The cardiovascular response to exercise with several groups of skeletal muscle suggests that work with the arms may decrease leg blood flow. This study evaluated whether intense exercise with the legs would have a similar effect on arm blood flow (Y(arm)) and O(2) consumption (V(O(2))(,arm)). Ten healthy male subjects (age 21 +/- 1 year; mean +/- S.D.) performed arm cranking at 80 % of maximum arm work capacity (A trial) and combined arm cranking with cycling at 60 % of maximum leg work capacity (A + L trial). The combined trial was a maximum effort for 5-6 min. Y(arm) measurement by thermodilution in the axilliary vein and arterial and venous blood samples permitted calculation of V(O(2))(,arm). During the combined trial, Y(arm) was reduced by 0.58 +/- 0.25 l min(-1) (19.1 +/- 3.0 %, P < 0.05) from the value during arm cranking (3.00 +/- 0.46 l min(-1)). The arterio-venous O(2) difference increased from 122 +/- 15 ml l(-1) during the arm trial to 150 +/- 21 ml l(-1) (P < 0.05) during the combined trial. Thus, V(O(2))(,arm) (0.45 +/- 0.06 l min(-1)) was reduced by 9.6 +/- 6.3 % (P < 0.05) and arm vascular conductance from 27 +/- 4 to 23 +/- 3 ml min(-1) (mmHg)(-1) (P < 0.05) as noradrenaline spillover from the arm increased from 7.5 +/- 3.5 to 13.8 +/- 4.2 nmol min(-1) (P < 0.05). The data suggest that during maximal whole body exercise in humans, arm vasoconstriction is established to an extent that affects oxygen delivery to and utilisation by working skeletal muscles.