Metopic suture of Taung (Australopithecus africanus) and its implications for hominin brain evolution

Abstract

The type specimen for Australopithecus africanus (Taung) includes a natural endocast that reproduces most of the external morphology of the right cerebral hemisphere and a fragment of fossilized face that articulates with the endocast. Despite the fact that Taung died between 3 and 4 y of age, the endocast reproduces a small triangular-shaped remnant of the anterior fontanelle, from which a clear metopic suture (MS) courses rostrally along the midline [Hrdlička A (1925) Am J Phys Anthropol 8:379–392]. Here we describe and interpret this feature of Taung in light of comparative fossil and actualistic data on the timing of MS closure. In great apes, the MS normally fuses shortly after birth, such that unfused MS similar to Taung’s are rare. In humans, however, MS fuses well after birth, and partially or unfused MS are frequent. In gracile fossil adult hominins that lived between ∼3.0 and 1.5 million y ago, MS are also relatively frequent, indicating that the modern human-like pattern of late MS fusion may have become adaptive during early hominin evolution. Selective pressures favoring delayed fusion might have resulted from three aspects of perinatal ontogeny: (i) the difficulty of giving birth to large-headed neonates through birth canals that were reconfigured for bipedalism (the “obstetric dilemma”), (ii) high early postnatal brain growth rates, and (iii) reorganization and expansion of the frontal neocortex. Overall, our data indicate that hominin brain evolution occurred within a complex network of fetopelvic constraints, which required modification of frontal neurocranial ossification patterns.

The Taung specimen consists of a natural endocast, a large fragment of fossilized face that articulates with the endocast, and a mandible that occludes with the maxillary dentition of the face (1). Taung died at ∼3.8 y of age (2), although an older age at death was suggested historically (1). Current estimates for Taung’s cranial capacity (uncorrected for age) of 382 cm³ (3) and 402–407 cm³ (4–6) are in fairly close agreement, especially compared with Dart’s earlier estimate of 520 cm³ (7). Acceptance of Australopithecus africanus as a legitimate early hominin took several decades, partly because Dart interpreted the endocast as representing an advanced humanlike pattern of convolutions despite its comparatively small volume (1, 8), which his colleagues found problematic (9). Despite the fact that Dart’s general interpretation of Taung is now accepted, debate continues about the morphology and implications of its endocast (10).

Adhering bony fragments and foramina on its basal surface that conduct nerves and vessels have been identified (3). The right temporal and both frontal poles that are embedded in the back of the facial fragment (11, 12) have been incorporated electronically into a virtual reconstruction of Taung’s entire endocast (3) (Fig. 1). Comparative measurements of the virtual endocast with other australopithecine endocasts reveal that Taung shares a number of shape features that align it more closely with other A. africanus specimens than with Paranthropus endocasts, including squared-off frontal lobes and the shape of the temporal poles (3, 13). This finding suggests that, with a brain size that falls within the range for great apes, but represents a higher degree of encephalization, A. africanus was neurologically reorganized compared with similarly sized Paranthropus brains (13, 14). It is not known, however, whether the brains of A. africanus had become reorganized in conjunction with a trend for brain enlargement, because their direct ancestors (and their brain sizes) remain unknown.

Fig. 1.

CT-based superior view of the Taung natural endocast (solid) and face (transparent). Note imprints of sagittal (S, interparietal), coronal (C), and metopic (M) sutures (arrowhead denotes endpoint), and fontanelle (F, area delimited by dashed line). (Scale bar, 30 mm.)

Here we describe a previously unanalyzed metopic suture (MS) on the Taung endocast, and compare it with the relevant frontal bone morphology of fetal to adult chimpanzees, bonobos, and Homo sapiens (SI Materials and Methods and Tables S1 and S2). Data are also provided regarding MS in 58 additional fossil hominins (SI Materials and Methods, MS in fossil hominins and Table S3), and we interpret the significance of an unfused or partially fused MS for hominin brain evolution.

Results

The natural endocast of Taung is well preserved and reproduces a small remnant of the anterior fontanelle, from which a clear imprint of a MS courses rostrally along the midline of the endocast (Fig. 1). In a 1925 description of Taung, Hrdlička noted “Posterior 3/5 of frontal (anti-bregmatic) brain surface partly covered by a remnant of bone, shows what appears to be patent metopic suture but no trace on bone (what there is of it) itself” (p. 390 in ref. 15). The MS imprint on the endocast terminates several millimeters above the preserved portion of the frontal squama. The squama does not exhibit traces of MS on its external and internal surfaces, as confirmed with medical CT imaging. The morphology of the MS and anterior fontanelle are consistent with observations from CT studies of humans, in which “fusion of the metopic suture commences at the nasion, proceeds superiorly in progressive fashion, and is completed at the anterior fontanelle in a manner analogous to a zipper closing” (p. 1213 in ref. 16). We thus classify Taung’s MS as partially fused. The sagittal (interparietal) and right and left coronal sutures are also clearly represented on the endocast and each intersects with the remnant of the anterior fontanelle.

In chimpanzees and bonobos, MS fusion typically occurs at a faster pace on the internal compared with the external table of the frontal squama, but in humans only minor differences between endocranial and exocranial MS fusion could be observed (Figs. S1 and S2). Endocranial and exocranial MS fusion states thus represent lower and upper ranges for the time of MS fusion. Accordingly, the partially fused MS on the Taung endocast indicates late fusion of MS in this 3- to 4-y-old individual.

In African great apes, the MS closes shortly after birth. In chimpanzees, it is completely fused in 65% of all cases before the eruption of the first deciduous molars (dm1) (Materials and Methods and Fig. 2A). In modern humans, the MS closes comparatively late, in 90% of all cases after the eruption of the dm1 (Fig. 2B). In concomitance with late MS closure, human infants and adults exhibit a high proportion of partially fused and unfused MS, but this condition is infrequent in the great apes (Fig. 2). These findings are in line with an earlier survey of “young” human, chimpanzee, and gorilla individuals [i.e., with any permanent teeth erupted, except the third molars (M3)], indicating that 33% of the human sample, but none of the great apes, retained MS (17). By adulthood (complete permanent dentition with appreciable wear), none of 172 chimpanzees or 185 gorillas had any trace of MS, whereas 16% of 2,104 humans had an unfused MS (17).

Fig. 2.

Prevalence of unfused (black), partially fused (gray), and completely fused (white) internal metopic sutures in ontogenetic series of P. troglodytes (A) and H. sapiens (B). Dental age classes are: pre (fetal preterm), neo (neonate to before dental eruption), dm1 (first deciduous molars erupted), dm2 (second deciduous molars erupted), M1/M2/M3: first, second, third permanent molars erupted. The Taung individual belongs to age class M1 (*). Note delayed MS fusion in humans compared with chimpanzees. Prevalence of unfused MS in adult humans varies between populations (the two M3 bars in B represent a global sample and a Swiss medieval sample).

Data gathered for state of MS fusion in fossil hominins (SI Materials and Methods, MS in fossil hominins and Table S3) reveal a clear trend in small-brained gracile hominins that lived in Africa between ∼3.0 and 1.5 million y ago (the term “gracile hominins” is used here to indicate fossils attributed to the genus Australopithecus and early Homo, as opposed to Paranthropus). Of five A. africanus specimens (Taung, Sts 5, Sts 71, Sts 60, and Sts No. 2 endocasts) and one early Homo cranium (SK 27) from South Africa, a partial MS is clearly present in two (Taung, SK 27), We have confirmed an earlier report (18) of partial MS in six East African specimens (KNM-ER 1805, KNM-ER 1813, KNM-ER 3733, KNM-ER 3883, KNM-WT 15000, and OH 24) that have been attributed to early Homo (e.g., Homo ergaster or Homo habilis). To our knowledge, unfused or partially fused MS have not been reported for Paranthropus aethiopicus, Paranthropus boisei, and Paranthropus robustus (for details see Table S3). Taken together, these data suggest that late MS fusion may have evolved in small-brained gracile hominins that lived between ∼3.0 and 1.5 million y ago. Comparatively high frequencies of partially fused or unfused MS in Homo erectus from Asia and in the Neanderthals indicate that the trend toward late MS fusion continued in mid-to-late Pleistocene hominins (Table S3).

Discussion

The MS normally becomes obliterated later in humans than in chimpanzees (Fig. 2). At the dental age of Taung (first permanent molar erupted), 35% of humans exhibit an unfused or partially fused MS; apes rarely retain a MS similar to Taung’s (17) (Fig. 2). Furthermore, unfused or partially fused MS in adults of Neolithic to contemporary human populations are relatively frequent, with a global average around 3–4% (19, 20). Early MS fusion likely represents a primitive feature of haplorhine primates, or even of euprimates (21), and is a feature uniting crown anthropoid primates (22). Late MS fusion in humans thus appears as a derived state, and early fusion, as observed in the great apes, appears as the primitive state.

We hypothesize that late MS fusion and an elevated proportion of partially fused or unfused MS in subadult and adult individuals result from one and the same developmental process. To test this hypothesis, we consider the proportion of fused MS as a function of age in chimpanzees and humans (Fig. 3A), and evaluate age-specific rates of MS fusion (Fig. 3B). Fig. 3B suggests that MS fusion occurs in a probabilistic manner, similar to a survival/death process (“survival” being similar to suture unfused; “death” being similar to suture fused). Peak values of the distributions in Fig. 3B indicate that MS fusion in chimpanzees is most likely to occur around birth, but fusion in humans is most likely to occur around the eruption of the second deciduous molar (dm2). Shifting the peak from birth to dm2 results in a longer tail of the probability distribution, and an elevated proportion of late-fusing MS or unfused MS in adult individuals (Fig. 3A). The presence of a still patent fontanelle and of a partially fused MS in the Taung child, and the incidence of unfused MS in five adult and two other younger Australopithecus/early Homo specimens (Table S3) is thus taken as evidence that a human-like pattern of late MS fusion was already present in mid-to-late Pliocene gracile hominins.

Fig. 3.

Metopic suture fusion, empirical and model data. (A) Percentage of fused sutures as a function of age class. (B) Age class-specific rate (probability) of suture fusion. Black/white bars: chimpanzees/humans; asterisk: dental age class of Taung. Note that A represents the cumulative probability distribution of B.

Already 70 y ago, persistence of the MS was hypothetically related to the evolutionary developmental transformation of the human skull (23, 24), including neurocranial expansion and thinning of the vault bones (23). Hypotheses on the genetic control of MS fusion have also been proposed (24). The large body of modern genetic evidence indicates that a network of at least 10 key genes mediates neurocranial suture fusion, and that these genes act differently on different sutures (25). It is intriguing that the recent sequencing of the Neanderthal genome provides evidence for positive selection in the modern human variant of one of these key genes, RUNX2, which is known to affect MS fusion (26). RUNX2-related disorders, such as cleidocranial dysplasia, result in delayed MS fusion and pathologies, such as extreme bulging of the forehead and hypertelorism (27). Premature closure of the metopic suture (metopic synostosis), on the other hand, typically results in trigonocephaly: that is, a narrow forehead with an external metopic ridge (keel) extending from glabella to the midforehead, relatively close-set orbits and no lateral browridge (28, 29).

Although various authors regard variation in MS fusion as an adaptively neutral feature (17, 20), the combined comparative evidence provided here for chimpanzees, humans, and fossil hominins supports the hypothesis that the late fusion of MS may have become adaptive relatively early during hominin evolution. Selective pressures favoring late fusion might have resulted from three different, but mutually nonexclusive, aspects of perinatal ontogeny: first, the obstetric dilemma; second, high early postnatal brain growth rates; and third, reorganization of the frontal neocortex.

Obstetric Dilemma.

As bipedalism was refined in conjunction with an evolutionary increase in neonate and adult brain sizes, the morphology of the birth canal constrained the size and shape of the neonate (30–32). Although exactly when during hominin evolution the obstetric dilemma arose has been a subject of debate (33, 34), this dilemma is especially severe in humans because of their large-headed (and relatively large-brained) neonates and relatively constricted birth canals. The anterior fontanelle and patent metopic suture of human neonates facilitate parturition. During delivery, contractions of the birth canal cause the edges of the neonate’s frontal and parietal bones to overlap and glide together in the region of the anterior fontanelle, which compresses the head and facilitates expulsion of the neonate from the birth canal (35). In early hominins, increased mobility of the neurocranial bones through delayed MS fusion might have represented an adaptive advantage facilitating birth (33).

High Early Postnatal Brain Growth Rates.

Compared with chimpanzees, human brains continue to grow at high fetal-like rates throughout the first postnatal year of life, which “may reflect the ontogeny of the ‘infrastructure’ required for rapid cognitive development” (p. 162 in ref. 36). We hypothesize that late MS closure in modern humans reflects an evolutionary adaptation of the growing frontal neurocranium to keep up with high brain growth rates. When sustained early brain growth appeared during hominin evolution is still a matter of debate (37). The large endocranial volume of the Mojokerto child at an age of < 2 y provides evidence for high early brain growth rates in H. erectus (37). Direct evidence for australopith early postnatal cranial ontogeny is currently not available, but evidence for delayed MS fusion in australopiths indicates that early brain growth may already have been fast before the emergence of the genus Homo. If so, rapid early postnatal brain growth preceded the increase in brain size in Homo, which could be because of the obstetric dilemma shifting prenatal brain growth rates postnatally in association with pelvic modifications for bipedalism, or because of an increase in relative brain size (i.e., increased encephalization) in australopiths compared with their (unknown) ancestors.

Reorganization of the Frontal Cortex.

The association of unfused MS with increased interorbital and frontal bone widths in extant humans (38, 39) is intriguing when one considers brain shape and possible neurological reorganization in early hominins (13) in conjunction with the frequencies of unfused MS in different taxa. As noted, a sample of Australopithecus and early Homo specimens that lived between ∼3.0 and 1.5 million y ago shows an unfused MS (Table S3), but no Paranthropus specimen does. In keeping with the tendency for an unfused MS in humans, A. africanus is characterized by increased interorbital and frontal bone widths compared with Paranthropus (40). Also consistent with findings for extant humans with unfused MS, endocasts of A. africanus have increased frontal widths in the region of the rostral prefrontal cortex (in addition to an expanded orbital frontal cortex) compared with endocasts of Paranthropus (13). It is therefore reasonable to hypothesize that, in addition to reflecting an adaptation to high postnatal brain growth rates, an unfused MS in Taung and other early gracile hominins may have been associated with the evolution of certain morphological (13) and cytoarchitectural features (41) of the prefrontal cortex, parts of which are differentially enlarged in humans and known to be crucial for their advanced cognitive capabilities. [As an aside, it is worth noting that humans also have a unique phase of shape change in their braincases before their deciduous teeth begin to erupt that results in a more general neurocranial globularization compared with chimpanzees (42–45).] If so, the evolution of increased rates of postnatal brain growth and neurological reorganization were probably entwined in (at least some) species of gracile early hominins.

Immature fossil hominins are currently playing a greater role in shaping the ways comparative data from living primates are interpreted (45, 46), and it is within this context that the suture morphology of Taung and other fossil hominins that lived more recently than 3 million y ago is interesting. Although it is beyond the scope of the present article, we hope that future researchers will test and extend the present findings by systematically collecting data on MS, anterior fontanelles, and endocranial size and shape in a wider sample of hominins, including those that lived before ∼3 million y ago (e.g., Australopithecus afarensis) as well as hominins that lived more recently than the fossils we have sampled. Such data are expected to contribute to a more detailed understanding of when, and in which hominin species, rates of postnatal brain growth first began to increase. This understanding, in turn, may contribute to a better grasp of the relationship between the evolutionary refinement of bipedalism and the evolution of brain size and shape.

Materials and Methods

Sample.

The state of MS fusion (unfused, partially fused, fully fused) was observed in ontogenetic samples of wild-shot Pan troglodytes (n = 407; CT data acquired on 78 specimens) and Pan paniscus (n = 136; CT data for 41 specimens), and in a worldwide ontogenetic sample of modern humans (n = 1060; CT data for 240 specimens). In addition, a sample of n = 376 individuals from Swiss medieval populations was examined. Pan data are from the Royal Africa Museum, Tervuren Belgium, and from the Collections of the Anthropological Institute and Museum, University of Zurich. Dental age classes used for developmental seriation were defined as follows: preterm fetal, neonate (from birth to before the eruption of the first teeth), dm1 (first deciduous molars erupted), dm2 (second deciduous molars erupted), M1/M2/M3: first, second, third permanent molars erupted. Details of sample structure and scoring procedures are provided in SI Materials and Methods and Tables S1 and S2. Because Taung only preserves endocranial evidence of MS, we use internal MS fusion scores for all comparative analyses in this study. Contrary to earlier studies (summarized in ref. 47, p. 478), we found no statistically significant difference in MS fusion patterns between P. troglodytes and P. paniscus (see SI Materials and Methods, MS in chimpanzees, bonobos, and modern humans and Tables S1 and S2). Suture fusion in Taung was studied on D.F.’s copy of the original endocast and on medical CT scans of the original fossil. Data on MS fusion status in fossil hominins were collected from the literature, D.F.’s collection of hominin endocasts and cranial casts, and from CT data (Table S3).