Recent studies have suggested that heparin associated thrombocytopenia (H.A.T.) is mediated by immune mechanisms. We studied 7 patients who developed thrombocytopenia while receiving heparin therapy. All of these patients had elevated platelet associated IgG (PAIgG) at the time of the thrombocytopenic episode which ranged from 6 to 47 fg/IgG per platelet (normal less than 5). In all patients the PAIgG returned to normal when the platelet count normalized. Ninety non-thrombocytopenic patients receiving heparin had normal PAIgG on multiple determinations. In four of the H.A.T. patients, recovery serum plus test platelets were incubated with varying concentrations of heparin (0.01-10 I.U. per ml) and the level of PAIgG measured on washed platelets. The PAIgG on the test platelets was within the normal range and less than 10 fg IgG per platelet for all heparin dilutions. In contrast, lo of 13 sera from patients with other types of drug induces thrombocytopenia had significant elevation of PAIgG on test platelets when the recovery sera was incubated with the causative drug and test platelets. Two of the four patients with H.A.T. who did not react in vitro to heparin were subsequently re-administered heparin in vivo and they did not develop thrombocytopenia or elevation of PAIgG. The lack of response to either in vitro or in vivo rechallenge indicates that although H.A.T. is associated with increased binding of IgG to platelets at presentation, the mechanism differs from that observed with the more usual drug induced thrombocytopenia.