The aim is to quantify the physiological cost of head-load carriage and to examine the 'free ride' hypothesis for head-load carriage in groups of women differing in their experience of head-loading. Twenty-four Xhosa women [13 experienced head-loaders (EXP), 11 with no experience of head-loading (NON)] attempted to carry loads of up to 70% of body mass on both their heads and backs whilst walking on a treadmill at a self-selected walking speed. Expired air was collected throughout. In a second study nine women, members of the British Territorial Army, carried similar loads, again at a self-selected speed. Maximum load carried was greater for the back than the head (54.7 +/- 15.1 vs. 40.8 +/- 13.2% BM, P < 0.0005). Considering study one, head-loading required a greater oxygen rate than back-loading (10.1 +/- 2.6 vs. 8.8 +/- 2.3 ml kg bodymass(-1) min(-1), P = 0.043, for loads 10-25% BM) regardless of previous head-loading experience (P = 0.333). Percentage changes in oxygen consumption associated with head-loading were greater than the proportional load added in both studies but were smaller than the added load for the lighter loads carried on the back in study 1. All other physiological variables were consistent with changes in oxygen consumption. The data provides no support for the 'free ride' hypothesis for head-loading although there is some evidence of energy saving mechanisms for back-loading at low speed/load combinations. Investigating the large individual variation in response may help in identifying combinations of factors that contribute to improved economy.