We tested the hypothesis that pharmacological reduction of the increase in cardiac output during dynamic exercise with a large muscle mass would influence the cerebral blood velocity/perfusion. We studied the relationship between changes in cerebral blood velocity (transcranial Doppler), rectus femoris blood oxygenation (near-infrared spectroscopy) and systemic blood flow (cardiac output from model flow analysis of the arterial pressure wave) as induced by dynamic exercise of large (cycling) vs. small muscle groups (rhythmic handgrip) before and after cardioselective beta 1 adrenergic blockade (0.15 mg kg-1 metoprolol i.v.). During rhythmic handgrip, the increments in systemic haemodynamic variables as in middle cerebral artery mean blood velocity were not influenced significantly by metoprolol. In contrast, during cycling (e.g. 113 W), metoprolol reduced the increase in cardiac output (222 +/- 13 vs. 260 +/- 16%), heart rate (114 +/- 3 vs. 135 +/- 7 beats min-1) and mean arterial pressure (103 +/- 3 vs. 112 +/- 4 mmHg), and the increase in cerebral artery mean blood velocity also became lower (from 59 +/- 3 to 66 +/- 3 vs. 60 +/- 2 to 72 +/- 3 cm s-1; P < 0.05). Likewise, during cycling with metoprolol, oxyhaemoglobin in the rectus femoris muscle became reduced (compared to rest: -4.8 +/- 1.8 vs. 1.2 +/- 1.7 mumol L-1, P < 0.05). Neither during rhythmic handgrip nor during cycling was the arterial carbon dioxide tension affected significantly by metoprolol. The results suggest that as for the muscle blood flow, the cerebral circulation is also affected by a reduced cardiac output during exercise with a large muscle mass.