Scaling the peak and steady-state aerobic power of running and walking humans

Publication Name

European Journal of Applied Physiology


Purpose: The first aim of this experiment was to evaluate the appropriateness of linear and non-linear (allometric) models to scale peak aerobic power (oxygen consumption) against body mass. The possibilities that oxygen consumption would scale allometrically across the complete metabolic range, and that the scaling exponents would differ significantly between basal and maximal-exercise states, were then evaluated. It was further hypothesised that the scaling exponent would increase in a stepwise manner with elevations in exercise intensity. Finally, the utility of applying the scaling exponent derived for peak aerobic power to another population sample was evaluated. Methods: Basal, steady-state walking and peak (treadmill) oxygen-consumption data were measured using 60 relatively homogeneous men (18–40 year; 56.0–117.1 kg), recruited across five mass classes. Linear and allometric regressions were applied, with the utility of each scaling method evaluated. Results: Oxygen consumption scaled allometrically with body mass across the complete metabolic range, and was always superior to both ratiometric analysis and linear regression. The scaling exponent increased significantly from rest (mass0.57) to maximal exercise (mass0.75; P < 0.05), but not between steady-state walking (mass0.87) and maximal exercise (P > 0.05). When used with an historical database, the maximal-exercise exponent successfully removed the mass bias. Conclusion: It has been demonstrated that the oxygen consumption of healthy humans scales allometrically with body mass across the entire metabolic range. Moreover, only two scaling exponents (rest and exercise) were required to produce mass-independent outcomes from those data. Accordingly, ratiometric and linear regression analyses are not recommended as scaling methods.

Open Access Status

This publication is not available as open access

Funding Sponsor

Defence Science and Technology Group



Link to publisher version (DOI)