Riboflavin is essential for normal erythropoiesis in rats, dogs, pigs, and monkeys. There is no evidence that this vitamin is required for normal erythropoiesis in man. The anemia in swine is normocytic.

Nicotinic acid deficiency is accompanied by a severe anemia in dogs. The type of anemia produced is normochromic and may be either macrocytic or normocytic and is associated with a mild reticulocytosis. Limited observations indicate that the bone marrow is hypoplastic and that erythropoiesis stops at the erythroblastic level. An anemia due to a deficiency of this vitamin has not been demonstrated in other species nor in man.

Pyridoxine is essential for normal erythropoiesis in chicks, rats, dogs, and pigs. The anemia is microcytic and slightly hypochromic in type. Anisocytosis, microcytosis, polychromatophilia, and normoblasts can be seen in the blood smear. An irregular reticulocytosis is present. The bone marrow is hyperplastic and there is an increase in the nucleated red blood cells. The anemia is accompanied by hemosiderosis of the tissues, an elevated serum iron level, and degeneration in the nervous system. There is no evidence of an increased rate of hemolysis. No relationship between pyridoxine and erythropoiesis has been demonstrated in man.

The "Lactobacillus casei group" includes the norite eluate factor, the L. casei factor from liver, folic acid, the Streptococcus lactis R factor of Keřesztesy et al., the yeast factor of Stokstad, the factor of Hutchings et al., vitamin M11, xanthopterin, vitamin Bc, vitamin Bc conjugate, vitamins B10 and B11, and pyracin.

The L. casei factor from liver has been identified as pteroylglutamic acid. The available evidence indicates that the norite eluate factor, folic acid, vitamin M, vitamin Bc, vitamin B10, and vitamin B11 are identical with pteroylglutamic acid. The Streptococcus lactic R factor of Keřesztesy et al. may be pteroic acid. The yeast factor of Stokstad is unidentified. The fermentation factor of Hutchings et al. has been identified as pteroyltriglutamic acid. Vitamin Bc conjugate is now known to be pteroylheptaglutamic acid. Thus the various members of this group are closely related chemically and represent minor alterations of a basic structure. The corresponding deficiency syndromes are probably identical. In the rat the deficiency is manifested by severe normocytic anemia, severe granulocytopenia, leukopenia, and thrombocytopenia. Nucleated red cells appear in the peripheral blood. Bone marrow studies suggest a maturation arrest in the early stage of development of all three of the cellular elements of the blood. The manifestations of the deficiency in the chick are macrocytic anemia, leukopenia, and thrombocytopenia. Again immature red cells are present in the peripheral blood. In the monkey the manifestations of the deficiency are normocytic anemia, leukopenia, and thrombocytopenia. In human beings the synthetic L. casei factor from liver (pteroylglutamic acid) has been shown to be effective in the treatment of various types of macrocytic anemia including pernicious anemia and sprue. The relation of this substance to the antipernicious anemia substance in liver remains to be determined.

The extrinsic factor of Castle is still unidentified. It now seems reasonable that it is related in some way to pteroylglutamic acid. It is unlikely that it is identical since the synthetic L. casei factor is effective even in the absence of normal gastric juice. A deficiency of the extrinsic factor in man results in an anemia which is identical with pernicious anemia and the bone marrow is cytologically indistinguishable. An accompanying neutropenia and thrombocytopenia are also frequently seen. The anemia responds rapidly to the parenteral administration of highly purified antipernicious anemia liver extracts and to pteroylglutamic (folic) acid. Achlorhydria is generally not present. Macrocytic anemia of nutritional origin occurring in the tropics varies from this anemia in one important aspect. It fails to respond to highly purified liver extracts. This strongly suggests that the factor responsible for the deficiency is distinct from that of the extrinsic factor of Castle. A deficiency of this factor has been produced in monkeys and the deficiency syndrome consists of a macrocytic anemia with a megaloblastic bone marrow. The anemia fails to respond to highly purified liver extracts which are effective in the treatment of pernicious anemia but does respond to crude liver extracts and to marmite, an autolyzed yeast extract. The relation between this factor and the L. casei factor has not been investigated.

The role of ascorbic acid in erythropoiesis is not clear. Although the scorbutic state in both guinea pigs and human beings is frequently accompanied by anemia it is questionable whether the anemia is due specifically to a deficiency of ascorbic acid. Much of the animal experimentation is inconclusive because pure ascorbic acid supplements were not used. Further work in animals is needed. In man it has been both asserted and denied that synthetic ascorbic acid is effective in relieving the anemia. It would seem, however, that there are some scorbutic patients who respond specifically to pure ascorbic acid. The anemia accompanying scurvy has been reported as macrocytic, normocytic, and microcytic. An induced, uncomplicated ascorbic acid deficiency in a human being did not result in anemia.

Pantothenic acid deficiency results in a normocytic anemia of moderate degree in pigs in about two-thirds of the animals. There is evidence which suggests that a deficiency of this vitamin in rats may result in anemia, granulocytopenia, and bone marrow hypoplasia. Not all animals show these changes and pantothenic acid, although completely preventive, does not exert a curative action in all animals. There seems to be a relation between pantothenic acid deficiency and a deficiency of the L. casei factor in the rat.

Choline deficiency in dogs results in a severe anemia. In many animals this change is irreversible. This may be explained by the irreversible liver damage which is present.

Biotin is necessary for the production of hemoglobin values greater than 14 grams per cent in dogs maintained on a highly purified ration. There is no evidence that biotin has an effect on erythropoiesis in other species.

In addition to the factors described above it has been shown that monkeys, pigeons, and guinea pigs require at least one more additional factor for normal erythropoiesis.

There is no evidence that thiamine, p-aminobenzoic acid, and inositol are concerned in erythropoiesis in any species.

Considering the relative size of the globin fraction of the hemoglobin molecule it is understandable that a deficiency of protein results in anemia. This has been demonstrated in rats and dogs. It has been pointed out that because of a marked reduction in the total blood volume only when the total circulating hemoglobin is determined and adjusted to a unit of surface can the true severity of the anemia be appreciated. Equine globin contains all ten of the "essential" amino acids and at least nine "nonessential" amino acids. Human globin has not been so extensively studied. It would be expected that a deficiency of any one of the "essential" amino acids would give rise to anemia. Actually, specific deficiencies of tryptophan, lysine, phenylalanine, and isoleucine have been produced in the rat and anemia developed in each instance. The morphological characteristics of these anemias have not been carefully investigated. The anemia due to tryptophan deficiency in the rat has been stated to be normocytic and normochromic. An anemia probably due to a lack of tryptophan has been produced in pigs. This anemia is normocytic, normochromic, and accompanied by a hypoplastic or normoplastic bone marrow and a normal level of iron in the serum. No increase of hemosiderin in the tissues has been noted. Whether the anemia produced in rats by feeding deaminized casein is due to a toxic substance rather than a deficiency of lysine is unsettled although large amounts of lysine prevent its development. Evidence that glycine is utilized in the synthesis of the pyrrole rings of protoporphyrin has been obtained by labeling this amino acid with N15 and feeding the labeled compound to rats. Pyrroles have also been synthesized in vitro from glycine. Similar evidence is available to indicate that acetic acid, or a derivative of it, is utilized for porphyrin synthesis.

Three mineral elements, iron, copper, and cobalt, have been shown to be essential for normal erythropoiesis in at least one species each. Iron is probably required for erythropoiesis in all mammals. A deficiency results, at least in the chronic stages, in a microcytic hypochromic anemia and is accompanied by a normoblastic, hyperplastic bone marrow and a low serum iron level, an increased amount of protoporphyrin in the erythrocytes, and an elevated serum copper level. Nucleated red blood cells are occasionally seen in the peripheral blood and the reticulocytes are increased.

The fundamental concepts of iron metabolism have changed greatly in recent years. These may be summarized. Iron is absorbed chiefly in the duodenum. In man it is absorbed principally as ferrous iron. Dogs absorb both valence forms well although some animals absorb the ferrous form more readily than the ferric form. Rats absorb both forms equally well. The absorption of iron is also dependent upon the concentration of the iron in the intestine, upon the solubility of the iron salt, and in the human being at least upon the presence of reducing substances in the diet as well as the reducing action of the gastric hydrochloric acid. In addition to these factors the need of the body for iron may determine, to a certain degree, the amount absorbed. This is known as the "selective absorption" theory. Recently it has been suggested that apoferritin acts as a receptor compound in the mucosal cell. As the concentration of the plasma iron falls, ferrous iron is removed from the mucosal cell resulting in a diminution of ferritin in the mucosa. When the ferritin has diminished to a point where the cell is no longer saturated with respect to ferrous iron, more iron is absorbed into the mucosal cell. Once absorbed the iron is transported in the plasma to the tissues where it is stored to a great extent as ferritin, a protein-iron complex. The iron is then used over and over again for hemoglobin synthesis. Iron is excreted from the body in only insignificant quantities. This theory requires substantiation.

Copper has been shown to be essential for normal erythropoiesis in chickens, mice, rats, rabbits, dogs, pigs, sheep, cattle, and infants. A deficiency of this mineral in rats is manifested by a microcytic hypochromic anemia and a moderate reticulocytosis. A condition due to a deficiency of copper, known as "enzootic ataxia," occurs in sheep in Western Australia. Anemia may be severe. In young lambs it is microcytic and hypochromic and is accompanied by demyelinization of the nervous system and hemosiderosis of the tissues. In adult sheep the anemia is slightly macrocytic and hypochromic. Blood smears reveal anisocytosis, poikilocytosis, Howell-Jolly bodies, normoblasts, numerous macrocytes, stippling, and polychromatophilia. Similar blood changes have been reported in copper-deficient cattle in Western Australia. In nutritional anemia in infants the rate of erythropoiesis is accelerated when copper is given in addition to iron. In adults supplemental copper therapy may be of value in a few cases. Such cases, if they occur, are rare. Most cases will respond if adequate doses of iron are given. This does not necessarily indicate that copper is not needed for erythropoiesis or that it is not a dietary essential but rather that the quantities needed are so small that sufficient copper is present in the body stores in adult life, in the diet, or as a contaminant in the iron used therapeutically to supply the needs. No case of uncomplicated copper deficiency has been reported in man. The manner in which copper is related to the formation of red cells is not understood.

The role of cobalt in erythropoiesis is unique. A deficiency results in anemia. The administration of small amounts to normal animals produces a polycythemia, whereas the administration of large amounts depresses erythropoiesis. The enzootic occurrence of cobalt deficiency in sheep and cattle has been reported from various regions of the world. Anemia is present and is oftentimes severe. The anemia is either normocytic or microcytic and usually hypochromic. Blood smears reveal anisocytosis and poikilocytosis. There is a hypoplasia of erythrogenic tissue in the bone marrow, hemosiderosis of the tissues and a reduction in reticulocytes in the blood. An experimental anemia due to cobalt deficiency has not been produced in either rats or dogs. There is no substantial or convincing evidence that cobalt is needed by human beings for normal erythropoiesis. The administration of small amounts of cobalt to normal rats, dogs, guinea pigs, frogs, mice, rabbits, chickens, pigs, and ducks produces a marked polycythemia which is accompanied by a reticulocytosis, hyperplasia of the bone marrow, and an increased erythropoietic activity in the spleen and liver. Larger doses of cobalt inhibit erythropoiesis. The metabolism of cobalt is unlike that of iron. The excretion of cobalt from the body once it is absorbed is exceedingly rapid and is principally through the kidneys.

In conclusion, certain vitamins, namely, riboflavin, nicotinic acid, pyridoxine, "folic acid," and the extrinsic factor, have been shown to be essential for normal erythropoiesis in at least one species each. It has been claimed that ascorbic acid, pantothenic acid, choline, and biotin play a role in erythropoiesis but these claims need substantiation. There is no substantial evidence that thiamine or inositol is concerned in red cell formation. The significance of p-aminobenzoic acid has yet to be determined. Protein is essential for normal red blood cell formation. The globin fraction of the hemoglobin molecule contains all ten of the "essential" amino acids as well as many of the "nonessential" ones. The stroma of the red cells also contains amino acids. It is logical, therefore, to assume that in the absence of any one of the so-called essential amino acids hemoglobin formation cannot take place normally. Actually specific deficiencies of tryptophan, lysine, phenylalanine, and isoleucine have been produced in the rat and anemia has developed in each instance. There is evidence to show that glycine and acetic acid, or a derivative of it, are utilized in the synthesis of the pyrrole rings of protoporphyrin. Three mineral elements, iron, copper, and cobalt, have been shown to be essential for normal erythropoiesis.

I gratefully acknowledge my indebtedness to Dr. Maxwell M. Wintrobe for his kind advice and aid in the preparation of this review as well as for the liberal use of his extensive reprint file.