Involvement of neural stimuli, central and/or peripheral in origin, in exercise ventilatory control was ascertained by examining the ventilatory responses to varying mechanical conditions of arm and leg cycle ergometries. Twelve untrained women underwent each of two modes of exercise at three levels of loading (0, 5 and 10 N), each at three levels of speed (30, 50 and 72 r.p.m.), during which steady-state values of minute ventilation (VE), tidal volume (VT), respiratory frequency (f) and CO2 excretion (VCO2) were measured. 2. Using the data obtained at the aerobic work intensities, the relationship of ventilatory responses (VE, VT and f) to the metabolic (VCO2) and mechanical (speed and load) variables were studied by multiple linear regression analysis. Coefficient of determination (r2) of the regression model was lowest (0.84) for f in the arm exercise and highest (0.99) for VE in the leg exercise. 3. Standardized partial regression coefficients of the model indicated that VE response is related to VCO2 at the rate of 94 +/- 3% (mean +/- S.E.M.) and to the pedal rate at 8 +/- 3% during the leg exercise, while it is closely related to VCO2 in the arm exercise. For f response, influence of the rate of limb movement was seen in the leg exercise but not in the arm exercise. The different effects of the rate of limb movement between the two exercise modes may be related to familiarity with the exercise modes, suggesting that a familiarity-related mechanism is involved in exercise ventilatory control. 4. A heavier load imposed on the limb muscles elicited a greater VT both in the arm and leg exercise and a lower f in the arm exercise. Postural control in the upper torso during increased limb muscle tension seems to affect VT and f.
