Platelet fibrinogen has been studied in normal, thrombasthenic, and hypofibrinogenaemic subjects. It has been differentiated into adsorbed (plasma) and extractable (intraplatelet) fractions. Isotopic studies suggest that exchange does not occur between intraplatelet and plasma fibrinogen and it appears possible that the intra-platelet fraction may be derived from the megakaryocyte. Six of nine thrombasthenic patients were found to have a severe deficiency of both adsorbed and extractable fibrinogen. Since the remaining three had near-normal platelet fibrinogen and all nine failed to aggregate it is improbable that the failure to adsorb fibrinogen is responsible for the defect in aggregation. Magnesium partially corrects adhesion to fibrin and clot retraction by these platelets, but has not been found to influence their fibrinogen adsorption. It is considered that the basic platelet surface defect, of varying severity, is responsible for the abnormalities of adsorption, aggregation, and adhesion in thrombasthenia. In the case of congenital hypofibrinogenaemia, fibrinogen transfusion corrects the long bleeding time, platelet-adsorbed fibrinogen, and the ability of platelets to spread on glass. It is possible that fibrinogen influences the surface properties of human platelets, although the final mechanism is not determined.