Orangutan positional behavior and the nature of arboreal locomotion in Hominoidea

Abstract

The Asian apes, more than any other, are restricted to an arboreal habitat. They are consequently an important model in the interpretation of the morphological commonalities of the apes, which are locomotor features associated with arboreal living. This paper presents a detailed analysis of orangutan positional behavior for all age-sex categories and during a complete range of behavioral contexts, following standardized positional mode descriptions proposed by Hunt et al. ([1996] Primates 37:363-387). This paper shows that orangutan positional behavior is highly complex, representing a diverse spectrum of positional modes. Overall, all orthograde and pronograde suspensory postures are exhibited less frequently in the present study than previously reported. Orthograde suspensory locomotion is also exhibited less often, whereas pronograde and orthograde compressive locomotor modes are observed more frequently. Given the complexity of orangutan positional behavior demonstrated by this study, it is likely that differences in positional behavior between studies reflect differences in the interplay between the complex array of variables, which were shown to influence orangutan positional behavior (Thorpe and Crompton 2005 Am. J. Phys. Anthropol. 127:58-78). With the exception of pronograde suspensory posture and locomotion, orangutan positional behavior is similar to that of the African apes, and in particular, lowland gorillas. This study suggests that it is orthogrady in general, rather than forelimb suspend specifically, that characterizes the positional behavior of hominoids.

Full text

Orangutan Positional Behavior and the Nature
of Arboreal Locomotion in Hominoidea
Susannah K.S. Thorpe
1
* and Robin H. Crompton
2
1
School of Biosciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
2
Department of Human Anatomy and Cell Biology, University of Liverpool, Liverpool L69 3GE, UK
KEY WORDS Pongo pygmaeus; posture; orthograde clamber; forelimb suspend
ABSTRACT The Asian apes, more than any other, are
restricted to an arboreal habitat. They are consequently an
important model in the interpretation of the morphological
commonalities of the apes, which are locomotor features
associated with arboreal living. This paper presents a de-
tailed analysis of orangutan positional behavior for all age-
sex categories and during a complete range of behavioral
contexts, following standardized positional mode descrip-
tions proposed by Hunt et al. ([1996] Primates 37:363–387).
This paper shows that orangutan positional behavior is
highly complex, representing a diverse spectrum of posi-
tional modes. Overall, all orthograde and pronograde sus-
pensory postures are exhibited less frequently in the pres-
ent study than previously reported. Orthograde suspensory
locomotion is also exhibited less often, whereas pronograde
and orthograde compressive locomotor modes are ob-
served more frequently. Given the complexity of orangu-
tan positional behavior demonstrated by this study, it is
likely that differences in positional behavior between
studies reflect differences in the interplay between the
complex array of variables, which were shown to influence
orangutan positional behavior (Thorpe and Crompton [2005]
Am. J. Phys. Anthropol. 127:58–78). With the exception
of pronograde suspensory posture and locomotion, orang-
utan positional behavior is similar to that of the African
apes, and in particular, lowland gorillas. This study sug-
gests that it is orthogrady in general, rather than fore-
limb suspend specifically, that characterizes the posi-
tional behavior of hominoids. Am J Phys Anthropol 000:
000–000, 2006. V
V
C2006 Wiley-Liss, Inc.
With the exception of ourselves, the living apes are
predominantly tropical forest dwellers (Dohlinow and
Fuentes, 1999), and their preferred foods (sensu Wrang-
ham et al., 1998) are predominantly ripe fruits (Pilbeam,
2002), harvested directly from the trees. The morphologi-
cal commonalities of all apes, such as a broad trunk, dor-
sally placed scapula, and an ability to raise the arm
above the head, are generally accepted as primarily loco-
motor features and confirm that the hominoids are, in
essence, an arboreal clade. Nevertheless, it is the Asian
apes that are now most restricted to locomotion in the
canopy. Chimpanzees, bonobos, mountain gorillas, and
even lowland gorillas spend a relatively large amount of
time on the ground, as reflected in their morphological
adaptations for quadrupedal knuckle-walking. While
orangutans may have further refined their morphological
adaptations to arboreal living since their split from the
common ancestor, as the only great ape to retain a pre-
dominantly arboreal lifestyle, they are an important model
in interpreting the shared morphological features of the
great apes.
However, orangutan locomotion has so far been only
very broadly described. Sugardjito and Van Hooff (1986)
provided a broad-ranging description of Sumatran orangu-
tan positional behavior during travel and rest, but grouped
48% of locomotion together under the category ‘‘quadruma-
nous scrambling.’’ Cant (1987a) refined the classification of
locomotion during travel and feeding in Bornean orangu-
tans by distinguishing between modes according to the
orientation of the body, and the direction of movement.
However, the study by Cant (1987a) was limited to two
adult females, and was conducted during an unusual
drought, which may be expected to have influenced the
ranging and foraging patterns of his subjects. Cant (1987b)
also studied the postures exhibited during feeding in two
types of fig trees by Sumatran adult males and females,
and found significant differences in positional behavior
based on both gender and food type. However, none of
these studies provided a comprehensive overview of loco-
motion and posture during feeding and travel for a sin-
gle study group. Nor were they detailed enough to allow
us to develop an understanding of locomotor anatomy.
Recent advances in the standardization of positional
mode classifications (Hunt et al., 1996), which advocate
approaching locomotion from a biomechanical perspective,
have opened the way for a comprehensive and meaning-
ful analysis of primate positional behavior. This is partic-
ularly appropriate for orangutans, as they may be ex-
pected to exhibit a diverse spectrum of positional modes,
due to their large body mass combined with the mechan-
ical constraints imposed on arboreal locomotion by spa-
tial discontinuity in the canopy and the fragility and
compliance of arboreal supports (Cant, 1992).
Grant sponsor: European Commission; Grant sponsor: Indonesian
Government; Grant sponsor: Leverhulme Trust; Grant sponsor:
Royal Society; Grant sponsor: L.S.B. Leakey Foundation; Grant
sponsor: Natural Environment Research Council.
*Correspondence to: Dr. S.K.S. Thorpe, School of Biosciences, Uni-
versity of Birmingham, Edgbaston, Birmingham B15 2TT, UK.
E-mail: S.K.Thorpe@bham.ac.uk
Received 29 June 2005; accepted 21 December 2005.
DOI 10.1002/ajpa.20422
Published online in Wiley InterScience
(www.interscience.wiley.com).
V
V
C2006 WILEY-LISS, INC.
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 000:000–000 (2006)

As argued by Thorpe and Crompton (2005), standard
nonparametric tests (such as v
2
tests on contingency
tables for examining relationships between variable fre-
quencies) are usually fairly robust for two-dimensional
analyses. However, such an approach may lead to mis-
leading conclusions, if we attempt to examine multidi-
mensional relationships by analyzing a series of combi-
nations of two-dimensional tables (Gilbert, 1981; Agresti,
1990). A further difficulty arises when cell frequencies
are low or zero, as happens very quickly when examin-
ing locomotor data, even where the overall data set is
numbered in tens of thousands of observations. Thorpe
and Crompton (2005) demonstrated that relationships
between locomotion and related ecological variables for
orangutans in this study group were more complex than
could be adequately represented by two-dimensional
analysis. They used log-linear modeling to analyze multi-
ple relationships among locomotor and substrate vari-
ables. This technique, although sophisticated, does not
lend itself to the detailed characterization of positional
behavior that is required here. Consequently, the analy-
sis in this paper is restricted to statements of frequency,
without recourse to statistical testing. This paper pro-
vides the complete positional behavior data set used for
the log-linear analysis of orangutan locomotion in Thorpe
and Crompton (2005).
While the African apes have, to a greater or lesser
extent, forsaken an arboreal lifestyle, it remains a gener-
ally accepted hypothesis (e.g., Fleagle, 1988) that the hom-
inoids are characterized primarily by adaptations for
forelimb suspension. This study therefore tests the above
hypothesis with reference to the most arboreal of great
apes, by a detailed qualitative and quantitative descrip-
tion of orangutan positional behavior, for all age-sex cat-
egories and during a complete range of behavioral contexts.
Our main findings are set in context by comparison with
data for other apes and for outgroup comparators.
METHODS
Field study
The field study took place in the Ketambe Research
Station (38410North, 978390East) in the Gunung Leuser
National Park (Leuser Ecosystem, Aceh Tenggara, Suma-
tra, Indonesia). The area is covered by pristine rain forest
on riverine terraces located along the course of the Alas
River, and was described in detail by Rijksen (1978) and
van Schaik and Mirmanto (1985). The study was carried
out between December 1998–December 1999, and all
observations were made by a single observer (S.K.S.T.)
to ensure consistency. Considerable self-training in esti-
mating height and identification of positional types and
support types, angles, and diameters was carried out in
the first 6 weeks of the study, and at intervals through-
out the study. The majority of data were collected between
February–November 1999.
Ten individuals were observed, including adult and
immature individuals of both sexes (see Table 1 for sub-
ject information). Focal instantaneous sampling on the
1-min mark was employed to enable collection of detailed
data on support use and contextual behavior, in addition
to a fine-grained classification of positional behavior. Once
located, the orangutans were followed for a maximum of
5 consecutive days from when they left their night nest
until they built their next night nest, and on at least
two separate occasions.
Positional behavior classification
We obtained 28,797 instantaneous observations of posi-
tional behavior, 25,986 being of postural behavior, and
2,811 of locomotion. Great ape positional behavior is highly
complex, and it is thus difficult to identify the optimum
level of detail at which classification should occur. It is obvi-
ously impossible to differentiate every movement, and as
Hunt et al. (1996) argued, neither is it desirable, as it is
equally impossible to analyze the anatomical implica-
tions of each of a near-infinite roster of positional classi-
fications. Nevertheless, the aim of this study was to
identify fine-grained differences in orangutan positional
behavior, to shed light on some of the key issues sur-
rounding fine-grained similarities and distinctions in
great ape morphology, and thus we opted for a detailed
classification system that separates submodes that are
kinematically and mechanically distinct, but pools them
into modes that may be used for comparative purposes
in the future. To enhance clarity in the text and tables,
submode labels for the present study are italicized, and
modes are in bold.
The classification of positional behavior follows the
theoretical approach detailed by Hunt et al. (1996). The
area of the base of support is a key factor in the classifi-
cation system, and we tried to avoid conflating postures
where the area of the base is different, except where ac-
curacy is very limited because of the difficulty of observ-
ing orangutans in the canopy (e.g., sit and sit-out; Hunt
et al., 1996). Classifications are defined in terms of the
number of weight-bearing body parts (e.g., limbs, or seg-
ments such as the ischia), and whether each body part
appears to be under predominantly tensile (in suspen-
sion under superstrates) or compressive (above the sub-
TABLE 1. Study subjects
Age-sex category Name
Age
(years)
No. focal
days Notes
Adult male Bobby Unknown 10 Dominant male
AM2 Unknown 5 Unidentified young adult male
Adult female Yet ca. 34 14 Dominant female, travels with immature
son Yossa (see below)
Sina ca. 29 11 Mid-rank, pregnant, travels with immature son
Ans 24 11 Low rank, travels with semidependent infant
Subadult male X ca. 31 5
Yop 21 3 Eldest offspring of Yet
Immature female Chris 12 12 Independent daughter of Ans
Immature male Eibert 8 13 Independent
Yossa 7 13 Son of Yet
2S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

strate) stress regimes. Body-part positions are included
in the definition if they appear to support more than
their own mass. Orangutans were frequently observed to
combine two or three postural components into hybrid
postures to maximize stability or reach. For example, an
orangutan might sit, with one forelimb in suspension
and one hindlimb in compression, with all parts appear-
ing to support more than their own mass. To reflect these
hybrid combinations, the positional component that ap-
peared to bear the most weight (>1/2 body mass where two
components were combined, and >1/3 body mass where
three components were combined) was recorded first in the
definition (component 1). The component that bore the next
largest proportion of body mass became component 2, and
the third largest became component 3 (Appendix A). An
indication of the weight borne by each limb may be drawn
from the level of deformation of the supports and the extent
to which they rebound when an animal removes its weight;
the general position of the body, such as the position of the
trunk relative to the limbs and the support; and the appear-
ance of the cheiridia and limbs (Hunt et al., 1996).
In their classification system, Hunt et al. (1996) sepa-
rated unimanual and bimanual forelimb-suspend and
cling in their equivalent of component 1, but they did
not separate single and dual use of the hindlimbs in
hindlimb suspension, and they recommended that stand-
ing on one leg be classified as a bipedal stand. Also, in
most cases when there was more than one component in
a submode, they did not distinguish single and dual use
of the forelimbs in the initial or subsequent components.
Here, we separate single and dual use of the forelimbs
and hindlimbs in component 1 for suspension and stand,
but not other modes, both in an attempt to highlight key
kinematic and mechanical information specific to orang-
utans but to avoid overanalyzing positional behavior,
and because the frequency of unaided and hand-assisted
bipedalism was of particular interest to us. However, we
did not separate single and dual use of hindlimbs or
forelimbs in components 2 and 3. Similarly, flexion and
extension are recorded for component 1, but not for sub-
sequent components.
We also drew a distinction in component 1 between
stand, where the legs are in a vertical and (if both are
used) similar kinematic position to each other, and com-
pression, where the legs may be angled in any position
below horizontal and may be positioned on substrates
at different levels, with different degrees of abduction/
adduction and flexion/extension to each other. This dis-
tinction is necessary for orangutan posture because of
the complexity of the environment and their capacity for
supporting body mass using multiple limbs and multiple
supports anywhere in their sphere of reach.
Distinction is not made on the basis of footfall pattern
or hand/foot grip, because foliage often made this diffi-
cult to record, and footfall patterns and grips are outside
the intended scope of this study. Joints are regarded as
extended only when in full extension, or very nearly so.
All other joint positions are recorded as flexion.
To prevent the term ‘‘climb’’ being used to describe
locomotion in a horizontal direction, Hunt et al. (1996)
advocated that ‘‘climb’’ be used to describe locomotion
within 458of true vertical. Thorpe and Crompton (2005)
showed that the mechanics of orangutan climbing
between 20–458of true vertical were fundamentally dif-
ferent from climbing within 208of true vertical. Their
log-linear analysis of the relationships between orangu-
tan locomotion and associated variables found that loco-
motion was best understood when angled climb/descent
(i.e., 20–458) was included in the torso orthograde suspen-
sion mode rather than vertical climb/descent. Conse-
quently, in Appendix B, we followed the recommendations
of Hunt et al. (1996) by classifying all climbing activity
above 458as vertical climb and vertical descent, but we
also provided values for angled climb in parentheses.
Overall, 129 biomechanically distinct positional sub-
modes were identified: 82 representing posture, and 47
representing locomotion. These conflate into 13 postural
and 14 locomotor modes. Positional behavior definitions
and frequencies are presented in Appendix A (posture)
and Appendix B (locomotion).
Statistical analysis
The interdependence of observations is a particular
problem in the analysis of positional behavior. Sequen-
tial observations using a small time interval are thought
to produce results in which assumptions of independence
may be violated, so that sampling procedures may be
required (Boinski, 1989; Hunt, 1992; Dagosto, 1994; War-
ren and Crompton, 1997). Of the 28,797 bouts observed
in this study, only 2,811 were of locomotion. Poor visibil-
ity in this study and the tendency of orangutans to rest
frequently during bouts of locomotion meant that se-
quential observations of locomotion were rarely obtained.
Consequently, all locomotor bouts were analyzed. How-
ever, sequential observations of postural behavior were
frequently obtained. To remove the problems of interde-
pendence, our analysis is simply based on a 25% random
sample (SPSS sample, n ¼7,155). Table 2 shows that
postural mode frequencies in the 25% sample do not dif-
fer substantially from those for the complete data set.
RESULTS
Posture
Due to the variability of orangutan positional behavior,
we found it necessary to distinguish more precisely
between orthograde and pronograde postures than did
Hunt et al. (1996). Thus, at the mode level, pronograde
stand subsumes all pronograde compressive postures,
including those classified under stand mode by Hunt
et al. (1996) and contralateral and ipsilateral compres-
sion, which are submodes novel to orangutans. Similarly,
orthograde stand subsumes both bipedal and mono-
TABLE 2. Comparison of percentages of postural modes
in full data set and 25% sample
Total data set 25% sample
Sit 59.7 58.1
Squat 0.0 0.4
Cling 0.1 0.1
Pronograde stand 3.6 4.3
Orthograde stand 7.4 5.8
Cantilever 0.0 0.0
Orthograde forelimb-suspend 3.5 3.3
Orthograde
quadrumanous-suspend
1.5 2.1
Forelimb/hindlimb suspend 9.6 9.1
Pronograde suspend 3.7 3.7
Hindlimb-suspend 0.8 0.8
Lie 9.9 12.2
Postural bridge 0.1 0.1
Total 100.0 100.0
3ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

pedal stand and compression if the trunk is orthograde.
Suspensory behavior is classified in the same manner.
Thus, pronograde suspend subsumes the quadruma-
nous suspend mode of Hunt et al. (1996) and other pro-
nograde suspensory behavior specific to orangutans.
Orthograde forelimb-suspend compares well with the
forelimb suspend of Hunt et al. (1996). However, orangu-
tans exhibit a number of additional orthograde suspen-
sory postures in which the hindlimbs support an equal
or greater body mass than the forelimbs, e.g., orthograde
hindlimb-suspend. Thus, we named an additional mode:
orthograde quadrumanous suspend.
At the submode level, most additional classifications
reflect hybrid combinations, normally between two pos-
tural components. However, a novel submode exhibited
relatively often (1.2%) was pronograde stand/forelimb-
suspend. In this posture, orangutans stand with their
trunk and hindlimbs in a pronograde stand position, but
one or both forelimbs bear significant body mass in sus-
pension, with the arm abducted to a vertical position
above the body. Forelimb-suspend/pronograde compres-
sion is a similar position, but with the majority of body
mass borne by the suspended forelimbs, and is thus within
the pronograde suspend mode. Orangutans also ex-
hibited a novel form of squat whereby the majority of
weight was supported by the feet with the hips and knees
strongly flexed, but substantial weight was also supported
by squatting with the distended tummy propped up on
another support, parallel to the main weight-bearing sub-
strate (stomach squat). Finally, pronograde bridging
behavior was previously documented in orangutans. How-
ever, on rare occasions in this study, the only female to have
a dependant infant was also observed to use an orthograde
bridge, in which she held supports on either side of a gap in
the canopy together to reduce the size of the gap, so that
her accompanying infant could cross independently.
Locomotion
In general, orangutan locomotion is described well by
the standardized positional modes of Hunt et al. (1996),
although we separated vertical descent from vertical
climb at the mode rather than the submode level, as the
effect of gravity results in very different demands on the
body during climbing and descending, and orangutans
exhibit demonstrably different mechanics in the two
modes (e.g., rump-first and head-first cascade descent,
and fire pole slide) (Appendix B). Two further additions
at the mode level were hindlimb swing and forelimb-
hindlimb swing. The latter mode describes suspensory
behaviors that are neither exclusively orthograde nor
pronograde, and which utilize forelimbs and hindlimbs
in any sequence (e.g., ipsilateral swing and cartwheel
swing). These submodes are mostly used as an interme-
diary form of movement to reorient the body between
two longer bouts of very different types of locomotion.
Cartwheel swing and cartwheel descent (under the verti-
cal descent mode), although infrequent (<1%), are both
submodes novel to orangutans. They are bouts of suspen-
sory locomotion resembling the sequence of limb usage
seen in human cartwheels, either on horizontal or nega-
tively inclined supports. Both are exhibited predomi-
nantly by adolescent individuals during times when they
appear playful.
Tree sway is a mode exemplified by orangutans (Sug-
ardjito and van Hooff, 1986). The definition, however, is
unique in that it refers to a technique that may encom-
pass a broad range of other positional behaviors. In this
study, tree sway was observed to utilize many different
postures, which alter as the center of gravity of the body
changes position in relation to the support, during oscil-
lations. Monopedal,bipedal, and tripedal stand accompa-
nied by forelimb suspend are common components seen
during tree sway, as are ipsilateral and quadrumanous
suspend, and sit/forelimb suspend. The animal may
sway the tree in such a way that it is on top of the sup-
port when it moves to the next tree, or in suspension
underneath. In this study, orangutans were also observed
to oscillate vertical lianas and vertical branches in a
movement reminiscent of a child swinging on the end of a
rope, in order to swing to a new support. We therefore
renamed the mode sway, as it is not reliant on trees.
DISCUSSION
Comparisons with other studies
Cant (1987a) and Sugardjito and van Hooff (1986) pre-
viously quantified orangutan positional behavior, but
there are notable differences between this study and pre-
vious results. Table 3 shows frequencies for the most
TABLE 3. Percentages of commonly observed (>1% overall total) postural modes according to behavioral
context in comparison to previous studies
Present study Cant (1987a),
feed
Sugardjito and
van Hooff
(1986)
1
Rest Feed Travel Social Look Other Total Travel Rest
Sit 42 63 59 53 45 60 58 49 41 59
Pronograde stand 0 4 14 4 19 11 4<1263
Orthograde stand 0 7 6 4 13 7 62
Orthograde forelimb
suspend
04 5 4 4 2 3One arm: <1
Orthograde quadrumanous
suspend
03 2 1 1 5 2Other: 11 33 4
Forelimb/hindlimb
suspend
012 6 4 11 5 9Ipsilateral: 30
Pronograde suspend 0 5 4 5 3 7 4Other: 6
Lie 58 1 1 16 4 1 12 0034
1
Mean values calculated from Tables 3 (pauses during travel) and 4 (rest) in Sugardjito and van Hoof (1986).
4S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

commonly occurring postural modes in the most fre-
quently occurring behavioral contexts, and in compari-
son to the feeding data of Cant (1987a) and the travel
and rest data of Sugardjito and van Hooff (1986). Sit
dominates the postural repertoire of the current subjects
in all behaviors except rest, which is dominated by lie.
Both pronograde and orthograde stand are most fre-
quently observed during the behavioral context ‘‘look,’’ in
which, in response to unfamiliar noises or movements,
orangutans will remain stationary and stare in the direc-
tion of the unfamiliar sound, often for some time. Pro-
nograde stand is also more associated with traveling
than with feeding, which suggests that the large dispar-
ity between the results of Cant (1987a) and Sugardjito
and van Hooff (1986) (Table 3) is a result of differences
in behavioral context, although the difference in the pre-
sent study is rather less than between the other two.
Taken as a whole, suspensory modes are more associated
with feeding than with other behaviors. However, overall
frequencies for suspensory postures in this study are ap-
proximately half those of previous studies. Cant (1987a)
found that suspensory postures accounted for 47% of
positional behavior during feeding, in comparison to 24%
here. Mean values calculated from the seven subjects of
Sugardjito and van Hooff (1986) for pauses during travel
were lower at 33%, but are still twice the 17% frequency
recorded in the current study. Cant (1987a) further found
that ipsilateral suspend accounted for 30% of observed
posture. Here, ipsilateral suspend was the most frequently
occurring suspensory submode, but accounted for only 7.3%
of total bouts, followed by pronograde quadrumanous sus-
pend at 3.4% (Appendix A). Most forms of suspend were
likely to be combined with one or two other submodes.
Quadrupedal locomotion shows a stronger association
with feeding than with traveling, whereas torso-ortho-
grade suspensory locomotion shows the reverse pattern
(Table 4). While frequencies for quadrupedal and tripe-
dal walk are similar to previous studies for travel, values
for feeding are twice those observed by Cant (1987a). This
may be partly related to classification, since Cant (1987a)
did not record any pronograde scramble, which in this study
accounted for 9% of overall locomotion. Torso-pronograde
suspensory locomotion is also higher in this study than
in Cant (1987a), which may suggest that Ketambe orangu-
tans simply exhibited greater levels of pronograde behavior
than their Bornean counterparts. This result might also
reflect differences in support availability (e.g., density, pro-
portion of pioneer species) between the two sites, related
perhaps to regrowth following tree-fall or environmental
degradation. Unfortunately, we lack the detailed compara-
tive and three-dimensional data on support availability
which would be required to test this proposition.
In this study, bipedal walk accounted for 7.3% of overall
locomotion, 75% of which (5.5% of total locomotion) was
TABLE 4. Percentages of commonly observed locomotor modes according to behavioral context and
in comparison to previous studies
1
Mode Submode
Present study Cant (1987a) Sugardjito and
van Hooff (1986),
travelTravel Feed Total Travel Feed
Quadrupedal and tripedal walk 15 24 18 12 12 12
Walk 7 11 8 12 12 ?
3
Pronograde scramble 8 13 9 ?? ?
Torso-orthograde suspensory
locomotion
39 26 35 62 53 ?
Brachiation and forelimb swing 16 12 15 11 9 19
Orthograde clamber and transfer 23 14 20 51 44
Torso-pronograde suspensory
locomotion
35412
Forelimb/hindlimb- swing 0 1 0 ? ? 46
2
Bipedal walk 7 8 7 #
4
#
Bipedal walk 2 2 2 00
Assisted bipedal walk 5 6 6 ##
Bridge 4 0 3 ? ?
Vertical climb 14 21 16 12 19
Vertical climb 12 18 14 12 19
Angled climb 2
5
3
5
2
5
––
Vertical descent 8 12 9 6 14 11
Vertical descent 6 10 7 6 14
Angled descent 2
5
2
5
2
5
––
Drop 2 2 2 ? ? ?
Sway 7 1 6 7 0 12
Ride 1 0 1 ? ? ?
1
Overall frequencies are shown for locomotor modes (in bold) and frequencies for submodes (in italics) to allow comparison where
classification system of previous authors is different from current study.
2
Quadrumanous suspend of Sugardjito and van Hooff’s (1986) appears to include all these modes/submodes.
3
?, not clear if mode was not observed, or if it was observed but combined with another mode.
4
#, Cant (1987a, p. 74) notes that assisted bipedalism was observed but was presumably recorded as orthograde clamber, and
frequencies are not provided.
5
Cant (1987a) defined vertical climbing as locomotion within 22.58of true vertical. For present study, overall frequencies for
vertical climb and descent are for within 458of true vertical, but values for true vertical (0–208) and angled climb (20–458) are
provided separately. Sugardjito and van Hooff (1986) did not provide angular distinction between vertical climb and quadrumanous
suspend.
5ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

hand-assisted. Extended hindlimb positions were exhibited
much more frequently than flexed, and bipedal scramble
was also observed on small, irregularly placed and vari-
ously angled supports. In contrast, neither of the previous
studies specified any bipedal locomotion. This is probably a
matter of classification, as this study employed many of the
standardized modes proposed by Hunt et al. (1996), includ-
ing a distinction between hand-assisted bipedality and
hindlimb-assisted orthograde clamber. Cant (1987a) noted
that bipedal behavior was observed, but was not recorded
as such because it was hand-assisted, and it was therefore
presumably classed as orthograde clamber. Similarly,
hand-assisted bipedality may have been incorporated into
the quadrumanous climb mode of Sugardjito (1982).
While orangutan locomotion is dominated by torso-
orthograde suspensory locomotion, it is exhibited
far less in the current study than in Cant (1987a). This
discrepancy is largely related to a reduction in ortho-
grade clamber and transfer: even when other submodes
or types that may have formed the orthograde clamber
of Cant (1987a) are included (assisted bipedalism and
angled climb/descent), frequencies here are still only 32%
and 25%, compared to 51% and 44% for traveling and
feeding, respectively, in Cant (1987a).
Traditionally, orangutans are renowned as cautious,
slow climbers. However, they clearly have the potential
for fast and acrobatic locomotion. Leap,inverted quad-
rupedal run,drop submodes, and rump-first and head-
first cascade descents all imply increased speed at the
expense of safety. Some of these behaviors appear to be
associated with fleeing aggressive situations (leap and
inverted quadrupedal run were both exhibited only once,
and in the context of escape from an aggressive adult
during competition for a favored food source). Rijksen
(1978) also observed leaping in fleeing subjects, and
MacKinnon (1971, 1974) described tumble descents (the
equivalent of cascade descents in this study) in the con-
text of escaping aggressive individuals. On the other
hand, other fast submodes appear to be part of normal
travel and feeding behavior: cartwheel swing,cartwheel
descent, and lunging and descending bridge appear more
dramatic than their travel and feeding contexts would
necessitate. Notably, most of these bouts were exhibited
by adolescents. Figures 1 and 2 show little difference be-
tween the positional mode profiles of the different age-
sex categories in this study, and Thorpe and Crompton
(2005) showed that the age-sex category of an individual
has less influence than habitat variables on the observed
locomotor profile, when analyzed in a multivariate anal-
ysis at the mode-level. However, the comparison by Payne
et al. (in press, a and b) of hindlimb muscle dimensions in
all the apes found that juvenile orangutans had the larg-
est proportion of total hindlimb muscle volume in the
pedal digital flexors, both in comparison to their adult
male orangutan and the other apes they studied. They
suggested that the difference between the adult and ju-
venile orangutans may be an artifact of increased levels
of terrestriality in captive adult orangutans. However,
the locomotor behavior profiles seen here indicate that a
secure pedal grip may be more important in wild juve-
niles than adults, as they exhibit fast and risky locomo-
tion more often than do more mature animals. This sug-
gests that orangutans become more cautious in their
locomotion as they age and/or increase body mass. While
positional modes are relatively comparable between age-
sex categories (Figs. 1, 2), it is possible that the more
subtle nuances of locomotion, described by the submodes
within each mode, may be found to differ with age. Unfortu-
nately, ape locomotion is a relatively small proportion of
their daily positional mode repertoire, and it is difficult
in the field to obtain enough locomotor observations to
quantify such subtlety.
Overall, this study shows greater levels of pronograde
and orthograde compressive locomotion and reduced levels
of orthograde suspensory locomotion than in other stud-
ies. Furthermore, all suspensory postures were exhibited
less frequently in this study. While differences in meth-
odology may certainly account for some differences be-
Fig. 1. Postural mode percentages according to age-sex cate-
gory. S, sit; SQ, squat; PST, pronograde stand; OS, ortho-
grade stand; OFLS, orthograde forelimb-suspend; OQS,
orthograde quadrumanous suspend; FHS, forelimb-hin-
dlimb suspend; PSU, pronograde suspend; HS, hindlimb
suspend;L,lie; PB, postural bridge. Definitions of age-sex
categories follow Thorpe and Crompton (2005), who found that
relationship between locomotion and related ecological variables
for this study group was best understood when subadult and
adult males were combined, with adult females and adolescents
as separate categories.
Fig. 2. Locomotor mode percentages according to age-sex
category. Q, quadrupedal walk;B,bipedal walk; VC, verti-
cal climb; VD, vertical descent; TOS, torso-orthograde
suspensory locomotion; TPS, torso-pronograde suspen-
sory locomotion; FHSW, forelimb-hindlimb swing; HSW,
hindlimb swing; BR, bridge;L,leap;D,drop; SWY, sway;
R, ride.
6S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

tween studies (in particular, that we set out to conduct a
detailed analysis, whereas previous authors aimed to
characterize broad-based trends), the results still indi-
cate large variability across studies. The study by Cant
(1987a) was conducted during an ‘‘unusual drought,’’ and
it is possible that orangutans were forced to access food
resources in less accessible places, such as the fragile
and unstable terminal ends of branches, where higher
levels of suspensory behavior would be expected. One
might then expect that the frequencies for vertical climb-
ing and descent would be greater in Cant (1987a),
whereas in fact the results are very similar (Table 4).
Unfortunately, Sugardjito and van Hooff (1986) did not
present data for food types, but Table 5 shows that com-
paratively high levels of fruit and ants/termites were
consumed in the present study, but considerably more
bark was consumed in Cant (1987a). This suggests a rich
supply of favored food sources in our study vs. a reduced
supply during that of Cant (1987a) (conducted during a
drought). However, the implications of this for positional
mode frequencies are difficult to predict quantitatively.
On the one hand, one might expect that the subjects of
Cant (1987a) would need to suspend more to reach the
remaining fruit on terminal branches. On the other
hand, orangutans use a variety of positional modes to
strip bark from trees and lianas, but compression is a
necessary component of most of these, as it is a requisite of
the leverage necessary to detach bark. As a consequence of
this factor alone, ‘‘stand’’ modes would be expected to be
higher in Cant (1987a), rather than lower (Table 3). Fur-
thermore, once orangutans have obtained a strip of bark,
they generally tend to sit to eat it; this would be expected to
skew the results toward sit, as the eating of bark tends to
be the longest part of the process (personal observations).
Differences may also reflect different subject profiles.
The study of Cant (1987a) consisted of two adult females,
whereas that of Sugardjito and van Hooff (1986) and the
present study report results for both sexes and all age
groups. Indeed, Cant (1987b) demonstrated that both food
type and gender are important predictors of postural be-
havior during feeding, as he observed that suspensory
behavior for males and females in two types of fig tree
ranged from 31% for males in the one fig species to 70%
for females in the other. However, Figures 1 and 2 show
that although there are slight variations in positional
behavior exhibited by adult females for particular posi-
tional behaviors in this study, overall adult female posi-
tional behavior does not differ substantially from the
other age-sex categories. Thorpe and Crompton (2005) also
demonstrated that when associations between all behav-
ioral and ecological variables are taken into account, the
age-sex category of an individual has limited influence on
its observed locomotor repertoire, although results for pos-
ture are not yet published.
The log-linear analysis by Thorpe and Crompton (2005)
concluded that support type and diameter had the strong-
est influence on observed locomotor repertoire of all the
variables they modeled. Height in the canopy and con-
textual behavior were not found to directly influence
locomotion; instead, their effect was modified by support
type and diameter, respectively. Therefore, it is possible
that the differences between studies simply reflect differ-
ent patterns of support use and differences in the overall
interplay between orangutan positional behavior and the
variables that were shown to influence it (Thorpe and
Crompton, 2005). The relationship between positional
behavior and support use is beyond the scope of this paper,
but will be addressed elsewhere. However, the results also
suggest that a key mechanism by which orangutans re-
solve complex habitat problems is substantially decreased
stereotypy.
Implications for ape specializations
Tables 6 and 7 present postural and locomotor frequen-
cies from the present study in comparison to previously
published data for other hominoids, cercopithecoids, and
atelines, updating the comparison by Hunt (1991a) with
new studies on lowland gorillas and mountain gorillas, and
expanded data sets on bonobos and orangutans. Carlson
(2005) also presented an updated overview of African ape
locomotion from studies by Doran (1989); Doran and Hunt
(1994); Hunt (1989, 1992); and Remis (1994, 1995, 1998).
While differences in some positional behavior classifica-
tions preclude his entire data set being used here, we
included some values in Table 7 to distinguish between
vertical climb and torso-orthograde suspensory loco-
motion, which were not separated in Doran (1996). Two
outgroup comparators, Papio and the atelines, were also
included. The inclusion of Papio allows us to reassess the
analysis by Hunt (1991a) of the distinctiveness of hominoid
positional modes in light of our expanded data set. Atelines
are included, following the argument by Larson (1998) that
postcranial features found in both hominoids and Ateles are
likely to be features that are highly responsive to function,
and consequently likely to be homoplastic characters for
hominoids. Unfortunately, only locomotor data are avail-
able for the atelines. Frequencies in Tables 6 and 7 are for
arboreal behavior only. Differences in methodology mean
that only a broad comparison may be made. Nevertheless,
Table 6 shows that orangutan postural behavior is more
similar to that of the African apes than was previously
thought. Hunt (1991a) found that the frequency of arm-
hanging was highest in orangutans of all the great apes. In
this study it is the presence of orthograde quadruma-
nous suspend,forelimb-hindlimb suspend,andpro-
nograde suspend that distinguishes the orangutans from
the other apes and baboons, but the difference is less
pronounced than implied by previous orangutan studies
(Tables 3 and 6). Otherwise, orangutans sit, lie, stand (both
quadrupedally and bipedally), and forelimb-suspend in very
similar frequencies to those exhibited during arboreal
behavior by lowland gorillas. Overall, the postural results
support the conclusion of Hunt (1991a) that with the ex-
ception of gorillas, orthograde suspensions (forelimb and
quadrumanous) are the postural behaviors of adaptive sig-
nificance that are exhibited by the apes in greater frequen-
cies than in Papio. However, orthograde stand may also
prove to be distinctive to apes, if future studies expand
data sets on species that currently have small sample sizes,
or undetermined frequencies of bipedal posture in Table 6.
Furthermore, since bonobos, mountain gorillas, and gibbons
all exhibit bipedal walking, it is probable that orthograde
stand will feature in their postural repertoire.
TABLE 5. Comparison of percentages of food types
consumed in Cant (1987a) and present study
Food type Present study Cant (1987a)
Fruit 64 46
Flowers 1 5
Leaves 14 16
Bark 10 31
Ants/termites 8 2
Other 2 0
4,717 bouts 66 hr
7ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

TABLE 6. Comparison of percentages of arboreal postures in hominoids and Papio
Sit Squat
Pronograde
stand
Orthograde
stand
Orthograde
forelimb
suspend
Orthograde
quadrumanous
suspend
Forelimb/
hindlimb
suspend
Pronograde
suspend Lie Other N
Pongo pygmaeus
1
49 0 <12.0<11136?
11
0<1 1,682 minutes
Pongo pygmaeus
2
58 0 4 6.0 3 2 9 4 12 1 7,155
Pongo pygmaeus
3
61 0 5 7.0 3 2 8 3 9 1 4,560
Pan troglodytes
schweinfurthil
4
81 2 1 0.3 12 0 0 0 3 0 2,518
Pan paniscus
5
90 0 2 0.0 5 0 0 0 3 0 Small
Gorilla gorilla
gorilla
6
50 29 4 5.0 1 0 0 0 10 0 473
Gorilla gorilla
berengei
7
p ? p ? p ? ? ? p ? Small
Gibbon
8
64 0 0 0.0 36 0 0 0 0 0 Small
Hylobates
syndactylus
9
47 0 0 0.0 53 0 0 0 0 0 Various
studies
Papio anubis
10
76 1 21 0.0 1 0 0 0 0 1 580
1
Cant (1987a); Table 8. Bornean adult females only. Frequencies represent proportions of time spent feeding.
2
Present study, all data. Other: hindlimb suspend, cling, splits, cantilever, postural bridge.
3
Present study, juveniles omitted. Other: hindlimb suspend, cling, splits, cantilever, postural bridge.
4
Hunt (1991b); Table 1. Percentage of 2-min instantaneous focal observations for Mahale and Gombe adults while feeding on fruit. Standardized by Hunt (1991b) for time of day.
5
Taken from Hunt (2004), Table 10.1. Percentage of 132 instantaneous time point surveys during arboreal feeding on fruit (Kano and Mulavwa, 1984).
6
Remis (1995); Table 9.
7
Doran (1996), 1-min instantaneous sampling. Modes are noted as P for present, since Doran (1996) did not separate arboreal and terrestrial data.
8
Taken from Hunt (2004), Table 10.1. Average of H. agilis and H. pileatus from Gittins (1983), 322 bouts sampled by 10-min scan surveys, and Srikosamatara (1984), 655 5-min
scan surveys.
9
Taken from Hunt (2004), Table 10.1. Average from Chivers (1972) (percentage of 234 5-sec instantaneous focal surveys) and Fleagle (1976) (percentage of 1,376 postural bouts
during feeding).
10
Hunt (1991b), Table 1. Instantaneous focal observations during feeding only.
11
?not clear if this mode was not observed or if it was observed but combined with another mode.

TABLE 7. Comparison of percentages of arboreal locomotion in hominoids, Papio, and atelines
Quadrupedal
and tripedal
walk
Vertical
climb and
descent
Bipedal
walk
Torso-orthograde suspensory
locomotion
Drop and
leap
Torso-pronograde
suspensory
locomotion Sway Ride Bridge
No.
bouts
Orthograde
clamber and
transfer
Brachiation
and forelimb
swing
Pongo pygmaeus
1
12 21 ? 49 11 0 1 6 ?
14
? 4,360
Pongo pygmaeus
2
18 25 7 21 14 2 4 6 1 3 1,504
Pongo pygmaeus
3
18 26 7 22 13 1 3 7 1 2 2,811
Pan troglodytes
schweinfurthii
4
36 49 7 5 5 0 ? ? ? ? 223
Pan troglodytes verus
5
22 68 3 ?
14
7 1 ? 0 ? 0 1,417
Female, 51/male, 59 Female, 8/male, 8
Pan paniscus
6
32 53 1 ? 9 4 ? 0 ? 0 1,461
Female, 24/male, 31 Female, 25/male, 26
Gorilla gorilla gorilla
7
19 48 5 17 3 ? 0 8 ? ? 122
Gorilla gorilla
beringei
8
53 40 2 ? 5 0 ? 0 ? 0 153
Female, 33/male, 27 Female, 3/male, 9
Gibbon
9
1 16 2 0 67 14 0 0 0 0 Small
Hylobates
syndactylus
10
0 32 8 0 59 2 0 0 0 0 Small
Lagothrix lagothricha
11
29 14 0 30 9 4 2 1 0 10 3,926
Ateles belzebuth
12
21 13 1 28 22 2 0 2 0 11 3,760
Papio anubis
13
68 21 0 0 0 10 0 0 0 0 26
1
Cant (1987a), calculated from Table 3. Locomotor bout sampling with distance for Bornean adult females only. Frequencies represent proportions of travel distance. Vertical
climb/descent values are underestimated, as Cant (1987a) defined vertical climb as climbing within 22.58of true vertical. Other studies cited here defined it within 458of true
vertical.
2
Present study, all data.
3
Present study, juveniles omitted.
4
Hunt (1991b), Table 3. Percentage of arboreal locomotion in all contexts for Gombe and Mahale subjects. Not standardized.
5
Doran (1996), calculated from Table 16.5. Locomotor bout sampling. Vertical climb/descent values include orthograde clamber/transfer and pronograde clamber. Individual values
for males and females are from Carlson (2005). See text for discussion.
6
Doran (1996), calculated from Table 16.5, 1-min instantaneous sampling. Vertical climb/descent values include orthograde clamber/transfer and pronograde clamber. Individual
values for males and females are from Carlson (2005). See text for discussion.
7
Remis (1995), Table 11. Wet season data only. Adults only. Data for orthograde clamber may overlap with bipedalism, as Remis (1995) noted that it was difficult to discriminate
relative proportions of weight borne by forelimbs and hindlimbs. Drop is included in orthograde suspend. Oscillation equates to ‘‘acrobatic behaviors’’ of Remis (1995), including
leap, tree sway, fireslide (equivalent to rump descent-firepole in this study), and bridging.
8
Doran (1996), calculated from Table 16.5. Instantaneous sampling. Vertical climb/descent values include orthograde clamber/transfer and pronograde clamber. Individual values
for males and females are from Carlson (2005). See text for discussion.
9
Taken from Hunt (2004), Table 10.3. Average of H. agilis,lar,andpileatus from Gittins (1983) (percent of 255 10-min scan surveys), Fleagle (1980) (percent of 211 pooled feeding
and travel bouts obtained with continuous focal sampling), and Srikosamatara (1984) (percent of 218 5-min scan surveys).
10
Taken from Hunt (2004), Table 10.3. Average from Fleagle (1980) (percent of 1,206 pooled feeding and travel bouts obtained with continuous focal sampling) and Gittins (1983)
(percent of 208 10-min scan surveys).
11
Cant et al. (2001, 2003). Adults only. Instantaneous focal animal sampling on 5-min mark. Vertical climb/descent values are slightly underestimated, as Cant et al. (2001, 2003)
defined vertical climb as climbing within 22.58of true vertical.
12
Cant et al. (2001, 2003). Instantaneous focal animal sampling on 5-min mark. Adults only. Vertical climb/descent values are slightly underestimated, as Cant et al. (2001, 2003)
defined vertical climb as climbing within 22.5of true vertical.
13
Hunt (1991b), Table 3. Percent of arboreal locomotion in all contexts. Adults only. Not standardized.
14
?, not clear if this mode was not observed, or if it was observed, but combined with another mode.

While vertical climbing accounts for a significant pro-
portion of ape arboreal locomotion, and must be energeti-
cally highly demanding since it is performed more or less
directly against the gravity vector, it is not unique to
apes (Table 7). Indeed, it is the single locomotor mode
common to all species compared in this paper. The Afri-
can apes generally show increased frequencies of vertical
climbing and descent in comparison to orangutans,
which is to be expected, given that they primarily travel
on the ground rather than through the canopy, and climb to
access arboreal food sources. In contrast, orangutans only
climb to access food on different levels in the canopy or to
access preferred substrates for locomotion (Thorpe and
Crompton, 2005).
Most locomotor frequencies are similar in orangutans
and the African apes, with a particular similarity in
many modes found between frequencies in orangutans
and lowland gorillas. Differences between these species
in ride and bridging behaviors are likely to be related to
methodology. Thus, torso-pronograde suspensory loco-
motion is probably the only mode that distinguishes
orangutans from the other apes studied. If so, the pres-
ence of pronograde postural and locomotory suspension
in Pongo but not the African apes is particularly inter-
esting, given that African ape positional behavior is
dominated by compressive pronogrady. Might this sug-
gest that adaptations for pronogrady evolved in parallel
in orangutans and the African apes, with orangutans de-
veloping both suspensory and compressive pronogrady,
and the African apes, due to their largely terrestrial na-
ture, only developing compressive pronogrady? This may
be viewed as a somewhat heretical concept, since many
still regard Proconsul as a suitable model for the locomo-
tor morphology of the ancestor of the living great apes
(e.g., Larson, 1998). However, in view of more recent
reviews by Harrison (2002) and Begun (2002), it may
nevertheless be a concept that merits some attention. Clari-
fication of whether the quadrumanous climbing and scram-
bling classification by Doran (1992a,b, 1993, 1996) for
chimps, bonobos, and mountain gorillas includes suspen-
sory pronogrady would shed light on this issue.
Similarity in the frequencies of orthograde compres-
sive postures and locomotion between lowland gorillas
and orangutans is paralleled by similarities in the hind-
limb kinematics of this posture (Hofstetter and Niemitz,
1997), which may be due to the energetic cost of main-
taining an upright posture in animals with a top-heavy
build (Payne et al., in press, a and b). However, similar-
ities in frequencies and kinematics do not seem to be re-
flected in clear similarities in hindlimb muscle dimen-
sions and maximum moments of force (Payne et al., in
press, a and b), implying that orthograde positional behav-
ior in adult male orangutans and gorillas is not necessar-
ily dynamically similar (Alexander and Jayes, 1983),
although it may be functionally similar in terms of kine-
matics and frequencies. Specifically, we would not expect
the hindlimb to be used so much as a rigid strut in goril-
las, as it undoubtedly is in orangutans (Buettner-Janusch,
1967; Hamanda, 1985; Tuttle and Cortwright, 1988; Payne,
2001; Payne et al., in press, a and b).
Table 7 shows that brachiation and forelimb swing
and bipedal walk are present in all apes, but not in
Papio.Orthograde clamber and transfer are present in
all great apes, and are similarly absent in Papio. While
Table 7 indicates that these submodes are also absent in
the hylobatids, the samples for these data are very
small. Fleagle (1976) noted that climbing in siamangs is
forelimb-dominated, and his Figure 2 seems to depict
orthograde clamber. Recent work by Collis (2001) also in-
dicates that siamangs exhibit orthograde clamber, although
the results have not yet been fully published. Gibbons may
also be found to exhibit orthograde clamber,althoughCart-
mill (1985) argued that their small body size and highly
derived brachiating specializations are likely to be derived
after the split from the common ancestor of the great apes,
and Young (2003) found that gibbon specializations skewed
the analysis of hominoid postcranial uniqueness and varia-
bility. Therefore, gibbons probably represent a poor model
for the common ape postcranial adaptation. Of these sub-
modes, orthograde clamber and transfer and brachiation
are also present in the atelines, which, according to the
argument of Larson (1998), would suggest that these
behaviors may have been acquired in parallel.
Overall, this study suggests that it is in fact orthograde
positional behavior in general, rather than forelimb sus-
pend specifically, that distinguishes apes from Old World
monkeys, and that may be regarded as characterizing the
positional behavior of the living hominoids. This idea is
supported by increasing evidence that orthogrady also
characterized the morphology of a number of Miocene
apes, including Morotopithecus, the proposed sister taxon
of the crown hominoid clade, dating from at least 20 mya
(MacLatchy et al., 2000). While Morotopithecus’s denti-
tion resembles that of Proconsul and Afropithecus, its
postcranium is distinct from the other Early and Middle
Miocene forms in its orthograde adaptations (MacLatchy,
2004). These include features of the vertebrae associated
with limited lower back morphology (MacLatchy, 2004);
features of the glenoid fossa indicative of loading over a
wide range of movements (MacLatchy, 2004); and high
cortical area relative to periosteal area of the femur, in-
dicating high axial rigidity (Ruff, 1989; Ruff and Runestad,
1992; Runestad, 1994). MacLatchy et al. (2000) concluded
that Morotopithecus possessed orthogrady and substan-
tial forelimb mobility, and a locomotor repertoire that
probably included forelimb suspension, slow brachiation,
cautious clambering, vertical climbing, and quadrupedal-
ism (MacLatchy, 2004; Young and MacLatchy, 2004). The
later Miocene taxa Oreopithecus and Dryopithecus, be-
lieved to belong to the crown hominoid clade (MacLatchy
et al., 2000), also possessed adaptations for suspensory,
orthograde behavior similar to those in modern homi-
noids (Harrison, 1987; Moya-Sola and Kohler, 1996; Har-
rison and Rook, 1997). Furthermore, Senut (2003) argued
that the Late Miocene Orrorin was adapted to bipedal-
ism, but was also a climber. For example, the anatomy of
the proximal femur in Orrorin (some 6 mya; Senut et al.,
2001) strongly suggests that the hip joint was already
being used habitually in extension (which implies at
least an orangutan-like degree of orthogrady). At about
twice this age, the truncal anatomy of Pierolapithecus
(Moya-Sola et al., 2004) was described as adapted for
upright posture, as of course was that of Oreopithecus,
where the structure of the inner ear is also consistent
with upright posture (Rook et al., 2004). Thus, the
weight of evidence, in our view, favors the temporal
primacy of orthogrady in the hominoids, rather than
suggesting a compressive pronograde locomotor repertoire
in the common ancestor of the hominoids.
Although positional-mode frequencies and distinctive-
ness are important in assessing the evolutionary origin
of ape anatomy, it is rather the muscular effort required
for a positional mode and the associated stresses that are
likely to determine the anatomical adaptations required for
10 S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

the effective performance of a given positional mode (Hunt,
1992). Some positional behaviors are, of course, inherently
more stress-inducing or energetically expensive than others.
But, all other things being equal, positional behaviors for
which an animal is well-adapted are expected to require
less muscle activity, and induce less stress in the skeleton
and ligaments, than behaviors for which the animal is
poorly adapted. It is argued that natural selection shapes
anatomy to reduce muscular activity and structural stress
in proportion to the frequency of the behavior (Basmajian,
1965; Cartmill et al., 1987; Hunt, 1991b, 1992). However,
Pontzer and Wrangham (2004) demonstrated that the rela-
tionship between muscle forces and stresses and frequen-
cies of positional modes is rather more complicated than
this suggests. They noted current evidence indicating that
chimpanzee terrestrial travel is energetically relatively
costly (Taylor et al., 1972; Heglund and Taylor, 1988). This
led them to test the hypothesis that arboreal adaptations
for climbing in common chimpanzees minimize total loco-
motor energy expenditure (climbing and walking) by lower-
ing the amount of energy spent climbing, a potentially
costly activity. They found, however, that morphological
specializations for arboreal locomotion incurred a signifi-
cant energy cost, and they concluded that nonenergetic
factors are more likely to be critical in shaping chimpanzee
locomotor anatomy. Cartmill and Milton (1977) pointed
out that a large animal is less likely to survive if it
falls from any great height, because the kinetic energy
which the tissues of its body must dissipate on impact
increases in proportion to the cube of its linear dimensions
and to the square of its terminal velocity (which is in
turn increased by the low surface-to-mass ratio). Accord-
ingly, Pontzer and Wrangham (2004) concluded that
chimpanzee postcranial adaptations should be seen as
adaptations to avoid falls, rather than to reduce the energy
cost of climbing. Given their higher degree of arboreality,
we would expect, therefore, that orangutan anatomical
adaptations should be particularly directed at avoidance
of falls.
Another measure of the stresses associated with posi-
tional behaviors is the dimensions of the bones and
muscles of the locomotor system, which Alexander (1974)
proposed should be understood largely in relation to the
most strenuous activities in which they are used. Climb-
ing and arboreal locomotion undoubtedly form the most
strenuous locomotor activities of the apes, due to the
effects of gravity and the discontinuous and three-dimen-
sional nature of the substrates. Thorpe et al. (1999)
showed that whereas human hindlimb muscles appear to
be designed to generate large forces over a narrow range
of joint positions (as required for bipedal running), chim-
panzee hindlimb muscles were designed to generate
smaller forces, but over a wider range of joint positions.
Payne (2001) and Payne et al. (in press, a and b) found
that these observations held true for all extant nonhuman
apes, and concluded that nonhuman apes have a greater
need than terrestrial mammals to vary muscle function
throughout the joint range of motion, due to the complex-
ities of their arboreal habitat. Orthograde postures involve
the exertion of significant muscle force in more extended
joint positions than either quadrupedalism or vertical
climbing (Crompton et al., 2003), supporting the theory
that orthograde behavior may be important in shaping the
musculoskeletal morphology of all apes. Indeed, it is by
some degrees that the muscular anatomy of humans is the
most distinctive of all the hominoids (Thorpe et al., 1999;
Payne, 2001; Payne et al., in press, a and b).
The ecological context of behavior is also an important
consideration when trying to identify the potential im-
portance of its influence on musculoskeletal design. The
benefits of forelimb suspension in hominoids are already
well-documented: suspension in general increases safety
for animals of large body mass (Cartmill, 1974, 1985); it
allows them to utilize small-diameter supports for feed-
ing and traveling (Cant, 1987a,b; Hunt, 1992); and it
extends their foraging radius (Grand, 1972), particularly
in the terminal branches where high-quality fruits abound
(Avis, 1962). However, the benefits provided by forelimb
suspension for each of these factors are surely magnified
by the use of both fore- and hindlimbs in supporting body
mass.
When arboreal animals stand bipedally or quadrupe-
dally, even on a horizontal branch, they tend to topple to
one side or the other because all the support points are
effectively colinear (Napier, 1967; Cartmill, 1985). Safety is
therefore likely to be increased by suspension, and animals
of large body mass were predicted to suspend more than
smaller ones (Cant, 1987b). While this prediction is sup-
ported in hylobatids (Hunt, 1991a), it is not fully borne out
by intraspecific studies of orangutans and lowland gorillas
(Cant, 1987b; Remis, 1995; Thorpe and Crompton, 2005) or
interspecific studies of gibbons and chimpanzees (Hunt,
1991a). Cartmill (1985) proposed that methods of resisting
the toppling tendency were: 1) having relatively short
limbs, thus keeping the body’s center of mass close to the
support; 2) having prehensile hands and feet that can grip
a branch and exert a torque that resists the toppling
moment; 3) suspension; or 4) reduction of body size. Clearly
the benefit of hindlimb-assisted suspension, as well as fore-
limb-assisted compression, is that it allows a combination
of options 1, 2, and 3, thus maximizing safety in a complex
and uncertain habitat.
Hunt (1992) identified the small size of available
weight-bearing supports in the most productive areas of
fruiting trees as the most significant evolutionary pres-
sure selecting for suspensory posture and locomotion.
Assisted and unassisted orthograde forelimb suspension
were seen in the context of feeding more than any other
posture, leading to the conclusion that arm-hang was the
key specialization that separated chimps from baboons.
We argue that in an arboreal context, this assessment is
based on an overly static analysis of positional behavior.
Few positional modes, other than a single leap, have a
clear beginning or end. Nearly all locomotor and postural
modes are abstracted by the observer from a continuum
of activity: such abstraction is a prerequisite of objective
analysis, and the observer’s judgment has to balance
conflation with oversplitting. Figure 3 illustrates this
point by showing five locomotor modes that might be iso-
lated from a single 18-sec video clip of completely unin-
terrupted orangutan locomotion.
In the terminal branches of trees, there is almost always
a multitude of additional supports available to support
other limbs in tension or compression, serving the function
of distributing body mass more widely and thus reducing
stress on a single limb. Indeed, for feeding to take place,
unimanual suspension will nearly always (even temporar-
ily) become bimanual suspension. To select another fruit, a
third or fourth limb will often make contact to bring about
a shift in body position. Thus, orthograde suspension with
formation of a foot contact can become orthograde suspen-
sion assisted by hindlimb compression, and this in turn can
become orthograde stand assisted by forelimb suspension,
simply by contact of an extremity and slight alterations in
11ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

the position of the body’s center of gravity. This, in turn,
may be driven by the movement of branches in response to
body weight, reaching, and environmental conditions.
A major problem with testing the hypothesis of Hunt
(1992) is that as a largely irregular rather than pat-
terned activity, the various submodes conflated under
torso-orthograde suspensory locomotion and bi-
pedal walk and similar postural modes are difficult to
analyze or describe in an experimental situation. Never-
theless, it is possible that we are mistaken in looking for
a single unique mode to account for the morphological
similarities of the apes. It is certainly plausible that
what underlies ape anatomy is the exact opposite: the
ability to respond to a complex array of ecological and
environmental conditions with one of a broad range (if
not a continuum) of orthograde-based postures and loco-
motion.
CONCLUSIONS
This paper shows that orangutan positional behavior
is highly complex, representing a continuum of posi-
tional forms rather than a collection of unique and
mutually exclusive positional modes, as may be seen in
arboreal animals of smaller body mass. Although orang-
utan positional behavior is broadly similar between age-
sex categories, there is some evidence to suggest that it
differs at the submode level, although a larger sample
size would be required to quantify this. Overall, this
study showed greater levels of pronograde and ortho-
grade compressive locomotion and reduced levels of
orthograde suspensory locomotion in comparison to other
studies. Suspensory postures were also exhibited less
frequently in this study than previously reported. Given
the complexity of orangutan positional behavior demon-
strated by this study, it is likely that differences in
positional behavior between studies reflect differences
in the interplay between the complex array of variables
which were shown to influence positional behavior
(Thorpe and Crompton, 2005).
Orangutans, alone among the apes, exhibit pronograde
suspensory positional behavior. Otherwise, most postural
and locomotor frequencies are similar in orangutans and
African apes, and in particular, lowland gorillas. This
study suggests that it is orthograde positional behavior in
general, rather than forelimb suspend specifically, which
characterizes the positional behavior of hominoids.
ACKNOWLEDGMENTS
We thank the Indonesian Institute of Sciences in
Jakarta, the Indonesian Nature Conservation Service in
Jakarta, Medan, and Kutacane, and the Leuser Develop-
ment Programme in Medan for granting permission and
giving support to conduct scientific research in the Ketambe
Research Station situated in the Leuser Ecosystem. We
thank the European Commission and the Indonesian Gov-
ernment for funding the Leuser Development Programme,
and our Indonesian counterpart Jito Sugardjito for his as-
sistance. We are grateful to two anonymous reviewers for
their comments on the manuscript.
Fig. 3. Five locomotor modes isolated from single 18-sec video clip of completely uninterrupted orangutan locomotion.
12 S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

APPENDIX A. Postural mode definitions
Postural mode % bouts
Postural submode (component 1/component 2/component 3)
Sit
Sit (P1a–P1f):
1
combines ‘‘sit,’’ ‘‘sit in,’’ ‘‘sit out,’’ ‘‘foot-prop sit,’’ ‘‘sit in/out,’’ ‘‘chair sit,’’ and ‘‘ischium sit,’’ as
these positions were difficult to distinguish when individuals were sitting on day nests/platforms.
50.50
Sit/forelimb-suspend (P1g). 5.91
Sit/hindlimb-suspend: more than half of body weight depends on one or both ischia, but one or both hindlimbs
grasp overhead substrate and support more than their own weight.
0.22
Sit/forelimb-hindlimb-suspend: as for ‘‘sit/hindlimb-suspend’’ but a forelimb and a hindlimb support more than
their own weight in either contralateral or ipsilateral suspensory combination.
0.17
Sit/forelimb-compression: as for ‘‘sit/forelimb-suspend,’’ but one or both forelimbs are below level of shoulders
and support body weight in compression.
0.21
Sit/hindlimb-compression: sit, with one or both hindlimbs supporting body weight in compression. 0.91
Sit/forelimb-hindlimb-compression: more than half of body weight is supported by ischia, but a forelimb and a
hindlimb support more than their own weight in contralateral or ipsilateral compression combination.
0.03
Sit/forelimb-suspend/hindlimb compression: more than half of body weight is supported by ischia, but weight
is also supported by a forelimb in suspension and a hindlimb in compression.
0.13
Squat
Squat (P2) 0.1
Squat/forelimb-suspend: bipedal or monopedal squat in which one or both forelimbs also support weight in
suspension.
0.25
Squat/forelimb-cling: as for ‘‘squat/forelimb-suspend,’’ but one or both forelimbs cling to a vertical or near
vertical substrate, supporting more than their own weight.
0.01
Stomach-squat: body weight is borne primarily by feet in a squat, but protruding stomach rests on additional
support and appears to support significant body mass.
0.06
Stomach-squat/forelimb-suspend: as above, but with some body weight borne by one or both forelimbs in
suspension.
0.03
Cling
Bimanual cling (P3a). 0.01
Cling/forelimb-suspend (P3c). 0.08
Cling/sit/forelimb-suspend: one hindlimb is in cling position and supports majority of body mass. This is aided
by a forelimb in suspension and one ischium.
0.04
Pronograde stand
Quadrupedal stand (P4a). 1.48
Tripedal stand (P4b). 1.26
Tripedal stand/forelimb-suspend: tripedal posture in which free forelimb is extended in arm-hanging fashion. 0.10
Tripedal stand/hindlimb-suspend: tripedal posture in which free hindlimb supports body weight in suspension. 0.01
Quadrupedal full-crouch (P4c1). 0.10
Quadrupedal forelimb-crouch (P4c2). 0.01
Quadrupedal hindlimb-crouch (P4c3). 0.01
Tripedal hindlimb-crouch: tripedal posture in which elbow is extended, but both hindlimbs are flexed. 0.07
Tripedal hindlimb-crouch/forelimb-suspend: as above, but with free forelimb supporting more than its own
weight in suspension.
0.01
Contralateral compression: standing with torso pronograde and face downward and weight supported by a
forelimb and a hindlimb on opposite sides of body.
0.01
Contralateral compression/hindlimb-suspend: as for ‘‘contralateral stand,’’ but with remaining hindlimb in
suspension.
0.03
Ipsilateral compression/forelimb-suspend: torso is pronograde and on its side. Majority of body weight is
supported in compression by a forelimb and hindlimb on same side of body. Significant body weight is also
supported by a forelimb under suspension.
0.01
Ipsilateral compression/hindlimb-suspend: as for ‘‘ipsilateral stand-forelimb suspension,’’ except that mass is
partially supported by a hindlimb in suspension rather than a forelimb.
0.01
Pronograde stand/forelimb-suspend: torso is in pronograde position, with one or both hindlimbs in quadrupedal
stand position, but one or both forelimbs are abducted and support body mass through suspension above head.
1.16
Orthograde stand
Extended bipedal stand (P5b). 1.58
Flexed bipedal stand (P5a). 0.18
Flexed bipedal stand/forelimb-suspend (P5c2). 0.24
Extended bipedal stand/forelimb-suspend (P5c1). 1.09
Extended bipedal stand/forelimb-compression: as for ‘‘extended bipedal stand,’’ but one or both forelimbs are
below level of shoulders and support body weight in compression.
0.07
Bipedal compression: bipedal posture where legs may be angled in any position below horizontal and may be
positioned on variously angled substrates, at different levels, and with different degrees of abduction/
adduction and flexion/extension to each other.
0.06
Bipedal compression/forelimb-suspend: as above, but with significant support from one forelimb in suspension. 0.03
Monopedal stand: body mass supported by standing on one leg, with insignificant contributions from other body
parts.
0.77
Monopedal stand/forelimb-suspend: as for ‘‘bipedal stand/forelimb-suspend,’’ but with only one hindlimb. 1.16
Monopedal stand/forelimb compression: as for ‘‘bipedal stand/forelimb-compression,’’ but with only one
hindlimb.
0.01
(Continued)
American Journal of Physical Anthropology—DOI 10.1002/ajpa
13ORANGUTAN POSITIONAL BEHAVIOR

APPENDIX A. (Continued)
Postural mode % bouts
Postural submode (component 1/component 2/component 3)
Monopedal stand/forelimb-hindlimb-suspend: majority of body weight is supported by one hindlimb, which in
most cases is rather abducted. Torso is often angled, and a fore-and hindlimb support weight in suspension,
either in ipsilateral or contralateral combination.
0.03
Monopedal compression/forelimb-suspend: as for ‘‘monopedal stand/forelimb-suspend,’’ but with the hindlimb
excessively abducted or adducted. Includes flexed and extended postures.
0.06
Monopedal compression/tripedal-suspend: majority of body weight is supported by one abducted or adducted
hindlimb. But torso is horizontal or near horizontal, and remaining three limbs support body weight in
suspension.
0.04
Monopedal compression/forelimb-hindlimb-suspend: majority of body weight is supported by one hindlimb,
which is abducted or abducted. Torso is often angled, and a fore- and hindlimb support weight in suspension,
either in ipsilateral or contralateral combination.
0.20
Monopedal compression/hindlimb-suspend: posture in which majority of body mass is supported by one
hindlimb under compression, but other hindlimb supports substantial body weight under suspension. Torso
can be in any position from horizontal to near vertical.
0.13
Monopedal stand/sit: more than half of body weight is supported by a hindlimb in compression. However, body
is learning in part-sitting posture against angled or horizontal support, and ischia contribute to supporting
body mass.
0.03
Monopedal stand/hindlimb-cling: orthograde posture in which majority of body weight is supported by one
hindlimb, but significant body weight is supported by other hindlimb in a ‘‘cling’’ position.
0.01
Cantilever
Extended cantilever (P7a). 0.01
Orthograde forelimb-suspend
Unimanual forelimb-suspend (P8a). 1.17
Bimanual forelimb-suspend (P8b). 0.15
Forelimb-suspend/sit (P8c). 0.46
Forelimb-suspend/sit/hindlimb compression: body mass is supported by one or both forelimbs, ischia, and one
or both hindlimbs in compression.
0.01
Forelimb-suspend/squat (P8d). 0.07
Forelimb-suspend/hindlimb compression: more than half of body weight suspended from one or both forelimbs.
Rest is supported by bipedal or monopedal compression. Trunk is held at least 458above horizontal. Distinct
from Forelimb suspend/hindlimb cling, because knees and hips may be extended or only slightly flexed, and
feet may be placed on supports of any size and orientation, and do not exhibit power grip typical of cling
postures.
1.27
Forelimb-suspend/tripedal compression: majority of body weight is suspended from one forelimb, and trunk is
held at least 458above horizontal. Rest is supported by one forelimb and both hindlimbs in compression.
Knees may be flexed or extended.
0.01
Forelimb suspend/hindlimb cling (P8f). 0.11
Forelimb suspend/lie (P8g). 0.01
Forelimb suspend/hindlimb-compression/hindlimb cling: one forelimb supports body weight in suspension, but
is aided by one hindlimb in compression and other hindlimb is flexed, grasping a support in cling posture,
and supporting more than its own weight.
0.01
Orthograde quadrumanous-suspend: orthograde suspend where body mass may be supported by one or both
hindlimbs in equal or greater proportion than one or both forelimbs.
Trunk-vertical suspend (P8h): orthograde suspension using one or both hindlimbs and one or both forelimbs in
tension in any combination.
1.95
Orthograde ipsilateral suspend/hindlimb compression: orthograde suspension by ipsilateral fore- and hindlimb,
with other hindlimb supporting more than its own weight in compression, and foot is below level of hip.
0.07
Orthograde hindlimb suspend: orthograde suspension by both hindlimbs, in which hindlimbs are abducted to
approximately 458above horizontal and support all body mass. Forelimbs may be used for balance, but not
weight-bearing.
0.03
Forelimb-hindlimb suspend
Ipsilateral suspend (P9a). 7.32
Ipsilateral suspend/hindlimb-compression: suspension by ipsilateral fore- and hindlimb, and compression with
remaining hindlimb. Body is relatively horizontal and on its side. All three support limbs bear approximately
equal body mass.
0.71
Ipsilateral-suspend/ipsilateral-compression: as above, but with ipsilateral fore- and hindlimb supporting body
weight in compression. Limbs in suspension appear to support most body weight.
0.11
Ipsilateral suspend/sit: majority of body weight is suspended by ipsilateral limbs, but ischia support some body
mass.
0.08
Contralateral suspend (P9b). 0.81
Contralateral suspend/hindlimb-compression: suspension by contralateral fore- and hindlimb, with body
relatively horizontal and facing downwards, and with other hindlimb supporting more than its own weight in
‘‘stand’’ posture.
0.07
Pronograde suspend
Quadrumanous-suspend (P10): inverted pronograde suspension involving both hindlimbs and one or both
forelimbs. Used as feeding posture, generally with one forelimb free to harvest food.
3.41
Quadrumanous-suspend/sit: as above, but with additional support from one or both ischia. 0.01
(Continued)
14 S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

APPENDIX A. (Continued)
Postural mode % bouts
Postural submode (component 1/component 2/component 3)
Quadrumanous-suspend/lie: as for ‘‘quadrumanous-suspend’’ but with additional support from back in
horizontal position.
0.04
Forelimb-suspend/pronograde-compression: as for ‘‘pronograde stand/forelimb-suspend,’’ but majority of body
mass is borne by one or both forelimbs in tension while abducted above head. Distinct from ‘‘forelimb-
suspend/hindlimb compression’’ because torso is pronograde. Head faces downward.
0.27
Hindlimb suspend
Extended bipedal hindlimb-suspend: suspension from both hindlimbs, with both hips and knees extended. 0.36
Extended monopedal hindlimb-suspend: suspension from one hindlimb, with extended hip and knee. 0.28
Hindlimb-suspend/forelimb-hindlimb compression: suspension from one extended hindlimb, with less than half
of body weight supported by a fore- and hindlimb in compression. In this study all fore- and hindlimb
combinations were ipsilateral.
0.06
Hindlimb-suspend/hindlimb compression: suspension from one extended hindlimb, with other hindlimb under
compression. Suspended limb supports majority of body weight, and torso is normally near vertical, with head
downward.
0.08
Lie
Lie (P13a, b, d). Includes ‘‘supine lie,’’ ‘‘lateral lie,’’ and ‘‘back lie,’’ as these were not possible to distinguish
when orangutan was in a nest.
10.37
Lie/forelimb-suspend: lie, with one forelimb supporting significant body weight. 0.03
Lie/forelimb-hindlimb-suspend: lie, with one forelimb and one hindlimb supporting significant body weight. 0.01
Sit/lie (P13e). 1.75
Sitllie/forelimb-suspend: as for ‘‘sit/lie,’’ but with-additional support from one forelimb in suspension. 0.03
Postural bridge
Pronograde bridge (P14). 0.08
Orthograde bridge: generally one forelimb and one hindlimb hold supports on each side of gap, with body
spanning gap in orthograde posture, in tension. Mothers used this posture to reduce size of gap to allow,
infant to cross independently at another level.
0.04
1
Where locomotor descriptions follow exact definition of Hunt et al. (1996), the code for definition in their paper is specified (e.g.,
P1a), and only classification details which differ in present study are provided. Where locomotor modes are specific to this study, we
provide a full definition.
APPENDIX B. Locomotor mode definitions
Locomotor mode,submode, description % bouts
Quadrupedal walk
Symmetrical gait walk (L1a)
1
8.00
Irregular gait walk (scramble) (also called pronograde scramble) (L1c1). 9.36
Tripedal walk
Tripedal walk (L2): in this study, nonocomoting limb is not specified, but was most often a forelimb.
Occasionally a forelimb was used in suspension to help support body weight.
0.22
Bipedal walk
Extended bipedal walk (L3a). 1.28
Flexed bipedal walk (L3b). 0.36
Hand-assisted extended bipedal walk: bipedal walk in which hindlimbs bear more than 50% of body mass in full
extension, but one or both forelimbs are used to assist, either in suspension or compression and bear more
than their own weight.
2.88
Hand-assisted flexed bipedal walk: as for ‘‘hand-assisted extended bipedal walk,’’ but with hindlimbs relatively
more bent.
0.57
Bipedal scramble: body is orthograde and majority of body mass is borne by hindlimbs, but hindlimb kinematics
are not characteristic of smooth bipedal gait. Typically, supports are small, irregularly placed, and variously
angled. Hindlimbs may utilize both extension and flexion during gait cycle.
0.14
Hand-assisted bipedal scramble: as above, but one or both forelimbs also bear more than their own weight,
either in compression or suspension. Similar to ‘‘orthograde clamber,’’ but majority of body mass is carried by
hindlimbs.
2.03
Vertical climb
2
Flexed-elbow vertical climb (L8a). 5.59 (0.60)
Inverted flexed-elbow vertical climb: ascent only on angled (20–458) supports, whereby orangutan is effectively
hanging underneath support while ascending.
(0.50)
Ladder climb (L8b). 0.18
Vertical scramble (L8c). 7.08 (0.64)
Extended-elbow vertical climbing (L8d). 1.17 (0.46)
Bimanual pull-up (L8f). 1.28 (0.04)
Vertical climb forelimbs only: vertical climbing in which body mass is borne only by forelimbs in typical forelimb
climbing pattern, but hindlimbs are not used for weight-bearing.
0.18 (0.07)
Vertical descent
2
Rump-first symmetrical descent (L8g1). 1.99 (0.28)
Rump-first scramble descent (L8g2). 4.45 (0.46)
Rump-first forelimbs only descent: rump-first descent in which only forelimbs are used. Hindlimbs may be used
for balance, but do not bear more than their own weight.
0.60 (0.14)
(Continued)
15ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

LITERATURE CITED
Agresti A. 1990. Categorical data analysis. New York: John Wiley &
Sons.
Alexander RM. 1974. The mechanics of jumping by a dog (Canis
familiaris). J Zool (Lond) [A] 173:549–573.
Alexander RM, Jayes AS. 1983. A dynamic similarity hypothesis
for the gaits of quadrupedal mammals. J Zool 201:135–152.
Avis V. 1962. Brachiation: the crucial issue for man’s ancestry.
Southwest J Anthropol 18:119–148.
Basmajian JV. 1965. Muscles alive: their functions revealed by
electromyography. Baltimore: Wilkins and Wilkins Co.
Begun DR. 2002. European hominoids. In: Hartwig WC, editor.
The primate fossil record. Cambridge: Cambridge University
Press. p 339–368.
Boinski S. 1989. The positional behavior and support use of squir-
rel monkeys: ecological implications. J Hum Evol 18:659–677.
Buettner-Janusch J. 1967. Origins of man. New York: John Wiley &
Sons.
Cant JGH. 1987a. Positional behavior of female Bornean orang-
utans (Pongo pygmaeus). Am J Primatol 12:71–90.
Cant JGH. 1987b. Effects of sexual dimorphism in body size on
feeding postural behavior of Sumatran orangutans (Pongo pyg-
maeus). Am J Phys Anthropol 74:143–148.
APPENDIX B. (Continued)
Locomotor mode,submode, description % bouts
Rump-first cascade descent: equivalent to ‘‘head-first cascade descent,’’ but rump-first. 0.18
Rump-first extended elbow descent: kinematically reverse of ‘‘vertical climb-extended elbow,’’ with limbs moving
in sequence, normally hand over hand, foot over foot.
0.28 (0.07)
Fire pole slide (L81). 0.18 (0.04)
Head-first descent (scramble) (L8h2). 0.36 (0.11)
Head-first descent (cascade) (L8h3). 0.07
Pronograde slide (L8k). 0.04 (0.04)
Sideways vertical descent (L8i). 1.07 (0.32)
Cartwheel descent: descent in which limbs grasp supports in motion which resembles limb sequence of human
cartwheels.
0.14 (0.07)
Torso-orthograde suspensory locomotion
Brachiate (L9a). 6.15
Forelimb swing (L9d). 8.25
Flexed-elbow forelimb swing (L9e). 0.18
Orthograde transfer (L9f). 6.05
Orthograde clamber (L9g). 14.37
Arrested drop (L9h). 0.85
Torso-pronograde suspensory locomotion
Inverted quadrupedal walk (L10a). 2.28
Inverted tripedal walk: as above, but with only three limbs. 0.11
Inverted quadrupedal run (L10b). 0.04
Inverted pronograde scramble (L10c). 1.28
Hindlimb swing: body is held upside-down, and animal swings on one or both hindlimbs. Often used as
intermediary form of locomotion to reorient body between two longer bouts of different locomotor modes.
0.14
Forelimb-hindlimb swing: suspensory locomotion which may or may not follow regular limb sequence,
utilizing both forelimbs and hindlimbs in both orthograde and pronograde positions.
Cartwheel swing: sequence of suspensory locomotion on horizontal or negatively inclined support which
resembles sequence of limb usage seen in human cartwheels.
0.18
Ipsilateral swing: swinging from ipsilateral fore- and hindlimb. Exhibited as single swing to join two other
modes of locomotion.
0.07
Bridge
Cautious pronograde bridge (L11a). 2.53
Inverted pronograde bridge: as above, except with body in inverted pronograde suspension. 0.11
Lunging bridge (L11b). 0.14
Supinograde bridge (L11d). 0.04
Descending bridge (L11e). 0.04
Leap
3
Pronograde leap (L12a). 0.04
Drop
3,4
0.36
Unimanual suspensory drop (L13c): as described in Hunt et al. (1996), but orangutans often tended to use one
limb to maintain contact with support throughout drop, although support does not bear any weight during
fall.
0.57
Bimanual suspensory drop (L13d). 0.04
Sway: based on tree sway (L16) of Hunt et al.’s (1996), but expanded to include any locomotion which relies on
oscillation of supports to progress forward. Also includes locomotion where orangutan swings on vertical
branch/liana (see text for further discussion).
4.55
Ride (L17): orangutans exhibit ‘‘ride’’ to move between different levels in canopy rather than from tree to
ground (Hunt et al., 1996).
0.50
1
Where locomotor descriptions follow exact definition of Hunt et al. (1996), the code for definition in their paper is specified (e.g.,
L1a), and only classification details which differ in present study are provided. Where locomotor modes are specific to this study, we
provide full definition.
2
For vertical climb and descent, values in parentheses are for angled climb.
3
Landings from leaps and drops were not recorded, but were generally pronograde.
4
Refers to drops in which semiposture assumed before drop was obscured from view.
16 S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa

Cant JGH. 1992. Positional behavior and body size of arboreal pri-
mates: a theoretical framework for field studies and an illustra-
tion of its application. Am J Phys Anthropol 88:273–283.
Cant JGH, Youlatos D, Rose MD. 2001. Locomotor behavior of
Lagothrix lagothricha and Ateles belzebuth in Yasuni National
Park, Ecuador: general patterns and non suspensory modes.
J Hum Evol 41:141–166.
Cant JGH, Youlatos D, Rose MD. 2003. Suspensory locomotion
of Lagothrix lagothricha and Ateles belzebuth in Yasuni
National Park, Ecuador. J Hum Evol 44:685–699.
Carlson KJ. 2005. Investigating the form-function interface in
African apes: relationships between principal moments of area
and positional behaviors in femoral and humeral diaphyses.
Am J Phys Anthropol 127:312–334.
Cartmill M. 1974. Pads and claws in arboreal locomotion. In:
Jenkins FA Jr, editor. Primate locomotion. New York: Aca-
demic Press. p 45–83.
Cartmill M. 1985. Climbing. In: Hildebrand M, Bramble DM,
Liem KF, Wake DB, editors. Functional vertebrate morphol-
ogy. Cambridge, MA: Harvard Belknap Press. p 73–88.
Cartmill M, Milton K. 1977. The lorisform wrist joint and the
evolution of brachiating adaptations in the Hominoidea. Am
J Phys Anthropol 47:249–272.
Cartmill M, Hylander WL, Shafland J. 1987. Human structure.
Cambridge, MA: Harvard University Press.
Chivers DJ. 1972. The siamang and the gibbon in the Malay
peninsula. In: Rumbaugh DM, editor. The gibbon and the sia-
mang, volume 1. Basel: Karger. p 103–135.
Collis A. 2001. Siamang locomotion: implications for the arbo-
real clambering hypothesis. Austral Primatol 14:9.
Crompton RH, Li Y, Thorpe SKS, Wang WJ, Savage R, Payne R,
Carey TC, Aerts P, Van Elsacker L, Hofstetter A, Gunther MM,
D’Aout K, DeClerq D. 2003. The biomechanical evolution of erect
bipedality. Cour Forsch Senckenberg 243:115–126.
Dagosto M. 1994. Testing positional behavior of Malagasy lemurs:
a randomization approach. Am J Phys Anthropol 94:189–102.
Dohlinow P, Fuentes A. 1999. The nonhuman primates. Moun-
tainview, CA: Mayfield Publishing Co.
Doran DM. 1989. Chimpanzee and pygmy chimpanzee posi-
tional behavior: the influence of environment, body size, mor-
phology, and ontogeny on locomotion and posture. Ph.D. dis-
sertation, State University of New York, Stony Brook.
Doran DM. 1992a. The ontogeny of chimpanzee and bonobo loco-
motor behavior: a case study of paedomorphism and its be-
havioral correlates. J Hum Evol 23:139–157.
Doran DM. 1992b. Comparison of instantaneous locomotor bout
sampling methods: a case study of adult male chimpanzee locomo-
tor behavior and substrate use. Am J Phys Anthropol 89:85–99.
Doran DM. 1993. The comparative locomotor behavior of chim-
panzees and bonobos: the influence of morphology on locomo-
tion. Am J Phys Anthropol 91:99–115.
Doran DM. 1996. The comparative positional behaviour of
the African apes. In: McGrew W, Nishida T, editors. Great
ape societies. Cambridge: Cambridge University Press. p 213–
224.
Doran DM, Hunt KD. 1994. Comparative locomotor behavior of
chimpanzees and bonobos: species and habitat differences.
In: Wrangham RW, McGrew WC, de Waal FBM, Heltne
PG, editors. Chimpanzee cultures. Cambridge, MA: Harvard
University Press. p 93–106.
Fleagle JG. 1976. Locomotion, posture and comparative anatomy
of Malaysian forest primates. Ph.D. dissertation, Harvard
University.
Fleagle JG. 1980. Locomotion and posture. In: Chivers DJ, edi-
tor. Malayan forest primates: ten years’ study in tropical rain
forest. New York: Plenum Press. p 191–207.
Fleagle JG. 1988. Primate adaptation and evolution. San Diego:
Academic Press.
Gilbert GN. 1981. Modelling society: an introduction to loglinear
analysis for social reseachers. London: George Allen & Unwin.
Gittins SP. 1983. Use of the forest canopy by the agile gibbon.
Folia Primatol (Basel) 40:134–144.
Grand TI. 1972. A mechanical interpretation of terminal branch
feeding. J Mammal 53:198–201.
Hamanda Y. 1985. Primate hip and thigh muscles: comparative
anatomy and dry weights. In: Hollihn U, editor. Bimanual sus-
pensory behavior: morphology, selective advantages and phy-
logeny. Tokyo: University of Tokyo Press. p 131–152.
Harrison T. 1987. The phylogenetic relationships of the early
catarrhine primates: a review of the current evidence. J Hum
Evol 16:41–80.
Harrison T. 2002. Late Oligocene to Middle Miocene catar-
rhines from Afro-Arabia. In: Hartwig WC, editor. The primate
fossil record. Cambridge: Cambridge University Press. p 311–
338.
Harrison T, Rook L. 1997. Enigmatic anthropoid or misunder-
stood ape? The phylogenetic status of Oreopithecus bambolii
reconsidered. In: Begun DR, Ward CV, Rose MD, editors.
Function, phylogeny, and fossils: Miocene hominoid evolution
and adaptations. New York: Plenum Press. p 327–362.
Heglund NC, Taylor CR. 1988. Speed, stride frequency, and
energy cost per stride: how do they change with body size and
gait? J Exp Biol 138:301–318.
Hofstetter AM, Niemitz C. 1997. Comparative gait analyses in
apes. In: Niemitz C, editor. Proceedings of the 5th International
Congress of the German Primatological Society. Berlin: Freie
Universitaet Berlin. p 73.
Hunt KD. 1989. Positional behavior in Pan troglodytes at the
Mahale Mountains and the Gombe Stream National Parks,
Tanzania. Ph.D. dissertation, University of Michigan, Ann
Arbor.
Hunt KD. 1991a. Positional behavior in the Hominoidea. Int J
Primatol 12:95–118.
Hunt KD. 1991b. Mechanical implications of chimpanzee posi-
tional behavior. Am J Phys Anthropol 86:521–536.
Hunt KD. 1992. Positional behavior of Pan troglodytes in the
Mahale Mountains and Gombe Stream National Parks, Tanzania.
Am J Phys Anthropol 87:83–105.
Hunt KD. 2004. The special demands of great ape locomotion
and posture. In: Russon AE, Begun DR, editors. The evolution
of thought: evolutionary origins of great ape intelligence.
Cambridge: Cambridge University Press. p 172–189.
Hunt KD, Cant JGH, Gebo DL, Rose MD, Walker SE, Youlatos
D. 1996. Standardised descriptions of primate locomotor and
postural modes. Primates 37:363–387.
Kano T, Mulavwa M. 1984. Feeding ecology of the pygmy
chimpanzees Pan paniscus of Wamba. In: Susman RL, editor.
The pygmy chimpanzee. New York: Plenum Press. p 233–274.
Larson SG. 1998. Parallel evolution in the hominoid trunk and
forelimb. Evol Anthropol 87:87–99.
MacKinnon J. 1971. The orang-utan in Sabah today. A study of
a wild population in the Ulu Segama reserve. Oryx 11:141–
191.
MacKinnon J. 1974. The behavior and ecology of wild orangu-
tans (Pongo pygmaeus). Anim Behav 22:3–74.
MacLatchy L. 2004. The oldest ape. Evol Anthropol 13:90–103.
MacLatchy L, Gebo D, Kityo R, Pilbeam D. 2000. Postcranial
functional morphology of Morotopithecus bishopi, with impli-
cations for the evolution of modern ape locomotion. J Hum
Evol 39:159–183.
Moya-Sola S, Kohler M. 1996. The first Dryopithecus skeleton:
origins of great-ape locomotion. Nature 379:156–159.
Moya-Sola S, Kohler M, Alba DM, Casanovas-Vilar I, Galindo J.
2004. Pierolapithecus catalaunicus, a new Middle Miocene
great ape from Spain. Science 306:1339–1344.
Napier JR. 1967. Evolutionary aspects of primate locomotion.
Am J Phys Anthropol 27:333–342.
Payne RC. 2001. Musculoskeletal adaptations for climbing in
hominoids and their role as exaptions for the acquisition of
bipedalism. Ph.D. dissertation, University of Liverpool.
Payne RC, Crompton RH, Isler K, Savage R, Vereecke EE,
Gu¨ nther MM, Thorpe SKS, D’Aou
ˆt K. In press. Morphological
analysis of the hindlimb in apes and humans. Part I: compar-
ative anatomy. J Anat.
Payne RC, Crompton RH, Isler K, Savage R, Vereecke EE,
Gu¨ nther MM, Thorpe SKS, D’Aou
ˆt K. In press. Morphological
analysis of the hindlimb in apes and humans. Part II:
Moment arms. J Anat.
17ORANGUTAN POSITIONAL BEHAVIOR
American Journal of Physical Anthropology—DOI 10.1002/ajpa

Pilbeam DR. 2002. Perspectives on the Miocene Hominoidea. In:
Hartwig WC, editor. The primate fossil record. Cambridge:
Cambridge University Press. p 301–310.
Pontzer H, Wrangham RW. 2004. Climbing and the daily energy
cost of locomotion in wild chimpanzees: implications for homi-
noid locomotor evolution. J Hum Evol 46:317–335.
Remis M. 1994. Feeding ecology and positional behavior of west-
ern lowland gorillas (Gorilla gorilla gorilla) in the Central
African Republic. Ph.D. dissertation, Yale University, New
Haven, CT.
Remis M. 1995. Effects of body size and social context on thear-
boreal activities of lowland gorillas in the Central African
Republic. Am J Phys Anthropol 97:413–433.
Remis M. 1998. The effects of body size and habitat on the posi-
tional behavior of lowland and mountain gorillas. In: Strasser
E, Fleagle J, Rosenberger A, McHenry HM, editors. Primate
locomotion—recent advances. New York: Plenum Press. p 95–
106.
Rijksen HD. 1978. A field study on Sumatran orangutans (Pon-
gopygmaeus abelli, Lesson 1827): ecology, behavior and con-
servation. Wageningen: Veenman.
Rook L, Bondioli L, Casali F, Rossi M, Kohler M, Moya-Sola S,
Macchiarelli R. 2004. The bony labyrinth of Oreopithecus
bambolii. J Hum Evol 46:349–356.
Ruff CB. 1989. New approaches to structural evolution of limb
bones in primates. Folia Primatol (Basel) 53:142–159.
Ruff CB, Runestad JA. 1992. Primate limb bone structural
adaptations. Annu Rev Anthropol 21:407–433.
Runestad JA. 1994. Humeral and femoral diaphyseal cross-sec-
tional geometry and articular dimensions in Prosimii and Platyr-
rhini (primates) with application for reconstruction of body mass
and locomotor behavior in Adapidae (primates, Eocene) Ph.D.
dissertation, Johns Hopkins University School of Medicine.
Senut B. 2003. Palaeontological approach to the evolution of
hominid bipedalism: the evidence revisited. Cour Forsch Sen-
kenberg 243:25–134.
Senut B, Pickord M, Gommery D, Mein P, Cheboi K, Coppens Y.
2001. First hominid from the Miocene (Lukeino formation,
Kenya). C R Acad Sci [IIA] 332:137–144.
Srikosamatara S. 1984. Notes on the ecology and behavior of
the hoolock gibbon. In: Preuschoft H, Chivers DJ, Brockelman
WY, Creel N, editors. The lesser apes. Edinburgh: Edinburgh
University Press. p 242–257.
Sugardjito J. 1982. Locomotor behavior of the Sumatran orang
utan (Pongo pygmaeus abelii) at Ketambe, Gunung Leuser
National Park. Malay Nat J 35:57–64.
Sugardjito J, van Hooff JARAM. 1986. Age-sex class differences
in the positional behavior of the Sumatran orang utan (Pongo
pygmaeus abelii) in the Gunung Leuser National Park, Indo-
nesia. Folia Primatol (Basel) 47:14–25.
Taylor CR, Caldwell SL, Rowntree VJ. 1972. Running up and
down hills: some consequences of size. Science 178:1096–1097.
Thorpe SKS, Crompton RH. 2005. Locomotor ecology of wild
orangutans (Pongo pygmaeus abelii) in the Gunung Leuser
Ecosystem, Sumatra, Indonesia: a multivariate analysis using
log-linear modeling. Am J Phys Anthropol 127:58–78.
Thorpe SKS, Crompton RHC, Gu¨ nther MM, Ker RF, Alexander
RM. 1999. Dimensions and moment arms of the hind- and
forelimb muscles of common chimpanzees (Pan troglodytes).
Am J Phys Anthropol 110:179–199.
Tuttle RH, Cortwright D. 1988. Positional behavior, adaptive
complexes, and evolution. In: Schwartz J, editor. Orang-utan
biology. New York: Oxford University Press. p 311–330.
van Schaik CP, Mirmanto E. 1985. Spatial variation in the
structure and litterfall of a Sumatran rainforest. Biotropica 17:
196–205.
Warren RD, Crompton RH. 1997. Locomotor ecology of Lepilemur
edwardsi and Avahi occidentalis. Am J Phys Anthropol 104:
471–486.
Wrangham RW, Conklin-Brittain NL, Hunt KD. 1998. Dietary
response of chimpanzees and cercopithecines to seasonal var-
tiation in food abundance. I. Antifeedants. Int J Primatol 19:
949–970.
Young NM. 2003. A reassessment of living hominoid postcranial
variability: implications for ape evolution. J Hum Evol 45:441–
464.
Young NM, MacLatchy L. 2004. The phylogenetic position of
Morotopithecus. J Hum Evol 46:163–184.
18 S.K.S. THORPE AND R.H. CROMPTON
American Journal of Physical Anthropology—DOI 10.1002/ajpa