Journal of Theoretical Biology

Transcription

1 Journal of Theoretical Biology 296 (2012) 6 12 Contents lists available at SciVerse ScienceDirect Journal of Theoretical Biology journal homepage: Relating ranging ecology, limb length, and locomotor economy in terrestrial animals Herman Pontzer n Department of Anthropology, Hunter College, 728 North Building, 695 Park Ave, NY 10065, USA article info Article history: Received 29 July 2011 Received in revised form 11 November 2011 Accepted 17 November 2011 Available online 29 November 2011 Keywords: Foraging efficiency Locomotion Energetics Evolution abstract Ecomorphological analyses have identified a number of important evolutionary trends in vertebrate limb design, but the relationships between daily travel distance, locomotor ecology, and limb length in terrestrial animals remain poorly understood. In this paper I model the net rate of energy intake as a function of foraging efficiency, and thus of locomotor economy; improved economy leads to greater net energy intake. However, the relationship between locomotor economy and net intake is highly dependent on foraging efficiency; only species with low foraging efficiencies experience strong selection pressure for improved locomotor economy and increased limb length. Examining 237 terrestrial species, I find that nearly all taxa obtain sufficiently high foraging efficiencies that selection for further increases in economy is weak. Thus selection pressures for increased economy and limb length among living terrestrial animals may be relatively weak and similar in magnitude across ecologically diverse species. The Economy Selection Pressure model for locomotor economy may be useful in investigating the evolution of limb design in early terrestrial taxa and the coevolution of foraging ecology and locomotor anatomy in lineages with low foraging efficiencies. & 2011 Elsevier Ltd. All rights reserved. 1. Introduction The selection pressures shaping limb length in most terrestrial animal lineages are no doubt numerous and often conflicting. Parsing the effects of different pressures requires an understanding of the relationships between limb design, locomotor performance, and evolutionary fitness (Arnold, 1983). Ecomorphological approaches linking anatomy to performance, and performance to fitness, have identified several important ecomorphological relationships among terrestrial animals. For example, since leaping performance is positively related to hind limb length it follows that saltatory species, such as leaping primates (Demes and Günther, 1989), have relatively long hind limbs. Similarly, swimming economy is enhanced in shortlimbed, streamlined body shapes (Fish et al., 2001), which predominate among semi-aquatic taxa. One aspect of ranging ecology that has remained surprisingly resistant to ecomorphological analysis is the relationship between limb length and daily travel distance. Previous work has demonstrated that longer limbs improve walking and running economy in terrestrial animals (Kram and Taylor, 1990; Pontzer, 2007). Yet species that travel farther each day, and would thus appear to benefit the most from greater economy, have similar limb lengths and locomotor costs to species with comparatively modest day n Tel.: þ ; fax: þ address: hpontzer@hunter.cuny.edu ranges (see Section 1.1). In this paper, I briefly discuss the effect of limb length on the energy cost of locomotion, and develop a simple model of the marginal effects of changing limb length on net energy intake, an indirect measure of fitness. I then use the model to examine the relationship between daily travel distance, locomotor economy, limb length, and fitness. Finally, I discuss the model s predictions in the context of observed limb lengths and ranging ecology in terrestrial animals Ranging ecology, locomotor economy, and limb length All organisms require energy to live and reproduce. For terrestrial animals, acquiring energy typically involves travel to find plant foods or prey. This travel incurs an energy cost, C (J/m), which must be outweighed, over the long term, by the rate at which energy is acquired, B (J/m) in order for an animal to meet its energy requirements. That is, for foraging to provide a net energy benefit, B must exceed C, and the net rate of energy return per distance traveled, E net (J/m) is E net ¼ðB2CÞ Over the course of a day, expected total net energy return, E net total (J/day) is a function of the average rate of return, (B C), and the distance traveled, D (m/day), such that E nettotal ¼ DðB2CÞ ð1þ ð2þ /$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi: /j.jtbi

2 H. Pontzer / Journal of Theoretical Biology 296 (2012) In principle, the net rate of energy gain ought to affect fitness by determining the energy available for growth, maintenance, and reproduction. This expectation is supported by studies examining the link between energy intake and reproductive output both within- and between-species. Among mammals, the rate of reproductive output, measured as grams of offspring produced per year, scales with maternal mass 0.75, as expected if maternal energy throughput is the primary determinant of reproductive output (Charnov and Ernest, 2006). Within species, studies in the wild (Altmann, 1991) and in captivity (Perrigo, 1987; Bautista et al., 1998; Deerenberg et al., 1998; Knott, 2001; Wiersma and Verhulst, 2005; Vaanholt et al., 2007) indicate that reproductive investment and output are positively correlated with net energy intake. Given the observation that net energy intake is positively correlated with reproductive output, Eqs. (1) and (2) predict that selection will target traits that increase D and B and decrease C. Indeed, there is some evidence that evolutionary increases in daily travel distance, D, are positively associated with reproductive output among mammals, as predicted by Eq. (2) (Pontzer and Kamilar, 2009). Similarly, while foraging strategies are complex and multidimensional, contingent on food availability and the counterstrategies of competitors, work in optimal foraging theory, specifically tests of Charnov s Marginal Value Theorem (Charnov, 1976), indicates that species generally pursue foods that improve the net rate of energy return (Stephens and Krebs, 1986). In contrast, there is little evidence to suggest that selection has acted to decrease locomotor cost, C, in species that range widely or appear to experience lower rates of return. One clear target of selection for improved locomotor economy (i.e., the distance traveled per unit energy spent on locomotion) would be limb length. Work over the past two decades, including recent experimental and comparative studies, has shown that limb length is one of four major anatomical determinants of locomotor cost in terrestrial animals 1 (Kram and Taylor, 1990; Pontzer, 2005, 2007; Roberts et al., 1998; Pontzer et al., 2009a). The metabolic cost of walking and running primarily derives from the volume of muscle activated each step to support body weight (Kram and Taylor, 1990; Pontzer, 2005; Pontzer et al., 2009a). Animals with longer limbs activate less muscle volume to support body mass during locomotion, and therefore expend less energy to cover a given distance (Kram and Taylor, 1990; Pontzer, 2005; Pontzer et al., 2009a). In fact, limb length is the primary anatomical determinant of locomotor economy in terrestrial animals, from ants to elephants (Pontzer, 2007). The mechanical work done to lift the body s center of mass (Minetti et al., 1999; Ruina et al., 2005) and to swing the limbs (Marsh et al., 2004) also contributes to cost, but does a relatively poor job predicting metabolic cost within or between species (Minetti et al., 1999; Full and Tu, 1991). Despite the importance of limb length in determining locomotor cost there is little empirical evidence suggesting that locomotor economy or limb length have been a primary target of selection in taxa that range widely. For example, despite their reduced digits and long metapodials, ungulates (artiodactyls and horses) are no more economical than generalized mammals (Taylor et al., 1982; Fig. 1). Similarly, despite the fact that carnivores travel an average of four times farther than herbivores each day (Garland, 1983; Carbone et al., 2005), the cost of locomotion (J/m) for carnivores is no different than that of other mammals (Taylor et al., 1982; Fig. 1), and limb lengths of carnivores are not exceptional; for example, a 25 kg goat and a 25 kg dog have similar limb lengths, about 40 cm 1 The others are body mass, the effective mechanical advantage of the limb joints, and the fascicle length of limb muscles (Roberts et al., 1998; Pontzer et al., 2009). These, along with limb length, determine the volume of muscle activated to support bodyweight during locomotion, and hence metabolic cost. Fig. 1. Locomotor Cost Among Terrestrial Mammals. The energy cost of locomotion, measured via oxygen consumption during treadmill trials and expressed as the mass-specific Cost of Transport (J kg 1 m 1 ), for terrestrial mammals (Taylor et al., 1982). Oxygen consumption was converted to Joules assuming 20.1 J/mlO 2.Trendline is for all species (n¼64; excludes kangaroos). ANCOVA with log-transformed data revealed that locomotor costs for carnivores and ungulates (artiodactyls and horses) are no different than for other mammals (F(62)¼1.05, p¼0.36). A test for heterogeneity of slopes similarly indicated that the relationship between body mass and cost does not differ among these groups (F(62)¼1.79, p¼0.18). (Pontzer, 2007). Within carnivores, average daily travel distance, D, is not correlated with limb bone length (Harris and Steudel, 1997). Given the evolutionary changes in limb length evident in saltatatory and semi-aquatic lineages it seems unlikely that the lack of increased limb length and locomotor economy in cursorial species is due to a genetic constraint in the vertebrate bauplan. Instead, the observation that limb length and locomotor economy are unrelated to ranging distance suggests that selection for improved locomotor economy, and thus increased limb length, is weak relative to other pressures on locomotor anatomy, even among species with long day ranges and low rates of energy acquisition. Weak selection on locomotor economy and limb length might be expected if other pressures on locomotor anatomy were exceptionally high. Alternatively, selection for improved locomotor economy in terrestrial animals might have pushed limb lengths to the point of diminishing returns long ago. In this case, selection pressure for further gains in locomotor economy and limb length in extant species is low, and other selection pressures acting on limb length prevail. Below, I model the marginal selection pressure for improved locomotor economy in terrestrial animals, and argue that selection for decreased locomotor cost is weak among modern taxa. I test this hypothesis using ranging data from extant mammals and lizards. Results suggest that modern species experience a sufficiently high rate of return that further improvement in locomotor economy would have a negligible effect on fitness, such that selection for further increase in limb length is held in check. I discuss the implications of the model for the evolution of limb length and the relationship between ranging ecology and locomotor performance in terrestrial animals. 2. Modeling selection pressure on locomotor economy and limb length As shown in Eq. (1), the net rate of energy intake per unit distance, E net,isequalto(b C). When holding daily travel distance,

3 8 H. Pontzer / Journal of Theoretical Biology 296 (2012) 6 12 D, constant,e net determines the net rate of energy acquired per day (Eq. (2)), and hence the energy available for maintenance, reproduction, and other critical activities. Thus, in principle, species can alter their net energy intake by changing D or E net. Previous work has examined the energetic consequences of evolutionary changes in D (Pontzer and Kamilar, 2009); the Economy Selection Pressure model here examines changes in locomotor economy on E net. In the simple case where C changes over evolutionary time for a species while B remains constant, change in E net will be equal to (B C 2 ) (B C 1 ), where C 1 is the initial locomotor cost and C 2 is the evolved locomotor cost. This term simplifies as DE net ¼ C 1 2C 2 ð3þ where DE net is the change in net intake per distance traveled. Let M be the percentage increase in locomotor economy, such that evolutionary decreases in locomotor cost result in positive changes in M, M ¼ðC 1 2C 2 Þ=C 1 ð4þ Combining Eqs. (3) and (4) gives DE net ¼ C 1 M ð5þ We can then express the change in E net as a percentage of its initial value, (B C 1 ), %DE net ¼ C 1M ð6þ ðb C 1 Þ The term (B C 1 ) can be rewritten as C 1 (B:C 1 1), where B:C 1 is the initial ratio of B over C, commonly referred to as foraging efficiency (Ydenberg et al., 1994). Eq. (6) can then be rewritten and simplified as M %DE net ¼ ð7þ ðb:c 1 1Þ This expression (Eq. (7)) provides the framework for the Economy Selection Pressure model for selection acting on locomotor economy. Changes in net energy intake, E net, are directly proportional to changes in locomotor economy, M. However, the slope of this relationship is itself a function of a species current foraging efficiency, B:C 1. For species with low foraging efficiencies, changes in locomotor economy will have relatively large effects on net energy intake, but in species with high foraging efficiencies, changes in locomotor economy will have relatively small effects on net energy intake. For example, following Eq. (7), for a species obtaining a foraging efficiency of B:C¼5, a 20% increase in economy will result in a 5% increase in E net, whereas for a species obtaining B:C¼41, a 20% increase in economy would result in only a 0.5% increase in E net. The shape of the relationships between B:C, M, and E net over a range of values is shown in Fig Modeling selection pressures on limb length in two hypothetical lineages As briefly discussed above, locomotor cost C is a function of limb length for terrestrial animals, decreasing with greater limb length as (limb length) 0.77 (Pontzer, 2007). Consider two species of equal body mass and limb length, one in which the ratio of B:C¼40, and one in which B:C¼12. In both species, a 10% increase in limb length is expected to improve locomotor economy (i.e., decrease cost, C) by 7%. As a result, given their respective foraging efficiencies, selection pressure for increased limb length is relatively low, less than 1% for both species (Fig. 3; Eq. (7)). Notably, the selection pressure for increased limb length in these two lineages is not a function of daily travel distance, D. If daily energy requirements (J/day) are similar for both species, as expected given their similar body mass, we expect D to be more than Fig. 2. Economy Selection Pressure Model. The change in net energy intake per distance traveled, %DE net, is a linear function of the change in locomotor economy, M. However, the slope of this function is highly dependent on the initial foraging efficiency, B:C 1. For species obtaining foraging efficiencies less than B:C 1 ¼10, small changes in economy have large effects on E net. For species obtaining foraging efficiencies greater than B:C 1 ¼30, even large changes in economy of 50% result in a less than 2% change in E net. Fig. 3. Economy Selection Pressure Model Applied to Limb Length. At time-i, species X obtains a foraging efficiency of B:C 1 ¼40; species Y obtains B:C 1 ¼12. The selection gradient for increased limb length, indexed as %DE net, is low for both species at time-i, as indicated by trajectories Xi and Yi. Over evolutionary time, B decreases by 50%, such that at time-ii B:C 1 ¼20 for species X and B:C 1 ¼6 for species Y. The selection gradient for increased limb length remains flat for species X, but becomes much steeper for species Y, as indicated by trajectories Xii and Yii. The effect of limb length on locomotor cost was calculated following Pontzer, (2007). 3.5-times greater in the low-b:c species in order to achieve similar E net total given its lower rate of return (Eq. (2)). But despite this substantial difference in D, the selection pressure for increased limb length is very small and of similar magnitude in both species (Fig. 3). Now consider a scenario in which the rate of return, B, is cut in half for both species over evolutionary time through a change in habitat or diet. For the first species, this halving of B lowers B:C to

4 H. Pontzer / Journal of Theoretical Biology 296 (2012) The selection pressure on locomotor economy and limb length remains low; a 10% increase in limb length would yield only a 0.4% increase in E net (Fig. 3). In contrast, for the second species, B:C is now 6, and improvements in locomotor economy will have a relatively large effect on E net and the energy available for reproduction. Selection pressure to improve locomotor economy, and therefore to increase limb length, will be relatively high; a 10% increase in limb length would yield a 1.4% increase in E net (Fig. 3). Selection for increased limb length and improved locomotor economy become dramatically higher as foraging efficiency drops below B:C¼10 and approaches B:C¼1 (Fig. 3). This hypothetical comparison may explain the lack of correspondence between ranging ecology and locomotor economy in terrestrial animals. In the first instance, when B:C for the two species were 40 and 12, neither experienced particularly strong selection on limb length and locomotor economy. Nonetheless, their foraging ecologies would appear quite different, with a nearly four-fold difference in the rate of return and a 3.5-fold difference in D, equivalent to the differences between modern carnivores and herbivores (Ruina et al., 2005; Marsh et al., 2004). Despite these ecological differences, limb lengths and locomotor economy would likely remain similar between the two species, since the selection pressure to improve economy would be minimal in both. Over evolutionary time, as foraging ecology for these two species changed and decreased B by 50%, the effect on their selection regimes would be quite different. In the high-b:c lineage (denoted X in Fig. 3), the selection pressure for improved economy would remain low, while in the low-b:c lineage (denoted Y in Fig. 3), selection for greater economy and increased limb length would become relatively strong. The difference in selection pressures between lineages would persist even if, as expected, both lineages double their daily travel distance to compensate for lower return rates. As a result, phylogenetically controlled studies examining the co-evolution of foraging ecology and locomotor economy would be unlikely to detect a clear signal; changes in daily travel distance would be correlated with changes in limb length and locomotor economy only in the (possibly few) lineages where B:C is exceptionally low. The Economy Selection Pressure model (Eq. (7); Fig. 2) does not predict that selection pressure for improved economy and increased limb length will ever reach 0. Instead, the model proposes that the magnitude of selection on economy and limb length will change dramatically as a function of B:C. As a result, the prevailing selection pressures acting on locomotor anatomy will depend on a species current ecology and evolutionary history. While even small fitness advantages can, in principle, have significant effects on morphology, in practice one must consider the relative strengths and combined effects of multiple selection pressures acting on a given trait (Arnold, 1983). As one selection pressure decreases, the relative strength of others will increase (Arnold, 1983). Thus, for species with B:C in excess of 10, further selection for increased limb length might be outpaced by competing pressures on limb growth and design. Costs of increased limb length, which are not considered in the present model, include the time and energy spent to grow and maintain longer limbs, as well as the potential for decreased performance in other aspects of locomotion such as acceleration (Biewener, 1989), or in other locomotor modes such as climbing or swimming (Fish et al., 2001; Pontzer and Wrangham, 2004). 3. Testing the Economy Selection Pressure model One prediction of the Economy Selection Pressure model is that extant terrestrial animals should obtain foraging efficiencies in excess of B:C¼10. Below B:C¼10, the model predicts that selection acting to maximize net energy will be relatively strong, and will act to improve locomotor economy (e.g., increase limb length) or to increase the rate of food acquisition (e.g., through a change in diet to a more abundant food source). Since the major terrestrial vertebrate Orders have all had 65 million years or more (much more, if one traces lineages back to the Devonian) of evolution on land, the Economy Selection Pressure model predicts that their foraging ecology and locomotor anatomy will have produced a foraging efficiency greater than B:C¼10 long ago. To test this prediction, data on daily travel distance, D, was collected from the literature for nine reptile and 228 mammal species (Garland, 1983, 1999; Carbone et al., 2005; Harris and Steudel, 1997; Appendix 1). Body mass was used to calculate the estimated Field Metabolic Rate, FMR (J/day) for each species using the All Lizards equation (reptiles) or the Eutherian Mammal equation (mammals) from a comprehensive FMR review (Nagy et al., 1999). Since FMR must equal gross metabolized food energy intake (J/day) over the long-term for weight-stable animals, estimated FMR was divided by observed D (m/day) for each species to provide an estimate of B (J/m). Estimating B in this way implicitly assumes that all ranging is directed toward foraging. This simplification is consistent with other analyses of ranging behavior in mammals (Carbone et al., 2005), but will have the effect of systematically underestimating B, since non-foraging travel is included in measurements of D for most, if not all, species. However, since the model predicts relatively high values of B relative to C, estimating B in a manner that likely underestimates its true value is a conservative approach. Estimating B using this manner also ignores short-term variance in the rate of return, including variance resulting from patchily distributed resources. Instead, measures of D and B are long-term (i.e., over multiple days) averages. For a small number of species (3 reptiles, 8 mammals), measurements of FMR and D were both available (Garland, 1983, 1999; Carbone et al., 2005; Harris and Steudel, 1997 ; Nagy et al., 1999). For these species, B was calculated by dividing observed FMR by D. For all species, body mass was used to calculate C using the general all-taxa equation for running in Taylor et al. (1982), C¼10.7Mass and the minimum walking cost equation in Rubenson et al. (2007), C¼17.8Mass Foraging efficiency, B:C, was then estimated for both running and walking estimates of C. These estimates of B:C were then compared to determine the effect that gait choice (walking or running) would have on net energy return. In addition to the error inherent on estimates of B, described above, the use of allometric equations to calculate C introduces additional error into estimations of B:C, since species vary in how closely they fit the allometric relationship (see Fig. 1). However, as with any least squares trendline, error in C is expected to be normally distributed about 0, and should not skew estimates of foraging efficiency in either direction. It should also be noted that estimates of C based on laboratory treadmill experiments (Taylor et al., 1982; Rubenson et al., 2007) may be somewhat lower than the cost of travel across the uneven terrain and varied substrates encountered by animals in their natural habitats; this issue is discussed below. Body mass, reported daily travel distance, and estimated foraging efficiency are given in Appendix A Results When C was calculated using the allometric equation for running (Minetti et al., 1999), nearly every species examined (94%) had an estimated foraging efficiency greater than B:C¼10, with most in excess of B:C¼40. Of the 237 species included in the analysis, only 14 (13 mammals, 1 reptile) had an estimated

5 10 H. Pontzer / Journal of Theoretical Biology 296 (2012) 6 12 B:Co10 (Fig. 4). One hundred and forty nine species (63%) in the dataset had an estimated foraging efficiency greater than B:C¼40. Estimated foraging efficiency was negatively correlated with body mass (n¼237, r 2 ¼0.07, po0.01), but this correlation was weak; the allometric exponent was low ( 0.13) and only 7% of the variation in B:C was explained by size (Fig. 4). Mean and median estimated foraging efficiencies for the entire dataset were and 54.0, respectively. These estimates were somewhat higher for mammals (141.0, 55.6) than reptiles (50.4, 27.0), but this apparent difference may well be spurious. The error induced by the allometric estimation of B and C, and the small size of the reptile sample, make statistically meaningful comparisons impossible. Without direct, empirical measures of B and C for any species in the dataset it is not possible to dissect foraging efficiencies of individual species with much certainty. However, there is reason to believe that extreme estimates of foraging efficiency, both high and low, result from compounded error in estimation. Four mammals (three rodents, one hyrax; Appendix A) had estimated efficiencies in excess of B:C¼1000; the largest was That these species acquire more than 1000 J of food energy for every Joule spent foraging is certainly possible, but it is likely that estimates of B or C for these species produce unreasonably high estimates of foraging efficiency. Similarly, four species (one reptile, three mammals) have estimated foraging efficiencies below B:C¼5.0. Such low estimates of foraging efficiency might result from the conservative approach taken here to estimate B, which includes non-foraging travel in estimating the net rate of return (see above). These low estimates might also result from underestimation of food intake. The lowest estimated foraging efficiency (B:C¼2.9) was seen in the spotted hyena, Crocuta crocuta. However, estimated FMR (18,382 kj/day) was far less than observed food intake in the wild (44,160 kj/day) (Smith et al., 2008). When observed food intake is used to calculate B, estimated foraging efficiency more than doubles, to B:C¼6.9. It is possible that intake is similarly underestimated in other seemingly low-efficiency species, which are predominantly carnivores (Appendix A), but direct observation of food acquisition or FMR will be needed to test this. Despite the errors associated with estimation of B:C, data from species in which observed FMR and D are available suggest that the estimates here are in the correct range. For these eleven species, median B:C¼16.2 (mean: 137.4, range: ). Using measured FMR instead of estimated FMR to calculate B had the effect of raising the lowest estimates and lowering the highest estimates of B:C, as expected (the range of estimated B:C for this group was when estimated FMR is used to calculate B). Here again, direct measures of B and C for more species would provide a more comprehensive test of the model. These results also appear to be robust to variation in locomotor cost due to differences in gait or substrate. Mean and median estimated foraging efficiencies calculated using walking cost (86.4, 42.4) were slightly below those calculated using running cost, because over three-quarters (77%) of the species in this dataset have a body mass below 20 kg. Animals below 20 kg have estimated walking costs that are greater than running costs (see allometric equations above, and Rubenson et al. (2007)) leading to higher estimates of C and hence lower estimates of B:C. This effect is particularly noticeable among reptiles in this sample, which are all small-bodied. The number of reptile species with estimated B:Co10 increases from one to four when estimated walking cost is used; all four weigh less than 40 g. By comparison, increasing estimated locomotor cost to account for the additional costs of traveling over uneven terrain and variable substrates has a modest effect on results. Increasing running-based locomotor costs by 15% for all species (n¼237) decreased mean and median B:C by 13%. Future work might examine whether smaller animals use running gaits rather than walking as a means of improving net energy gain (Rubenson et al., 2007), and how lab-based measures of locomotor cost compare to travel costs encountered in natural habitats. 4. Discussion Fig. 4. Foraging Efficiencies in Terrestrial Animals (A) estimated foraging efficiencies for terrestrial mammals (black) and reptiles (gray). Species with measured, rather than estimated, FMR values are indicated with a þ. Dashed line indicates B:C¼10. (B) Histogram of estimated foraging efficiency, B:C, for 228 mammals (black) and 9 reptiles (gray). The bar labeled 440 includes all species with B:C440. Foraging ecology is known to have strong effects on limb design in terrestrial animals. Non-human primates and other arboreal species have grasping hands and feet that are useful in the trees (Fleagle, 1998); semi-aquatic species have shortened limbs that appear to improve swimming performance (Fish et al., 2001); felids have retractable claws that improve their ability to capture prey. Yet despite the primary importance of limb length in determining locomotor economy and foraging efficiency, there appears to be little correspondence between limb length and daily travel distance in terrestrial animals. Species with the longest day ranges do not have the longest limbs or the most economical locomotion (Taylor et al., 1982; Harris and Steudel, 1997).

6 H. Pontzer / Journal of Theoretical Biology 296 (2012) The Economy Selection Pressure model for locomotor economy outlined here provides an explanation for this surprising observation. Given that several selection pressures are likely acting on limb length in all species, the model presented here predicts that selection for improved economy and greater limb length will only prevail when foraging efficiency is quite low, generally below B:C¼10 (Fig. 2). Since most taxa obtain efficiencies well in excess of this benchmark, there is little reason to expect species or population differences in ranging ecology to be reflected in limb length. The estimated energetic advantage of further increase in limb length is quite low for nearly all species examined (Fig. 4). The approach taken here uses a different currency for foraging cost and benefit than many previous studies, focusing on the energy return or cost per unit distance rather than per unit time (Charnov, 1976; Ydenberg et al., 1994). In principle this is a relatively minor difference, as the cost of locomotion can be expressed as energy expended per unit time or per unit distance (Taylor et al., 1982). Indeed, Eq. (7), the framework for the model present here, would not change if B, C, ande net were expressed as energy/time rather than energy/distance. However, in practice, it is simpler to calculate the energy cost per distance while traveling, because calculating the energy cost per unit time requires information on the organism s rate of travel (Pontzer, et al.). Regardless of the currency used, by linking anatomical design to locomotor performance and foraging ecology, the model presented here may provide a useful bridge between current biomechanical investigations of locomotor anatomy and established models of foraging ecology such as Charnov s Marginal Value Theorem (Charnov, 1976). The Economy Selection Pressure model presented here was developed and tested in living terrestrial species, but it may also shed light on the evolutionary pressures shaping vertebrate anatomy and foraging strategies as these lineages transitioned from water to land. The ranging ecologies and foraging efficiencies of extinct taxa will always remain somewhat speculative, but there is some evidence to suggest that early terrestrial species were under strong selection to decrease locomotor costs and increase the rate the food acquisition as they transitioned to life on land. Semiaquatic transitional forms likely had short limbs relative to modern terrestrial taxa (Daeschler et al., 2006) that would have improved their swimming economy but decreased their terrestrial economy (Fish et al., 2001; Pontzer, 2007). As terrestrial travel grew to account for a greater proportion of travel, selection for increased limb length and lower cost would have followed. As a result, selection pressures on early terrestrial taxa may have long ago pushed limb lengths to the proportions common among modern terrestrial species. The similarity in limb lengths (relative to body mass) across a broad range of taxa, including extinct reptiles (Fig. 5), suggests limb lengths increased rapidly before stabilizing at the proportions seen in most species today. This pattern of early change followed by stasis in limb proportions is expected if selection for increased locomotor economy reaches a point of diminishing returns, as proposed by the model presented here (Fig. 2). The Economy Selection Pressure model provides an explanation for the lack of correspondence between daily travel distance and limb length in terrestrial species; most species obtain such high foraging efficiencies that selection for further economy is negligible. Conversely, the model provides a framework for examining the coevolution of foraging ecology and limb design in species and lineages with low foraging efficiencies. Empirical measures of return rates and locomotor cost in species with low foraging efficiencies are needed to test how selection for economy has increased limb length in some lineages. These data may improve reconstructions of evolutionary history in lineages that have experienced increases in limb length, including our own (Pontzer et al., 2009a, 2010 ). Fig. 5. Limb Lengths Among Extant and Extinct Terrestrial Species. Data from Pontzer, (2007) and Pontzer et al. (2009b). Trendline only includes living taxa. Acknowledgments I thank Carel van Schaik, Jennifer E. Smith, Campbell Rolian, David Raichlen, and four anonymous reviewers for comments and conversations that improved this manuscript. Support was provided by Hunter College. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at doi: /j.jtbi References Altmann, S.A., Diets of yearling female primates (Papio cynocephalus) predict lifetime fitness. Proc. Natl. Acad. Sci. USA 88, Arnold, S.J., Morphology, performance, and fitness. Am. Zool. 23, Bautista, L.M., Tinbergen, J., Wiersma, P., Kacelnik, A., Optimal foraging and beyond: how starlings cope with changes in food availability. Am. Nat. 152, Biewener, A.A., Scaling body support in mammals: limb posture and muscle mechanics. Science 245, Carbone, C., Cowlishaw, G., Isaac, N.J.B., Rowcliffe, J.M., How far do animals go? Determinants of day range in mammals. Am. Nat. 165, Charnov, E.L., Optimal foraging: the marginal value theorem. Theor. Populat. Biol. 9, Charnov, E.L., Ernest, S.K.M., The offspring-size/clutch-size trade-off in mammals. Am. Nat. 167, Daeschler, E.B., Neil, H., Shubin, N.H., Jenkins Jr, F.A., A Devonian tetrapodlike fish and the evolution of the tetrapod body plan. Nature 440, Deerenberg, C., Overkamp, G.J.F., Visser, G.H., Daan, S., Compensation in resting metabolism for experimentally increased activity. J. Comp. Physiol. B 168, Demes, B., Günther, M.M., Biomechanics and allometric scaling in primate locomotion and morphology. Folia Primatol. 53, Fish, F.E., Frappell, P.B., Baudinette, R.V., MacFarlane, P.M., Energetics of terrestrial locomotion of the platypus Ornithorhynchus anatinus. J. Exp. Biol. 204, Fleagle, J.G., Primate Adaptation and Evolution, second ed. Academic Press, NY. Full, R.J., Tu, M.S., Mechanics of rapid running insects: two-, four-, and six-legged locomotion. J. Exp. Biol. 156, Garland Jr, T., Laboratory endurance capacity predicts variation in field locomotor behaviour among lizard species. Anim. Behav. 58, Garland, T., Scaling the ecological cost of transport to body mass in terrestrial mammals. Am. Nat. 121, Harris, M., Steudel, K., The selection of hind limb length in the Carnivora: the influence of daily movement distance, home range area, prey size, latitude, and prey capture method. J. Zool. London 241, Knott, C.D., In: Ellison, P.T. (Ed.), Reproductive Ecology and Human Evolution, Aldine de Gruyter, New York, pp Kram, R., Taylor, C.R., Energetics of running: a new perspective. Nature 346,

7 12 H. Pontzer / Journal of Theoretical Biology 296 (2012) 6 12 Marsh, R.L., Ellerby, D.J., Carr, J.A., Henry, H.T., Buchanan, C.I., Partitioning the energetics of walking and running: swinging the limbs is expensive. Science 303, Minetti, A.E., Ardigo, L.P., Reinach, E., Saibene, F., The relationship between mechanical work and energy of locomotion in horses. J. Exp. Biol. 202, Nagy, K.A., Girard, I.A., Brown, T.K., Energetics of free-ranging mammals, reptiles, and birds. Ann. Rev. Nutr. 19, Perrigo, G., Breeding and feeding strategies in deer mice and house mice when females are challenged to work for their food. Anim. Behav. 35, Pontzer, H., A new model predicting locomotor cost from limb length via force production. J. Exp. Biol. 208, Pontzer, H., Limb length and the scaling of locomotor cost in terrestrial animals. J. Exp. Biol. 210, Pontzer, H., Wrangham, R.W., Climbing and the daily energy cost of locomotion in wild chimpanzees: implications for hominoid locomotor evolution. J. Hum. Evol. 46, Pontzer, H., Kamilar, J.M., Great ranging associated with greater reproductive investment in mammals. PNAS USA 106, Pontzer, H., Raichlen, D.A., Sockol, M.D., 2009a. The metabolic cost of walking in humans, chimpanzees, and early hominins. J. Hum. Evol. 56, Pontzer, H., Allen, V., Hutchinson, J.R., 2009b. Biomechanics of running indicates endothermy in bipedal dinosaurs. PLoS ONE 4, e7783. Pontzer, H., Rolian, C., Rightmire, G.P., Jashashvili, T., Ponce de León, M.S., Lordkipanidze, D., Zollikofer, C.P.E., Locomotor anatomy and biomechanics of the Dmanisi hominins. J. Hum. Evol. 58, Pontzer, H., Raichlen, D.A., Sockol, M.D., From treadmill to tropics: calculating ranging cost in chimpanzees. In: D Aout, K., Vereecke, E.E. (Eds). Studying Primate Locomotion: Linking Laboratory and Field Research. New York, Springer. pp Roberts, T.J., Chen, M.S., Taylor, C.R., Energetics of bipedal running: II. limb design and running mechanics. J. Exp. Biol. 201, Rubenson, J., Heliams, D.B., Maloney, S.K., Withers, P.C., Lloyd, D.G., Fournier, P.A., Reappraisal of the comparative cost of human locomotion using gaitspecific allometric analyses. J. Exp. Biol. 210, Ruina, A., Bertram, J.E.A., Srinivasan, M., A collisional model of the energetic cost of support work qualitatively explains leg sequencing in walking and galloping, pseudo-elastic leg behavior in running and the walk-to-run transition. J. Theor. Biol. 237, Smith, J.E., Kolowski, J.M., Graham, K.E., Dawes, S.E., Holekamp, K.E., Social and ecological determinants of fission-fusion dynamics in the spotted hyena. Anim. Behav. 76, Stephens, D.W., Krebs, J.R., Foraging Theory. Princeton University Press, Princeton. Taylor, C.R., Heglund, N.C., Maloiy, G.M.O., Energetics and mechanics of terrestrial locomotion: I. Metabolic energy consumption as a function of speed and body size in birds and mammals. J. Exp. Biol. 97, Vaanholt, L.M., De Jong, B., Garland Jr., T., Daan, S., Visser, H., Behavioural and physiological responses to increased foraging effort in male mice. J. Exp. Biol. 210, Wiersma, P., Verhulst, S., Effects of intake rate on energy expenditure, somatic repair and reproduction of zebra finches. J. Exp. Biol. 208, Ydenberg, R.C., Welham, C.V.J., Schmid-Hempel, R., Schmid-Hempel, P., Beauchamp, G., Time and energy constraints and the relationships between currencies in foraging theory. Behav. Ecol. 5,