During the last decade, the cellular and plasmatic factors as well as fluid-dynamic conditions responsible for the 'anomalous' rheological behavior of blood have been clarified. Hypotheses about the hemodynamic significance of the hemorheological factors in microvascular perfusion can be formulated and have now been subjected to in vivo tests. Red cells have bipotential flow properties: as a consequence of passive, fluid-drop-like adaptation of RBC and membrane 'tank-treading' rapidly flowing mammalian blood has properties akin to an emulsion with extremely high fluidity - while slowly flowing blood becomes aggregated and assumes the properties of a concentrated reticulated suspension of low fluidity. In addition, in the normal microcirculation the fluidity of blood is maintained by rheological and vasomotor factors despite the fact that the cells are larger than the capillaries. Here, fluidity of blood flow thence critically depends a. on the presence of a minimum of shear forces sufficient to disperse aggregates and deform single cell, b. on the biochemical integrity of the vascular content which is prerequisite for normal RBC deformability, c. on the factors that keep microvascular hematocrit low. Under abnormal biophysical and/or biochemical conditions slowly flowing blood may suffer local loss of fluidity, under conditions of fluidity being abolished without coagulation. In hypoperfused microvascular beds, the variability of flow conditions, as well as the resulting variability of flow behavior is submitted as the cause of non-homogeneous perfusion ('collateral blood viscidation'). Hemorheological active drugs (such as pentoxifylline) cannot improve the deformability (fluidity) of normal erythrocytes, but they can protect the cells against the effect of rigidifying metabolites such as lactate, pH-drop, hyperosmolarity etc. Thus, they can maintain blood fluidity and thence residual flow despite abnormal flow behavior and flow conditions in hypoperfused organs. In vivo results corroborate the theory that factors which increase macroscopic blood viscosity in vitro are of small significance in the normal circulation (where they are likely to be compensated by vasodilatation). However, these same factors are critical to perfusion in the abnormal or decompensated circulation in the absence of adequate flow forces and following the recruitment of the limited vasodilator reserve of vital organs.