There is no convincing evidence that metabolic CO₂ production rate (QCO₂) during muscular exercise of moderate intensity is directly controlled: its rate of change, rather, reflects the influence of the controlled rates of O₂ consumption (QO₂) and the substrate mixture being catabolised. Its response to exercise of moderate intensity (i.e. below the lactate threshold, θ_L) is therefore mono-exponential with a time constant (τ) similar to that for QO₂. Following a delay that incorporates the muscle-lung vascular transit time, the change in QCO₂ is expressed at the lungs (VCO₂) with an appreciably-longer τ (e.g. Whipp, 1987) as a result of the influence of the intervening high-capacitance CO₂ stores: (a) a transient alkalosis caused by proton trapping during phosphocreatine hydrolysis and (b) increased muscle tissue and muscle-venous PCO₂ and bicarbonate concentration ([HCO₃^–]). This capacitative effect results in a transient decrease of the respiratory exchange ratio (R) which, with very rapid work-rate incrementation and/or prior CO₂-stores depletion (by volitional or anticipatory hyperventilation) can yield a false positive non-invasive estimation of θ_L (‘pseudo-threshold’) (e.g. Ozcelik et al., 1999). Above θ_L, the components of the VCO₂ kinetics are far more complex, reflecting: (a) the translation of the non-linear pulmonary O₂ uptake (VO₂) slow component (e.g. Whipp, 1987) into a corresponding VCO₂ reponse; (b) production of supplemental CO₂ from the rate (rather than the amount) at which muscle and blood [HCO₃^–] decrease consequent to buffering of the H⁺ associated with the [lactate] increase (the amount of CO₂ evolved from these buffering reactions is more than double that of the equivalent aerobic yield); and (c) the time course of the compensatory hyperventilation for the metabolic acidaemia. For heavy-intensity exercise (i.e. the work-rate range within which arterial [lactate] and [H⁺] can be stabilized at constant, although elevated, levels), VCO₂ kinetics often evidence an overshoot before subsequently stabilising (Ozyener et al., 2002). This early overshoot reflects the influence of the rapid phase of the stores wash-out of CO₂ consequent to a rapidly-falling [HCO₃^–], but with little or no early recruitment of compensatory hyperventilation. Despite augmented peripheral chemoreceptor stimulation of ventilation (V_E) at these work rates (from increased [H⁺] and [K⁺], for example), PaCO₂ is typically slightly elevated (e.g. 3-4 mm Hg) during the initial on-transient phase, suggesting that the kinetics of respiratory compensation for the acidosis are long, relative to those of peripheral chemoreceptor responsiveness (Rausch et al., 1991). At higher work rates (for which arterial [lactate] and [H⁺] increase throughout the exercise to the limit of tolerance), VCO₂ kinetics have been shown to revert to a mono-exponential-like form (Casaburi et al., 1989; Ozyener et al., 2002), but are slower than for sub-θ_L exercise. Interestingly, no VCO₂ ‘slow phase’ is evident, despite this being clearly discernible in VO₂. This apparent steady state-like behaviour of VCO₂ at these intensities is associated with the offsetting effects of a slowing of the rate of [HCO₃^–] decrease and the progressive hyperventilatory decline in PaCO₂. In conclusion, therefore, VCO₂ kinetics are not consistently mono-exponential above the lactate threshold, but even when so, they should not be considered reflective of simple-compartment dynamics. Rather, the profile of VCO₂ conflates the influences of the differing rates of HCO₃^– breakdown and degrees of compensatory hyperventilation with that of the underlying aerobic component.