Revisiting classical experiments on the impact of air resistance on metabolic rate, we aimed to overcome limitations of previous research, notably: low participant numbers (n=1-3), highly turbulent wind, and confounding effects of rising body temperature. In a custom-built wind tunnel with reduced turbulence, 14 participants (8 males, 6 females) walked (5 km.h-1) and ran on a treadmill (70%V̇O2max) at 0, 2, 4 and 6 m.s-1 headwind or tailwind in a counterbalanced design, with rest-breaks between each exposure to avoid rises in body core temperature. Oxygen consumption (V̇O2) exhibited strong linear relationships versus wind direction, dynamic pressure and air speed squared(Vwr2), lower in magnitude for headwind than tailwind. A moderate linear relationship was observed between heart rate, wind direction, dynamic pressure and Vwr2. Below 4 m⸱s-1, the effect of wind was well within inter- and intra-individual variation and equipment uncertainty, and only at wind speeds ≥4 m⸱s-1 did the differences in physiological responses reach statistical significance. Our data indicate that at running speeds below 4 m⸱s-1 (14.4 km/h), indoor treadmill and outdoor running are comparable in terms of the metabolic impact of air movement relative to the person. However, this does not extend to the thermoregulatory effect of wind, with outdoor running providing a higher cooling rate due to the self-generated wind created during running. By removing the confounding impact of core temperature rises, the observed effects of headwind were lower and those of tailwind larger than observed previously. In the context of middle-distance running, headwind created by running at 21.5 km.h-1 would result in a 2.2% increase of V̇O2. A relative tailwind of the same speed would lead to a 3.1% reduction.