Duffin and Mateika (Duffin J. & Mateika, 2013) state that their rebreathing data could also be fitted to a parabola. If they consider a parabolic shape an indication of hyper‐additive interaction, it is surprising that they did not compare the quality of the fits of linear regression vs. parabolic or hyperbolic fits in order to decide which interaction mode would fit their data best. Apart from this, we have doubts about several assumptions underlying the modified rebreathing technique : 1. the absence of carotid body activity in hyperoxia; 2) that the hypoxic response would be a modified acidic response; 3. due to variable changes in CBF during the manouever, the tissue‐arterial PCO2 relationship cannot be constant (Battisti‐Charbonney et al., 2011); and 4. absence of cortical influences on ventilation following five minutes of voluntary hyperventilation. This article is protected by copyright. All rights reserved.