Mechanical constraints on the functional morphology of the gibbon hind limb

Subject data

The material used in this study comprised 11 gibbon cadavers of known age and sex (Table 1). Specifically, we employed three white-handed gibbons (H. lar: L1–L3), two pileated gibbons (H. pileatus: P1 and P2), two moloch gibbons (H. moloch: M1 and M2) and four siamangs (Symphalangus syndactylus: S1–S4). All specimens were frozen until required for this study and were eviscerated prior to dissection. Specimens were obtained from The Royal Zoological Society of Antwerp (L1, L3 and S2) and The National Museums of Scotland, Edinburgh (L2, P1, P2, M1, M2, S1, S3 and S4). Most cadavers were eviscerated during post-mortem examination and body mass (prior to evisceration) was not available for all specimens. Therefore, hind limb muscle mass (HLMM) was used to normalize the animals for size. Unfortunately, the iliopsoas muscle was unavailable for dissection, because of its use in a prior study, and was not included in any of the analyses. Data from Payne et al. (2006a) were used to indicate where the iliacus muscle (part of the iliopsoas and an important hip flexor) would be positioned on a graph of PCSA against FL (see below and Table 1).

Scaling and functional muscle groups

The data were normalized assuming geometric similarity (Alexander et al. 1981; Thorpe et al. 1999; Payne et al. 2006a; Williams et al. 2008). Because of post-mortem evisceration, body mass was unavailable for some subjects and so HLMM was used as a normalizing factor. Masses were scaled directly to HLMM, lengths to HLMM^1/3 and areas to HLMM^2/3. HLMM correlated significantly with body mass for the subjects where body mass was known (linear regression, P = 0.002, Fig. 1).

For part of the analysis, muscles were categorized into functional groups, which are given in Table 2, together with the abbreviations used for the hind limb muscles used in the analyses. A weighted harmonic mean was used to calculate group averages of FL; this technique takes the mass of each muscle into account when calculating a mean (for a more detailed description see: Alexander et al. 1981; Thorpe et al. 1999; Payne et al. 2006a).

Anatomical measurements

Hind limb muscles were removed systematically and measurements of isolated muscles were taken. Measurements of mass were taken to the nearest 0.1 g, using an electronic scale (Radwag, Poland, accurate to 0.01 g), whereas measurements of length were taken to the nearest 0.1 mm using a set of digital Vernier callipers (Mitutoyo, UK, accurate to 0.01 mm). Measurements included: muscle tendon unit (MTU) mass and MTU length, muscle belly length and mass. The muscle was then cut along its tendon in order to determine the orientation of the muscle fascicles and the length of internal tendon. FL was measured at three points along the muscle belly and the mean was calculated. Photographs of pennate muscles were taken using a digital camera (Nikon D40) so that the pennation angle (θ) could be measured using custom-written software (LabVIEW 8.2, National Instruments). The pennation angle was measured at 10 points along the muscle belly (to account for internal variation) and the mean was calculated. For muscles with an external tendon, the tendon was removed and a uniform section of known length was weighed; this weight was divided by section length and the density of mammalian tendon (1.12 g cm⁻³; Ker et al. 1988) in order to estimate the cross-sectional area. The tendon length (TL) was measured from its most proximal fibres (in the muscle) to its most distal fibres (insertion on the bone). Muscle function was estimated from the position of the muscle on the skeleton and its line of action.

Muscle physiological cross-sectional area and fascicle length

The PCSA of a muscle is affected by its pennation angle (Alexander, 1968; Burkholder et al. 1994; Thorpe et al. 1999; Payne et al. 2006a). It can be directly related to the maximum isometric force (F_MAX) generating capacity by multiplying it by the maximum isometric stress of vertebrate skeletal muscle (0.3 MPa; Wells, 1965; Fukunaga et al. 2001; Medler, 2002). Although this method is commonly used in functional anatomy it should be noted that the value of maximum isometric stress has been shown to vary between muscles of mammalian species (0.1–0.3 Mpa, Josephson, 1993; Hiroyuki et al. 1996) and so caution should be exercised when making hypotheses based on estimations of F_MAX.

The PCSA was estimated using:

where m is the muscle belly mass, ρ is muscle density (1.06 g cm⁻³; Mendez & Keys, 1960) and l is the muscle FL. Previous studies (Thorpe et al. 1999; Payne et al. 2006a) have observed that pennation angles are close to 20° in most ape limb muscles, suggesting that θ has little effect on the PCSA [as Cos(20°) ≈ 1]. However, pennation angle was included in our calculation of PCSA because the pennation angle of many gibbon hind limb muscles exceeded 20° (maximum θ = 39°, implying a 22% reduction in PCSA), which was considered substantial enough to be taken into account.

Muscle function can be estimated by the position of the muscle or muscle group on a graph of PCSA against FL (Fig. 2; see also Williams et al. 2008). Muscles at the top of the graph, with high PCSA, can produce high levels of force (as PCSA is directly related to F_MAX). The maximum contraction distance is proportional to FL, so muscles on the right-hand side of the graph, where FL is high, can contract over a wide range of motion. Muscles with both high PCSA and long fascicles are capable of producing high levels of work (force × distance); these sit in the middle or at the top on the right-hand side of the graph. Contraction velocity, and therefore power (as power = work ÷ time), can also be said to increase with FL (Zajac, 1989) but only where all other variables are fixed, e.g. of two muscles with identical physiological properties but different lengths, the longer muscle should contract more rapidly (and hence produce more power) as they have a greater number of sarcomeres in series (Zajac, 1989, 1992). In reality, however, the muscle fibre type has a much greater influence on contraction velocity and a myriad of other factors have a profound effect on muscle power output (temperature: Marsh & Bennett, 1986; activation pattern: Biewener, 1998; muscle fibre type: Altringham & Johnson, 1990; Widrick et al. 1996; architectural gear ratio and pennation angle: Alexander, 1996; Azizi et al. 2007).

Muscles on the bottom right-hand side of the graph, with low PCSA and high FL, produce a modest force over a wide range of motion. Finally, the function of muscles on the bottom of the left-hand side of the graph (low PCSA, short fascicles) is debatable. Their function may be related to the stabilization of joints (Williams et al. 2008), they may be used for precision movements or it may simply be that this is the ‘default’ position of unspecialized muscle and muscles specialized for force, power or work deviate from this.

Tendon function

The function of a tendon can be estimated from its gross morphology, where long thin tendons are indicative of elastic energy storage (‘compliant’ MTU) and short thick tendons imply large amounts of muscular contraction and therefore work (i.e. ‘stiff’ MTUs; Ker et al. 1988; Williams et al. 2007; McGowan et al. 2008). The safety factor is an index that links the force-producing capabilities of the muscle to the force-resisting capabilities of the tendon. MTUs with a large PCSA : tendon cross-sectional area (TCSA) ratio can place the tendon under large amounts of stress, eliciting large tendon strain and therefore enabling elastic energy storage. MTUs with a relatively smaller PCSA : TCSA ratio have a lower propensity to stretch the tendon and, hence, are less likely to be associated with elastic energy storage. The safety factor gives an indication of how close the tendon comes to rupture when the muscle undergoes F_MAX. A safety factor of 1 implies that if the muscle contracts with F_MAX this will be just enough to cause the tendon to fail, whereas a safety factor of 2 suggests that F_MAX is half of the force required to rupture the tendon, etc. We estimated the safety factor as follows:

where the maximum tendon stress is 100 MPa (again, this value has been shown to vary between species; Pollock & Shadwick, 1994) and the maximum isometric stress of skeletal muscle is 0.3 MPa (Wells, 1965; Medler, 2002).

A second estimate of tendon function can be made by devising a ratio of TL : FL. Muscles with long fascicles and short tendons possess a large amount of control over the tendon as the fascicles can negate any tendon strain by muscular contraction. In this context, muscles with low TL : FL ratios can be termed ‘stiff’. MTUs with short fascicles and long tendons (high TL : FL) are less able to contract to negate tendon strain because of the relatively shorter fascicles. These MTUs can be termed ‘compliant’ and are more likely to be associated with elastic energy storage. The TL : FL ratio was calculated using:

By including the pennation angle (θ) in our results we calculate an ‘effective fascicle length’ (EFL) by which we divide TL to give a ratio of TL : EFL. In muscles with parallel fibres θ was 0.

Descriptive anatomy

There were few qualitative differences in organization of the hind limb musculature between the different species and between individuals of the same species. One obvious variation was the presence of a plantaris muscle, which was either absent (4 of 11 specimens), completely or partially fused with the lateral head of the gastrocnemius (5/11), or completely separate (2/11, Fig. 3). When present, the thin tendon ran at the medial side of the Achilles tendon and inserted separately on the posterior side of the tuber calcanei.

The gluteus superficialis (called gluteus maximus in humans) was irregularly shaped and had fascicles that were orientated in different directions relative to the insertion tendon (Fig. 4). The muscle had a thin, sheet-like origin, originating on the posterior side of the gluteus medius, across the width of the ilium. It became progressively thicker around the hip joint. A small portion of the belly inserted with an internal tendon onto the greater trochanter and another small part passed down the lateral side of the hip until the proximal end of the femoral diaphysis where in some specimens it was associated with a thickened fascia, probably homologous to the tensor fascia lata in humans. Most of the muscle belly passed posterior to the hip and inserted directly onto the posterior aspect of the femur, without a tendon.

The adductor muscles were fused to varying degrees in all of the specimens, making identification and separation difficult. The adductor longus was quite distinct and the easiest to identify, whereas the adductor brevis was very difficult to identify as a separate muscle and was therefore treated as part of the adductor magnus in the analyses.

A detailed qualitative description of the hind limb anatomy of gibbons has been published previously (Bischoff, 1870; Kanagsuntheram, 1952; Sigmon & Farslow, 1986; Vereecke et al. 2005) and we refer to those publications for a full anatomical description of the gibbon hind limb musculature. Mean anatomical data are presented for each species in Table 3 and Appendix 1.

Hind limb muscle volume

The gluteals (hip extensors), adductors and quadriceps (knee extensors) muscles made up the majority of the hind limb muscle volume (HLMV), together accounting for 58 ± 4% of the total volume (Fig. 5). The lar and moloch gibbons both had larger quadriceps than gluteals (lar: 29% vs. 23% of HLMV; moloch: 26% vs. 19% of HLMV for quadriceps vs. gluteals, respectively), a trend not seen in the pileated gibbon or siamang (pileated: 19% vs. 22% of HLMV; siamang: 14% vs. 21% of HLMV for quadriceps vs. gluteals, respectively). There were few other interspecific differences in muscle volume make-up. The adductor group was the third largest functional muscle group in all of the species, despite including the adductor magnus, which was the largest single hind limb muscle (regardless of whether it was fused with the other adductors) in all species (14.6 ± 0.2% of HLMV). The largest muscle group on the distal limb segment was the plantarflexor group (9.8 ± 1.4% of HLMV), consisting of the triceps surae. The other muscle groups on the shank made up 5% or less of HLMV.

Muscle physiological cross-sectional area and fascicle length

The gibbon hip (gluteals) and knee (quadriceps) extensors showed a relatively higher PCSA and relatively shorter fascicles than the other functional muscle groups of the hind limb (Fig. 6). The knee extensors of the siamang had a relatively lower PCSA than the other gibbon species, suggesting that the quadriceps of the siamangs have a lower propensity for force production. The muscles with the longest fascicles were the bi-articular knee and hip flexors (i.e. sartorius and gracilis), which had a small PCSA. There were no muscles with both high PCSA and long fascicles. The majority of muscle groups (plantarflexors, dorsiflexors, knee flexors, digital flexors, hip rotators and digital extensors) had short fascicles and relatively small PCSAs, putting them on the lower left-hand side of Fig. 6.

The iliacus muscle taken from Payne et al. (2006a) had relatively short fascicles and an intermediate PCSA, positioning it at the middle/left-hand side of the graph (red star, Fig. 6).

Tendon anatomy

The tendons of the hind limb muscles of gibbons display a range of safety factors, implying varying tendon function throughout the hind limb (Fig. 7). The knee flexors and hip extensors (semimembranosus and semitendinosus), dorsal flexors (tibialis anterior), digital extensors (extensor digitorum longus and extensor hallucis longus) and digital flexors (flexor tibialis and flexor fibularis) all had safety factors above 4 (semimembranosus, 6.4; semitendinosus, 11.4; tibialis anterior, 6.6; extensor digitorum longus, 13.5; extensor hallucis longus, 9.8; flexor tibialis, 4.1; flexor fibularis, 6.0). Tendons with lower safety factors include the patellar tendon of the quadriceps (Pat., 2.2) and the Achilles tendon of the triceps (Achilles, 3.1), as well as the tendon of origin of the soleus (2.9) and the tendon of insertion of tibialis posterior (3.1). The safety factor was highly variable between species, and no pattern was observed suggesting that one species had consistently higher or lower safety factors than any other.

Generally, MTUs on the pelvis and thigh had TL : EFL ratios of around or less than 1 (Fig. 8). There was one notable exception to this: the quadriceps had relatively longer tendons and more pennate and shorter fascicles than other muscles on the thigh (e.g. adductor magnus), giving a higher TL : EFL ratio (interspecific means of 3.87, 3.84, 4.36 and 3.12 for rectus femoris, vastus medius, vastus intermedius and vastus lateralis, respectively vs. 0.88 interspecific mean for all other thigh muscles). Muscles on the distal limb segment (shank and foot) had higher TL : EFL ratios than MTUs on the hip and thigh (interspecific mean for all muscles: pelvis, 0.88; thigh, 1.99; shank, 5.15). The highest TL : EFL ratios were seen in tibialis posterior and peroneus longus (interspecific means of 9.23 and 9.13, respectively).

Gluteus superficialis and adductor magnus

The morphology of the gluteus superficialis was similar to that described by Sigmon & Farslow (1986) and Stern (1972), where the authors correlated specific regions of the muscle with human equivalents (e.g. the presence of a pars tensorica is hypothesized as an equivalent of tensor fascia lata). The gibbon's gluteus superficialis had a similar morphology to that of the African great apes (Sigmon & Farslow, 1986; Payne et al. 2006a). Because the muscle is made up of several functional parts with varying fibre orientations, its function is likely to be diverse. The position of the muscle in gibbons (i.e. mainly posterior to the hip joint) suggests that its main role is hip extension, although it may also play some role in abduction and/or stabilizing the hip joint because it also covers the lateral aspect of the hip joint.

The high degree of fusion in the adductor muscles observed in most of our gibbon specimens has not been described by other authors (Sigmon & Farslow, 1986; Payne et al. 2006a). All of the adductor muscles appear to perform the same role (thigh adduction) and probably work together, allowing varying degrees of fusion.

The unpredictability of the presence of a plantaris muscle in gibbons has also been documented by other authors (Sigmon & Farslow, 1986; Vereecke et al. 2005). This muscle is known to be vestigial in many ape species (Sigmon & Farslow, 1986). The plantaris is a dedicated plantarflexor and inverter of the foot, yet its small size and variable presence suggest that it is of minor importance for foot motion.

Muscle volume: adaptations for vertical climbing, orthograde clambering and leaping?

The gluteals, quadriceps and adductors made up the majority of the HLMV of gibbons. Having large (voluminous) muscles in these areas might give insight into the specialization of the gibbon hind limb. Gluteals and quadriceps are hip and knee extensors, respectively, which might be important in several activities such as bipedalism, vertical climbing and leaping. The volume of muscle dedicated to knee and hip extension in the gibbon is likely to be useful for some or all of these activities. Moreover, having a large proportion of muscle mass situated proximally (hip and thigh) will minimize the inertia of the swinging limb during locomotion, thus reducing metabolic work (Steudel, 1996 and see below for further discussion).

Adductor muscles are traditionally associated with keeping the limbs underneath the body (Alexander, 1996). During vertical climbing and orthograde clambering the limbs are often outside the projection of the body's centre of gravity. Therefore, arboreal animals should possess enlarged adductor muscles to cope with the increased muscular demand of such activities (Preuschoft, 2002; Isler, 2005). In gibbons, in which these modes form up to 35% of locomotion (Fleagle, 1976), the adductors made up a similar proportion of the HLMV as reported for other non-human apes (interspecific mean; gibbons: 16.4 ± 5.5% HLMV; bonobos: 18.5%; chimpanzee: 18.8%; gorilla: 22.5%; orang-utan: 16.7%; data on great apes from Thorpe et al. 1999 and Payne et al., 2006a), suggesting that limb adduction has a similar importance across the non-human apes. Humans have significantly smaller adductor muscles (≈ 7% of HLMV, based on estimates from Thorpe et al. 1999 and assuming that each hind limb in humans makes up 19% of the total body mass; Zihlman, 1992), which is probably associated with osteological adaptations to bipedal walking, e.g. the bicondylar or valgus angle of the knees (i.e. medio-distal inclination of the femur, Jones et al. 1992).

Functional implications of the gibbon's hind limb anatomy

The short-fascicled, large-PCSA anatomy of the hip and knee extensors implies that the muscles are suitable for high muscular force production but not for exerting much muscular work. It is possible that gibbons use high levels of force (as distinct from power) for a number of activities, including the dissipation of energy during landing. In this case, the muscles have to work eccentrically to decelerate the gibbon during landing, reducing the magnitude of the forces associated with landing, although our estimations suggest that the patellar tendon would be close to rupture during maximal eccentric loading (see ‘Tendon anatomy: elastic energy storage or ideal mass distribution?’ below and Westing et al. 1991; Demes et al. 1999). However, the short-fascicled large PCSA hip and knee extensors may still be able to produce high levels of power at the joint by means of a power-amplifying mechanism, as observed in several primate genera (Galago: Aerts, 1998; bonobo: Scholz et al. 2006). Power amplifiers usually take one of two forms; some galagos are very proficient leapers and use a tendinous mechanism, where the patellar tendon stores elastic strain energy during pre-stretch, which is released rapidly prior to push-off, amplifying power generation (Alexander, 1995; Aerts, 1998). Bonobos utilize short-fascicled hip and knee extensors (Payne et al. 2006a) coupled with small muscle moment (lever) arms at the hip and knee joints (Payne et al. 2006b) to turn relatively small fascicular contractions into relatively large joint movements (see also Alexander, 1995; Fukunaga et al. 2001). Recent research has shown that bonobos are expert leapers and it is suggested that they use this ‘amplifying’ mechanism for propulsion generation in leaping (Scholz et al. 2006). Our results indicate that gibbons, which are also very able leapers (Fleagle, 1976; Gittins, 1983), have hip and knee extensors that fall into positions on a PCSA against FL graph that are similar to bonobos (Fig. 9). This leads us to hypothesize that both gibbons and bonobos may use a similar mechanism, coupling their short-fibred, large PCSA hip and knee extensors to short muscle moment arms, in order to enhance leaping performance. It is interesting to note that none of the other apes, of which none are remarkable jumpers, have hind limb muscles with similar relative PCSAs as the bonobo or gibbon. Also, although all of the published data (Fig. 9) are scaled in the same way as our gibbon data, these take no account of pennation angle, meaning that the gap between the gibbon's and bonobo's relative PCSA is probably exaggerated, which further strengthens our hypothesis.

Within gibbons, the knee extensors of the siamang have a relatively smaller PCSA than those of other gibbons (Fig. 6). If knee extensors are indeed used to power leaping in gibbons this would suggest that siamangs are less adept or less frequent leapers than the other gibbon species. In support of this, field reports (Fleagle, 1976) indicate that siamangs indeed spend proportionally less time leaping than other species of gibbon (6% vs. 15% for lar gibbons; see ‘Interspecific differences’ below for further discussion).

The bi-articular hip and knee flexors (gracilis and sartorius) in gibbons have relatively longer fascicles than those of any of the non-human ape species (Fig. 9), which may reflect a higher propensity for positioning the hind limb in a wider range of postures. The rapid locomotion of the gibbons through an unstable three-dimensional environment may mean that being able to move the limbs over a wide range of motion has advantages in reaching a branch and avoiding a fall. They are also vertical climbers and orthograde clamberers; it is likely that limb placement is highly variable during this form of locomotion and long-fascicled muscles should provide some aid to this. Indeed, the long-fascicled, low-PCSA muscles of the hind limb of orang-utans (where orthograde clambering is a major activity; Thorpe & Crompton 2005, 2006) have been attributed to varied limb placement during orthograde clambering by Hunt et al. (1996) and Payne et al. (2006a), although when scaled to mass^1/3 the FLs do not seem extraordinary in comparison to other ape species (Fig. 9). The hind limb position is also thought to be important in powering brachiation (Bertram & Chang, 2001) where ‘leg-lift’ raises the centre of mass during the swing phase, resulting in an increase in mechanical energy or a reduction in collisional energy loss on the next swing (Usherwood & Bertram, 2003). Long-fascicled muscles allow a greater range of hind limb motion, enabling a greater upward displacement of the body's centre of mass during the swing and a greater mechanical energy benefit for brachiating gibbons. Hip flexors probably play an important role in this leg-lift. However, data on a major hip flexor, the iliopsoas, were not available, which means that we have underestimated the mass of muscle used to flex the hip and the force involved in these movements. Published data indicate that the iliacus (part of the iliopsoas) is short fascicled with a large PCSA in comparison with the other hip flexors (i.e. rectus femoris, sartorius and gracilis), suggesting that the muscle is unlikely to increase the range of motion of the hip significantly but that it will increase the amount of force available for leg-lift (Fig. 6).

Tendon anatomy: elastic energy storage or ideal mass distribution?

Overall, the tendons in the distal hind limb segment were relatively longer (with respect to FL; high TL : EFL) than those in the proximal limb segment (Fig. 8 and Appendix 1). Longer tendons allow muscle force to be transmitted to the distal limb without the burden of extra muscle mass placed distally, which is detrimental to efficient locomotion (through increased limb inertia; Steudel, 1996). Long tendons also allow short-fascicled muscles to produce force more efficiently by combining isometric muscle contraction with tendon strain, thus keeping the muscle fascicles at optimum length for efficiency (Alexander, 1996). The thickness of tendons with respect to PCSA is shown by the safety factor, where relatively thick tendons have a high safety factor and thinner tendons have a lower safety factor.

One method of power amplification for muscles is the sudden release of elastic energy previously stored in a tendon (Alexander, 1995; Aerts, 1998) and the safety factor can be used to estimate whether this is likely to be the case. The safety factors of the tendons in the gibbon hind limb varied greatly, suggesting different functions (Fig. 7). The lowest safety factor, and hence highest potential tendon stress, was found for the patellar tendon, suggesting that it may be used for elastic energy storage. As the patella tendon is associated with the knee extensors (quadriceps), a low safety factor in this tendon supports the hypothesis that leaping may be powered by a tendinous mechanism. This hypothesis is further supported by the relatively long tendons and relatively short fascicles of the vasti and rectus femoris (Fig. 8), suggesting a relatively ‘compliant’ MTU. Interestingly, our estimations of the safety factor are based on maximum isometric stress, which may be exceeded during eccentric loading (e.g. during cyclical locomotion or landing), further reducing the safety factor (Ker et al. 1988; Westing et al. 1991) and potentially making the patellar tendon vulnerable to rupture under high eccentric loads.

The perceived ‘compliance’ (based on a high TL : EFL) of the MTUs in the distal hind limb could simply be a by-product of minimizing inertia in the distal limb. By using the TL : EFL in combination with the safety factor we can gain some insight into whether or not the distal MTUs may be used to store elastic strain energy or are merely a by-product of limb inertia optimization. The Achilles tendon in gibbons has a very low safety factor and a high TL : EFL ratio, which is due to the remarkably long length of the Achilles tendon in gibbons compared with that of other non-human apes (Payne et al. 2006a; Vereecke & Aerts, 2008). A compliant triceps MTU may play a number of roles during gibbon locomotion including energy storage during bipedalism or leaping (Vereecke et al. 2006b; Vereecke & Aerts, 2008). Alternatively, it may be used to transfer force to the distal limb from the powerful vasti, as in the Galago (Aerts, 1998). It is difficult to know from gross anatomy alone what the function of the Achilles tendon is, especially as it is likely to have a variety of roles given the gibbon's extensive locomotor repertoire. However, the large PCSA of the triceps surae suggests a significant role in hind-limb-dominated locomotion. Further data on tendon properties (Young's modulus, ultimate tensile strength, etc.) and muscle fibre type are needed to yield further insight into the triceps’ role in power production, force transfer and weight support.

Interspecific differences

The few studies investigating the locomotor behaviour of wild gibbons (Whitmoor, 1975; Fleagle, 1976; Gittins, 1983) indicate that there are few interspecific differences in locomotor repertoire, which could explain why only a few interspecific differences in myology were observed in our study. One notable difference was the smaller mean PCSA of the knee extensors of the siamang compared with other gibbon genera. The volume of the knee extensor muscles of siamangs was less than the volume of hip extensor muscles, a pattern also observed in the pileated gibbon but not in the lar gibbon or moloch gibbon. Like siamangs, pileated gibbons spend relatively little time leaping (ca. 5% of their locomotor time; Whitmoor, 1975), yet pileated gibbons had the highest PCSA in their knee extensor muscles of any of our subjects. This suggests that the quadriceps PCSA might not be as good an indicator of leaping frequency as muscle volume, although our sample size is too small to draw any definite conclusions about this. Of the gibbons in our sample, those that spend a greater proportion of time leaping (lar gibbons and moloch gibbons; Fleagle, 1976) had a large muscle volume dedicated to knee extension, and those that spend proportionally less time leaping (siamangs and pileated gibbons; Whitmoor, 1975; Fleagle, 1976) had less muscle volume associated with knee extension.

The lack of significant interspecific differences in the locomotor anatomy of our population could also be attributable to the limited sample size and age of the specimens. Although this was the largest sample number in any quantitative myological study on gibbons to date (11 individuals), the sample number for each species was still relatively low for the purpose of addressing interspecific variation (four siamangs, three lar gibbons, two pileated gibbons and two moloch gibbons). Hence the primary aim of this study was not to investigate interspecific differences in myology but to quantify the general anatomy of the gibbon hind limb and link it to the major locomotor modes utilized by these species. The small interspecific differences that were noted indicate that the presented anatomical data are valid for all studied gibbon species and we would expect all hylobatids to present a similar hind limb musculature as quantified in this study.

All of our animals were kept in captivity, and the area in which they lived is small compared with their natural home range (Milton & May, 1976), so it is likely that they were not as physically active as wild animals. As all of our cadavers were captive animals it is probable that they were subject to similar limitations of activity. At least four of our specimens were over 25 years old (three were of unknown age) and it is likely that this has some effect on the absolute values of some muscle masses, although none died of musculoskeletal pathologies. Unfortunately, these are unavoidable limitations when working with endangered species (the gibbons represented in this study are specified as endangered or critically endangered on the IUCN Red List, IUCN, 2008), as specimens are very difficult to obtain. We would like to underline that, due to these limitations, the provided anatomical data are very valuable as they provide a quantitative database of the hind limb musculature of the gibbon. Such databases are valuable tools for a number of studies investigating comparative anatomy, evolutionary biomechanics and human evolution.

Conclusion

This study has investigated how the gibbon hind limb may cope with the varying mechanical demands placed upon it by the gibbon's varied locomotor repertoire. The short-fascicled, high-PCSA hip and knee extensors are likely to play a role in leaping, potentially via a power-amplifying mechanism using the relatively compliant patellar tendon, whereas the long-fascicled knee and hip flexors enable a wide range of limb positions for support and centre of mass position. Further analyses of moment arms and tendon properties, as well as kinematics and kinetics of gibbon locomotion, are needed to provide further evidence of this hypothesis.