Artificial blood is herein defined as consisting of red cell substitutes. Red cell substitutes are solutions intended for use in patients whose red cells are either not available or their use is to be avoided for other reasons. Despite enormous effort, more than 100 years of research have not produced a solution that can be used safely in humans. Hemoglobin can exist in either of two structural conformations, corresponding to the oxy (
R
, relaxed) or deoxy (
T
, tense) states. The key differences between these two structures are that the constrained
T
state has a much lower oxygen affinity than the
R
state and the
T
state has a lower tendency to dissociate into subunits that can be filtered in the kidneys. Therefore, stabilization of the
T
conformation would be expected to solve both the oxygen affinity and renal excretion problems. The clinical consequences of stabilization are not known. Differences in the reactions considered to be useful in the production of hemoglobin‐based blood substitutes are determined by the dimensions and reactivity of the cross‐linking reagents. Even small differences among structures of the reagents can yield products having very different properties. In addition, the conditions of the reaction are very important, not only in regard to the state of ligation, ie, oxygen saturation, but also in regard to the presence of agents or molecules that block or compete for certain reactive sites. It has been difficult to produce a pure modified hemoglobin for toxicity studies because most processes start with relatively crude, stroma‐free hemoglobin. Hemoglobin is provided by the red blood cell in highly purified form. However, red cell membranes contain many components that could potentially cause toxicity problems. Furthermore, plasma proteins and other components could cause toxic reactions in recipients of hemoglobin preparations. Rabiner's method for the filtration purification of hemoglobin was thought to be a significant advance over older centrifugation methods. However, hemoglobin prepared in this way still caused unwanted reactions in human recipients. If clinical efficacy and safety of hemoglobin solutions can be shown, the demand for product would soon outstrip the supply of outdated human blood. One solution to the hemoglobin supply problem is to use as a starting material blood from nonhuman sources. For example, bovine hemoglobin is being developed for use. The ultimate success of bovine, or any other hemoglobin, depends on demonstration of safety, not on supply. One problem in using bovine hemoglobin is the fear of bovine spongiform encephalitis (BSE) virus. The FDA is concerned about bovine products of all types. An alternative and novel source of hemoglobin for modification is from microorganisms the genome of which has been modified to contain globin genes for recombinant hemoglobin (rHb) production. Significant strides have been made in this approach. Several of the products discussed herein are under intense development. One product, based on recombinant hemoglobin, was in early human trials as of this writing. Other hemoglobin‐based solutions are also under review at the FDA.