AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 151:544–557 (2013)

Bite Force and Occlusal Stress Production in Hominin

Evolution
Carolyn M. Eng,* Daniel E. Lieberman, Katherine D. Zink, and Michael A. Peters

Department of Human Evolutionary Biology, Harvard University, Cambridge, MA 02138

KEY WORDS masticatory biomechanics; fossil hominins; diet

ABSTRACT Maximum bite force affects craniofacial not confirm our hypothesis of isometry. Specimens in the
morphology and an organism’s ability to break down genus Homo fall below the regression line describing the
foods with different material properties. Humans are relationship between bite force and molar area for non-
generally believed to produce low bite forces and spend human anthropoids and australopiths. These results
less time chewing compared with other apes because suggest that Homo species generate maximum bite
advances in mechanical and thermal food processing forces below those predicted based on scaling among
techniques alter food material properties in such a way australopiths and nonhuman primates. Because this
as to reduce overall masticatory effort. However, when decline occurred before evidence for cooking, we
hominins began regularly consuming mechanically proc- hypothesize that selection for lower bite force production
essed or cooked diets is not known. Here, we apply a was likely made possible by an increased reliance on
model for estimating maximum bite forces and stresses nonthermal food processing. However, given substantial
at the second molar in modern human, nonhuman pri- variability among in vivo bite force magnitudes meas-
mate, and hominin skulls that incorporates skeletal data ured in humans, environmental effects, especially varia-
along with species-specific estimates of jaw muscle archi- tions in food mechanical properties, may also be a factor.
tecture. The model, which reliably estimates bite forces, The results also suggest that australopiths had ape-like
shows a significant relationship between second molar bite force capabilities. Am J Phys Anthropol 151:544–
bite force and second molar area across species but does 557, 2013. VC 2013 Wiley Periodicals, Inc.

Apes such as chimpanzees spend 40% of the day intensive over the lower paleolithic (e.g., in the Acheu-
feeding and chewing, whereas humans today, including lian), but it certainly became more elaborate during the
hunter-gatherer populations, spend less than 5% of the middle paleolithic (Klein, 2009). Wrangham et al. (1999)
day chewing foods with a higher energy density that posit that cooking began at least two million years ago,
have been cooked or mechanically processed (Organ but the oldest evidence for controlled fire is dated to 1.0
et al., 2011). Because processed foods are usually softer Ma (Berna et al., 2012), and cooking hearths with burnt
and less tough, they usually require less time and force bones become common only in the middle paleolithic
to comminute. Yet, despite much interest in how and (James, 1989; Rowlett, 2000; Goldberg et al., 2001).
when the transition from more to less mechanically Interpretations of the fossil record have led to less
demanding diets occurred, we lack reliable data on bite consensus about changes in diet and the origin of cook-
force and stress generation capabilities in nonhuman ing and mechanical processing. One line of evidence is
primates and fossil hominins. Here, we build on existing postcanine tooth size, which increases in Australopithe-
models to estimate bite forces and occlusal stresses in a cus, becomes even larger in robust species such as Au.
large sample of nonhuman primate and hominin skulls. boisei, and then decreases in Homo Postcanine crown
These estimates incorporate human and nonhuman pri- size is similar in H. habilis and Au. africanus but is
mate masticatory muscle architecture data to account about 25% smaller in early African H. erectus, and then
for variations in both skeletal geometry and muscle decreased roughly 10% over the Pleistocene (Wood, 1991;
architecture. Additionally, we explore how different cor- McHenry and Coffing, 2000; Lieberman, 2011). Several
rection factors (CFs) for the relationship between the
skeletal estimate of jaw adductor muscle cross-sectional
area (CSA) and muscle physiological cross-sectional area Additional Supporting Information may be found in the online
(PCSA) influence estimates of bite force generation. version of this article.
Archaeological and comparative data provide the most
direct evidence for the transition from unprocessed to Grant sponsor: American School for Prehistoric Research
mechanically processed or cooked diets. Chimpanzees (Peabody Museum, Harvard University).
occasionally use rocks to break open hard foods like nuts
and palm seeds (Goodall, 1986; Boesch and Boesch, *Correspondence to: Carolyn M. Eng, Department of Human
Evolutionary Biology, Harvard University, 11 Divinity Ave.,
1990), and although such behaviors leave little archaeo-
Cambridge, MA 02138. E-mail: [email protected]
logical trace, early hominins likely used rocks to pound
open some foods, and perhaps to grind them before Received 9 March 2012; accepted 24 April 2013
chewing. Evidence for the use of stone tools is at least
2.6 Ma old (de Heinzelin et al., 1999; Semaw et al., DOI: 10.1002/ajpa.22296
2003), and possibly extends to 3.4 Ma (McPherron et al., Published online 11 June 2013 in Wiley Online Library
2010). Food processing may have become slightly more (wileyonlinelibrary.com).

Ó 2013 WILEY PERIODICALS, INC.

 BITE FORCE AND STRESS IN HOMININS 545
researchers suggest that these decreases were made pos- size and lever arms that are proportional to actual meas-
sible by cooking (Wrangham et al., 1999; Lucas, 2004; ures of force) increased isometrically with occlusal area,
Carmody and Wrangham, 2009), but mechanical food suggesting that molar stress was relatively constant
processing techniques such as grinding and pounding across extant primates, humans, and hominins. If so,
also alter food mechanical properties considerably and then the relatively larger jaw adductor muscles of robust
may explain these decreases. The hypothesis that the ge- australopiths increased bite force generation, but in pro-
nus Homo consumed less tough food is supported by den- portion to increases in molar surface area, resulting in
tal microwear analyses, which document a decrease in no change in bite stresses. Similarly, decreases in bite
the prevalence of features such as scratches and pits force in the genus Homo coupled with smaller molar sur-
between the genus Australopithecus and Homo (Ungar face areas, resulted in bite stresses similar to those of
and Scott, 2009). fossil hominins and nonhuman primates. Demes and
Another line of evidence used to make inferences Creel’s seminal analysis, however, has several limita-
about shifts in hominin diet and bite force production tions. First, masseter CSAs were not measured directly,
has been variations in skull morphology. Mastication but instead were estimated as 80% of temporalis CSA
generates bending and twisting strains within ranges based on relative sizes of the masseter and temporalis
known to stimulate bone growth and remodeling measured in orangutans and humans. This estimate
(Hylander, 1979; Hylander and Johnson, 1997; Hylander may be problematic because of variation in the relative
et al., 1998). Australopiths, especially Au. boisei and Au. size of the masseter and temporalis across species. In
robustus, have more robust mandibular corpora than addition, Demes and Creel did not include the medial
humans or apes, a likely adaptation for resisting high pterygoid, and they estimated muscle moment arms only
and repeated bending forces in the transverse and sagit- from their origin on the cranium without using their
tal planes as well as longitudinal twisting (Hylander, mandibular insertions. Finally, Demes and Creel did not
1988; Daegling and Grine, 1991). Mandibular robusticity account for variation in muscle architecture across spe-
relative to body mass is similar in H. habilis and gracile cies, particularly PCSA, a measure of the sum of the
australopiths such as Au. africanus, but declines consid- CSA of all muscle fibers, which is directly proportional
erably in H. erectus (Chamberlain and Wood, 1985; to a muscle’s maximum isometric force (Powell et al.,
Wood and Aiello, 1998). Experiments examining facial 1984). Because pennation allows more fibers to be
strains (Hylander et al., 1991; Hylander and Johnson, packed in a space and increases a muscle’s PCSA, esti-
1994; Ross and Hylander, 1996; Ross, 2001) along with mating muscle CSAs from skeletal dimensions ignores
finite element models (Strait et al., 2009) indicate that the effects of pennation and likely underestimates
the relatively taller, wider, and more buttressed face of PCSAs, and hence maximum muscle force. Finally, esti-
australopiths, especially the more robust forms, would mating unitless BFEs rather than absolute values of
have effectively resisted vertical bending and shearing, force complicates interpretations of scaling relationships
as well as twisting in the coronal plane. Although reduc- between force and tooth area and comparison with in
tions in face and jaw robusticity in Homo are often inter- vivo measures of bite force from other studies.
preted as evidence for a less mechanically demanding O’Connor et al. (2005) expanded on Demes and Creel
diet, facial flattening likely reduces in vivo stresses gen- (1988) to compare anterior bite force production capabil-
erated by masticatory forces (Lieberman et al., 2004; ities in late Homo. O’Connor et al. estimated absolute
Wroe et al., 2010). bite forces by measuring masticatory lever arms and
A final line of evidence, considered here, is to estimate muscle CSAs from crania and mandibles. A muscle
maximum bite force capabilities from craniofacial mor- architecture CF was calculated for each muscle by divid-
phology. The most frequently assessed parameter is ing the average muscle CSAs of modern humans by
effective mechanical advantage (EMA, the ratio of a PCSA values measured from computed tomography (CT)
muscle’s in-lever and out-lever) of the jaw adductors, images. The muscle-specific CFs were applied to all taxa
especially in the sagittal plane (e.g., Rak, 1983; Demes and were multiplied by muscle CSA to estimate muscle
and Creel, 1988; Lieberman, 2011). These studies docu- PCSA. Muscle PCSA estimates were then used to esti-
ment that anterior advancement of the zygomatic arch, mate muscle forces, which were combined with lever
lowering the occlusal plane relative to the TMJ, and fa- arms to estimate maximum bite forces. O’Connor et al.
cial retraction augment the EMA of the masseter and found that Neanderthals and modern humans have simi-
temporalis muscles in Australopithecus relative to great lar anterior bite force capabilities, and concluded that
apes, and in robust versus gracile australopiths. A com- overall skull size and robusticity best predicts maximum
mon inference is that natural selection will favor more bite force. O’Connor et al. (2005) is a substantial
efficient torque production in species that chew more advance over Demes and Creel (1988), but they focused
mechanically demanding diets. EMA estimates, however, primarily on anterior bite forces in only the genus
quantify only efficiency rather than actual bite force. Homo. In addition, muscle architecture CFs were based
Additionally, most bite force analyses estimate maximum on estimates of muscle architecture measured from CT
bite force, rather than chewing force, the forces used to images, but the accuracy of imaging-based methods of
break down foods. These caveats aside, maximum bite estimating PCSA is reduced across muscles (Weijs and
force estimates are useful for evaluating the maximum Hillen, 1984; Goto et al., 2005; Eng et al., 2007).
capabilities of the masticatory system, especially the Finally, Wroe et al. (2010) further elaborated the
ability to process mechanically demanding foods. methods used by Demes and Creel (1988) and O’Connor
Inferences about the evolution of food processing must et al. (2005) to estimate bite forces in a human, several
also consider that comminution is governed by the nonhuman primates, and two fossil hominins, Au. africa-
amount of stress (force per unit area) applied to a food nus and Au. boisei, using a 3D finite element analysis
item (Preuschoft, 1989; Lucas et al., 1994). In a land- (FEA). They found that the large EMA in humans allows
mark study, Demes and Creel (1988) showed that bite them to produce relatively high bite forces while exert-
force equivalents (BFEs, relative measures of muscle ing relatively low muscle forces, especially in the

American Journal of Physical Anthropology

546 C.M. ENG ET AL.

anterior dentition. Furthermore, scaled to produce the do not need to generate high bite forces and would have
same bite force, peak transitory forces were relatively lower PCSAs and thus, higher CFs. Therefore, we pre-
low in the human cranium and mandible relative to dict that humans will have relatively higher CFs com-
other taxa. Wroe et al. argue that these data provide evi- pared with nonhuman primates. This hypothesis is
dence that the human jaw improves performance for important to test because if muscle architecture differs
hard object feeding. Although 3D FEA has advantages between humans and nonhuman primates, and only spe-
over models that use 2D data and approximate muscles cies in the genus Homo had human-like jaw muscle
as a single vector, Wroe et al. modeled only one female architecture, then bite force estimates that do not
per group, and just two fossil hominins, neither from the account for muscle architecture differences may under-
genus Homo. Additionally, because Wroe et al. use the or over-estimate differences between Homo and other
muscle CFs from O’Connor et al. (2005), their estimates primates.
of bite force do not account for the differences in muscle
architecture across humans and nonhuman primates. Hypothesis 2
Here, we build upon the approaches used by Demes
Following Demes and Creel (1988), molar bite force is
and Creel (1988), O’Connor et al. (2005), and Wroe et al.
predicted to correlate significantly and scale isometri-
(2010) to estimate molar bite force and stress generation
cally with molar occlusal area across anthropoids. The
in a broad sample of extant apes and fossil hominins.
basis for this prediction is that bite force is proportional
We use physical measures of muscle mass, fiber length,
to PCSA and thus scales to the power of 0.67 against
and pennation angle to calculate muscle PCSA and esti-
body mass. Using measures of M2 area from this study
mate muscle architecture CFs for each extant primate.
along with body mass estimates from Smith and Jungers
Species-specific CFs allow us to account for differences
(1997), we found that in our sample of primates, M2
in muscle architecture related to the wide variation in
area scales with body mass to the power of 0.46, so the
feeding behaviors across primates. Our goal is to esti-
predicted scaling coefficient of bite force relative to occlu-
mate molar bite force using both skeletal and muscular
sal area should be close to 1.46 (0.67/0.46).
parameters known to influence comminution efficiency
in vivo. We also test predictions for the scaling of maxi- Hypothesis 3
mum bite force and occlusal area (i.e., occlusal stress)
among taxa. Because of recent changes in food processing, we
predict that Homo species generate relatively low max-
BIOMECHANICAL MODEL AND HYPOTHESES imum bite forces compared with other primates and
fall below the regression line between molar bite force
Torque (t), the rotational force about an axis, is the and molar occlusal area for nonhuman primates and
product of an exerted force (F) and its moment arm (r). australopiths.
In a biological system, muscle torque (tin) must offset a
resistant torque (tout) to maintain equilibrium: tin 5 tout,
or Fin 3 rin 5 Fout 3 rout. Applying this equation to mas-
tication, the maximum magnitude of bite force produced MATERIALS AND METHODS
at any location along the dental row (FBite) is equal to Skull Sample
the muscle resultant force, FMuscle, times the ratio of the
muscle resultant in-lever arm (rMuscle) divided by the We calculated bite forces for male and female skulls
out-lever arm (rBite), using the primary jaw adductors, (crania plus mandibles) of six species of extant homi-
the masseter (mass.), medial pterygoid (MP), and tempo- noids, one old world monkey (Macaca fascicularis), and
ralis (temp.). Another factor to consider is differential eight fossil hominin species in Australopithecus and
recruitment of the working side (WS) and balancing side Homo (Table 1). Sample sizes in the extant species were
(BS) muscles. After Spencer (1998), we assume that BS chosen to be commensurate with the samples from fossil
muscle forces average 80% of WS forces for a typical taxa. In extant species, adult individuals with a full den-
molar. Therefore, overall bite force is estimated as: tal row and no apparent pathology were chosen from the
collections at the Museum of Comparative Zoology and
1:8ððF3rÞWS Mass: 1ðF3rÞWS Temp: 1ðF3rÞWS MP Þ Peabody Museum, Harvard University. The modern
FBite 5 (1) Homo sapiens sample consists of 12 late agricultural/
rBite early industrial individuals from the Campania and
Lazio regions of Italy. Hominin measurements were
The model and background information summarized obtained from casts. Presumptive males and females
above lead to three general hypotheses. were not available for all hominin species. Additionally,
no mandible was available for several fossil crania. In
Hypothesis 1 these instances, we measured mandibles of similar-sized
We predict that muscle architecture CFs (the ratio of individuals; for H. habilis, the OH 13 mandible was
skeletally measured CSA to muscle PCSA) differ paired with the KNM-ER 1813 cranium. For H. heidel-
between humans and nonhuman primates. Because a bergensis, the Mauer mandible was paired with the
muscle architecture CF indicates how a muscle’s mass is Kabwe and Petralona 1 crania and for H. erectus, the
arranged to modify its force-generating properties, spe- BK67 mandible was matched with the KNM-ER 3733
cies that comminute fracture resistant foods would bene- cranium.
fit from packing more fibers in their jaw adductor
muscles to increase muscle PCSA, hence, muscle force. Occlusal Area
These species would have relatively low CFs compared
to species with smaller PCSAs. Conversely, species that Occlusal area of the second upper molar (M2) was cal-
regularly consume extremely soft and/or processed diets culated as the product of buccolingual and mesiodistal

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 547

TABLE 1. Specimens used to estimate M2 bite force in this study

Fossil specimens Australopiths

Au. afarensis (n 5 1) Hadar composite (male)a
Au. africanus (n 5 1) STS 5 (female)b
Au. boisei (n 5 1) OH 5 (male)c
Homo species
H. habilis (n 5 1) KNM-ER 1813c with jaw from OH 13 (male)
H. heidelbergensis (n 5 3) Kabwed with Mauer jaw, Petralona 1e with Mauer jaw,
Atapuerca 5f
H. erectus (n 5 4) Zhoukoudian composite,b KNM-WT 15000,b KNM-ER 3733b
with the BK67 jaw, D2700e
H. neanderthalensis (n 5 6) La Ferrassie,d La Chapelle,g Amud I,h Le Moustier,h
Shanidar-1,h Gibraltar-1h
Pleistocene Grotte des Enfants 6,h Skhul V,h Predmosti 3h and 4,h
H. sapiens (n 5 7) Qafzeh 9,h Brno 3,h Dolni Vestonice 3h
Extant specimens M. fascicularis (n 5 8) 4 males, 4 females
H. lar (n 5 8) 4 males, 4 females
P. pygmaeus (n 5 5) 3 males, 2 females
G. gorilla (n 5 8) 4 males, 4 females
P. paniscus (n 5 2) 1 male, 1 female
P. troglodytes (n 5 12) 6 males, 6 females
H. sapiens (n 5 12) 6 males, 6 females
a
Individual molar area measure from Kimbel et al. (2004).
b
Mean molar area from Wood (1991).
c
Individual molar area measure from Wood (1991).
d
Molar area measured from cast.
e
Individual molar area measure from Rightmire (1993).
f
Molar area measure from cast (Rubas, personal communications).
g
Mean molar area from Brace and Mahler (1971).
h
Individual molar area measure from Frayer (1976).

diameters. In non-fossil specimens, these measurements The height and width of the medial and lateral pterygoid
were taken using digital calipers (accuracy, 0.01 mm) fol- plates were measured with calipers and modeled as an
lowing Wood (1991). Whenever possible, fossil measure- ellipse to estimate the medial pterygoid CSA. The CSA of
ments were taken from the literature (Brace and the masseter was estimated as the maximum projected dis-
Mahler, 1971; Frayer, 1976; Wood, 1991; Rightmire, tance (in superior view) between the zygomatic arch and
1993; Kimbel et al., 2004). Because molar area could not the mandibular ramus multiplied by the length of the mus-
be measured in the edentulous STS5 individual, second cle origin on the zygomatic arch. The reliability of muscle
molar area values for Au. africanus were used. area and lever arm measures were tested using repeated
measures made by a single observer. For all measure-
ments, the intraclass correlation coefficient for measures
Lever Arm Estimates made on a single skull was greater than 0.95 (P < 0.05).
Bite force out-levers were measured using digital cali- Muscle PCSAs were estimated from skeletal estimates
pers on each skull from the glenoid fossa center to the of CSA using species-specific muscle architecture CFs.
center of the occlusal surface of M2. Using scaled photo- For chimps, the masseter, temporalis, and medial ptery-
graphs of the skulls, in-levers were measured with goid muscles were dissected from a formalin-fixed adult
ImageJ software (NIH, Bethesda, MD) as the perpendic- male (Harvard University MCZ 61063). Dissected muscles
ular distance between the approximate line of action of were weighed and fiber length (Lf0 ) measured from three
each major jaw adductor and the glenoid fossa center. regions in each muscle using digital calipers. Pennation
Following Hiiemae (1971), the masseter’s line of action angle was measured with a goniometer as the angle
was estimated in lateral view from the midpoint of the between the fibers and the distal tendon. Under magnifi-
muscle origin on the zygomatic arch to the angle of the cation, muscle fiber bundles were isolated from each fasci-
mandible; the line of action of the temporalis was esti- cle and mounted on slides. Following Lieber et al. (1990),
mated from the centroid of the muscle’s origin (measured bundle sarcomere length (Ls0 ) was determined by laser
with ImageJ) to the coronoid process; and the line of diffraction and normalized fiber length (Lf) calculated as:
action of the medial pterygoid was estimated in infero-  
2:7lm
medial view (perpendicular to the muscle’s plane) Lf 5L0 f (2)
L0 S
between the lateral and medial pterygoid plates and the
centroid of the muscle’s insertion on the mandible. where Lf0 is measured fiber length and 2.7 lm is the
optimal sarcomere length for human muscle (Lieber
Muscle Cross-Sectional and Physiological et al., 1994). PCSA was calculated as:
Cross-Sectional Areas
M  cosu
For the temporalis, modeling clay was used to fill the PCSA5 (3)
q  Lf
temporal fossa between the zygomatic arch and the lateral
wall of the skull. The clay was then removed and scanned where M is muscle mass, u is pennation angle, and q is
on a flatbed scanner and the area measured in ImageJ. muscle density (1.0564 g/cm3; Mendez and Keys, 1960).

American Journal of Physical Anthropology

548 C.M. ENG ET AL.

These data were combined with PCSA values from addi- representing H. habilis precludes making statistical
tional adult chimps for the masseter (n 5 1; Spencer, comparisons with other species.
personal communications), temporalis, and medial ptery-
goid (n 5 2; Spencer, personal communications; Taylor, RESULTS
personal communications) to obtain an average PCSA
for each muscle. For human CFs, PCSA was calculated Differences in masticatory morphology among taxa
using measures of muscle mass, fiber length, and penna- vary in terms of the average EMA at M2 as well as
tion angle from the temporalis, masseter, and medial uncorrected total jaw adductor muscle CSA. Average
pterygoid measured by van Eijden et al. (1997) in 8 EMA of the jaw adductor muscles at M2 in australopiths
humans. Temporalis and masseter PCSA values reported is not significantly different from gorillas, orangutans,
in Taylor and Vinyard (2013) for P. pygmaeus (n 5 2), chimps, bonobos, gibbons, and macaques and most Homo
Hylobates sp. (n 5 3), G. gorilla (n 5 4), P. paniscus (n species except H. heidelbergensis (Table 2 and Support-
5 1), and Macaca sp. (n 5 4) were used to calculate CFs ing Information Table S1). Additionally, the uncorrected
in the remaining anthropoid species. To compare mass- total CSA of the masticatory muscles in australopiths is
eter PCSA between primates and humans, we used the not significantly different from all species (Table 2 and
sum of the masses of the deep and superficial masseter Supporting Information Table S2). The pattern in Homo
along with the superficial masseter normalized fiber is different. While the uncorrected total jaw adductor
length and pennation angle reported in Taylor and Vin- muscle CSA in modern H. sapiens is similar to that of
yard (2013) to calculate total masseter PCSA. all Homo species (Table 2 and Supporting Information
CSAs measured on the primate skulls were averaged Table S2), modern humans have a significantly greater
and divided by the average PCSA for each species to cal- average masticatory muscle EMA at M2 than H. heidel-
culate muscle architecture CFs for the temporalis, mass- bergensis and H. neanderthalensis (Table 2 and Support-
eter, and medial pterygoid. Temporalis PCSA was not ing Information Table S1). In addition, average jaw
reported for P. pygmaeus and P. paniscus, so the chimp adductor muscle EMA at M2 in modern humans (0.58 6
temporalis CF was used in these species. A medial ptery- 0.01 cm) is significantly larger than that of most other
goid PCSA is available only for chimps, so the chimp extant apes (Gorillas: 0.49 6 0.01 cm; Chimps: 0.51 6
medial pterygoid CF was used for all nonhuman primates. 0.01; Gibbons: 0.44 6 0.01; Supporting Information Ta-
Species-specific CFs were used to calculate muscle PCSA ble S1). Modern humans, orangutans, and bonobos have
for extant primates from skeletally measured CSA. Aus- similar average EMAs at M2 (Table 2 and Supporting In-
tralopith muscle PCSAs were calculated using the chimp formation Table S1).
CF, whereas Homo PCSAs were calculated using the Dissections and muscle architecture data from Taylor
human CFs. We consider using the human CFs for Homo and Vinyard (2013) were used to calculate muscle archi-
to be the most conservative choice for phylogenetic rea- tecture CFs as the muscle CSA measured on the skull
sons as well as the general morphological similarities divided by the average PCSA for temporalis, masseter,
between Homo and later humans. Estimating Homo and medial pterygoid muscles in each primate group
PCSAs using the chimp CFs allows us to test the effect of (Table 3). With the exception of the human masseter,
this assumption on bite force estimates. The normalized these CFs are considerably less than 1.0 (Fig. 1), indicat-
muscle CSAs were multiplied by muscle specific tension ing that the skeletally measured CSAs underestimate
(35 N/cm2) to calculate muscle force (Zajac, 1989). PCSAs, hence muscles’ capability to generate force. Note
also that the CFs vary across muscles and between spe-
Statistical Analyses cies. As predicted (H1), the human CFs fall two standard
deviations above the nonhuman ape CFs for the mass-
To test the hypothesis (H1) that muscle architecture eter (1.48 vs. 0.73 6 0.20; mean 6 2 SDs) and tempora-
CFs differ between humans and nonhuman primates, we lis (0.65 vs. 0.38 6 0.12). As a consequence of the high
examined whether the human masseter and temporalis CFs in humans, the uncorrected total muscle CSA esti-
CFs fell two standard deviations outside the mean CFs mates are similar between P. troglodytes and all Homo
for nonhuman apes. Least squares regression of log- species (Table 2 and Supporting Information Table S2)
transformed M2 bite force plotted against log-trans- but after correcting for muscle architecture differences,
formed M2 occlusal area were used to test the hypothesis P. troglodytes generates significantly larger muscle
(H2) that bite force scales isometrically with molar area. forces than all Homo species (Table 4 and Supporting In-
Prediction intervals on the regression line of M2 molar formation Table S3). If we assume that muscle size is
area and bite force for australopiths and nonhuman proportional to skeletally measured CSA and that the
anthropoids were used to test the hypothesis (H3) that accuracy of skeletal measures of CSA does not vary
Homo species generate relatively low bite forces. Multi- among taxa, these results suggest that the greater bite
ple stepwise regression was then used to examine the force capability of apes compared with humans is largely
relative contributions of muscle force and EMA to bite achieved through altered muscle architecture as opposed
force. ANOVA was used to compare total uncorrected to a gross increase in muscle size.
muscle CSA, total muscle force, M2 bite force, M2 area, Maximum estimated M2 bite forces vary widely across
and average EMA in the masseter, temporalis, and species (Table 4). Gorillas, orangutans, and australopiths
medial pterygoid across species in H. sapiens, P. troglo- generate the highest estimated maximum M2 bite forces,
dytes, P. pygmaeus, M. fascicularis, H. lar, and G. go- and gibbons and H. heidelbergensis generate the lowest
rilla. Pairwise comparisons were made using Dunnett’s M2 bite forces. In terms of estimated molar occlusal
T3 post hoc test, which accounts for unequal variance stress, chimps, bonobos, orangs, and macaques generate
across groups and is suitable for multiple comparisons the highest estimated M2 bite stresses, whereas species
(Dunnett, 1980). Significance was set to a 5 0.05. in the genus Homo generate the lowest estimated M2
Although we can make general comparisons between H. stresses (Table 5). Note that gorillas and australopiths
habilis and other species, the single individual have large second molar areas compared with other

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 549

TABLE 2. Masseter, temporalis, and medial pterygoid skeletally measured cross-sectional area and effective mechanical advantage
at M2 for the species used in this study
Skeletally measured cross-sectional area (cm2) Effective mechanical advantage
Masseter Temporalis Medial pterygoid Masseter Temporalis Medial pterygoid
H. sapiens 11.14 6 0.63 6.83 6 0.37 1.95 6 0.05 0.61 6 0.01 0.49 6 0.02 0.63 6 0.01
Male 10.60 6 0.75 6.65 6 0.61 2.00 6 0.08 0.60 6 0.02 0.47 6 0.03 0.62 6 0.02
Female 11.79 6 1.07 7.05 6 0.40 1.90 6 0.05 0.62 6 0.02 0.51 6 0.03 0.64 6 0.02
Pleistocene H. sapiens 12.68 6 0.29 7.21 6 0.35 2.91 6 0.15 0.53 6 0.04 0.43 6 0.04 0.35 6 0.03
H. habilis 12.04 7.34 2.18 0.60 0.53 0.29
H. heidelbergensis 10.40 6 0.17 6.17 6 0.33 3.74 6 0.15 0.47 6 0.01 0.35 6 0.02 0.26 6 0.02
H. erectus 10.41 6 0.68 6.68 6 0.54 2.54 6 0.16 0.54 6 0.04 0.48 6 0.02 0.45 6 0.06
H. neanderthalensis 11.11 6 0.33 7.93 6 0.50 3.25 6 0.18 0.48 6 0.04 0.39 6 0.02 0.38 6 0.04
Australopith average 20.26 6 4.87 14.45 6 4.06 3.71 6 0.74 0.62 6 0.02 0.43 6 0.02 0.48 6 0.02
Au. afarensis 22.46 16.35 3.26 0.62 0.41 0.52
Au. africanus 10.95 6.67 2.70 0.65 0.48 0.47
Au. boisei 27.37 20.35 5.15 0.59 0.41 0.45
P. troglodytes 12.72 6 0.76 9.95 6 0.65 2.06 6 0.03 0.52 6 0.01 0.45 6 0.01 0.57 6 0.01
Male 13.89 6 1.18 11.43 6 0.84 2.13 6 0.04 0.51 6 0.01 0.45 6 0.02 0.56 6 0.01
Female 11.55 6 0.78 8.48 6 0.53 2.00 6 0.05 0.54 6 0.01 0.44 6 0.01 0.58 6 0.01
P. paniscus 10.18 6 0.34 7.18 6 1.14 1.01 6 0.07 0.50 6 0.00 0.45 6 0.03 0.64 6 0.02
Male 10.53 8.33 1.08 0.51 0.49 0.63
Female 9.84 6.04 0.94 0.50 0.42 0.66
G. gorilla 27.74 6 2.31 20.70 6 2.09 4.68 6 0.44 0.48 6 0.01 0.38 6 0.01 0.61 6 0.01
Male 33.40 6 1.30 26.03 6 1.19 5.74 6 0.33 0.45 6 0.01 0.37 6 0.02 0.59 6 0.02
Female 22.08 6 1.39 15.37 6 0.20 3.62 6 0.18 0.51 6 0.01 0.38 6 0.01 0.64 6 0.02
P. pygmaeus 17.68 6 2.67 15.63 6 2.18 3.79 6 0.33 0.55 6 0.01 0.44 6 0.01 0.60 6 0.02
Male 21.70 6 1.70 18.99 6 1.30 4.29 6 0.16 0.53 6 0.01 0.42 6 0.01 0.59 6 0.03
Female 11.66 6 1.47 10.59 6 0.47 3.04 6 0.23 0.57 6 0.00 0.46 6 0.02 0.62 6 0.01
H. lar 4.37 6 0.19 2.80 6 0.13 1.06 6 0.04 0.42 6 0.01 0.44 6 0.01 0.46 6 0.01
Male 4.75 6 0.20 3.01 6 0.18 1.17 6 0.03 0.40 6 0.02 0.44 6 0.02 0.45 6 0.02
Female 4.00 6 0.17 2.60 6 0.15 0.96 6 0.02 0.43 6 0.02 0.44 6 0.02 0.47 6 0.01
M. fascicularis 4.50 6 0.23 4.00 6 0.30 1.07 6 0.07 0.46 6 0.02 0.46 6 0.01 0.57 6 0.03
Male 5.07 6 0.13 4.55 6 0.35 1.21 6 0.07 0.44 6 0.01 0.47 6 0.01 0.52 6 0.01
Female 3.94 6 0.14 3.46 6 0.32 0.93 6 0.06 0.47 6 0.04 0.44 6 0.00 0.62 6 0.03

TABLE 3. Muscle physiological cross-sectional area (PCSA), skeletally measured cross-sectional area (CSA), and muscle architec-
ture correction factors (CFs; skeletally measured cross-sectional area/muscle physiological cross-sectional area) for the masseter,
temporalis, and medial pterygoid muscles
Masseter Temporalis Medial pterygoid
2 2 2 2 2
CSA (cm ) PCSA (cm ) CF CSA (cm ) PCSA (cm ) CF CSA (cm ) PCSA (cm2) CF
a
Human 11.14 7.53 1.48 6.83 10.47 0.65 1.95 5.28 0.37
Chimpanzeeb 12.72 18.67 0.68 9.95 23.92 0.42 2.06 13.77 0.15
Bonoboc 10.18 16.77 0.61 7.18 NA 0.42d 1.01 NA 0.15
Gorilla 27.74 37.74 0.74 20.70 47.24 0.44 4.68 NA 0.15
Orangutan 17.68 20.25 0.87 15.63 NA 0.42d 3.79 NA 0.15
Gibbon 4.37 5.38 0.81 2.80 8.55 0.33 1.06 NA 0.15
Macaque 4.50 6.99 0.64 4.00 12.26 0.33 1.07 NA 0.15
a
PCSA values calculated from mass, fiber length, and pennation angle (n 5 8) reported in Van Eijden et al. (1997).
b
PCSA values measured in this study and averaged with PCSAs for the masseter (Spencer, personal communications), temporalis,
and medial pterygoid (Spencer and Taylor, personal communications).
c
PCSA values for masseter and temporalis in other nonhuman primates reported in Taylor and Vinyard (2013). See text for details.
d
In the absence of P. paniscus and P. pygmaeus temporalis PCSA data, the chimp temporalis CF was used for these species.

taxa, keeping their bite stresses comparatively low de- bite forces are not significantly different from any of the
spite high bite forces. On the other hand, M2 bite forces taxa examined due to the small sample size and large
are concentrated on smaller molar areas in chimps, variation among australopiths. Australopiths have the
bonobos, and macaques, leading to relatively higher M2 capacity to generate similar muscle forces compared
occlusal stresses (Table 5). with gorillas (3,113.97 6 743.48 N vs. 4,051.98 6 367.40
It is impossible to know when human-like muscle N) and orangutans (2,759.30 6 342.76 N). Additionally,
architecture evolved, but if we assume that the human australopiths generate similar magnitudes of maximum
CFs apply to the genus Homo and chimp CFs apply to M2 bite forces (Table 4) and bite stresses (Table 5) com-
australopiths, comparison of fossil hominins with extant pared with gorillas and orangutans. If we apply human
primates suggests that the masticatory capabilities of CFs, australopith muscle forces (1,608.49 6 395.84 N)
australopiths are similar to, or greater than, extant non- are more similar to chimp muscle forces (1,966.84 6
human apes. However, australopith muscle forces and 86.20 N) and they generate similar maximum M2 bite

American Journal of Physical Anthropology

550 C.M. ENG ET AL.

TABLE 4. Total masseter, temporalis, and medial pterygoid muscle force and molar bite force at M2, both calculated using muscle
cross-sectional areas measured on the skull and corrected using muscle architecture correction factors
Muscle force (Newtons) M2 bite force (Newtons)
H. sapiensa 816.49 6 32.22 825.75 6 45.17
Male 798.21 6 46.30 789.22 6 56.49
Female 838.41 6 47.69 869.57 6 74.53
Pleistocene H. sapiensa 963.95 6 33.14 754.72 6 51.18
H. habilisa 886.02 790.58
H. heidelbergensisa 932.56 6 24.12 582.34 6 17.88
H. erectusa 846.49 6 53.39 753.45 6 76.34
H. neanderthalensisa 997.99 6 25.94 738.52 6 61.90
Australopith averageb 3,113.97 6 743.48 2,796.01 6 595.45
Au. afarensis 3,281.32 2,989.65
Au. africanus 1,750.73 1,681.57
Au. boisei 4,309.85 3,716.82
P. troglodytes 1,966.84 6 86.20 1,780.18 6 71.42
Male 2,164.28 6 105.94 1,928.60 6 98.65
Female 1,769.40 6 76.68 1,631.76 6 62.43
P. paniscus 1,354.46 6 121.34 1,244.65 6 135.01
Male 1,475.80 1,379.66
Female 1,233.12 1,109.65
G. gorilla 4,051.98 6 367.40 3,420.10 6 251.55
Male 4,991.04 6 184.50 4,067.33 6 97.59
Female 3,112.92 6 89.41 2,772.86 6 80.55
P. pygmaeus 2,759.30 6 342.76 2,559.69 6 295.84
Male 3,287.97 6 187.37 2,978.83 6 250.14
Female 1,966.28 6 146.44 1,930.99 6 169.41
H. lar 734.04 6 26.33 578.19 6 19.93
Male 796.81 6 22.15 618.90 6 25.81
Female 671.27 6 10.83 537.48 6 9.09
M. fascicularis 920.92 6 56.67 806.93 6 45.39
Male 1,042.14 6 50.90 898.42 6 40.88
Female 799.70 6 51.01 715.45 6 48.60
a
Bite forces in all Homo species were normalized using a correction factor calculated from human muscle architecture data.
b
Bite forces in australopith species were normalized using a chimpanzee-derived correction factor.

Fig. 1. Muscle architecture correction factors (skeletally measured cross-sectional area/muscle physiological cross-sectional
area) for the human, chimpanzee, bonobo, gorilla, orangutan, gibbon, and macaque masseter, temporalis, and medial pterygoid
muscles.

forces compared with chimps (1,413.75 6 308.94 N vs. Australopith bite stresses are significantly smaller than
1,780.18 6 71.42 N). Australopiths are also more similar gorilla, chimp, and orangutan bite stresses when using
to chimps when using average chimp and human CFs. the human CFs.

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 551

TABLE 5. M2 molar area and M2 bite stress for the species used significance (P < 0.10; Supporting Information Tables S3
in this study and S4).
If we apply the chimp CFs to fossil Homo, Homo bite
Molar area M2 bite stress force estimates are significantly greater than modern H.
at M2 (cm2) (N/cm2)
sapiens bite forces (except for H. erectus, whose are simi-
H. sapiens 1.08 6 0.02 768.42 6 48.12 lar) and not significantly different from most nonhuman
Male 1.08 6 0.03 732.70 6 52.41 ape bite forces including chimps, bonobos, and orangu-
Female 1.08 6 0.04 811.28 6 88.17 tans (P > 0.05). However, Homo bite forces are signifi-
Pleistocene H. sapiens 1.27 6 0.08 613.12 6 61.02 cantly smaller than gorilla bite forces (P < 0.05). With
H. habilis 1.64 482.06 chimp CFs, bite stresses are similar between fossil
H. heidelbergensis 1.30 6 0.23 478.49 6 83.47
H. erectus 1.56 6 0.06 485.17 6 48.47
Homo and modern H. sapiens, except H. neanderthalen-
H. neanderthalensis 1.31 6 0.05 561.51 6 32.84 sis who has significantly greater bite stresses than mod-
Australopith average 2.61 6 0.50 1,078.92 6 178.74 ern H. sapiens (P < 0.05). With chimp CFs, bite stresses
Au. afarensis 2.12 1,410.21 are similar between fossil Homo specimens and nonhu-
Au. africanus 2.11 796.95 man apes. Finally, applying average chimp-human CFs
Au. boisei 3.61 1,029.59 to fossil Homo, we find that fossil Homo bite forces and
P. troglodytes 1.18 6 0.03 1,510.09 6 59.52 stresses are similar to modern H. sapiens bite forces and
Male 1.22 6 0.04 1,591.25 6 92.02 stresses and significantly smaller than chimp and gorilla
Female 1.15 6 0.04 1,428.93 6 66.96 bite forces and stresses (P < 0.05). Smaller bite forces
P. paniscus 0.85 6 0.08 1,462.32 6 15.19
Male 0.93 1,477.50
and stresses in fossil Homo compared with orangutans
Female 0.77 1,447.13 only approaches significance (P < 0.10) with the average
G. gorilla 2.60 6 0.15 1,312.05 6 41.05 chimp-human CFs.
Male 2.93 6 0.10 1,389.02 6 39.16 The regression between maximum M2 bite force and
2
Female 2.26 6 0.14 1,235.09 6 48.81 M molar area (Fig. 2) in all non-Homo species is highly
P. pygmaeus 1.71 6 0.09 1,478.38 6 107.79 significant with a slope of 0.90 (95% confidence interval:
Male 1.84 6 0.05 1,613.56 6 109.73 0.83–0.97; R2: 0.93), but does not confirm H2 that bite
Female 1.51 6 0.01 1,275.62 6 107.44 force scales isometrically (predicted slope of 1.46) with
H. lar 0.43 6 0.01 1,346.83 6 43.95 molar area in nonhuman primates. When Homo is
Male 0.46 6 0.02 1,355.06 6 90.87
Female 0.40 6 0.01 1,338.61 6 26.63
included in the analysis, the strength of the relationship
M. fascicularis 0.45 6 0.01 1,785.59 6 99.84 and slope both decrease (95% CI: 0.55–0.92; R2 5 0.45).
Male 0.47 6 0.01 1,921.62 6 122.32 As hypothesized (H3), species in the genus Homo fall
Female 0.44 6 0.02 1,649.56 6 138.65 below the regression line of M2 area and bite force for
australopiths and nonhuman primates, indicating a
reduced ability to generate M2 occlusal stress (Fig. 2).
Species in the genus Homo fall below the regression line
When compared with most earlier Homo, modern if we apply the average chimp-human CFs but do not
humans have a greater average EMA at M2, but similar fall below the line with chimp CFs.
maximum jaw adductor muscle forces and M2 bite forces.
Modern H. sapiens generate similar total jaw adductor DISCUSSION
muscle force compared with Pleistocene H. sapiens, Comparison With In Vivo Measures of Bite Force
H. erectus, and H. heidelbergensis, and significantly
lower muscle force compared with H. neanderthalensis Archaeological evidence suggests that advances in
(Table 4 and Supporting Information Table S3). Modern food processing techniques began with the Oldowan, and
H. sapiens generate similar M2 bite forces (Table 4 and that the regular use of cooking probably began in the
Supporting Information Table S4) and bite stresses middle Pleistocene. Both changes may have decreased
(Table 5 and Supporting Information Table S6) compared the selective advantage for having a large, robust masti-
with Pleistocene H. sapiens, H. neanderthalensis, and H. catory complex to generate and resist high bite stresses
erectus. Modern H. sapiens generate significantly higher during mastication. However, testing these and other
bite forces (Table 4; P < 0.05) but not significantly differ- hypotheses requires accurate and reliable estimates of
ent bite stresses compared with H. heidelbergensis (Ta- maximum bite force from fossils.
ble 5 and Supporting Information Table S6). It is worth Although we cannot test the accuracy of bite force esti-
noting that our modern human bite force estimates are mates for fossil hominins, the predicted M2 bite force
slightly higher in females than males. However, this dif- values of extant hominoids including humans correspond
ference is not statistically significant and is likely due to well with published empirically measured maximum bite
our small modern human sample size. forces (Table 6). Of note, our estimates of maximum M2
Modern H. sapiens have absolutely lower maximum bite force for humans fall within the range of experimen-
M2 bite forces and occlusal stresses than most extant tal measures of maximum bite force measured using a
apes. Compared with modern H. sapiens, chimps and broad array of techniques on humans from western pop-
gorillas have significantly higher M2 bite forces (Table 4 ulations (Klatsky, 1942; Pruim et al., 1980; Gibbs et al.,
and Supporting Information Table S4) and M2 occlusal 1981; Van Spronsen et al., 1989; Dean et al., 1992;
stresses (Table 5 and Supporting Information Table S6) Braun et al., 1995; Sinn et al., 1996). The only sample
because of significantly higher muscle forces in chimps with significantly higher bite forces is a Yupik (formerly
and gorillas compared with humans (Table 4 and Sup- known as Alaskan Eskimo) population (Waugh, 1937),
porting Information Table S3). Although bite stresses which is reported to have maximum bite forces nearly
are significantly higher in orangutans than humans twice as high as those of other human populations.
(Supporting Information Table S6), higher muscle and Yupik and Inuit skulls are reported to have enormous
bite forces in orangutans versus humans only approach masticatory muscles, anteriorly positioned jaw adductor

American Journal of Physical Anthropology

552 C.M. ENG ET AL.

Fig. 2. Logarithmic plot of M2 molar area and M2 bite force. The black symbols represent australopith species, white symbols
represent Homo species, and gray symbols represent extant nonhuman primates. Where identified, males are plotted as triangles
and females as squares. Specimens with unidentified sex are plotted as circles. The regression line (fit on nonaveraged data) is rep-
resented by the gray solid line and the gray dashed lines show the 95% prediction intervals. The scaling equation of bite force at
the second molar (BF) and second molar area (MA) is BF 5 3.02 3 MA0.74 when all specimens are included and BF 5 3.16 3
MA0.90 when Homo is excluded. The regression line shown here does not include Homo.

muscles, and low prognathism (Hylander, 1977). The lat- consider is the remarkable range of published in vivo
ter two features indicate a high EMA, whereas the for- human bite forces. In vivo bite forces not only range
mer would increase masticatory muscle force generation, across studies but also differ by as much 3-fold within
together producing extremely high occlusal forces that the same population (Fig. 3). Although this large vari-
may be adaptive for well-documented Yupik and Inuit ability is hypothetically unique to humans, we lack
feeding behaviors including chewing seal skin, eating enough in vivo bite force measures from other primates
frozen food and bone, and using the jaws as pliers (de to make such a conclusion. Variation among in vivo bite
Poncins, 1941). forces across studies may also be due to methodological
Although Yupik and Inuit populations may be adapted differences in transducer type and size and whether bite
for generating large bite forces, individual bite force force is measured unilaterally or bilaterally.
capabilities are also strongly affected by environmental In terms of nonhuman primates, there are few in vivo
factors during development. Modern humans that measurements available to test our results. Lucas et al.
engage in vigorous masticatory activity associated with (1994) estimated maximum bite force in orangutans by
bruxism (i.e., teeth grinding) and have related severe measuring the amount of force required to break Macad-
dental attrition generate relatively high bite forces (911 amia seeds that they observed orangutans fracturing in
N; Waltimo et al., 1994). However, our bite force esti- a single bite. Lucas et al. found that orangutans would
mates from a late agricultural/early industrial modern need 2,000 N of force to break the seeds, which is
human sample are probably not lower than Yupik and below our maximum M2 bite force estimate of 2,559.7 6
Inuit bite forces solely from chewing a diet of processed 295.8 N (Table 6). Our estimates of M2 bite force for
foods because similar maximum bite forces have been macaques, 806.9 6 45.4 N, are higher than measures
measured in other indigenous populations include the reported in the literature (Table 6). Hylander (1979)
Shipibo of Peru (601 N; n 5 24), the Otomi of Mexico measured molar bite force in macaques (Macaca fascicu-
(738 N; n 5 43), and army recruits from the Aymara of laris) using a transducer-mounted plastic block.
Peru (818 N; n 5 51; Neumann and DiSalvo, 1958). Hylander’s average maximum bite force measure of
While tortillas are a dietary staple for the Otomi, the 256.3 6 39.3 N (range: 205.0–333.43 N) falls below the
Shipibo and Aymara eat both domesticated and wild maximum M2 bite force estimated here. Dechow and
foods, highlighting how bite force capabilities are influ- Carlson (1990) measured molar forces while maximally
enced by a complex interplay of genetic, dietary, and stimulating the masticatory muscles in anesthetized
extramasticatory factors. Another cautionary note to macaques (Macaca mulatta). Using unilateral measures

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 553
to estimate bilateral forces, they estimated maximum Tests of Study Hypotheses
forces for males (369.3 6 26.3 N) and females (286.2 6
7.1 N), also below the values reported here. Our first hypothesis (H1) is that CFs between skele-
tally measured muscle CSAs and PCSAs measured on
TABLE 6. Comparison of empirical maximum bite force values jaw adductor muscles would be higher in humans than
with our M2 bite force estimates nonhuman primates. As hypothesized, nonhuman pri-
mate CFs were more than two standard deviations lower
Empirically measured than human CFs, suggesting that nonhuman primates
bite force values M2 bite force
have 55–147% larger PCSAs relative to skull size.
(Newtons) estimate (N)
Larger PCSAs indicate that nonhuman primate mastica-
H. sapiens 721 (245–1,246)a 825.8 6 45.2 tory muscles pack more fibers in the same amount of
(659.3–1,105.7) space than human muscles, allowing them to achieve
b
Male 500 6 22 higher bite forces despite lower EMAs. Among primates,
517.0 (245–934)c muscle architecture CFs varied between jaw adductor
652 6 426d
muscles most likely because these muscles (especially
756 6 109 (403–1,206)e 789.2 6 56.5
814.0 6 22.5 (417–1,280)f (659.3–986.6) the masseter and temporalis) evolved different roles
911 6 79 (660–1,150)g (Vinyard and Taylor; 2010). Conversely, variation in CFs
1,277 (890–1,548)h among muscles may reflect differences in accuracy of the
Female 368.1 6 20.0i 869.6 6 74.5 skeletal estimates of muscle size, particularly for the
377.3 (111–734)c (662.3–1,105.7) medial pterygoid whose complex attachments make it
412 6 23b difficult to approximate muscle CSA on the skull. Simi-
615.0 6 18.4 (342–914)f larly, differences in skull shape between chimps and
1,188 (890–1,450)h humans may lead to variability in our ability to accu-
P. pygmaeus 2,000–3,000j 2,559.7 6 295.8
rately measure CSA on the skull, ultimately causing dif-
(1,761.6–3,464.0)
M. fascicularis 256.3 6 39.3 (205–333)k 806.9 6 45.4 ferences in CFs between these species.
(617.7–988.0) The human masseter has an exceptionally high CF,
Male 369.3 6 26.3l 898.4 6 40.9 1.48, indicating that the skeletally measured CSA over-
(808.4–988.0) estimates the muscle’s PCSA. In contrast, the highly
Female 286.2 6 7.1l 715.5 6 48.6 pennate arrangement of the medial pterygoid leads to a
(617.7–850.0) low CF in chimps and humans (0.15 and 0.37, respec-
tively). As noted previously, the low medial pterygoid CF
Bite force values are mean 6 s.e.m. (range).
a may also be due to the difficulty of estimating its CSA
Gibbs et al. (1981).
b
Dean et al. (1992). on the skull. Low CF values in the nonhuman primate
c
Klatsky (1942). jaw adductor muscles suggest that estimates of mastica-
d
Van Spronsen et al. (1989). tory muscle force using skeletal CSAs underestimate
e
Pruim et al. (1980). muscle force capabilities by up to 85%. Therefore, care
f
Braun et al. (1995). should be taken when estimating muscle force from skel-
g
Waltimo et al. (1994). etal measurements of muscle size. Muscle architecture
h
Waugh (1937). CFs can be used to account for pennate muscle architec-
i
Sinn et al. (1996). ture, but CFs should be both species- and muscle-spe-
j
Lucas et al. (1994).
k cific. Previous bite force estimates from fossils should be
Hylander (1979).
l
Dechow and Carlson (1990). re-examined using human and nonhuman anthropoid
specific muscle architecture CFs. Additionally, care

Fig. 3. Mean and ranges of maximum bite forces estimated in this study for humans (white markers) and nonhuman apes
(gray markers), and in vivo bite force estimates (mean and range; black markers) from Gibbs et al. (1981) (1); Klatsky (1942) (2);
Pruim et al. (1980) (3); Braun et al. (1995) (4); Waltimo (1994) (5); and Waugh (1937) (6). Where identified, males are plotted as tri-
angles and females as squares. Circles are the average of male and female bite forces. The ranges of in vivo bite forces demonstrate
that bite force can vary up to 3-fold within populations.

American Journal of Physical Anthropology

554 C.M. ENG ET AL.

should be taken when choosing CFs for fossil species as size to estimate maximum molar bite forces for fossil
it influences bite force estimates. Using the human CFs and extant species. While our estimates of bite force
or the chimp-human average CFs, fossil Homo species from extant nonhuman primates are similar, we esti-
generate similar maximum M2 bite forces compared with mate lower bite forces for modern H. sapiens (825 vs.
modern humans but significantly greater bite forces 1,470 N), H. erectus (753 vs. 2,075 N), and Au. africanus
than modern humans when using the chimp CFs. (1,711 vs. 2,598 N) and higher estimates for Au. afaren-
Our results do not confirm our hypothesis (H2) that sis (3,047 vs. 2,305 N) and Au. boisei (3,783 vs. 3,471 N).
M2 bite force scales isometrically and with a high degree When comparing our average bite force values with
of correlation with molar occlusal area. As hypothesized the estimates of Wroe et al. (2010), our estimates fall
(H3), species in the genus Homo fall below the regres- above their values for female G. gorilla (2,773 N vs.
sion line between bite force and molar area in nonhu- 1,723 N), Au. boisei (3,717 vs. 2,161 N), female P. pyg-
man primates and australopiths, suggesting a reduced maeus (1,931 vs. 1,031 N), Au. africanus (1,682 vs. 831
ability to generate occlusal stress. N), and female P. troglodytes (1,632 vs. 1,511 N) but our
If cooking did not begin regularly until one million bite force estimates fall below their average estimated
years ago or more recently, then changes in the skull in bite force for female H. sapiens (870 vs. 1,317 N). There
early Homo such as smaller teeth, reduced facial but- are several reasons why our results may differ. First,
tressing and less robust mandibles, could be due to Wroe et al.’s estimates are based on only one female
increases in nonthermal food processing (Wood and Col- individual per species and thus may under represent av-
lard, 1999; Klein, 2009; Lieberman, 2011). Cutting, slic- erage bite forces. Additionally, because Wroe et al. used
ing, mechanical pounding, and other methods of extra- human CFs to estimate muscle PCSA in anthropoids,
oral mechanical processing decrease the size of food par- they likely underestimate nonhuman anthropoid muscle
ticles before ingestion and mastication. These processing PCSAs and subsequently, muscle forces (Taylor and Vin-
techniques also break down cell walls of plants and fi- yard, 2013). On the other hand, Wroe et al. (2010) esti-
brous connective tissue in meat, decreasing stiffness and mated H. sapiens bite force from a San hunter-gatherer,
toughness. Although later Homo was hypothesized to who may generate higher bite forces than our late agri-
fall further from the regression line than early Homo, cultural/early industrial modern human sample that
modern H. sapiens generate similar bite forces and likely consumed a highly processed diet. Additionally,
stresses compared to early Homo. These results are not they estimate three-dimensional bite forces, whereas
consistent with a two-stage shift in masticatory capabil- this study’s moment arms are solely in the sagittal plane
ities with the first related to first nonthermal food proc- and 3D bite force estimates may be more accurate than
essing and then later to cooking. Reduced variation in 2D estimates. Finally, Wroe et al. represented muscles
bite forces across Homo, however, limits the ability to as multiple force vectors, which may more accurately
infer major dietary transitions from occlusal or craniofa- represent muscles with complex attachments.
cial features. Larger sample sizes as well as additional Nonetheless, the results presented here agree with
sources of data such as enamel thickness and occlusal Wroe et al. (2010) that human bite force estimates are
topography will be necessary to test for changes in mas- influenced by their relatively greater EMAs compared
ticatory abilities within Homo. with nonhuman apes. While Wroe et al. found the great-
est relative difference in EMA at the anterior dentition
Comparison With Previous Studies and further discussed the implications for using teeth as
tools in Homo, our study is limited to molar bite force
These results help refine previous estimates of bite estimates. Wroe et al. argued that the large EMAs allow
force capabilities among hominins. Although Demes and humans to generate high bite forces with relatively low
Creel (1998) used unitless BFEs, there is good agree- muscle forces that may have permitted gracilization of
ment between our data and Demes and Creel, which the mandible and cranium in later Homo. Wroe et al.
both estimate bite force values in australopiths to be did not include earlier Homo species in their analysis,
similar to those of G. gorilla and P. pygmaeus. Similarly, but the results presented here indicate that modern
both studies found bite force values in Homo to be less humans generate similar or slightly higher bite forces
variable and slightly below what would be expected in than earlier Homo, suggesting that high bite force effi-
species of similar molar size based on values from the ciency in modern humans may not apply to other Homo
other primates and hominins. In addition, O’Connor species. Furthermore, it should be evident that mechani-
et al. (2005) calculated a maximum M1 bite force of 815 cal advantage is a poor proxy for bite force. Plotting sec-
N for H. sapiens, 790 N for H. neanderthalensis, and ond molar bite force against total jaw adductor muscle
1,650 N for P. pygmaeus. The average maximum second force and average EMA at M2 demonstrates that muscle
molar bite force calculated in this study was 825.75 N in force covaries significantly with M2 bite force across this
H. sapiens, 738.52 N for H. neanderthalensis, and sample (Fig. 4). Multiple stepwise regression confirms
2,559.69 N for P. pygmaeus. These broadly comparable that muscle force is the best predictor of bite force.
estimates differ primarily because of muscle architecture When EMA is added to the regression equation, the R2
CFs. The human CFs used here are similar to O’Connor value changes from only 0.966 to 0.976, suggesting that
et al. in the temporalis (0.65 vs. 0.64) and medial ptery- M2 bite force is increased by increasing muscle force
goid (0.37 vs. 0.26), but our masseter CF is 50% larger rather than muscle mechanical advantage.
(1.48 vs. 0.99). Additionally, we used a masseter CF cal-
culated from orangutan data and chimp muscle-derived Limitations
temporalis and medial pterygoid CFs to estimate orang-
utan bite force. These were 12–42% lower than the mus- Several limitations of this study merit consideration.
cle CFs used by O’Connor et al. (2005) and thus we One is that we used only a single static estimate of
estimated higher orangutan muscle forces and bite EMA, even though this ratio changes dynamically
forces. Constantino et al. (2010) use molar tooth chip throughout the jaw joint’s range of motion (Koolstra,

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 555
predicted based on scaling relationships among nonhu-
man primates and australopiths. We hypothesize that
the advent of nonthermal food processing technologies in
the lower paleolithic before cooking became common
decreased the selective pressure acting on cranial form
related to the generation of high bite stresses during
mastication. Our data also support earlier interpreta-
tions that australopiths had ape-like bite force capabil-
ities. Finally, M2 bite force values are more highly
correlated with muscle force than EMA across homi-
noids. Therefore, researchers should use caution when
using EMA as a proxy for bite force in fossil hominins,
and species-specific muscle architecture CFs should be
used to account for differences in muscle architecture
across primates.
Fig. 4. Total jaw adductor muscle force and average jaw
adductor muscle effective mechanical advantage plotted against ACKNOWLEDGMENTS
bite force for all specimens used in this study. Muscle force is a bet-
ter predictor of bite force than effective mechanical advantage. We thank the Associate Editor and two anonymous
reviewers for greatly improving the content of this manu-
script. We also thank Ian Tattersall, Gary Sawyer, Giselle
2002). Using a single EMA for each muscle also does not Garcia, and the rest of the staff at the American Museum
account for muscles having distinct functional compart- of Natural History, Michele Morgan in the Peabody Mu-
ments with different fiber directions, attachment areas, seum, and Hopi Hoekstra, Judy Chupasko, and Mark
and lines of action (van Eijden et al., 1997; Davis et al., Omura at the Museum of Comparative Zoology for their
2010; Wroe et al., 2010). Furthermore, we used 2D help in procuring specimens used in this study. We are
rather than 3D estimates of lever arms, which may grateful to Andrea Taylor and Mark Spencer for providing
cause us to under- or over-estimate bite force (McHenry muscle architecture data.
et al., 2007). Another limitation is that when estimating
M2 bite force, we assumed maximum activation of jaw LITERATURE CITED
adductor muscles, which is unlikely to occur commonly
in normal mastication. Berna F, Goldberg P, Horwitz LK, Brink J, Holt S, Bamford M,
Another limitation is sample size. Larger sample sizes Chazan |M. 2012. Microstratigraphic evidence of in situ fire
both across hominoids, and within fossil hominins are in the Acheulean strata of Wonderwerk Cave, Northern Cape
necessary not only for more accurate estimates of mean province, South Africa. Proc Natl Acad Sci USA 109:E1215-
molar bite forces, but also to better understand ranges of E1220.
variation. Furthermore, given the changes in diet that we Bodine SC, Roy RR, Eldred E, Edgerton VR. 1987. Maximal
force as a function of anatomical features of motor units in
assume occurred in Homo, using modern human muscle
the cat tibialis anterior. J Neurophysiol 6:1730-1745.
architecture to represent all Homo species may be errone-
Boesch C, Boesch H. 1990. Tool use and tool making in wild
ous. Muscle architecture data from hunter-gatherer popu-
chimpanzees. Folia Primatol (Basel) 54:86-99.
lations who do not chew highly processed foods may be
Brace CL, Mahler PE. 1971. Post-Pleistocene changes in the
more representative of earlier Homo, but these data are human dentition. Am J Phys Anthropol 34:191-203.
not currently available. Additionally, reduced variation in
Braun S, Bantleon HP, Hnat WP, Freudenthaler JW, Marcotte
our Homo sample does not reflect the large range of varia- MR, Johnson BE. 1995. A study of bite force, Part 2: relation-
tion in vivo bite forces that have been measured in ship to various cephalometric measurements. Angle Ortho-
humans. This suggests that our bite force estimates may dont 65:373-377.
be representative of our specific study sample and care Carmody RN, Wrangham RW. 2009. The energetic significance
should be taken when using these data to predict bite of cooking. J Hum Evol 57:379-391.
forces in other human populations. Chamberlain AT, Wood BA. 1985. A reappraisal of variation in
Another limiting factor to consider is muscle fiber types. hominid mandibular corpus dimensions. Am J Phys Anthro-
Stedman et al. (2004) found that a gene encoding a muscle pol 66:339-405.
protein isoform was inactivated in masticatory muscles Constantino PJ, Lee JJ-W, Chai H, Zipfel B, Ziscovici C, Lawn
around 2.4 Ma in the genus Homo, and may have led to a BR, Lucas PW. 2010. Tooth chipping can reveal the diet and
shift in muscle fiber type, muscle fiber size, and even whole bite forces of fossil hominins. Biol Lett 6:826-829.
muscle size. However, because there is conflicting evidence Daegling DJ, Grine FE. 1991. Compact bone distribution and
for a difference in muscle stress in fast versus slow muscle biomechanics of early hominid mandibles. Am J Phys Anthro-
fibers (Bodine et al., 1987; Lucas et al., 1987; Schiaffino pol 86:321-339.
and Reggiani, 2011), we would not expect a large influence Davis JL, Santana SE, Dumont ER, Grosse IR. 2010. Predicting
on muscle force of fiber type alone. Nonetheless, more bite force in mammals: two-dimensional versus three-dimen-
detailed analyses of the muscle fiber type and muscle fiber sional lever models. J Exp Biol 213:1844-1851.
size profile in human and nonhuman primate masticatory de Heinzelin J, Clark JD, White T, Hart W, Renne P, WoldeGab-
muscles may reveal other factors influencing bite force var- riel G, Beyene Y, Vrba E. 1999. Environment and behavior of
2.5-million-year-old Bouri hominids. Science 284:25-629.
iability among these species.
de Poncins G. 1941. Kabloona. New York: Reynal.
CONCLUSIONS Dean JS, Throckmorton GS, Ellis E, Sinn DP. 1992. A prelimi-
nary study of maximum voluntary bite force and jaw muscle
Efforts to model bite forces suggest that taxa in the efficiency in pre-orthognathic surgery patients. J Oral Maxil-
genus Homo generate maximum bite forces below those lofac Surg 50:1284-1288.

American Journal of Physical Anthropology

556 C.M. ENG ET AL.
Dechow PC, Carlson DS. 1990. Occlusal force and craniofacial craniofacial growth in a retrognathic face. J Hum Evol
biomechanics during growth in rhesus monkeys. Am J Phys 46:655–677.
Anthropol 83:219-237. Lieberman DE. 2011. The evolution of the human head. Cam-
Demes B, Creel N. 1988. Bite force, diet, and cranial morphol- bridge, MA: Belknap Press.
ogy of fossil hominids. J Hum Evol 17:657-670. Lucas PW. 2004. Dental functional morphology: how teeth
Dunnett CW. 1980. Pairwise multiple comparisons in the work. Cambridge: Cambridge University Press.
unequal variance case. J Am Stat Assoc 75:796-800. Lucas PW, Peters CR, Arrandale SR. 1994. Seed-breaking forces
Eng CM, Abrams GD, Smallwood LR, Lieber RL, Ward SR. exerted by orang-utans with their teeth in captivity and a
2007. Muscle geometry affects accuracy of forearm volume new technique for estimating forces produced in the wild. Am
determination by magnetic resonance imaging (MRI). J Bio- J Phys Anthropol 94:365-378.
mech 40:3261-3266. Lucas SM, Ruff RL, Binder MD. 1987. Specific tension measure-
Frayer DW. 1976. Evolutionary dental changes in upper paleo- ments in single soleus and medial gastrocnemius muscle
lithic and mesolithic human populations. Ann Arbor, MI: Uni- fibers of the cat. Exp Neurol 95:142-154.
versity of Michigan. McHenry CR, Wroe S, Clausen PD, Moreno K, Cunningham E.
Gibbs CH, Mahan PE, Lundeen HC, Brehnan K, Walsh EK, 2007. Supermodeled sabercat, predatory behavior in Smilodon
Holbrook WB. 1981. Occlusal forces during chewing and swal- fatalis revealed by high-resolution 3D computer simulation.
lowing as measured by sound transmission. J Prosthet Dent Proc Natl Acad Sci USA 104:16010-16015.
46:443-449. McHenry HM, Coffing K. 2000. Australopithecus to Homo:
Goldberg P, Weiner S, Bar-Yosef O, Xu Q, Liu J. 2001. Site for- transformations in body and mind. Annu Rev Anthropol
mation processes at Zhoukoudian, China. J Hum Evol 41:483- 29:125-146.
530. McPherron SP, Alemseged Z, Marean CW, Wynn JG, Reed D,
Goodall J. 1986. The chimpanzees of Gombe: patterns of behav- Geraads D, Bobe R, Bearat HA. 2010. Evidence for stone-tool-
ior. Cambridge, MA: Belknap Press. assisted consumption of animal tissues before 3.39 million
Goto TK, Yahagi M, Nakamura Y, Tokumori K, Langenbach years ago at Dikika, Ethiopia. Nature 466:857-860.
GEJ, Yoshiura K. 2005. In vivo cross-sectional area of human Mendez J, Keys A. 1960. Density and composition of mamma-
jaw muscles varies with section location and jaw position. J lian muscle. Metabolism 9:184-188.
Dent Res 84:570-575. Neumann H, DiSalvo N. 1958. Caries in Indians of the Mexican
Hiiemae K. 1971. The structure and function of the jaw muscles Cordillera, the Peruvian Andes and at the Amazon head-
in the rat (Rattus norvegicus). Zool J Linn Soc 50:111-132. waters. Brit Dent J 104:13-16.
Hylander WL. 1977. The adaptive significance of Eskimo cranio- O’Connor CF, Franciscus RG, Holton NE. 2005. Bite force pro-
facial morphology. In: Dahlberg AA, Graber TM, editors. Oro- duction capability and efficiency in Neandertals and modern
facial growth and development. The Hague: Mouton. p 129- humans. Am J Phys Anthropol 127:129-151.
169. Organ C, Nunn CL, Machanda Z, Wrangham RW. 2011. Phylo-
Hylander WL. 1979. Mandibular function in Galago crassicau- genetic rate shifts in feeding time during the evolution of
datus and Macaca fascicularis: an in vivo approach to stress Homo. Proc Natl Acad Sci USA 108:14555-14559.
analysis of the mandible. J Morphol 159:253-296. Powell PL, Roy RR, Kanim P, Bello M, Edgerton VR. 1984. Pre-
Hylander WL. 1988. Implications of in vivo experiments for dictability of skeletal muscle tension from architectural deter-
interpreting the functional significance of “robust” australopi- minations in guinea pig hindlimbs. J Appl Physiol 57:1715-
thecine jaws. In: Grine FE, editor. Evolutionary history of the 1721.
“robust” australopithecines. New York: Aldine de Gruyter. p Preuschoft H. 1989. Biomechanical approach to the evolution of
55-83. the facial skeleton of hominoid primates. Fortschr Zool
Hylander WL, Johnson KR. 1994. Jaw muscle function and 35:421-431.
wishboning of the mandible during mastication in macaques Pruim G, De Jongh H, Ten Bosch J. 1980. Forces acting on the
and baboons. Am J Phys Anthropol 94:523-547. mandible during bilateral static bite at different bite force
Hylander WL, Johnson KR. 1997. In vivo bone strain patterns levels. J Biomech 13:755-763.
in the zygomatic arch of macaques and the significance of Rak Y. 1983. The australopithecine face. New York: Academic
these patterns for functional interpretations of craniofacial Press.
form. Am J Phys Anthropol 102:203-232. Rightmire GP. 1993. The evolution of Homo erectus: compara-
Hylander WL, Johnson KR, Picq PG. 1991. Masticatory stress tive anatomical studies of an extinct human species. Cam-
hypotheses and the supraorbital region of primates. Am J bridge: Cambridge University Press.
Phys Anthropol 86:1-36. Ross C, Hylander W. 1996. In vivo and in vitro bone strain
Hylander WL, Ravosa MJ, Ross CF, Johnson KR. 1998. Mandib- analysis of the circumorbital region of Aotus and the function
ular corpus strain in primates: further evidence for a func- of the postorbital septum. Am J Phys Anthropol 101:183-215.
tional link between symphyseal fusion and jaw adductor Ross CF. 2001. In vivo function of the craniofacial haft: the
muscle force. Am J Phys Anthropol 107:257-271. interorbital “pillar.” Am J Phys Anthropol 116:108-139.
James SR. 1989. Hominid use of fire in the lower and middle Rowlett RM. 2000. Fire control by Homo erectus in East Africa
pleistocene: a review of the evidence. Curr Anthropol 30:1-26. and Asia. Acta Anthropol Sin 19 (Suppl):198-208.
Kimbel WH, Rak Y, Johanson DC. 2004. The skull of Australo- Schiaffino S, Reggiani C. 2011. Fiber types in mammalian skele-
pithecus afarensis. New York: Oxford University Press. tal muscles. Physiol Rev 91:1447-1531.
Klatsky M. 1942. Masticatory stresses and their relation to den- Semaw S, Rogers MJ, Quade J, Renne PR, Butler RF, Domi-
tal caries. J Dent Res 21:387-390. nguez-Rodrigo M, Stout D, Hart WS, Pickering T, Simpson
Klein RG. 2009. The human career: human biological and cul- SW. 2003. 2.6-million-year-old stone tools and associated
tural origins, 3rd ed. Chicago: University of Chicago Press. bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia. J Hum
Koolstra J. 2002. Dynamics of the human masticatory system. Evol 45:169-177.
Crit Rev Oral Biol Med 13:366-376. Sinn DP, de Assis EA, Throckmorton GS. 1996. Mandibular
Lieber RL, Fazeli BM, Botte MJ. 1990. Architecture of selected excursions and maximum bite forces in patients with tempo-
wrist flexor and extensor muscles. J Hand Surg Am 15A:244- romandibular joint disorders. J Oral Maxillofac Surg 54:671-
250. 679.
Lieber RL, Loren GJ, Frid en J. 1994. In vivo measurement of Smith RJ, Jungers WL. 1997. Body mass in comparative prima-
human wrist extensor muscle sarcomere length changes. J tology. J Hum Evol 32:523-559.
Neurophysiol 71:874-881. Spencer MA. 1998. Force production in the primate masticatory
Lieberman DE, Krovitz GE, Yates FW, Devlin M, St. Claire M. system: electromyographic tests of biomechanical hypotheses.
2004. Effects of food processing on masticatory strain and J Hum Evol 34:25-54.

American Journal of Physical Anthropology

 BITE FORCE AND STRESS IN HOMININS 557
Stedman HH, Kozyak BW, Nelson A, Thesier DM, Su LT, Low electromyography during chewing across primates. Anat Rec
DW, Bridges CR, Shrager JB, Minugh-Purvis N, Mitchell MA. 293:572-582.
2004. Myosin gene mutation correlates with anatomical Waltimo A, Nystr€ am M, K€ an€
anen M. 1994. Bite force and den-
changes in the human lineage. Nature 428:415-418. tofacial morphology in men with severe dental attrition. Eur
Strait DS, Weber GW, Neubauer S, Chalk J, Richmond BG, J Oral Sci 102:92-96.
Lucas PW, Spencer MA, Schrein C, Dechow PC, Ross CF. Waugh L. 1937. Dental observations among Eskimos. J Dent
2009. The feeding biomechanics and dietary ecology of Aus- Res 16:355-356.
tralopithecus africanus. Proc Natl Acad Sci USA 106:2124- Weijs WA, Hillen B. 1984. Relationship between the physiologi-
2129. cal cross-section of the human jaw muscles and their cross-
Taylor AB, Vinyard CJ. 2013. The relationships among jaw- sectional area in computer tomograms. Acta Anat (Basel)
muscle fiber architecture, jaw morphology and feeding behav- 118:129-138.
ior in extant apes and modern humans. Am J Phys Anthropol Wood BA. 1991. Koobi fora research project: hominid cranial
151:120-134. remains. Oxford: Clarendon Press.
Wood BA, Aiello LC. 1998. Taxonomic and functional implica-
Ungar PS, Scott RS. 2009. Dental evidence for diets of early
tions of mandibular scaling in early hominins. Am J Phys
Homo. In: Grine FE, Leakey RE, Fleagle JG, editors. The first
Anthropol 105:523-538.
humans: origin and early evolution of the genus Homo. New
Wood BA, Collard M. 1999. The changing face of genus Homo.
York: Springer-Verlag. p 121-134.
Evol Anthropol 8:195-207.
van Eijden TMGJ, Korfage JAM, Brugman P. 1997. Architec- Wrangham RW, Jones JH, Laden G, Pilbeam D, Conklin-Brit-
ture of the human jaw-closing and jaw-opening muscles. Anat tain NL. 1999. The raw and the stolen: cooking and the ecol-
Rec 248:464-474. ogy of human origins. Curr Anthropol 40:567-594.
Van Spronsen P, Weijs W, Valk J, Prahl-Andersen B, Van Ginkel Wroe S, Ferrara TL, McHenry CR, Curnoe D, Chamoli U. 2010.
F. 1989. Comparison of jaw-muscle bite-force cross-sections The craniomandibular mechanics of being human. Proc R Soc
obtained by means of magnetic resonance imaging and high- London Ser B 277:3579-3586.
resolution CT scanning. J Dent Res 68:1765-1770. Zajac FE. 1989. Muscle and tendon: properties, models, scaling,
Vinyard CJ, Taylor AB. 2010. A preliminary analysis of the and application to biomechanics and motor control. Crit Rev
relationship between jaw-muscle architecture and jaw-muscle Biomed Eng 17:359-411.

American Journal of Physical Anthropology