An unusual patient with post-transfusion purpura is described who had a rapidly rising platelet count with high dose steroid treatment in the presence of an increasing anti-PlA1 antibody titre. A platelet autoantibody could not be detected by a radioactive platelet antiglobulin test, but autologous platelet survival data indicated a compensated thrombolytic state in the recovery period. Thus in spite of the failure to demonstrate a platelet autoantibody, the rapid response to steroids in association with a reduced platelet survival is compatible with an autoimmune platelet destructive process. In addition to PIA1 platelet antibodies, this patient had an associated anamnestic recall of multispecific red cell and lymphocytotoxic antibodies.