Oxygen-carrying plasma expanders (blood substitutes) have been sought for over a century. Development of current products is a result of evolution in the understanding of proteins in general, of hemoglobin in particular, and of how cell-free hemoglobin interacts with the control of local blood flow to ensure adequate tissue oxygenation. Hemoglobin-based products are considered in four "generations" corresponding to major improvements. First-generation products consisted of hemoglobin, freed of red cell membranes (stroma-free hemoglobin [SFH]), which was renal toxic and vasoactive. Second-generation products were polymerized with aldehyde reagents to reduce or eliminate the renal toxicity, but the products were heterogeneous and still vasoactive. Third-generation products employed more specific intramolecular crosslinking to eliminate polymerization and promote homogeneity, but they also remained vasoactive. Fourth-generation products are based on a new understanding of the way in which microvascular blood flow is controlled and the influence of O(2) delivery to vascular walls. After more than a century of research, one of these new solutions should find use as an alternative to red cells for transfusion in certain clinical settings.