The metabolic cost of walking on an incline in the Peacock (Pavo cristatus)

Introduction

Locomotion is an integral part of every organism’s life, and requires a significant proportion of daily energy expenditure (Karasov, 1981; Pontzer & Wrangham, 2004). Understanding how locomotor activity influences the daily energy budget of an animal is important as it can provide insights into adaptations that have evolved to alleviate the costs of moving around (Carrier, 1987; Webb, 1984). Research into the adaptations that relate to locomotor performance has provided valuable insight into the overall costs of animal movement (Bramble & Carrier, 1983; Cavagna, Heglund & Taylor, 1977; Parker, Robbins & Hanley, 1984). Relatively few studies, however, have sought to investigate how locomotion up gradients may influence the overall energy budget.

Movement over non-level terrain is interesting as animals rarely move over consistently level surfaces, as natural terrain is uneven (Lees et al., 2013). Increases in speed usually account for elevations in the energetic cost of terrestrial locomotion. This increase occurs as when speed increases, the force required to move the COM must be generated in less time, which is achieved by recruiting a higher volume of muscle and/or faster-acting muscle fibres (Ellerby et al., 2003; Full, 1987; Kram & Taylor, 1990; Taylor, Schmidt-Nielsen & Raab, 1970). However, locomotion on an incline can also increase metabolic rate when compared to moving on level ground (Warncke, Bandholtz & Schultze-Motel, 1988; Wickler et al., 2000; Wunder & Morrison, 1974). The increased cost of incline locomotion has been explained by the additional mechanical workload required to move the centre of mass (CoM) against gravity (Roberts et al., 1997; Wickler et al., 2005). Increasing the gravitational potential energy of the CoM is achieved by either recruiting a larger volume of muscle to generate the required energy (Pierotti et al., 1989), or by increasing the active shortening, and therefore strain, of a smaller volume of muscle (Daley & Biewener, 2003; Gillis & Biewener, 2002).

In an attempt to unify the costs associated with incline locomotion, Taylor, Caldwell & Rowntree (1972) deduced the cost to lift one kilogram one metre vertically as $15.5\mathrm{J}{\mathrm{kg}}^{-1}{\mathrm{m}}_{\mathrm{v}}^{-1}$ across four mammalian species. However, subsequent empirical investigations have failed to find a similar cost of lifting (CoL) across a range of animals (Cohen, Robbins & Davitt, 1978; Farley & Emshwiller, 1996; Parker, Robbins & Hanley, 1984). The mass specific cost of incline locomotion has been found to be lower in larger animals, with efficiencies of metabolic energy use generally increasing with body mass (M_b) (Snyder & Carello, 2008; Tullis & Andrus, 2011). Disparity among studies investigating the relationship between M_b, speed and angle of inclination remains, with the relationship between efficiency (defined as the conversion of metabolic energy into mechanical work) and gradient also following no distinct pattern (Cohen, Robbins & Davitt, 1978; Warncke, Bandholtz & Schultze-Motel, 1988).

In terms of the energetic costs of locomotion broad similarities in the kinematics of terrestrial vertebrate groups are observed, but subtle differences in posture and gait give a proposed explanation for the species-specific variation (Nudds, Codd & Sellers, 2009; Reilly, McElroy & Biknevicius, 2007). Changing from a walking gait to an aerial running gait, where the in-phase relationship between E_p + E_kv (the sum of the gravitational potential and vertical kinetic energies of the CoM, used to raise the body’s CoM) and E_kh (the kinetic energy of the CoM used to reaccelerate the body for forward motion) results in no pendular exchange between potential and kinetic energy, is associated with lower energy economy (Heglund et al., 1982). However, the aerial running gait does afford some metabolic savings by increasing elastic potential energy storage in the tendons (Taylor, 1985). The stored elastic potential energy can be used to increase E_p + E_kv and E_kh and reaccelerate the CoM (Cavagna, Heglund & Taylor, 1977; Rubenson et al., 2004). Interestingly, bipedal birds often exhibit a grounded running gait (McMahon, Valiant & Frederick, 1987; Nudds et al., 2010). During grounded running, duty factor (DF, showing the ratio of limb contact with the ground during a stride) remains above 0.5 (Alexander, 2004), but E_p + E_kv and E_kh are in-phase (Heglund et al., 1982; Rubenson et al., 2004). Maintaining a higher DF increases the time available to generate the force required to raise the CoM (Roberts et al., 1998). The long tendons of birds are also stretched at low forces, allowing adequate elastic potential energy storage despite the in-phase relationship between E_p + E_kv and E_kh (Gatesy & Biewener, 1991; Rubenson et al., 2004; Taylor, 1985). Regardless of the apparent advantages, the effect of grounded running on the overall metabolic cost of locomotion, on level or gradient ground, has yet to be elucidated.

Studies investigating the effects of incline locomotion on kinematic parameters have found differences in DF (Dutto et al., 2004), stride frequency (Gillis & Biewener, 2002) and t_stance (the relative amount of time a foot is on the ground during a stride (Hoyt, Wickler & Cogger, 2000)) between level and incline locomotion. Force generation during t_stance is a direct measure of the metabolic cost of locomotion, and stride frequency has been directly associated with energetic cost, yet relatively few studies have investigated the metabolic cost of incline locomotion with the associated kinematic changes (Heglund et al., 1982; Lees et al., 2013; Nudds & Codd, 2012). Because the rate of force production is the parameter of interest, the inverse of contact time (1/t_stance) is used for analysis and gives a measure of force application (Kram & Taylor, 1990). Combining information about the energetic and kinematic responses to locomotion on changing gradients may therefore provide valuable insight into the species-specific differences observed.

Here we investigated the energetic and kinematic response to two grades of incline locomotion in the peacock. We hypothesise that the energetic cost of locomotion in the peacock is higher on an incline than on the level gradient, and that movement on gradients influences locomotor kinematics.

Materials and Methods

Results

Energetics

Net-P_met increased linearly with increasing U at the same incremental rate for each gradient (Incline × U, F_2,39 = 5.03, P = 0.08) and net-P_met increased with increasing gradient as indicated by differences between the intercepts of the slopes (Fig. 1A, LME: U, F_1,41 = 84.48, P < 0.001; incline, F_2,41 = 36.34, P < 0.001;). A Tukey post-hoc test indicated a significant difference at the 0.05 confidence level between net-P_met on the level and at each incline (level and 5°, Z = 5.19, P < 0.001; level and 7°, Z = 5.67, P < 0.001). No difference was found in net-P_met between the two inclines 5° and 7°(Z = − 0.73, P = 0.74). There was no significant effect of incline on resting P_met (ANOVA: F_2,15 = 3.13, P = 0.07).

Kinematics

DF decreased linearly with increasing U at the same incremental rate for each gradient (Incline × U, F_2,39 = 0.15, P = 0.93) and DF increased with increasing gradient as indicated by differences between the intercepts of the slopes (Fig. 2A, LME: U, F_1,41 = 80.74, P < 0.001; Incline, F_2,41 = 25.48, P < 0.001). A Tukey post-hoc test indicated a significant difference at the 0.05 confidence level between DF on the level and at each incline (level and 5°, Z = 2.94, P = 0.009; level and 7°, Z = 5.02, P < 0.001). No difference was found in DF between the two inclines 5°and 7°(Z = − 2.13, P = 0.08). DF remained above 0.5 in all treatments, which showed the birds did not reach an aerial running phase. The minimum DF recorded was 0.65 ± 0.02 at 1 m s⁻¹ when walking on the level gradient.

1/t_stance increased linearly with increasing U at the same incremental rate for each gradient (F_2,39 =1.93, P =0.38) and 1/t_stance decreased with increasing gradient as indicated by differences between the intercepts of the slopes (Fig. 2B, LME: U, F_1,41 = 1813.6, P < 0.001; Incline, F_2,41 = 35.37, P < 0.001). A Tukey post-hoc test indicated a significant difference at the 0.05 confidence level between 1/t_stance on the level and at each incline, as well as between the incline gradients (level and 5°, Z = − 3.55, P = 0.001; level and 7°, Z = − 5.9, P < 0.001; 5° and 7°, Z = 2.41, P = 0.04).

f_stride increased linearly with increasing U at the same incremental rate for each gradient (Incline × U, F_2,39 = 0.15, P = 0.93) and there were no differences in f_stride between the gradients (Fig. 2C, LME: U, F_1,41 = 129.61, P < 0.001; Incline, F_2,41 = 1.87, P = 0.39. l_stride increased linearly with increasing U at the same incremental rate for each gradient (Incline × U, F_2,39 = 1.5, P = 0.47) and there were no differences in l_stride between the gradients (Fig. 2D, LME: U, F_1,41 = 98.89, P < 0.001; Incline, F_2,41 = 1.87, P = 0.39).

t_swing decreased linearly with increasing U at the same incremental rate for each gradient (Incline × U, F_2,39 = 0.004, P = 1.00) and there were no differences in t_swing between the gradients (LME: U, F_1,41 = 110.89, P < 0.001; Incline, F_2,41 = 1.68, P = 0.43).

On the level gradient, E_kh and E_p + E_kv were out of phase indicating that the peacocks were using a walking gait at each of the three speeds (0.5 m s⁻¹, 0.75 m s⁻¹, 1 m s⁻¹). At a 7° incline, E_kh and E_p + E_kv were in-phase at 0.75–1.0 m s⁻¹, illustrating that two of the four birds analysed were grounded running at 1 m s⁻¹, with one of these birds grounded running at 0.75 m s⁻¹. Overall, the CoM results suggest that peacocks may change from a walking to a grounded running gait at lower speeds on an incline than on the level.

Discussion

The energetic cost of locomotion in the peacock is greater while moving on an incline gradient than on a level gradient, consistent with previous investigations in birds (Ellerby et al., 2003; Lees et al., 2013; Rubenson et al., 2006), reptiles (Farley & Emshwiller, 1996; Zani & Kram, 2008), mammals (Cohen, Robbins & Davitt, 1978; Eaton et al., 1995; Fancy & White, 1987) and some invertebrates (Full & Tullis, 1990; Tullis & Andrus, 2011). Energetic cost also increased linearly with U in each treatment, as in most studies, with only a few exceptions where the energetic cost of locomotion increases curvilinearly with U (Langman et al., 1995; Snyder & Farley, 2011; Steudel-Numbers & Wall-Scheffler, 2009; Wickler et al., 2005).

The elevated metabolic cost associated with incline locomotion has been attributed to increases in muscle activity required to raise the CoM against gravity, while accounting for the apparent reduction in elastic potential energy (Gabaldón, Nelson & Roberts, 2004; Roberts et al., 1997; Snyder & Farley, 2011). During incline movement, increased muscle shortening also raises metabolism as eccentric muscle contractions are replaced by more costly concentric muscle contractions (Daley & Biewener, 2003; Wickler et al., 2005). Concentric muscle contractions generate less tension than eccentric muscle contractions (Davies & Baknes, 1972; Proske & Morgan, 2001), resulting in the recruitment of a larger volume of muscle to generate the force required for incline locomotion (Gabaldón, Nelson & Roberts, 2008). Increased blood flow to stance phase muscles (Rubenson et al., 2006) and elevated stance phase muscle activity (Gabaldón, Nelson & Roberts, 2004) have also been found, which may further suggest that incline locomotion requires more positive muscular work than level movement.

We found no difference in the energetic cost of locomotion between the two incline gradients studied, although these differed only by 2 degrees. Previously, the energetic cost of locomotion has been shown to increase as incline rises (Eaton et al., 1995; Full, 1987; Raab, Eng & Waschler, 1976) or stay constant between inclines (Chassin et al., 1976; Lees et al., 2013; Lipp, Wolf & Lehmann, 2005), with no clear differences between bipedal, quadrupedal or hexapedal animals demonstrated. However, relating the energetic cost of locomotion to total force output is required to account for factors other than net mechanical work (e.g., isometric muscle contractions and muscle efficiency) influencing metabolic rate (Biewener, 1990; Roberts et al., 1997). Alterations in posture, gait, joint mechanics and footfall events can also affect the magnitude of force required for locomotion (Birn-Jeffery & Higham, 2014; Hesse et al., 2014; Lammers, Earls & Biknevicius, 2006; Reilly, McElroy & Biknevicius, 2007).

One of the most relevant investigations emphasised the importance of the rate of force application on the overall cost of transporting body weight. The comparative approach across a number of mammalian species found the larger the animal, the lower the rate of force application (due to a larger l_stride and lower f_stride allowing a larger t_stance) and subsequently, the lower the transport costs per gram of body weight (Kram & Taylor, 1990). The effects of incline on the energetic cost of locomotion are less well understood; however, it is apparent that the kinematic results may provide a clearer understanding of the metabolic costs associated with incline locomotion.

In a number of investigations, few kinematic parameter changes are found between horizontal and incline locomotion (Eaton et al., 1995; Full & Tullis, 1990; Lees et al., 2013). In the present study, however, peacocks maintained higher DF’s on an incline than on the level (Nudds & Codd, 2012). The increase in DF, and the decrease in 1/t_stance found when comparing level and incline movement (Oldruitenborgh-Oosterbaan, Barneveld & Schamhardt, 1997), indicated that the increased cost of incline locomotion in the peacock was not influenced by a higher rate of muscle force development (Kram & Taylor, 1990; Roberts et al., 1998). The lack of difference in t_swing between level and incline movement also suggested that the energetic cost of incline locomotion was not influenced by increased swing phase muscle activity to power more rapid foot movements (Heglund & Taylor, 1988; Roberts et al., 1997; Rubenson et al., 2006). When comparing the two inclines, DF did not differ considerably, but 1/t_stance was lower on a 7° incline than on a 5° incline. The lower rate of force development found at a 7° incline may compensate for some of the elevated cost associated with increasing incline. Here, a lower volume of muscle with slower rates of activity would be recruited, which may reduce the energetic cost of locomotion relative to the gradient ascended (Biewener, 1990; Heglund & Cavagna, 1987).

Although costly compared to level locomotion (Minetti, Ardigo & Saibene, 1994), it has been suggested that changing gait at lower speeds on an incline can afford some metabolic savings (Wickler et al., 2003), which may contribute to the lack of considerable difference in the energetic cost found between the two inclines. However, further analyses of the CoM mechanics and kinematic parameters with increasing U and incline are needed to determine whether peacocks change gait at lower speeds with increasing incline. Changing gait also has important physiological implications, such as reducing limb mechanical stress (Farley & Taylor, 1991), thereby allowing longer endurance and reduced risk of injury (Biewener & Taylor, 1986; Smith & Wilson, 2013). As bone and muscle structure is highly conserved (Biewener, 1991), changing to grounded running at lower speeds on an incline may have same effect of reducing peak muscle and bone stress as changing gait with U on the level (Perry et al., 1988).

The CoM results indicated variability in the U at which gait changes occurred in the peacock on both level ground and at an incline. Although preliminary in comparison with other studies, it has previously been demonstrated that there is overlap in the speeds used for walking, grounded running and aerial running and therefore great variability in the U at which a gait transition occurs (Nyakatura et al., 2012). This can be explained by the gradual transitional nature of gait changes in birds, whereby shifts from walking (vaulting) to aerial running gaits occur via vaulting-like and bouncing-like phases (Hancock, Stevens & Biknevicius, 2007). The higher variations seen in phase shifts from walking to running gaits at lower speeds have also been rationalised by the structural and functional locomotor requirements of an avian species. For example, for birds with a crouched posture, maintaining stiff limbs for walking is more costly than adopting a crouched grounded running gait (Nyakatura et al., 2012).

The MCoT values at each incline were considerably lower than the values reported for small animal groups (Farley & Emshwiller, 1996; Full & Tullis, 1990; Lipp, Wolf & Lehmann, 2005; Snyder & Carello, 2008). Similarly, the MCoT value for level movement was lower than the predicted value of 6.62 J kg⁻¹ m⁻¹ for a 4.6 kg bird (using the equation MCoT = $10.8M_b^{-0.32}$ Nudds et al., 2010). The MCoT values presented at the intermediate U, and the curvilinear relationship shown between MCoT and U, were consistent with previous findings in humans (Snyder & Farley, 2011) and horses (Wickler et al., 2000) moving on inclines. The MCoT was also higher on an incline compared to moving on the level, as found in a number of previous investigations (Farley & Emshwiller, 1996; Minetti et al., 2002; Warncke, Bandholtz & Schultze-Motel, 1988; Wunder & Morrison, 1974). The CoL at each incline did not differ considerably, corresponding with the results for net-P_met. The CoL found in both treatments was larger than that predicted (Taylor, Caldwell & Rowntree, 1972), but similar to the $27\mathrm{J}{\mathrm{kg}}^{-1}{\mathrm{m}}_{\mathrm{v}}^{-1}$ found in elk calves (Cohen, Robbins & Davitt, 1978). The efficiencies of converting metabolic energy into mechanical work were also similar between the two inclines, differing from the marked increases in efficiencies (Cohen, Robbins & Davitt, 1978; Lees et al., 2013), and the reduced efficiencies (Full & Tullis, 1990; Taylor, Schmidt-Nielsen & Raab, 1970; White & Yousef, 1978), previously found with increasing incline. However, further comparisons are difficult due to a large number of studies using only one incline gradient. Dissimilarities in other methodologies (e.g., incline gradients used) also highlight the difficulties in finding a consistent pattern of the costs of incline locomotion across species, emphasising the need for future work on incline locomotion.

The peacock has the highest efficiencies, in terms of converting metabolic energy into metabolic power, and lowest CoT values of bipedal birds previously studied (Ellerby et al., 2003; Lees et al., 2013; Snyder & Carello, 2008; Warncke, Bandholtz & Schultze-Motel, 1988), closely followed by the similarly-sized Marabou stork (Bamford & Maloiy, 1980). For comparison with other bipeds and quadrupeds, efficiency data for a number of species are shown in Fig. 3, where larger animals (1–200 kg) tend to have higher efficiencies than smaller animals (>1 kg). However, the relationship between size and efficiency is not linear, suggesting structural and postural differences play an important role in determining the costs and efficiencies of locomotion (Reilly, McElroy & Biknevicius, 2007). For example, being larger is associated with longer stride lengths and lower stride frequencies (Heglund et al., 1982; Heglund & Taylor, 1988; Kram & Taylor, 1990; Taylor, Heglund & Maloiy, 1982). Larger animals also tend to have straighter limbs, longer tendons with a smaller cross-sectional area compared to the muscles, and ungulate or digitigrade foot placement. Therefore, larger animals have less crouched locomotor positions, increased elastic potential energy savings and require lower force generation per unit mass than smaller, more crouched animals (Biewener, 1990; Pontzer, 2005; Reilly, McElroy & Biknevicius, 2007).

In summary, we have found that the energetic cost of locomotion is significantly higher on an incline than on the level ground, with changes in DF and 1/t_stance corresponding with the energetic results. Interestingly, the peacock had the highest efficiencies of locomotion of any bipedal bird studied, yet no other kinematic changes were observed with increasing incline. Taken together, our results accentuate the importance of combining energetic and kinematic responses, while also acknowledging allometric and phylogenetic differences, to gain a clearer understanding of the intrinsic costs associated with incline locomotion across the animal kingdom.

Supplemental Information