- Open Access
Scale effects and morphological diversification in hindlimb segment mass proportions in neognath birds
© Kilbourne; licensee BioMed Central Ltd. 2014
- Received: 18 October 2013
- Accepted: 25 April 2014
- Published: 6 May 2014
In spite of considerable work on the linear proportions of limbs in amniotes, it remains unknown whether differences in scale effects between proximal and distal limb segments has the potential to influence locomotor costs in amniote lineages and how changes in the mass proportions of limbs have factored into amniote diversification. To broaden our understanding of how the mass proportions of limbs vary within amniote lineages, I collected data on hindlimb segment masses – thigh, shank, pes, tarsometatarsal segment, and digits – from 38 species of neognath birds, one of the most speciose amniote clades. I scaled each of these traits against measures of body size (body mass) and hindlimb size (hindlimb length) to test for departures from isometry. Additionally, I applied two parameters of trait evolution (Pagel’s λ and δ) to understand patterns of diversification in hindlimb segment mass in neognaths.
All segment masses are positively allometric with body mass. Segment masses are isometric with hindlimb length. When examining scale effects in the neognath subclade Land Birds, segment masses were again positively allometric with body mass; however, shank, pedal, and tarsometatarsal segment masses were also positively allometric with hindlimb length. Methods of branch length scaling to detect phylogenetic signal (i.e., Pagel’s λ) and increasing or decreasing rates of trait change over time (i.e., Pagel’s δ) suffer from wide confidence intervals, likely due to small sample size and deep divergence times.
The scaling of segment masses appears to be more strongly related to the scaling of limb bone mass as opposed to length, and the scaling of hindlimb mass distribution is more a function of scale effects in limb posture than proximo-distal differences in the scaling of limb segment mass. Though negative allometry of segment masses appears to be precluded by the need for mechanically sound limbs, the positive allometry of segment masses relative to body mass may underlie scale effects in stride frequency and length between smaller and larger neognaths. While variation in linear proportions of limbs appear to be governed by developmental mechanisms, variation in mass proportions does not appear to be constrained so.
- Stride Frequency
- Limb Segment
- Trait Change
- Positive Allometry
- Terrestrial Locomotion
The relative proportions of limb segments are one of the most conspicuous aspects of whole limb morphology. In terms of segment lengths, the proportions of limbs have been extensively studied in major amniote groups, including mammals[1–5], non-avian dinosaurs[3, 6–10], pterosaurs, birds[6, 8, 12–15], lizards[16–19], and turtles. Within these groups, the relative lengths of limb segments have been linked to specializations for predominant habitat[12, 15, 20], biomechanical demands[3–5, 7, 21], and functional diversity[6, 8].
Though bone masses have been studied in mammals and birds, previous studies on masses of whole limb segments inclusive of both hard and soft tissues have focused primarily on ungulates and primates[24–33], otherwise receiving little attention. Thus, it remains unknown how size influences changes in the mass proportions of limbs within most amniote lineages. Yet the relative masses within limb segments are likely critical to terrestrial locomotion. Less massive distal limb segments give the limb a more proximal concentration of mass and, consequently, a reduced cost to swing[35–38]. Savings in the metabolic cost of swinging the limbs may be of high importance in terrestrial locomotion, as the swinging of limbs can account for as much as 24% of the total metabolic energy expended. However, the morphology of distal limb segments, and perhaps consequently their relative mass, may be strongly influenced by functions apart from terrestrial locomotion. The morphology of distal limb segments can be specialized for functions as varied as swimming, climbing, prey capture, and digging while still being able to meet the demands of terrestrial locomotion[12, 40]. It therefore remains possible that functions of the limb outside of terrestrial locomotion influence the mass of the distal segments. The scaling of limb segment masses and their potential to influence locomotor costs merits investigation.
Cubo & Casinos found that femoral, tibiotarsal, and tarsometatarsal mass all scale with positive allometry relative to body mass (the authors did not study phalangeal mass). Thus, the skeletal contribution to segment mass is positively allometric. Yet it remains unknown whether soft tissue mass scales in parallel to bone mass. Thus a discrepancy between the scaling of bone mass and total segment mass may exist. In light of this, studying the scaling of segment mass can provide first insights into whether segment mass is tightly coupled to bone mass.
Diversification of hindlimb morphology
Here I test for departures from isometry with regards to the scaling of the masses of the femoral, tibiotarsal, tarsometatarsal segments and digits relative to body mass and hindlimb length, which are measures of body size and hindlimb size, respectively. My sample of neognath birds encompasses a high diversity in terms of hindlimb morphology and function. By using such a sample, I aim to discern general trends in the scaling of segment mass in neognath birds. Negative allometry of segment masses, particularly those of distal segments would indicate changes in mass proportions beneficial for locomotor economy of larger neognaths. By using an ecologically diverse, taxonomically wide sample of neognath birds, I will also make initial inferences into how segment masses have diversified in this lineage. Through use of branch length transformation models, I will investigate models of trait change in relation to branch lengths of the phylogeny. Study of the scale effects in the mass proportions of the neognath hindlimb can serve as a platform for future work on the scaling of limb and limb segment design in birds in general and other tetrapod clades. Likewise, applying models of trait evolution to limb morphology may serve as a means for understanding how morphological changes in the locomotor system enable diversification of amniote clades.
Sampled neognath taxa, following the taxonomy of Hackett et al.
Body mass (g)
Hindlimb length (cm)
Thigh mass (g)
Shank mass (g)
Pes mass (g)
Tars. mass (g)
Digit mass (g)
Falco peregrinus 1
Corvus brachyrhynchos 1
Accipiter striatus 1
Buteo jamaicensis 1
Bubo virginianus 1,2
Cardinalis cardinalis 1
Cathurus guttatus 1
Zonotrichia albicollis 1
Scolopax minor 1
Ardea herodias 1
Pavo cristatus 1
Branta canadensis 1
Data were collected from specimens slated to be prepared as skeletal specimens and stored wholly intact and frozen in airtight bags in deep freezers as they awaited preparation. Though specimens were not freshly dead (i.e., < 24 hours dead), the use of airtight bags of the specimens insured against desiccation/freeze-drying. Inspection of specimens after thawing and manipulation of limb joints also prevented use of desiccated specimens.
Prior to cutting the limb into segments, hindlimb length was measured in its passively flexed position. To determine the passively flexed length of the limb, the limb was manually extended to its maximum and then allowed to flex on its own accord. This particular method of determining limb length was chosen given the known differences in posture between smaller and larger bodied birds. As smaller bodied birds maintain a crouched, more flexed limb posture, and larger bodied birds maintain an upright, less flexed posture, measuring the passively flexed length of the limb takes into account these differences in posture and does not distort the data for small or large bodied neognaths. Note, however, that my measure of hindlimb length specifically reflects size-dependent differences in posture and is not an attempt to mimic in vivo limb movements or standing posture for each species.
Scaling relationships were assessed using Reduced Major Axis (RMA) model II bivariate regression. Prior to regression analysis, data consisting of species means were log-transformed. RMA regressions were my chosen method of analysis, as Model II regressions assume that both x and y variables contain some degree of error (either measurement errors and/or biological variation) and are not independent in the strict sense[53, 54]. Moreover, RMA regressions are ideal for testing slope values against null model predicted values. Additionally, to identify how segment masses co-diversified against body mass and hindlimb length, I also separately performed phylogenetic generalized least squares (PGLS) regressions (see below).
Under isometric scaling, segment mass should scale as (body mass)1.0 and (hindlimb length)3.0. Log transformation of the data changes the scaling relationship from its normal power function expression of y = a(x)b into a linear linear function: log(y) = log(a) + log(b)(x). Thus, according to isometry, the regression slope should be 1.0 and 3.0 when scaling against body mass or hindlimb length, respectively[56, 57]. To test for departures from isometry, two methods were used. The first method was an F-test to test whether the empirical value significantly deviates from isometry’s prediction, with deviations being significant if P < 0.05. The second method, utilizing effect size statistics, employed 95% confidence intervals for the regression slope. If slope confidence limits exclude the predicted value, then isometry was rejected. F-tests were performed and confidence intervals were calculated in R version 2.15.1 using the module SMATR. Through a combination of effect size and significance value based statistics, an increase in Type I error rates due to repeated testing bias is avoided, as is a decrease in statistical power due to Bonferroni corrections.
For both the entire Neognath sample and Land Bird subsample, differences in slope and intercept (i.e., ‘elevation’ following the terminology of Warton et al.) of RMA regressions were identified by using common slope tests and Wald’s test, respectively. If P < 0.05, then differences in regression slope and intercept were considered significant. Tests for common slope and intercept were performed in the R module SMATR.
Species data is not independent due to hierarchically structured phylogenetic relationships among species. As such, conventional statistical methods are not suited for estimating evolutionary models of trait evolution or inferring the evolutionary processes that produce empirical trait values. For segment masses, body mass, and hindlimb length, I tested two models of trait evolution by transforming branch lengths with Pagel’s λ and Pagel’s δ.
λ is a branch length transformation that models the dependence of observed trait variation on phylogenetic relationships of a given tree[48, 62]. It should be noted that λ is a often used as a direct measure of phylogenetic signal – the tendency for increased phenotypic similarity with increasing phylogenetic relatedness[62, 63] – within each trait. A multiplicative factor of a tree’s internal branches, λ of 0.0 indicates a complete absence of phylogenetic signal and that traits evolved independently among the individual sampled taxa; in contrast, a λ of 1.0 indicates that traits evolved by constant-rate Brownian motion along the branches of the tree[48, 64]. In theory, a value of 1.0 indicates that rates of trait change have remained constant across the tree; however, inferring a relationship between phylogenetic signal and rates of trait change is highly problematic and should be avoided.
δ is a branch length transformation that models whether rates of trait change are greater towards the root or the tips of the tree[48, 65], acting as a multiplicative factor of both shared and internal branches lengths on the tree. δ > 1.0 indicates that more recent evolution within a clade has had a greater influence on trait diversification. In contrast, δ < 1.0 indicates that early evolution within a clade has a had a greater influence upon trait diversification. δ = 1.0 indicates that a trait diversified under a model of Brownian motion and the branch lengths remain unchanged. It is important to note that δ represents only a monotypic increase or decrease in rates of trait change across the tree. In all likelihood though, rates of trait change differ amongst the different branches of the tree, and there are existing methods to check for such differing rates (e.g., auteur:). However, given my sample size, my data is poorly suited to methods such as auteur, which is ideally suited by datasets and phylogenies with at least ~ 60 taxa. In spite of this, using a δ transform can reveal whether rates of change in segment masses are not monotypic along the tree.
To test whether trait diversification in terms of both λ and δ departed from a Brownian motion model, 95% confidence intervals were generated for both of these parameters for each trait studied, with an exclusion of 1.0 indicating a departure from Brownian motion. λ , δ, and accompanying confidence limits were estimated using the module pmc (Phylogenetic Monte Carlo;) in R.
Furthermore, the fit of each model was compared using a Monte Carlo-based method in pmc. First, the likelihood ratio was calculated as the difference between the log likelihood of observing the data under maximum likelihood models of λ and δ. Then under the λ model, a given trait was simulated as evolving along the specified phylogeny over 1000 iterations. For each iteration, a λ and δ model were fit to the data and the likelihood ratio between the two fits was calculated. From the 1000 iterations, a distribution of likelihood ratios was calculated with a 95% confidence interval. If the confidence interval excluded the observed likelihood ratio, here acting as a critical value, then the λ model is rejected (i.e., the observed likelihood ratio is not the result of applying both the λ and δ models to a trait that has evolved in line with a λ model). This procedure is then repeated simulating a given trait as evolving under a δ model – a likelihood ratio distribution and accompanying 95% confidence interval are generated by applying the two models to simulated data evolving under Pagel’s δ. As in the test of Pagel’s λ, the observed likelihood ratio is used as a critical value in combination with the confidence limits to test this second model of trait evolution. For a more detailed explanation of the pmc method, see Boettiger et al.. Given that the pmc module only allows pair-wise comparisons of models, I compared each model to a Brownian motion model using confidence limits for λ and δ as described above, whereas to directly to compare these two models and test their fit of the data, I used the Monte Carlo-based methods in pmc.
I also applied the models of Pagel’s λ and δ to residuals from bivariate generalized least squares regressions of segment masses against body mass and hindlimb length. Even though individual traits may follow a given evolutionary model, it does not necessarily guarantee that the traits have co-diversified under such a model. To determine if segment mass traits have co-diversified with measures of body and limb size in line with the two trait diversification models, I generated 95% confidence intervals for λ and δ and used the Monte Carlo based method of Boettiger and colleagues to test the fit of these models to regression residuals. In addition to Monte Carlo-based methods of model fit, I also repeated each bivariate regression for the entire neognath sample as a Phylogenetic Generalized Least Squares (PGLS) regression. Performing PGLS regressions can illuminate whether the co-diversificaiton of segment masses and measures of size has been either allometric or isometric.
To test for diversification models for each trait, I used the phylogeny of Jetz et al. with internal nodes based upon the phylogeny of Hackett et al.. A tree consisting of the sampled species was generated using the website http://birdtree.org. Branch lengths were based upon divergence times in absolute time.
Body mass scaling
Results of regressions of segment masses against body mass and hindlimb length for the entire neognath sample
Body mass scaling
Hindlimb length scaling
Results of regressions of segment masses against body mass and hindlimb length for the Land Bird subsample
Body mass scaling
Hindlimb length scaling
Limb length scaling
Regarding the entire neognath sample, segment masses scale isometrically with hindlimb length (Figure 4B and Table 2). Slopes ranged between 3.01 (thigh segment) and 3.24 (tarsometatarsal segment). Among the limb segments, the slopes do not differ (P = 0.8433); however, as when scaling against body mass, the proximal and distal pairs of limb segments differ in intercept (P < 0.0001). Yet, as is the case scaling when against body mass, the segments comprising each pair do not differ in slope from one another (P > 0.05).
For the Land Bird subsample, segment masses scale isometrically or with positive allometry when scaled against hindlimb length (Figure 5B and Table 3). Slopes range from 3.06 to 3.54. Thigh segment and digit mass are isometric with hindlimb length, not significantly differing from a slope of 3.0, whereas shank segment and tarsometatarsal segment mass are both positively allometric (Table 3). However, directly comparing slopes across limb segments finds no significant difference in slope (P = 0.4780). Post hoc tests uncover that the thigh and shank segments significantly differ in intercept from the tarsometatarsal segment and digits (P < 0.0001). However, the thigh and shank segments do not differ in slope (P = 0.9950), just as the tarsometatarsal segment and digits do not (P = 0.8396).
Models of trait diversification fit to each trait
λ C. L.
Models of trait co-diversification fit to each trait alongside body mass or hindlimb length
λ C. L.
Body mass scaling
Hindlimb length scaling
Regarding the co-diversification of segment masses alongside limb length, likelihood ratio confidence limits do not reject either model (Table 5). Both λ (with the exception of the thigh) and δ indicate that phylogeny significantly influences variation in residuals with confidence intervals excluding a value of 0.0. Neither λ and δ can reject a model of Brownian motion for the co-diversification of segment masses with limb length, as all confidence intervals include a value of 1.0.
Results of PGLS regressions
Body mass scaling
Hindlimb length scaling
All segment masses scale with positive allometry relative to body mass, whereas they scale isometrically relative to hindlimb length. Thus, limb segment masses do not scale with negative allometry or with increasingly lower scaling exponents distally along the limb, which would reduce the cost of swinging the limbs in larger neognath species. Rather, scale effects of individual hindlimb segments parallel the scaling of whole hindlimb mass relative to body mass (i.e., positive allometry) and hindlimb length (i.e., isometry). In light of these results, the scale effects in the hindlimb’s mass proportions do not afford a lowered cost of swinging the limbs in neognath birds with respect to increasing size. However, the differences in regression elevation indicate that the pes and its constituent segments have less mass than the more proximal segments (in absolute terms) for a given body mass or limb length (Figures 4 and5). Consequently, absolute differences in mass between proximal and distal segments result in limbs with a lowered cost of swinging compared to limbs with a more even distribution of mass between proximal and distal segments.
The scaling of segment masses differs from the scaling of segment lengths, which are determined by limb bone lengths. Amongst the limb segments, only the lengths of the tibiotarsus and tarsometatarsus scale with positive allometry relative to body mass; the lengths of the femur and digit III in contrast scale with isometry[12, 14, 15, 44, 45]. Thus, the scaling of limb segment mass is not necessarily tied to the scaling of limb segment length. In contrast – and perhaps not surprisingly – scale effects in segment mass may be more strongly tied to the scaling of hindlimb bone mass. Much like the masses of their respective segments, the masses of the femur, tibiotarsus, and tarsometatarsus all scale with positive allometry relative to body mass. The positive allometry of tibiotarsal and tarsometatarsal mass is likely due to the relatively greater lengths of these long bones in larger avian species, whereas the positive allometry of femoral mass is likely due to the relatively greater femoral robusticity in larger avian species[44, 69]. It also worth noting that cross-sectional area and second moment of area of these three elements all scale with positive allometry[14, 70], which also likely contributes to the positive allometry of bone mass and, consequently, segment mass. However, the allometry present in the second moment of area is in part due to distribution of bone tissue about the cross-section’s neutral axis. With regards to the digits, aside from data on the length of digit III or the longest digit[44, 45] and total digit mass (current study), scale effects in digit morphology remain unexplored (though see Pike & Maitland for scale effects in claw shape). Given that many functional specializations occur in the pes – such as webbed feet and raptorial claws – future studies of scale effects in segment masses should investigate scale effects both within and across individual functional groups. However analysis of scaling trends within individual functional groups requires larger sample sizes than included in this study and must wait until subsequent studies with higher within-group sampling.
With regards to how muscle mass may contribute to the scaling of segment mass, the picture is somewhat murkier. There is no available data on how total hindlimb muscle mass scales against body or hindlimb length. However, the masses of the biceps femoris group, iliotibialis, femorotibialis, gastrocnemius, and digital flexors all tend to scale isometrically with body mass or with slight positive allometry. If the isometry between muscle mass and body mass is characteristic of the remaining muscles of the hindlimb, then it would indicate that the positive allometry of the mass of the muscled segments (e.g., the thigh and shanks) is due principally to the scaling of bone mass. It should be noted though that the studies of Maloiy et al. and Bennett used functionally based (e.g., cursorial/non-cursorial birds) samples of Aves, and the results of their study may not necessarily reflect scaling patterns for a more inclusive species sampling.
With specific regards to the mass of the pes and its constituent segments, bone mass is almost certainly the primary determinant of segment mass. Given that the flexors and extensors of the intertarsal joint and digits are concentrated on the thigh and shanks segments[72–74], the pedal segments are comprised of predominantly bone, tendon, and integument. In some species, digital extensors are weakly developed, and in these taxa these muscles might make minor – though significant – contributions to pedal mass. The overall concentration of muscle mass on the thigh and shank segments likely underlies differences in regression elevation between the more proximal (thigh and shank) and more distal segments (pes, tarsometatarsal segment, and digits) (Figures 4 and5).
Regarding scale effects as segment masses co-diversified with body and limb size, segment mass is positively allometric with body mass and isometric with body mass and hindlimb length (Table 6). Furthermore, inspection of the confidence intervals for the intercept reveal that, as in the raw regressions, the two proximal-most segments differ in their mass from the two distal-most segments. It thus appears that the between species differences in segment mass scale effects (i.e., raw regression results) only partly reflect how segment masses have co-diversified with body and limb size.
Whole limb mass distribution
Like individual limb segment masses, whole hindlimb mass is positively allometric alongside body mass and isometric alongside hindlimb length (Figure 6). Thus, the scaling of total limb mass is not the result of conflicting scaling trends among individual limb segments. The hindlimb’s mass distribution – as reflected by the hindlimb’s center of mass position and radius of gyration– is also positively allometric with body mass. However, the positive allometry of the hindlimb’s mass distribution is not due to more distal limb segments possessing greater allometric exponents, given the lack of differing scale effects among the hindlimb’s comprising segments (Figures 4,5 and6). Instead, the positive allometry of mass distribution traits is likely due to the scale effects in hindlimb posture. Larger-bodied birds have a more upright posture affording the extensor muscles a greater mechanical advantage across their joints; in contrast smaller-bodied birds have a more crouched limb posture[45, 52] (though see Ref.). As a result of these size-related limb postures, the mass of the distal limb segments is extended farther from the hindlimb’s pivot, and, as a consequence, the hindlimb’s center of mass shifts distally with increasing body size. It should be noted also that the measure of hindlimb length used in Kilbourne specifically reflected the postural differences between small and large-bodied neognath species. Though not a significant departure from isometry’s predicted exponent (1/3), the allometric exponent relating the scaling of hindlimb length to body mass is higher than the prediction (0.41 from Ref.; 0.37 when reanalyzed with current sample, PF-test = 0.1837). As the masses of hindlimb segments and the hindlimb mass distribution are all isometric with hindlimb length, it seems all the more plausible that postural differences between smaller and larger bodied neognaths underlie how the hindlimb’s mass distribution traits scale with body mass.
Negative allometry of limb segment masses would be beneifical for relatively lower costs, given that relatively less mass would need to be accelerated to swing the limb[37, 38]. Likewise, negative allometry of distal limb segment masses would shift the limb’s center of mass proximally along the limb, also resulting in a relative reduction in the cost of swinging the limb[37, 38]. However, in spite of the potential benefits, I found that limb segment masses scale either with positive allometry (relative to body mass) or isometry (relative to hindlimb length). The mass of body segments may be minimized in neognath species in order to minimize the cost of flight. Birds possess a number of traits that can contribute to a lowered metabolic cost of flight, including smaller body masses (though see), pneumatized bones[79–81], and long bones with a more efficient distribution of bone tissue about their cross-section[14, 82] (though some of these traits could be exaptations enabling flight). Thus, neognaths and other birds having hindlimb segments of minimal mass is not implausible; however, isometry or negative allometry of segment mass could result in larger-bodied birds having limbs with too little mass to withstand the mechanical loads occurring not only during terrestrial locomotion but also in other functions, such as prey capture, swimming, or climbing. Conversely, negative allometry or isometry of segment masses could result in small-bodied bird having hindlimbs of greater mass, which could increase the metabolic cost of flight.
The notion that negative allometry of segment mass – particularly bone mass – may result in structurally weak limbs coincides with how bone flexural modulus scales with body mass. Among avian long bones, flexural modulus, the resistance to bending owing to both a bone’s structure and material, decreases with increasing body mass. Additionally, avian long bones are not optimized to be of minimum mass. In a survey of long bone cross-sections within amniotes, Currey & Alexander found that the greater minimization of bone mass in birds may result in long bones more prone to mechanical failure due to the ‘rough-and tumble’ lives of birds. Given that the predominant tissue of the distal limbs segments is bone, it seems highly possible that the negative allometry of segment masses may render the distal limb more susceptible to mechanical failure.
Alternatively, negative allometry of hindlimb segment masses may not be pivotal to neognath locomotion in light of their ability of flight. Notably flight is a cheaper means of locomotion than walking or running to cover long distances, though it is highly costly on a basis of per unit time.
Implications for terrestrial locomotion
The lack of negative allometry of segments masses may act to hamper the terrestrial locomotor ability of larger-sized neognaths by result of limbs that are costly to swing relative to body and limb size. Consequently, larger neognaths may be restricted in how quickly they can oscillate their limbs during terrestrial locomotion. In an examination of scale effects in avian terrestrial locomotion, Gatesy & Biewener found that in larger avian species stride frequency increases with speed at a shallower rate than in smaller avian species, whereas stride length increases at a steeper rate with speed in larger species. Comparing species locomoting at their top speed on a treadmill, the authors found that stride frequency decreases alongside body mass, being proportional to (body mass)-0.18. Though this exponent is greater than the predicted exponent for isometric scaling (-1/3;), it must be noted that stride frequency still overall decreases relative to increasing body size. In contrast to stride frequency, stride length for birds locomoting at their top speed increases alongside body mass, scaling as (body mass)0.39 and well above isometry’s predicted exponent of 1/3. Thus it seems that larger-bodied neognaths may ameliorate any detrimental consequences of scale effects in segment mass by favoring longer strides and relatively lower limb oscillations (i.e., stride frequencies). It should be noted that terrestrially locomoting birds also tend to increase speed by predominantly lowering stance duration. In contrast, swing duration remains invariant or undergoes only minor decreases with increasing speed not only in birds[86–95] but also in mammals[96–106]. The limiting factor on decreasing swing duration could likely be the mass and moment of inertia of the limb and its segments.
Additionally, larger bodied birds may try to allay negative consequences of segment mass scaling by changing their hindlimb kinematics relative to smaller bodied birds. Applying leg weights to the tarsometatarsal segments of running turkeys (Meleagris gallopavo) and guinea fowl (Numida meleagris) has been found to elicit a kinematic response, such as smaller limb segment excursion angles and/or longer swing durations, in light of increased energy expenditure[107, 108]. Thus, in response to limb segments with relatively greater mass, especially those distal on the limb, larger birds may differ in their hindlimb kinematics relative to smaller birds by decreasing limb segment excursions or increasing joint flexion during swing phase. However, to test this hypothesis, detailed data on hindlimb joint kinematics are needed for a sample of birds diverse in both body size and limb function and locomoting over a range of speeds.
Likely owing to limited sample size[67, 109], the two models of trait diversification were plagued with wide confidence intervals, indicating that caution is needed when interpreting these results. Though λ is fairly robust to increasing species sample size, it is highly unlikely that my limited sample (N = 38) fully reflects and encapsulates trait evolution within Neognathae, especially given this clade’s high species richness and complex evolutionary history (~10,000 species[41, 42]). However, by sampling a diverse assemblage of limb specializations, I sought to highlight the role of species poor lineages with distinct hindlimb morphologies in neognath diversification.
With few exceptions, confidence intervals for λ and δ indicate that phylogeny influences variation in segment masses and scale effects by exclusion of a value of 0.0 (Tables 4 and5). However, for both individual traits and hindlimb regression residuals, neither model for segment diversification could be rejected, highlighting the uncertainty in the data (Tables 4 and5). Even when δ could be rejected for the co-diversification of thigh, shank, and pes mass alongside body mass, λ exhibits wide confidence limits nearly spanning bounds upon this parameter (0.0,1.0).
Though sample size likely plays a factor in these results, it is also probable that deep divergences within the phylogeny of my sampled taxa are an additional factor. Examining node ages from my phylogeny reveals that the major lineages constituting my sampled taxa diverged tens of millions of years ago, such as Galloanseraes (109.5 mya), Land Birds (82.1 mya), and Apodiformes (82.9 mya) (divergence times from Jetz et al.). As diversification events occur farther back in the past, the less information is retained in the tree. Given the species richness of neognaths, it is also a distinct possibility that phylogenetic signal varies across the branches of the tree or that rates of evolution do not increase or decrease linearly or monotypically. Furthermore, a lack of data from fossil taxa may also increase the difficulty of identifying a model of trait evolution for hindlimb segment masses in neognaths. Ultimately, as the quality of the data does not provide enough power to distinguish between models, let alone reject either model or both, any interpretations of parameter estimates should be treated with extreme caution without greater sampling of taxa.
Linear vs. mass proportions
With regards to the linear proportions in the avian hindlimb, variation appears to be constrained by embryonic development patterning and postnatal functional demands, both of which limit the variation of zeugopodal (i.e., tibiotarsus) length. Proximo-distal patterning of the amniote limb through activation-inhibition dynamics results in a trade-off in length between the stylopod and autopod, whereas the zeugopod exhibits reduced variation, being approximately 1/3 of total limb segment length. Thus, through developmental pathways, there is decreased variation in the ‘middle segment’ of not only birds but also mammals[5, 15]. However, factors other than activation-inhibition dynamics, such as postnatal growth and/or functional specializations may promote increased variation in zeugopod length. In spite of the apparent influence of developmental mechanisms, it should be noted that the reduced variation in zeugopod length still likely confers a biomechanical advantage to amniotes[5, 15, 113]. Yet does the variation in relative proportions of segments lengths apply as well to the mass proportions of the avian hindlimb?
The differences in limb segment variation with regards to mass vs. length suggest that between segment variation in mass and length are decoupled. It thus appears that while activation-inhibition dynamics likely restrict the relative proportions of segment lengths in birds and other amniotes, such mechanisms do not influence the abundance or perhaps the density of the different tissues comprising the limb segments. This suggests that while developmental mechanisms influence segment lengths relative to one another, other aspects of limb design, such as muscle architecture and bone robustness, may be under greater influence from functional demands and specializations. Alternatively, between segment variation in segmental traits apart from lengths may be under the influence of differing developmental mechanisms or a combination of developmental and functional constraints.
Scale effects within hindlimb segment masses of neognath birds are either positively allometric (when scaled against body mass) or isometric (when scaled against hindlimb length). These scale effects are paralleled within the subclade Land Birds, apart from shank, pedal, and tarsometatarsal segment masses scaling with positive allometry relative to hindlimb length. These results for Neognathae are at odds with previously reported scaling relationships between segment lengths and body mass, in which femur length and digit III length scale with isometry and tibiotarsal and tarsometatarsal length scale with positive allometry. Rather, the scaling of segment mass relative to body mass appears to have stronger ties to the scaling of long bone mass relative to body mass, especially in the case of more distal limb segments. The scaling of hindlimb segment masses likely explains the scaling of stride frequency with body mass and how large-bodied birds increase speed, whereas the negative allometry of hindlimb segment masses may be precluded by the mechanical demands placed upon the limb by locomotor and ecological function. Modeling trait evolution by branch length scaling reveals the influence of phylogeny on segment mass values; however, inherent uncertainty in the fitting of evolutionary models curtails any robust inferences of trait evolution. In spite of recent work indicating that developmental patterning through activation-inhibition dynamics governs limb linear proportions, variation in relative segment masses does not appear to be under the influence of activation-inhibition dynamics.
I would like to thank John Bates, Matt Friedman, John Nyakatura, Sushma Reddy, and Tom Schulenberg, as well as two anonymous reviewers, for comments and discussion that greatly improved the quality of this manuscript. I would also like to thank Matthias Krüger for assistance with specimens, and Ben Marks, Thomas Gnoske, and David Willard for access to specimens and assistance on short notice. This study was supported by DFG grant Fi 410/15-1.
- Gregory WK: Notes on the principles of quadrupedal locomotion and on the mechanism of limbs in hoofed animals. Ann NY Acad Sci. 1912, 22: 267-294.Google Scholar
- Howell AB: Speed in Animals: Their Specializations for Running and Leaping. 1944, Chicago: University of Chicago PressGoogle Scholar
- Coombs WP: Theoretical aspects of cursorial adaptations in dinosaurs. Quart Rev Biol. 1978, 53: 393-418.Google Scholar
- Christiansen P: Locomotion in terrestrial mammals: the influence of body mass, limb length and bone proportions on speed. Zool J Linn Soc. 2002, 136: 685-714.Google Scholar
- Schmidt M, Fischer MS: Morphological integration in mammalian limb proportions: dissociation between function and development. Evolution. 2009, 63: 749-766.PubMedGoogle Scholar
- Gatesy SM, Middleton KM: Bipedalism, flight, and the evolution of theoropod locomotor diversity. J Vert Paleont. 1997, 17: 3-08-329.Google Scholar
- Carrano MT: What, if anything, is a cursor? Categories vs. continua for determining locomotor habit in mammals and dinosaurs. J Zool. 1999, 247: 29-42.Google Scholar
- Middleton KM, Gatesy SM: Theropod forelimb design and evolution. Zool J Linn Soc. 2000, 128: 149-187.Google Scholar
- Benson RBJ, Choiniere JN: Rates of dinosaur limb evolution provide evidence for exceptional radiation in Mesozoic birds. Proc R Soc B. 2013, 280: 20131780-http://dx.doi.org/10.1098/rspb.2013.1780,PubMedPubMed CentralGoogle Scholar
- Dececchi TE, Larsson HCE: Body and limb size dissociation at the origin of birds: uncoupling allometric constraints across a macroevoluitonary transition. Evolution. 2013, 67: 2741-2752.PubMedGoogle Scholar
- Dyke GJ, Nudds RL, Rayner JMV: Limb disparity and wing shape in pterosaurs. J Evol Biol. 2006, 19: 1339-1342.PubMedGoogle Scholar
- Zeffer A, Johansson LC, Marmebro Å: Functional correlation between habitat use and leg morphology in birds (Aves). Biol J Linn Soc. 2003, 79: 461-484.Google Scholar
- Nudds RL, Dyke GJ, Rayner JMV: Forelimb proportions and the evolutionary radiation of Neornithes. Proc R Soc B. 2004, 271: S324-S327.PubMedPubMed CentralGoogle Scholar
- Doube M, Yen SCW, Klosowski MM, Farke AA, Hutchinson JR, Shefelbine SJ: Whole-bone scaling of the avian pelvic limb. J Anat. 2012, 221: 21-29.PubMedPubMed CentralGoogle Scholar
- Stoessel A, Kilbourne BM, Fischer MS: Morphological integration versus ecological plasticity in the avian pelvic limb skeleton. J Morph. 2013, 274: 483-495.PubMedGoogle Scholar
- Snyder RC: The anatomy and function of the pelvic girdle and hindlimb in lizard locomotion. Amer J Anat. 1954, 95: 1-45.PubMedGoogle Scholar
- Christian A, Garland T: Scaling of limb proportions in monitor lizards (Squamata: Varanidae). J Herpet. 1996, 30: 219-230.Google Scholar
- Blob RW: Interspecific scaling of the hindlimb scaling in lizards, crocodilians, felids and canids: does limb bone shape correlate with posture?. J Zool. 2000, 250: 507-531.Google Scholar
- Zaaf A, Van Damne R: Limb proportions in climbing and ground-dwelling geckoes (Lepidosauria, Gekkonidae): a phylogenetically informed analysis. Zoomorph. 2001, 121: 45-53.Google Scholar
- Joyce WG, Gauthier JA: Paleoecology of Triassic stem turtles sheds new light on turtle origins. Proc R Soc B. 2004, 271: 1-5.PubMedPubMed CentralGoogle Scholar
- Garland T, Janis CM: Does metatarsal/femur ratio predict maximal running speed in cursorial mammals?. J Zool. 1993, 229: 133-151.Google Scholar
- Christansen P: Mass allometry of the appendicular skeleton in terrestrial mammals. J Morph. 2002, 251: 195-209.Google Scholar
- Cubo J, Casinos A: Scaling of skeletal element mass in birds. Belg J Zool. 1994, 124: 127-137.Google Scholar
- Grand TI: Body weight: its relation to tissue composition, segment distribution, and motor function. Amer J Phys Anthro. 1977, 47: 211-240.Google Scholar
- Grand TI: Adaptations of tissues and limb segments to facilitate moving and feeding in arboreal folivores. The ecology of arboreal folivores. Edited by: Montgomery GG. 1978, Washington DC: Smithsonian Institution Press, 231-242.Google Scholar
- Grand TI: Body weight: its relationship to tissue composition, segmental distribution of mass, and motor function. Part III. The Didelphidae of French Guyana. Austral J Zool. 1983, 31: 299-312.Google Scholar
- Grand TI: How muscle mass is part of the fabric of behavioral ecology in East African bovids (Madoqua, Gazella, Damaliscus, Hippotragus). Anat Embryol. 1997, 195: 375-386.PubMedGoogle Scholar
- Sprigings E, Leach D: Standardized technique for determining the centre of gravity of body and limb segments of horses. Equine Vet J. 1986, 18: 43-49.Google Scholar
- Crompton RH, Li Y, Alexander RM, Wang W, Gunther MM: Segment intertial properties of primates: new techniques for laboratory and field studies of locomotion. Amer J Phys Anthro. 1996, 99: 547-570.Google Scholar
- Buchner HHF, Savelberg HHCM, Schamhardt HC, Barneveld A: Inertial properties of Dutch warmblood horses. J Biomech. 1997, 30: 653-638.PubMedGoogle Scholar
- Isler K, Payne RC, Günther MM, Thorpe SKS, Li Y, Savage R, Crompton RH: Inertial properties of hominoid limb segments. J Anat. 2006, 209: 201-218.PubMedPubMed CentralGoogle Scholar
- Schoonaert K, D’Août K, Aerts P: Morphometrics and inertial properties in the body segments of chimpanzees (Pan troglodytes). J Anat. 2007, 210: 518-531.PubMedPubMed CentralGoogle Scholar
- Nauwelaerts S, Allen WA, Lane JM, Clayton HM: Inertial properties of equine limb segments. J Anat. 2011, 218: 500-509.PubMedPubMed CentralGoogle Scholar
- Hildebrand M, Hurley JP: Energy of the oscillating legs of a fast-moving cheetah, pronghorn, jackrabbit, and elephant. J Morph. 1985, 184: 23-31.PubMedGoogle Scholar
- Wickler SJ, Hoyt DF, Clayton HM, Mullineaux DR, Cogger EA, Sandoval E, McGuire R, Lopez C: Energetic and kinematic consequences of weighting the distal limb. Equine Vet J. 2004, 36: 772-777.PubMedGoogle Scholar
- Browning RC, Modica JR, Kram R, Goswami A: The effects of adding mass to the legs on the energetics and biomechanics of walking. Med Sci Sports Exer. 2007, 39: 515-525.Google Scholar
- Kilbourne BM: On birds: scale effects in the neognath hindlimb and differences in the gross morphology of wings and hindlimbs. Biol J Linn Soc. 2013, 110: 14-31.Google Scholar
- Kilbourne BM, Hoffman LC: Scale effects between body size and limb design in quadrupedal mammals. PLoS One. 2013, 8: e78392-PubMedPubMed CentralGoogle Scholar
- Marsh RL, Ellerby DJ, Carr JA, Henry HT, Buchanan CI: Partitioning the energetics of walking and running: swinging the limbs is expensive. Science. 2004, 303: 80-83.PubMedGoogle Scholar
- Samuels JX, Meachen JA, Sakai SA: Postcranial morphology and locomotor habits of living and extinct carnivorans. J Morph. 2013, 274: 121-146.PubMedGoogle Scholar
- Gill FB: Ornithology. 2006, San Francisco: W.H. Freeman, 3Google Scholar
- Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO: The global diversity of birds in space and time. Nature. 2012, 491: 444-448.PubMedGoogle Scholar
- Abourachid A, Höfling E: The legs: a key to bird evolutionary success. J Ornith. 2012, 153: 193-198.Google Scholar
- Maloiy GMO, Alexander RM, Njau R, Jayes AS: Allometry of the legs of running birds. J Zool. 1979, 187: 161-167.Google Scholar
- Bennett MB: Allometry of the leg muscles of birds. J Zool. 1996, 238: 435-444.Google Scholar
- Gatesy SM, Dial KP: Locomotor modules and the evolution of avian flight. Evolution. 1996, 50: 331-340.Google Scholar
- O’Meara BC: Evolutionary inferences from phylogenies: a review of methods. Ann Rev Ecol Evol Syst. 2012, 43: 267-285.Google Scholar
- Pagel M: Inferring historical patterns of biological evolution. Nature. 1999, 401: 877-884.PubMedGoogle Scholar
- Hackett SJ, Kimball RT, Reddy SR, Bowie RCK, Braun EL, Braun MJ, Chojnowski JL, Cox WA, Han K-L, Harshman J, Huddleston CJ, Marks BD, Miglia KJ, Moore WS, Sheldon FH, Steadman DW, Witt CC, Yuri T: A phylogenomic study of birds reveals their evolutionary history. Science. 2008, 320: 1763-1768.PubMedGoogle Scholar
- Bertram JEA, Biewener AA: Differential scaling of the long bones in the terrestrial Carnivora and other mammals. J Morph. 1990, 204: 157-169.PubMedGoogle Scholar
- Field DJ, Lynner C, Brown C, Darroch SAF: Skeletal correlates for body mass estimation in modern and fossil flying birds. PLoS One. 2013, 8: e82000-PubMedPubMed CentralGoogle Scholar
- Gatesy SM, Biewener AA: Bipedal locomotion: effects of speed, size and limb posture in birds and humans. J Zool. 1991, 224: 127-147.Google Scholar
- Ricker WE: Linear regression in fishery research. J Fish Res Board Can. 1973, 30: 409-434.Google Scholar
- Rayner JMV: Linear relations in biomechanics – the statistics of scaling functions. J Zool. 1985, 206: 415-439.Google Scholar
- Warton DI, Wright IJ, Falster DS, Westoby M: Bivariate line-fitting methods for allometry. Biol Rev. 2006, 81: 259-291.PubMedGoogle Scholar
- McMahon TA: Using body size to understand the structural design of animals: quadrupedal locomotion. J App Physiol. 1975, 39: 619-627.Google Scholar
- McMahon TA: Muscles, Reflexes, and Locomotion. 1984, Princeton University Press: PrincetonGoogle Scholar
- Nakagawa S, Cuthill IC: Effect size, confidence interval and statistical significance: a practical guide for biologists. Biol Rev. 2007, 82: 591-605.PubMedGoogle Scholar
- R Development Core Team: R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 2012, Available at: http://www.R-project.org. ISBN 3-900051-07-0Google Scholar
- Nakagawa S: A farewell to Bonferroni: the problems of low statistical power and publication bias. Behav Ecol. 2004, 15: 1044-1045.Google Scholar
- Felsenstein J: Phylogenies and the comparative method. Amer Nat. 1985, 125: 1-15.Google Scholar
- Revell LJ, Harmon LJ, Collar DC: Phylogenetic signal, evolutionary process, and rate. Syst Biol. 2008, 57: 591-601.PubMedGoogle Scholar
- Blomberg SP, Garland T: Tempo and mode in evolution: phylogenetic inertia, adaptation and comparative methods. J Evol Biol. 2002, 15: 899-910.Google Scholar
- Freckleton RP, Harvey PH, Pagel M: Phylogenetic analysis and comparative data: a test and review of evidence. Amer Nat. 2002, 160: 712-726.Google Scholar
- Pagel M: Modelling the evolution of continuously varying characters on phylogenetic trees: the case of Hominid cranial capacity. Morphology, Shape, and Phylogeny. Edited by: Macleod N, Forey PL. 2002, London: CRC Press, 269-286.Google Scholar
- Eastman JM, Alfaro ME, Joyce P, Hipp AL, Harmon LJ: A novel comparative method for identifying shifts in the rate of character evolution on trees. Evolution. 2011, 65: 3578-3589.PubMedGoogle Scholar
- Boettiger C, Coop G, Ralph P: Is your phylogeny informative? Measuring the power of comparative methods. Evolution. 2012, 66: 2240-2251.PubMedPubMed CentralGoogle Scholar
- Revell LJ: Phylogenetic signal and linear regression on species data. Methods Ecol Evol. 2010, 1: 319-329.Google Scholar
- Gatesy SM: Hindlimb scaling in birds and other theropods: implications for terrestrial locomotion. J Morph. 1991, 209: 83-96.Google Scholar
- Cubo J, Casinos A: Biomechanical signifiance of cross-sectional geometry of avian long bones. Eur J Morph. 1998, 36: 19-28.Google Scholar
- Pike AVL, Maitland DP: Scaling of bird claws. J Zool. 2004, 262: 73-81.Google Scholar
- Verstappen M, Aerts P, De Vree F: Functional morphology of the hindlimb musculature of the black-billed magpie, Pica pica (Aves, Corvidae). Zoomorph. 1998, 118: 207-223.Google Scholar
- Paxton H, Anthony NB, Corr SA, Hutchinson JR: The effects of selective breeding on the architectural properties of the pelvic limb in broiler chickens: a comparative study across modern and ancestral populations. J Anat. 2010, 217: 153-166.PubMedPubMed CentralGoogle Scholar
- Picasso MBJ: The hindlimb muscles of Rhea americana (Aves, Paleognathae, Rheidae). Anat Histol Embryol. 2010, 39: 462-472.PubMedGoogle Scholar
- McKitrick MC: Phylogenetic analysis of avian hindlimb musculature. Misc Pub Mus Zool Univ Mich. 1991, 179: 1-85.Google Scholar
- Usherwood JR: Constraints on muscle performance provide a novel explanation for the scaling of posture in terrestrial animals. Biol Lett. 2013, 9: 20130414-doi:10.1098/rsbl.2013.0414PubMedPubMed CentralGoogle Scholar
- Maurer BA, Brown JH, Dayan T, Enquist BJ, Ernest SKM, Hadly EA, Haskell JP, Jablonski D, Jones KE, Kaufman DM, Lyons SK, Niklas KJ, Porter WP, Roy K, Smith FA, Tiffney B, Willig MR: Similarities in body size distributions of small-bodied flying vertebrates. Evol Ecol Res. 2004, 6: 783-797.Google Scholar
- Butler RJ, Goswami A: Body size evolution in Mesozoic birds: little evidence for Cope’s rule. J Evol Biol. 2008, 21: 1673-1682.PubMedGoogle Scholar
- O’Connor PM: Pulmonary pneumaticity in the postcranial skeleton of extant Aves: a case study examining Anseriformes. J Morph. 2004, 261: 141-161.PubMedGoogle Scholar
- O’Connor PM: Evolution of archosaurian body plans: skeletal adaptations of an air-sac-based breathing apparatus in birds and other archosaurs. J Exp Zool. 2009, 311A: 504-521.Google Scholar
- Smith ND: Body mass and foraging ecology predict evolutionary patterns of skeletal pneumaticity in the diverse “waterbird” clade. Evolution. 2011, 66: 1059-1078.PubMedGoogle Scholar
- Currey JD, Alexander RM: The thickness of the walls of tubular bones. J Zool. 1985, 206: 453-468.Google Scholar
- Cubo J, Casinos A: Mechanical properties and chemical composition of avian long bones. Eur J Morph. 2000, 38: 112-121.Google Scholar
- Schmidt-Nielsen K: Locomotion: energy cost of swimming, flying, and running. Science. 1975, 177: 222-228.Google Scholar
- Harrison JF, Roberts SP: Flight respiration and energetics. Annu Rev Plant Physiol Plant Mol Biol. 2000, 62: 179-205.Google Scholar
- Gatesy SM: Guineafowl hindlimb function. I: cineradiographic analysis and speed effects. J Morph. 1999, 1240: 115-125.Google Scholar
- Reilly SM: Locomotion in the quail (Coturnix japonica): the kinematics of walking with increasing speed. J Morph. 2000, 243: 173-185.PubMedGoogle Scholar
- Verstappen M, Aerts P: Terrestrial locomotion in the black-billed magpie. I. Spatio temporal gait characteristics. Motor Control. 2000, 4: 150-164.PubMedGoogle Scholar
- Abourachid A: Kinematic parameters of terrestrial locomotion in cursorial (ratites), swimming (ducks), and striding birds (quail and guinea fowl). Comp Biochem Physiol A. 2001, 131: 113-119.Google Scholar
- Van C, Aerts P: Terrestrial locomotion in the white stork (Ciconia ciconia): spatio-temporal gait characteristics. Anim Biol. 2004, 54: 281-292.Google Scholar
- Rubenson J, Heliams DB, Lloyd DG, Fournier PA: Gait selection in the ostrich: mechanical and metabolic characteristics of walking and running with and without an aerial phase. Proc R Soc B. 2004, 271: 1091-1099.PubMedPubMed CentralGoogle Scholar
- White CR, Martin GR, Butler PJ: Pedestrian locomotion energetics and gait characteristics of a diving bird, the great cormorant, Phalacrocorax carbo. J Comp Physiol. 2008, 178: 745-754.Google Scholar
- Nudds RL, Gardiner JD, Tickle PG, Codd JR: Energetics and kinematics of walking in the barnacle goose (Branta leucopsis). Comp Biochem Physiol A. 2010, 156: 318-324.Google Scholar
- Nyakatura JA, Andrada E, Grimm N, Weise H, Fischer MS: Kinematics and center of mass mechanics during terrestrial locomotion in northern lapwings (Vanellus vanellus, Charadriiformes). J Exp Zool A. 2012, 317: 580-594.Google Scholar
- Stoessel A, Fischer MS: Comparative intralimb coordination in avian bipedal locomotion. J Exp Biol. 2012, 215: 4055-4069.PubMedGoogle Scholar
- Goslow GE, Reinking RM, Stuart DG: The cat step cycle: hind limb joint angles and muscle lengths during unrestrained locomotion. J Morph. 1973, 141: 1-42.PubMedGoogle Scholar
- Pridmore PA: Trunk movements during terrestrial locomotion in the Marsupial Monodelphis domestica (Didelphidae). J Morph. 1992, 211: 137-146.Google Scholar
- Hoyt DF, Wickler SJ, Cogger EA: Time of contact and step length: the effect of limb length, running speed, load carrying and incline. J Exp Biol. 2000, 203: 221-227.PubMedGoogle Scholar
- Gillis GB, Biewener AA: Hindlimb muscle function in relation to speed and gait: in vivo patterns of strain and activation in a hip and knee extensor of the rat (Rattus norvegicus). J Exp Biol. 2001, 204: 2717-2731.PubMedGoogle Scholar
- Courtine G, Roy RR, Hodgson J, McKay H, Raven J, Zhong H, Yang H, Tuszynski MH, Edgerton VR: Kinematic and EMG determinants in quadrupedal locomotion of a non-human primate (Rhesus). J Neurophysiol. 2005, 93: 3127-3145.PubMedGoogle Scholar
- Gillis GB, Flynn JP, McGuigan P, Biewener AA: Patterns of strain activation in the thigh muscles of goats across gaits during level locomtion. J Exp Biol. 2005, 208: 4599-4611.PubMedGoogle Scholar
- Rocha-Barbosa O, Fiuza de Castro Loguerico M, Renous S, Gasc J-P: Limb joint kinematics and their relation to increasing speed in the guinea pig Cavia porcellus (Mammalia: Rodentia). J Zool. 2005, 266: 293-305.Google Scholar
- Hutchinson JR, Schwerda D, Famini DJ, Dale RHI, Fischer MS, Kram R: The locomotor kinematics of Asian and African elephants: changes with speed and size. J Exp Biol. 2006, 209: 3812-3827.PubMedGoogle Scholar
- Robilliard JJ, Pfau T, Wilson AM: Gait characterisation and classification in horses. J Exp Biol. 2007, 210: 187-197.PubMedGoogle Scholar
- Maes LD, Herbin M, Hackert R, Bels VL, Abourachid A: Steady locomotion in dogs: temporal and associated spatial coordination patterns and the effect of speed. J Exp Biol. 2008, 211: 138-149.PubMedGoogle Scholar
- Nyakatura JA, Fischer MS, Schmidt M: Gait parameter adjustments of cotton-top tamarins (Saguinus oedipus, Callitrichidae) to locomotion on inclined arboreal substrates. Amer J Phys Anthro. 2008, 135: 13-26.Google Scholar
- Marsh RL, Ellerby DJ, Henry HT, Rubenson J: The energetic costs of trunk and distal-limb loading during walking and running in guinea fowl Numida meleagris. I. Organismal metabolism and biomechanics. J Exp Biol. 2006, 209: 2050-2063.PubMedGoogle Scholar
- Nelson FE, Roberts TJ: Task-dependent force sharing between muscle synergists during locomotion in turkeys. J Exp Biol. 2008, 211: 1211-1220.PubMedGoogle Scholar
- Blomberg SP, Garland T, Ives AR: Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution. 2003, 57: 717-745.PubMedGoogle Scholar
- Münkemüller T, Lavergne S, Bzeznik B, Dray S, Jombart T, Schiffers K, Thuiller W: How to measure and test phylogenetic signal. Methods Ecol Evol. 2012, 3: 743-756.Google Scholar
- Slater GJ, Harmon LJ, Alfaro ME: Integrating fossils with molecular phylogenies improves inference of trait evolution. Evolution. 2012, 66: 3931-3944.PubMedGoogle Scholar
- Young NM: Macroevolutionary diversity of amniote limb proportions predicted by developmental interactions. J Exp Zool B. 2013, 320: 420-427.Google Scholar
- Fischer MS, Blickhan R: The tri-segmented limbs of therian mammals: kinematics, dynamics, and self-stabilization – a review. J Exp Zool. 2006, 305A: 935-952.Google Scholar
- Sokal RR, Braumann CA: Significance tests for coefficients of variation and variability profiles. Syst Zool. 1980, 29: 50-66.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.