Orangutans (genus Pongo, subfamily Ponginae) are the third closest living relative of modern humans and are equally related to all extant and extinct species nested within the African ape-human clade, the subfamily Homininae (Prado-Martinez et al., 2013; Groves, 2018). Although the present geographic distribution of orangutans is restricted to the Southeast Asian islands of Sumatra and Borneo, fossil evidence indicates that their prehistoric range extended south to Java and as far north as southern China (Harrison et al., 2014; Tschen, 2016). Two extant species are typically recognized, Pongo abelii in Sumatra and Pongo pygmaeus in Borneo (Groves, 1986,Lesson, 1827; Groves, 2001), but recently a third species has also been proposed, Pongo tapanuliensis, living in a small, isolated area of Sumatra (Nater et al., 2017). Differences in the behavioral ecology (e.g., degree of sociality) between P. abelii and P. pygmaeus are well documented (MacKinnon, 1974; Sugardjito and Van Hooff, 1986; Galdikas, 1988; Delgado and van Schaik, 2000) and genetic studies indicate that molecular divergence between these two species extends well into the Pleistocene, e.g., ∼1 Ma; (Zhi et al., 1996; Xu and Arnason, 1996; Steiper, 2006; Arora et al., 2010; Locke et al., 2011, 2022; Banes et al., 2022).

Orangutan habitats in Sumatra are at lower altitudes with higher forest productivity and more soft pulp fruit than in Borneo (Delgado and van Schaik, 2000). These factors may explain why P. abelii has historically had a larger effective population size and are rarely observed on the ground (Delgado and van Schaik, 2000; Ancrenaz et al., 2014). Pongo pygmaeus, however, spend considerably more time on the ground (MacKinnon, 1974; Sugardjito and Van Hooff, 1986;Galdikas, 1988; Delgado and van Schaik, 2000; Thorpe and Crompton, 2009). For example, at the field site of Suaq Balimbing, P. abelii individuals are observed on the ground an average of 0.02% of the time or <0.25% for flanged males (Caroline Schuppli, pers. comm.), whereas at Tuanan, P. pygmaeus individuals spend an average of 2.29% of the time on the ground or ∼5% (ranging up to 13%) for flanged males, who travel, eat, and sometimes even sleep on the ground (Ashbury et al., 2015; Caroline Schuppli, pers. comm.). This is especially notable for travel, as terrestriality could make up 20% of all distance travel in P. pygmaeus (Loken et al., 2013), and 22% of 30-min observed travel bouts include ground travel for flanged males (Ashbury et al., 2015).

It remains unclear why P. abelii are observed to come to the ground less often than P. pygmaeus, but the main explanatory hypotheses relate to habitat continuity, canopy structure, food availability, and/or predation pressure. For example, the preferred food of orangutans is fruit, especially ripe fruit (Galdikas, 1988; Knott, 1998; van Schaik et al., 2009), and it is energetically more efficient for large-bodied apes to travel long distances in between fruiting trees terrestrially rather than arboreally (Ashbury et al., 2015). Over the past several million years there have been increasingly severe periods of low fruit availability in orangutan habitats due to El Nino Oscillations (Harrison and Chivers, 2007). As ripe fruit is typically widely dispersed, it would be more efficient for orangutans to travel between fruiting trees terrestrially rather than arboreally, consistent with observed evidence of greater terrestriality in Bornean orangutans (MacKinnon, 1974; Sugardjito and Van Hooff, 1986; Galdikas, 1988; Delgado and van Schaik, 2000; Thorpe and Crompton, 2009; Ashbury et al., 2015; Caroline Schuppli, pers. comm.), which experience more severe periods of low fruit availability than do their counterparts on Sumatra (Delgado and van Schaik, 2000; Wich et al., 2011). Alternatively, the greater frequency of terrestriality observed in P. pygmaeus could be due to the anthropogenically-caused and dramatic habitat loss that has occurred in recent decades on Borneo (Delgado and van Schaik, 2000; Ancrenaz et al., 2014). However, camera trap studies suggest that habitat continuity does not predict terrestrial locomotor behavior in P. pygmaeus, who use the ground even in primary forest with closed canopy and ample opportunities to move through trees (Loken et al., 2013, 2015).

It is also possible that the presence of tigers on Sumatra and their absence on Borneo could explain the differences in orangutan terrestriality (Medway, 1977; Cant, 1987; Meijaard, 1999; van Schaik, 1999; Ashbury et al., 2015; Wennemann et al., 2022). If true, the possibility that terrestriality may be a longstanding part of Bornean orangutan behavior would be interesting given that P. pygmaeus and P. abelii are otherwise generally thought to be extremely similar to one another in postcranial anatomy (Ashbury et al., 2015). However, because it is only within roughly the past 25 years that it has become broadly accepted to consider P. pygmaeus and P. abelii as distinct species (Xu and Arnason, 1996; Zhi et al., 1996; Arora et al., 2010; Locke et al., 2011), a majority of past comparative studies that included orangutans either did not distinguish between these two taxa or only sampled one or the other (e.g., Corruccini and Ciochon, 1976; Oxnard, 1977; Tocheri et al., 2005, Deane and Begun, 2008; Shaw and Ryan, 2012; Bello-Hellegouarch et al., 2013; Young et al., 2015; Marchi et al., 2016; Sarringhaus et al., 2022; Patel, 2023; but see e.g., Patel et al., 2020; Wennemann et al., 2022). Thus, many aspects of potential variation in orangutan skeletal anatomy still remain unexplored at the species level, including long bone cross-sectional geometry, and there is a clear need for more research examining the functional morphology of orangutans.

Bone strength reflects the mechanical loads placed on the skeleton during life (Wolff, 1892), especially during growth and development when bone formation is rapid and bones are more responsive to the forces to which they are exposed, but it is also influenced by genetic and hormonal variation, overall activity levels, and more (e.g., Wainwright et al., 1976; Lovejoy et al., 1976; Burr et al., 1981; Ruff and Hayes, 1983; Ruff and Hayes, 1983, Schaffler et al., 1985; Burr et al., 1989; Trinkaus et al., 1991; Ruff and Runestad, 1992; Martin et al., 1998; Polk et al., 2000; Currey, 2002; Ruff, 2002; Ruff et al., 2006; Demes, 2007; Habib and Ruff, 2008; Kikuchi and Hamada, 2009; Shaw and Ryan, 2012; Patel et al., 2013; Marchi et al., 2016; Kralick and Zemel, 2020). Across primates, measurements of long bones and their cross-sectional properties reflect the kinds of loads supported during life with, for example, primarily terrestrial taxa having hindlimbs that are relatively stronger, more rigid, and more robust than primarily arboreal taxa (e.g.,Schaffler et al., 1985; Burr et al., 1989; Schultz, 1953; Kimura, 2003; Ruff et al., 2006; Marchi et al., 2016). Thus, studies of cross-sectional geometry extend from the premise that the mechanical loading environment causes bone to adapt functionally and rely on a variety of bone diaphyseal properties, including but not limited to cortical bone thickness, cross-sectional area, and polar section modulus (Zpol; e.g., Demes and Jungers, 1993; Trinkaus et al., 1994; Ruff, 2002). The polar section modulus, which is equal to polar moment of area divided by the maximum distance from the section centroid to the outer bone surface, has been used in studies of long bone structure in primates as it is considered a strong indicator of torsional and (twice) average bending strength as well as average bending strength (Ruff, 2003; Ruff et al., 2013, Young et al., 2010; 2018). Another useful measure of bone cross-sectional geometry is the index of cross-sectional shape using the ratio of maximum to minimum area moments of inertia (Imax/Imin; Daegling, 2002; Patel et al., 2013), as ratios near 1 are more circular and those that deviate from 1 are more elliptical (Patel et al., 2013). This ratio is appealing as it does not require information about anatomical orientation or body mass/size and can be measured on even fragmentary bones (Patel et al., 2013).

Previous research by Ruff et al. (2013) demonstrated that bending strength (measured by Zpol) of skeletal hindlimb-to-forelimb bones (e.g., femur-to-humerus) strength ratios1 corresponds with levels of relative terrestriality in western lowland (Gorilla gorilla gorilla) and mountain gorillas (Gorilla beringei beringei), the latter of which are more terrestrial and have relatively stronger hindlimb bones. Moreover, leg-to-arm strength ratios show a developmental trajectory in comparisons among infant (<2 years old), juvenile (i.e., 2–11 years old), and adult mountain gorillas (Ruff et al., 2013). These results correspond with behavioral observations showing that adult and juvenile mountain gorillas are more terrestrial than infant mountain gorillas and adult western lowland gorillas (Remis, 1994; 1995; Doran, 1996, 1997; Remis, 1998). A similar study found that femoral-humeral strength, based on ratios of polar moment of area (J), relates to relative levels of terrestriality when compared across humans, African apes, and Sumatran orangutans, with the latter showing the relatively strongest arms (Sarringhaus et al., 2022). Sarringhaus (2013) also provides Zpol values for femur-humerus strength proportions in Pan troglodytes, and these data suggest that chimpanzees have considerably stronger femoral relative to humeral strength compared to even the most terrestrial of gorillas. In other words, chimpanzee mean femur-to-humerus Zpol ratio (1.27; Sarringhaus, 2013) appears to be roughly five times the magnitude of the difference between the corresponding mean ratios for adult mountain (0.50) and western lowland gorillas (0.26; Ruff et al., 2013). This somewhat unexpected difference between chimpanzees and gorillas, given their broadly similar locomotor repertoires, may be due in part to greater bending moments that likely act on the forelimb bones during climbing given the larger body masses of gorillas.

In the present study, we build on these earlier studies of relative limb bone strength in extant great apes and humans by focusing specifically on orangutans. Demonstrating that cross-sectional morphology varies between and within species of great ape and other primates in ways that relate to observed differences in arboreal and terrestrial locomotor behavior patterns is critical for forming a baseline from which to reconstruct the locomotor repertoires of extinct taxa (e.g., Deane and Begun, 2008; Ruff et al., 2016; Wennemann et al., 2022). If differences are found among orangutans that correlate to species and sex level differences in terrestriality, this nuance could help facilitate efforts to reconstruct nuanced biomechanical interpretations of the limbs of fossil hominids. Using skeletons of wild unhabituated adult orangutans collected in Borneo and Sumatra between 1905 and 1937, we quantify relative leg-to-arm strength in order to assess the level of functional adaptation to arboreal and terrestrial locomotion in each of these taxa and in relation to published data on other extant hominids. Because orangutans are the most arboreal of all extant great apes (Table 1), we predict that orangutan skeletons should show less relative leg-to-arm strength than previously published data on other great apes (e.g., Ruff et al., 2013; Marchi et al., 2016; Sarringhaus et al., 2016; Burgess, 2018). More specifically, we predict that orangutan leg-to-arm strength ratios will be lower than seen in all gorillas but nearest in value to those that climb the most (i.e., adult western gorillas and infant mountain gorillas), and furthest from those that climb the least (i.e., juvenile and adult mountain gorillas). We also predict that P. pygmaeus will exhibit stronger relative leg-to-arm strength ratios than those of P. abelii based on the observed behavioral differences in relative amounts of terrestrial versus arboreal locomotion between these two species (MacKinnon, 1974; Sugardjito and Van Hooff, 1986; Galdikas, 1988; Thorpe and Crompton, 2009) and the fact that proximal manual and pedal phalanges are more curved in P. abelii than in P. pygmaeus (Wennemann et al., 2022). Finally, flanged males are expected to have stronger relative legs than adult females because they have been observed on the ground more often than adult females and smaller unflanged males (Galdikas, 1979; Rodman and Mitani, 1987; Thorpe and Crompton, 2009; Ancrenaz et al., 2014; Ashbury et al., 2015), although we note that recent camera trap data has found terrestriality to be equally employed by males and females (Loken et al., 2013, 2015).