ESPITE CONSIDERABLE progress in assessing in vitro platelet function, in vivo function is relatively poorly understood, because few methods exist to study platelets after transfusion. Thus the in vivo efficacy of the various hemostatic products, developed for the prevention or treatment of thrombocytopenic bleeding, has been difficult to evaluate. This has sometimes led to confusion about optimal conditions for the preparation and storage of human platelets for transfusion.~ This is true not only for random donor platelet concentrates prepared by standard and plateletapheresis methods, but also for platelet products that have been cryopreserved, lyophilized, or prepared and stored in other ways. In addition, various putative platelet substitutes, which represent an emerging group of hemostatic agents, require not only toxicity testing but also test systems to evaluate in vivo efficacy. A requirement for hemostatic products is that they prevent or stop bleeding in thrombocytopenic humans, with minimal toxicity. Contemporary human platelet concentrates must therefore assume the gold standard position against which other hemostatic preparations can be measured. How hemostatic efficacy can be measured has been problematic, and a surrogate in vivo measure for the correction of thrombocytopenic bleeding in humans has been difficult to define. 1 Many of the available in vitro tests of platelet function have been validated against in vivo survival of the transfused product. Survival for many years has represented the in vivo 'measure of platelet function. However, a new platelet product that has equal in vivo survival to a conventional platelet concentrate does not imply that the two necessarily have equal hemostatic function. In the same way, two platelet products that have equivalent in vivo hemostatic efficacy may not have equivalent in vivo survival. Thus, the ability of in vitro tests to predict in vivo hemostatic function is wanting.