From January, 1982, to July, 1990, 51 children with recurrent infections were investigated for immunodeficiency in this department by testing neutrophil function, lymphocyte subsets and serum immunoglobulin and complement concentrations. The prevalence of immune dysfunction within the group was 39% (20 of 51). A previously described clinical scoring system, which aims to identify children with a history of recurrent infection who merit investigation for immunodeficiency was also applied to all 51 children. The scoring system identified only 55% (11 of 20) of those with laboratory-proved immunodeficiency and had a false negative rate of 45% (9 of 20). This latter group included 2 children with severe combined immunodeficiency and 1 with hypogammaglobulinemia, diagnoses that one cannot afford to miss. The system was not sufficiently sensitive to be of use in deciding which child to test for immunodeficiency.