After years of hope and frustation, research on artificial oxygen carriers seems to have exited from an apparent stagnation. Several products have recently been the subject of a certain publicity: one is a synthetic compound based on perfluorocarbons, while others are solutions of modified human hemoglobin (Hb). These latter transporters are prepared from outdated bank blood or from recombinant Hb. Much progress have been made in controlling the oxygen affinity of Hb solutions. Several strategies, based on mutants or cross-links, have lead to Hb solutions with properties similar to those of natural erythrocytes. The lack of the physiological effector 2.3-diphosphoglycerate in the purified Hb solutions can be compensated for by a combination of mutations, or by the type of crosslinking. Unfortunately, less progress have been made to compensate the loss of the reducing system, and the rate of oxidation may limit the useful lifetime of these solutions. A major obstacle in the utilisation of the Hb based oxygen carriers (HBOC) is the appearance of a vasoconstriction, which may be overcome through a better understanding of the mechanism of interaction of NO with Hb. Apart from this effect, the initial clinical tests have not revealed a toxicity to the exposure of Hb solutions. There is thus some optimism that the functional properties can be improved to produce a useful HBOC.