Shouldering the Burdens of Locomotion and Posture: Glenohumeral Joint Structure in Prosimians

Authors

  • Adrian S. Wright-Fitzgerald,

    1. Department of Health Sciences, Sargent College of Health and Rehabilitation Sciences, Boston University, Boston, Massachusetts
    2. Department of Athletic Training, Duquesne University, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Mark D. Balceniuk,

    1. Department of Physical Therapy, Duquesne University, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Anne M. Burrows

    Corresponding author
    1. Department of Physical Therapy, Duquesne University, Pittsburgh, Pennsylvania
    2. Department of Anthropology, University of Pittsburgh, Pittsburgh, Pennsylvania
    • Department of Physical Therapy, 600 Forbes Avenue, Duquesne University, Pittsburgh, PA 15282
    Search for more papers by this author
    • Fax 412.396.4399


Abstract

Despite its importance in movement of the upper limb, the soft-tissue morphology of the shoulder joint complex (the acromioclavicular, coracoclavicular, and glenohumeral joints) across primates is poorly understood. This study compares soft-tissue morphology of these three shoulder joint components among broad phylogenetic, locomotor, and postural behavior ranges in prosimian primates. Two adult specimens of Galago moholi (a vertical clinger and leaper) were dissected for study, along with one adult each of Cheirogaleus medius (an arboreal quadruped), Eulemur macaco (an arboreal quadruped that also frequently engages in suspensory behavior), and Tarsius syrichta (a vertical clinger and leaper). Because of their role in glenohumeral joint movement and stabilization, the rotator cuff muscles were also dissected and weighed among the species. Results showed that muscle mass of individual components of the rotator cuff musculature may be adaptive to locomotor and postural behaviors of the taxa in this study. Two soft-tissue components of the glenohumeral joint, but not the acromioclavicular and coracoclavicular joints, were also considered adaptive. The quadrupedal species, C. medius and E. macaco, both had glenohumeral ligaments and E. macaco had a relatively deeper glenoid articular surface for the humerus because of the shape of the glenoid labrum. Additionally, this study noted a lack of a teres minor muscle in G. moholi, C. medius, and E. macaco despite previous studies describing them. A relatively robust teres minor muscle was found in T. syrichta. Even with the limited sample dissected here, these results suggest that soft-tissue joint morphology itself may be as adaptive to locomotory and postural styles as osseous morphology. Anat Rec, 293:680–691, 2010. © 2010 Wiley-Liss, Inc.

INTRODUCTION

The comparative and functional morphology of the bones composing the primate shoulder complex (clavicle, scapula, and humerus) are relatively well studied and have been shown to be adapted to both locomotor style and postural behavior of a given species (Ashton and Oxnard, 1964; Oxnard, 1967; Rodman, 1979; Kimes et al., 1981; Larson, 1993; Gebo and Sargis, 1994; Taylor, 1997; Voisin and Balzeau, 2004; Voisin, 2006). Additionally, both the gross and the fiber architectural characteristics of primate shoulder muscles have been shown to be similarly adapted (Tuttle and Basmajian, 1978; Larson, 1988; Larson and Stern, 1992; Anapol and Gray, 2003; Higurashi et al., 2006; Schmidt and Schilling, 2007; Michilsens et al., 2009). However, there is a surprising lack of comparative studies examining bony articulations of the shoulder joint complex (e.g., glenohumeral, acromioclavicular, and coracoclavicular joints) alongside their associated soft-tissue structures. Given that the motions of the “shoulder” itself occur at these separate joints, a detailed understanding of the comparative, functional, and adaptive morphology of this joint complex may inform our insight into morphological adaptations to both locomotor style and postural behavior. Moreover, an increased insight into the adaptive morphology of the shoulder joint complex in primates may assist our efforts in comprehending the selective forces that drive the evolution of the primate shoulder and the process of its modification from a weight-bearing joint to a relatively weight-free structure that is more closely associated with manipulative functions in Homo.

Primate species are usually broadly assigned to a single locomotor category and to typical postural behaviors (e.g., Napier and Napier, 1994; Fleagle, 1999). However, these often do not describe fully the motions that occur at limb joints because of within-species variation of joint motion resulting from seasonal food availability, the presence of dependent offspring, ontogenetic stages, and so forth. Broad categories of typical locomotor and postural behaviors, however, can be assigned to particular species based upon the frequency of time spent in a given behavior within a 24 hr period. Thus, primate species are generally assigned to one of four broad locomotor categories, based on the frequency of observed locomotor activities: arboreal quadrupedalism, terrestrial quadrupedalism, leaping, and brachiation (Fleagle, 1999; Nowak, 1999). Whereas each of these broad categories can be further subdivided into specific locomotor activities (e.g., Hunt et al., 1996; Thorpe and Crompton, 2006), and they are commonly used as gross descriptors of primate locomotion.

Postural behaviors are recognized as activities performed by an individual where there is no displacement of the individual relative to its surroundings (Rose, 1979). All primates spend more of their time in these behaviors than in locomotor behaviors; thus, postural behaviors may play a great adaptive role in shaping morphology of various limb elements (Rose, 1979; McGraw, 1998).

Among extant primates, prosimians are perhaps the least understood in terms of shoulder adaptations (both osteological and muscular) to locomotor and postural behaviors (Larson, 1993; Fleagle, 1999). However, prosimians (lorises, galagos, lemurs, and tarsiers) consist of species that occupy all locomotor categories except for brachiation, and they practice a wide range of postural behaviors such as vertical clinging, suspension, and palmigrade wrist positioning (Napier and Napier, 1994; Fleagle, 1999). In addition, prosimians are widely considered to be the best extant representation of the stem primates (Cartmill, 1972; Martin, 1990; Fleagle, 1999; Soligo and Martin, 2006; Silcox, 2007). Thus, an increased understanding of the adaptive and functional morphology of shoulder joints across prosimians may aid our efforts at reconstructing the morphotype of the earliest primates and their locomotor and postural behaviors.

The aim of this study is to compare the soft-tissue morphological structures of the glenohumeral, acromioclavicular, and coracoclavicular joints among prosimian species chosen for differences in phylogenetic relationships, locomotor style, and postural behaviors. Whereas the rotator cuff muscles themselves are not intrinsically part of these joints; the presence of their tendons is often cited as a major stabilizer to the glenohumeral joint and are, therefore, examined in this study.

Species Included in This Study

Major locomotor style and postural behaviors of each species used in this study are described here, along with a description of the phylogenetic relationships among the species. All taxonomic designations are based upon Groves (2001).

Galago moholi (Lemuriloriformes : Loriformes : Galagonidae).

This small (140–225 g), nocturnal galago primarily inhabits Acacia thornveld savannas of central and southern Africa and remains in an arboreal setting for the majority of its time. G. moholi is categorized as a vertical clinger and leaper, but also spends some time quadrupedally running along branches in a palmigrade posture (Charles-Dominique, 1977 [referring to both G. senegalensis and G. moholi in “lesser galagos”]; Harcourt and Bearder, 1989). When G. moholi lands from a leap, it lands hindlimbs first, later grabbing the substrate with its hands. This species feeds primarily on exudates and invertebrates and obtains these through a vertical clinging posture (Harcourt and Bearder, 1989).

Eulemur macaco (Lemuriloriformes : Lemuriformes : Lemuroidea : Lemuridae).

These medium-sized (1.0–2.5 kg) lemurs are found in the semideciduous forests of Madagascar and are active both in the trees and on the ground. E. macaco is cathemeral (being active during day and night periods) and is reported to use both terrestrial and arboreal quadrupedal running and suspension by both the forelimbs and hindlimbs in feeding (Colquhoun, 1993, 1998; Mittermeier et al., 2008). This species feeds on a tremendously wide variety of foods, but fruits seem to be the greatest percentage of resources (Mittermeier et al., 2008).

Cheirogaleus medius (Lemuriloriformes : Lemuriformes : Cheirogaleoidea : Cheirogaleidae).

The fat-tailed, dwarf lemur is a small (142–217 g) nocturnal species that is primarily a generalized arboreal quadruped found in the dry secondary forests of Madagascar (Lahann, 2007). C. medius feeds primarily on fruits and is not reported to engage in any clinging or suspensory behavior (Lahann, 2007; Mittermeier et al., 2008).

Tarsius syrichta (Tarsisimiiformes : Tarsiiformes : Tarsiidae).

The Philippine tarsier is an exceptionally small (117–134 g) nocturnal primate that is categorized as a forest-dwelling vertical clinger and leaper (Kappeler, 1991; Dagosto et al., 2001). T. syrichta, like other tarsiers, feeds primarily on invertebrates and is not reported to use any quadrupedal or suspensory behaviors.

Study Aims

Using the above taxa, this study aims to assess the gross morphology of the soft-tissue structures of the prosimian shoulder joint complex (glenoid labrum, tendon of the long head of the biceps brachii muscle, joint capsule of the glenohumeral joint, glenohumeral ligaments, acromioclavicular joint capsule, and coracoclavicular ligament) and to collect weights of the “rotator cuff” muscles (supraspinatus, infraspinatus, teres minor, and subscapularis muscles) to: 1) provide the first morphological and quantitative data on these structures in prosimian species, 2) to compare relative weights of the rotator cuff muscles among taxa with respect to phylogenetic position, locomotor style, and postural behavior, and 3) to compare gross morphology of the joint structure among taxa relative to phylogenetic position, locomotor style, and postural behavior. Whereas the sternoclavicular joint is typically included in the shoulder joint complex, we do not include it in this study because of lack of availability.

MATERIALS AND METHODS

Two adult cadaveric specimens of G. moholi, and one adult specimen each of C. medius, E. macaco, and T. syrichta were dissected for study. All specimens were acquired from Duke Lemur Center/Duke University Primate Center, where they were kept in a seminatural environment. All animals died of natural causes. After death, each cadaver was immersed in 10% buffered formalin and stored in this fashion.

All dissections were done on either the right or left upper limb, depending upon availability and condition, and were carried out using 2.5x magnifying loupes except for E. macaco, which was large enough to be dissected without magnifiers. Using microdissection tools, skin and superficial fascia was excised away from the shoulder region to reveal the dorsally located rotator cuff muscles. All other dorsally located musculature attaching to the scapula around the rotator cuff muscles was dissected away (i.e., latissimus dorsi, trapezius, and deltoid muscles). Whereas the latissimus dorsi muscle was not attached to the scapula in any species used in this study (Jouffroy, 1962); it was in the general region of interest and was cleared for an unobstructed view of the rotator cuff area. Muscle identification followed Murie and Mivart (1872), Woollard (1925), and Jouffroy (1962). Once the rotator cuff muscles were located, the upper limb was disarticulated from the trunk by cutting through the midpoint of the clavicle with scissors and cutting connections between the scapula and the rhomboid, latissimus dorsi, serratus anterior, trapezius, pectoralis major, and pectoralis minor muscles. In this fashion, all rotator cuff muscles plus the teres major muscles were left attached to the scapula and humerus (Figs. 1,2).

Figure 1.

(a) Left upper limb of adult Galago moholi,(b) dorsal view, (c) dorsal view with infraspinatus muscle visible, and (d) costal view of rotator cuff components.

Figure 2.

(a) Right upper limb of adult Tarsius syrichta,(b) dorsal view, (c) dorsal view with infraspinatus muscle visible, (d) costal view of rotator cuff components, and (e) dorsal view showing teres minor muscle (Tminor). Abbreviations: TB = triceps brachii muscle; LD = latissimus dorsi muscle; SA = serratus anterior muscle; IS = infraspinatus muscle; IF = infraspinous fossa; BB = biceps brachii muscle; TM = teres major muscle.

All rotator cuff muscles were detached individually from their scapular and humeral attachments using microdissection tools and then blotted dry using a paper towel (Atzeva et al., 2007). Weights were recorded to the nearest 0.01 g using a digital Ohaus Analytical Plus scale.

The acromioclavicular joint was dissected free from overlying musculature (deltoid and trapezius muscles) so that the joint capsule was visible. The coracoclavicular joint/ligament was defined from surrounding fascia and the subclavius muscle, and the glenohumeral joint capsule was defined both dorsally and ventrally in a similar fashion. The tendon of the long head of the biceps brachii muscle was followed through the glenohumeral joint capsule. The joint capsule was opened in all specimens to view the internal aspect of the glenohumeral joint, glenoid labrum, glenohumeral ligaments, and termination of the tendon of the long head of the biceps brachii muscle. The joint capsule was opened by cutting longitudinally through the dorsal surface of the joint capsule using microdissection tools. Once the joint capsule was excised, the humeral head was rotated away from the glenoid fossa of the scapula so that the ventral portion of the capsule, glenoid labrum, and termination of the tendon of the long head of the biceps brachii muscle were visible. The dorsal portion of the capsule was chosen for incision because any glenohumeral ligaments that were present were expected to be located ventrally (Standring, 2004). As no comparative data exist for the presence of the glenohumeral ligament in nonhuman primates, we use humans as a reference in this study.

Limited sample sizes in this study inhibit the ability to perform statistical analyses on rotator cuff musculature weights. However, percentage contribution of each muscle to the entire rotator cuff mass was calculated for each species to allow comparison. Average adult body weights were collected from the literature for all taxa except T. syrichta, where the body weight of the specimen was known, and total rotator cuff muscle mass was compared to these adult body weight values.

RESULTS

Rotator Cuff Muscles

Figures 1–4 show the rotator cuff musculature of the four species in this study. All musculature attached in manners were previously described (e.g., Murie and Mivart, 1872; Woollard, 1925; Jouffroy, 1962). Whereas Murie and Mivart (1872), document the presence of a teres minor muscle in G. crassicaudatus, G. alleni, and some lorisids and lemurids; this study did not locate this muscle in any strepsirrhine species (Figs. 1–3). Jouffroy (1962) describes its distinct, clear presence in all prosimians, but Woollard (1925) describes this muscle in lemuroids as being only “feebly” present. The teres minor muscle was clearly distinct in the T. syrichta specimen used in this study (Fig. 4). Only in this specimen, there was a clear separation between the infraspinatus muscle and the teres minor muscle; all other species had muscle fibers arising from the inferior border of the scapula, as the teres minor muscle is described, but these fibers were inextricably bound to the infraspinatus muscle and had no separate tendon of attachment to the greater tubercle of the humerus (Figs. 1–4).

Figure 3.

(a) Left upper limb of adult Cheirogaleus medius,(b) dorsal view, and (c) costal view of rotator cuff components. Abbreviations: LD = latissimus dorsi muscle; Tr = trapezius muscle; SA = serratus anterior muscle; SS = subscapularis muscle; TM = teres major muscle.

Figure 4.

(a) Right upper limb of adult Eulemur macaco,(b) dorsal view, (c) dorsal view with infraspinatus muscle visible, and (d) costal view of rotator cuff components. Abbreviations: TR = trapezius muscle; SS = subscapularis muscle.

In general gross appearance, both the supraspinatus and infraspinatus muscles appeared in all specimens as robust muscles with unipennate fibers. The teres minor muscle in T. syrichta was very small and had unipennate fibers. However, the subscapularis muscle in all specimens appeared as a robust multipennate muscle (Figs. 1–4 for all muscles).

Table 1 and Fig. 5 display the raw weights of individual muscles in the rotator cuff group, and the ratios of weights of individual rotator cuff muscles to entire muscle mass of the rotator cuff. Whereas there is only one specimen for each species (except G. moholi), it is apparent that the subscapularis muscle makes up the majority of the rotator cuff muscle mass relative to the supraspinatus and infraspinatus muscles in all species. In G. moholi and C. medius the infraspinatus and supraspinatus make up 23.5% and 30%, respectively, of the total muscle mass of the rotator cuff. In both T. syrichta and E. macaco the supraspinatus muscle makes up ∼ 20% of the total mass of the rotator cuff group. The infraspinatus muscle of T. syrichta accounts for roughly 20% of the total rotator cuff mass, and that of E. macaco makes up 26% of total rotator cuff mass (Fig. 5).

Figure 5.

Pie charts showing the relative contributions of each individual muscle of the rotator cuff group to the entire mass of the rotator cuff group in each species. “*”–In the figure for T. syrichta, the infraspinatus muscle is shown as making up 20.5% of the total mass of the rotator cuff group for this species. Here, the diminutive teres minor muscle has been grouped with the infraspinatus muscle component.

Table 1. Individual muscle masses (in g) of rotator cuff group components (percentage of each muscle relative to entire mass of the rotator cuff muscle group) and percentage of total body weight of the species accounted for by the rotator cuff mass (RC%)
SpeciesSupraspinatus (%)Infraspinatus (%)Teres Minor (%)Subscapularis (%)RC%
  • a

    Values for all muscles in this species are the average of two specimens: individual weights of supraspinatus muscle −0.19 and 0.13 g, infraspinatus muscle −0.13 and 0.12 g, subscapularis muscle −0.26 and 0.23 g. “*”: The teres minor muscle of T. syrichta was so small that it accounts for close to 0% of the total mass of the rotator cuff muscle group (sec also Fig.5). Values in parentheses beside the species name are adult body weight pulled from the literature (G. moholi: Harcount and Bearder, 1989; C. medius: Atzeva et al., 2009: E. macaco: Kappeler, 1991) expect for T. syricha. Body weight for this specimen was known in this study.

Galago moholi (200 g)0.16 (30%)a0.125 (23.5%)NA0.245 (46.5%)0.27%
Cheirogaleus medius (156 g)0.18 (30%)0.14 (23%)NA0.29 (%)0.39%
Eulemur macaco (2,445 g)1.08 (19%)1.52 (26%)NA3.20 (55%)0.24%
Tarsius syrichta (120 g)0.08 (20%)0.08 (20%)0.005 (x)*0.25 (60%)0.34%

Figures 6,7 display morphology of the acromioclavicular, coracoclavicular, and glenohumeral joints from specimens used in this study. All species had an acromioclavicular joint capsule that appeared to be grossly similar to each other and to the acromioclavicular joint capsule of humans (e.g., Standring, 2004).

Figure 6.

Shoulder joint complex structures from (top row) Galago moholi and (bottom row) Tarsius syrichta. Abbreviations: GL = glenoid labrum; BB = biceps brachii muscle; AC = acromioclavicular; CC = coracoclavicular; GF = glenoid fossa; GH = glenohumeral; SS = subscapularis muscle; HH = humeral head. Unlabeled arrow at bottom of (a) is pointing to the distal portion of the long head of the biceps brachii muscle tendon.

Figure 7.

Shoulder joint complex structures from (top row) Cheirogaleus medius and (bottom row) Eulemur macaco. Abbreviations: GF = glenoid fossa; HH = humeral head; SS = subscapularis muscle; AC = acromioclavicular; BB = biceps brachii muscle; TB = triceps brachii muscle. Unlabeled arrow in (a) is pointing to the acromioclavicular joint capsule; unlabeled arrows in (b) are pointing to the glenoid labrum; asterisks (*) indicate the position of the glenohumeral ligaments in (a) and (d).

The coracoclavicular joint was clearly visible in G. moholi and E. macaco as dense connective tissue between the coracoid process of the scapula and the deep surface of the clavicle (Figs. 6,7). Whereas in humans it is reported as consisting of two separate ligaments, the trapezoid and conoid ligaments (Standring, 2004); it was observed in this study to be a single united band of dense connective tissue. There was no visible dense connective tissue between the coracoid process of the scapula and the deep surface of the clavicle in either C. medius or T. syrichta.

The glenohumeral joint capsule and associated structures (tendon of the long head of the biceps brachii muscle, glenoid labrum, and glenohumeral ligaments) were clearly visible in most specimens (Figs. 6,7). In all species, the posterior portion of the glenohumeral joint capsule was thickened, where it was reinforced by the tendons of the supraspinatus and infraspinatus muscles (plus the teres minor muscle in T. syrichta). In C. medius and E. macaco the anterior portion of the joint capsule was clearly reinforced on the deep surface by glenohumeral ligaments (Fig. 6). Two ligaments were located in each of these species. No such ventrally located thickenings of the glenohumeral joint capsules were found in G. moholi or T. syrichta (Fig. 7).

The glenoid labrum appeared to be similar across species as a relatively narrow, continuous band of fibrocartilage attached to the periphery of the glenoid fossa of the scapula and, rostrally, to the tendon of the long head of the biceps brachii muscle (Figs. 6,7). Whereas the labrum seems to be similar across these species it appeared to be relatively more “cup-like” in E. macaco than in both G. moholi and T. syrichta, where it seems to be relatively “flatter.”

The tendon of the long head of the biceps brachii muscle was similar in all species. It traveled through the intertubercular groove of the humerus, pierced the glenohumeral joint capsule, and attached into the glenoid labrum and the supraglenoid tubercle of the scapula. There were no apparent differences among species.

The osteological features of the glenoid fossa, acromion process, and proximal portion of the humerus have been well documented in numerous primate taxa (Larson, 1993, for a review) and are not detailed here. However, it bears mentioning that the coracoid process position relative to the humeral head varies among taxa in this study. In E. macaco, the coracoid process covers the majority of the ventral surface of the humeral head (Fig. 7), a condition that is not seen in any other species in this study.

DISCUSSION

Rotator Cuff

Primates, relative to other mammals, are noted for relying primarily on their hindlimbs for propulsion in nonsuspensory locomotion relative to forelimbs (Larson et al., 2000; Schmitt and Lemelin, 2002). Thus, morphology of forelimb structures cannot be interpreted strictly with respect to locomotory function as can be done more heavily with the hindlimb. Instead, primates use the forelimb at least as much in positional, nonlocomotory behaviors that would likely be under separate selection. In this conceptualization of forelimb adaptive function, morphology of the shoulder joint complex and the rotator cuff components (musculature and tendons) may reflect postural and other nonlocomotory behaviors such as food procurement and food handling more heavily than locomotor behaviors (Larson, 1998; Larson et al., 2000; Hanna et al., 2006; Stevens, 2008; Wright et al., 2008).

Results from this study reflect these previous findings regarding forelimb function and adaptive morphology in primates. Whereas sample sizes in this study are limited, and we can make some cautious inferences in light of locomotor style and postural behavior for each species used. Comparison of relative contribution of each rotator cuff muscle to the entire mass of the rotator cuff group among species reveals some interesting trends. In all species the subscapularis muscle makes up the greatest percentage of the total rotator cuff mass, a pattern seen among many comparative primate studies (Inman et al., 1944; Ashton and Oxnard, 1963; Anapol and Gray, 2003; Hirugashi et al., 2006; Potau et al., 2009). Similar to previous studies, this study notes the multipennate nature of this muscle relative to the unipennate nature of the supraspinatus and infraspinatus muscles (and teres minor muscle in T. syrichta).

In quadrupedal primates, the subscapularis muscle medially rotates the shoulder and, perhaps more importantly, stabilizes the glenohumeral joint against the shearing force experienced by the joint capsule during locomotion (Tuttle and Basmajian, 1978; Larson and Stern, 1987; Larson, 1993). Based On this observation, we would have expected to see the relatively greatest subscapularis muscle mass in the quadrupedal species, C. medius and E. macaco. E. macaco did in fact have a high relative subscapularis muscle mass (55%) but T. syrichta, a vertical clinger and leaper, had the highest relative subscapularis muscle mass at 60%. C. medius had a very low relative subscapularis muscle mass. Larson (1988) found that the subscapularis muscle participated in the free arm movements of gibbons, a brachiating suspensory species. Thus, the high percentage of the subscapularis muscle in E. macaco in this results may also reflect their use of suspensory behavior.

Vertical clinging and leaping is not associated in the literature with a high subscapularis muscle mass, high levels of medial rotation of the shoulder, or with the need to support the glenohumeral joint capsule. Although it is possible, the clinging component of this locomotion category requires medial rotation of the shoulder to hold onto the substrate with the hands, which may explain the high relative subscapularis muscle mass in T. syrichta. The other vertical clinger and leaper used in this study, G. moholi, showed the lowest relative subscapularis muscle mass. However, G. moholi is reported to spend some time on the ground bipedally hopping (Harcourt and Bearder, 1989) whereas T. syrichta is not reported to engage in this behavior (Gursky, 2007). This difference in locomotion may explain the differences in relative subscapularis muscle mass between these two vertical clingers and leapers.

The primate supraspinatus and infraspinatus muscles also act to stabilize the glenohumeral joint but with great functional differences otherwise. The infraspinatus muscle is a lateral rotator of the shoulder joint, an activity important during quadrupedal walking and in suspensory behaviors (Whitehead and Larson, 1994; Larson, 1995). In this study, E. macaco had the greatest percentage attributed to infraspinatus muscle (26%). This may reflect the combination of quadrupedal and suspensory behavior seen in E. macaco. The supraspinatus muscle in primates does not seem to have a great active role in locomotion but does stabilize the glenohumeral joint during quadrupedal locomotion, and it is reported to be important in shoulder elevation during reaching activities (Tuttle and Basmajian, 1978; Larson and Stern, 1989, 1992). In this study, both G. moholi and C. medius had the greatest percentage of supraspinatus muscle mass (30%). Whereas quadrupedal walking may not involve a great deal of overhead reaching, it may require stabilization of the glenohumeral joint during progression along terminal branches (Schmitt and Lemelin, 2002; Schmitt, 2003).

Interestingly, E. macaco had the greatest ratio of total rotator cuff muscle mass relative to body weight among all species in this study. This species averages in excess of 2 kg total body mass and is by far the largest animal in this study. Larson (1988) found that gibbons have a high participation of the subscapularis muscle in suspension activities. It is possible that E. macaco is using not only the subscapularis muscle for suspension in a similar manner but, being so large, has a higher relative rotator cuff muscle mass to stabilize the glenohumeral joint from distraction because of the suspension of its great body weight.

Finally, this study did not locate any independent teres minor muscle in any species except for T. syrichta, contrary to previous reports (e.g., Murie and Mivart, 1872; Jouffroy, 1962). In this study, the more caudal aspect of the infraspinatus muscle did appear to be somewhat separated from the remaining portion of the muscle (Figs. 1, 3, and 4) but this semi-independent portion always attached onto the greater tubercle of the humerus with the tendon of the infraspinatus muscle rather than separately onto the greater tubercle where teres minor is reported to insert in related taxa (e.g.,Swindler and Wood, 1982). Thus, this cannot be considered a separate teres minor muscle. Findings of a separate teres minor muscle in galagos and lemurs in previous studies may be because of individual variation within these taxa or to phylogenetic factors. Tarsius is routinely problematic in phylogenetic analyses, being placed either in Haplorrhini with monkeys and apes or in Prosimii with lorises and lemurs (Le Gros Clark, 1949; Groves, 2001). All monkeys and apes for which the rotator cuff has been studied are reported to possess a distinct teres minor muscle (e.g., Ashton and Oxnard, 1963; Swindler and Wood, 1982). The presence of a distinct teres minor muscle in T. syrichta may be a phylogenetically valuable character in the placement of Tarsius, possibly strengthening the case for its inclusion with monkeys and apes with the Haplorrhini.

As a concluding remark to this section, it must be remembered that almost all studies on primate rotator cuff function and gait have been done on anthropoids, not on prosimians (e.g., Tuttle and Basmajian, 1978; Larson and Stern, 1987, 1989, 1992; Whitehead and Larson, 1994). It is possible that the biomechanical inferences afforded by these previous studies do not apply to the results of this study. Future studies that concentrate specifically on prosimians are necessary to definitively interpret the present results on the rotator cuff muscle group.

Joint Structure

This study provides the first comparative data on the soft-tissue structures of the glenohumeral, acromioclavicular, and coracoclavicular joints in primates. Whereas previous studies have demonstrated that osseous morphology of the scapula, clavicle, and proximal part of the humerus is adaptive to locomotory and postural styles; this study found mixed results in the adaptive soft-tissue morphology of the shoulder joint complex.

The capsules of the acromioclavicular joint were without apparent variation among species, and the structure of this joint was very similar to that seen in humans (e.g., Standring, 2004). Differing locomotor and postural behaviors (arboreal quadrupedalism, suspension, and leaping) and phylogenetic position did not appear to be associated with variation in this joint.

The coracoclavicular joint was noted only in G. moholi and E. macaco in this study. In these species the coracoclavicular ligament was observed as a single band of dense connective tissue, unlike the condition in humans where it consists of two separate bands, the trapezoid and conoid ligaments (Standring, 2004). G. moholi and E. macaco practice very different locomotor and postural behaviors and are phylogenetically divergent. The apparent lack of this ligament in C. medius and T. syrichta in this study cannot be explained at this time, and we have no reason to believe that our dissection methods failed to locate an existing structure. The coracoclavicular ligament in humans prevents excessive rotation of the clavicle away from the scapula (Standring, 2004), but its function in nonhuman primates is uncertain. If it acts similarly in nonhuman primates, its apparently unique presence in one leaper and one quadruped to the exclusion of the other leaper and quadruped is mysterious. An increased sample size may help to answer this question.

The glenohumeral joint capsule did vary both among species and among locomotor/postural behaviors. Dorsally, the capsule was similar among species and was reinforced by tendons of the supraspinatus and infraspinatus muscles (and of the teres minor muscle in T. syrichta). Ventrally, it was reinforced by the tendon of the subscapularis muscle. Additionally, glenohumeral ligaments were located in the two quadrupedal species, C. medius and E. macaco, and may act as joint stabilizers. These ligaments were not found in the clinging and leaping species, G. moholi and T. syrichta. Leaping in these species involves landing with the hindlimb first and using the forelimbs primarily in “clinging” and procuring insects for food. Glenohumeral joint stabilization on the ventral surface may not be as important in vertical clinging and leaping activities as in quadrupedal activities where there is a great amount of shearing force on the joint capsule (Tuttle and Basmajian, 1978; Larson and Stern, 1987; Larson, 1993).

The course and attachment of the tendon from the long head of the biceps brachii muscle did not vary among species and appeared to be similar to human morphology. As in humans, it may be adapted for stabilizing the position of the humeral head in overhead activities (Standring, 2004).

Lastly, the glenoid labrum appeared to vary among the species from this study. In E. macaco the labrum appeared to be more “cup-like,” creating a deeper articulating site for the humeral head than in G. moholi, T. syrichta, and C. medius (Figs. 6,7). Such a deeper articulating surface may increase the stability of the glenohumeral joint during activities that would distract the humeral head from the glenoid fossa such as suspensory behavior (e.g., Roberts, 1974).

CONCLUSIONS

Overall, this study found that individual muscle mass of the rotator cuff muscle components in the prosimian species used are generally reflective of locomotor and postural behavior. However, the nonosseous components of the shoulder joint complex present a mixed bag of both adaptive and nonadaptive morphologies. This may be due to the limited sample size in this study, but it may also indicate that soft-tissue joint structures are highly conserved throughout phylogeny and are not subjected to evolutionary selection in the same manner as osseous characters. Future studies using expanded primate taxa with larger sample sizes should be carried out to further explore the question of adaptive morphology of the primate shoulder joint complex.

Acknowledgements

The authors wish to thank Jason Organ for his kind invitation to submit this article. The authors thank the two reviewers for their comments that greatly enhanced the quality of this manuscript. The authors also wish to thank Leanne Nash and Tim Smith for useful discussion on drafts of this manuscript. This is Duke Lemur Center publication # 1166.

Ancillary