During exercise in severe acute or chronic hypoxia maximal cardiac output (Qmax) is reduced. With a lower Qmax, maximal systemic oxygen delivery is also reduced in hypoxia and hence, VO₂ max and exercise capacity. The reduction in Qmax is associated with a lower peak heart rate (HRpeak), particularly in chronic hypoxia. However, increasing HRpeak to nomoxic values with glycopyrrolate (a muscarinic antagonist) did not increase cardiac output above the value observed during maximal exercise in chronic hypoxia. Lowering the afterload by reducing blood viscosity by isovolemic haemodilution or by peripheral vasodilation (intra-arterial ATP infusion at peak exercise) had almost no influence on maximal exercise Q. Although blood volume is reduced during the first 2-3 months of residence at altitude, plasma volume expansion with 1 litre of 6% dextran did not cause changes in peak exercise capacity nor Q. Structural changes at the level of the myocardium can be excluded since sea level values of Qmax can be achieved in chronic hypoxia with hyperoxia. Thus, it seems that the reduction of Qmax in chronic hypoxia is caused by a neural mechanism that senses PaO₂ and downregulates the maximal pumping capacity of the heart. In agreement, we have observed that during submaximal exercise in acute hypoxia (FiO₂=0.11, 100 W) cardiac output is barely increased to counteract the effect of adenosine-induced hypotension (mean arterial pressure ~72 mmHg), despite the fact that an increase of exercise Q to a value 85-90% of Qmax would have allowed for a preserved mean arterial pressure.