The hemoglobin solution covalently linked to dextran benzene tetracarboxylate (dex-BTC) described in Part 1, offered the following characteristics : [Hb] = 70 g/l, non modified Hb < 15%, metHb < 5%, P50 close to blood value, viscosity and oncotic pressure near to physiological values (37-degrees-C). It was biocompatible, stable in plasma, non hypotensive in rat model, non pyrogenic and did not present abnormal toxicity (on mice, according to the french Pharmacopea). Nevertheless, on account of the presence of dextran in the conjugate and due to the literature, we observed an anaphylactoid reaction in rats consecutive to the injections : absence of hypotension but presence of a massive oedema resulting from an important increase in the capillary permeability with harmful consequences to the vascular retention and survival time. As guinea pigs are insensitive to dextran, we are beginning to assess this conjugate solution in this animal. The first results are very promising : a lack of mortality after large injection, an almost complete vascular retention, a weak urinary loss (only non conjugated Hb), and a plasmatic half-disappearance time close to 7h. Results from haemorrhagic shocks to a final hematocrit < 0.10 l/l and total and isovolemic exchange transfusions seem to prove a real oxygen-carrying capacity of this new Hb solution.