Quantification of platelet-bound immunoglobulin is widely used in the evaluation of thrombocytopenia. Several methods have been devised among which labelled ligand-binding assays seem to be most appropriate. In series of adult patients such assays have been shown to be superior in separating immune-thrombocytopenia from thrombocytopenia of non-immune causes. We studied 62 children with thrombocytopenia of various causes, using radiolabelled protein A as a ligand to measure platelet-surface bound IgG. The test was highly sensitive (93%) in detecting immune-thrombocytopenia. The specificity, however, was only 57%, which is less than in published studies of adults. In a number of cases presumed to be non-immune-thrombocytopenia, notably a few patients with leukaemia and bone marrow aplasia, we found increased amounts of platelet surface-bound IgG. The significance of this finding is not clear. An indirect assay measuring platelet-specific IgG in plasma was less sensitive (46%) but highly specific for immune-thrombocytopenia (89%). The measurements of platelet-surface-bound IgG and platelet-specific IgG in plasma are of limited diagnostic value in childhood thrombocytopenia but are useful in following the treatment in chronic ITP.