The Robust Australopithecine Face: A Morphogenetic Perspective

See allHide authors and affiliations

Science  09 Apr 1999:
Vol. 284, Issue 5412, pp. 301-305
DOI: 10.1126/science.284.5412.301


The robust australopithecines were a side branch of human evolution. They share a number of unique craniodental features that suggest their monophyletic origin. However, virtually all of these traits appear to reflect a singular pattern of nasomaxillary modeling derived from their unusual dental proportions. Therefore, recent cladistic analyses have not resolved the phylogenetic history of these early hominids. Efforts to increase cladistic resolution by defining traits at greater levels of anatomical detail have instead introduced substantial phyletic error.

Robust australopithecines are conventionally sorted into three species: a single species from South Africa, Australopithecus robustus [1.8 to 1.5 million years ago (Ma)] (1), and in East Africa, A. aethiopicus (2.7 to 2.3 Ma) and A. boisei(2.3 to 1.3 Ma) (2). All are characterized by extreme postcanine megadontia, premolars with molarized roots, lower molars with accessory cuspules, and thick molar enamel. All also have sagittal and compound temporal/nuchal extracranial crests, a zygomatic arch positioned high above the occlusal plane, a forward placement of the zygoma, and a robust mandible with an absolutely and relatively tall ramus and a correspondingly tall posterior face (Fig. 1). Robust australopithecines also display markedly small incisors and canines, a thickened hard palate (that part of the upper jaw formed by the palatine process of the maxilla and the horizontal plate of the palatine), a vertically tall infraorbital region, low infraorbital foramina, a face hafted high onto the neurocranium, a frontal region depressed behind the supraorbital torus and between anteriorly converging temporal lines [the frontal trigone (3)], and strong postorbital constriction.

Figure 1

Characteristic features of the robust australopithecine face. 1, Vertically elongated infraorbital region; 2, low position of the infraorbital foramina; 3, high hafting of face onto neurocranium; 4, frontal trigone; 5, vertically tall mandibular ramus and posterior face; 6, sagittal crest; and 7, anteriorly placed zygomae (“dished” face).

Many of these features are not unique to the robust australopithecines. For example, A. africanus shares large second and third molar crowns, a vertically tall mandibular ramus, and, as compared to the more primitive condition displayed by A. afarensis, a somewhat more forward placement of the zygoma (4). A fragmentary Homo rudolfensissample from East Africa also shows postcanine megadontia (though not as extreme as in robust australopithecines), molarized premolar roots, and thick enamel (5, 6). This repetitious pattern of postcanine megadontia in early hominids has frustrated attempts to resolve their phylogenetic history. Recently, impressive lists of craniodental features shared by robust australopithecines have been cited as overwhelming evidence in support of their monophyletic origin (6, 7). Unfortunately, simply defining a character does not constitute evidence that it is independent and not the incidental effect of another. Which of the robust australopithecine features are truly independent? Resolution of these kinds of issues requires consideration of (i) the underlying genetic program that is ultimately responsible for their morphology and (ii) the process by which that program is interpreted and expressed.

Recent findings from developmental biology have demonstrated that homologous developmental pathways occur in a variety of embryonic processes and organisms, but within discrete units or modules (8, 9). Such modules display a number of distinct properties (9). They (i) have an autonomous genetically isolated organization, (ii) contain hierarchical units and may be parts of others, (iii) occupy specific physical locations within the developing system, (iv) exhibit varying degrees of connectivity to other modules, and (v) undergo sequential transformations during individual development. Modules span a hierarchy from molecules to organ primordia and body segments, in such a way that a single complex anatomical structure (for example, the tetrapod forelimb or the vertebrate skull) can be viewed as the independent product of a hierarchical expression of individual embryonic units. Changes either in the attributes of individual modules (state, number, or location) or in the timing of their interactions with one another alter the way organisms develop, and result in modified sets of features being presented to the filtering action of natural selection.

The skull arises and matures as part of an integrated complex of relatively independent functional modules (10,11). Each consists of all of the tissues, organs, glands, spaces, and supportive structures necessary to carry out a single function (such as mastication, respiration, or olfaction). Like genetically defined embryonic modules, the functional modules of the skull are readily definable, are relatively autonomous [in terms of both development and evolution (11)], and interact with one another within a developmental cascade. Much of the adult craniofacial form can therefore be understood with reference to the independent function (but integrated development) of (i) the brain and its associated sensory capsules, (ii) the nasal airway, and (iii) the oral apparatus (11, 12).

The integration of functional modules is achieved through growth remodeling and displacement. Growth remodeling is bone deposition and resorption on the endosteal and periosteal surfaces of bony laminae by the osteoblasts and osteoclasts that reside in their surface membranes (13). Such active growth is regulated by the endosteum and periosteum, whose activities are themselves mediated by signals (mechanical, electrical, hormonal, chemical, and so on) received from adjacent tissues. As an example, the hard palate is the structural interface between the nasal and oral cavities, and it therefore consists of functionally independent nasal and oral surfaces (Fig. 2A) (14). In response to signals received from the growing tissues of the nasal cavity, the nasal surface (lamina) of the hard palate drifts downward through a combined process of resorption on its outer surface and deposition on its inner surface. Similarly, the oral surface (lamina) of the hard palate drifts inferiorly (through the combined processes of periosteal deposition and endosteal resorption) in response to signals from the growing and functioning oral tissues (Fig. 2A). Ancillary to this active downward relocation of the two palatal surfaces is their passive downward and forward displacement in association with sutural growth. With expansion of the nasal tissues, the bony vomer is displaced away from its articulations with the perpendicular plate of the ethmoid superiorly and the hard palate inferiorly (Fig. 2A). In response to the tension within the sutures created by this displacement (13), bone is deposited at the vomeroethmoidal and vomeromaxillary sutures, and the hard palate is relocated downward and anterior to the cranial base. Additional sutural growth associated with expansion of the neuro-orbital and oral cavities displaces the hard palate even farther downward and forward.

Figure 2

Principles of facial growth. (A) Displacement of the vomer away from its contacts with the perpendicular plate of the ethmoid and hard palate. Sutural deposition in the “spaces” created by this displacement permanently relocates the hard palate and upper jaw inferiorly. The detail of the upper jaw illustrates the independent remodeling and inferior drift of the nasal and oral laminae of the hard palate (maxilla plus palatine) and premaxilla. Plus signs indicate deposition; minus signs indicate resorption. (B) Vertical growth of the mandibular condyle displaces the posterior maxilla and hard palate inferiorly, thus eliciting relatively greater bone deposition at the sphenopalatine and posterior ethmomaxillary sutures. The detail illustrates remodeling activities and cortical drift associated with maxillary rotation. Plus signs indicate deposition; minus signs indicate resorption; downward-pointing arrows indicate drift. Symbol size reflects the relative extent and degree of remodeling activity (remodeling and drift). Black areas are air sinuses.

The distinctive cranial morphotype of robust australopithecines must ultimately be interpretable in the context of growth remodeling and displacement. With respect to their distinctive palatal morphology, the traditional view holds that a thickened palate developed in response to some mechanical demand, and in particular in response to the need to reinforce the midpalatal suture against elevated masticatory stress (15). However, experimental studies demonstrate that an increase in sutural area in response to elevated extrinsic force is achieved through increased interdigitation rather than the accumulation of thick cortical bone (16). Thus, a thickened palate must be a function of some other constraint on rostral form. The unusual palatal morphology characteristic of the robust taxa may instead be a product of extreme maxillary rotation during ontogeny (17). Maxillary rotation is a normal aspect of anthropoid facial ontogeny (18). It results from differential growth of the sutures that attach the midface to the basicranium [the sphenopalatine and ethmomaxillary sutures (Fig. 2B)]. Relatively greater sutural deposition posteriorly typically occurs in order to maintain functional occlusion as the mandibular condyle expands vertically (13, 18). To keep pace with the skeletal changes taking place posteriorly, the anterior hard palate (both laminae) and premaxilla (forming the anteriormost nasal cavity floor and holding the incisors) undergo relatively greater remodeling, and the entire anterior midface and dentition drift downward (Fig. 2B).

In the robust australopithecines, an unusually tall mandibular ramus would have been associated with an extreme degree of maxillary rotation during ontogeny (19). Inferior drift of the oral surface of the hard palate would ultimately be determined by the height of the mandibular ramus and the extent of this rotation. Drift of the hard palate's nasal surface, however, would be regulated primarily by the size of the nasal airway, itself a parameter that is largely determined by body size (12). Because mandibular ramus height in the robust australopithecines far exceeds that of other hominoids of similar body size (20), a general thickening of the hard palate (a gradual separation of its nasal and oral surfaces) would occur. Indeed, this is precisely what occurs in modern chimpanzees, in which an increase in the height of the posterior midface is accompanied by a dramatic thickening of the hard palate and its concurrent invasion by the maxillary air sinus (Fig. 3) (21). The fact that palatal morphogenesis in the robust australopithecines was similar to that of the chimpanzee is indicated by the presence of a palatal component of the maxillary sinus, a “recessus palatinus,” in several robust australopithecine crania (3, 17). In addition, palatal thickness values of the fossil crania are those that would be expected in chimpanzees with posterior midfaces “grown up” to robust australopithecine size (Fig. 3). Therefore, the thickened hard palate of the robust australopithecines would appear to be a simple by-product of a vertically expanded mandibular ramus.

Figure 3

Palatal thickness compared to posterior facial height in an ontogenetic series of chimpanzees and a number of early hominid crania. LN, natural log. Chimpanzee data points represent sex-specific means (n = 68) of developmental age groups defined by stage of dental eruption and suture closure (deciduous → M1 erupted → M2 erupted → M3erupted → patent spheno-occipital synchondrosis → fused spheno-occipital synchondrosis). Posterior facial height is measured as the vertical distance separating the articular eminence from the occlusal plane. Number and letter designations beside data points indicate museum accession numbers of fossil crania.

Australopithecus africanus also displays a relatively tall mandibular ramus suggestive of extreme maxillary rotation (20), and although some specimens, including the famous Taung child, possess a recessus palatinus (22), adult palatal thickness in this taxon rarely approaches that typical of the robust taxa (Fig. 3). The A. africanus morphology therefore suggests that some additional factors may have promoted palatal thickening in robust crania or deterred extreme palatal thickening in nonrobust crania such as A. africanus (or both).

Growth remodeling of the nasal surface of the hard palate is regulated almost exclusively by the spatial demands of the nasal cavity (13). In comparison, the nasal surface of the premaxilla (Fig. 2) must satisfy not only the requirements of the nasal cavity but also those of the permanent incisors developing within it. Consequently, the resorptive capacity of the premaxilla is more restricted than that of the hard palate. In A. africanus, the vomer only contacts the nasal cavity floor at the hard palate (Fig. 4A) (23). There it projects below the premaxilla and into the incisive canal (the communicating passage between the nasal and oral cavities). This isolation of the vomer from the premaxilla provides some developmental (remodeling) independence of the two components of the nasal cavity floor. As a result, resorption of the hard palate's nasal surface could continue even after the resorptive limits of the premaxilla had been reached. Unlike the arrangement in A. africanus, the vomer of the robust australopithecines extends onto the nasal surface of the premaxilla (Fig. 4B) (23). This configuration constrains the nasal surface of the hard palate to maintain the same transverse level as the adjacent premaxilla throughout growth (21). As a consequence, all resorption of the anterior nasal cavity floor would cease once the allowable limits of resorption of the premaxilla were reached. With inferior drift of its nasal surface thus limited, continued inferior drift of the hard palate's oral surface (in response to extreme maxillary rotation) would result in the consistent development of a greatly thickened hard palate (Fig. 4B).

Figure 4

Proposed variation in nasomaxillary modeling and craniofacial morphogenesis in australopithecine taxa. (A) Isolation of the vomer from the premaxilla in A. africanus allows continued resorption of the palatal component of the nasal cavity floor after the resorptive limits of the premaxilla have been reached. (B) Extension of the vomer onto the premaxilla reduces the resorptive capacity of the palatal component of the nasal cavity floor. With inferior drift of the nasal floor thus constrained, orbitonasal and oral cavities become displaced in opposite directions (large arrows at right). Plus signs indicate deposition; minus signs indicate resorption; downward-pointing arrows indicate drift. Symbol size reflects the relative extent and degree of activity (remodeling and drift). Black areas are air sinuses.

The attachment of the vomer along the nasal cavity floor, through its influence on the pattern of nasal floor remodeling, may therefore be the pivotal factor responsible for the divergent palatal morphologies characteristic of australopithecine taxa (17). However, is the vomer's influence on australopithecine cranial morphology confined to the subnasal region? Probably not. If the model of australopithecine palatal development described above is indeed correct (24), then substantial downward remodeling of the nasal cavity floor in the robust specimens could not have occurred (Fig. 4B). Instead, continued expansion of the nasal cavity would set up a competition between the nasal and oral cavities for space within the midfacial skeleton. When such competition exists, an additional displacement must take place until the positions of the competing elements become sufficiently modified (13). In the robust australopithecine cranium, the point of contact of the two expanding elements would have been the nasal cavity floor (Fig. 4B). From this interface, the oral cavity would have been displaced inferiorly and the nasal cavity superiorly. Because the circumorbital elements are intimately associated with the perpendicular plate of the ethmoid, superior displacement of the nasal cavity would secondarily displace the entire circumorbital region superiorly.

With a displacement of the oral and orbito-nasal skeletons in opposite directions, any structures that span their connecting interface, such as the infraorbital region, must be vertically elongated. Also expected with a relatively greater superior displacement of the upper facial skeleton would be a higher hafting of the facial skeleton onto the neurocranium and a corresponding low frontal region. Strong postorbital constriction and a low position of the infraorbital foramina within the face would also be expected with such displacement of the upper face relative to the neural capsule. A suite of features can therefore be identified as the expected morphological correlates of a pattern of facial ontogeny in which the oral and upper facial capsules are displaced relative to a more stationary nasal cavity floor.

Table 1 lists the shared derived features of robust australopithecine crania as identified in a recent cladistic analysis (7). Most of these 20 traits are simply the most divisible elements of the more comprehensive masticatory features of (i) postcanine megadontia, (ii) a small anterior dentition, and, (iii) a large masticatory musculature. However, a number are not readily interpretable with respect to the masticatory apparatus. But, as has been demonstrated here, all are interpretable with respect to just two features of the robust australopithecine cranium: (i) a tall mandibular ramus and (ii) a vomeral insertion on the nasal surface of the premaxilla. The first is known to be functionally integrated with the postcanine occlusal area (25). The underlying basis of the second feature has yet to be explored, but it is critical to note that a similar vomeral insertion is found in modern humans (17,21), and we share with the robust australopithecines a relatively small anterior dentition. It therefore seems reasonable to suggest that the vomeral morphologies of both groups reflect a small anterior dentition. If so, and if the model of robust australopithecine craniofacial morphogenesis outlined above is correct, then all of the skeletal traits identified as synapomorphies of aParanthropus clade are merely developmental by-products of dental size and proportions.

Table 1

Synapomorphies of Paranthropus(7).

View this table:

Although all robust australopithecines combine extremely large postcanine teeth with small anterior dentitions, they do not share identical tooth morphologies. Rather, A. boisei exhibits a number of non–size-related features of its postcanine dentition (for example, distinctive morphology of the lower fourth premolar and distinct lower molar cusp proportions) that are not observed inA. robustus (2). In addition, the postcanine teeth of A. robustus are notably smaller than those of both of the geologically older East African taxa. Although it is possible that a reduction in postcanine tooth size occurred during the evolution of A. robustus (2), it is no less probable that the East and South African forms had separate phyletic origins. Therefore, despite their fundamentally similar cranial morphologies, the phylogenetic history of the robust australopithecines remains unresolved.


View Abstract

Navigate This Article