained by the appearance of a number of plasma cells secreting V186.2+Gly95+ antibodies that acquired high affinity by multiple SHMs as well as plasmablasts secreting V186.2+Tyr95+ antibodies. However, we did not detect any plasmablasts secreting V186.2+Gly95+ antibodies, showing that plasmablasts and plasma cells have a different antibody repertoire, i.e. their respective repertoires are asymmetric. On the basis of these findings, we discussed the relationship between the BCR affinity of memory B cells and plasmablasts as well as plasma cells as pertaining to their ontogeny.