Recently we determined the absolute frequencies of lipopolysaccharide (LPS) and Nocardia-delipidated cell mitogen (NDCM) -sensitive B lymphocytes from BALB/c mice secreting immunoglobulin (Ig) molecules bearing any one of three different idiotopes originally found on a monoclonal anti-fl-galactosidase antibody (1). These three idiotopes, 66, 137 and 395, are defined by syngeneic monoclonal antibodies against the monoclonal Ig 174 produced by a BALB/c mouse (C. Le Guern, E. Barbier, and D. Juy. Idiotypic heterogeneity of monoclonal anti-fl-galactosidase antibodies. Manuscript in preparation). Extensive studies by conventional immunization clearly demonstrated that these three determinants are not normally expressed at detectable level when BALB/c mice are immnized with fl-galactosidase and, therefore, can be classified as nonrecurrent idiotopes (Le Guern et al., see above). Our results, however, indicate that these specificities are not only part of the idiotypic repertoire of LPS-and NDCM-sensitive BALB/c B cells, but that they can be induced to expression with a frequency similar to that of a recurrent idiotype, i.e., M-460 (1, 2). Because the frequency of a given specificity (either antibody or idiotypic specificity) determined by mitogenic stimulation does not require the presence of antigen, this analysis was not dependent on a particular protocol of immunization, but rather reflected the absolute frequencies of competent B cells in a steady state. Thus the discrepancy found between antigenic and polyclonal activation with regard to idiotope(s) expression may very well reflect independent expression of these V region markers and antigen-binding sites. The existence of defined idiotopes on immunoglobulins without known antigen specificity has been fully documented for some considerable time (3-6) but, up until now, it has been difficult to establish the possible existence of well-defined rules that govern the relationship between idiotopes and antigen-binding sites. Such studies have been hampered by the difficulty of selecting B cell clones or hybridomas that are positive or negative for any possible combination of a number of well-defined sets of V regions. Here, however, we were able to address this problem, because we knew the absolute frequencies of each of our three idiotopes and therefore could select and study individual clones of mitogenreactive B cells positive for any possible combination of these determinants and correlate them to anti-/~-galactosidase activity. The results presented here indicate