To study association of platelets with factor VIII, the purified protein was 125I-labelled with Bolton-Hunter reagent to a specific activity of 243,000-360,000 cpm/U. Autoradiographs of SDS polyacrylamide gels revealed polypeptides of VIII at Mr about 240 kDa, 90 kDa and intermediate values, as well as some radioactive contaminants, but the light chain (Mr 78/76) seen with silver stain was not labelled. After 2.5-5-fold activation with thrombin, the higher radioactive Mr band disappeared, the band at 90 kDa became more intense, and a band appeared at about 45 kDa. The radioactivity associated with platelets, studied in the presence of haemophilic BaSO4-treated plasma, was maximal after 6-8 min and increased 3-15-fold on activation with thrombin. With activated VIII, autoradiographs of platelet pellets showed only VIIIa but results are expressed as units of unactivated VIII bound. At 0.3-0-0.7 U/ml, 10(8) platelets bound 0.0008-0.004 U VIIIa. The amount bound was not affected by the ratio of unlabelled VIII to VIII labelled in the presence of a 50-fold molar excess of unlabelled Bolton-Hunter reagent. Binding increased to 1.5 U VIIIa/10(8) platelets (about 13,600 molecules per platelet) at 140 U/ml, with no evidence of saturation. Binding was not affected by monoclonal antibodies to platelet glycoproteins IIb/IIIa or Ib, quenching the thrombin before adding platelets, or aggregating the platelets with A23187 in the presence of thrombin. Qualitatively, binding of labelled VIIIa and factor Va studied by others are similar. Binding of 125I-VIIIa to aggregated and unaggregated platelets was normal in patient M.S. whose platelets were shown by others to be deficient in their ability to bind radiolabelled factor Xa and generate coagulant activity. This difference. and the fact that platelet coagulant activity is increased by platelet activation and/or aggregation, suggest that binding of Va or VIIIa alone does not determine the assembly of active proteolytic complexes on the platelet surface.