n vitro, rheological studies establish that whole blood viscosity and yield stress are high in patients with an erythrocytosis. However, a number of factors ensure that these patients, under physiological conditions, do not show the clinical features observed in other hyperviscosity states. These include red cell axial migration in flowing blood and "plug flow" in the largest vessels. In addition, a small increase in vessel diameter leads to large increases in blood flow, and generally high blood flows produce the lowest blood viscosity values. The increased hemoglobin levels and the increase in oxygen-carrying capacity at high hematocrit values compensate for the tissue hypoxia. In the "non-hypoxemic" erythrocytoses (polycythemia vera, idiopathic and apparent erythrocytosis), there is an increased incidence of vascular occlusion in untreated patients. The reasons for this include reduced peripheral blood flow, increased platelet-vessel wall interactions, and the demonstrated in vitro hyperviscosity which comes into play with abnormally low flow, seen in vivo under pathological conditions. In the erythrocytosis of hypoxemic lung disease and its associated hypoxemia, pulmonary vasoconstriction enhances susceptibility to hyperviscosity effects in particular. Moreover, the vasoconstriction caused by the hypoxemia prevents the normal adaptive changes of increased vessel diameter. Microcytic hypochromic red cell changes of iron deficiency do not cause a higher viscosity value at any given hematocrit value compared with normal red cells. However, in hypoxemic states oxygen-carrying capacity should be maximized, since the hemoglobin value is disproportionately lower at any given hematocrit in the presence of microcytic hypochromic cells compared with normal red cells.