Journal of Human Evolution 130 (2019) 61e71

Contents lists available at ScienceDirect

Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol

The mechanical origins of arm-swinging

* Michael C. Granatosky a, , Daniel Schmitt b a Department of Organismal Biology and Anatomy, University of Chicago, Chicago, IL, USA b Department of Evolutionary Anthropology, Duke University, Durham, NC, USA article info abstract

Article history: Arm-swinging is a locomotor mode observed only in , in which the hindlimbs no longer have a Received 9 July 2018 weight bearing function and the forelimbs must propel the body forward and support the entirety of the Accepted 2 February 2019 's mass. It has been suggested that the evolution of arm-swinging was preceded by a shift to inverted quadrupedal walking for purposes of feeding and balance, yet little is known about the me- chanics of limb use during inverted quadrupedal walking. In this study, we test whether the mechanics of Keywords: inverted quadrupedal walking make sense as precursors to arm-swinging and whether there are Arboreal locomotion fundamental differences in inverted quadrupedal walking in primates compared to non- mam- Brachiation Primates mals that would explain the evolution of arm-swinging in primates only. Based on kinetic limb-loading Sloths data collected during inverted quadrupedal walking in primates (seven species) and non-primate Bats (three species), we observe that in primates the forelimb serves as the primary propulsive Suspensory locomotion and weight bearing limb. Additionally, heavier individuals tend to support a greater distribution of body weight on their forelimbs than lighter ones. These kinetic patterns are not observed in non-primate mammals. Based on these findings, we propose that the ability to adopt arm-swinging is fairly simple for relatively large-bodied primates and merely requires the animal to release its grasping foot from the substrate. This study fills an important gap concerning the origins of arm-swinging and illuminates previously unknown patterns of primate locomotor evolution. © 2019 Elsevier Ltd. All rights reserved.

1. Introduction (Hunt et al., 1996; Usherwood et al., 2003). Despite the funda- mental importance of arm-swinging to the evolutionary history of Arm-swinging is a form of suspensory locomotion unique to primates, its origins remain unknown. While the adaptive advan- primates among mammals in which only the forelimbs are used for tages of suspensory locomotion have been discussed in great detail weight-support and forward propulsion. This dynamic and me- (see Grand, 1972; Cartmill and Milton, 1977; Cartmill, 1985; chanically challenging form of locomotion has evolved multiple Granatosky, 2016), no hypothesis has yet to be proposed for why times within primates: at least once in atelines (Jones, 2008; primates utilize arm-swinging while all other mammals are Rosenberger et al., 2008), once in Pygathrix (Byron and Covert, restricted to inverted quadrupedal walking. 2004; Granatosky, 2015; Byron et al., 2017), and once, if not more From a paleontological perspective, investigating the origins of times, in hominoids (Avis, 1962; Lewis, 1971; Tuttle, 1975; Larson, arm-swinging is difficult. Much of these challenges arise from the 1998). Furthermore, even anatomically non-specialized primates fact that, although there are many anatomical modifications pre- have been observed occasionally arm-swinging (Supplementary sent within the postcranial skeleton of specialized arm-swinging Online Material [SOM] Table S1). It should be noted that, primates (e.g., Johnson and Shapiro, 1998; Larson, 1998; Rein although arm-swinging is often used synonymously with brachia- et al., 2015; see list below), these features are not required to tion (e.g., Avis, 1962; Tuttle, 1975; Byron and Covert, 2004; Jones, adopt arm-swinging behavior. It is easy to recognize habitual arm- 2008), in this manuscript we reserve the term brachiation to swingers among extant and extinct taxa with their long forelimbs describe the ricochetal suspensory movements of extant and long manual digits. But the earliest instances of arm-swinging as a critical adaptive behavior may have involved a functional transition in the role of the forelimb without major anatomical change. The number of anatomically non-specialized primates * Corresponding author. observed occasionally arm-swinging (SOM Table S1) supports the E-mail address: [email protected] (M.C. Granatosky). https://doi.org/10.1016/j.jhevol.2019.02.001 0047-2484/© 2019 Elsevier Ltd. All rights reserved. 62 M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 notion that the ability to adopt arm-swinging locomotion does not craniocaudally elongated and dorsoventrally short spinous pro- require any particular anatomical necessity beyond a mobile cesses, dorsally oriented transverse processes, and dorsoventrally shoulder joint and the ability to flex the digits of the hand into a and mediolaterally elongated vertebral bodies (Johnson and functional hook (Granatosky, 2016). Such features are considered Shapiro, 1998; Granatosky et al., 2014). Such profound anatomical synapomorphic for primates (Larson, 1998; Bloch and Boyer, 2002; similarities have led to the as yet untested proposition that the Schmitt and Lemelin, 2002; Boyer et al., 2013); thus, all primates evolution of arm-swinging could be associated with an increase in have the potential to move by arm-swinging. Additionally, Byron body size in arboreal primates that commonly adopt inverted et al. (2017) demonstrated that there are limited mechanical solu- quadrupedal walking (Grand, 1972; Cartmill and Milton, 1977; tions to arm-swinging, and as such anatomical convergence and Cartmill, 1985). parallelism are to be expected, an idea that reflects arguments In support of this notion, it is worth noting that the relative made by Larson (1998). Such evolutionary processes make accurate importance of suspensory locomotion and posture in arboreal reconstructions of phylogenetic relatedness and trait evolution mammals' behavioral repertoires varies as a function of its body tenuous. It should also be noted that many derived postcranial weight; for larger it is mechanically easier to hang below a features of extant hominoids are suitable for both suspensory be- relatively small branch rather than balancing atop it (Grand, 1972; haviors and vertical climbing, so it is uncertain whether they Cartmill, 1985). In this light, it is possible to hypothesize that the evolved first as an adaptation of one of these behaviors and were evolution of arm-swinging in primates proceeds from arboreal subsequently co-opted for the other, or whether the two special- quadrupedal walking to inverted quadrupedal walking and then to ized behaviors evolved simultaneously (for a relevant example, see arm-swinging, with changes in hindlimb and forelimb function at Moya-Sol a et al., 2004; Alba et al., 2010; Alba, 2012). Taken each stage. That explanation, which is intuitively appealing, re- together, a purely paleontological approach to investigating the mains untested and is not a sufficient explanation in light of origins of arm-swinging is problematic because without clear available data. Other mammals use inverted quadrupedal walking diagnostic anatomical features linked with specific behaviors, it is in their normal locomotor repertoire (Fujiwara et al., 2011; impossible to accurately reconstruct the locomotor repertoire of Granatosky, 2016), but arm-swinging has only evolved in pri- the earliest supposed arm-swinging primates. mates. Moreover, the fact that multiple lineages of primates As complement to paleontological analysis, an experimental evolved arm-swinging independently suggests the presence of approach using extant animals has proven greatly informative in some shared mechanical precursor to arm-swinging present in all reconstructing patterns in locomotor evolution across tetrapods primate lineages and not observed in other mammals (Byron et al., (Schmitt and Lemelin, 2002; Reilly et al., 2006; Grossi et al., 2014; 2017). The questions that arise then are: do the mechanics of Nyakatura et al., 2014; Karantanis et al., 2015). Mechanically, the inverted quadrupedal walking make sense as precursors to arm- evolution of arm-swinging must have required a functional reor- swinging and are there fundamental differences in inverted ganization of the forelimb to become the primary propulsive and quadrupedal walking in primates compared to non-primate load-bearing limb (Fleagle et al., 1981; Granatosky, 2016; Byron mammals that would explain the evolution of arm-swinging in et al., 2017). This arrangement presents an especially difficult primates only? biomechanical challenge for primates, which predominantly rely Data on limb-loading patterns of primates during inverted on the hindlimb for both weight support and propulsion during quadrupedal walking are currently limited to three closely-related arboreal quadrupedal walking, climbing, and leaping (Kimura et al., strepsirrhine species. During inverted quadrupedal walking for 1979; Reynolds, 1985; Hirasaki et al.,1993; Hanna et al., 2017). Thus, these species, the forelimbs carry more body weight than the the origin of arm-swinging in primates, and not other animals, is hindlimbs and serve as the primary propulsive limb, while the simultaneously central to understanding the evolution of our order hindlimbs serve primarily braking functions (Ishida et al., 1990; and profoundly difficult to explain. Granatosky et al., 2016), a pattern opposite to what they exhibit One locomotor mode seen as a potential precursor to arm- during above-branch walking. Unfortunately, due to the narrow swinging is inverted quadrupedal walking; a form of suspensory taxonomic breadth of this data it remains unclear whether the movement where animals walk quadrupedally upside down and patterns observed during inverted quadrupedal walking in pri- utilize all four limbs for support and progression (Cartmill and mates reflect strepsirrhine-specific patterns, primate-specific pat- Milton, 1977; Mendel, 1979; Granatosky et al., 2016; Byron et al., terns, or mechanical requirements of below branch quadrupedal 2017). Switching between arboreal quadrupedal walking above locomotion for all mammals. branches and inverted quadrupedal walking below branches is The goal of this study is to test the hypothesis that inverted thought to require very little anatomical and neuromuscular reor- quadrupedal walking in primates is characterized by forelimb- ganization, and as a result has evolved numerous times among dominated features of weight support and propulsion that would varying mammalian clades (Demes et al., 1990; Ishida et al., 1990; make this form of locomotion a potential precursor to arm- Fujiwara et al., 2011; Nyakatura, 2012). Slow, deliberately moving swinging. To address this, we collected measures of weight sup- arboreal quadrupeds, like the lorises and sloths, that commonly port distributiondi.e., peak vertical force (Vpk) and vertical im- adopt inverted quadrupedal walking exhibit considerable pulse (VI)dand whether a limb applies a net propulsive or braking anatomical similarities with arm-swinging primates (e.g., highly force (net fore-aft force impulse) on the forelimbs and hindlimbs mobile wrist and ankle with and deep carpal and tarsal tunnels; during arboreal quadrupedal walking and inverted quadrupedal Grand, 1967; Cartmill and Milton, 1977; Mendel, 1979; Jenkins, walking for seven primate species across a wide range of body mass 1981; Read, 2001), degree of phalangeal curvature in the hands (0.47e5.09 kg). Limb-loading data were collected as animals moved and feet (Jungers et al., 1997), elongation of limbs (Mendel, 1981; at self-selected speeds across a stimulated arboreal runway inte- Preuschoft and Demes, 1984; Swartz, 1989), medially compressed grated with force plates (Granatosky et al., 2016). In addition, and as scapula (Green et al., 2016), globular humeral head and capitulum an essential test of this model, these data were compared to pat- (Miller, 1935; Ashton and Oxnard, 1964; Jenkins et al., 1978; Rose, terns of inverted quadrupedal walking we collected on three spe- 1988), short olecranon process (Rein et al., 2015), longer femoral cies of non-primate (bats and sloths) across a range in neck and greater trochanter that lies inferior to the femoral head body mass (0.03e7.25 kg) moving on similar instrumented run- (Simons et al., 1992; Godfrey and Jungers, 2003; Gebo, 2014), ways (SOM Table S2). M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 63

2. Methods the entire stride and calculating instantaneous speed at each in- terval based on a known distance marked on the runway used to 2.1. Experimental design calibrate the image space. Only walking steps (i.e., duty factor over 50%) in which the animal was traveling in a straight path and not Kinetic gait data were collected from animals housed at the accelerating or decelerating (i.e., steady-state locomotion) were Duke Lemur Center (Durham, NC, USA), North Carolina Zoological selected for analysis. Steady-state locomotion was determined by Park (Asheboro, NC, USA), Central Florida Zoo and Botanical Gar- calculating the instantaneous velocity between subsequent video dens (Sanford, FL, USA), Lubee Bat Conservancy (Gainesville, FL, frames throughout the entire stride, and then using regression USA), and Monkey Jungle Dumont Conservancy (Miami, FL, USA). analysis to determine whether velocity changed throughout the All animal use was approved by the Duke Lemur Center (DLC stride (Granatosky, 2015; Granatosky et al., 2016). Only strides with Research Project #MO-10-11-3) and Duke's Institutional Animal no detectable change in speed (i.e., slope not significantly different Care and Use Committee (IACUC protocol # A270-11-10 and A245- from zero) were analyzed. Additionally, only steps with single-limb 14-10). All animals were adults and were clear of any pathologies or contacts on the plate or those steps in which the forelimb and gait abnormalities (SOM Table S2). hindlimb forces were clearly differentiated were analyzed. Prior to the first trial, animals were weighed by animal care staff. From these data, three variables were calculated for each step: Forces for each day of trials were normalized to the weight recor- (1) peak vertical (Vpk) force; (2) vertical impulse (VI); and (3) net ded from that first day of data collection. The total sampling period fore-aft force impulse. The Vpk force was measured as the highest for any individual lasted no longer than two weeks, so fluctuations magnitude point measured on the vertical force component of the in body weight between the first and last day of sampling were substrate reaction force. The VI and net fore-aft force impulse were likely low. Forelimb and hindlimb single-limb forces were collected measured as the area under the force-time curve in the vertical and while animals walked above [ruffed lemur (Varecia variegata), ring- horizontal component of the substrate reaction force, respectively. tailed lemur (Lemur catta), Coquerel's sifaka (Propithecus coquereli), The net fore-aft force impulse provides a means for differentiating aye-aye (Daubentonia madagascariensis), common squirrel monkey the overall braking or propulsive role of the limb during particular (Saimiri sciureus), Nancy Ma's night monkey (Aotus nancymaae), locomotor behaviors (Demes et al., 1994). The net fore-aft force and white-faced capuchin (Cebus capucinus)] and below impulse for each limb was calculated by subtracting the braking [(V. variegata, L. catta, P. coquereli, D. madagascariensis, S. sciureus, impulse from the propulsive impulse. Positive values indicate a net A. nancymaae, C. capucinus), large flying fox (Pteropus vampyrus), propulsive limb, while negative values indicate a net braking limb. Linnaeus's two-toed sloth (Choloepus didactylus), and common Values approximating zero represent single limb contact forces vampire bat (Desmodus rotundus)] instrumented runways. The were braking and propulsive impulses are approximately equal. instrumented portion of the runways consisted of two Kistler force All force data were normalized for the direction of travel, plates (model 9317B; Kistler, Amherst, NY, USA). Force plate output differing body mass, and orientation. This resulted in comparable was sampled at 12,000 Hz, and imported, summed, and processed force curves that all displayed vertical force as positive value on using BioWare™ v. 5.1 software (Kistler, Amherst, NY, USA), and the vertical axis, braking force as a negative value on the fore-aft then filtered (Butterworth, 60 Hz) and analyzed using MATLAB axis, and propulsive force as a positive value on the fore-aft axis. (MathWorks, Natick, MA, USA). In order to make comparisons between subjects of differing body Locomotor data on below branch quadrupedal locomotion in masses, Vpk force was compared in multiples of body weight (% De. rotundus was collected in a custom-made plexiglass enclosure bw), and VI and net fore-aft force impulse in body weight seconds (0.48 m 0.15 m 0.11 m) with an instrumented ceiling following (%bws). the runway design outlined in previous studies (Riskin and Hermanson, 2005; Riskin et al., 2006; Granatosky, 2018b). The 2.2. Statistical analyses use of this enclosure was required to prevent De. rotundus from flying away during trials. Animals were encouraged to walk back All statistical tests were conducted using MATLAB. Shapiro-Wilk and forth on the ceiling by lightly blowing on them through a straw and Levene's test were used to determine normality and homo- (Riskin and Hermanson, 2005). The material makeup of the ceiling scedasticity within the data (Sokal and Rohlf, 2012). Prior to any consisted of a wire mesh, in which animals were able to grasp with statistical comparisons, body weight-normalized Vpk, VI, and net their claws. The force plates were secured to section of stiff wire fore-aft force impulse for each limb, species, and orientation were mesh (0.025 m 0.15 m). compared with speed (m/s) and contact time (s) using a regression For all other species, locomotor data was collected on an analysis to determine whether the variables of interest were instrumented pole measuring 3.66 m in length and 3.10 cm in influenced by variation in speed or contact time within the sample. diameter that was positioned in their home enclosures. A small The magnitude of substrate reaction forces is thought to be influ- section of dowel was secured on one end of each force plate enced by speed and contact time (Demes et al., 1994). In order to measuring the same diameter as the rest of the runway and large account for the effect of speed and contact time, all Vpk, VI, and net enough to accommodate the fore-aft length of the entire hand or fore-aft force impulse data that demonstrated a significant corre- foot (~10 cm). These instrumented sections were mounted in the lation were examined using an analysis of covariance (ANCOVA) middle of the runway flush with, but separated by a small gap from, with speed and contact time as the covariate to compare across the rest of the runway. limbs and orientation (Olejnik and Algina, 1984; Vickers, 2005). For The animals were video-recorded during trials from a lateral those variables that showed no association with speed or contact view using a GoPro camera (Hero 3 þ Black Edition; GoPro, San time, we used a Kruskal-Wallis test, a non-parametric ANOVA, to Mateo, CA, USA) modified with a Back-Bone Ribcage (Ribcage v1.0; determine whether there were statistically significant differences Back-Bone Gear Inc, Ottawa, ON, Canada), which allows the GoPro across limbs and orientation. No interspecific comparisons were cameras to be outfitted with interchangeable lenses and eliminates conducted. image distortion inherent to the camera (Granatosky, 2015; To assess whether variation in body size between species could Granatosky et al., 2016; Byron et al., 2017). All videos were recor- influence the tendency for forelimb-biased weight support, we ded at 120 frames/s. For each step, the subject's speed was calcu- conducted a regression analysis comparing weight distribution on lated by digitizing a point on the subject's head at each field over the forelimbs (R) during inverted quadrupedal walking in primate 64 M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 and non-primate mammals to body weight. We calculated R for primates the forelimb served a net braking function, while the each individual as: hindlimb was net propulsive (Fig. 3). In sharp contrast, during inverted quadrupedal walking all primate species in this study ¼ = R VIFL VItotal demonstrated higher values for weight support on the forelimbs compared to the hindlimbs. In addition, the forelimb now served as where VIFL is the mean VI in the forelimb for each individual, and the primary propulsive limb, while the hindlimb had a net braking VItotal is the sum of the mean VI in the forelimb and hindlimb for function (Table 1; Fig. 3). This pattern, which is a profound reversal each individual (Raichlen et al., 2009; Larson and Demes, 2011; from the arboreal quadrupedal walking condition, is independent of Granatosky et al., 2018). Values equal to 0.5 represent equal phylogenetic relationships, or the relative frequency at which each weight distribution between the limbs. Values lower than 0.5 species utilizes suspensory locomotor behaviors during their indicate greater weight distribution borne by the hindlimbs, and normal locomotor repertoires (SOM Table S1). All non-primate values greater than 0.5 indicate greater weight distribution borne mammals demonstrated the same pattern of a primarily propul- by the forelimbs. In this study, R was calculated based on mean VI in sive forelimb and braking hindlimb during inverted quadrupedal the forelimb and hindlimb for each individual. This calculation of R walking, but none of the non-primate species demonstrated was necessary, because animals in this study tended to walk with a forelimb-biased weight support. Both bat species and sloths showed diagonal sequence gait (i.e., each hindlimb footfall is followed by a equal contributions weight support between the forelimb and contralateral forelimb footfall) and placed the hindlimb in close hindlimb, or a tendency for the hindlimb to bear a majority of body proximity to the forelimb. Therefore, likelihood of getting single weight. limb forces from both the forelimb and the hindlimb during a single Additionally, during inverted quadrupedal walking in primates stride was low. We also assessed whether R varied with the average the proportion of weight distribution on the forelimbs (R) increased of speed and contact time of the individual during inverted significantly (y ¼ 0.031x þ 0.522; R2 ¼ 0.546; p 0.001) with body quadrupedal walking using a regression analysis to determine if mass (Fig. 4). While all primates demonstrated forelimb-biased fi this association could confound our interpretations. No signi cant weight support (SOM Fig. S1), heavier individuals tended to support relationship was observed between R and speed or contact time. a greater distribution of body weight on their forelimbs than lighter Data reported in this paper are provided as SOM Table S3. individuals. No such relationship (y ¼ 0.009x þ 0.394; R2 ¼ 0.177; p ¼ 0.197) was observed in the non-primate mammals in this study. 3. Results

During arboreal quadrupedal walking, all primates displayed 4. Discussion lower values for weight supportdi.e., lower Vpk (Fig. 1) and VI (Fig. 2)don the forelimbs compared to the hindlimbs (Tables 1 and Arm-swinging is a form of locomotion observed only in pri- 2). Furthermore, during arboreal quadrupedal walking in all mates, in which the forelimbs alone must serve to both propel the

Figure 1. Mean and standard deviation values of peak vertical forces in the forelimbs (yellow) and hindlimbs (blue) during above branch quadrupedal locomotion in primates and inverted quadrupedal walking in primates and non-primate mammals. All data presented as a percentage of body weight (%bw). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 65

Figure 2. Mean and standard deviation values of vertical impulse in the forelimbs (yellow) and hindlimbs (blue) during above branch quadrupedal locomotion in primates and inverted quadrupedal walking in primates and non-primate mammals. All data presented as a percentage of body weight seconds (%bws). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

body forward and support the animal's mass. Yet it is unknown remain balanced above the support. As the body size to support how this form of locomotion evolved. The data here show a size ratio increases, arboreal quadrupedal walking becomes per- parsimonious mechanical pathway for the evolution of arm- ilous (Napier,1967; Cartmill,1985; Lammers and Gauntner, 2008). swinging. In stark contrast to the well-reported pattern seen dur- One solution to this balance problem may be for arboreal animals ing arboreal quadrupedal walking, climbing, and leaping (Kimura to move their center of mass below the support, thereby adopting et al., 1979; Reynolds, 1985; Hirasaki et al., 1993; Hanna et al., suspensory behaviors (Napier, 1967; Cartmill, 1985). While the 2017), when walking quadrupedally below branches all of the pri- relationship between suspensory locomotion and body mass and/ mate species sampled so far rely primarily on the forelimb for both or increasing the feeding sphere has been established in a number weight support and propulsion. This finding indicates that the of sources (Napier, 1967; Grand, 1972; Cartmill, 1985), the ques- fundamental reorganization of forelimb function, which is neces- tion remains: why do primates adopt arm-swinging instead of sary for the evolution of arm-swinging (Byron et al., 2017), occurs simply adopting inverted quadrupedal walking like all other during the adoption of inverted quadrupedal walking. In contrast, mammalian taxa? non-primate mammals do not show forelimb-biased loading pat- The most common answer to this question usually comes down terns during inverted quadrupedal walking and rely much more to an argument of the energetically efficient movement associated heavily on the hindlimb for support. This suggests that non-primate with the pendular mechanics of arm-swinging (Erikson, 1963; mammals are faced with a mechanical constraint that makes them Fleagle, 1974; Chang et al., 2000). However, as has been evi- less likely to adopt arm-swinging behavior. It should be noted that denced by a number of studies, arm-swinging primates rarely our sample of non-primate mammals was limited to only three adopt kinematic and kinetic patterns consistent with pendular species. Therefore, we propose caution in interpreting the patterns mechanics (Bertram et al., 1999; Chang et al., 2000; Usherwood observed in non-primate mammal until data from a greater taxo- et al., 2003; Bertram, 2004; Byron et al., 2017). Furthermore, nomic breadth can be collected. work by Parsons and Taylor (1977) demonstrated relatively high locomotor costs of transport during arm-swinging compared to 4.1. Body size and the origins of arm-swinging arboreal quadrupedal walking or inverted quadrupedal walking. Our data on the association between forelimb functional roles and There are two non-mutually exclusive hypotheses as to why body weight in primates demonstrate that heavier individuals tend animals may adopt suspensory positional behaviors. From an to support a greater proportion of their body weight on their ecological perspective, suspensory movement allows an animal to forelimbs. The mechanism for why this pattern occurs is beyond effectively double its feeding sphere (i.e., animals are able to ac- the scope of this study and remains an area of active investigation, cess resources both above and below the support; Grand, 1972). A but these data suggest that for larger-bodied primates the hindlimb second theory considers a biomechanical model that predicts the is contributing little to overall weight support during inverted ratio of body size to support varies inversely with the ability to quadrupedal walking. As such, we propose that the ability to adopt 66 M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71

Table 1 Summary statistics (mean ± SD) for kinetic variables during above branch quadrupedal locomotion and inverted quadrupedal walking in primates and non-primate mammals.

Species Orientation Speed Limb n Contact time (s) Net fore-aft Peak vertical Vertical impulse Weight distribution (m/s) impulse (%bws) force (%bw) (%bws) on the forelimbs (R)a

Varecia variegata Above 0.82 ± 0.17 FL 40 0.50 ± 0.13 3.00 ± 0.96 56.37 ± 7.34 16.08 ± 4.31 0.39 ± 0.03 HL 31 0.63 ± 0.15 2.31 ± 1.38 77.14 ± 8.96 25.96 ± 6.43 Below 0.69 ± 0.15 FL 38 0.60 ± 0.08 3.51 ± 1.81 82.54 ± 14.95 28.36 ± 4.57 0.59 ± 0.01 HL 33 0.58 ± 0.07 1.78 ± 1.35 59.64 ± 10.94 19.43 ± 3.92 Lemur catta Above 0.57 ± 0.15 FL 28 0.69 ± 0.13 2.92 ± 1.05 44.28 ± 5.10 18.36 ± 3.86 0.36 ± 0.02 HL 25 0.69 ± 0.12 2.58 ± 1.29 78.63 ± 9.86 31.89 ± 4.82 Below 0.55 ± 0.13 FL 52 0.49 ± 0.11 2.09 ± 2.23 70.12 ± 9.32 19.34 ± 4.62 0.61 ± 0.05 HL 41 0.49 ± 0.11 2.36 ± 1.75 43.79 ± 10.84 12.16 ± 5.75 Propithecus coquereli Above 0.64 ± 0.18 FL 14 0.47 ± 0.19 3.07 ± 1.84 50.85 ± 13.77 14.09 ± 6.12 0.32 ± 0.08 HL 15 0.57 ± 0.12 3.40 ± 1.60 83.81 ± 15.11 27.57 ± 8.76 Below 0.51 ± 0.07 FL 59 0.59 ± 0.12 2.71 ± 1.91 79.90 ± 10.71 28.52 ± 5.38 0.67 ± 0.08 HL 51 0.65 ± 0.13 2.74 ± 2.00 34.98 ± 7.80 13.48 ± 4.85 Daubentonia madagascariensis Above 0.61 ± 0.23 FL 46 0.47 ± 0.09 2.79 ± 1.65 53.48 ± 9.11 14.91 ± 4.71 0.40 ± 0.01 HL 28 0.59 ± 0.16 3.98 ± 2.42 73.54 ± 8.19 22.52 ± 6.62 Below 0.44 ± 0.14 FL 26 0.81 ± 0.26 4.32 ± 4.80 66.96 ± 8.33 31.09 ± 9.03 0.61 ± 0.05 HL 22 0.73 ± 0.20 2.98 ± 2.55 49.04 ± 10.14 19.18 ± 7.10 Cebus capucinus Above 0.80 ± 0.20 FL 15 0.44 ± 0.21 2.15 ± 1.50 53.13 ± 16.09 14.88 ± 9.79 0.32 ± 0.05 HL 14 0.59 ± 0.20 2.22 ± 1.47 82.37 ± 14.33 28.16 ± 11.02 Below 0.57 ± 0.11 FL 16 0.62 ± 0.17 3.14 ± 2.91 83.05 ± 9.04 32.95 ± 9.97 0.65 ± 0.05 HL 16 0.63 ± 0.21 2.40 ± 2.68 59.24 ± 12.80 20.15 ± 7.12 Aotus nancymaae Above 0.63 ± 0.08 FL 13 0.68 ± 0.19 1.88 ± 1.17 39.73 ± 11.37 14.73 ± 6.01 0.26 ± 0.03 HL 16 0.88 ± 0.30 2.10 ± 1.32 76.58 ± 7.94 42.51 ± 16.76 Below 0.71 ± 0.16 FL 16 0.68 ± 0.25 1.83 ± 1.79 69.10 ± 11.86 28.62 ± 9.99 0.56 ± 0.02 HL 14 0.74 ± 0.29 2.55 ± 1.65 48.21 ± 12.52 22.93 ± 13.12 Saimiri sciureus Above 0.59 ± 0.13 FL 79 0.21 ± 0.09 0.80 ± 0.64 47.82 ± 7.29 5.49 ± 2.46 0.38 ± 0.04 HL 52 0.23 ± 0.08 1.32 ± 0.82 65.41 ± 7.30 8.85 ± 3.43 Below 0.56 ± 0.07 FL 19 0.42 ± 0.19 2.46 ± 2.11 59.29 ± 8.78 11.80 ± 3.84 0.54 ± 0.04 HL 14 0.52 ± 0.26 2.10 ± 1.19 45.92 ± 4.46 9.67 ± 3.09 Choloepus didactylus Below 0.10 ± 0.03 FL 25 5.10 ± 1.73 14.53 ± 10.90 61.64 ± 15.74 130.72 ± 23.19 0.47 ± 0.00 HL 18 5.17 ± 1.99 13.49 ± 21.12 59.75 ± 15.41 146.40 ± 46.04 Pteropus vampyrus Below 0.17 ± 0.04 FL 16 1.77 ± 0.89 3.79 ± 9.54 43.07 ± 12.26 34.81 ± 17.20 0.38 ± 0.07 HL 25 1.94 ± 0.79 4.23 ± 8.09 65.81 ± 19.14 56.63 ± 22.74 Desmodus rotundus Below 0.18 ± 0.05 FL 22 0.93 ± 0.37 5.92 ± 10.87 61.32 ± 13.76 36.51 ± 18.69 0.41 ± 0.05 HL 27 0.82 ± 0.36 5.33 ± 9.70 68.59 ± 13.33 48.31 ± 16.55

Abbreviations: FL ¼ forelimb; HL ¼ hindlimb. a Weight distribution on the forelimbs (R) calculated as R ¼ VIFL/VItotal, where VIFL is the mean VI in the forelimb for each individual, and VItotal is the sum of the mean VI in the forelimb and hindlimb for each individual. arm-swinging is fairly simple for relatively large-bodied primates, utilizing both arboreal quadrupedal walking and inverted and merely requires the animal to release its grasping foot from the quadrupedal walking during arboreal movement (Doran, 1996; substrate. This view is reinforced by the finding that the two Thorpe and Crompton, 2006; Granatosky, 2018a). In regards to largest-bodied primate species analyzed in this study (i.e., the continued use of arboreal quadrupedal walking at large body C. capucinus and P. coquereli) both occasionally used arm-swinging sizes, this can easily be explained as a function of substrate to cross the runway (SOM Fig. S2), and have been observed arm- diameter. As has been shown in a number of studies, evenly swinging in the wild (SOM Table S1). highly specialized arm-swinging primates commonly utilize As appealing as the relationship between body size and arboreal quadrupedal walking on relatively large diameter sub- forelimb-biased loading may appear, important issues remain. We strates (Cant, 1986; Doran, 1992, 1996; Doran and Hunt, 1996; have thrown around the phrase ‘relatively large-bodied primates’ Workman and Covert, 2005; Thorpe and Crompton, 2006). As as if these animals could be categorized as distinct from other illustrated by Cartmill (1985), balancing on top of substrates only species. Unfortunately, we cannot provide an exact body weight at becomes a problem when the diameter of the body is greater than which primates should switch from inverted quadrupedal walking the diameter of the substrate. Even then, arboreal animals may to arm-swinging. If one was to rely solely on our regression equa- utilize other mechanisms to maintain upright postures (i.e., tion (see Results) then 100% of body weight should be supported by powerful grasping extremities and muscular compensation of the forelimbs at ~ 15 kg. Obviously, this is a gross overestimate as toppling torques) rather than switching to suspensory locomotion much smaller bodied primates, such as gibbons (~5.5 kg), spider (Cartmill, 1985; Lammers and Gauntner, 2008). As for the monkeys (~7.5 kg), and douc langurs (~9.5 kg), all frequently utilize continued use of inverted quadrupedal walking in large-bodied arm-swinging as part of their locomotor repertoire (Granatosky, taxa, we see no mechanical reason to suggest this is not 2018a). Interestingly, despite our attempts to collect data during possible or that this pattern negates our hypothesis. It may be the inverted quadrupedal walking in these taxa, the only form of sus- case that these animals maintain hindlimb support primarily as a pensory locomotion that could be elicited was arm-swinging (see means of increased security rather than weight support. Alter- Byron et al., 2017). Understanding why primates tend to favor arm- natively, models by Nyakatura and Andrada (2013) propose that swinging versus inverted quadrupedal walking at ~5 kg remains an during inverted quadrupedal walking some cautious arboreal open and important question. quadrupeds, such as sloths, utilize the hindlimb to slow down and To make this discussion even more difficult, there are some control the pendular movement of the center of mass. Data relatively large-bodied primates (e.g., , , collected during inverted quadrupedal walking from an and some colobus monkeys; see SOM Table S1) that are capable of would allow one to address these hypotheses. Table 2 Statistical comparisons between forelimb and hindlimb fore-aft impulse, vertical peak force, and vertical impulse. a

Species Orientation Forelimb/hindlimb net fore-aft impulse statistical Forelimb/hindlimb vertical peak force statistical Forelimb/hindlimb vertical impulse statistical comparison comparison (F; p) comparison (F; p) (F; p) ..Gaaok,D cmt ora fHmnEouin10(09 61 (2019) 130 Evolution Human of Journal / Schmitt D. Granatosky, M.C. No Adjusted for covariate Adjusted for No Adjusted for covariate Adjusted for No Adjusted for covariate Adjusted for covariate with contact time (s) covariate with speed covariate with contact time (s) covariate with speed covariate with contact time (s) covariate with speed observed (m/s) observed (m/s) observed (m/s)

Varecia variegata Above e 267.30; <0.001 304.70; <0.001 e 167.70; <0.001 148.50; <0.001 e 64.35; <0.001 61.08; <0.001 Below 186.00; eeee49.99; <0.001 e 76.69; <0.001 e <0.001 Lemur catta Above e 290.00; <0.001 ee288.40; <0.001 298.00; <0.001 e 322.40; <0.001 193.60; <0.001 Below e 109.80; <0.001 eee161.20; <0.001 e 118.5; <0.001 52.45; <0.001 Propithecus coquereli Above e 157.60; <0.001 eee36.22; <0.001 e 29.45; <0.001 30.05; <0.001 Below e 203.00; <0.001 213.30; <0.001 e 587.60; <0.001 607.6; <0.001 e 630.00; <0.001 250.50; <0.001 Daubentonia Above e 167.60; <0.001 207.90; <0.001 e 94.70; <0.001 ee15.66; <0.001 e madagascariensis Below e 40.7; <0.001 eee48.96; <0.001 e 43.85; <0.001 28.58; <0.001 Cebus capucinus Above e 30.7; <0.001 eee24.94; <0.001 e 13.59; <0.001 15.32; <0.001 Below e 69.51; <0.001 e 36.95; ee 17.48; ee <0.001 <0.001 ee ee e< < Aotus nancymaae Above 72.11; 105.30; 52.26; 0.001 42.39; 0.001 < 0.001 <0.001 Below 48.02; eeee19.97; <0.001 e 5.34; 0.029 2.23; 0.147 <0.001 Saimiri sciureus Above e 264.60; <0.001 ee183.50; <0.001 184.80; <0.001 e 111.80; <0.001 50.32; <0.001 Below e 51.37; <0.001 eee26.91; <0.001 e 7.63; <0.001 e Choloepus didactylus Below 32.31; eee0.24; 0.630 ee3.12; 0.085 e <0.001 ee e < e ee e Pteropus vampyrus Below 7.60; 0.009 21.03; 0.001 10.75; 71 0.003 Desmodus rotundus Below 14.61; ee3.50; 0.068 ee e6.77; 0.012 e <0.001 a There is one degree of freedom for all statistical tests. 67 68 M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71

Figure 3. Mean and standard deviation values of fore-aft force impulse in the forelimbs (yellow) and hindlimbs (blue) during above branch quadrupedal locomotion in primates and inverted quadrupedal walking in primates and non-primate mammals. All data presented as a percentage of body weight seconds (%bws). Positive values indicate a net propulsive limb, while negative values indicate a net braking limb. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

4.2. Evolutionary implications and paleontological evidence generate hypotheses about the evolutionary history of arm- swinging in the atelines. Based on ancestral state reconstructions, Based on data collected in this study, we propose that the Jones (2008) proposed that the last common ancestor of the ate- forelimb loading patterns observed during inverted quadrupedal lines was an animal with several traits intermediate between walking in primates would have promoted the eventual adoption of Alouatta and Lagothrix (e.g., biceps lever arm, humeral bituberosity arm-swinging. In our sample, all primate species, regardless of index, humeral head shape), but not yet anatomically specialized body size and taxonomic affiliation, rely primarily on the forelimb for arm-swinging. This suggests that this ancestor was primarily an for weight support and propulsion. As such, we argue that all pri- arboreal quadruped, but likely had had some suspensory abilities mate lineages that commonly utilize inverted quadrupedal walking intermediate between Alouatta and Lagothrix. as part of their behavioral repertoire would have the tendency to The catarrhine fossil record provides a much more indirect view evolve arm-swinging locomotion. Furthermore, based on the rela- of the evolution and specialization of suspensory behaviors. Such tionship between body size and the degree of forelimb-biased complexity arises from rampant parallelism between lineages weight support during inverted quadrupedal walking (discussed (Tuttle, 1975; Larson, 1998; Alba, 2012; Alba et al., 2015), and an above) it is likely that arboreal lineages characterized by relatively attempt to reconstruct trait evolution and phylogenetic relation- large body size would be especially likely to adopt arm-swinging. ships between these lineages is beyond the scope of this study. Paleontological evidence for this hypothesis is supported by There are, however, a few fossils that are promising candidates for trends in body size evolution of the three primate lineagesdi.e., linking inverted quadrupedal walking and arm-swinging. Epi- hominoids (Fleagle, 1978; Grabowski and Jungers, 2017), atelines vindobonensis is a middle Miocene catarrhine that is (Rosenberger and Strier, 1989; Jones, 2008; Rosenberger et al., characterized by a mix of anatomical features that have made 2008), and Pygathrix (Delson et al., 2000)dwhere arm-swinging reconstructing its positional behavior difficult (Zapfe, 1958; Fleagle, represents the dominant form of suspensory locomotion. 1975; Pina Miguel, 2016). Features of both the forelimbs (e.g., a While the tendency for large body size and arm-swinging to be globular articular surface of the humeral head and a degree of associated has been inferred from evidence in the fossil record, a humeral torsion paired with relatively long olecranon process of direct paleontological link between inverted quadrupedal walking the ulna and relatively large tubercles of the humerus relative to the and arm-swinging is more tenuous. However, a closer look at articular surface) and hindlimbs (e.g., a femoral head extending evolutionary history of atelids and hominids reveals that both lin- above the greater trochanter on a slender staff-like femoral shaft eages have taxa that appear to have experimented with fore- and paired with a long facet for the sustentaculum and a basin-shaped hindlimb suspensory positional behaviors prior to evolving more concave facet for the tibial malleolus on the talus) do not match a specialized anatomy associated with arm-swinging. singular locomotor behavior (Zapfe, 1958; Fleagle, 1975; Arias- Jones (2008) conducted a phylogenetic study that incorporated Martorell et al., 2015; Rein et al., 2015; Pina Miguel, 2016). the fossil genera Caipora (~10 ka) and Protopithecus (~600 ka) to Instead, this odd combination of features suggests an animal M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 69

Figure 4. The relationship between the proportion of body weight on the forelimb (R) and body mass for primates (stars) and non-primate mammals (circles) during inverted quadrupedal walking. Across primates the proportion of weight distribution on the forelimbs increased significantly (y ¼ 0.031x þ 0.522; R2 ¼ 0.546; p < 0.001) with body mass. No such relationship (p ¼ 0.197) was observed in the non-primate mammals. Values lower than 0.5 indicate greater weight distribution borne by the hindlimbs, and values greater than 0.5 indicate greater weight distribution borne by the forelimbs. capable of generalized arboreal (or even terrestrial) quadrupedal weight-support and propulsive limb. Furthermore, our data locomotion, combined with climbing, fore- and hindlimb suspen- demonstrate that heavier individuals tend to support a greater sory locomotion, and perhaps arm-swinging to some extent proportion of their body weight on their forelimbs. These patterns (Fleagle, 1975; Arias-Martorell et al., 2015; Rein et al., 2015; Pina are not observed in non-primate mammals. As such, we propose Miguel, 2016). that the ability to adopt arm-swinging is fairly simple for relatively A similar mix of mosaic features are present in the partial large-bodied primates, and merely requires the animal to release its skeleton of the extinct dryopithecine Hispanopithecus laietanus grasping foot from the substrate. These data provide a testable (Almecija et al., 2007; Alba, 2012; Alba et al., 2012; Pina Miguel, hypothesis to explain why arm-swinging locomotion is observed 2016). In the forelimb, the highly curved phalanges (Almecija only in primates, and how it evolved convergently across primates. et al., 2007) and an elbow complex suitable for enabling consid- We hope that these findings will inspire future works, and we see erable pronation and supination (Alba et al., 2012) are consistent ample opportunities to explore the: (1) relationship between body with suspensory positional behaviors. However, H. laietanus also size and the use of arm-swinging and inverted quadrupedal demonstrates an ulnar olecranon morphology consistent with walking in extant primates; (2) anatomical correlates of inverted arboreal quadrupedal walking (Alba, 2012; Alba et al., 2012). quadrupedal walking and whether these can be separated from Similarly, the femora and tibia are indicative of a locomotor arm-swinging; and (3) paleontological record for concrete links repertoire combining suspensory locomotion (e.g., femoral neck- between inverted quadrupedal walking and arm-swinging. shaft angle and relative width of the tibial medial malleolus) with arboreal quadrupedal walking (e.g., the shape of the tibial articular Author contributions surface; Tallman et al., 2013; Pina Miguel, 2016). Taken together, H. laietanus supports the hypothesis that an animal capable of M.C.G. and D.S. designed the initial experiment. M.C.G. collected adopting inverted quadrupedal walking may have represented an and analyzed the data. M.C.G. and D.S. wrote the manuscript. intermediate locomotor stage during the transition from a gener- alized arboreal quadruped locomotion to arm-swinging. Acknowledgments

5. Conclusion We thank all those that helped with animal care and use. Without their help, we would not have been able to complete this In this study, we demonstrate that when primates adopt study. We thank Pierre Lemelin, Christine Wall, Jandy Hanna, Cal- inverted quadrupedal walking the forelimb becomes the primary lum Ross, Zeray Alemseged, Gabriel Yapuncich, David Green, Angel 70 M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71

Zeininger, Zhe-Xi Luo, David Alba, and the anonymous three re- Doran, D.M., Hunt, K.D., 1996. Comparative locomotor behavior of chimpanzees and viewers and associate editor for their comments and inspiration bonobos. In: McGrew, W.C., de Waal, F., Wrangham, R. (Eds.), Cultures. Harvard University Press, Cambridge, pp. 93e108. that improved the overall quality of this work. We also thank Erikson, G., 1963. Brachiation in New World monkeys and in anthropoid . Amanda Carr for providing illustrations. This research was funded Symposia of the Zoological Society of London 10, 135e164. in part by the Leakey Foundation, Force and Motion Foundation, Fleagle, J.G., 1974. Dynamics of a brachiating siamang [ (Symphalangus) syndactylus]. Nature 248, 259e260. and National Science Foundation's Graduate Fellowship Program. Fleagle, J.G., 1975. A small -like hominoid from the Miocene of Uganda. Folia Primatologica 24, 1e15. Fleagle, J.G., 1978. Size distributions of living and fossil primate faunas. Paleobiology Appendix A. Supplementary Online Material 4, 67e76. Fleagle, J.G., Stern, J.T., Jungers, W.L., Susman, R.L., Vangor, A.K., Wells, J.P., 1981. Supplementary online material related to this article can be Climbing: a biomechanical link with brachiation and with bipedalism. Sym- posia of the Zoological Society of London 48, 359e375. found at https://doi.org/10.1016/j.jhevol.2019.02.001. Fujiwara, S., Endo, H., Hutchinson, J.R., 2011. Topsy-turvy locomotion: biomechan- ical specializations of the elbow in suspended quadrupeds reflect inverted gravitational constraints. Journal of Anatomy 219, 176e191. References Gebo, D., 2014. Primate Comparative Anatomy. Johns Hopkins University Press, Baltimore. Alba, D.M., 2012. Fossil apes from the Valles-Pened es Basin. Evolutionary Anthro- Godfrey, L.R., Jungers, W.L., 2003. The extinct sloth lemurs of Madagascar. Evolu- pology 21, 254e269. tionary Anthropology 12, 252e263. Alba, D.M., Almecija, S., Moya-Sol a, S., 2010. Locomotor inferences in Pierolapithecus Grabowski, M., Jungers, W.L., 2017. Evidence of a chimpanzee-sized ancestor of and Hispanopithecus: Reply to Deane and Begun (2008). Journal of Human humans but a gibbon-sized ancestor of apes. Nature Communications 8, 880. Evolution 59, 143e149. Granatosky, M.C., 2015. Kinetic and kinematic patterns of arm-swinging in the red- Alba, D.M., Almecija, S., Casanovas-Vilar, I., Mendez, J.M., Moya-Sol a, S., 2012. shanked douc langur (Pygathrix nemaeus). Journal of Vietnamese Primatology 2, A partial skeleton of the fossil great Hispanopithecus laietanus from Can Feu 33e40. and the mosaic evolution of crown-hominoid positional behaviors. PLoS One 7, Granatosky, M.C., 2016. A mechanical analysis of suspensory locomotion in primates e39617. and other mammals. Ph.D. Dissertation, Duke University. Alba, D.M., Almecija, S., DeMiguel, D., Fortuny, J., Ríos, M.P. de los, Pina, M., Granatosky, M.C., 2018a. Forelimb and hindlimb loading patterns during quadru- Robles, J.M., Moya-Sol a, S., 2015. Miocene small-bodied ape from Eurasia sheds pedal locomotion in the large flying fox (Pteropus vampyrus) and common light on hominoid evolution. Science 350, aab2625. vampire bat (Desmodus rotundus). Journal of Zoology 305, 63e72. Almecija, S., Alba, D.M., Moya-Sol a, S., Kohler,€ M., 2007. Orang-like manual adap- Granatosky, M.C., 2018b. A review of locomotor diversity in mammals with analyses tations in the fossil hominoid Hispanopithecus laietanus: first steps towards exploring the influence of substrate-use, body mass, and intermembral index in great ape suspensory behaviours. Proceedings of the Royal Society B 274, primates. Journal of Zoology 306, 207e216. 2375e2384. Granatosky, M.C., Miller, C.E., Boyer, D.M., Schmitt, D., 2014. Lumbar vertebral Arias-Martorell, J., Alba, D.M., Potau, J.M., Bello-Hellegouarch, G., Perez-P erez, A., morphology of flying, gliding, and suspensory mammals: Implications for the 2015. Morphological affinities of the proximal humerus of Epipliopithecus vin- locomotor behavior of the subfossil lemurs Palaeopropithecus and Babakotia. dobonensis and Pliopithecus antiquus: Suspensory inferences based on a 3D Journal of Human Evolution 75, 40e52. geometric morphometrics approach. Journal of Human Evolution 80, 83e95. Granatosky, M.C., Tripp, C.H., Schmitt, D., 2016. Gait kinetics of above and below Ashton, E.H., Oxnard, C.E., 1964. Functional adaptations in the primate shoulder branch quadrupedal locomotion in lemurid primates. Journal of Experimental girdle. Journal of Zoology 142, 49e66. Biology 219, 53e63. Avis, V., 1962. Brachiation: the crucial issue for man's ancestry. Southwestern Granatosky, M.C., Fitzsimons, A., Zeininger, A., Schmitt, D., 2018. Mechanisms for the Journal of Anthropology 18, 119e148. functional differentiation of the propulsive and braking roles of the forelimbs Bertram, J.E., 2004. New perspectives on brachiation mechanics. American Journal and hindlimbs during quadrupedal walking in primates and felines. Journal of of Physical Anthropology 125, 100e117. Experimental Biology 221, 1e11. Bertram, J.E., Ruina, A., Cannon, C.E., Chang, Y.H., Coleman, M.J., 1999. A point-mass Grand, T.I., 1967. The functional anatomy of the ankle and foot of the slow loris model of gibbon locomotion. Journal of Experimental Biology 202, 2609e2617. (Nycticebus coucang). American Journal of Physical Anthropology 26, 207e218. Bloch, J.I., Boyer, D.M., 2002. Grasping primate origins. Science 298, 1606e1610. Grand, T.I., 1972. A mechanical interpretation of terminal branch feeding. Journal of Boyer, D.M., Yapuncich, G.S., Chester, S.G.B., Bloch, J.I., Godinot, 2013. Hands of early Mammalogy 53, 198e201. primates. American Journal of Physical Anthropology 152, 33e78. Green, D.J., Spiewak, T.A., Seitelman, B., Gunz, P., 2016. Scapular shape of extant Byron, C.D., Covert, H.H., 2004. Unexpected locomotor behaviour: brachiation by an hominoids and the African ape/modern human last common ancestor. Journal old world monkey (Pygathrix nemaeus) from Vietnam. Journal of Zoology 263, of Human Evolution 94, 1e12. 101e106. Grossi, B., Iriarte-Díaz, J., Larach, O., Canals, M., Vasquez, R.A., 2014. Walking like Byron, C.D., Granatosky, M.C., Covert, H.H., 2017. An anatomical and mechanical dinosaurs: chickens with artificial tails provide clues about non-avian theropod analysis of the douc monkey (genus Pygathrix), and its role in understanding locomotion. PLoS One 9, e88458. the evolution of brachiation. American Journal of Physical Anthropology 164, Hanna, J.B., Granatosky, M.C., Rana, P., Schmitt, D., 2017. The evolution of vertical 801e820. climbing in primates: evidence from reaction forces. Journal of Experimental Cant, J.G.H., 1986. Locomotion and feeding postures of spider and howling mon- Biology 220, 3039e3052. keys: field study and evolutionary interpretation. Folia Primatologica 46, 1e14. Hirasaki, E., Kumakura, H., Matano, S., 1993. Kinesiological characteristics of vertical Cartmill, M., 1985. Climbing. In: Hildebrand, M., Bramble, D.M., Liem, K.F., climbing in Ateles geoffroyi and Macaca fuscata. Folia Primatologica 61, 148e156. Wake, D.B. (Eds.), Functional Vertebrate Morphology. Belknap, Cambridge, Hunt, K.D., Cant, J.G.H., Gebo, D.L., Rose, M.D., Walker, S.E., Youlatos, D., 1996. pp. 73e88. Standardized descriptions of primate locomotor and postural modes. Primates Cartmill, M., Milton, K., 1977. The lorisiform wrist joint and the evolution of 37, 363e387. “brachiating” adaptations in the Hominoidea. American Journal of Physical Ishida, H., Jouffroy, F., Nakano, Y., 1990. Comparative dynamics of pronograde and Anthropology 47, 249e272. upside down horizontal quadrupedalism in the slow loris (Nycticebus coucang). Chang, Y., Bertram, J., Lee, D., 2000. External forces and torques generated by the In: Jouffroy, F., Stack, M., Niemitz, C. (Eds.), Gravity, Posture and Locomotion in brachiating white-handed gibbon (Hylobates lar). American Journal of Physical Primates. Editrice Il Sedicesimo, Firenze, pp. 209e220. Anthropology 113, 201e216. Jenkins, F.A., 1981. Wrist rotation in primates: a critical adaptation for brachiators. Delson, E., Terranova, C.J., Jungers, W.L., Sargis, E.J., Jablonski, N.G., 2000. Body mass Symposia of the Zoological Society of London 48, 429e451. in Cercopithecidae (Primates, Mammalia): Estimation and scaling in extinct and Jenkins, F.A., Dombrowski, P.J., Gordon, E.P., 1978. Analysis of the shoulder in extant taxa. Anthropological Papers of the American Museum of Natural History brachiating spider monkeys. American Journal of Physical Anthropology 48, 83, 1e159. 65e76. Demes, B., Jungers, W.L., Nieschalk, U., 1990. Size-and speed-related aspects of Johnson, S.E., Shapiro, L.J., 1998. Positional behavior and vertebral morphology in quadrupedal walking in slender and slow lorises. In: Jouffroy, F., Stack, M., atelines and cebines. American Journal of Physical Anthropology 105, 333e354. Niemitz, C. (Eds.), Gravity, Posture and Locomotion in Primates. Editrice Il Jones, A.L., 2008. The evolution of brachiation in ateline primates, ancestral char- Sedicesimo, Firenze, pp. 175e197. acter states and history. American Journal of Physical Anthropology 137, Demes, B., Larson, S.G., Stern, J.T., Jungers, W.L., Biknevicius, A.R., Schmitt, D., 1994. 123e144. The kinetics of primate quadrupedalism: “hindlimb drive” reconsidered. Journal Jungers, W.L., Godfrey, L.R., Simons, E.L., Chatrath, P.S., 1997. Phalangeal curvature of Human Evolution 26, 353e374. and positional behavior in extinct sloth lemurs (Primates, Palaeopropithecidae). Doran, D.M., 1992. The ontogeny of chimpanzee and pygmy chimpanzee locomotor Proceedings of the National Academy of Sciences USA 94, 11998e12001. behavior: a case study of paedomorphism and its behavioral correlates. Journal Karantanis, N.-E., Youlatos, D., Rychlik, L., 2015. Diagonal gaits in the feathertail of Human Evolution 23, 139e157. glider Acrobates pygmaeus (Acrobatidae, Diprotodontia): Insights for the evo- Doran, D.M., 1996. Comparative positional behavior of the African apes. In: lution of primate quadrupedalism. Journal of Human Evolution 86, 43e54. McGrew, W.C., Marchant, L.F., Nishida, T. (Eds.), Great Ape Societies. Cambridge Kimura, T., Okada, M., Ishida, H., 1979. Kinesiological characteristics of primate University Press, Cambridge, pp. 213e224. walking: its significance in human walking. In: Morbeck, M., Preuschoft, H., M.C. Granatosky, D. Schmitt / Journal of Human Evolution 130 (2019) 61e71 71

Gomberg, N. (Eds.), Environment, Behaviour, Morphology: Dynamic In- Rein, T.R., Harvati, K., Harrison, T., 2015. Inferring the use of forelimb suspensory teractions in Primates. Gustav Fischer, New York, pp. 297e311. locomotion by extinct primate species via shape exploration of the ulna. Journal Lammers, A.R., Gauntner, T., 2008. Mechanics of torque generation during of Human Evolution 78, 70e79. quadrupedal arboreal locomotion. Journal of Biomechanics 41, 2388e2395. Reynolds, T., 1985. Mechanics of increased support weight by the hindlimbs in Larson, S.G., 1998. Parallel evolution in the hominoid trunk and forelimb. Evolu- primates. American Journal of Physical Anthropology 67, 335e349. tionary Anthropology 6, 87e99. Riskin, D.K., Hermanson, J.W., 2005. Biomechanics: independent evolution of Larson, S.G., Demes, B., 2011. Weight support distribution during quadrupedal running in vampire bats. Nature 434, 292. walking in Ateles and Cebus. American Journal of Physical Anthropology 144, Riskin, D.K., Parsons, S., Schutt, W.A., Carter, G.G., Hermanson, J.W., 2006. Terrestrial 633e642. locomotion of the New Zealand short-tailed bat Mystacina tuberculata and the Lewis, O.J., 1971. Brachiation and the early evolution of the Hominoidea. Nature 230, common vampire bat Desmodus rotundus. Journal of Experimental Biology 209, 577e579. 1725e1736. Mendel, F.C., 1979. The wrist joint of two-toed sloths and its relevance to brachi- Rose, M.D., 1988. Another look at the anthropoid elbow. Journal of Human Evolution ating adaptations in the hominoidea. Journal of Morphology 162, 413e424. 17, 193e224. Mendel, F.C., 1981. Foot of two-toed sloths: its anatomy and potential uses relative Rosenberger, A.L., Halenar, L., Cooke, S.B., Hartwig, W.C., 2008. Morphology and to size of support. Journal of Morphology 170, 357e372. evolution of the spider monkey, genus Ateles. In: Campbell, C.J. (Ed.), Spider Moya-Sol a, S., Kohler,€ M., Alba, D.M., Casanovas-Vilar, I., Galindo, J., 2004. Pier- Monkeys: Behaviour, Ecology and Evolution of the Genus Ateles. Cambridge olapithecus catalaunicus, a new Middle Miocene great ape from Spain. Science University Press, Cambridge, pp. 19e49. 306, 1339e1344. Rosenberger, A.L., Strier, K.B., 1989. Adaptive radiation of the ateline primates. Miller, R.A., 1935. Functional adaptations in the forelimb of the sloths. Journal of Journal of Human Evolution 18, 717e750. Mammalogy 16, 38e51. Schmitt, D., Lemelin, P., 2002. Origins of primate locomotion: gait mechanics of the Napier, J.R., 1967. Evolutionary aspects of primate locomotion. American Journal of woolly opossum. American Journal of Physical Anthropology 118, 231e238. Physical Anthropology 27, 333e341. Simons, E.L., Godfrey, L.R., Jungers, W.L., Chatrath, P.S., Rakotosamimanana, B., 1992. Nyakatura, J.A., 2012. The convergent evolution of suspensory posture and loco- A new giant subfossil lemur, Babakotia, and the evolution of the sloth lemurs. motion in tree sloths. Journal of Mammalian Evolution 19, 225e234. Folia Primatologica 58, 197e203. Nyakatura, J.A., Andrada, E., 2013. A mechanical link model of two-toed sloths: no Sokal, R.R., Rohlf, F.J., 2012. Biometry: the Principles and Practice of Statistics in pendular mechanics during suspensory locomotion. Acta Theriologica 58, Biological Research. Freeman, San Francisco. 83e93. Swartz, S., 1989. Pendular mechanics and the kinematics and energetics of Nyakatura, J.A., Andrada, E., Curth, S., Fischer, M.S., 2014. Bridging “Romer's gap”: brachiating locomotion. International Journal of Primatology 10, 387e418. limb mechanics of an extant belly-dragging lizard inform debate on tetrapod Tallman, M., Almecija, S., Reber, S.L., Alba, D.M., Moya-Sol a, S., 2013. The distal tibia locomotion during the early . Evolutionary Biology 41, 175e190. of Hispanopithecus laietanus: More evidence for mosaic evolution in Miocene Olejnik, S.F., Algina, J., 1984. Parametric ANCOVA and the rank transform ANCOVA apes. Journal of Human Evolution 64, 319e327. when the data are conditionally non-normal and heteroscedastic. Journal of Thorpe, S.K.S., Crompton, R.H., 2006. Orangutan positional behavior and the nature Educational and Behavioral Statistics 9, 129e149. of arboreal locomotion in Hominoidea. American Journal of Physical Anthro- Parsons, P., Taylor, C., 1977. Energetics of brachiation versus walking: a comparison pology 131, 384e401. of a suspended and an inverted pendulum mechanism. Physiological Zoology Tuttle, R., 1975. Parallelism, brachiation, and hominoid phylogeny. In: Luckett, W.P., 50, 182e188. Szalay, F.S. (Eds.), Phylogeny of the Primates. Plenum Press, New York, Pina Miguel, M., 2016. Unravelling the positional behaviour of fossil hominoids: pp. 447e480. Morphofunctional and structural analysis of the primate hindlimb. Ph.D. Usherwood, J.R., Larson, S.G., Bertram, J.E., 2003. Mechanisms of force and power dissertation, Universitat Autonoma de Barcelona. production in unsteady ricochetal brachiation. American Journal of Physical Preuschoft, H., Demes, B., 1984. Biomechanics of brachiation. In: Preuschoft, H., Anthropology 120, 364e372. Chivers, D.J., Brockelman, W.Y., Creel, N. (Eds.), The Lesser Apes: Evolutionary Vickers, A.J., 2005. Parametric versus non-parametric statistics in the analysis of and Behavioral Biology. Edinburgh Press Publishing, Edinburgh, pp. 96e118. randomized trials with non-normally distributed data. BMC Medical Research Raichlen, D.A., Pontzer, H., Shapiro, L.J., Sockol, M.D., 2009. Understanding hind limb Methodology 5, 35. weight support in chimpanzees with implications for the evolution of primate Workman, C., Covert, H.H., 2005. Learning the ropes: The ontogeny of locomotion in locomotion. American Journal of Physical Anthropology 138, 395e402. red-shanked douc (Pygathrix nemaeus), Delacour's (Trachypithecus delacouri), Read, C.S., 2001. A comparative analysis of the wrist and ankle morphology of and Hatinh langurs (Trachypithecus hatinhensis) I. positional behavior. American hominoids and lorisids, with implications for the evolution of hominoid loco- Journal of Physical Anthropology 128, 371e380. motion. Master's Thesis, Durham University. Zapfe, H., 1958. The skeleton of Pliopithecus (Epipliopithecus) vindobonensis Zapfe Reilly, S.M., McElroy, E.J., Odum, R.A., Hornyak, V.A., 2006. Tuataras and salaman- and Hürzeler. American Journal of Physical Anthropology 16, 441e457. ders show that walking and running mechanics are ancient features of tetrapod locomotion. Proceedings of the Royal Society of London B 273, 1563e1568.