TYPE Original Research

PUBLISHED 15 January 2024

DOI 10.3389/feart.2023.1301411

Human lower limb muscle cross

OPEN ACCESS sectional area scales with positive
EDITED BY
Nicholas Baird Holowka,
University at Buffalo, United States
allometry reflecting bipedal
REVIEWED BY
Steven Lautzenheiser,
evolutionary history
The University of Tennessee, Knoxville,
United States
Marta Pina,
Anna Warrener*
London South Bank University, United Kingdom
Department of Anthropology, University of Colorado Denver, Denver, CO, United States
*CORRESPONDENCE
Anna Warrener,
[email protected]

RECEIVED 24 September 2023 Introduction: Muscle cross-sectional area is an important variable for
ACCEPTED 20 December 2023
PUBLISHED 15 January 2024
understanding force generating potential and locomotor adaptation.
Geometric scaling predicts area scales proportionally to body mass2/3.
CITATION
Warrener A (2024), Human lower limb muscle Previous research has quantified scaling relationships between hindlimb
cross sectional area scales with positive musculature and mass in apes, but these relationships have not been clearly
allometry reflecting bipedal
established in humans. Scaling in the human lower-limb is likely influenced by
evolutionary history.
Front. Earth Sci. 11:1301411. bipedalism and dimorphism in lean and total body mass between the sexes.
doi: 10.3389/feart.2023.1301411
Methods: To investigate these relationships, cross-sectional area in 20 muscles
COPYRIGHT
of the lower-limb were obtained through MRI in twenty-eight (14 female, 14
© 2024 Warrener. This is an open-access article
distributed under the terms of the Creative male) participants and measured in Analyze 14.0. Log transformed muscle cross-
Commons Attribution License (CC BY). The use, sectional areas were grouped by function (gluteals, knee extensors, hamstrings,
distribution or reproduction in other forums is
biarticulate knee/hip flexors, plantarflexors), and least-squares regressions were
permitted, provided the original author(s) and
the copyright owner(s) are credited and that the calculated for each group against log-body mass.
original publication in this journal is cited, in
accordance with accepted academic practice. Results: All muscle groups were significantly (p < 0.001) correlated with mass
No use, distribution or reproduction is (0.56 < r2 < 0.70) and, except for the gluteals, all groups scaled with positive
permitted which does not comply with these
allometry with slopes outside the 95% CI reported in the literature for apes.
terms.
Correlations and slopes were lower for females (0.05 < r2 < 0.62; 0.4 < b < 1.0),
than males (0.56 < r2 < 0.79: 0. 7< b < 1.9) in all muscle groups, but there were no
statistically significant differences in slope except for the tensor fasciae latae and
the hamstrings. However, including sex as a predictor in multiple regression
analysis increased the explained variance in cross-sectional area by 1-18% across
functional muscle groups.

Discussion: These results suggest human lower-limb muscle scaling has

responded to force production requirements of bipedalism, but differences in
lean and total mass do not clearly impact allometric equations in males
and females.

KEYWORDS

bipedalism, sexual dimorphism, muscle scaling, geometric similarity, muscle

crosssectional area

1 Introduction
Bipedalism is the defining characteristic of the hominin lineage (Stamos and Alemseged,
2023). Despite a shared upright, bipedal form of locomotion, fundamental differences in
skeletal architecture between numerous hominin species indicate variation in the kinematic
and kinetic profiles of bipedalism during hominin evolution (Lovejoy et al., 2009; DeSilva

Frontiers in Earth Science 01 frontiersin.org

Warrener 10.3389/feart.2023.1301411

et al., 2013; McNutt et al., 2018). The skeleton is a key biomechanical importance of manual and pedal dexterity in arboreal environments
component of the movement system in all vertebrates, providing a (Alexander et al., 1981). Muscle mass is also a key driver of basal
ridged structure upon which muscles, ligaments, and tendons act, metabolic rate when accounting for differences in body mass and
and is also frequently the only biological system that can be assessed phylogeny in mammals (Raichlen et al., 2010). Furthermore, the
in extinct taxa. However, muscle architectural properties are equally association between relative muscle mass and climate variables
important in determining movement profiles, and are therefore of strongly suggests that muscle mass phenotype has been under
great importance when evaluating gait mechanics. selection for thermoregulation across the mammalian order
The force production profile of skeletal muscle is dictated by its (Raichlen et al., 2010). These examples highlight the importance of
architectural properties. Muscles with greater mass and investigating variation in muscle mass and architecture to understand
physiological cross-sectional area (PCSA) produce higher forces, morphology, locomotor function, and behavior in a wide range of taxa.
while fascicle length is associated with range of motion and muscle In the hindlimb specifically, muscle cross-sectional area
shortening velocity (Wickiewicz et al., 1984; Fukunaga et al., 2001). (generally measured as physiological cross-sectional area PCSA)
These characteristics have consequences for locomotor mode, has been investigated across mammals as a means of assessing the
metabolic demand, and ecology across organisms. Interpretations force generating capacity of specific functional muscle groups.
of differences in muscle properties must account for the effect of Among terrestrial mammals, total hindlimb PCSA scales as
body size in both interspecific and intraspecific comparisons across mass0.88 and among primates as an order as mass0.8 (Alexander
animals. When anthropometric variables increase in direct et al., 1981; Myatt et al., 2011). However, in great apes (Pan, Gorilla,
proportion to body mass, this results in geometric similarity in and Pongo) specifically, hindlimb PCSA has a somewhat lower
animals of different sizes, referred to as isometry (Alexander et al., scaling exponent (mass0.7) but still exceeds that predicted by
1981; Biewener, 2003). Isometry predicts that linear variables isometry (Myatt et al., 2011). Musculoskeletal modeling has
increase as mass1/3, areas as mass2/3, and masses and volumes as demonstrated the close relationship between PCSA and
mass1. These relationships can be expressed as the locomotor function in some apes. In chimpanzees, the gluteus
allometric equation: maximus has the largest PCSA of the lower limb muscles,
corresponding with the large hip extensor moments generated
y
axb during quadrupedal locomotion (O’Neill et al., 2013).
Furthermore, the large PCSA of the tibialis posterior reflects its
where y is the independent variable, a and b are constants, and x is
role in foot inversion during climbing (O’Neill et al., 2013). Overall,
body mass (Biewener, 2003).
the macro-muscle architectural properties of apes are functionally
This equation can be linearized by taking the natural log (ln) of
tied to the ability to generate forces over a wide range of limb
the independent and dependent variables resulting in the
positions related to greater degrees of arboreal locomotor behavior
following equation:
(Myatt et al., 2011).
ln y
ln a + b ln x In humans, the adoption of habitual bipedal locomotion has
substantially changed the functional demands of hindlimb muscles,
The regression has a slope (b) equivalent to the scaling factor which are recruited to produce much larger forces over smaller ranges
between the variable of interest and body mass. If the slope exceeds of motion. During human running, ground reaction forces (GRF) may
that expected for geometric similarity, the dependent variable scales exceed three-times body mass (Lieberman et al., 2010), demonstrating
with positive allometry, and if the slope is less than would be the high forces that must be repetitively generated by lower limb
expected, the dependent variable scales with negative allometry. muscles. Vertical, braking, and propulsive GRFs all scale with
Allometric scaling may be the result of functional adaptation, positive allometry in humans (Stickley et al., 2018), therefore, we
phylogeny, or structural demands (Alexander et al., 1981). might expect lower limb functional muscle groups, most associated
Relative muscle mass and architectural properties show with vertical and fore-aft GRF production, like the hip and knee
considerable differences between mammalian taxonomic groups that extensors, and plantarflexors, to also scale with positive allometry.
engage in varying forms of locomotion. Generally, primates are Muscles associated with these functions, the soleus, gluteus medius,
characterized by lower total muscle mass than other mammals gluteus maximus and vastus lateralis, have the largest PCSA of the
related to the number of arboreal species within the order, but muscles in the human lower limb (O’Neill et al., 2013). Sylvester et al.
among primates, terrestrial species have significantly greater relative (2021) simulated muscle force production during walking using a
muscle mass than arboreal taxa (Muchlinski et al., 2012). By forward dynamic approach. Their results showed the highest muscle
comparison, terrestrial mammals have higher proportions of muscle forces occurring in the hip abductors, knee extensors, and ankle
mass relative to body mass, particularly in the hindlimb, which is plantarflexors, with reduced intersubject variability in the force
associated with the ability to produce larger forces at faster speeds on a profiles of these muscles during the stance phase of walking
stable substrate (Muchlinski et al., 2012). Locomotor demand is also (Sylvester et al., 2021). These findings support the notion that
represented in muscle architectural properties across taxonomic groups. bipedality generates substantial force production requirements for
In cursorial mammals, distal muscles are generally smaller, but have specific functional muscle groups associated with vertical support
shorter muscle fiber lengths enabling high force production while and acceleration of the body center of mass. Additionally, because of
maintaining a lighter distal limb which is advantageous for the dimorphism between males and females in body composition, where on
swing phase of stride (Alexander et al., 1981; Lieberman et al., average females have ~10% greater fat mass than males (Wells, 2012),
2015). By comparison, primates have relatively larger distal fore and differences in the scaling exponents relating body mass and muscle size
hindlimb muscle mass with long fiber lengths in association with the may exist between the sexes.

Frontiers in Earth Science 02 frontiersin.org

Warrener 10.3389/feart.2023.1301411

Investigating the evolution of muscle properties in the hominin are supported by previous inverse dynamics models utilizing skeletal
lineage is obviously complicated by the lack of any preserved soft parameters of AL 288–1 (Australopithecus afarensis) and KNM-WT
tissue for extinct species. However, recent computational advances 15000 (Homo erectus) that suggested higher mass and distance specific
have allowed the first representations of complete musculoskeletal muscle power requirements in australopiths than Homo during walking
models of the pelvis and lower-limb of an extinct hominin. Wiseman (Wang et al., 2004). Therefore, a transition in body proportions (skeletal
(2023) reconstructed 36 muscles of the Australopithecus afarensis and muscular) may have been required for the enhancement of long
specimen AL 288–1 using polygonal modelling. This approach distance or high speed bipedal locomotion in the hominin lineage
creates three-dimensional muscle models, accounting for the (Wang et al., 2004).
entirety of each muscle attachment site, and wraps muscle bellies The purpose of this study is to provide additional insight into the
around adjacent musculature to generate realistic muscle placement evolution of human lower limb muscle architecture specifically
within the limb. While PCSA was not reported for this model, through the investigation of the allometric relationship between
muscle mass and moment arm calculations were used to address body mass and muscle cross-sectional area of the lower limb. Muscle
functionally relevant locomotor aspects of australopith lower limb cross-sectional area is an important variable for determining the
anatomy. The largest differences in relative lower limb segment force production capacity of muscles and is more easily obtainable in
muscle masses between the australopith model and humans were living subjects from imaging than muscle volume. Individual
found in the foot and the thigh, with australopiths having a muscles and muscle group allometry is assessed in a cohort of
substantially greater percentage of both segment masses young, active female and male participants with the goal of
comprised of muscle (52.5% versus 15.5% in the foot, 73.7% comparing scaling factors between humans and apes reported in
versus 49.6% in the thigh) (Wiseman, 2023). In the pelvis, the literature, as well as assessing the role of variation in body
muscles were more transversely oriented in the australopith but composition by sex on resulting allometric equations in humans.
maintained muscle moment arms comparable with humans The following hypotheses were tested:
throughout much of hip joint range of motion (Wiseman, 2023).
O’Neill et al. (2023) have also recently developed a three- 1) Individual muscle cross-sectional areas within a functional
dimensional musculoskeletal model of Australopithecus afarensis, muscle group will have the same slope relative to body mass.
allowing estimations of pelvis and lower limb muscle moment arms, 2) Total muscle cross-sectional area of a functional muscle group
PCSA, and isometric joint moments for comparison with human and scales isometrically with body mass.
chimpanzee models. Muscle moment arms throughout joint ranges of 3) Slopes of muscle cross-sectional area functional muscle groups
motion were similar between species with some key exceptions. The relative to body mass will be the same in apes and humas.
hamstrings had smaller moment arms compared to chimpanzees Alternatively, given the increased force production demands of
throughout the flexion/extension range of motion, but were larger human hindlimb muscles for bipedalism, slopes are expected
than those predicted for humans (O’Neill et al., 2023). Knee extensor to be greater for human regressions than those of apes.
moments arms were lower in A. afarensis than humans, while 4) Slopes for muscle cross-sectional area functional muscle
superficial ankle plantarflexors were intermediate in A. afarensis groups relative to body mass will be the same in human
relative to the modern taxa. Hip abductor moment arms were females and males. Alternatively, greater average fat mass in
strongly similar in pattern between humans and A. afarensis, females may produce lower slope values or correlations with
although humans showed a slightly higher magnitude, but strikingly body mass for functional muscle groups than in males.
dissimilar to chimpanzees (O’Neill et al., 2023). These differences are in
accordance with expectations for locomotor and osteometric
distinctions between the species, namely, continued reliance on 2 Methods
climbing (ischial and calcaneal tuber lengths) coupled with
developed bipedal locomotion (pelvic shape and hip abductor 2.1 Subjects
function) (Prang, 2015; O’Neill et al., 2023). PCSA and muscle mass
estimates were sensitive to the modeling parameters (human-like versus Twenty-eight volunteers (14 self-identified males, 14 self-
chimp-like fiber lengths and pinnation angles), but O’Neill et al. (2023) identified females) signed informed consent to participate in this
argue a chimp-like reconstruction is more likely given the importance of study, which was approved by the Washington University in St.
longer muscle fibers for force production over a wide range of motion Louis Internal Review Board (#09–0216). All participants were
during climbing. Muscle mass estimated from these parameters yields physically active recreational runners between 18–40 years of age.
strong agreement between this study and that of Wiseman (2023), with Participants completed an MRI screening form and were paid
the unilateral pelvis and single lower limb accounting for ~10% $25 for participation in the study.
of body mass.
Data obtained from these reconstructions provide additional
evidence of competent bipedal locomotion coupled with retained 2.2 Imaging
capabilities for arboreal climbing in Australopithecus (O’Neill et al.,
2023; Wiseman, 2023). However, differences in the relative masses and Magnetic resonance imaging (MRI) of the participants was
moment arms of muscles in some lower limb segments suggest an obtained on a Siemens Avanto 1.5T scanner at the Center for
evolutionary transition in muscle architectural scaling in concordance Clinical Imaging Research, Mallinckrodt Institute of Radiology.
with skeletal indicators associated with the changing functional Four overlapping sections from the pelvis (approximate L4 level)
demands of walking and running in the genus Homo. These finding to the foot (mid metatarsal) were scanned in coronal orientation

Frontiers in Earth Science 03 frontiersin.org

Warrener 10.3389/feart.2023.1301411

FIGURE 1
Example of axial images of the thigh and leg with individual muscles outlined. Vastus lateralis (VL), vastus intermedius (VI), vastus medialis (VM) rectus
femoris (RF) semimembranousus (SM), semitendinosus (ST), biceps femoris long head (BFLH), adductor magnus (AD), gracilis (G), sartorious (S),
gastrocnemius medialis (GM), gastrocnemius lateralis (GL), soleus (Sol), tibialis posterior (TP). Not shown are biceps femoris short head, adductor longus,
tensor fasciae latae, and gluteal muscles.

with 1.7 mm isotropic voxel resolution. During scanning, superior to and inferior to the original measurement. If a slice
participants were supine and a leg board, foot board, and divider area was greater than the original measurement, the procedure was
maintained the participants’ lower limbs in anatomical position. repeated around the new ACSAmax slice until the largest cross-
Scans lasted approximately 45 min. Analyze 14.0 software (Analyze section was identified. To assess measurement repeatability, the
Direct, KS, USA) was used to reconstruct and visually inspect the procedure was repeated for four muscles (gluteus maximus, rectus
images. Axial coordinates of each scan field were used to assess the femoris, adductor longus, soleus) on four participants. Average
length of overlap between the images which were then cropped and ACSAmax pooled error across participants was 1.6%. All
appended to create a single anatomical image of the measurements were performed by the author.
entire lower limb. The comparative ape sample consists of 12 individuals of both
sexes (4 chimpanzees, 2 bonobos, 4 gorillas, and 2 orangutans)
reported in Myatt et al. (2011). Functional muscle groups were
2.3 Muscle cross-sectional area defined following Myatt et al. (2011) where possible to facilitate
measurements comparison with their published ape data (Table 1). Differentiation
between adductor brevis, adductor longus, and pectineus in the
A muscle’s ability to generate force is frequently determined by superior thigh was insufficient in several participants to reliably
its PCSA which requires knowledge of muscle volume, pinnation measure ACSAmax of adductor brevis and pectineus, so those
angle, and fascicle length (Biewener, 2003). These are difficult muscles are excluded from analysis of the adductor compartment
parameters to determine in living humans without multiple and no comparison between humans and apes is made. ACSAmax of
imaging modalities or estimations from cadaveric specimens. adductor longus and adductor magnus occurs inferior to the
However, studies have documented strong correlations between insertion of adductor brevis and pectineus, so they could be
maximum anatomical cross-sectional area (ACSAmax), PCSA, and reliably measured. Participant specific data for all muscles can be
force production capacity (Fukunaga et al., 2001; Albracht et al., found in the Supplementary Table S1).
2008), and because pinnation angle in the muscles of the lower limb
are less than 30°, ACSAmax should closely approximate PCSA (Myatt
et al., 2011). Therefore, ACSAmax was used for analysis in this study. 2.4 Statistics
ACSAmax was measured in 20 muscles of the right pelvis and
lower limb of each participant using the spline trace and object mask Statistics were performed using SPSS 29.0 software.
tools in Analyze 14.0 (Figure 1). Estimated ACSAmax for each muscle Anthropometrics were compared between males and females
was first visually identified, the contours of the muscle were using independent sample t-tests. Variables were then converted
manually traced and the resulting area was recorded. This to natural log (ln) and ordinary least squares regression was used to
procedure was repeated for no fewer than three axial slices assess the scaling relationship between body mass and individual

Frontiers in Earth Science 04 frontiersin.org

Warrener 10.3389/feart.2023.1301411

TABLE 1 Muscles measured for this study grouped by function.

Functional Muscle Groups

Gluteals gluteus maximus, gluteus minimus, gluteus medius

Adductors adductor magnus, adductor longus

Knee extensors rectus femoris, vastus lateralis, vastus intermedius, vastus medialis

Hamstrings semimembranosus, semitendinosus, biceps femoris long head, biceps femoris short head

Bi-articular knee and hip flexors gracilis, sartorius

Plantarflexors Gastrocnemius lateralis, gastrocnemius medialis, soleus, tibialis posterior

Other Tensor fasciae latae

TABLE 2 Anthropometrics. p-values are for mean differences between males and females.

Total Female mean ± 1SD Male mean ± 1SD Difference p-value

Body Mass (kg) 65.4 ± 9.1 60.8 ± 7.1 70.0 ± 8.7 NS

Height (m) 1.71 ± 0.08 1.65 ± 0.03 1.78 ± 0.06 <0.05

Maximum Anatomical Cross-Sectional Area (cm2)

G. maximus 52.8 ± 9.3 46.3 ± 5.37 59.16 ± 7.96 NS

G. medius 30.7 ± 5.5 27.72 ± 4.22 33.63 ± 5.07 NS

G. minimus 17.6 ± 3.3 15.34 ± 2.02 19.91 ± 2.69 NS

Tensor Fasciae Latae 7.3 ± 2.4 5.8 ± 1.2 8.8 ± 2.4 <0.001

A. magnus 33.5 ± 8.0 28.12 ± 4.29 38.88 ± 7.17 NS

A. longus 15.5 ± 4.4 12.25 ± 1.92 18.74 ± 3.80 <0.05

Semimembranosus 14.0 ± 3.3 12.31 ± 2.0 15.48 ± 3.60 NS

Semitendinosus 9.2 ± 2.7 7.5 ± 1.1 10.9 ± 2.8 NS

Biceps femoris long 14.0 ± 2.6 13.1 ± 1.3 14.8 ± 3.2 <0.05

Biceps femoris short 6.8 ± 1.9 5.4 ± 0.7 8.1 ± 1.9 <0.05

Vastus lateralis 32.0 ± 7.2 27.3 ± 4.4 36.7 ± 6.3 NS

Vastus intermedius 23.5 ± 5.6 19.5 ± 3.5 27.6 ± 4.2 NS

Vastus medialis 24.2 ± 5.3 20.7 ± 3.1 27.7 ± 4.7 NS

Rectus femoris 13.9 ± 3.8 10.9 ± 1.6 16.9 ± 2.9 NS

Sartorius 3.8 ± 0.9 3.2 ± 0.6 4.5 ± 0.8 NS

Gracilis 4.7 ± 1.5 3.8 ± 0.9 5.6 ± 1.4 NS

Gastrocnemius lateralis 9.6 ± 2.2 8.8 ± 1.9 10.4 ± 2.2 NS

Gastrocnemius medialis 15.6 ± 3.4 14.4 ± 2.8 16.8 ± 3.7 NS

Soleus 23.6 ± 4.2 22.7 ± 4.6 24.5 ± 3.7 NS

Tibialis posterior 4.9 ± 1.0 4.8 ± 1.0 5.1 ± 1.0 NS

muscle ACSAmax, functional muscle group ACSAmax, and sex b1 − b2

t







specific functional muscle group ACSAmax as well as total lower s21 + s22
limb ACSAmax. Significant differences between regression slopes for
individual muscles within a functional group, and for muscle groups where b is the slope and s is the standard error of the slope, and the
and summed muscle ACSAmax differences between males and degrees of freedom are equal to (n1 + n2)—4 (Currell, 2015).
females were determined by t-test using the following formula: Multiple regressions were also performed with body mass and

Frontiers in Earth Science 05 frontiersin.org

Warrener 10.3389/feart.2023.1301411

FIGURE 2
Regression plots of natural log transformed body mass (kg) versus ACSAmax (cm2) for individual muscles within functional muscle groups. All
regressions are significant at the p < 0.001 level. (A) gluteals and tensor fasciae latae, (B) adductors and bi-articulate hip-knee flexors, (C) hamstrings, (D)
knee extensors, (E) plantarflexors. There are no significant differences in the slope of the regression lines for individual muscles within a functional group
except for the biceps femoris long head and the soleus. Regression parameters are listed in Table 3.

sex as independent variables to determine the influence of each ACSAmax on body mass were calculated for each muscle group
factor on natural log transformed ACSAmax independently. To and for summed muscle ACSAmax. The average difference in
determine the magnitude of difference in ACSAmax controlling z-scores between the sexes is reported as a measure of effect size.
for body mass between the sexes, standardized residuals of Results were considered statistically significant at α = 0.05.

Frontiers in Earth Science 06 frontiersin.org

Warrener 10.3389/feart.2023.1301411

TABLE 3 Intercept, slope, and 95% CI for allometric equations relating natural log transformed body mass and ACSAmax. All regressions are significant at the
p < 0.001 level. Slopes and 95% CI for ape regressions provided by Myatt et al. (2011). Slopes in bold indicate that the 95% CI does not overlap the isometric
expectation of 0.67. * indicates the human slope for a functional muscle group does not overlap the 95% CI for the ape regression.

Muscle group Individual muscles Intercept ± 95% CI Slope ± 95% CI R2 Ape Slope ± 95% CI
Gluteals 1.02 ± 1.28 0.86 ± 0.26 0.56 0.97 ± 0.29

G. maximus 0.14 ± 1.49 0.91 ± 0.36 0.52

G. medius 0.19 ± 1.77 0.77 ± 0.425 0.35

G. minimus −0.59 ± 1.84 0.82 ± 0.44 0.36

Adductors −2.11 ± 1.57 1.45±0.372 0.70 NA

A. magnus −2.12 ± 1.70 1.34±0.41 0.63

A. longus −4.14 ± 2.06 1.64±0.49 0.64

Tensor Fasciae Latae −4.91 ± 2.97 1.64±0.71 0.46

Hamstrings −0.99 ± 1.40 1.14±0.33* 0.65 0.68 ± 0.20

Semimembranosus −2.4 ± 1.81 1.19±0.43 0.55

Semitendinosus −4.2 ± 2.01 1.53±0.48 0.62

Biceps femoris long −0.51 ± 1.64 0.75 ± 0.39 0.37

Biceps femoris short −3.72 ± 2.36 1.34 ± 0.57 0.48

Knee Extensors −0.83 ± 1.52 1.28±0.36* 0.67 0.77 ± 0.31

Vastus lateralis −2.38 ± 1.49 1.4±0.36 0.71

Vastus intermedius −2.24 ± 2.04 1.29±0.49 0.53

Vastus medialis −1.42 ± 1.88 1.1 ± 0.45 0.49

Rectus femoris −3.25 ± 2.51 1.4±0.60 0.47

Bi-articulate hip and knee flexors −4.68 ± 2.04 1.63±0.49* 0.64 0.52 ± 0.20

Sartorius −5.2 ± 1.95 1.56±0.47 0.64

Gracilis −5.6 ± 2.69 1.7±0.64 0.53

Plantarflexors −0.34 ± 1.23 1.03±0.29* 0.66 0.62 ± 0.36

Gastrocnemius lateralis −2.88 ± 1.64 1.23±.39 0.61

Gastrocnemius medialis −2.94 ± 1.27 1.36±0.30 0.76

Soleus −0.13 ± 1.81 0.79 ± 0.43 0.35

Tibialis posterior −1.80 ± 2.32 0.81 ± 0.55 0.26

The bold values are outside confidence intervals.

3 Results significant at p < 0.001. Body mass explains between 35%–52% of the
variation in individual gluteal muscle (maximus, medius, and
Mean height for all participants was 1.71 ± 0.08 m (range minimus) ACSAmax, and all regressions have slopes greater than
1.58–1.87 m). Mean body mass was 65.4 ± 9.1 kg (range predicted by isometry, but the 95% confidence intervals overlap 0.67,
47–91.4 kg) (Table 2). so isometric scaling cannot be precluded. There are no significant
differences between individual gluteal muscle ACSAmax regression
slopes relative to body mass. Body mass explains 63% and 64% of
3.1 Combined sample individual ACSAmax adductor magnus and adductor longus ACSAmax respectively, both
scaling with body mass individual muscles scale with positive allometry, and there is no
significant difference between the slopes of the two individual
Figure 2 shows scatterplots for all individual muscles regressed muscle ACSAmax relative to body mass.
on body mass for the combined sample, and Table 3 shows In the hamstrings, all muscles scale with positive allometry and the
regression coefficients for individual and function muscle groups. 95% CI does not include 0.67, except for the biceps femoris long head
All regressions of individual muscle ACSAmax on body mass are (b = 0.75), and body mass explains between 37%–62% of the variation

Frontiers in Earth Science 07 frontiersin.org

Warrener 10.3389/feart.2023.1301411

FIGURE 3
Regression plots of natural log transformed body mass (kg) versus ACSAmax (cm2) for functional muscle groups (A) gluteals, hamstrings, and knee
extensors, and (B) plantarflexors and biarticulate knee and hip flexors. All regressions are significant at the p < 0.001 level. Solid lines are for human data.
Dashed lines of the same color as functional muscle groups are regression lines based on slope and intercept values from Myatt et al. (2011)*. All slope
values for humans fall outside the 95% CI for ape regression lines except for the gluteal muscles. Regression parameters are listed in Table 3.

in individual muscle ACSAmax in this functional group. The biceps and the 95% confidence intervals for the slopes of each group
femoris long head has a significantly lower regression slope value than exclude 0.67. The slope value for gluteal ACSAmax is also greater
the regression slope for semitendinosus (p < 0.05) and approaches than 0.67 (b = 0.86), but the confidence intervals include the
significance for the semimembranosus and biceps femoris short head isometric slope. Body mass explains between 56%–70% of the
(p = 0.08 and p = 0.12 respectively), but no other slope comparisons variation in muscle group ACSAmax with the gluteals having the
within the muscle group are significantly different. Among the knee lowest correlation coefficient and the adductors the highest. All
extensors, the vasti muscles and the rectus femoris all scale with positive functional muscle group ACSAmax values, except the gluteals, have
allometry, but the rectus femoris 95% CI includes 0.67. R2 values for this slope values greater than the 95% confidence intervals for
muscle group are higher in the vastus lateralis, where 71% of variation in comparable muscle groups in great apes as reported by Myatt
ACSAmax is explained by body mass and lowest in the rectus femoris et al. (2011) (Figure 3; Table 3).
(R2 = 0.47), and there are no significant differences in regression slopes
between any pair of muscles in the functional group.
Both biarticulate flexors crossing the hip and knee scale with 3.3 Functional group ACSAmax scaling by sex
positive allometry. Sixty-four percent and 53% of the variation in the
sartorius and gracilis respectively is explained by body mass with no Males and females in the sample differed statistically in height
significant difference in the regression slope of these muscles. (p < 0.05) but not body mass (Table 2). The only muscle ACSAmax
Among the plantarflexors, gastrocnemius medialis and lateralis that were statistically significantly different between the sexes were
scale with positive allometry (b = 1.36 and 1.23 respectively), the tensor fasciae latae, biceps femoris long head, short head, and the
while soleus and tibialis posterior have lower slope values, but grouped knee flexors (Table 2). All regressions of group muscle
the only statistically significant difference is between the soleus ACSAmax and summed lower limb ACSAmax on body mass were
and gastrocnemius medialis (p <0.05). Body mass explains 76% and significant at the p < 0.05 level in both sexes, except for the tensor
61% of the variance in ACSAmax in the medial and lateral fasciae latae and the gluteals in females (Table 4, Figure 4).
gastrocnemius respectively, but a much lower percentage in the Regression slopes are universally higher in males than females,
soleus (35%) and tibialis posterior (26%). Finally, body mass but the only statistically significant slope differences are for the
explains 46% of the variation tensor fasciae latae ACSAmax with tensor fasciae latae and the hamstrings. Body mass also explains a
an allometrically positive slope of 1.64. greater percentage of muscle ACSAmax variance in each functional
group and for summed muscle ACSAmax in male regressions
compared to females, indicating a stronger correlation between
3.2 Combined sample functional group body mass and muscle ACSAmax generally in males. Including
ACSAmax scaling sex as an independent variable with body mass in a multiple
regression model increases explained variance in functional
All regressions of functional muscle group ACSAmax on body ACSAmax across all muscle groups except in the plantarflexors
mass are significant at p < 0.001. Except for the gluteal muscle group, where there is not a significant relationship between sex and
all functional group ACSAmax values scale with positive allometry group ACSAmax (Table 5). When controlling for body mass,

Frontiers in Earth Science 08 frontiersin.org

Warrener 10.3389/feart.2023.1301411

FIGURE 4
Regression plots of natural log transformed body mass (kg) versus ACSAmax (cm2) for functional muscle groups (A) gluteals and hamstrings, (B) knee
extensors, plantarflexors, and biarticulate hip and knee flexors, (C) adductors and tensor fasciae latae and (D) summed total ACSAmax by sex. Closed
symbols and solid lines are male data points and regression lines, and open symbols and dotted lines are female data points and regression lines. All
regressions are significant at the p < 0.05 level except for the tensor fasciae latae and the gluteals in females. Males and females have significantly
different regression slopes for the tensor fasciae latae and hamstrings. Regression parameters are listed in Table 4.

muscle ACSAmax values were larger in males (0.6 < 1.1 standard within the group. Each functional group scaled with positive
deviations) than in females in all muscle groups except the ankle allometry and, except for the gluteals, 95% confidence intervals
plantarflexors where females where 0.3 standard deviations greater did not overlap the isometric slope of 0.67, rejecting the second
than males, and the tensor fascia latae where there was no average hypothesis.
difference in size between the sexes (Table 5).
Hypothesis 3: Slopes of muscle cross-sectional area functional
muscle groups relative to body mass will be the same in apes and
4 Discussion humas. Alternatively, given the increased force production demands
of human hindlimb muscles, slopes are expected to be greater for
human regressions than those of apes.
Hypothesis 1 and 2: Individual muscle cross-sectional areas In support of the third hypothesis alternative, the hamstrings,
within a functional muscle group will have the same slope knee extensors, bi-articulate hip and knee flexors, and plantarflexors
relative to body mass, and total muscle cross-sectional area of a regression slopes for the combined human sample all exceeded those
functional muscle group scales isometrically with body mass. reported for apes and were outside the ape regression 95%
Individual muscles within a functional group generally showed confidence intervals (Myatt et al., 2011). The greater slope values
the same scaling relationship to body mass except for biceps femoris of the functional muscle groups analyzed here indicate substantial
long head and the soleus. This supports the first hypothesis and deviation away from an ape-like pattern of muscle scaling in our
suggests scaling analysis of muscles in functional groups is evolutionary history. Demand on lower limb muscle force
appropriate and does not mask scaling differences of muscles production is increased in humans due to the redistribution of

Frontiers in Earth Science 09 frontiersin.org

Warrener 10.3389/feart.2023.1301411

TABLE 4 Intercept, slope, and 95% CI for allometric equations relating log transformed body mass and ACSAmax by sex. Slopes in bold indicate that the 95%
CI does not overlap the isometric expectation of 0.67.

Muscle group Sex Intercept ± Slope ± R2 p-value p-value Slope difference

95% CI 95% CI Regression F/M
Gluteals F 2.87 ± 1.87 0.4 ± 0.45 0.23 0.083 0.26

M 1.7 ± 1.7 0.71 ± 0.4 0.56 <0.001

Adductors F 0.36 ± 2.01 0.81 ± 0.49 0.52 0.004 0.18

M −1.15 ± 1.96 1.22±0.46 0.71 <0.001

Tensor Fasciae Latae F 0.05 ± 4.6 0.41 ± 1.21 0.05 0.44 0.04

M −5.97 ± 4.53 1.91±1.07 0.56 0.002

Hamstrings F 1.61 ± 1.57 0.5 ± 0.38 0.40 0.015 0.009

M −1.85 ± 2.5 1.35±0.6 0.67 <0.001

Knee Extensors F 1.37 ± 2.31 0.73 ± 0.56 0.40 0.016 0.34

M 0.38 ± 1.38 1.01±0.32 0.79 <0.001

Bi-articulate hip and knee F −2.1 ± 4.0 1.0 ± 0.97 0.30 0.044 0.32
flexors

M −4.11 ± 2.24 1.51±0.53 0.76 <0.001

Plantarflexors F −1.03 ± 2.41 1.2 ± 0.59 0.62 <0.001 0.61

M −0.32 ± 2.01 1.03 ± 0.47 0.65 <0.001

Summed F 2.90 ± 1.44 0.70 ± 0.35 0.60 0.001 0.1

M 1.62 ± 1.13 1.03±0.27 0.85 <0.001

The bold values are outside confidence intervals.

TABLE 5 Results of multiple regression analysis with body mass and sex as independent variables and muscle group ACSAmax as the dependent variable.
p-values are for the slopes of the independent variables and change R2 shows the increase in explained variance of the dependent variable when sex is
added to the regression model. Z-score difference measures the average difference in ACSAmax controlling for body mass measured in units of standard
deviations. Positive values indicate males are larger than females.

Muscle Group Intercept ± Slope ± 95% CI p-value Slope±95% p-value Change Z-score
95% CI body mass body mass CI sex sex R2 Diff
Gluteals 2.05 ± 1.14 0.56 ± 0.28 <0.001 0.15 ± 0.08 <0.001 0.18 1.0

Adductors −0.71 ± 1.28 1.02 ± 0.33 <0.001 0.21 ± 0.08 <0.001 0.146 1.1

Tensor Fasciae Latae −3.34 ± 3.1 1.18 ± 0.78 0.005 0.23 ± 0.21 0.03 0.09 0

Hamstrings −0.29 ± 1.48 0.93 ± 0.38 <0.001 0.10 ± 0.10 0.04 0.05 0.6

Knee Extensors 0.56 ± 1.21 0.87 ± 0.30 <0.001 0.20 ± 0.08 <0.001 0.17 1.2

Bi-articulate hip and −3.43 ± 2.06 1.26 ± 0.52 <0.001 0.18 ± 0.14 0.01 0.08 0.8
knee flexors

Plantarflexor −0.06 ± 1.95 1.11 ± 0.35 <0.001 −0.40 ± 0.26 ns 0.01 −0.3

Summed 2.04 ± 0.86 0.86 ± 0.22 <0.001 0.14 ± 0.06 <0.001 0.11 1.1

body mass exclusively to the lower limbs during bipedalism. Muscle increases in either gait (Wall-Scheffler et al., 2010; Franz and Kram,
activation, and presumably force demand, is also influenced by gait, 2012). In some muscles, stance phase peak activation amplitudes
incline, and speed (Wall-Scheffler et al., 2010; Franz and Kram, during faster walking speeds may exceed those during slower
2012). Studies have documented consistent and substantial increases running suggesting optimization of muscle performance may
in electromyographic (EMG) amplitude in muscles of the initiate gait transition (Cappellini et al., 2006). Interestingly,
hamstrings, quadriceps, adductors, abductors, and plantarflexors despite the major differences in kinematics and kinetics between
during stance phase of inclined walking and running or as speed the two gaits, muscle activation patterns are largely similar during

Frontiers in Earth Science 10 frontiersin.org

Warrener 10.3389/feart.2023.1301411

both stance and swing phase across the gaits with the major The impact of greater average fat mass relative to body mass in females
distinction being a shift in timing of the activation of ankle (~10%) (Wells, 2012) has implications for sports performance.
plantarflexion to earlier in the support phase during running Difference in performance times for elite male and female athletes in
(Cappellini et al., 2006). There has been particular focus on the both running and swimming events corresponds closely with differences
activation of the human gluteus maximus during bipedal in BMI, implicating relative muscle mass as the key explanatory variable
locomotion given the substantially larger size of this muscle in in race time disparities (Gagnon et al., 2018). Future studies that assess
humans compared to apes. Lieberman et al. (2006) argue that the the scaling relationship between lean body mass and muscle mass, with
gluteus maximus is significantly more active during stance and larger samples of both males and females, will be needed to confirm the
swing phase of running than walking, suggesting increases in current findings.
muscle size occurred with the development of endurance running The relative contributions of lean and fat mass in relation to
capabilities in the genus Homo (although they do not rule out the muscle ACSAmax, and the evolutionary history of sexual
evolutionary importance of inclined walking on gluteus maximus dimorphism in the hominin lineage, both have implications for
size). However, Wall-Scheffler et al. (2010) suggest that the interpreting muscle architecture in the fossil record. Human life-
reduction in variability in gluteus maximus activation during history is characterized by an energetic paradox with substantial
higher speed or incline walking compared to running is more energy demands required for encephalization, large offspring,
compatible with a model of optimization of a walking gait (at shortened interbirth intervals, high activity rates, and longevity
varying intensities) being a primary driver of increased gluteus (Pontzer et al., 2016). To fuel these demands, human total energy
maximus size in humans. More recent data suggests that the expenditure (TEE) is accelerated compared to other apes, and
gluteus maximus is recruited for fast and powerful movements in humans (particularly females) also possess an increased capacity
a variety of locomotor contexts from sprinting to climbing (Bartlett to store energy as fat (Wells, 2007; Pontzer et al., 2016). This
et al., 2014). While both human and ape gluteals scale with positive metabolic strategy likely arose with the evolution of hunting and
allometry (albeit with CI overlapping 0.67) in the current study, the gathering early in the genus Homo (Pontzer, 2012), and modern
higher regression intercept in humans (1.05 vs. 0.35) supports a dimorphism in relative fat mass is a likely consequence of the greater
grade shift in gluteal size in our evolutionary history, even if the rate metabolic burden of reproduction on human females. The fossil
of increase relative to body mass within species is relatively similar record also suggests that some early hominins such as
across all apes. The shared activation patterns of the two gaits, and Australopithecus had a more ape-like pattern of body size sexual
similar demands for increased force with speed or incline makes it dimorphism compared to modern humans (Gordon, 2013).
difficult to argue that either walking or running has exclusively Unfortunately, current data for apes are insufficient to directly
driven adaptive responses in muscle architecture. address the allometric relationship between body mass and
ACSAmax in highly dimorphic primate taxa given the rarity of
Hypothesis 4: Slopes for muscle cross-sectional area functional cadaveric specimens of these species (Myatt et al., 2011). Future
muscle groups relative to body mass will be the same in human research exploring sex specific scaling relationships between body
females and males. Alternatively, greater average fat mass in females mass and muscle size in primates generally are needed to determine
may produce lower slope values or correlations with body mass for the role of selection, allometry, growth duration, and growth rate in
functional muscle groups than in males. muscle dimorphism in our evolutionary lineage.
The fourth hypothesis is supported by the current data and suggests
similarity in relative force production demands across muscle groups in
both males and females. However, males universally had higher slope 5 Limitations
values than females but they did not reach statistical significance (except
for tensor fascia latae and hamstrings). The current data, suggest that in One limitation of this study is the comparison of ACSAmax from
both sexes muscle ACSAmax increases at a faster rate than body mass, the human sample with PCSA measured in apes by Myatt et al.
giving larger individuals greater ability to produce force relative to body (2011). Theoretically, in non-pinnate muscles, measurements of
mass. These results are in partial agreement with previous studies of ACSA pass perpendicular to parallel muscle fibers and these two
allometric scaling of thigh and leg girth, used as a proxy for muscle cross- measures should be the same (Gadeberg et al., 1999; Akagi et al.,
sectional area (Nevill et al., 2004). For both proximal and distal limb 2009). However, discrepancies in PCSA and ACSAmax measured on
segments, Nevill et al. (2004) found girth scaled with positive allometry the same individuals and in reported muscle force production
in both athletes and non-athletes. The scaling factors did not differ capacity relative to different measures of area has created
between males and females at the thigh, but leg girth scaling was affected uncertainty about the best measurement for predicting muscle
by sex (Nevill et al., 2004). Lower R2 values for all female regressions in force (Fukunaga et al., 1992; Jones et al., 2008). This should be
this study also indicate more variability in muscle mass relative to body considered when evaluating the results of the current study.
mass compared to males. This is supported by the multiple regression Additionally, varying levels of sexual dimorphism and different
analyses, where including sex as a covariate with body mass increased locomotor behaviors in the ape taxa analyzed by Myatt et al.
explained variance across muscle groups by 1%–18%, and demonstrates (2011) may impact the relationship between body mass and
a significant independent effect of sex on ACSAmax for all muscle groups PCSA reported in the ape regressions. For example, orangutans
except the plantarflexors. Furthermore, controlling for body mass, males engage in more cautious climbing and are rarely terrestrial in the
had greater muscle ACSAmax values in most muscle groups. These wild, while gorillas transition from mixed arboreal/terrestrial
differences could be attributed to sex alone, or could be a combination of locomotion as juveniles to strictly terrestrial behavior in
effects from training, sex, and other variables not measured in this study. adulthood. However, given the scarcity of ape cadaveric

Frontiers in Earth Science 11 frontiersin.org

Warrener 10.3389/feart.2023.1301411

specimens, sample sizes are currently insufficient to provide species, Author contributions
or sex, specific allometric relationships. The human sample size may
also be insufficient in cases where the 95% confidence intervals for AW: Conceptualization, Data curation, Formal Analysis,
regression slopes overlapped with isometric expectations. Larger Funding acquisition, Investigation, Methodology, Software,
sample sizes may ultimately show isometric scaling in these cases, Visualization, Writing–original draft, Writing–review and editing.
however obtaining MRI or CT muscle data on a large mixed-sex
sample of young, healthy adults is difficult. Finally, due to limitations
on software and personnel access, all measurements of ACSAmax Funding
were performed solely by the author, so no estimates of inter-
observer error are available for these data. The author(s) declare financial support was received for the
research, authorship, and/or publication of this article. Funding for
this project was provided by the National Science Foundation grant
6 Conclusion NSF#0850841, The L.S.B. Leakey Foundation, Wenner-Gren
Foundation for Anthropological Research, and the University of
This research addresses the relationship between lower-limb Colorado Denver.
muscle cross-sectional area and body mass to enhance
understanding of the role of bipedalism in shaping muscle
architecture in human evolution. Across functional muscle groups, Acknowledgments
human cross-sectional area scales with positive allometry, reflecting
the force production demands of bipedal walking and running. In all Thank you to Glenn Foster and Scott Love at the Center for
functional groups but the gluteal muscles, human scaling factors are Clinical Imaging for running MRIs for all participants and to two
outside the 95% confidence intervals for comparable muscle groups in reviewers for comments that improved this manuscript.
apes reported in the literature. Despite differences in average fat mass
between males and females, this study did not find statistically
different scaling factors for most lower-limb muscles between the Conflict of interest
sexes. However, the overall lower slopes and R2 values for female
regressions, as well as the independent effect of sex in multiple The author declares that the research was conducted in the
regression analysis, suggest that additional data that explicitly absence of any commercial or financial relationships that could be
account for fat mass are needed to confirm these findings. construed as a potential conflict of interest.

Data availability statement Publisher’s note

The original contributions presented in the study are included in All claims expressed in this article are solely those of the authors and
the article/Supplementary Material, further inquiries can be directed do not necessarily represent those of their affiliated organizations, or
to the corresponding author. those of the publisher, the editors and the reviewers. Any product that
may be evaluated in this article, or claim that may be made by its
manufacturer, is not guaranteed or endorsed by the publisher.
Ethics statement
The studies involving humans were approved by the Supplementary material
Washington University in St. Louis Internal Review Board (#09-
0216). The studies were conducted in accordance with the local The Supplementary Material for this article can be found online
legislation and institutional requirements. The participants provided at: https://www.frontiersin.org/articles/10.3389/feart.2023.1301411/
their written informed consent to participate in this study. full#supplementary-material

References
Akagi, R., Takai, Y., Ohta, M., Kanehisa, H., Kawakami, Y., and Fukunaga, T. (2009). Biewener, A. (2003). Animal locomotion. New York: Oxford University Press.
Muscle volume compared to cross-sectional area is more appropriate for evaluating muscle
Cappellini, G., Ivanenko, Y. P., Poppele, R. E., and Lacquaniti, F. (2006). Motor
strength in young and elderly individuals. Age Ageing 38, 564–569. doi:10.1093/ageing/afp122
patterns in human walking and running. J. Neurophysiology 95, 3426–3437. doi:10.
Albracht, K., Arampatzis, A., and Baltzopoulos, V. (2008). Assessment of muscle 1152/jn.00081.2006
volume and physiological cross-sectional area of the human triceps surae muscle in vivo.
Currell, G. (2015). Scientific data analysis. Oxford University Press.
J. Biomechanics 41, 2211–2218. doi:10.1016/j.jbiomech.2008.04.020
DeSilva, J. M., Holt, K. G., Churchill, S. E., Carlson, K. J., Walker, C. S., Zipfel, B., et al.
Alexander, R. M. N., Jayes, A. S., Maloiy, G. M. O., and Wathuta, E. M. (1981).
(2013). The lower limb and mechanics of walking in Australopithecus sediba. Science
Allometry of the leg muscles of mammals. J. Zoology 194, 539–552. doi:10.1111/j.1469-
340, 1232999. doi:10.1126/science.1232999
7998.1981.tb04600.x
Franz, J. R., and Kram, R. (2012). The effects of grade and speed on leg muscle
Bartlett, J. L., Sumner, B., Ellis, R. G., and Kram, R. (2014). Activity and functions of
activations during walking. Gait Posture 35, 143–147. doi:10.1016/j.gaitpost.2011.
the human gluteal muscles in walking, running, sprinting, and climbing. Am. J. Phys.
08.025
Anthropol. 153, 124–131. doi:10.1002/ajpa.22419

Frontiers in Earth Science 12 frontiersin.org

Warrener 10.3389/feart.2023.1301411

Fukunaga, T., Miyatani, M., Tachi, M., Kouzaki, M., Kawakami, Y., and Kanehisa, H. O’Neill, M. C., Lee, L.-F., Larson, S. G., Demes, B., Stern, J. T., and Umberger, B. R.
(2001). Muscle volume is a major determinant of joint torque in humans: muscle (2013). A three-dimensional musculoskeletal model of the chimpanzee (Pan
volume and joint torque. Acta Physiol. Scand. 172, 249–255. doi:10.1046/j.1365-201x. troglodytes) pelvis and hind limb. J. Exp. Biol. 216, 3709–3723. doi:10.1242/jeb.079665
2001.00867.x
O’Neill, M. C., Nagano, A., and Umberger, B. R. (2023). A three-dimensional
Fukunaga, T., Roy, R. R., Shellock, F. G., Hodgson, J. A., Day, M. K., Lee, P. L., et al. musculoskeletal model of the pelvis and lower limb of Australopithecus afarensis.
(1992). Physiological cross-sectional area of human leg muscles based on magnetic Am. J. Biol. Anthropol., 1. –22. doi:10.1002/ajpa.24845
resonance imaging. J. Orthop. Res. 10, 926–934. doi:10.1002/jor.1100100623
Pontzer, H. (2012). Ecological energetics in early Homo. Curr. Anthropol. 53,
Gadeberg, P., Andersen, H., and Jakobsen, J. (1999). Volume of ankle dorsiflexors and S346–S358. doi:10.1086/667402
plantar flexors determined with stereological techniques. J. Appl. Physiology 86,
Pontzer, H., Brown, M. H., Raichlen, D. A., Dunsworth, H., Hare, B., Walker, K., et al.
1670–1675. doi:10.1152/jappl.1999.86.5.1670
(2016). Metabolic acceleration and the evolution of human brain size and life history.
Gagnon, C. M., Steiper, M. E., and Pontzer, H. (2018). Elite swimmers do not exhibit a Nature 533, 390–392. doi:10.1038/nature17654
body mass index trade-off across a wide range of event distances. Proc. R. Soc. B Biol. Sci.
Prang, T. C. (2015). Calcaneal robusticity in Plio-Pleistocene hominins: implications
285, 20180684. doi:10.1098/rspb.2018.0684
for locomotor diversity and phylogeny. J. Hum. Evol. 80, 135–146. doi:10.1016/j.jhevol.
Gordon, A. D. (2013). “Sexual size dimorphism in Australopithecus: current understanding 2014.09.001
and new directions,” in The Paleobiology of Australopithecus vertebrate paleobiology and
Raichlen, D. A., Gordon, A. D., Muchlinski, M. N., and Snodgrass, J. J. (2010). Causes
paleoanthropology. Editors K. E. Reed, J. G. Fleagle, and R. E. Leakey (Dordrecht: Springer
and significance of variation in mammalian basal metabolism. J. Comp. Physiol. B 180,
Netherlands), 195–212. doi:10.1007/978-94-007-5919-0_13
301–311. doi:10.1007/s00360-009-0399-4
Jones, E. J., Bishop, P. A., Woods, A. K., and Green, J. M. (2008). Cross-sectional area and
Stamos, P. A., and Alemseged, Z. (2023). Hominin locomotion and evolution in the late
muscular strength. Sports Med. 38, 987–994. doi:10.2165/00007256-200838120-00003
miocene to late pliocene. J. Hum. Evol. 178, 103332. doi:10.1016/j.jhevol.2023.103332
Lieberman, D. E., Raichlen, D. A., Pontzer, H., Bramble, D. M., and Cutright-Smith, E.
Stickley, C. D., Andrews, S. N., Parke, E. A., and Hetzler, R. K. (2018). The
(2006). The human gluteus maximus and its role in running. J. Exp. Biol. 209,
effectiveness of scaling procedures for comparing ground reaction forces.
2143–2155. doi:10.1242/jeb.02255
J. Biomechanics 77, 55–61. doi:10.1016/j.jbiomech.2018.06.021
Lieberman, D. E., Venkadesan, M., Werbel, W. A., Daoud, A. I., D’Andrea, S., Davis, I.
Sylvester, A. D., Lautzenheiser, S. G., and Kramer, P. A. (2021). Muscle forces and the
S., et al. (2010). Foot strike patterns and collision forces in habitually barefoot versus
demands of human walking. Biol. Open 10, bio058595. doi:10.1242/bio.058595
shod runners. Nature 463, 531–535. doi:10.1038/nature08723
Wall-Scheffler, C. M., Chumanov, E., Steudel-Numbers, K., and Heiderscheit, B. (2010).
Lieberman, D. E., Warrener, A. G., Wang, J., and Castillo, E. R. (2015). Effects of stride
Electromyography activity across gait and incline: the impact of muscular activity on human
frequency and foot position at landing on braking force, hip torque, impact peak force
morphology. Am. J. Phys. Anthropol. 143, 601–611. doi:10.1002/ajpa.21356
and the metabolic cost of running in humans. J. Exp. Biol. 218, 3406–3414. doi:10.1242/
jeb.125500 Wang, W., Crompton, R. H., Carey, T. S., Günther, M. M., Li, Y., Savage, R., et al.
(2004). Comparison of inverse-dynamics musculo-skeletal models of AL 288-
Lovejoy, C. O., Suwa, G., Spurlock, L., Asfaw, B., and White, T. D. (2009). The pelvis
1 Australopithecus afarensis and KNM-WT 15000 Homo ergaster to modern
and femur of Ardipithecus ramidus: the emergence of upright walking. Science 326,
humans, with implications for the evolution of bipedalism. J. Hum. Evol. 47,
71–71e6. doi:10.1126/science.1175831
453–478. doi:10.1016/j.jhevol.2004.08.007
McNutt, E. J., Zipfel, B., and DeSilva, J. M. (2018). The evolution of the human foot.
Wells, J. C. K. (2007). Sexual dimorphism of body composition. Best Pract. Res. Clin.
Evol. Anthropol. 27, 197–217. doi:10.1002/evan.21713
Endocrinol. Metabolism 21, 415–430. doi:10.1016/j.beem.2007.04.007
Muchlinski, M. N., Snodgrass, J. J., and Terranova, C. J. (2012). Muscle mass scaling in
Wells, J. C. K. (2012). Sexual dimorphism in body composition across human
primates: an energetic and ecological perspective. Am. J. Primatology 74, 395–407.
populations: associations with climate and proxies for short- and long-term energy
doi:10.1002/ajp.21990
supply. Am. J. Hum. Biol. 24, 411–419. doi:10.1002/ajhb.22223
Myatt, J. P., Crompton, R. H., and Thorpe, S. K. S. (2011). Hindlimb muscle
Wickiewicz, T. L., Roy, R. R., Powell, P. L., Perrine, J. J., and Edgerton, V. R. (1984).
architecture in non-human great apes and a comparison of methods for analysing
Muscle architecture and force-velocity relationships in humans. J. Appl. Physiology 57,
inter-species variation: comparing hindlimb muscle architecture in non-human great
435–443. doi:10.1152/jappl.1984.57.2.435
apes. J. Anat. 219, 150–166. doi:10.1111/j.1469-7580.2011.01383.x
Wiseman, A. L. A. (2023). Three-dimensional volumetric muscle reconstruction of
Nevill, A. M., Stewart, A. D., Olds, T., and Holder, R. (2004). Are adult physiques
the Australopithecus afarensis pelvis and limb, with estimations of limb leverage. R. Soc.
geometrically similar? The dangers of allometric scaling using body mass power laws.
Open Sci. 10, 230356. doi:10.1098/rsos.230356
Am. J. Phys. Anthropol. 124, 177–182. doi:10.1002/ajpa.10351

Frontiers in Earth Science 13 frontiersin.org