Tuesday, May 23, 2017

Middle Pleistocene Homo Crania from Broken Hill and Petralona: Morphology, Metric Comparisons, and Evolutionary Relationships

G. Philip Rightmire
Human Paleontology and Prehistory
Part of the series Vertebrate Paleobiology and Paleoanthropology pp 145-159
25 January 2017



A fossilized human cranium was discovered by miners quarrying at Broken Hill (now Kabwe ) in 1921. Broken Hill is one of the best preserved hominins ever recovered from a later Middle Pleistocene locality. Remarkably, no comprehensive descriptive or comparative account has been published since 1928. Overall, Broken Hill resembles Homo erectus . The frontal is flattened with midline keeling, the vault is low, and the massive face is “hafted” to the braincase in such a way as to accentuate facial projection. At the same time, there are apomorphic features shared with later humans. Brain size is 1280 cm3, the temporal squama is arch-shaped, and the upper scale of the occipital is expanded relative to its lower nuchal portion. Specialized characters of the temporomandibular joint region include a raised articular tubercle and a sphenoid spine. Reorientation of the nasal aperture and placement of the incisive canal suggest that the face may be more nearly vertical than in H. erectus. It is apparent that Broken Hill is similar to other African crania from Bodo, Ndutu, and Elandsfontein as well as European fossils including Arago and Petralona. However, the systematic position of these hominins remains controversial. The material has been grouped into a series of grades within a broad H. sapiens category. A very different reading of the record recognizes multiple, distinct taxa and suggests that speciation must have occurred repeatedly throughout the Pleistocene. Still another perspective holds that differences among the African and European specimens are minor and can be attributed to geography and intragroup variation. It is argued that many of the fossils belong together in one widely dispersed taxon. If the Mauer mandible is included within this hypodigm , then the appropriate name is H. heidelbergensis . Treated in a broad sense, H. heidelbergensis is ancestral to both H. neanderthalensis and H. sapiens . This study will provide a detailed account of the morphology of Broken Hill and its similarities to other Middle Pleistocene hominins from Africa. Comparisons will include Arago, Petralona, and assemblages such as Sima de los Huesos. My approach will address the taxonomic utility of characters of the vault, cranial base and face, species-level systematics, and evolutionary relationships.

from the chapter:


The cranium from Broken Hill (now Kabwe) remains one of the treasures of prehistory. It was found in 1921, when miners quarrying for lead ore broke into the lower part of an extensive cavern containing quantities of mineralized bones and stone artifacts. Accounts of the circumstances surrounding this discovery are contradictory (Hrdlička 1930). Several additional human fossils, along with animal bones, were collected from the cave fill, but claims for the association of any of these elements with the original cranium remain incompletely documented. Comparative studies of the fauna have demonstrated similarities with the large assemblage from Elandsfontein in South Africa, indicating an early Middle Pleistocene age (Klein 1994; Klein et al. 2006). However, more recent efforts to date individual bones directly using Electron Spin Resonance (ESR) suggest that the Broken Hill material may be of late Middle Pleistocene antiquity (Stringer 2011).

The cranium was described initially by Woodward (1921), who saw resemblances to the Neanderthals then known from Europe but attributed the find to a new species (‘Homo rhodesiensis’). More comprehensive studies were published several years later by Pycraft (1928) and Mourant(1928). While pointing to differences in certain features, Mourant (1928) again argued for a close relationship between Broken Hill and Late Pleistocene Neanderthals. It is now recognized that this comparison was inappropriate.  Broken Hill lacks the specialized characters of Neanderthals but resembles other crania from Elandsfontein in South Africa and Bodo from the Middle Awash of Ethiopia. As a group, these African fossils are broadly similar to hominins from Middle Pleistocene localities in Europe including the Sima de los Huesos in Spain, Arago Cave in France, and Petralona in Greece.

Interpreting this record has been problematic. The number of taxa represented is disputed, and phylogenetic relationships remain to be clarified. In one view, African and European mid-Pleistocene populations can be grouped with later humans within a broad Homo sapiens category. Archaic and modern grades are defined by advances in brain size and skull form. Although changes to the vault and face accumulate in a mosaic pattern, early and late groups are said to follow one another seamlessly, as segments of a single evolving lineage (Bräuer 2007, 2008). A very different reading of the record recognizes multiple, distinct taxa as evidence for speciation occurring repeatedly throughout the Pleistocene (Tattersall and Schwartz 2008; Schwartz and Tattersall 2010). At least two lineages are identified, in addition to Homo erectus and recent humans. A European branch can be traced back via Petralona, Arago, and Sima de los Huesos, deeply into the Middle Pleistocene. Proponents of this view (Arsuaga et al. 1989, 1997; see also Hublin 2009) claim that even the oldest European hominins share apomorphies with Homo neanderthalensis and can reasonably be attributed to this species. A variation on this phylogenetic scheme has been proposed by Martinón-Torres et al. (2012), who find that the Sima de los Huesos teeth are “more Neanderthal” in form than the Mauer or Arago dentitions.

Given this result, Martinón-Torres et al. (2012) suggest that along with an ancient lineage linking the Sima hominins directly with Neanderthals, a second population including Mauer and Arago was present in Europe. This second species must be called Homo heidelbergensis.

A key question is how these European lineages are related to the hominins in Africa. If all of the European fossils are subsumed within Homo neanderthalensis, then the species represented by Broken Hill, Elandsfontein, and Bodo can be called Homo rhodesiensis, following the nomenclature proposed by Woodward (1921). Some (chronologically late) members of this group exhibit morphology that is archaic, coupled with characters suggestive of a link to anatomically modern humans. Still another perspective holds that morphological differences among the most ancient European and African specimens are minor and can be attributed to geography and intragroup variation (Stringer 1983, 1993; Rightmire 1990, 1996, 1998, 2008; Mounier et al. 2009). It can be argued that many of the fossils belong together in one geographically dispersed taxon. If the Mauer mandible is included within the hypodigm, then the appropriate name for this species is Homo heidelbergensis. Treated in this broad sense, Homo heidelbergensis must be ancestral to both Homo neanderthalensis and Homo sapiens.

While many of the Middle Pleistocene fossils are incomplete, Broken Hill is clearly one of the most informative specimens. Another is the cranium from Petralona. It is possible to document the extent to which these African and European fossils differ in their craniofacial morphology.

Over the course of nearly a century, Broken Hill has been treated in numerous comparisons involving modern humans, Neanderthals, and earlier Homo. Following Mourant (1928), many of these studies have been based on measurements or, more recently, cranial landmarks used in morphometric analysis (Friess 2010; Harvati et al. 2010, 2011). Another approach has emphasized anatomical description, with attention to the relevance of individual characters. Since it was discovered in 1959, the Petralona cranium has also been studied in detail (Stringer et al. 1979, 1983). This research has produced many useful data, but there is (still) no firm consensus as to the evolutionary significance of either specimen. In this review, I introduce further evidence from measurements and comparative anatomy. My goal is to clarify the relationship of Broken Hill to Petralona, with the goal of testing the null hypothesis that these individuals can be grouped together in one taxon.

Body of chapter omitted here.

Discussion and Conclusion of the paper below:


Hominin fossils are known from numerous Middle Pleistocene localities. It is recognized that these individuals display traits that are derived in comparison to Homo erectus. At the same time, the skulls retain numerous primitive features that set them apart from modern humans. How these diverse mid-Pleistocene assemblages should be classified, and how they fit into the “tree” of human evolution, are important questions. The crania from Broken Hill and Petralona are key specimens, from which inferences concerning the morphology of larger African and European regional populations can be drawn. Of course, all biological populations display variation, and the extent of this variation cannot be gauged adequately from small samples. Particularly for Africa, few complete skulls are available. Nevertheless, the detailed anatomical and metric comparisons conducted here provide information that is useful in evaluating the null hypothesis that Broken Hill and Petralona represent paleodemes of one species.

The two crania are similar in many aspects of form. Both are long with relatively low vertices, and both display massive and projecting supraorbital tori, flattened frontals heavily invaded by complex air sinuses, postorbital narrowing, and occipitals that are flexed relative to those of modern humans. Petralona differs from Broken Hill in having a wider cranial base, a reduced vol/aub ratio, massive supramastoid crests, and a less prominent torus crossing the occipital bone. The well preserved Broken Hill basicranium presents derived (sapiens-like) features including an increased petrotympanic angle associated with (coronal) alignment of the petrous and tympanic axes, “erosion” of the pyramid apex leading to enlargement of the foramen lacerum, a projecting sphenoid spine, and clear definition of an articular tubercle at the anterior margin of the mandibular fossa (Rightmire 1990, 2001, 2008). Insofar as the (damaged and partly obscured) cranial base of Petralona can be evaluated, its morphology resembles that of Broken Hill.

In forward placement of the facial skeleton relative to the anterior cranial fossa, Broken Hill and Petralona are comparable to Homo erectus. At the same time, the lateral margin of the nose is vertical, rather than forward sloping as in Homo erectus. The lower terminus of this border is set back below the overhanging nasal saddle. This reorientation suggests that the facial profile is less prognathic than in Homo erectus. In facial forwardness at subspinale, Broken Hill and Petralona are similar, and the angle at subspinale is reduced in relation to that of the flat-faced Bodo cranium. Both Broken Hill and Petralona possess prominent anterior nasal spines, coupled with spinal crests separating the sill from the subnasal portion of the maxilla.

Elsewhere in the facial skeleton, there is more variation. The Petralona face is broader at the zygomatic arches and exhibits a more robust cheek region than does Broken Hill. The orbital cavities are low and relatively broad. The walls of the maxillae appear to be inflated. Infraorbital foramina are not associated with any grooves or furrows. Neither here nor elsewhere in the cheek region is there much indication of hollowing. Petralona thus stands in some contrast to Broken Hill, where there are localized depressions of the infraorbital surface, even if no canine fossa is developed. These observations have been taken to indicate that not only Petralona but also Arago and other European mid-Pleistocene hominins anticipate the distinctive midfacial morphology of later Neanderthals (Hublin 1998, 2009). As described by Arsuaga et al. (1997), the cheek region of SH 5 is not inflated in the extreme manner of Neanderthals, but it can be interpreted as intermediate in form. How such facial features are evaluated (whether any of them can be judged to be true Neanderthal apomorphies) is critical to determining how the Petralona, Arago, and Sima de los Huesos individuals are related to populations outside of Europe and how the fossils should be treated in phylogenetic schemes.

Information relevant to these questions is advanced by Harvati et al. (2010), who have carried out a geometric morphometric study designed to quantify craniofacial shape in Middle Pleistocene hominins, Neanderthals, and modern humans. This analysis is based (in part) on landmarks situated on the supraorbital torus, the orbits and nasal aperture, the zygomatic bone, and the maxilla. After superimposition with generalized Procrustes analysis (GPA), mean configurations of the groups are assessed visually. When viewed in the transverse plane, configurations confirm that “classic” Neanderthals have “a more convex maxilla” and a more receding infraorbital profile than do recent populations. Importantly, Petralona, Arago, and SH 5 are essentially indistinguishable from Broken Hill and Bodo. Both European and African groups are said to approach (but not match) the Neanderthal condition. Midfacial prognathism is explored by comparing orientation of midsagittal landmarks in relation to lateral portions of the face. The Middle Pleistocene crania appear to have “less anteriorly placed” faces than Neanderthals and to be “nearly identical” to one another in (mean) shape. These findings suggest that there is little basis for claiming that the European mid-Pleistocene hominins are more similar to Neanderthals than their African counterparts. Consequently, Harvati et al. (2010) hypothesize that some facial attributes of Petralona and Arago commonly regarded as “incipient” Neanderthal features may instead be plesiomorphic states.

 A different interpretation invokes the effects of scaling. Size has long been recognized as contributing to variation in craniofacial shape (Lahr and Wright 1996). Maddux and Franciscus (2009) have used a geometric morphometric approach to explore the influence of allometry on the infraorbital region in Middle Pleistocene, Late Pleistocene, and recent populations of Homo. The authors project a grid onto each specimen, fitting it to the boundaries of the infraorbital plate. Landmarks are digitized as the intersections of grid lines and are intended to capture the topography of the underlying curvilinear surface. GPA serves to superimpose the landmarks of all specimens, aligning them to the mean configuration and allowing quantification of size and shape. Principal components analysis suggests that Neanderthals share with European and African mid-Pleistocene crania (and some Upper Paleolithic anatomically modern individuals) a relatively flat infraorbital surface topography.  Most recent human skulls exhibit relatively depressed infraorbital plates. It can be established that the degree of infraorbital depression is clearly correlated with cheek size. There is thus “a growing body of evidence” that changes in facial shape are, at least in part, secondary allometric consequences of reduction in overall size during the evolution of Homo. Maddux and Franciscus (2009) caution against treating features such as an inflated maxilla or a canine fossa as discrete phylogenetic traits. A “puffy” maxilla may not be a Neanderthal apomorphy, and the “canine fossa” of later humans may be a result of decreasing facial size.


If the Broken Hill and Petralona midfacial contours are indeed nearly coincident, and if differences in orbit shape, nasal aperture size, and palatal proportions are taken as indications of the variation to be expected within (all) hominoid populations, then there is little basis in facial form for distinguishing these mid-Pleistocene individuals.

Both the African and the European crania seem to approach Neanderthals in flatness of the infraorbital profile and shape of the maxilla, but neither conforms fully to the Neanderthal condition. Harvati et al. (2010) are inclined to view the Middle Pleistocene morphology as plesiomorphic. But size and scaling must also be considered. It is probable that Broken Hill and Petralona share with Neanderthals (and some Upper Paleolithic humans) a relatively flat infraorbital topography because they have larger faces than recent Homo sapiens (Maddux and Franciscus 2009).

These findings can be read to show that Petralona does not evince true Neanderthal apomorphies in the midface.

Neither Petralona nor Arago can be linked more closely to later European populations than can the African mid-Pleistocene hominins. At the same time, Petralona and Broken Hill share many aspects of facial form, vault proportions, and discrete anatomy. The same conclusion can be drawn from studies of the skull base. It follows that the initial null hypothesis cannot be rejected. The fossils represent paleodemes of a single evolutionary lineage widely dispersed across Africa and Europe. Just how this lineage is related to the Neanderthals, and when the latter emerged as a distinct species, are key questions that remain unresolved.  But it is likely that European and African populations of Homo heidelbergensis did not separate until relatively late in the Middle Pleistocene.

No comments: