Role of locomotor economy in the origin of bipedal posture and gait

It was recently argued (Leonard and Robertson, 1995, 1997) that there are significant differences between human males and females in the energetic cost of locomotion (or transport), and that human females in particular reap a great energetic savings as compared to a typical quadruped. This purported difference is used to bolster the idea that the energetic advantages of human, particularly female, locomotion might have been important in the origin of bipedality. Discussions of this topic rely heavily on measurements of the energetic cost of locomotion in a variety of mammals compiled by a number of authors. Work on this subject is based on the fact that ongoing submaximal locomotion is usually powered almost entirely by ATP generated by aerobic pathways. Experimenters consequently measure the rate of oxygen consumed by a subject under a variety of locomotor conditions, often on a treadmill. Protocols usually contain elements that ensure that locomotion will be fueled aerobically (e.g., Taylor et al., 1982). Often the cost of a given locomotor activity is simply reported in ml O2, or, alternately, converted to Joules or Kcal. The cost of an activity can be expressed as a rate, called the “cost of locomotion,” i.e., the cost to engage in the activity for a particular amount of time. It is also common to report instead the cost to travel a particular distance, called the “cost of transport.” It has been demonstrated that the amount of energy that an animal consumes in traveling is highly dependent on its mass (for a summary, see Taylor et al., 1982). Larger animals typically expend a greater total number of calories because they are doing more work (moving a greater mass). On the other hand, the cost per kilogram of body mass is typically smaller in a larger animal, and this reduction has been well-documented (e.g., Taylor et al., 1982). Because body mass is rarely a constant in most interesting comparisons of locomotor energetics, nearly all studies involving comparisons of locomotor costs have reported the cost per kilogram. Leonard and Robertson (1995, 1997) cite data that they believe have been “recently synthesized by the World Health Organization” (FAO/WHO/UNU, 1985). In fact, the findings on locomotor energetics in that publication are based exclusively on a much earlier synthesis (McDonald, 1961, see footnote on p. 184). That early work had the benefit of a very large sample size. Its disadvantage, however, is that the sample was drawn from a wide variety of still earlier works in which a number of methodologies were used. Some studies used Douglas bags for collecting all expired gases (closed circuit), while others used open circuit techniques. Some protocols involving Douglas bags required subjects to carry equipment while their cost of locomotion was determined; in other protocols, the equipment was carried in some other manner (McDonald, 1961). Thus the data from the many previous studies were not always comparable. As McDonald (1961) points out, it was not always clear whether the weights given for subjects were for the subject alone or whether it included the weight of the equipment carried by some subjects in some of the protocols. Thus the potential for error was substantially greater than in a study conducted under a single experimental protocol in a single laboratory. Leonard and Robertson (1995, 1997) appear to be unaware that there has been considerable discussion in the 40 years since the work by McDonald (1961) on the question of gender differences in locomotor energetics. Numerous studies have been done using experimental designs very much more carefully controlled than that of McDonald (1961), in that male and female subjects in each study had their costs measured using the same experimental protocol. These have resulted in other conclusions. The vast majority of recent studies on gender differences in the cost of human walking have reported that there is no difference between the sexes in the cost/kg (Bhambani and Singh, 1985, 12 males, 12 females; Miller and Stamford, 1987, 4 males, 3 females; Waters et al., 1988, 39 adult males, 34 adult females; Pivarnik and Sherman, 1990, 12 males, 12 females; Zamparo et al., 1992, 6 males, 3 females; Sherman, 1998, 22 males, 22 females). These studies