We tested the hypothesis that reflexes arising from working respiratory muscle can elicit increases in sympathetic vasoconstrictor outflow to limb skeletal muscle, in seven healthy human subjects at rest. We measured muscle sympathetic nerve activity (MSNA) with intraneural electrodes in the peroneal nerve while the subject inspired (primarily with the diaphragm) against resistance, with mouth pressure (PM) equal to 60 % of maximal, a prolonged duty cycle (TI/TTot) of 0.70, breathing frequency (fb) of 15 breaths min-1 and tidal volume (VT) equivalent to twice eupnoea. This protocol was known to reduce diaphragm blood flow and cause fatigue. MSNA was unchanged during the first 1-2 min but then increased over time, to 77 +/- 51 % (s.d.) greater than control at exhaustion (mean time, 7 +/- 3 min). Mean arterial blood pressure (+12 mmHg) and heart rate (+27 beats min-1) also increased. When the VT, fb and TI/TTot of these trials were mimicked with no added resistance, neither MSNA nor arterial blood pressure increased. MSNA and arterial blood pressure also did not change in response to two types of increased central respiratory motor output that did not produce fatigue: (a) high inspiratory flow rate and fb without added resistance; or (b) high inspiratory effort against resistance with PM of 95 % maximal, TI/TTot of 0.35 and fb of 12 breaths min-1. The heart rate increased by 5-16 beats min-1 during these trials. Thus, in the absence of any effect of increased central respiratory motor output per se on limb MSNA, we attributed the time-dependent increase in MSNA during high resistance, prolonged duty cycle breathing to a reflex arising from a diaphragm that was accumulating metabolic end products in the face of high force output plus compromised blood flow.