Human immunoglobulin G (IgG) can be divided into four subclasses that are selectively expressed. For instance, carbohydrate antigens preferentially elicit IgG2 antibodies, whereas protein antigens usually elicit IgG1 and IgG3. Elucidating the biological basis of the selective expression of these IgG subclasses is important to our understanding immunodeficiencies and B lymphocyte development. To investigate clinical importance of IgG subclass deficiencies, a sensitive and specific assay has been developed for IgG subclasses using particle concentration fluorescence immunoassay. Preliminary clinical studies have already shown that infection-prone individuals often have selective IgG2 subclass deficiency. Normal levels of IgG2, however, do not rule out an immunodeficiency in the infection-prone individuals because some individuals have normal levels of IgG subclasses and are poorly responsive to antigens of bacteria. Based on animal studies, two contrasting models of B cell development have been advanced. One model of B cell development proposes a single lineage and proposes that a B cell can successively switch and produce any IgG subclass. The other model proposes multiple lineages and proposes that a B cell can express only some IgG subclasses. It has been found by us that anti-PC antibodies are mostly IgG2 with some IgG1, and that the V region of IgG1 anti-PC antibody is different from that of IgG2 antibody. Our finding, therefore, suggests that B cells producing anti-PC antibodies are progeny of not one ancestral B cell that has successively switched, but two independent ancestral B cells. Cellular studies using polyclonal activators also suggest that regulatory mechanisms for IgG1 and IgG3 are different from those of IgG2 and IgG4. Taken together, we favor the multi-lineage model better than the single lineage model of human B cell development.