Functional and Phylogenetic Considerations
The muscles of the forelimb offer several insights into the functional anatomy and phylogenetic relationships of hippos relative to other artiodactyls. Among the extrinsic muscles, m. brachiocephalicus is noteworthy. Choeropsis demonstrates typical artiodactyl origins and insertions for this muscle; however, a clear intersectio clavicularis is absent. Published accounts of common hippos are ambiguous regarding this trait. Interestingly, among artiodactyls, the intersectio clavicularis is also absent in Hyemoschus and Tragulus. Thus, hippos are similar to tragulids in this regard.
The pectoral muscles of hippos also differ from those of other artiodactyls. While published descriptions of common hippos are unclear on the matter, illustrations indicate the pectoral muscles are similarly arranged in the two species. Hippos resemble Babyrousa, ruminants, and camels in lacking a clear subdivision of the superficial pectoral. In contrast, m. pectoralis superficialis is clearly divided into two parts (mm. pectoralis descendens and pectoralis transversus) in Sus and Pecari. Most importantly, hippos are distinct among the artiodactyls in having a more extensive origin for m. pectoralis superficialis; in other artiodactyls, the muscle is restricted to the first three to six costal cartilages, while in hippos, m. pectoralis superficialis extends up to 8 cm caudal to the xiphisternum. M. pectoralis profundus is also characterized by a broad origin in hippos compared with other artiodactyls, extending up to 35 cm caudal to the xiphoid process. Both pectoral muscles are crucial parts of the suspensory apparatus of the trunk. The pectoral muscles support the trunk on the limbs, retract the unsupported limb, and advance the trunk over the fixed limb. They are also capable of moving the trunk laterally over a fixed limb. The well-developed pectoral muscles in hippos reflect a capacity for powerful retraction of the forelimb; this may be an adaptation for charging (a staple of intraspecific interaction), climbing the steep inclines of muddy river and lake banks, and propelling the body forward against the resistance of water during underwater locomotion.
The four muscles that stabilize the hippo shoulder joint (mm. infraspinatus, supraspinatus, subscapularis, and teres minor) are similar to those of most artiodactyls, although there are a few points of difference. For example, in pygmy and common hippos, m. infraspinatus inserts by means of a single tendon onto the caudal aspect of the greater tubercle of the humerus. Babyrousa and Camelus resemble the hippos; however, in Sus and ruminants, the tendon divides into a deeper, fleshy insertion and a superficial, tendinous insertion. The deeper limb inserts on the medial surface of the humerus in pigs, and on the caudal aspect of the greater tubercle in ruminants, whereas the more superficial inserts on a depression lateral to the caudal aspect of the greater tubercle. M. subscapularis is partially subdivided in pygmy hippos and may have up to three divisions in the common hippo. Suids and camels lack subdivisions; however, ruminants may have three or four parts. Thus, pygmy hippos appear to be intermediate between suids and ruminants, whereas common hippos approximate the ruminant condition.
M. deltoideus is subdivided into a pars acromialis and a pars scapularis in both pygmy and common hippos. Due to the reduction of the acromion process in suids, m. deltoideus is undivided in these artiodactyls. However, ruminants and camels have distinct acromial and scapular parts, thus resembling hippos. M. articularis humeri is not described in accounts of common hippos, suggesting that it is absent in all hippos. The muscle is also absent in ruminants. In contrast, m. articularis humeri is constant in Babyrousa and variably present in Sus. It is also present in Camelus. Yet again, the hippos resemble ruminants more closely than other artiodactyls.
The muscles of the hippo brachium differ from those of most artiodactyls in only a few significant ways. M. coracobrachialis consists of a single belly in pygmy and common hippos. In contrast, two bellies are present in suids, ruminants, and camels. In addition, pygmy and common hippos have an undivided medial head of m. triceps brachii. In Sus, the medial head is divided into two parts by the conjoined tendon of mm. latissimus dorsi and teres major. This tendon also divides the medial head in Babyrousa, although the two parts fuse distally. In ruminants, an accessory medial belly is also present to a variable extent. In this case, hippos resemble Camelus, which also lacks an accessory triceps muscle. Lastly, in the pygmy hippo, m. anconeus is fused with the medial head of m. triceps brachii. However, it forms a distinct muscle in the common hippo. M. anconeus is also independent in Sus and Camelus. In Babyrousa, it is fused with both medial and lateral heads, and in ruminants it is fused with the lateral head.
The flexor muscles of the hippo antebrachium show some interesting contrasts to other artiodactyls. M. palmaris longus is present in pygmy hippos and variably present in common hippos. In the pygmy hippos dissected for this study, m. palmaris longus inserted onto the palmar aponeurosis, contributed fibers to the superficial flexor tendon to digit IV, and formed the entire superficialis tendon to digit V. However, Macalister (1873) and Campbell (1935, 1936) recorded an insertion only onto the palmar aponeurosis in their pygmy hippos. Although Gratiolet and Alix (1867) claimed it was absent in the common hippo, Macdonald et al. (1985) and Kajava (1923) observed m. palmaris longus in the larger species. Kajava (1923) describes m. palmaris longus inserting onto the palmar aponeurosis, with extensions to the long flexor tendon sheaths and the base of the proximal phalanx of digit II. M. palmaris longus is absent in all other artiodactyls and all perissodactyls; however, it is present in elephants and hyraxes. Thus, hippos are unique among ungulates in retaining this feature.
In both hippo species, m. flexor carpi ulnaris exhibits two separate bellies: ulnar and humeral. These bellies may give rise to a common tendon or independent tendons of insertion. The primary insertion is the os carpale accessorium; however, this may be accompanied by a secondary insertion onto the metacarpal or proximal phalanx of digit V. In Sus, the ulnar head can be absent, or, if present, much reduced. The ulnar head is also absent in Potamochoerus and Babyrousa, while it is much reduced in Pecari and Camelus. In ruminants, there are two heads of origin, but the ulnar head is weak and largely tendinous. Therefore, m. flexor carpi ulnaris is more developed in hippos versus suids and tayassuids, and to a lesser extent, ruminants. Perissodactyls also vary in the degree of development of the ulnar head; it is reduced in Tapirus, variably absent in Equus, and present in Dicerorhinus. Both heads are present and well developed in hyraxes and elephants. Thus, compared with other artiodactyls, hippos demonstrate a more primitive condition in retaining the ulnar head of m. flexor carpi ulnaris.
The digital flexors of hippos also show several interesting features. M. flexor digitorum superficialis usually consists of a single muscle belly in pygmy and common hippos. However, in suids, tayassuids, and ruminants, the muscle has superficial and deep parts. M. flexor digitorum superficialis is very derived in Camelus, lacking a muscle belly altogether; this tendinous structure originates from the os carpale accessorium and metacarpal bone. In hippos, the muscles contributing to the long flexor tendons differ from those in other artiodactyls. M. flexor digitorum superficialis does not provide a tendon to digit V in Choeropsis; instead, the superficial tendon to this digit is provided by m. palmaris longus (this study) or m. flexor digitorum brevis V (Campbell, 1935, 1936). A digit V contribution from m. flexor digitorum superficialis is usually present in the common hippo; however, when absent, this tendon is provided by m. flexor digitorum brevis V (Macalister et al., 1985). In addition, m. flexor digitorum profundus supplies tendons to all four digits in hippos. Thus, a pair of long flexor tendons serves each of the digits in the Hippopotamidae. In suids and tayassuids, there are strong superficial and deep tendons to digits III and IV, and weak deep tendons to digits II and V only. However, deep tendons do serve all four digits in Potamochoerus. In camels and most ruminants, there are strong superficial and deep tendons to digits III and IV only; however, Ovibos, Tragulus, and Hyemoschus have deep tendons to all four digits. Thus, the flexor tendons of hippos resemble those of Potamochoerus, Ovibos, and tragulids. In most artiodactyls, digits II and V are not typically weight-bearing, as they are in hippos (Fig. 8). In addition, the presence of long flexor tendons to all four digits in hippos would assist in adducting the toes to help prevent splaying, especially on soft terrain. The condition in Tapirus and Dicerorhinus parallels that of hippos, with superficial and deep tendons serving digits II–IV. In contrast, single superficial and deep tendons serve the central digit in Equus. Interestingly, the mm. interflexorii are absent in pygmy and common hippos. Mm. interflexorii are present in suids, tayassuids, and ruminants; however, they are absent in camels. Mm. interflexorii are also absent in perissodactyls. In short, the digital flexor apparatus of hippos is relatively primitive compared with that of other artiodactyls.
Figure 8. A–C: Lateral view of forelimb in the (A) ox, (B) common hippo, and (C) pig. Whereas hippos have four weight-bearing digits, the pig and ruminant have reduced lateral digits.
Download figure to PowerPoint
In all artiodactyls, pronators and supinators are reduced or absent. In rare instances, these muscles are variably expressed, as is the case for m. pronator teres in pygmy and common hippos. The muscle was present in this and Campbell's (1935, 1936) study; however, it was absent in Macalister's (1873). Similarly, m. pronator teres was present in Kajava's (1923) common hippo, but absent in those dissected by Cuvier and Laurillard (1849) and Gratiolet and Alix (1867). M. pronator teres is also variably developed in Sus and Potamochoerus; when present, it inserts on the middle to distal thirds of the radius. However, in Babyrousa, m. pronator teres is reduced to a completely tendinous state and inserts more proximally on the radius. In tayassuids, it is a mixture of fleshy and tendinous fibers, and inserts on the midshaft of the radius. With the exception of the tragulids, m. pronator teres is poorly developed in ruminants, and inserts on the proximal third of the radius. The muscle is absent altogether in Camelus. M. pronator teres is variably developed in the perissodactyls, being fibrous or absent in equids, weakly developed or absent in tapirs, and absent in rhinos. M. supinator and m. pronator quadratus are absent in pygmy and common hippos, although Macalister (1873, p. 498) claimed a “barely detectible” m. pronator quadratus was present in his pygmy hippo. Apart from this report, the muscle is absent in all artiodactyls, and there is only a single account of its presence among perissodactyls, in the tapir (Murie, cited in Windle and Parsons, 1901). M. supinator is occasionally present in reduced form in Sus. When present, it is a thin slip of muscle that blends with the radial head of m. flexor digitorum profundus. However, m. supinator is absent in other artiodactyls, and in perissodactyls. Interestingly, m. brachioradialis is absent in the pygmy hippo but variably present in the common hippo. M. brachioradialis is absent in all other artiodactyls and among perissodactyls, it is present in the tapir and rhino. Therefore, the common hippo once again retains a primitive feature that has been lost by other artiodactyls.
As noted above, the antebrachial extensor muscles show considerable variation, both within species and across the Artiodactyla. M. extensor carpi ulnaris is similarly disposed in both species of hippo, although it may be entirely tendinous with a secondary insertion onto the base of the fifth metacarpal in the common hippo. In suids, the muscle is composed of a superficial tendinous and a deep fleshy part; the superficial tendinous part is homologous to the muscle in hippos, but the fleshy belly inserts on the fifth metacarpal, or more rarely, the fourth, and is absent in the hippo (although common hippos can have a similar secondary insertion). In Pecari, the muscle is similar to suids, although the fleshy belly can insert on the os carpale IV and fourth and fifth metacarpals in some individuals. In ruminants, m. extensor carpi ulnaris also inserts more distally, on the fourth or the rudimentary fifth metacarpal, and in Camelus, the insertion reaches the fused metacarpal bone. Therefore, compared with other artiodactyls, m. extensor carpi ulnaris is reduced in the pygmy hippo, losing its connection to the fifth metacarpal. A more distal insertion indicates the muscle acts on the carpometacarpal joint as well as the carpal joint, whereas in the pygmy hippo, it only acts on the latter. This arrangement may allow the carpometacarpal joint to be independent of the carpal joint, allowing for more controlled movement of the digits in the pygmy hippo.
M. extensor digitorum lateralis is incompletely subdivided in the pygmy hippo. Published accounts of the common hippo do not comment on the relative degree of subdivision. In suids and Pecari, the muscle is composed of two distinct bellies, while in ruminants and camels, the muscle is undivided. M. extensor digitorum lateralis sends tendons to digits IV and V in the pygmy hippo, but tendon insertions vary in the common hippo, with fourth and fifth digit tendons, or a single fourth digit tendon. In suids and tayassuids, insertion is onto the fourth and fifth digits; however, a single fifth digit tendon may be present in Sus. In ruminants, it inserts on the fourth digit (giraffids, cervids, most bovids), fourth and third digits (Bos), or fourth and fifth digits (tragulids). Finally, in Camelus, the tendon serves digit IV. Thus, the hippo m. extensor digitorum lateralis most closely resembles that of the tragulids.
M. abductor digiti I longus is similar in its morphology in the common and pygmy hippo, although it may take sole origin from the ulna in the larger species. Campbell (1935, 1936) and Macalister (1873) claim that m. abductor digiti I longus inserts on the vestige of the first metacarpal in the pygmy hippo, but this study revealed an insertion on os carpale II. Humphry (1872, p. 373) also lists the “rudimentary trapezium” as the site of insertion in the common hippo. M. abductor digiti I longus inserts further distally in other artiodactyls, terminating on the base of the second metacarpal in suids, tayassuids, and tragulids, and on the fused third and fourth metacarpals in camels and most ruminants. The proximal insertion of m. abductor digit I longus in hippos mirrors that seen for m. extensor carpi ulnaris. Both muscles flex and abduct the carpal joint; in the case of m. abductor digiti I longus, this function is quite unusual, as it is a carpal extensor in other artiodactyls. As noted above, decoupling the carpal and carpometatacarpal joints may allow for more controlled digital movement in hippos.
M. extensor digiti II is absent in hippos and Camelus, but a rudimentary version of this muscle is present in Sus, where it arises from the ulna, deep to m. extensor digitorum communis, and gives rise to a tendon that fuses with that of the common extensor or inserts on or near the second digit. In Babyrousa, the muscle is quite small and originates from the lateral aspect of the ulna, proximal to the origin of m. abductor digiti I longus and deep to the common extensor. It gives rise to two tendons, one joining a tendon from m. extensor digitorum communis to digit II, and the other inserting onto the fascia covering the third metacarpal. In ruminants, m. extensor digiti II helps form the lateral belly of m. extensor digitorum communis, sending tendons to digits III and IV.
M. flexor digitorum brevis is present in pygmy hippos, although it appears to vary in form. Campbell (1935, 1936) describes this muscle as a small, flat muscle belly arising from the tendon of m. palmaris longus, forming the perforated tendon to the fifth toe, and inserting on the proximal phalanx. He also observed tendinous fibers originating from the os carpi accessorium and joining the m. flexor digitorum superficialis tendon to digit IV as well as the m. flexor digitorum brevis tendon to digit V, and includes these tendinous fibers in the m. flexor digitorum brevis complex. In the two specimens dissected for this study, m. flexor digitorum brevis was represented by muscle fibers on the deep aspect of the m. palmaris longus tendon; a brevis tendon was not present. Elements of m. flexor digitorum brevis are also present in common hippos. Kajava (1923) describes three accessory muscles arising from the m. flexor digitorum superficialis tendons to digits II–IV, as well as m. flexor digitorum brevis digiti V arising from the tendon of m. palmaris longus. The description by Macdonald et al. (1985) is less complete. According to these authors, the muscle originates on “the carpal bone” and inserts on the fifth toe. M. flexor digitorum brevis is absent in other artiodactyls, and among the perissodactyls, it is variably developed in Tapirus. M. flexor digitorum brevis is constant in the hyrax and elephant. This muscle is thus only present in those taxa that retain the more primitive arrangement of lateral weight-bearing digits, but even among these ungulates it is reduced.
The intrinsic muscles of the manus are much more developed in hippos than in other artiodactyls, although the patterns of reduction in different groups are complex. M. lumbricalis IV is present in pygmy and common hippos. In common hippos, the lumbrical may arise from the m. flexor digitorum profundus tendon to digit IV, or before the division of the profundus tendon. Interestingly, Kajava (1923) describes a single lumbrical with insertions on the fourth and fifth digits in his common hippo. For Sus, many authors describe m. lumbricalis II arising from either the common tendon of m. flexor digitorum profundus or the profundus tendon to digit II. However, according to Nickel et al. (1986) and Sisson (1975a), the mm. lumbricales are represented by a few muscle fibers running between the deep flexor tendon and the deep belly of the superficial flexor. Kneepkens et al. (1989) argue that m. lumbricalis II is present in Sus, but is mislabeled “m. flexor digiti II (brevis)” (e.g., sensu Nickel et al., 1986). M. lumbricalis II is also present in Potamochoerus, but mm. lumbricales are absent in Babyrousa and Pecari. In addition, they are absent in camels and most ruminants, although Getty (1975a) describes muscular bundles in domestic ruminants that may be regarded as mm. lumbricales; these arise from m. flexor digitorum profundus and insert onto the digital tendon at the carpus. Although most ruminants lack these muscles, Tragulus may have a lumbrical to the third and fourth digits, or solely to the fourth, while Hyemoschus has m. lumbricalis II. Thus, once again, hippos most closely resemble tragulids, particularly Tragulus. Among perissodactyls, Tapirus has three mm. lumbricales (one to each digit), while Equus has two, each arising from the single profundus tendon. Mm. lumbricales are not present in Dicerorhinus.
M. abductor digiti V is similarly arranged in the pygmy and common hippo; however, Kajava (1923) records an insertion onto the fibrous digital sheath of digit V, rather than a bony insertion in his common hippo. In Sus and Potamochoerus, m. abductor digiti V is similar to that seen in the pygmy hippo. However, Nickel et al. (1986) describe an abductor of the fifth digit in the domestic pig that originates from the palmar aspect of the fifth metacarpal and inserts on the lateral aspect of the metacarpophalangeal joint and proximal phalanx; this muscle could represent m. abductor interosseus digiti V. In Babyrousa, m. abductor digiti V has superficial and intermediate heads. These muscles represent the m. abductor digiti V and the abductor interosseus digiti V described here for the pygmy hippo. M. abductor digiti V is present in tayassuids and resembles the condition present in the pygmy hippo, although Campbell (1935, 1936) did not observe a fleshy representative in his collared peccary. M. abductor digiti V is absent in camels and in all of the ruminants except the tragulids. Among the perissodactyls, m. abductor digiti V is present in Tapirus and Dicerorhinus.
Mm. adductores digiti II et V are present in pygmy and common hippos; however, Kajava (1923) indicates a more distal origin off the metacarpals in his common hippo. Adductors for the lateral digits are present in suids and Pecari. According to some accounts, Sus demonstrates origins and insertions similar to those in the pygmy hippo (Kajava, 1923; Campbell, 1935, 1936). However, Nickel et al. (1986) describes the adductors as arising from the third and fourth metacarpals, as in Kajava's (1923) common hippo. These muscles are absent in Camelus and most ruminants; however, mm. adductores digiti II et IV are present in Tragulus. Among the perissodactyls, the tapir has three mm. adductores, to digits II, IV, and V.
Pygmy and common hippos have eight mm. interossei (two per digit). Campbell (1935, 1936) describes four mm. interossei in Sus, including one to each toe (radial side of digits II and III; ulnar side of digits IV and V); these represent the digit II–V abductores interossei. Kajava (1923) provides a similar account for Sus, although he also describes a digit II adductor interosseus. Getty (1975) and Nickel et al. (1986) describe two mm. interossei in Sus, for digits III and IV. However, according to these authors, the second and fifth digits each have a flexor, adductor, and abductor. Most likely, the adductors to digits II and V represent mm. adductores digiti II et V, rather than mm. interossei. In addition, as the flexor of the second digit arises from the deep flexor tendon, this represents m. lumbricalis II (sensu Kneepkens et al., 1989). In addition, the flexor of the fifth digit described by these authors represents a second muscle belly of m. abductor interosseus digiti V, as described by Campbell (1935, 1936). In light of these considerations, the abductors to digits II and V would represent true mm. interossei. According to these calculations, Sus lacks an adductor interosseus for the second (variable), third, fourth, and fifth digits. Whatever the precise homologies, Sus has a reduced number of intrinsic muscles compared with the hippo manus.
In Potamochoerus, the mm. adductores interossei are entirely tendinous, as is m. abductor interosseus digiti IV. Compared with Sus and Potamochoerus, Babyrousa demonstrates only a slight reduction, as it lacks m. adductor interosseus digiti II. In Pecari, m. adductor interosseus digiti II is absent; mm. adductores interossei digiti III et V are entirely tendinous, and mm. abductor and adductor interosseus digiti IV are reduced. The intrinsic musculature is further reduced in ruminants; there is only a single middle interosseous muscle. This muscle divides into five tendons of insertion, inserting on the sesamoid bones of digits III and IV as well as uniting with the tendons of the superficial flexor, contributing to the cuff around the deep flexor tendon. An exception to this rule, once again, are the tragulids, which have four interossei muscles, corresponding to each digit. In camels, a single, entirely tendinous interosseous element is present, arising from the distal carpus and proximal metacarpus, and inserting onto the proximal sesamoid bones. Thus, compared with other artiodactyls, hippos have more extensively developed intrinsic muscles of the manus, including a full complement of mm. interossei (two per digit).
Although hippos are semiaquatic, the musculature does not show extreme modifications as they have maintained typical terrestrial patterns of locomotion; hippos walk underwater rather than use lateral or dorsoventral movements of the trunk to swim. Nonetheless, hippos do display some unique characteristics related to their lifestyle. Because of their large size and the increased resistance of movement against water, the musculature shows several adaptations for increased power, including increased muscle size and fusions and reductions in number of muscle bellies. Especially noteworthy are the modifications of the pectoral muscles to propel the trunk forward and the arrangement of the long flexor tendons to serve each of the four weight-bearing digits, providing increased control of splay on soft ground. In addition, many of the intrinsic muscles of the manus are retained in hippos, including m. lumbricalis IV, mm. adductores digiti II et V, and a full complement of interossei, once again indicating that control of the digits is functionally important in these animals.
This study also indicates that, although the forelimb musculature of hippos shares some similarities with suids, hippos more closely resemble ruminants, especially tragulids. This finding supports the molecular data (e.g., Shimamura et al., 1997; Ursing and Arnason, 1998; Nikaido et al., 1999) (Fig. 9). Moreover, pygmy and common hippos are unique among artiodactyls in retaining several primitive features in the forelimb (e.g., mm. palmaris longus and flexor digitorum brevis). These findings are consistent with the hypothesis that hippos diverged from other Artiodactyla early in the history of this group. Additional studies of hindlimb, facial, and trunk muscles may help determine whether this unique family of artiodactyls was indeed closely allied to the Cetacea (see Cooper et al., 2007, this issue).
Figure 9. Cladogram for the Artiodactyla. This cladogram is supported by numerous molecular analyses, and is consistent with the findings in the current study (Nikaido et al., 1999).
Download figure to PowerPoint