AbstractBlood viscosityin vivo(“apparent viscosity”) and its variations with flow rate was analyzed in the maximally dilated calf muscle vascular bed of the cat by comparing pressure‐flow relationships for blood and a Newtonian fluid (dextran‐Tyrode) over a flow range between 60 and 0.2 ml/min × 100 g tissue. Viscosityin vitrofor the same perfusates was measured in a cone‐plate viscometer.—Apparent viscosity was much lower (approximately 50 %) thanin vitrovalues at high shear rates, with less variation between animals. It increased with decreased flow but was as a maximum only doubled, which occurred at flows around 0.5 ml/min × 100 g. Since such small flows normally occur in constricted vessels with higher flow velocities than at maximal dilatation, the range of viscosity changes with flow in the intact circulation is probably decidedly smaller. The steep rise of viscosityin vitroat quite low shear rates had no counterpartin vivo;in fact, viscosity then tended to fall again. The discrepancies between blood viscosityin vivoandin vitroseem to be related to vascular dimensions, favouring “bolus flow” and hence low regional viscosity in the most narrow vessels, which may become more pronounced at further luminal reduction, active or passive.Addition of high molecular weight dextran (HMD) raised apparent viscosity to seemingly high levels at low flows. However, to a considerable extent this appears to be due to cell aggregation and plugging of microvessels rather than to a genuine increase o