New High-Resolution Computed Tomography Data of the Taung Partial Cranium and Endocast and Their Bearing on Metopism and Hominin Brain Evolution

Home , Anterior fontanelle, Nasion

New high-resolution computed tomography data of the Taung partial cranium and endocast and their bearing on metopism and hominin brain evolution

Ralph L. Hollowaya,1, Douglas C. Broadfieldb, and Kristian J. Carlsonc,d,e

aDepartment of Anthropology, Columbia University, New York, NY 10027; bDepartment of Anthropology, Florida Atlantic University, Boca Raton, FL 33431; cEvolutionary Studies Institute, Palaeosciences Centre, and dSchool of Geosciences, University of Witwatersrand, Johannesburg 2050, South Africa; and eDepartment of Anthropology, Indiana University, Bloomington, IN 47405

Edited* by C. Owen Lovejoy, Kent State University, Kent, OH, and approved July 16, 2014 (received for review February 17, 2014) Falk and colleagues [Falk D, Zollikofer CP, Morimoto N, Ponce de Falk and colleagues (2) recently argued in support of such León MS (2012) Proc Natl Acad Sci U S A 109(22):8467–8470] hy- relationships. They proposed that delayed fusion of the metopic pothesized that selective pressures favored late persistence of suture coupled with delayed closure of the anterior fontanelle a metopic suture and open anterior fontanelle early in hominin provided a selective advantage during hominin evolution. The evolution, and they put an emphasis on the Taung Child (Austral- advantages, they reasoned, stemmed from mutually nonexclusive opithecus africanus) as evidence for the antiquity of these adaptive evolutionary pressures resulting from an “obstetric dilemma,” features. They suggested three mutually nonexclusive pressures: an high early postnatal brain growth rates, and reorganization in the “obstetric dilemma,” high early postnatal brain growth rates, and frontal neocortex (2: 8469). Falk and colleagues placed substantial neural reorganization in the frontal cortex. To test this hypothesis, emphasis on the Taung Child (Australopithecus africanus)intheir we obtained the first high-resolution computed tomography (CT) arguments, as of the 63 fossil hominin crania they examined for data from the Taung hominin. These high-resolution image data a metopic suture (2, Table S3), Taung was the only australopith in and an examination of the hominin fossil record do not support which they observed a partial or unfused metopic suture. Therefore, the metopic and fontanelle features proposed by Falk and col- their argument that a persistent metopic suture and open anterior leagues. Although a possible remnant of the metopic suture is ob- fontanelle constitute an adaptation that extends substantially ANTHROPOLOGY served in the nasion–glabella region of the Taung partial cranium backward in time within the hominin lineage (i.e., into pre-Homo (but not along the frontal crest), this character state is incongruent lineages) rests entirely on the status of these features in the Taung with the zipper model of metopic closure described by Falk and Child. Given that they play such a prominent role in these recent colleagues. Nor do chimpanzee and bonobo endocast data support claims (2), we suggest it is prudent to critically reexamine the state the assertion that delayed metopic closure in Taung is necessary of these features in the Taung Child. In doing so, we offer high- because of widening (reorganization) of the prefrontal or frontal resolution computed tomography (CT) image data of the specimen. cortex. These results call into question the adaptive value of delay- Prefrontal widening, although perhaps indicating a volume ing metopic closure, and particularly its antiquity in hominin evolu- increase (assuming no proportional decreases in lobe/region tion. Further data from hominoids and hominins are required to volumes elsewhere), is an insufficient indicator of cortical re- support the proposed adaptive arguments, particularly an obstetric organization. Rather, documenting cortical (prefrontal/frontal) dilemma placing constraints on neural and cranial development reorganization in brains of early hominin taxa requires the dem- in Australopithecus. onstration of differences in endocast convolutional details that can be traced to underlying cytoarchitectonic changes. Logically, it prefrontal reorganization | cranial capacity | human evolution Significance he Taung natural endocast, a cast of the endocranial surface Tof the cranium, has always been problematic because of The Taung Child (Australopithecus africanus) has historical and preservation and preparation. The endocast is distinguished by scientific importance in the fossil record as the first and best numerous calcite deposits on its surface. Indeed, the midline example of early hominin brain evolution. It was recently region is marked by numerous calcite deposits, indicating that proposed that Taung exhibits adaptive morphology (e.g., per- the true surface of the endocranium lies below these deposits sistent metopic suture and open anterior fontanelle), permit- (Fig. 1). In addition, there are remnants of the endocranial table ting important postnatal brain growth late into infancy. As adhering to the surface of the natural endocast, obscuring many Taung provides the only purported pre-Homo fossil evidence of its morphological features, including portions of its left frontal for the suggested adaptive mechanism, we test the hypothesis lobe (Fig. 1). Finally, Dart’s exuberant use of his wife’s knitting that it displays these features. Using new high-resolution needles to extricate the fossil from the surrounding breccia may images and in silico exploration, we did not observe the hy- have scarred the surfaces, forever obliterating some of the pothesized form of these features. If delayed metopic suture original features of the natural endocast. closure were adaptive, for example, permitting substantial The metopic suture is a synostosis between the two halves of postnatal brain growth and alleviating an obstetric dilemma, the frontal bone, and although it is present in all mammals there is no evidence this mechanism evolved before Homo. during development, it normally fuses in humans shortly after birth. Biological anthropologists most often refer to its persis- Author contributions: R.L.H., D.C.B., and K.J.C. designed research, performed research, tence in adults as a nonmetric trait (Fig. S1), the frequency of analyzed data, and wrote the paper. which varies among different human populations (1). Its pres- The authors declare no conflict of interest. ence in primates has no documented relationship to the size and *This Direct Submission article had a prearranged editor. form of the underlying prefrontal cortex, nor to corresponding Freely available online through the PNAS open access option. areas on endocasts. Specifically, known correlations between the 1To whom correspondence should be addressed. Email: [email protected]. presence of metopism/open anterior fontanelle and prefrontal/ This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. frontal lobe reorganization have not yet been demonstrated. 1073/pnas.1402905111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1402905111 PNAS Early Edition | 1of6 Downloaded by guest on October 4, 2021 the image data available to Falk and colleagues (2) and less affected by the aforementioned well-known preservation issues. Specifically, we use these data to assess whether the sharp-edged remnant of cortical bone covering much of the left half of the frontal region of the natural endocast reflects an open metopic suture, as has been suggested (2, 7). If direct evidence for such a metopic suture is to be found, then we would predict that any remaining evidence of a metopic suture in the existing frontal should be found in more caudal regions and should be continuous with the extent of the metopic suture suggested by Falk and colleagues on the natural endocast (i.e., support a “zipper” model of anterior to posterior closure), and that evidence of a protruding ridge along the midline of the natural endocast surface extending rostrally from bregma (or rostrally from the purported open anterior fontanelle) should be observable to some unknown extent, indicating the presence of an open metopic suture that became filled by intrusive matrix during formation of the natural endocast. Results A metopic suture is visible neither on the external table of the frontal bone (Fig. 1) nor on the rendering of the partial cranium, with the possible exception of the region between nasion and glabella (Fig. 2 A and B). A coronally oriented virtual section passing approximately through the rostral-most region of the cranium receiving the frontal poles of the endocast (Fig. 2C), which are visualized as two asymmetric elliptical regions (darker areas) surrounded by diploic bone (lighter areas), displays a thin suture internally connecting the periosteal surface near nasion with the endocranial surface near the left frontal pole. Note that the frontal squama in this virtual section (Fig. 2C) is uninterrupted by such a feature. This internal structure appears to be manifested superior to nasion on the external table of the cranial rendering as a remnant metopic suture (Fig. 2B). Additional internal evidence Fig. 1. Superior view of the original Taung partial cranium and endocast. of a metopic suture (Fig. S1) in the frontal of the Taung partial The frontal portion faces toward the bottom of the image. The bregmatic cranium is absent in a virtual section through glabella and through region of the natural endocast is roughly 3 cm posterior to the broken edge virtual sections between glabella and the posterior break (Fig. 2 C of the frontal bone in this image. Calcite deposits (black arrows) filling su- and D). tural voids during the fossilization process can be seen on the surface of the natural endocast, most visibly along the right coronal suture. The left hemi- Fig. 3A illustrates the location of representative coronally sphere of the natural endocast is covered by remnants of the endocranial table oriented virtual sections through the Taung natural endocast. to varying extents and thicknesses. It appears that remnant bone covers the The thin remnant of the endocranial table covering the left side same type of calcite infill associated with the left coronal suture, whereas such of the frontal region can be differentiated from the underlying infill is not as visible at the medial-most margin of the remnant bone anterior surface of the natural endocast in the high-resolution image data to bregma (i.e., location of the purported metopic suture). (Fig. 3B). In each of the illustrated sections, the remnant layer comes to an abrupt end near the midline (i.e., immediately left of the leftmost of the white arrows). In other words, there is not a must be remembered that internal prefrontal/frontal reorgani- separate superiorly projecting ridge of matrix at each terminus zation (e.g., cytoarchitectonic changes) may occur without any rising from the natural endocast beneath the bone, as would be external evidence manifested on the surface of a brain or on its expected if an open (or partially open) metopic suture was representation via the endocast surface. Nonetheless, the pre- present during the formation of the natural endocast. Rather, frontal region of the Taung endocast has not been previously ar- uniform material contrast throughout the remnant layer until its gued to share any demonstrable reorganizational features in sulcal sudden cessation at the midline suggests a protruding ridge was morphology with Homo to the exclusion of chimpanzees or other absent, unlike what is observable on the natural endocast at the apes, particularly in the third inferior frontal convolution (3, 4). right coronal suture (Fig. 1). Indeed, some have steadfastly claimed that the prefrontal region High-magnification stereomicrographic pictures also were of the Taung endocast surface was apelike (5). Neural re- taken of the midline surface of the Taung natural endocast at organization of the australopith brain toward a derived Homo multiple locations between its midfrontal region and bregma to configuration remains an intriguing possibility, however, given independently assess evidence for the hypothesized protruding the recent claim (6) that the presumably later australopith ridge-like structure (Fig. S2). These images also provide no con- from Malapa (i.e., MH1), Australopithecus sediba, has an endo- clusive evidence of a protruding ridge similar to that appearing at cast that may display some reorganization in surface features of the right coronal suture. In general, these images corroborate the its third inferior convolution. coronally oriented virtual sections (Fig. 3B) in showing no support Because the Taung Child is the only australopith put forth in for a superiorly projecting ridge of calcite rising from the midline the adaptationist argument for metopism in hominins (2), critical of the natural endocast rostral to bregma or rostral from the evaluation of this specimen is necessary to verify the accuracy purported open anterior fontanelle. and evolutionary significance of these claims. Here we offer di- It is important to appreciate that a large degree of variation rect evidence for verifying the purported metopic suture and characterizes the small sample of additional A. africanus endocasts. open anterior fontanelle (2, 7), evidence that is less limited than Comparing them with a sample of 40 bonobo (Pan paniscus)

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1402905111 Holloway et al. Downloaded by guest on October 4, 2021 ANTHROPOLOGY

Fig. 2. Rendering of the Taung partial cranium produced from high-resolution image data obtained in this study, as well coronally oriented virtual sections through the frontal of the specimen. The rendering (A) and close-up of the nasion–glabella region (B) were generated from volume data consisting of isotropic 55.5 μm voxels. Three black arrows (B) arranged vertically indicate the approximate extent of a potential metopic suture persisting in the frontal. Coronally oriented virtual sections through the partial cranium obtained at glabella (g), nasion (n), and midfrontal (mf) locations are depicted in C. The black arrow in the virtual section through nasion indicates internal evidence of the potential persisting metopic suture. A coronally oriented virtual section through the posterior region of the frontal, just anterior to its broken edge, is depicted in D. Note the absence of internal evidence (e.g., see C and Fig. S1)for a persistent metopic suture in the midfrontal (mf) and posterior frontal (pf) sections.

endocasts from the collection of R.L.H. is instructive. Maximum argument of Falk and colleagues regarding prefrontal widening width measurements of australopith prefrontal regions across the indicating reorganization (2), one would have to conclude that breadth of third inferior convolutions are: Taung, 68 mm; the bonobo prefrontal was reorganized and more advanced than Sterkfontein (Sts) 60, 72 mm; Sts 71, 72 mm; Sterkfontein type 2, the prefrontal of several australopiths. Moreover, when aug- 78 mm; Sts 5, 80 mm; and Sterkfontein (StW) 505, 80 mm. menting the bonobo sample with an additional 35 chimpanzee Where only a single side is available, width was doubled to (Pan troglodytes) endocasts from the collection of R.L.H., the obtain these full breadths. Two endocasts [endocranial volumes combined mean EV becomes 364 mL (SD, 43 mL), with a com- (EVs) of 359 and 372 mL] randomly chosen from among the 40 bined mean maximum prefrontal breadth between Broca’s caps bonobos have maximum prefrontal widths of 73 and 74 mm, respectively. A third bonobo endocast (EV, 403 mL) with (third inferior convolutions) of 68.5 mm (SD, 4.3 mm). The ’ roughly the same EV as Taung has a prefrontal width of 77 mm. Spearman s rho correlation (rs) between EV and prefrontal None of the 40 bonobo endocasts in the sample display third breadth in this sample of 75 Pan endocasts is 0.678 (P < 0.001). inferior convolutional reorganization beyond what is usually Clearly, convolutional patterns provide a more conclusive means observed in anthropoids. If one compares these bonobo endocasts of comparing prefrontal/frontal organization than widening with those of A. africanus mentioned earlier and applies the or EV.

Holloway et al. PNAS Early Edition | 3of6 Downloaded by guest on October 4, 2021 Discussion State of Metopic Suture in Taung. The form of metopic suture in- dicated in the high-resolution image data set of the Taung partial cranium is inconsistent with the form of metopic suture and open anterior fontanelle described by Falk and colleagues (2). A persistent metopic suture in the region between nasion and glabella would contradict an anterior to posterior zipper model of closure suggested by Falk and colleagues. Moreover, using the high-resolution image data set of the Taung natural endocast, we find no evidence of a distinct superiorly protruding ridge on the Taung natural endocast indicating a possible open (or partially fused) metopic suture rostral to bregma. Thus, neither of the predicted features necessary to support the form of metopic suture in the Taung Child suggested by Falk and colleagues (2) are observed in the present study. Standard medical CT images do not provide spatial resolution equivalent to that of the high-resolution image data set generated in the present study (Fig. S3), and thus it would appear that Falk and colleagues (2) may have been unknowingly constrained by this deficiency. According to surface renderings produced from medical CT data (e.g., see their figure 1), it would appear possible that they may have interpreted the midline border of the remnant endocranial table as a superiorly protruding ridge indicative of an open (or partially fused) metopic suture. In modern humans, the metopic suture normally closes 3–9mo postpartum (8, 9), whereas in Pan, all of the sutures (including the metopic suture) normally close before birth (10, 11), notably earlier than in Homo. This is likely because suture fusion is more strongly related to dental development than it is to brain development (11, 12). With an estimated age between 3.73 and 3.93 y at death, Taung is well beyond the age for normal closure of the metopic suture exem- plified by modern humans or chimpanzees and other primates (11, 13), unless the suture were to persist as a nonmetric trait (1). Given the results of the study and the accepted age of Taung, the latter scenario would seem more likely than would a metopic suture persisting as a selectively advantageous trait implicated in brain development. Indeed, reorganization and development of the prefrontal and frontal cortex in humans exhibits little influence on fusion of the metopic suture, as development of these areas continues well beyond the age at which the metopic suture normally fuses. The high-resolution image data generated in the present study are no more definitive than the medical CT image data used by Falk and colleagues (2) in determining whether there was an open anterior fontanelle in the Taung Child, except that the higher spatial resolution permits more precise visualization of borders of the purported structure. The borders are not as sharp as one would expect if the anterior fontanelle remained open (Fig. 3A), but the aforementioned damage during extraction of the fossil or decades of subsequent wear could be contributing factors to the observed lack of sharpness. It is also worth noting that an open anterior fontanelle would be extraordinarily rare if Taung was indeed 3.5–4 y of age (13, 14), particularly if Taung followed a growth pattern assumed to be more similar to chimpanzees than humans (15, 16). In addition to refuting the presence of an open metopic suture and anterior fontanelle in the Taung Child, this evidence calls Fig. 3. Rendering of the Taung natural endocast produced from high-res- into question the proposed adaptive significance of delayed olution image data obtained in this study, as well as coronally oriented metopic suture and anterior fontanelle closure in hominin evo- sections through the specimen. The rendering (A) was generated from vol- lution, particularly in australopiths. We suggest that a more ume data consisting of isotropic 65 μm voxels. Coronally oriented virtual reasonable and parsimonious explanation for the observed form sections through the natural endocast (B) were obtained at two frontal of metopic suture in Taung (i.e., possibly retained between locations, one anterior (fa) and one posterior (fp). White arrows in each of nasion and glabella) would be its manifestation as an epigenetic the two virtual sections point to the remnant endocranial table adhering to the left hemispheric surface of the natural endocast. Note that the remnant trait, as in modern humans (1) and many other primates. layer varies in thickness and appears to abruptly end immediately to the left of the leftmost white arrows in each section. There is no discernible ridge of Frontal Widening and Reorganization. Prefrontal widening in iso- material protruding superiorly from the endocast surface in either location, lation, although perhaps indicating a volume increase (assuming as occurs along the coronal suture. no proportional decreases in lobe/region volumes elsewhere),

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1402905111 Holloway et al. Downloaded by guest on October 4, 2021 should not be considered indicative of cortical reorganization. As Prat neither demonstrates nor illustrates the metopic suture demonstrated here, some great ape endocasts (e.g., those of claimed to be present in KNM-ER 1805 and other fossils. We bonobos and chimpanzees) display prefrontal widening that also note the absence of references to a metopic suture or open exceeds prefrontal breadths of Taung, Sts 60, and Sts 71 (Fig. S4 anterior fontanelle in several original descriptive works, such as and Fig. S5) while simultaneously exhibiting EVs smaller than those of Day and colleagues (26), Walker and Leakey (27), or that of Taung. These same Pan endocasts exhibit convolutional Schwartz and Tattersall (28), for KNM-ER 3883 or KNM-ER details that are reflective of ape patterns and that are not more 3733. Although the latter two specimens are described by Leakey derived toward Homo patterns than are the patterns exhibited by and Walker (29) as retaining traces of a metopic suture, these the australopith endocasts. features are described as occurring in the glabellar region, which Even considering an exclusively hominin sample of endocasts, contradicts the explicit locations (i.e., near bregma) suggested by the relationship between prefrontal breadth and cortical orga- Falk and colleagues (2). In contrast to Falk and colleagues (2), nization is uncertain. Although illustrations of prefrontal outlines, R.L.H. and D.C.B. find no evidence for an open metopic suture, as shown in ref. 17, clearly indicate a greater degree of “point- even ectocranially, in the Sambungmacan 3 Homo erectus cranium, edness” to prefrontal regions of “robust” australopith endocasts having studied the original fossil and made endocast ourselves (30). compared with prefrontal regions of “gracile” australopith endo- Given the relevance of location to the zipper closure model en- casts (e.g., those of A. africanus), these outlines cannot be directly dorsed by Falk and colleagues (2), as it was with the Taung Child, it translated into neurological/functional differences in cortical re- would be worthwhile to study these (and other) specimens, using organization, given the considerable differences in cranial and facial more appropriate means (e.g., high-resolution CT) of determining development and architecture of these groups (18, 19). Convolu- the state of metopic suture closure. Such analyses should precede tional details of both robust and gracile forms also are simply mute relying on these specimens to support any adaptive significance of and do not allow conclusions regarding potential or actual differ- delayed metopic suture closure in hominin evolution. Ultimately, ences in their neural organization. even if future in silico techniques are successful in corroborating the suggested elevated metopic suture frequencies in original Homo Obstetric Dilemma and Metopism. As suggested by Falk and col- fossils, a connection between metopic suture frequency and the leagues (2), it might seem reasonable to implicate persistent proposed adaptive benefits of delayed closure must be demonstrated. metopism as an adaptive trait when facing an obstetric dilemma. For example, whether delayed closure facilitates reorganization Tague and Lovejoy (20) and subsequent others (e.g., see ref. 21), through convolutional changes that are distinctly non-ape must still however, suggest that although hominin parturition may have be corroborated in many early Homo endocasts (31). ANTHROPOLOGY become more difficult compared with that of extant anthropoids In sum, we believe the claim of high frequencies of metopic by the time of A. afarensis, the crucial dilemma would not have sutures in early hominins (2) is premature, and thus the propo- occurred until encephalization began to increase dramatically in sition that delayed metopic suture closure may have conferred Homo. Reconstructions of hominin pelves (e.g., Sts 14 and MH2) a selective advantage in early hominin evolution is equally pre- support the absence of a crucial obstetric dilemma in additional, mature. Rigorous analysis of a lynchpin specimen in this argu- Australopithecus later species of (20, 22). The absence of an adap- ment (i.e., the Taung Child) provides no support for the notion tive metopic suture in Taung also supports the notion that this di- that australopiths may have delayed metopic suture fusion, pos- lemma was not present in australopiths. sibly for adaptive reasons. We suggest that the scenario hypothe- Although it is true that pelvic dimensions changed from Aus- sized by Falk and colleagues (2) requires substantially more tralopithecus to Homo, in part to accommodate the larger-brained evidence to support the anatomic, neurologic, and adaptive asser- babies of the latter, there is little reason to suggest that fetal and tions promulgated therein. Where instances of direct disagreement postpartum development was under any more selective pressure in exist with original descriptions, the state of metopic sutures sug- australopiths than in anthropoids and modern humans. Also, given gested by Falk and colleagues should be precisely illustrated to the average size of the birth canal in Australopithecus (20), the support the new characterizations. To this end, we suggest that estimated size of the neonate brain, and the growth trajectory in australopiths, based on dental and osteological data (23), there is high-resolution image data sets, such as obtained here from the no evidence that Australopithecus experienced selective pressures Taung specimens, could be extremely helpful, if not necessary, to necessitating delayed brain or skull development greater than substantiate claims of delayed metopic suture closure. Relatively those observed in modern humans. Rather, the recently described low-resolution images (e.g., medical CT) likely do not offer A. sediba pelvis (22), despite an EV of 420 mL (6), suggests that enough spatial resolution to provide definitive evidence, as was the physical constraints of even the late Australopithecus birth the case with the Taung Child. Also, location discrepancies ob- canal probably had less effect on parturition and the timing of served in Taung, and seemingly in other endocasts such as KNM-ER suture closure than in modern humans. It is also worth noting that 1813, KNM-ER 3733, and KNM-ER 3883, must be reconciled with an additional contributing factor to an obstetric constraint on fetal the zipper suture closure pattern suggested by Falk and colleagues brain size has been suggested by Dunsworth and colleagues (24), (2) (i.e., a closure process concluding with fusion at bregma, rather who posited that a metabolic limit exists that restricts the ability of than nasion/glabella). Not all purported instances of partially fused a female to keep a fetal brain growing. or open metopic sutures in hominin fossils support this closure model, but some may instead reflect the underlying mechanisms Metopism in Homo. Although the present study focused on the manifested in epigenetic or nonmetric traits. issue of metopism in australopiths, and specifically in the Taung Child, our experience with this specimen leaves us skeptical of Methods claims that many early Homo specimens (e.g., OH 24, KNM-ER The Taung (UW-1) partial cranium and natural endocast were separately 1805, KNM-KNM-ER 1813, KNM-ER 3883, 3733, and so on, as scanned by K.J.C. in the Paleosciences Centre Microfocus X-ray CT facility of listed in table S3 of ref. 2) clearly and without ambiguity also the University of the Witwatersrand. The system is a Nikon Metrology XTH demonstrate persistent (or partially fused) metopic sutures or 225/320 LC dual-source industrial CT system (see www.wits.ac.za/microCT for additional information on the system). Each specimen was securely wrapped open anterior fontanelles. Falk and colleagues (2) do not de- in standard bubble wrap to prevent movement during scanning and to scriptively substantiate or illustrate these examples and identify protect it from being damaged by the walls of the container. The container support as coming from low-resolution CT data, which in the was an acrylic tube with an external diameter of 110 mm and a wall thick- case of Taung, proved to be insufficient (Fig. S3). Although they ness of 5 mm. It was closed off on the bottom by a horizontal acrylic piece (2) claim to have “confirmed” features described by Prat (25), glued to the bottom of the tube. Once a specimen was secured inside the

Holloway et al. PNAS Early Edition | 5of6 Downloaded by guest on October 4, 2021 acrylic container, the container was attached to the rotating sample holder After importing volume data, the same procedure was used for pro- platform of the scanner by beeswax. ducing renderings and virtual sections of both specimens in the Virtual The partial cranium was scanned using the following parameters: 135 kV; Imaging in Paleontology laboratory of the University of the Witwatersrand. 400 μA; 6,000 projections, where each projection represented the average of Using the region growing tool, the object of interest was separated from two frames; an acquisition time of two frames per second; and copper (1.2 other voxels in the volume containing background and noise. Once the mm) and aluminum (3.6 mm) filtration. Data were acquired with an isotropic region of interest was identified, a volume rendering of that region was μ voxel size of 55.5 m. Projections were reconstructed into volume data using generated. The volume rendering was positioned in Frankfurt horizontal, proprietary Nikon metrology software (i.e., CT Pro; Nikon Metrology, Inc). after which coronal sections were obtained by scrolling through the Automated protocols in CT Pro were applied for correcting beam hardening repositioned volume. Each section was saved as a separate .tif file, along and reducing noise. Thirty-two-bit float volume data produced from the with an image of the volume rendering indicating the location of the scanning were imported into VG Studio Max 2.1 (Volume Graphics GmbH) virtual section. and transformed into 16-bit unsigned volume data. While importing volume data, the gray value histogram was calibrated to background and material values of the specimen scanned in air. ACKNOWLEDGMENTS. We are grateful to Dr. Bernhard Zipfel and Brendon Billings for facilitating access to and assisting with transport of the Taung The natural endocast was scanned using the following parameters: 190 kV, specimens and modern human crania from the Dart Collection. We also are 330 uA, 5,000 projections, an acquisition time of four frames per second, and grateful to the Department of Science and Technology of South Africa and copper (1.2 mm) and aluminum (3.6 mm) filtration. Data were acquired with the National Research Foundation of South Africa for multiple funding an isotropic voxel size of 65 μm. The ring artifact correction option in Nikon awards that permitted the creation of the Palaeosciences Centre Microfocus proprietary software (i.e., Inspect-X; Nikon Metrology, Inc) was used, which X-ray CT facility and the Virtual Imaging in Palaeontology laboratory of the essentially averages three frames to reduce noise accumulation in the center University of the Witwatersrand. We thank Prof. Alan Morris for providing of the field of view. Projections were reconstructed into volume data, using information on examples of adult modern humans with metopic sutures in proprietary Nikon metrology software (i.e., CT Pro). Automated protocols in the Dart collection. We also thank Dr. Tea Jashashvili for assistance with scanning the modern human cranium (A 3163) and Kimberley Chapelle for CT Pro were applied for correcting beam hardening and reducing noise. assistance with segmenting Taung images. We are grateful to Dr. Owen Thirty-two-bit float volume data produced from the scanning were imported Lovejoy for useful critical comments and, in particular, to Dr. Tim White into VG Studio Max 2.1 and transformed into 16-bit unsigned volume data. for critical comments and useful suggestions regarding future research on While importing volume data, the gray value histogram was calibrated to the Taung specimen. We also thank two anonymous reviewers for their background and material values of the specimen scanned in air. constructive suggestions.

1. Hauser G, De Stefano GF (1989) Epigenetic Variants of the Human Skull 15. Coqueugniot H, Hublin J-J, Veillon F, Houët F, Jacob T (2004) Early brain growth in (Schweizerbart, Stuttgart). Homo erectus and implications for cognitive ability. Nature 431(7006):299–302. 2. Falk D, Zollikofer CPE, Morimoto N, Ponce de León MS (2012) Metopic suture of 16. Zuckerman D (1928) Age-changes in the chimpanzee, with special reference to Taung (Australopithecus africanus) and its implications for hominin brain evolution. growth of brain eruption of teeth and estimation of age with a note on the Taung Proc Natl Acad Sci USA 109(22):8467–8470. ape. Proc Zool Soc Lond 1:1–42. 3. Neubauer S, Gunz P, Weber GW, Hublin J-J (2012) Endocranial volume of Austral- 17. Falk D (2012) Hominin brain evolution, 1925-2011: An emerging overview. African opithecus africanus: New CT-based estimates and the effects of missing data and Genesis: Perspectives on Hominin Evolution, eds Reynolds SC, Gallagher, A. (Cambridge small sample size. J Hum Evol 62(4):498–510. University Press, Cambridge), pp 145–162. 4. Falk D, Clarke R (2012) Letter to the Editor: Response to Holloway and Broadfield’s 18. Kimbel WH, Rak Y, Johanson JC (2004) The Skull of Australopithecus afarensis (Oxford Critique of Our Reconstruction of the Taung Virtual Endocast. Am J Phys Anthropol University Press, New York). 148:483-485. 19. Rak Y (1983) The Australopithecine Face (Academic Press, New York). 5. Falk D (2009) The natural endocast of Taung (Australopithecus africanus): Insights 20. Tague RG, Lovejoy CO (1986) The obstetric pelvis of A.L. 288-1 (Lucy). J Hum Evol 15: from the unpublished papers of Raymond Arthur Dart. Am J Phys Anthropol 140 237–256. (Suppl 49):49–65. 21. Wells JCK, DeSilva JM, Stock JT (2012) The obstetric dilemma: An ancient game of 6. Carlson KJ, et al. (2011) The endocast of MH1, Australopithecus sediba. Science Russian roulette, or a variable dilemma sensitive to ecology? Am J Phys Anthropol 149 333(6048):1402–1407. (Suppl 55):40–71. 7. Hrdlicka A (1925) The Taung ape. Am J Phys Anthropol 8:379–392. 22. Kibii JM, et al. (2011) A partial pelvis of Australopithecus sediba. Science 333(6048): 8. Weinzweig J, et al. (2003) Metopic synostosis: Defining the temporal sequence of 1407–1411. normal suture fusion and differentiating it from synostosis on the basis of computed 23. DeSilva JM, Lesnik JJ (2008) Brain size at birth throughout human evolution: A new tomography images. Plast Reconstr Surg 112(5):1211–1218. method for estimating neonatal brain size in hominins. J Hum Evol 55(6):1064–1074. 9. Vu HL, Panchal J, Parker EE, Levine NS, Francel P (2001) The timing of physiologic 24. Dunsworth HM, Warrener AG, Deacon T, Ellison PT, Pontzer H (2012) Metabolic hy- closure of the metopic suture: A review of 159 patients using reconstructed 3D CT pothesis for human altriciality. Proc Natl Acad Sci USA 109(38):15212–15216. scans of the craniofacial region. J Craniofac Surg 12(6):527–532. 25. Prat S (2002) Anatomical study of the skull of the Kenyan specimen KNM-ER 1805: A 10. Patten CJ (1911) Cranium of a chimpanzee, showing metopic suture; also fontanelle re-evaluation of its taxonomic allocation? C R Palevol 1:27–33. and sutural bone-plates. J Anat Physiol 46(Pt 1):33–35. 26. Day MH, Leakey REF, Walker AC, Wood BA (1976) New hominids from East Turkana, 11. Cray J, Jr, Cooper GM, Mooney MP, Siegel MI (2012) Ectocranial suture fusion in Kenya. Am J Phys Anthropol 45(3 PT 1):369–435. primates: As related to cranial volume and dental eruption. J Med Primatol 41(6): 27. Walker AC, Leakey REF (1993) The Nariokotome Homo erectus skeleton (Harvard 356–363. University Press, Cambridge, MA). 12. Cray J, Jr, Mooney MP, Siegel MI (2010) Timing of ectocranial suture activity in Pan 28. Schwartz JH, Tattersall I (2005) The Human Fossil Record (Wiley-Liss, New York), Vol 4. troglodytes as related to cranial volume and dental eruption. Anat Rec (Hoboken) 29. Leakey REF, Walker AC (1985) Further hominids from the Plio-Pleistocene of Koobi 293(8):1289–1296. Fora, Kenya. Am J Phys Anthropol 67(2):135–163. 13. Lacruz RS, Rozzi FR, Bromage TG (2005) Dental enamel hypoplasia, age at death, and 30. Broadfield DC, et al. (2001) Endocast of Sambungmacan 3 (Sm 3): A new Homo weaning in the Taung child. S Afr J Sci 101:567–569. erectus from Indonesia. Anat Rec 262(4):369–379. 14. Bromage TG, Dean MC (1985) Re-evaluation of the age at death of immature fossil 31. Holloway RL, Broadfield DC, Yuan MS (2004) Brain Endocasts: The Paleoneurological hominids. Nature 317(6037):525–527. Evidence. The Human Fossil Record (Wiley–Liss, Hoboken, NJ), Vol 3.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1402905111 Holloway et al. Downloaded by guest on October 4, 2021