Baab 2016 Herectusneurocranialshape

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/283477977
Defining Homo erectus
Article · December 2015

DOI: 10.1007/978-3-642-39979-4_65
CITATIONS READS
9 6,281
1 author:
Karen L Baab
Midwestern University
72 PUBLICATIONS 764 CITATIONS
SEE PROFILE
All content following this page was uploaded by Karen L Baab on 25 May 2016.
The user has requested enhancement of the downloaded file.

Journal of Human Evolution 92 (2016) 1e21
Contents lists available at ScienceDirect
Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol
The role of neurocranial shape in defining the boundaries of an

expanded Homo erectus hypodigm
Karen L. Baab
Department of Anatomy, Arizona College of Osteopathic Medicine, Midwestern University, Glendale, AZ 85308, USA
a r t i c l e i n f o a b s t r a c t
Article history: The main goals of this study were to evaluate the distinctiveness of Homo erectus neurocranial shape
Received 13 December 2014 relative to other closely related species, and assess the likelihood that particular fossils were correctly
Accepted 6 November 2015 attributed to H. erectus given how shape variation related to geography, time and brain size. This was
Available online xxx
accomplished through analyses of several sets of landmarks designed to maximize the fossil sample,
including 24 putative H. erectus fossils. The question of taxonomic differentiation was initially assessed
Keywords:
for the type specimen (Trinil II) and morphologically similar Sangiran fossils and subsequently for
Geometric morphometrics
increasingly inclusive definitions of H. erectus. Results indicated that H. erectus fossils from China,
Homo heidelbergensis
Cranial morphology
Indonesia, Georgia and East Africa shared a neurocranial shape that was distinct from that of other Plio-
Dmanisi Pleistocene Homo taxa, a pattern only partially accounted for by brain size. Early Indonesian H. erectus
Daka formed a morphological “bridge” between earlier and later populations assigned to H. erectus from Africa
KNM-ER 42700 and Asia, respectively. These results were combined with discrete characters to create a more complete
species definition for H. erectus. There were two notable exceptions to the general pattern of H. erectus
uniqueness. The 0.8e1.0 Ma (millions of years ago) Daka calvaria from Ethiopia consistently grouped
with mid-Pleistocene Homo, including Bodo and Kabwe, rather than African or Asian H. erectus. In
addition, Daka also exhibited several traits derived for mid-Pleistocene Homo, and its scaling pattern
mirrored mid-Pleistocene Homo rather than H. erectus. Daka may have belonged to an “advanced”
H. erectus population close to the root of Homo heidelbergensis sensu lato (s.l.), or to an early population of
H. heidelbergensis s.l.. The 1.5 Ma KNM-ER 42700 specimen from Kenya exhibited a unique calvarial shape
distinct from H. erectus despite the exclusion of problematic landmarks from the frontal bone. These
unique aspects of shape were not present in two other subadult fossils, KNM-WT 15000 and D2700.
© 2015 Elsevier Ltd. All rights reserved.
1. Introduction single H. erectus species that includes fossils from Indonesia

(including the younger Ngandong, Sambungmacan, and Ngawi
The taxonomic integrity and composition of Homo erectus has fossils), China, East Africa, and possibly South Africa (e.g.,
been the subject of continuous debate since the erection of Pith- Rightmire, 1990; Anto n, 2003; Baab, 2008b), or restrict H. erectus to
ecanthropus erectus by Dubois (1894). The most restricted concep- the Asian fossils and assign the African and probably the Georgian
tion of H. erectus limits the species to fossils from Trinil and fossils to the less derived Homo ergaster (e.g., Wood, 1994; Rosas
Sangiran on the island of Java (Schwartz, 2004), while the most and Bermudez De Castro, 1998; Vekua et al., 2002). This plethora
inclusive definitions incorporate fossils typically assigned to the of views reflects different species definitions and interpretations of
species from Asia, Africa and Georgia, as well as fossils commonly the evolutionary dynamics and relationships within Homo as well
assigned to earlier Homo species (namely Homo habilis and Homo as the complexity of operationalizing species definitions in the
rudolfensis; Lordkipanidze et al., 2013). It has even been suggested paleontological record, all deeply entrenched systematics issues
that H. erectus is not a distinct taxon, and should be sunk into Homo that are unlikely to be resolved by any single analysis (for additional
sapiens (Wolpoff et al., 1994). Most workers ascribe to either a reviews of H. erectus systematics, see Dunsworth and Walker, 2002;
Anton, 2003; Baab, 2014). Yet, some of the debate also relates to
whether H. erectus is a distinct taxon that can be defined relative to
other closely related and morphologically similar species, and how
E-mail address: kbaab@midwestern.edu.
http://dx.doi.org/10.1016/j.jhevol.2015.11.004
0047-2484/© 2015 Elsevier Ltd. All rights reserved.
2 K.L. Baab / Journal of Human Evolution 92 (2016) 1e21
individual fossils or fossil samples relate to one another (based on There are, however, disadvantages inherent in a GM approach to
patterns of variation and covariation). defining species and assessing taxonomic assignments, and the
Several studies have addressed the degree of variation within the application of shape information to alpha taxonomy has been
broadly defined H. erectus hypodigm or the degree of difference criticized (Wood quoted in Switek, 2013). By evaluating overall
among subsets of this sample corresponding to temporal/ cranial shape rather than individual aspects of shape, incomplete
geographic groupings or proposed species boundaries (Kidder and fossils are often excluded from the analysis, thus reducing sample
Durband, 2000, 2004; Villmoare, 2005; Terhune et al., 2007; Baab, sizes. Also, by focusing on principal components of shape variation,
2008b). The results mostly (but not exclusively) support a single this study is biased in favor of large scale shape features rather than
species interpretation, but it was clear that the pattern of within small-scale differences in localized morphological structures.
sample variation is complex. This study contributes to the current Moreover, shape analyses are ill-suited for evaluating truly discrete
debate by: (1) evaluating whether neurocranial shape alone can be characters that are present or absent because capturing discon-
used to define and differentiate H. erectus, and (2) assessing how tinuous variation using a set of landmarks that must correspond
different samples/specimens assigned to H. erectus relate to each across taxa remains challenging (Bookstein, 2002; Klingenberg,
other and to other taxa in terms of vault shape. A related goal is to 2008; Polly, 2008; Go mez-Robles et al., 2011). Finally, similarity
describe and visualize shape differences between H. erectus and in shape reflects a variety of factors that may overprint a species-
other Plio-Pleistocene Homo taxa. The first goal is addressed by specific pattern, including scaling effects and sexual dimorphism.
examining increasingly more inclusive definitions of H. erectus sensu Therefore, the information obtained from shape analysis should be
lato (s.l.), beginning with just the geochronologically older Asian married with information from other datasets to form a fuller
fossils and ending with the most recent additions to the hypodigm picture of the alpha taxonomy of H. erectus s.l..
from the Caucasus and East Africa. It is important to evaluate if these To address the aims of this study, a series of comparative ana-
taxonomic assignments are correct as they collectively expand the lyses of cranial shape among Homo species was performed based on
temporal and spatial bounds of H. erectus and impact our under- different subsets of neurobasicranial landmarks. These subsets
standing of the evolutionary history of this species. were designed to include the maximum number of potential
Common sources of intraspecific craniometric variation include H. erectus fossils. Twenty-two fossils assigned to H. erectus were
geography, diachronic evolutionary trends and allometric variation. studied from Koobi Fora/Ileret, Bouri, Olduvai Gorge, Olorgesailie,
The second goal is therefore evaluated by asking whether the Dmanisi, Trinil, Sangiran, Sambungmacan, Ngawi, Ngandong and
pattern of morphological variation accords with the pattern of Zhoukoudian, in addition to representatives of H. habilis,
geographic, temporal and size (specifically endocranial volume) H. rudolfensis, mid-Pleistocene Homo s.l., Homo neanderthalensis,
variation in the sample. In other words, are fossils that are similar in Homo floresiensis, and H. sapiens.
their geographic origin, geochronological age, and brain size more
similar to one another than other fossils that are more distinct for 2. Background
these parameters? Operationally, these patterns can only be
assessed on a broad scale as these three factors are not an exclusive 2.1. Is Homo erectus morphologically distinct?
list of variables impacting intraspecific variation. To a more limited
extent, these patterns should also hold between species, such that A central theme in this debate concerns whether H. erectus is
earlier members of H. erectus will more closely resemble early morphologically distinct from earlier and later Homo species. The
Homo species than later and more derived species. However, the issue of sinking H. erectus into H. sapiens or, conversely, of sinking
magnitude of interspecific differences may be greater, as species earlier Homo species into H. erectus speaks to the fuzzy nature of
may diverge more strongly along these axes. Along these lines, a fossil species boundaries. Additionally, researchers have raised
major divergence from observed geographic, temporal and allo- concerns that H. erectus is a “grade,” morphologically intermediate
metric patterning may indicate a species-level distinction. Impor- between H. habilis and mid-Pleistocene Homo, rather than a “good”
tant insights regarding the taxonomic validity and composition of species with clear morphological boundaries (Andrews, 1984;
H. erectus should emerge from the integration of these results with Stringer, 1984). Other important questions include how divergent
previous studies based on qualitative morphological descriptions. the various subsets of fossils assigned to H. erectus are, and what the
Morphological definitions of H. erectus rely heavily on both significance of these differences is. This study will primarily address
discrete cranial traits and aspects of cranial shape. However, it is the first issue (whether H. erectus is distinct), but will also provide
possible to improve the qualitative descriptions of cranial shape context relevant for addressing the other two questions.
using a geometric morphometric (GM) approach, which is specif- The older African fossils may not exhibit the full suite of H. erectus
ically designed to preserve shape throughout the analysis. This al- traits, which blurs the boundary between early Homo and H. erectus.
lows for cranial shape to be treated as a continuous variable rather There is also overlap between the cranial morphologies of traditional
than a series of discrete traits. Multivariate evaluation of shape H. erectus fossils and mid-Pleistocene Homo fossils from Europe, Af-
quantified by 3D landmarks avoids subjective decisions about rica and Asia (Wolpoff et al., 1994), and descriptions of fossils
which aspects of shape are distinct features by analyzing the shape currently assigned to Homo heidelbergensis s.l. (which itself may
of the cranium as a single, integrated unit. Further, it clarifies how consist of multiple species) often emphasized similarities to
aspects of shape co-vary within or between groups and how this H. erectus (e.g., Ndutu: Clark, 1976; Bodo: Conroy et al., 1978; Omo
patterning may correspond to taxonomic boundaries. Finally, a GM Kibish II: Day and Stringer, 1982). In addition, the larger and possibly
approach may better capture the relatively subtle aspects of shape geochronologically younger Ngandong (Solo) fossils have been var-
that differentiate closely-related Plio-Pleistocene Homo species iably assigned to both H. erectus and “archaic H. sapiens,” suggesting a
(rather than the more visually apparent differences between “fuzzy” boundary between these taxa.
H. erectus and H. sapiens). Many species definitions for H. erectus These issues have continued to plague more recent additions to
focus on differentiating it from modern H. sapiens (Wood, 1984; the H. erectus hypodigm. The attribution of the ~1.8 Ma (millions of
Wolpoff et al., 1994), but including species that abut H. erectus years ago) fossils from Dmanisi, Georgia to H. erectus has been
closely in time and that are more similar morphologically than challenged on the basis of morphological features shared with
H. sapiens is a more conservative test of whether H. erectus cranial H. habilis (Rosas and Bermudez De Castro, 1998; Gabunia et al.,
shape is unique. 2000; Martino n-Torres et al., 2008; Wood, 2011). At the other end
K.L. Baab / Journal of Human Evolution 92 (2016) 1e21 3
of the spectrum, fossils recently attributed to H. erectus from the Weidenreich (1943), Le Gros Clark (1955), Howell (1978), Rightmire
Dakanihylo (“Daka”) member of the Bouri Formation in the Middle (1990) and Anto n (2003), with particular attention paid to the
Awash of Ethiopia (BOU-VP-2/66) and especially the Dandiero diagnostic value of traits with regard to differentiating H. erectus
(Buia) Basin in the Danakil Depression of Eritrea (UA-31) present from H. habilis and H. heidelbergensis s.l.. The emphasis is on fea-
traits seemingly more derived than those seen in H. erectus (Anto n, tures affecting neurobasicranial shape. More localized (discrete)
2003). For the Buia cranium (~1.0 Ma), which was provisionally traits often included in H. erectus species definitions are summa-
assigned to H. erectus, these included vertical lateral cranial walls, a rized in Table 1. There is significant variation in the expression of
high height: breadth ratio for the neurocranium, no frontal keel, these features across and even within putative H. erectus sites, and
massive supraorbital tori and thinner parietal bones (Abbate et al., many of these traits are primitive for the genus Homo or are shared
1998; Macchiarelli et al., 2004). The initial description of the derived characters with later Homo species. Very few features, if
~1.0 Ma Daka cranium emphasizes the resemblance of this fossil to any, are clearly diagnostic for H. erectus, but many may contribute
the Buia fossil (Asfaw et al., 2002), and it too has parallel vault walls to a combination definition for the species.
and other derived features such as an arched temporal squama, The earliest descriptions of H. erectus crania were those of Dubois
double-arched browridges, lack of a transverse occipital torus, and (1894) and Black (1929,1931). Dubois' (1894) description of the Trinil
a longer upper than lower scale of the occiput. Similarly, some calotte (type specimen of Pithecanthrous [Homo] erectus) focused in
aspects of the Sambungmacan 3 (Sm 3) cranial morphology, large part on comparisons with apes and modern humans, with only
including its more globular neurocranium, steeper frontal bone and a brief reference to the Neanderthal and Spy skulls. He noted several
less angled occipital, are atypical of classic H. erectus morphology features which remain relevant to defining H. erectus: elongated
and could represent additional taxonomic diversity in Java (Delson vault that is higher than that of chimpanzees (Pan) in lateral view,
et al., 2001; Marquez et al., 2001). There is no definitive geological sharply angled occipital bone with a strongly inclined nuchal plane, a
context for Sm 3, but is suggested to be of Middle Pleistocene age transverse occipital torus, supraorbital tori (superciliary arches) less
(Kaifu et al., 2006). developed than in Pan but more than in the typical Hylobates, less
Baab (2008a) raised concerns about the attribution by Spoor post-orbital constriction than in Pan, and more extensive develop-
et al. (2007) of KNM-ER 42700 to H. erectus. These concerns were ment of the nuchal bone inferior to the inferior nuchal line than in
based on the finding that its cranial vault shape fell outside the
bounds of variation observed in other H. erectus fossils, and inter- Table 1
mediate between H. erectus and mid-Pleistocene Homo/H. nean- Features often cited in H. erectus species definitionsa.
derthalensis. Spoor et al. (2008) countered that the calvarial shape Frontal bone
of KNM-ER 42700 was affected by 1) minor plastic deformation Sagittal keel
and/or 2) its older subadult or young adult status. Moreover, they Coronal keel
Sagittal and coronal keels contributing to a “4-sided hump” (cruciate
expressed concerns about the ability of 3D landmarks to distin-
eminence)
guish among Homo taxa. The first of these concerns will be
addressed here by careful landmark selection and “mirroring” Parietal bone
Sagittal keel (often with parasaggital flattening)
landmarks from the undistorted side. It is also worth noting that a
Angular torus
recent analysis confirmed that the shape of the KNM-ER 42700
calvaria falls outside the H. erectus range of variation even after a Occipital bone
Transverse occipital torus
virtual reconstruction removing taphonomic distortion (Bauer and
Vertical separation of inion above endinion
Harvati, 2015). The second concern cannot be addressed entirely in Occipitomastoid crest
this study, but two other immature H. erectus fossils that are likely Lacks a true external occipital protuberance (has only a linear tubercle)
younger than KNM-ER 42700 will also be examined to see whether
Temporal bone
a similar set of shape differences characterizes this sample. The Well-developed mastoid and supramastoid crests (sometimes separated by a
power of 3D landmarks to differentiate among Homo taxa will also supramastoid sulcus)
be examined. A wide digastric fossa
Regardless of species definition, there is an expectation that Inwardly-angled tips of the mastoid processes
Thick tympanic bone
paleontological species can be identified relative to other closely
Less vertically oriented tympanic bone (more prone)
related and potentially morphologically similar species. This Thick petrosal crest with a prominent spine
distinctiveness need not result from the presence of autapomor- Petrosal crest terminates medially in a tubercle (processus supratubarius/
phies, but may instead be due to a unique combination of primitive infratubarius)
No styloid process
and shared derived features e the combination species definition
No postglenoid process
discussed by Wood (1984). In fact, the loss of primitive features and Presence of a processus vaginalis variable
the addition of more derived features (which may be either unique Surface of petrous pyramid smooth
or shared with later species) through time is plausible if H. erectus is Long axis of petrosal bone is more antero-posteriorly angled than in humans
a temporally extensive species. Nevertheless, fossils at both ends of Flat articular tubercle medially
Entoglenoid formed by the tympanic squama (i.e., no sphenoid spine)
the timespan need to hang together on the basis of sufficient shared
Fissure between the tympanic plate and the entoglenoid process medially
morphology as to be a cohesive entity. It is therefore valuable to Mastoid fissure (between mastoid and tympanic plate)
identify if, and in what ways, fossils assigned to H. erectus differ Arborized sigmoid sinus
from both earlier and later Homo species. Various authors have
Sphenoid bone
addressed this question in the past based on qualitative de- No infratemporal crest
scriptions of cranial (and, to a more limited extent, mandibular, Restricted foramen lacerum
dental and postcranial) morphology.
Generalized
Endocranial volume intermediate between apes and humans (both
2.2. Defining and diagnosing Homo erectus absolutely and relatively)
Elevated cranial vault thickness
The following discussion focuses on the features of traditional a
Characters taken from: Dubois (1894), Black (1929, 1931), Weidenreich (1940,
H. erectus species definitions codified by workers such as 1943, 1950), Santa Luca (1980), Schwartz (2004).
apes. Dubois (1894) also described a bregmatic (or cruciate) Indonesian fossils. Moreover, a supraorbital and supraglabellar sul-
eminence (frontal, sagittal and coronal keels that meet at bregma) cus have been described for H. habilis (e.g., KNM-ER 1813: Wood,
and a feature that probably corresponds to an angular torus that is 1991) and are present in H. heidelbergensis fossils (Asfaw et al.,
continuous with the lateral transverse occipital torus. 2008). The generally flat and receding frontal squama is also not
Black (1929, 1931) described the first cranial remains of Sinan- clearly diagnostic relative to earlier or later taxa (Stringer, 1984;
thropus pekinensis, the genus and species name he created for Rightmire, 1988), and frontal curvature indices based on published
fossils now assigned to H. erectus from Zhoukoudian, China. In values evince substantial overlap among Plio-Pleistocene Homo taxa
addition to some of the same features described above, he also (Lordkipanidze et al., 2006; Kaifu et al., 2008; Rightmire, 2008).
noted the flat superior border of the temporal squama, absence of A projecting supraorbital torus is common to all Plio-Pleistocene
postglenoid and styloid processes, a supraorbital sulcus (described Homo and values for supraorbital thickness overlap between
as supraglabellar and more lateral depressions posterior to the H. erectus and later Homo (Rightmire, 2008). Nevertheless, shape
supraorbital torus), greatest cranial breadth across the prominent analyses have recovered differences in the frontal bone generally
supramastoid crests, laterally diverging vault walls inferior to the and supraorbital region in particular between H. erectus and mid-
parietal tuberosities and convergence of the parietals above the Pleistocene Homo (Baab, 2007; Athreya, 2009; Freidline et al.,
squamosal suture, vertical separation between inion and endinion, 2012). Postorbital constriction was viewed by Wood (1984) as a
and a marked petrosal crest ending medially in a rounded tubercle primitive retention shared with early Homo, but was actually
(probably corresponding to Weidenreich's processus supra- reduced in the African and Asian fossils assigned to H. erectus
tubarius/infratubarius; Black, 1931). He also discussed the elevated relative to the early Homo (Lordkipanidze et al., 2006). However,
cranial vault thickness (CVT) of the juvenile Zhoukoudian (Zkd) III there was overlap among the early Homo, putative H. erectus
(but notes that the bones are thinner in Zkd II). He described the (including Dmanisi) and H. heidelbergensis ranges (Lordkipanidze
glenoid fossa as deep and antero-posteriorly short but did not find et al., 2006; Rightmire, 2008). Several of these features may be
anything about the temporomandibular joint especially distinct influenced by brain size, including relative height of the vault,
from the modern human form. Black (1931) found that Zkd III postorbital constriction and brow ridge development.
differed from Kabwe (H. heidelbergensis), Neanderthals and modern The sharply angled occipital bone and longer nuchal than oc-
humans in its particularly low cranial profile and narrow frontal cipital plane differentiate some H. erectus s.l. from both earlier and
bone based on comparative linear dimensions and indices. later Homo, but at least one mid-Pleistocene Homo fossil, Petralona,
Weidenreich (1940, 1943, 1951) expanded on Black's original is also highly angled, and some of the Dmanisi fossils have a more
description of Zhoukoudian H. erectus based on a larger sample, and rounded occipital contour (Lordkipanidze et al., 2006; Rightmire,
compared them to fossils of Indonesian Pithecanthropus. 2008, 2013; Stringer, 2012). The parietal bone has been described
Weidenreich (1951), as well as Oppenoorth (1932a, 1932b, 1937), also as longitudinally flatter and more rectangular (the four borders
described the Ngandong fossils. Features identified by Weidenreich being more similar in length) in outline than in modern humans
(1943) arguably formed the core of most subsequent H. erectus def- but more curved transversely (Weidenreich, 1943). These may be
initions, which have been refined and expanded by many workers, primitive retentions (Andrews, 1984; Stringer, 1984), but published
including: Le Gros Clark (1940, 1955), MacIntosh and Larnach (1972), values for the parietal sagittal arcs and chords (used to calculate the
Jacob (1975), Howell (1978), Howells (1980), Santa Luca (1980), parietal sagittal curvature) vary substantially across authors for the
Andrews (1984), Rightmire (1984, 1990, 1998), Stringer (1984), same fossils, making it difficult to quantitatively assess sagittal
Wood (1984, 1991), Franciscus and Trinkaus (1988), Picq (1990), curvature from these values (Tobias, 1991; Lordkipanidze et al.,
n (2003), and Kaifu et al. (2008).
Wolpoff et al. (1994), Anto 2006; Kaifu et al., 2008; Rightmire, 2008).
The low and elongated vault in norma lateralis may differentiate The low, flat and postero-inferiorly sloping superior border of
H. erectus from early Homo, but conflicting values for the altitudinal the temporal squama was identified by Andrews (1984) as a
index (height/length) complicate this interpretation (Wood, 1984; primitive retention shared with great apes. However, Terhune and
Tobias, 1991; Lordkipanidze et al., 2006). Maximum cranial Deane (2008) found that H. erectus had a temporal squama that was
breadth is usually across the well-developed supramastoid crests in lower relative to its length than in Australopithecus afarensis,
H. erectus (or mastoid processes in some African H. erectus), above H. habilis and later Homo taxa, but taller than in African apes and
which the walls converge superiorly. Weidenreich (1943) further similar to Australopithecus africanus. Temporal squama shape may
describes two inward “bends” in the coronal contour of the mid- therefore be autapomorphic for H. erectus within Homo (see also
vault, one just above the supramastoid crest and one at the parie- Martınez and Arsuaga, 1997), although it is perhaps related to
tal tuberosity. Some early Homo also exhibit maximum breadth overall vault shape (Terhune and Deane, 2008). The glenoid fossa
across the mastoid processes but have relatively vertically aligned was described by Black (1931) as deep and antero-posteriorly short
parietals (Tobias, 1991; Wood, 1991; Villmoare, 2005). Mid- but not substantially different from modern humans. Weidenreich
Pleistocene Homo fossils often have maximum breadth across the (1943), on the other hand, argued that the fossa was deeper relative
supramastoid crests, but the walls of the vault are more vertical to its antero-posterior length than in humans or apes.
(Bra€uer, 2008; Stringer, 2012). Therefore, a well-developed supra- A more acute petrotympanic angle compared to modern
mastoid crest may have originated during the evolution of humans has long been part of the H. erectus species definition
H. erectus, while the degree of medial convergence of the vault (Weidenreich, 1943, 1951; Howell, 1978). This single angle captures
walls could be autapomorphic for the species. variation in the orientation of the tympanic and petrous elements
Black (1931) discussed the narrow frontal bone with a supra- of the temporal bone. Weidenreich (1943) identified a more
glabellar depression and more lateral depressions separating the sagittally oriented petrous pyramid in Asian H. erectus compared to
supraorbital torus and the bilateral frontal tuberosities present on H. sapiens (though both were more coronal than in apes). Dean and
Zhoukoudian Skull III. Weidenreich (1943) described this as a single Wood (1982) confirmed this for African H. erectus by showing that
tuberosity spanning between the temporal lines. In either case, the they had values at the upper end of the H. sapiens range (more
particularly narrow frontal bone and frontal tuberosity are mostly sagittally oriented), but nonetheless overlapped both this species
restricted to the Zhoukoudian fossils. The supraorbital sulcus is and early Homo (see also Tobias, 1991). The African H. erectus values
variably expressed across fossils broadly assigned to H. erectus. The for tympanic angle also overlapped early Homo and H. sapiens in
supraglabellar depression, for example, is not present in most this study. Lahr (1996) argued that a less coronally oriented
Table 2
Samples analyzed in this study.
Table 3
Landmarks and definitions.
Landmark Definition
Inion Where superior nuchal lines merge in midsagittal plane

Lambda Apex of the occipital bone at its junction with the parietals, in the midline
Bregma Posterior border of the frontal bone in the midsagittal plane
Midline post-toral sulcus Minima of concavity on midline post-toral frontal squama
Glabella Anterior-most point on frontal bone in Frankfort horizontal
Nasion Junction of frononasal and internasal sutures
Dacryon Where lacrimo-maxilary suture meets the frontal bone
Supraorbital notch Greatest projection of notch into the orbital space, taken on the medial side of the notch
Orbitale Inferior-most point on margin of orbit
Frontomalare-temporale Where the fronto-zygomatic suture crosses the temporal line
Frontomalare-orbitale Where the fronto-zygomatic suture crosses the inner orbital rim
Mid-torus inferior Inferior margin of superior margin of orbit roughly at the middle of the orbit
Mid-torus superior Superior aspect of supraorbital torus, directly above mid-torus inferior on anterior aspect of torus
Anterior pterion Where coronal suture intersects spheno-frontal or spheno-parietal suture
Porion Uppermost point on the margin of the external auditory meatus
Auriculare Point vertically above the center of the external auditory meatus at the root of the zygomatic process
Frontotemporale Where the temporal line reaches its most anteromedial position on the frontal
Parietal notch Junction of parietomastoid and squamosal sutures
Asterion Point where the temporal line reaches its most antero-medial position on the frontal
Opisthion Midline point at the posterior margin of the foramen magnum
Tympanomastoid junction Where tympanic tube and mastoid fissure meet laterally
Stylomastoid foramen Posterior border of sylomastoid foramen
Postglenoid process Infralateral-most point posterior to glenoid fossa and anterior to ectotympanic tube (postglenoid tuberosity or crest)
Inferior entoglenoid Most inferior point on the entoglenoid pyramid
Lateral articular fossa Superior-most point on the lateral margin of the articular eminence
Temporo-sphenoid suture Point where temporo-sphenoid suture passes from squama to cranial base (often on infratemporal crest)
Metopion Midway between glabella and bregma in midline, calculated a posterieri
Mid-temporal squama Point midway between glabella and bregma in midline, calculated a posterieri
Mid-parietal Midway between bregma and lambda in midline, calculated a posterieri
tympanic bone distinguished Ngandong H. erectus from H. sapiens. and historical considerations as indicated by the dashed lines in
In contrast, Martınez and Arsuaga (1997) described H. erectus (as Table 2. An analysis of the frontal bone performed to investigate the
well as early Homo) as having more coronal tympanic bones fragmentary Olorgesailie fossil, KNM-OL 45500, is presented in
compared to their modern human and African (and some of their Supplementary Online Material (SOM).
Eurasian) mid-Pleistocene Homo sample. Additional analysis using The majority of fossil specimens are adults, with the exception
a single set of measurements and more thorough sampling of fossils of D2700 from the Eurasian H. erectus site of Dmanisi, KNM-ER
is necessary to establish the utility of this trait in delineating among 42700 from Ileret, KNM-WT 15000 from West Turkana, and
fossil Homo species. possibly Zkd 3. The D2700 subadult fossil has an unfused spheno-
occipital synchondrosis and M3s that are erupted but not in oc-
clusion (Vekua et al., 2002; Rightmire et al., 2006). Previous in-
3. Material and methods
vestigations demonstrated that its cranial shape was within the
range of other H. erectus (Baab and McNulty, 2009) and its neuro-
3.1. Samples
cranium was within the range of the expected shapes for an adult
H. erectus of the equivalent size (Baab, 2008b). KNM-ER 42700 was
The analyses presented here include fossils usually assigned to
described as a subadult or young adult based on the spheno-
the following species: H. sapiens, H. neanderthalensis, H. erectus s.l.,
occipital synchondrosis being two-thirds fused (Spoor et al.,
H. floresiensis, H. habilis, and H. rudolfensis. Mid-Pleistocene Homo
2007). Age estimates for KNM-WT 15000 range from 7.5 to 15
fossils were also examined, which may belong to H. heidelbergensis
years of age based on dental and epiphyseal development (as
s.l., or to several species, such as H. heidelbergensis and Homo rho-
reviewed by Graves et al., 2010). The adult status of Zkd 3 is un-
desiensis or even the enigmatic “Denisovans.” They will be referred n,
certain (Black, 1929, 1931; Weidenreich, 1943; Mann, 1971; Anto
to here as mid-Pleistocene Homo for simplicity. Original fossils
2001), but previous geometric morphometric analysis of
were examined when available and casts when the originals were
Asian H. erectus indicated that this specimen fits well within the
not accessible, as summarized in Table 2.
range of cranial morphology exhibited by other Zhoukoudian fossils
The fossils were subjected to several analyses designed to
(Baab, 2010).
address two questions: (1) Is the calotte/calvarial shape of H. erectus
distinct from other closely related Homo species? (2) How do
́
different fossils attributed to H. erectus expand the magnitude and 3.2. Landmark acquisition and superimposition
direction of variation in this group? These questions were
addressed through interspecific comparisons of increasingly in- Three-dimensional cranial landmarks were acquired using a
clusive samples of fossils attributed to H. erectus. The most Microscribe 3D digitizer. The full landmark protocol has been
restricted definition of H. erectus examined here was that consisting described elsewhere (Baab, 2007), and different subsets of this
of just the Trinil and Sangiran fossils, because Trinil II is the type protocol have been analyzed previously (Baab and McNulty, 2009;
specimen of H. erectus and the Sangiran fossils are the most com- Baab, 2010; Baab et al., 2010, 2013). Landmark definitions are in
parable in their geography, geochronology and, possibly, Table 3. Six separate landmark superimpositions were performed
morphology (cf. Schwartz, 2004). Subsequently, individual fossils by generalized Procrustes analysis (Rohlf and Slice, 1990), one for
or fossil samples were added based on spatial, geochronological each of the landmark sets described below.
3.3. Estimation of missing landmarks 3.4. Corrections for distortion in fossils
Bilateral landmarks missing on one side due to damage or A number of the hominin fossils included in the analyses have
distortion of the original were estimated based on the position of its suffered some degree of postmortem taphonomic deformation. In
antimere (if present) using reflected relabeling (Gunz et al., 2009). order to minimize the effects of this distortion, several strategies
In a few cases, midline landmarks were estimated based on sur- were employed. For one, landmarks were excluded from regions of
rounding morphology or by reference to closely related specimens localized displacement and reflected from the opposite side if
that preserved a similar morphology in the surrounding (local) preserved. Examples of this include KNM-ER 1813 where the left
region. Details of this approach have been described previously for supraorbital torus landmarks were reflected from the right side,
Kabwe (Broken Hill), Zkd 5, and KNM-ER 1813 (see Baab and KNM-WT 15000 where the right lateral supraorbital torus land-
McNulty, 2009 for more details), and are described here for the marks as well as right parietal notch and asterion were reflected
Ceprano and Petralona fossils. from the left side, and KNM-ER 3883 where the left basicranial
The Ceprano calvaria is a crucial fossil to include in the landmarks and left porion, parietal notch, and asterion were
comparative analyses due to its mix of primitive, H. erectus-like mirrored from the right side (see Wood [1991] for discussion of
features and more derived traits. To do so, the mid-parietal land- distortion in this fossil). Asfaw et al. (2008) document relatively
mark was estimated for Ceprano based on the Zkd 5 morphology minor distortion in the Daka fossil, most of which was localized
because both fossils presented a similar curvature of the remaining rather than systemic. The landmarks most likely to be affected were
portion of the midsagittal suture and the superior median plane of the right fronto-temporale, fronto-orbitale, and the superior and
the occipital bone after superimposition. The mid-parietal land- inferior torus landmarks. No specific corrections were made
mark was positioned midway between bregma and lambda as because this distortion was judged unlikely to affect landmark
identified in Ceprano. It is unfortunate that Zkd 5 is the closest placement to a great degree.
anatomical match as this specimen has plaster bridging a gap that The mis-alignment of the frontal bone in KNM-ER 42700 (Spoor
begins slightly before bregma and extends to ~50% of the length of et al., 2008) appears to have mainly affected the right side and
the parietal in the median plane. Therefore, the minimum land- midline of the frontal bone (see also Bauer and Harvati, 2015).
mark analyses were also re-run without the mid-parietal landmark Therefore, bilateral frontal bone landmarks (e.g., frontomalare
to assess the effect of including this landmark in the analysis. orbitale, frontomalare temporale, frontotemporale, anterior pte-
To further increase the sample size for mid-Pleistocene Homo, rion) were reflected from the left while the midline landmarks
the tympanomastoid landmark was estimated for Petralona. (glabella, post-toral sulcus) were excluded from analysis. The right
Because there was no curve that incorporated this landmark, side of the cranial base is also offset more anteriorly than the left
tympanomastoid was estimated as the average position of this side, so the postglenoid process, temporosphenoid, and tympano-
landmark in 22 Plio-Pleistocene Homo crania. Using the average mastoid on the right were also reflected from the left.
position for all Homo avoided biasing the analysis in the direction of
any particular taxon, but this method of landmark estimation was 3.5. Landmark sets
judged inferior to those approximations discussed above because it
was not possible to use the original morphology of Petralona as a The first landmark set was restricted to the calotte in order to
guide. allow for the inclusion of the type specimen of H. erectus, Trinil II
Table 4
Composition of landmark sets.
Landmark Analysis
Trinil Maximum Fossils Maximum Landmarks OH 9 42700 Bodo
Inion X X X X X
Lambda X X X X
Bregma X X X X X
Midline post-toral sulcus X X X X X
Glabella X X X X
Nasion X X
Dacryon X
Supraorbital notch X X X X X
Frontomalare-temporale X X X X X
Frontomalare-orbitale X X X X X
Mid-torus inferior X X X X
Mid-torus superior X X X X
Anterior pterion X X X X
Porion X X X X
Auriculare X X X X X
Frontotemporale X X X X X X
Parietal notch X X X X
Asterion X X X X
Opisthion X X X
Tympanomastoid junction X X X X
Stylomastoid foramen X
Postglenoid process X X X X X
Inferior entoglenoid X X X X
Lateral articular fossa X
Temporo-sphenoid suture X X X X
Metopion X X X
Mid-temporal squama X X X X
Mid-parietal X X X X
(Trinil analysis; Table 4) (Dubois, 1894). The first step in this anal- Although PC axes are statistical constructs that are not a priori
ysis was to assess whether the particular set of landmarks pre- biologically meaningful, it is probable that the most variable as-
served on Trinil II was sufficient to delineate a H. erectus calotte pects of morphology in an interspecific sample of primarily adult
shape that is distinct from other Homo species, and whether San- specimens will align with taxonomic differences. However,
giran fossils share the Trinil morphology (as suggested by Schwartz, between-group principal components analyses (BG-PCA), which
2004). If so, the more complete S 17 and S 2 fossils can be used as emphasize differences among groups rather than among in-
proxies for the morphology of the type specimen in subsequent dividuals, were also performed. In a BG-PCA, only the group aver-
analyses. The Trinil analysis includes a large sample of putative ages are used to calculate PCs, and the full sample is then projected
H. erectus fossils (Table 2), but the restricted landmark set can onto these components (axes; Mitteroecker and Bookstein, 2011).
provide only limited information about the H. erectus cranial Between group-PCA was chosen in preference to canonical variates
morphology. Thus, two additional analyses allow for comparison of analysis as this is more appropriate for analyses involving small
overall calvarial shape. The first of these maximizes the fossil samples and large numbers of variables (Mitteroecker and
sample size while the second maximizes the density of landmark Bookstein, 2011). Two BG-PCAs were performed for the 42700
coverage and therefore the amount of morphology captured and Bodo landmark sets which contained a good balance between
(termed the maximum fossils analysis and the maximum land- sample sizes, landmark coverage and focal fossils. One BG-PCA was
marks analysis, respectively; Tables 2 and 4). based on five a priori groups: early Homo (H. habilis and
Two landmark sets were designed to include key H. erectus or H. rudolfensis), Neanderthals, mid-Pleistocene Homo, African/
potential H. erectus fossils that did not preserve the landmarks in Georgian H. erectus and Asian H. erectus (see Table 1). The second
the maximum fossils analysis. One set of landmarks was designed BG-PCA was the same, except the two early Homo species were
to include the OH 9 fossil (OH 9 analysis; Tables 2 and 4). OH 9 is separated (only relevant to the Bodo analysis) and fossils assigned
often viewed as the African fossil with the greatest resemblance to to H. erectus were divided into the following temporo-geographic
the Asian H. erectus sample and was also included by Asfaw et al. groups: early Indonesian, late Indonesian, Chinese, African, and
(2002) and Gilbert et al. (2008) in the same paleodeme as the Georgian. Daka and KNM-ER 42700 were not categorized a priori in
Daka fossil for their cladistic analyses. The second was restricted to either of these analyses and were not included in the calculation of
landmarks present on the KNM-ER 42700 fossil from Kenya (42700 the axes.
analysis; Tables 2 and 4), which also allowed for the inclusion of Daka and KNM-ER 42700 were not assigned to a group a priori
KNM-WT 15000. This is a significant addition as both fossils are because results from the standard PCAs demonstrated significant
from Kenya, are subadult individuals, and are of a similar morphological differences between these two fossils and the other
geochronological age (KNM-ER 42700: 1.55 Ma; KNM-WT 15000: African fossils assigned to H. erectus. Results were summarized by
1.51e1.56 Ma) (Brown and McDougall, 1993; Spoor et al., 2007). The an unweighted pair-group average (UPGMA) cluster analysis of
final set of landmarks captured the shape of the frontal bone and Euclidean distances based on the four non-zero eigenvectors that
the lateral aspect of the temporal bone to allow for inclusion of the completely described differences among the five groups.
Bodo fossil (Bodo analysis; Tables 2 and 4). Bodo provides a con- Using the same two landmark sets, the pairwise Procrustes
servative test of whether H. erectus and mid-Pleistocene Homo can distances among fossils assigned to H. erectus, mid-Pleistocene
be distinguished on the basis of cranial shape, as it is one of the Homo, and H. neanderthalensis were calculated. The range of
oldest mid-Pleistocene Homo fossils. interspecific values based on distances among fossils assigned to
these three species (but not within each group) was presented.
3.6. Statistical analyses Then, different subsets of intraspecific distances were calculated:
distances within the mid-Pleistocene Homo group (including the
Each set of landmarks was subjected to a standard principal distance between the two Neanderthal fossils), distances among all
components analysis (PCA), which summarizes the main axes of fossils assigned to H. erectus (with the exception of KNM-ER 42700
variation across the entire sample. Both fossils and modern humans and Daka), distances among just the Asian fossils assigned to
were analyzed in the maximum fossils analysis to provide a broad H. erectus and distances among just the African/Georgian fossils
taxonomic context. Only the fossils were analyzed in the other assigned to H. erectus. Distances of the Daka fossil to fossils assigned
analyses as including H. sapiens in the analysis reorients the axes in to H. erectus from Africa and Georgia, to fossils assigned to H. erectus
a way that could mask distinctions among the extinct groups. from Asia and to mid-Pleistocene Homo were also presented.
A convex hull was drawn around the most restrictive (i.e., Finally, distances from KNM-ER 42700 to the African/Georgian
exclusive) definition of H. erectus e the Trinil and Sangiran fossils e fossils assigned to H. erectus were computed.
and was then extended to accommodate each addition to the
hypodigm as outlined in Table 2. In some instances, the fossils 4. Results
added were within the distribution of previously included fossils. In
this case, the convex hull was not extended. A shaded convex hull 4.1. Trinil landmark set: PCA
was also used to delineate the mid-Pleistocene Homo sample.
Principal component (PC) scores were regressed on the natural Part of the H. erectus hypodigm was distinct from other Homo
logarithm of endocranial volume (EV) when axes reflected taxo- species based on the restricted anatomical regions preserved in
nomic variation to explore whether the aspects of shape separating the H. erectus type specimen (Trinil II). The Asian fossils assigned
groups could be attributed to simple scaling effects or whether a to H. erectus, as well as KNM-ER 3733 and KNM-ER 3883 were
species deviated from a common size-shape trajectory. This rep- differentiated from mid-Pleistocene Homo, early Homo (H. habilis
resents a difference from the more common practice of regression and H. rudolfensis) and H. neanderthalensis on PCs 1 and 4 com-
scores on the logarithm of centroid size. Endocranial volume was bined (Fig. 1). The Trinil and Sangiran fossils occupied a central
used because cranial superstructures and cranial vault thickness position within the Asian H. erectus scatter while the Zhoukoudian
affect the centroid size values. However, the natural logarithm of and Turkana Basin fossils were isolated at the positive end of PC 1.
centroid size was used rather than EV for the analysis that included The younger Indonesian and some Dmanisi fossils (D2700 and
modern humans, as EVs were not available for the recent human D3444) extended the range in the opposite direction (the negative
sample. end of PC 1). The former scored high on PC 4, thus separating them
Ng11
0.04
Sm1
Ng6 D2280 PC 4: 11.5%
Ng10 0.02 S17
Sm3
Ng12 Ngaw Trnl Zkd11
Petr
Zkd5
-0.04 Kabw Cepr 0.02
3777 3883 0.06 0.08
D3444
D2700 Zkd12
Dali 1813 Zkd3
-0.02
Daka
SH5
1470
Ferr 42700
-0.04
Chap PC 1: 28.0%
Figure 1. Principal components ordination based on Trinil landmark set. Principal components 1 and 4, which showed the best distinction between H. erectus and other fossil taxa,
are shown. Symbols are as follows, black squares: fossils assigned to Asian H. erectus; gray squares: fossils assigned to African and Georgian H. erectus; circles: mid-Pleistocene
Homo; triangle: early Homo; rectangle: H. neanderthalensis. Convex polygons circumscribe increasingly inclusive definitions of H. erectus as follows, solid black: Trinil and San-
giran fossils; large black dashes: þ Zhoukoudian fossils; small black dashes: þ Ngandong, Sambungmacan and Ngawi fossils; solid gray: þ African fossils; large gray dashes: þ
Dmanisi fossils; small gray dashes: þ Daka fossil (see Table 2). The solid gray convex hull surrounds mid-Pleistocene Homo. Wireframes connecting landmarks are used to show
shape change from the negative to the positive extremes of the axes in left lateral view.
from other species, while the latter scored lower on PC 4 and thus differentiation was roughly orthogonal to the direction of maximal
overlapped the distribution of mid-Pleistocene Homo and intraspecific variation for all groups except Neanderthals, which
H. habilis. The Dmanisi fossils were distinct from these species on were, however, represented by only two specimens.
PC 2. Daka was positioned close to the mid-Pleistocene Homo/ Daka consistently clustered with the mid-Pleistocene Homo and
H. habilis cluster on PCs 1, 2 and 4. KNM-ER 42700 scored lower H. neanderthalensis groups; Sm 3 approached this distribution but
than all H. erectus fossils on PC 4 and was not associated with any did not overlap it. The two Neanderthals scored slightly lower than
of the taxonomic clusters in the subspace spanned by PCs 1 and 4. any other fossils except KNM-ER 1813 (H. habilis) on PC 3 (not
Size did not account for a significant proportion of the variance on figured). PC 4 can also be viewed as an archaic-to-derived trajec-
PC 1 or 4. tory, but only within the fossil taxa. In this context, the shape of the
The second PC strongly contrasts KNM-ER 42700, and, to a lesser D3444 vault was more similar to Asian H. erectus than other
extent, D3444, from the remainder of the sample (SOM Fig. 1). All Dmanisi fossils and Sm 3 and Daka overlapped the mid-Pleistocene
fossils assigned to H. erectus were distinct from H. habilis/ Homo/Neanderthal distributions.
H. rudolfensis on PC 3 except D2280 (Dmanisi) (SOM Fig. 1). KNM-ER Regressions of the first and third components on the logarithm
3733 and D3444 also plotted close to early Homo on this compo- of centroid size were statistically significant, but the amount of
nent. Size accounted for 19% (p ¼ 0.02) and 32% (p < 0.01) of variance explained by size on PC 3 was trivial (R2 ¼ 0.04, p < 0.01).
variance in PC 2 and 3 scores, respectively. The scaling relationship The relationship on PC 1 (R2 ¼ 0.11, p < 0.01) was also weak, and
on PC 2 is strongly influenced by the low score of the small KNM-ER modern humans consistently scored higher than fossil hominins
42700 fossil; the significant relationship between PC 2 score and even when centroid sizes overlapped.
size disappears when it is excluded from the regression. This sug- Excluding the H. sapiens sample resulted in clearer distinction
gests that shape differences between KNM-ER 42700 and other among the extinct groups. H. habilis and most fossils attributed to
Homo species do not follow size-shape trends seen in the genus H. erectus were separated from the more derived Homo taxa on PC 1,
Homo generally. although the later Indonesian H. erectus sample approached the
mid-Pleistocene Homo range (Fig. 3). In contrast, the sole H. habilis
4.2. Maximum fossils landmark set: PCA fossil, KNM-ER 1813, was very distinct from the remainder of the
archaic Homo sample on the second component. The Afro-Georgian
Landmarks included in this analysis captured overall cranial portion of the H. erectus sample was positioned closer to H. habilis
vault shape, with the exception of the nuchal plane of the occipital than was the Asian portion, with most later Indonesian and
bone. When extinct Homo taxa were analyzed alongside recent Zhoukoudian fossils being equidistant from H. habilis. The third
H. sapiens, there was an archaic-to-derived trajectory of cranial component separated the late Indonesian from the Zhoukoudian
shape variation along a combination of PCs 1 and 2 wherein the sole and Turkana Basin fossils, while PC 4 distinguished the mid-
representative of H. habilis, KNM-ER 1813, was contrasted with Pleistocene Homo sample from Neanderthals (SOM Fig. 2). All
modern humans. H. erectus, mid-Pleistocene Homo and Neander- Dmanisi fossils scored low on PC 4.
thals were arrayed between them, with H. erectus positioned The somewhat extreme position of the Zkd 5 cranium on PC 2
closest to H. habilis (Fig. 2). The direction of maximal interspecific could be affected by its reconstruction, as the large frontal bone
Ng6 Zkd5
Zkd3 Kabw
0.03
D2280 S17
D3444 Dali
3733
3883 Chap
Zkd11 Ng12
Zkd12 Sm3
Ng11
-0.12 -0.09 Daka -0.06 -0.03 0.03 0.06
1813 SH5
D2700 Ferr
Cepr -0.03
PC 1: 24.5%
-0.06
Figure 2. Principal components ordination based on maximum fossils landmark set of fossils and recent H. sapiens. Principal components 1 and 4, which showed the best
distinction between H. erectus and other fossil taxa, are shown. Legend same as Figure 1, with the addition of diamonds: H. sapiens. Wireframes connecting landmarks are used to
show shape change from the negative to the positive extremes of PC 1 in left lateral and superior views.
fragment does not directly articulate with the posterior neuro- 4.3. Maximum landmarks landmark set: PCA
cranium (the two pieces were discovered during separate excava-
tions in the 1930s and 1960s [Weidenreich, 1943; Qiu et al., 1973]). The maximum landmarks analysis included an additional three
For example, a rotation of the two elements to create a slightly bilateral and two midline landmarks which more completely
greater height and less receding frontal squama might bring its captured frontal, temporal and occipital morphology. The primary
shape closer to the remainder of the H. erectus sample. However, axis of variation differentiated early Homo from mid-Pleistocene
other elements of shape captured by PC 2, such as the wider mid- Homo and Neanderthals, with most fossils assigned to H. erectus
vault relative to the temporomandibular joint or the greater ante- occupying an intermediate position between these two extremes
roposterior distance between asterion and inion, are unlikely to be but closer to the latter (Fig. 4A). A large proportion of the variance
affected by the reconstruction. in scores along the first axis was attributable to size differences
Twenty eight per cent of variance in PC 1 scores (p ¼ 0.01), (R2 ¼ 75%; p < 0.01) (Fig. 4B). While the mid-Pleistocene Homo
which distinguished between most fossils assigned to H. erectus fossils scored higher than similarly-sized H. erectus, this difference
(except Daka) and later Homo taxa, was accounted for by differ- was negligible for SH 5 (European mid-Pleistocene Homo). Daka
ences in EV. At the same EV, H. erectus consistently scored higher and D3444 had larger positive residuals, while KNM-ER 1813 and
than Neanderthals and particularly mid-Pleistocene Homo. This 3733 had larger negative residuals, indicating that not all early
pattern was most marked for the Zhoukoudian H. erectus and least H. erectus fossils are characterized by a similar scaling relationship.
apparent for D3444 and Sm 3. Daka behaved like the more derived A regression line based on only fossils assigned to H. erectus (with
Homo fossils rather than H. erectus. The second component, which the exception of Daka) further highlights that both earlier and later
separated H. habilis from other Homo fossils, also reflected allo- Homo species differ from the allometric trajectory within this
metric variation (R2 ¼ 0.39, p < 0.01). Superficially, this suggests group.
that some of the differences in shape between H. habilis and The sole Sangiran fossil (S 17) occupied an intermediate position
H. erectus are related to the larger size of the latter. However, along PCs 1 and 2 among fossils assigned to H. erectus (and there-
KNM-ER 1813 (H. habilis) scored much higher than predicted for fore lower on PC 2 than the other species). The other Asian fossils
its size. The two Neanderthals and SH 5 (Sima de los Huesos) to a extended the H. erectus range toward mid-Pleistocene Homo on PC
lesser extent also score higher on PC 2 than predicted by their 1 and even further away from the rest of the Homo sample on PC 2.
brain size. Therefore, a common allometric scaling effect cannot Most African and Georgian fossils extended the range toward
explain the shape differences between H. erectus and other Homo H. habilis, but D3444 extended the H. erectus range much higher on
taxa. PC 2. Daka was positioned very close to the mid-Pleistocene Homo
As discussed in Materials and methods above, the analysis was range on PCs 1 and 2. Size accounted for less than 10% of shape
re-run without the mid-parietal landmark since its estimation in differences on PC 2, but showed a clear distinction between large
Ceprano may be problematic. The same basic patterns emerged, H. erectus and mid-Pleistocene Homo (as well as Daka). D3444 also
with H. habilis scoring very high on PC 2, in particular contrast to scored higher than predicted for its brain size. The third component
Zkd 5, and H. erectus differed from later Homo species along PC 1. (13.2%) further distinguished KNM-ER 1813 from other Homo fossils
The most notable difference was the higher scores on PC 2 for most and separated the single Neanderthal fossil (La Chapelle-aux-
mid-Pleistocene Homo/Neanderthal/Daka fossils, with the excep- Saints) from the mid-Pleistocene Homo sample. Both of these fossils
tion of Ceprano. One result of this shift was to bring the Dali and were most strongly contrasted on this axis from the Dmanisi fossils
Kabwe fossils closer to Ceprano. and S 17. The relationship with size was not statistically significant.
reported for the previous analysis (Fig. 4B), with OH 9 plotting very
close to S 17.
1813
4.5. 42700 landmark set: PCA
0.08
Principal components 1 and 3 together provided the greatest
separation among the taxonomic clusters (Fig. 6) and the first
component again contrasted H. habilis and mid-Pleistocene Homo/
H. neanderthalensis, with most H. erectus arrayed between these
0.06 endpoints. The two Sangiran fossils were not as centrally posi-
PC 2: tioned on PC 1 as in the maximum landmarks or OH 9 analyses due
to their higher scores. The Zhoukoudian and particularly the
15.7% younger Indonesian fossils extended the putative H. erectus range
0.04 toward later Homo on PC 1 and away from other Homo fossils on PC
3. The African and Georgian fossils scored higher on the third
component than the Asian fossils with the exception of Zkd 12.
D2700 3733 D3444 and KNM-WT 15000 overlapped Asian H. erectus on the first
component, while KNM-ER 3733, KNM-ER 3883, D2280 and D2700
0.02
3883 extended the range toward H. habilis on this component. Daka
SH5 overlapped the mid-Pleistocene Homo range on PCs 1e3. KNM-ER
D2280
Ferr Daka D3444 42700 plotted close to mid-Pleistocene Homo in the subspace
Ng12 spanned by PCs 1 and 3, but exhibited fairly extreme scores on PCs
Chap
2, 3 and 4 (especially PC 2). KNM-WT 15000 generally occupied a
-0.04 -0.02 0.02 0.04 more central position, but was at the opposite extreme to KNM-ER
S17 42700 on PC 2 (not figured). This is significant as both of these
Cepr Zkd12 fossils are approximately the same geochronological age and both
Sm3 Zkd11
are immature individuals. Omo 2 overlapped the edge of the Asian
Dali -0.02
H. erectus range. Size accounted for 47% (p < 0.01) of the variance in
Ng11 Zkd3
Kabw Ng6
PC 1 scores, but there was substantial dispersion around the
regression line. In general, fossils assigned to H. erectus had higher
scores than mid-Pleistocene Homo when their EVs overlapped in
-0.04 size. However, D3444, Daka, Sm 3 and, particularly, KNM-ER 42700
had lower scores.
As discussed above, some of the frontal bone landmarks were
excluded from this analysis due to plastic deformation of the frontal
bone. While frontomalare temporale, frontomalare orbitale and
-0.06
frontotemporale were recorded only from the left side, where
Zkd5 distortion appeared minimal (and mirrored to the right side), it
PC 1: 17.4% remains possible that these landmarks were still affected by the
plastic deformation that affected the rest of this region. Thus, the
Figure 3. Principal components ordination based on maximum fossils landmark set of
fossils only. Principal components 1 and 2, which showed the best distinction between PCA was re-run without these three landmarks. The Homo species
H. erectus and other fossil taxa, are shown. Legend same as Figure 1. Wireframes were fairly well separated in the subspace of PCs 1 and 2. Although
connecting landmarks are used to show shape change from the negative to the positive KNM-ER 42700 overlapped both H. erectus and mid-Pleistocene
extremes of the axes in left lateral and superior views.
Homo on PC 1, it was isolated by its uniquely high score on PC 2,
in contrast to LB1 (H. floresiensis; SOM Fig. 3). The two
H. neanderthalensis crania were also more distinct on the first
4.4. OH 9 landmark set: PCA component because the relatively high and wide vault of La
Chapelle-aux-Saints was emphasized by this particular set of
The ordination along the first two PCs of the OH 9 landmarks landmarks.
bore resemblances to that of the maximum landmarks analysis
discussed above despite the absence of midline landmarks from the 4.6. Bodo landmark set: PCA
superior vault. The primary axis of variation contrasts KNM-ER
1813 (H. habilis) and mid-Pleistocene Homo/Neanderthals, with The Bodo analysis included landmarks from the frontal and
most fossils assigned to H. erectus positioned between these ex- temporal bones that addressed the shape of these bones as well as
tremes but closer to the latter (Fig. 5). Again, S 17 was positioned their position and orientation relative to one another. Overall, the
centrally within the scatter of potential H. erectus fossils. Other Bodo analysis resembled the maximum fossils analysis already
Asian fossils scored low on PC 2 and some scored closer to more presented, in that the first axis separated the H. erectus samples and
derived Homo on PC 1. African and Georgian fossils were positioned later Homo, while early Homo (in this case H. habilis and
between S 17 and H. habilis. OH 9 (as well as KNM-ER 3883 and H. rudolfensis) were distinguished along the second axis. Daka
D2280) clustered near to S 17 while KNM-ER 3733 and the other overlapped the mid-Pleistocene Homo range on both axes. It
Dmanisi fossils scored higher on PC 2. Daka was positioned close to differed, however, in that the two Neanderthals were inserted be-
mid-Pleistocene Homo on PCs 1 and 2. OH 9 had the highest score tween H. erectus and mid-Pleistocene Homo on PC 1. S 17 occupied
on PC 3, but was again closest to other African H. erectus (not an intermediate position along PC 2 but had a higher position on PC
shown). Variation along PC 1 was related to overall size (R2 ¼ 0.70; 1 than many putative H. erectus fossils. Most African and Georgian
p < 0.01). The regression analysis was very comparable to that fossils extended the range toward early Homo on PC 2. The Chinese
A
Daka
D3444 SH5 Kabw
0.02
1813
D2700 Dali
3733
-0.08 -0.06 -0.04 D2280 0.02 0.04
3883 S17 Chap
-0.02
PC 2: Ng12
14.9% Zkd11
-0.06 Ng6
PC 1: 22.4%
0.06 Dali
Kabw
Chap
0.04
Daka SH5
0.02 Ng6
Ng12
S17
PC 1 Score
0.00
D3444
D2280
Zkd11
-0.02 3883
D2700
3733
-0.04
-0.06
1813
-0.08
6.2 6.4 6.6 6.8 7.0 7.2 7.4

Logged Endocranial Volume
Figure 4. Principal components ordination based on maximum landmarks landmark set (A) and regression of PC1 scores on logged endocranial volume (B). Principal components 1
and 2, which showed the best distinction between H. erectus and other fossil taxa, are shown. Legend same as Figure 1. Wireframes connecting landmarks are used to show shape
change from the negative to the positive extremes of the axes in left lateral and superior views. The solid regression line is based on the entire sample whereas the dotted regression
line was based on only those fossils assigned to H. erectus with the exception of Daka.
H. erectus were strongly contrasted with later Homo while Ngan- captured mid- and posterior vault morphology and the Bodo
dong/Sambungmacan were more clearly contrasted with early landmark set, which captured frontal and temporal bone
Homo. The latter group did not approach mid-Pleistocene Homo/ morphology. The three sets of trees were based on: 1) scores from
Neanderthals, as they did in other analyses. the standard PCA (i.e., individual variation), 2) scores from a BG-
Size had a moderate influence on the position of fossils along PCA based on eight (42700 landmark set) or nine (Bodo landmark
both axes (PC 1: R2 ¼ 0.17; p ¼ 0.04; PC 2: R2 ¼ 0.34; p < 0.01). On the set) a priori groups, and 3) scores from a BG-PCA based on five a
first axis, mid-Pleistocene Homo and Daka scored well above the priori groups. In the second case, the trees reflect variation among
regression line (and therefore higher than comparably sized taxa as it relates to the shape differences among the a priori
H. erectus) while most Zhoukoudian fossils and Ng 6 scored lower identified groups.
than predicted by this scaling relationship. Indonesian H. erectus and The distances among individuals and clusters increased as more
Dali scored higher-than-predicted for their size on PC 2 while early groups (or single individuals) were the basis of analysis, and fossils
Homo, Bodo and Neanderthals scored particularly low for their size. were more likely to cluster in the “wrong” taxon in the analysis of
In other words, groups occupying the extremes of PC 1 or PC 2 had individuals. H. erectus formed a single cluster in the five-group PCA
large residuals from the size-shape regression line for that axis. (Fig. 8 a,d) but was split into two clusters in the other analyses. The
late Indonesian H. erectus fossils formed a cluster that was more
4.7. UPGMA clustering similar to late Homo than other H. erectus based on the 42700
landmarks (Fig. 8 b,c), perhaps related to allometric trends. All
As a way of examining between group variation, three sets of Indonesian and Georgian fossils assigned to H. erectus form a
UPGMA trees were generated for the 42700 landmark set, which cluster that was more similar to later Homo taxa when the Bodo
PC 1: 20.2%
Daka
0.04
3733 D3444
Kabw
SH5
0.02
D2700
1813 D2280
OH9
-0.08 -0.06 -0.04 -0.02 0.02 0.04 Dali
S17
3883 Zkd12
-0.02
PC 2:
-0.04 Chap
16.3%
Zkd11
-0.06
Ng6
Figure 5. Principal components ordination based on OH 9 landmark set. Principal components 1 and 2, which showed the best distinction between H. erectus and other fossil taxa,
are shown. Legend same as Figure 1. Wireframes connecting landmarks are used to show shape change from the negative to the positive extremes of the axes in left lateral and
superior views.
landmarks were used (Fig. 8 e,f), a pattern unlikely to be related to fossils and S2 with the Zhoukoudian fossils in all analyses except
size. Mid-Pleistocene Homo and the Daka fossil were in the same the 5-group BG-PCA (Fig. 8 b,c, e,f), thus confirming their
cluster, linked to a Neanderthal cluster, in most analyses (Fig. 8). morphologically intermediate position within the larger H. erectus
KNM-ER 42700 also clustered with the former group in the trees s.l. sample. The major division among fossils assigned to H. erectus
based on the BG-PCAs but was on its own branch in the tree based in the 42700 trees was broadly along size-lines, with the exceptions
on individuals (Fig. 8 aec). Many of the temporogeographic clusters of Daka, Sm 3, and KNM-WT 15000 that clustered with the larger
within H. erectus s.l. (i.e., paleodemes) were consistently recovered, Indonesian fossils.
including the later Indonesian fossils, the Georgian fossils and
KNM-ER 3733/3883. Interestingly, S 17 grouped with the Georgian 4.8. Magnitude of shape distances
The patterns of pairwise Procrustes distances based on the

42700 42700 and Bodo landmark sets were similar. The median inter-
1813
Kabw specific Procrustes distance was higher than any of the median
D3444
Ferr SH5 intraspecific values (Fig. 9). The median intraspecific H. erectus
0.02 D2700 value was higher than the median intraspecific value calculated for
Daka
the later Homo species, which was likely driven by the higher
3883
Zkd12 15000 3733 within African/Georgian and between Asian and African/Georgian
Dali
D2280 distances, as the within Asian distances were lower. There was,
-0.04 -0.02 0.02 0.04 0.06 however, extensive overlap in the ranges of inter- and intraspecific
S17 distances.
Chap Omo2 It is apparent that KNM-ER 42700, despite being from the same
Sm3 S2
-0.02 LB1 geographic region and general time period as the African fossils
Sm1 assigned to H. erectus, and despite the inclusion of the KNM-WT
15000 juvenile in the sample, is very distinct from this group.
PC 3: Zkd11
10.6% Ng11 Ng12 The distances of Daka to members of mid-Pleistocene Homo were
-0.04 very low, lower than from Daka to fossils assigned to H. erectus. The
PC 1: 16.5% median distance of Daka to African/Georgian H. erectus was higher
Ng6
than the distance to Asian fossils based on the 42700 landmarks,
-0.06 but the pattern was reversed using the Bodo landmarks. This latter
result is probably due to the more extreme positions of the Zhou-
Figure 6. Principal components ordination based on 42700 landmark set. Principal koudian fossils on PC1 and the Ngandong/Sambungmacan fossils
components 1 and 3, which showed the best distinction between H. erectus and other on PC 2 in the Bodo PCA. This difference presumably related to
fossil taxa, are shown. Legend same as Figure 1, with the addition of the inverted
triangle: H. floresiensis. Wireframes connecting landmarks are used to show shape
differences in landmark composition as the 42700 landmark set
change from the negative to the positive extremes of the axes in left lateral and su- was missing many landmarks from the frontal bone while the Bodo
perior views. analysis was primarily frontal bone landmarks.
and there was some overlap among these groups in the individual
PCAs and UPGMA clustering based on individual variation. How-
Ng11
ever, the separation between these groups was clearer when the
0.06
frontal bone was considered (e.g., the maximum landmarks and
Ng6 Bodo analyses). In all cases, the 1.0e0.8 Ma Daka fossil grouped
PC 1: 19.4% with mid-Pleistocene Homo. Moreover, some of the overlap was
Ng12 0.04 Ng10 due to Omo 2, a fossil assigned here to mid-Pleistocene Homo
Sm3 Dali (Bra€uer, 2008; Rightmire, 2008) but whose taxonomic affinities are
uncertain. McDougall et al. (2005) assigned Omo 2 to early
0.02
H. sapiens based on their inference that it derived from the same
S17 Kabw stratigraphic level as the more modern-looking Omo 1 fossil, and
Zkd11 Zkd5
Ferr Chap was therefore dated to ~195 ka (thousands of years ago). Omo 2
-0.06 -0.04 0.02 0.04 0.06 was, however, a surface find and its archaic morphology has been
D2280 Zkd12 Daka documented in several studies (Friess, 2003; Rightmire, 2008;
Petr
Cepr Gunz et al., 2009). The interpretation of this result is uncertain
3883 D3444 SH5 given its unknown geological age and the fact that Omo 2 was only
Zkd3 D2700 Bodo complete enough to include in one analysis. This result could
3733 indicate morphological overlap between species, the presence of
-0.04
“an archaic, late-surviving lineage, present alongside anatomically
modern humans” (Rightmire, 2008: 12), or that this landmark set
was insufficient to distinguish among taxa.
-0.06 PC 2: 17.1%
1470 H. erectus (with the exception of Daka) differed from mid-
Pleistocene Homo in its more postero-inferiorly angled (i.e., less
1813 vertical) and shorter occipital plane, less superiorly expanded vault,
-0.08
more inferiorly projecting entoglenoid process, longer and lower
temporal squama, greater postorbital constriction and relatively
Figure 7. Principal components ordination based on Bodo landmark set. Principal
components 1 and 2, which showed the best distinction between H. erectus and other narrower vault at mid-temporal squama. The supraorbital torus
fossil taxa, are shown. Legend same as Figure 1. Wireframes connecting landmarks are was also flatter across its superior margin, narrower medio-
used to show shape change from the negative to the positive extremes of the axes in laterally and thinner (supero-inferiorly) at mid-torus. Consistent
left lateral and superior views.
with these observations, characterizations of mid-Pleistocene
Homo frequently include references to its high, arched temporal
5. Discussion squama, laterally expanded and more vertical parietals, a less
angled occipital and strong supraorbital tori (e.g., Rightmire, 2007;
5.1. Distinctiveness of H. erectus cranial shape Br€auer, 2008; Stringer, 2012). However, there is considerable vari-
ation in the expression of these features and overlap between
A major goal of this study was to assess the distinctiveness of H. erectus and mid-Pleistocene Homo in individual traits (Rightmire,
H. erectus on the basis of neurocranial shape. H. erectus differed 2008; Mbua and Bra €uer, 2012). For example, Rightmire (2008)
from modern humans due to the more globular shape of the human observed a number of traits that differed between H. erectus and
cranial vault with a relatively long and vaulted parietal bone, more mid-Pleistocene Homo, but the ranges of the two taxa often over-
vertical frontal squama, diminutive supraorbital torus, minimal lapped (e.g., relative brain size, postorbital narrowing, occipital
constriction behind the orbits, high temporal squama, narrower proportions and angulation and possibly cranial base flexion). The
occipital bone, antero-inferiorly angled occipital plane, wide mid- analyses presented here indicate that overall neurocranial shape
vault, and narrow cranial base. This characterization of the hu- distinguishes between H. erectus and mid-Pleistocene Homo even if
man neurocranium accords well with previous descriptions (Day individual measurements overlap, and this was particularly
and Stringer, 1982; Lieberman et al., 2002; Bruner et al., 2003; apparent in the frontal bone. Compared to Neanderthals, the vault
Trinkaus, 2006; Mounier et al., 2011), but did not capture features was lower and the supraorbital torus was straighter (less arched
such as parietal bossing. over each orbit). The H. erectus vault was also wider posteriorly but
The H. erectus s.l. sample examined here was also consistently more constricted posterior to the orbits.
and widely separated from early Homo on the basis of its relatively
antero-posteriorly longer and vertically shorter cranial vault, long 5.2. Scaling
and flat frontal bone, greater posterior projection of inion, less
inferiorly projecting entoglenoid process and greater breadth of the Differentiation of fossils assigned to H. erectus and to later Homo
neurocranium compared to both the supraorbital torus and the taxa occurred on PC 1 and was always correlated with endocranial
cranial base. Very few studies have addressed the distinction be- volume. This relationship was especially strong (R2 0.70) when
tween the cranial vault of early Homo and H. erectus, although early Homo was also differentiated from putative H. erectus along
Wood (1991) noted the evolution of a longer cranium in H. erectus this axis (in a direction opposite to later Homo). Even when early
(and H. sapiens) relative to early Homo. Other features that separate H. erectus was distinguished from early Homo and later Homo on
the two taxa include localized traits such as an angular torus or a orthogonal axes, endocranial volume accounted for a significant
prominent petrotympanic crest with a process supratubarious, and proportion of variation on both axes. In both cases, most H. erectus
aspects of the facial skeleton (e.g., broader nasal bones and a more (with the exception of D3444) had residuals from the regression
convex lateral malar; Turner and Chamberlain, 1989; Rightmire, line in the opposite direction to mid-Pleistocene Homo. Increased
1990; Anto n, 2004) not captured by the landmarks used here. endocranial volume was related to features including a decrease in
The distinction between fossils assigned to H. erectus and later postorbital constriction, a more robust supraorbital torus, a rela-
Homo, particularly mid-Pleistocene Homo, was also apparent, but tively wider mid-vault, and a more inferiorly positioned posterior
the ranges of the two groups more closely approached each other temporal squama (parietal notch). Therefore, the shape variation
UPGMA trees based on PCA of individuals (no a priori groups)
a) Mid-Pleist. Late Indo. Afr./Georg./ d) Mid-Pleist. Neand.
42700 } 42700
Homo, Nean. H. erectus Early Indo./Chin. Early Afr./Chin. Georg./Indo. Homo + Zkd12, Early
+ Daka + 15000 H. erectus + Omo2 Homo H. erectus H. erectus + Daka Dali Homo
}
}
}
}
}
}
}
}
1813 }
D2280
D2700
D3444
Zkd11
Zkd12
D2700
D3444
D2280
Zkd11
Zkd12
Omo2
15000
Kabw
Kabw
Ng10
Ng12
Ng11
Zkd3
Zkd5
Daka
Bodo
Ng11
Ng12
Chap
Daka
Chap
3733
3883
1470
1813
Cepr
3733
3883
Sm3
SH5
Ng6
Ferr
Sm1
Sm3
Dali
SH5
Ng6
Ferr
Dali
Petr
S17
S17
S2
0.016
0.032
0.048
Euclidean Distance
0.064
0.080
0.096
0.112
0.128
UPGMA trees based on BG-PCA with 8-9 a priori groups

b) Mid-Pleist. Afr./Georg./ e) Mid-Pleist.
Early Late Indo. Homo + Early Indo./Chin. Early Georg./Indo. Homo Afr./Chin.
Homo H. erectus Daka, 42700 Neand. H. erectus Homo H. erectus + Daka Neand. H. erectus
}
}
}
}
}
}
}
}
1813 }
D2280
D2700
D3444
Zkd11
Zkd12
D2280
D3444
D2700
Zkd11
Zkd12
Omo2
42700
15000
Kabw
Daka
Kabw
Daka
Ng11
Ng12
Ng10
Ng12
Ng11
Chap
Zkd3
Zkd5
Chap
3883
3733
1813
1470
3883
3733
Cepr
Sm1
Sm3
SH5
Ng6
Sm3
Ferr
SH5
Dali
Ng6
Ferr
Dali
S17
Petr
S17
S2
0.016
0.032
0.048
Euclidean Distance
0.064
0.080
0.096
0.112
0.128
UPGMA trees based on BG-PCA with 5 a priori groups

c) Mid-Pleist.
H. erectus f) Mid-Pleist.
H. erectus
Homo + Early Homo Early
}
}}
Neand. Daka, 42700 Asian Afr./Georg. Homo Neand. + Daka Asian Afr./Georg. Homo
}
}
}
}
}
}
}
}
}
D2700
D2280
D3444
Zkd12
Zkd11
D3444
D2700
D2280
Zkd12
Zkd11
Omo2
42700
15000
Kabw
Kabw
Daka
Daka
Ng11
Ng10
Ng12
Zkd3
Zkd5
Ng11
Ng12
Chap
3733
3883
Chap
Cepr
1470
3733
3883
1813
Sm3
SH5
Ng6
1813
Sm1
Sm3
SH5
Ferr
Dali
Ng6
Ferr
Dali
Petr
S17
S17
S2
0.016
0.032
0.048
Euclidean Distance
0.064
0.080
0.096
0.112
0.128
Figure 8. UPGMA trees based on scores from standard PCA of individuals (a, d), a BG-PCA with numerous a priori groups (b, e) and a BG-PCA of a small number of a priori groups (c,
f). The trees in the left column (a, b, c) are based the 42700 landmark set while the right column (d, e, f) are based on the Bodo landmarks.
0.16 caused by geographic isolation and/or local adaptation of the

spatially widespread H. erectus populations (Anto n, 2002).
0.14
Included in the later Indonesian fossils were Ngawi and Sm 3,
both more recent additions to the hypodigm. The Ngawi fossil fell
comfortably within the H. erectus range of variation and showed
Procrustes Distance
0.12
affinities to other H. erectus fossils from Ngandong and Sambung-
macan, in line with previous descriptions and analyses of the Ngawi
0.1
fossil (Sartono, 1991; Widianto et al., 2001; Widianto and Zeitoun,
2003; Durband, 2006). Sm 3 generally fell at the edge of the
0.08 H. erectus distribution in the direction of the mid-Pleistocene Homo
sample due to its more globular neurocranium, but was more
0.06 closely aligned with H. erectus in frontal bone shape (Fig. 7; SOM
Fig. 4). Sm 3 was very distinct from H. sapiens in its 3D calvarial
.
.e.
o
er
.e.
.
sia .
.e.
.e.
fic
.e.
f-G H.e
.e
om
nd
a& nH
H
a& gH
H
H
H
ci
H
ea
rg
n
shape in the present study despite having a (2D) midsagittal profile

rg
pe
g
n
i
ak eor
r
P
N
sia
ith
eo
eo
eo
rs
M
A
in
f-G
te
f-G
A
G
ith
In
intermediate between fossil Homo and H. sapiens (Delson et al.,

-
in
A
A
A
W
ith
ak
&
&
&
in
W
&
D
ith
2001) and a median frontal squama profile like that of recent

0
Be ka
n
70
sia
o
n
W
a
om
ee
ee
D
42
nA
tw
tw
H
humans (Bruner et al., 2013). Importantly, it was always near the

n
n
ee
Be
ee
ee
P
M
tw
tw
tw
n
Be
Ngandong, Sm 1 and Ngawi fossils in morphospace (see also Anto

in
Be
Be
ith
W
et al., 2002), and likewise shares many discrete traits with these
Figure 9. Box-and-whisker plots of Procrustes distances. Distances were calculated groups (Delson et al., 2001; Anto n et al., 2002). There is thus no
using the 42700 landmark set (black outlines) and the Bodo landmark set (gray out- compelling evidence from this study to exclude any of these fossils
lines). Medians, first and third quartiles and the minimum and maximum values are from a species that includes Sangiran/Trinil.
presented. H. erectus is abbreviated H.e.
The addition of KNM-ER 3733, 3883 and OH 9 extended the
shape variation toward early Homo and away from both of the later
among taxa and within the H. erectus sample was correlated with Homo species as well as the Zhoukoudian and later Indonesian
differences in endocranial volume, but differently signed residuals fossils assigned to H. erectus. This basic pattern is consistent with
from the regression line in H. erectus and mid-Pleistocene Homo previous work that has emphasized the more primitive and
indicated that changes in endocranial volume are insufficient to generalized nature of early African H. erectus and the more distinct
explain all shape differences between these groups. In other words, and derived morphology of Asian H. erectus, and is also an expected
brain size increases in the H. erectus sample led to different cranial result if early African H. erectus is near the stem of the species. The
shape evolution than did brain size increases in mid-Pleistocene closer resemblance of KNM-ER 3733 and 3883 to early Homo was
Homo. also due in part to their small size. The larger OH 9 fossil was more
similar to S 17, which was similar in size and possibly geochrono-
5.3. Spatiotemporal variation in the H. erectus sample logical age (Larick et al., 2001; but see; Hyodo et al., 2011) despite
its geographic distance. KNM-WT 15000 grouped with H. erectus
Regional and chronological variation in the H. erectus sample but had a higher vault and less projecting occipital bone than other
was apparent in several analyses, but the analysis of Procrustes Turkana Basin H. erectus, presumably related to its immature status.
distances confirmed that intraspecific differences were of a lesser Importantly, the Sangiran/Trinil fossils bridged the gap between
magnitude than interspecific ones. To some extent, intraspecific older but more geographically distant fossils from Africa and more
variation is better assessed in an analysis restricted to H. erectus geographically proximate but geochronologically younger fossils
fossils, as the PC axes here are influenced by interspecific variation. from Asia. The other African fossils considered here (Daka and
What is of interest in this study is how different temporo- KNM-ER 42700) defy the basic patterns just described and are
geographically circumscribed populations assigned to H. erectus discussed subsequently.
relate to one another in the context of interspecific variation. Of the three Georgian fossils considered, D2280 and D2700
The Sangiran fossils are a useful starting point for thinking about generally fell in or near the range of variation of the H. erectus
alpha taxonomy of H. erectus because they were similar to Trinil II sample, and overlapped the Koobi Fora fossils on the first few
(the less complete type specimen), positioned close to the center of components. Despite favorable comparisons with early Homo
the distribution of the fossils assigned to H. erectus, and consistently (Gabounia et al., 2002; de Lumley et al., 2006), these fossils were
differed from the other Homo taxa (see also Schwartz, 2004). Adding more similar in overall vault shape to the Koobi Fora H. erectus, and
the Zhoukoudian fossils to this most conservative definition of on higher components often grouped with Indonesian H. erectus
H. erectus generally expanded the direction of shape variability away (e.g., PC 3 in SOM Fig. 2). Of the three Dmanisi fossils, D2700 was
from other Homo fossils. The late Indonesian fossils (e.g., Ngandong the most similar to early Homo, but was morphologically closer to
and Sambungmacan) further increased shape variation in the group, H. erectus than early Homo. Therefore, although they introduce
usually in the direction of later Homo along PC 1 but further from all additional variation, they behaved as expected for small and
other fossils along PC 2. The greater affinity of later Indonesian geochronologically older members of the species positioned be-
fossils and mid-Pleistocene Homo along PC 1 is due in part to the tween Africa and Asia. Similarities to both African and Asian
larger size of both groups of fossils despite some shape divergence (particularly Sangiran) H. erectus have been noted elsewhere with
between the two groups at overlapping sizes. Portions of the Asian regard to both the crania and mandibles from Dmanisi (Br€ auer and
H. erectus hypodigm, either from later Indonesian sites or from Schultz, 1996; Rosas and Bermudez De Castro, 1998; Rightmire
Zhoukoudian, were often isolated along either PC 1 or 2, confirming et al., 2006).
previous analyses that have highlighted distinctions among the D3444 showed more affinities with mid-Pleistocene Homo than
Asian paleodemes (Anto n, 2002; Kidder and Durband, 2004; Kaifu did the other African/Georgian fossils. This was a result of a more
et al., 2008; Zeitoun et al., 2010). The shape differences that distin- vertical occipital plane, less posteriorly projecting inion (due in part
guished these populations were not clearly related to variation in to its low transverse occipital torus) and a more robust supraorbital
endocranial volume and may instead result from genetic drift torus. Despite these similarities, D3444 clustered with H. erectus,
and specifically with other Georgian fossils, and retained primitive indicator of modern human population history than several other
features lost in more derived Homo, including a low squamosal cranial bones (von Cramon-Taubadel, 2009). The Bodo landmark set
suture, inferiorly projecting entoglenoid process, a narrower mid- included information from only the frontal and temporal bones,
vault, marked postorbital constriction, sagittal keeling on the pa- both good indicators of human population history, and possibly
rietals, and a reduced foramen lacerum (Lordkipanidze et al., 2006). hominin phylogeny more generally (von Cramon-Taubadel, 2009).
When the frontal bone and the mid- and posterior vault were Therefore, the Bodo landmark set may be a better indicator of
evaluated separately (SOM Figs. 3 and 4), D3444 was more clearly population history. In this context, different evolutionary scenarios
associated with H. erectus, suggesting it is the shorter and higher may be invoked to explain the Indonesian/Georgian and the Afri-
vault that drives its convergence on the mid-Pleistocene vault can/Zhoukoudian clusters including ancestoredescendant re-
shape. Thus, both D3444 and Sm 3 approached the condition of lationships and gene flow, but homoplasy cannot be ruled out.
mid-Pleistocene Homo more closely than their geographic con- More explicit population genetic models may help to discern be-
temporaries, but they did so in different ways. In common with tween these scenarios.
later Homo, both had relatively short and tall vaults, but D3444 also Although the median within-group distance for H. erectus was
had a more robust supraorbital torus and proportionately shorter less than the interspecific distance, the ranges overlapped sub-
frontal bone (antero-posteriorly), whereas Sm 3 had greater stantially (Terhune et al., 2007), and it was higher than the intra-
breadth across the mid-vault and less constriction across specific distance for later Homo species. H. erectus also occupied a
frontotemporale. greater region of morphospace than the equally geographically
Taken as a whole and in conjunction with evidence from diverse mid-Pleistocene Homo sample, possibly due to its greater
discrete traits, the Dmanisi sample should probably be included in time depth or stronger population structure or, alternatively,
H. erectus, with the recognition that these fossils expand variation because it contains more than one species. Finally, neurocranial
in H. erectus, and not always in the direction of early Homo. Some shape is only one aspect of anatomy that needs to be considered.
features of D3444 would then be interpreted as individual varia- For example, the vault shape of LB 1 overlapped H. erectus but its
tions that superficially converge on the mid-Pleistocene Homo mandible and dentition differ in meaningful ways from this species
condition. In fact, some of these features are likely related e the (e.g., Brown and Maeda, 2009).
proportionately shorter frontal bone and less projecting inion Therefore, although the early African/Georgian fossils differed
together result in a shorter vault, which, when scaled to the same only subtly from later Asian ones in their vault shape, these results
size as other H. erectus, appears relatively tall. Given that D3444 do not specifically refute the two species solution. If H. ergaster is
and the newly described D4500 are the most robust of the five recognized because discrete traits indicate it is ancestral to later
Dmanisi crania and quite different in shape from one another, it is Homo whereas Asian populations were an evolutionary “dead end”
unlikely that either pattern can be entirely attributed to sexual (e.g, Wood, 1984), then the Georgian sample could be assigned to
variation. H. ergaster based on the extensive overlap in neurocranial shape.
This study was not designed to address the question of multiple Cranial vault shape did not support the idea that some African
species, but some of the results are nevertheless relevant to this fossils represent H. ergaster and others H. erectus s.s. (e.g., OH 9
question. The results described above accord with descriptions of [Wood, 1994]) nor that Daka bridged the two species and thus
H. erectus as a species that shared a “total morphological pattern” blurred this particular species boundary (Asfaw et al., 2002).
(sensu Le Gros Clark, 1959) distinct from other species, with
intraspecific variation in qualitative and metric traits across its 5.4. Consideration of problematic fossils
range (e.g., Weidenreich, 1951; Rightmire, 1990; Anto n, 2003; Kaifu
et al., 2008). The fact that median distance among fossils assigned Two African fossils, Daka and KNM-ER 42700, are more prob-
to H. ergaster (excluding KNM-ER 42700 and Daka) was lower than lematic as members of H. erectus, even broadly defined, and thus
the median interspecific distance is compatible with a single spe- deserve more careful consideration. Including Daka in H. erectus not
cies model. Previous studies have shown that the magnitude of only expands the range of variation encompassed by H. erectus, but
shape variation in H. erectus is within the bounds of some single effectively erases the boundary between H. erectus and later Homo
neontological species despite its greater time depth, but exceeds species such as H. heidelbergensis s.l. or H. rhodesiensis in Africa.
others (Kramer, 1993; Villmoare, 2005; Baab, 2008b), particularly in Daka exhibited a tall cranial vault relative to cranial length, a pro-
the temporal bone (Terhune et al., 2007). Perhaps more important portionally shorter frontal bone with wide, tall and arched supra-
is the fact that the Trinil/Sangiran fossils were equally distant from orbital tori, a less posteriorly projecting inion, and proportionally
other Asian fossils as early African ones sometimes assigned to greater breadth at mid-vault and frontotemporale relative to the
H. ergaster. The relationship among paleodemes of H. erectus cor- posterior vault. Rightmire (2013) recently performed an analysis
responded with geographic, temporal and allometric differences, aimed at assessing patterns of covariation within the cranium and
common sources of intraspecific variation. For example, the early their relationships to endocranial volume and ectocranial di-
and late Indonesian fossils differed from early African/Georgian mensions in H. erectus (including Daka) and mid-Pleistocene Homo.
fossils in the same direction. The greater distance of the latter can Although difficult to compare directly to the results presented here,
be explained by their larger size and/or longer genetic isolation. it appears that while Daka conformed to the H. erectus condition in
This interpretation is not without some problems. UPGMA most aspects, it differed from H. erectus in its relatively tall cranial
clusters based on individual variation failed to uncover a single vault above porion (relative to cranial length and breadth) and its
H. erectus clade, indicating very real variation within this sample. thick supraorbital torus (relative to overall cranial size), in agree-
Interestingly, the UPGMA analysis highlighted different patterns of ment with some of the results presented here.
geographic clustering based on the mid- and posterior vault versus Asfaw et al. (2002) used a cladistic analysis of 22 characters to
the frontal and temporal regions. This could reflect a complex evaluate the position of a paleodeme (sensu Howell, 1999) that
pattern of population history or mosaic evolution of the vault. In consisted of Daka, OH 9, and the Buia cranium from Eritrea. The
neither case were the divisions among populations along tradi- result of the latter analysis (and several other variants presented in
tional H. ergaster e H. erectus sensu stricto (s.s.) lines. The 42700 Gilbert et al. [2008]) did not support separate African and Asian
landmark set was more heavily weighted toward the posterior H. erectus clusters. Together, these results led Asfaw et al. (2002:
cranium, including the occipital bone, which is a less reliable 317) to argue that “the ‘Daka’ calvaria's metric and morphological
attributes centre it firmly within H. erectus. Daka's resemblance to Pleistocene Homo pattern in a population otherwise more similar
Asian counterparts indicates that the early African and Eurasian to H. erectus. Although convergences in cranial shape do occur, the
fossil hominids represent demes of a widespread palaeospecies.” presence of both a more derived cranial shape and cranial non-
While it is true that the cladistic analyses did not support a division metric traits suggests that the resemblance of Daka to mid-
between H. erectus s.s. and H. ergaster, these analyses also did not Pleistocene Homo reflects evolutionary change in that direction
support a separation between H. erectus s.l. and fossils attributed to rather than convergence. Most likely, Daka was a member of an
more derived Homo species and therefore cannot be read as “advanced” population of H. erectus that was ancestral to a later
unambiguously supporting the assignment of Daka to H. erectus. Homo species (e.g., H. heidelbergensis or H. rhodesiensis) and whose
The use of paleodemes may also mask important variation among cranial shape strongly foreshadowed that group or an early mem-
individuals (Manzi et al., 2003), and the analysis may have been ber of a later Homo species that includes individuals like Bodo and
strongly influenced by the inclusion of endocranial volume (Anto n, Kabwe that were studied here, as well as Saldanha and Ndutu that
2003). Clustering based on phenetic distances derived from pres- were too incomplete to include.
ence/absence data for 22 H. erectus and mid-Pleistocene Homo KNM-ER 42700, the 1.5 Ma fossil from Kenya, did not cluster
fossils grouped Daka with the East Turkana H. erectus, but also with other small early Pleistocene fossils from Africa or Georgia.
indicated that Daka was distinct from the typical H. erectus pattern Instead, this fossil often occupied a distinct region of shape space
in the direction of mid-Pleistocene Homo (Manzi et al., 2003). that did not overlap with other hominin taxa (Fig. 6 and SOM Fig. 1),
However, the assessment of character states for some fossils was and was more likely to converge on the more derived Homo species
problematic (e.g., postorbital constriction [Gilbert et al., 2003]). rather than early Homo due to its higher cranial vault with reduced
Although the Daka calvaria undoubtedly shares anatomical constriction behind the supraorbital tori. Issues of distortion in the
features with H. erectus, a number of these are also retained in some landmark set were minimized in this analysis, but still confirmed
members of mid-Pleistocene Homo, including midline keeling of an earlier analysis that did not adjust for distortion in the fossil
the frontal bone, an angular torus and greatest width in the (Baab, 2008a), as well as an analysis that corrected for deformation
supramastoid/mastoid region (Rightmire, 1996, 2008; Mbua and by performing a virtual reconstruction of the calvaria (Bauer and
Bra€uer, 2012). Moreover, some traits seen in Daka, such as more Harvati, 2015). Therefore, deformation is unlikely to explain the
vertical parietal walls with parietal bossing, a vertical occipital position of this fossil outside of the range of H. erectus calvarial
plane, a longer upper than lower scale of the occiput, a sphenoid shape.
spine, and a high arched temporal squama (Asfaw et al., 2008; It remains possible, however, that the seemingly unusual
Gilbert et al., 2008), are typically considered derived for mid- morphology of KNM-ER 42700 is attributable to its immature sta-
Pleistocene Homo relative to H. erectus (Rightmire, 1996, 2007, tus. Two other juvenile/subadult fossils assigned to H. erectus,
2008; Terhune and Deane, 2008; Rightmire, 2009; Stringer, 2012). D2700 and KNM-WT 15000, did not behave similarly to KNM-ER
Although Asfaw et al. (2008) compared the arched condition of the 42700 in this analysis. D2700 generally clustered with other
supraorbital tori in Daka to that seen in KNM-ER 3733, the tori of Georgian fossils. KNM-WT 15000 appeared more similar to Asian
the latter is much less vertically expanded and the superior margin H. erectus than did other African fossils from the Turkana Basin on
is flatter. The tori of Daka more closely resembles the condition in PC 1 (Fig. 6), and was particularly similar in its position to D3444.
certain mid-Pleistocene Homo, including Kabwe, Saldanha, The more rounded vault of KNM-WT 15000 was likely due in part to
Ceprano, Bodo, and Petralona. The presence of derived features in its particularly young age (~8 years of age based on enamel his-
an African fossil that may pre-date the documented time range of tology [Dean et al., 2001]) and minimally developed cranial su-
H. heidelbergensis or H. rhodesiensis by as little as 200 thousand perstructures. The spheno-occipital synchondrosis of KNM-ER
years (Bodo is ~0.6 Ma and Daka was found in deposits 1.0e0.8 Ma) 42700 is 2/3 fused and it was originally described as a young adult
is perhaps not entirely surprising. Yet, analyses of discrete features or late subadult (Spoor et al., 2007). Based on this evidence, KNM-
have come to conflicting conclusions regarding the phylogenetic ER 42700 was older than D2700 (whose synchondrosis was less
position of Daka e Mounier et al. (2011) found that Daka fell within fused), which was itself likely older than KNM-WT 15000 based on
the H. erectus s.l. clade and was most similar to OH 9 and KNM-ER M3 eruption (Rightmire et al., 2006). The immature status of KNM-
3883 based on distances calculated from discrete traits while Argue ER 42700 is an unlikely explanation for the distinctiveness of its
(2015) argued that Daka grouped with mid-Pleistocene Homo, calvarial shape compared to other small early Pleistocene H. erectus
particularly Bodo and Ceprano, based on cladistic analysis of cranial unless the spheno-occipital synchondrosis is a poor indicator of
traits. age. Additional research regarding ontogenetic shape change in
The relatively small endocranial volume of 992e995 cm3 for the H. erectus and the true developmental age of the KNM-ER 42700
Daka calvaria (Asfaw et al., 2002; Gilbert et al., 2008) is lower than fossil could further clarify this issue.
those recorded for mid-Pleistocene Homo, which range from 1100 The original description of KNM-ER 42700 included a multi-
to 1430 cm3 (Rightmire, 2013). However, the Sale fossil from variate analysis of linear dimensions and a comparison of discrete
Morocco, also proposed to be an early member of mid-Pleistocene features among Plio-Pleistocene Homo and this fossil. Specific traits
Homo (Hublin, 1985, 2001), has a comparably small endocranial listed in support of its H. erectus attribution were keeling on the
volume of 930e960 (Jaeger, 1975) or 880 cm3 (Holloway, 1981). The frontal and parietal bones, a medio-laterally narrow mandibular
analyses presented here include most well-preserved H. erectus fossa, coronally oriented tympanic and sagittally oriented petrous
fossils, including fossils with endocranial volumes that overlap the parts of the temporal bone, a short occipital scale, and close
mid-Pleistocene Homo range, yet the Daka calvaria is the only focal approximation of opisthocranion and lambda. As discussed previ-
fossil to consistently group with mid-Pleistocene Homo. A value of ously, additional work is necessary to establish the utility of either
~995 cm3 is plausible if Daka belonged to an early population of the petrous or tympanic angles in differentiating among fossil
H. heidelbergensis s.l. or H. rhodesiensis that was descended from Homo taxa. More problematic is the fact that early African and
African H. erectus, whose documented endocranial volumes range Georgian fossils assigned to H. erectus have previously been
from ~727 to 1067 cm3 (Holloway, 1975, 1978). described as having a “glenoid fossa … wide mediolaterally”
The overlapping calvarial shape shared between Daka and (Anto n, 2003, p. 137) and opisthocranion is nearly coincident with
members of mid-Pleistocene Homo can be interpreted as an inion (not lambda) in H. erectus (e.g., Weidenreich, 1943; Wood,
example of individual variation that converged on the mid- 1984). Therefore, only two (midline keeling and short occipital
scale) or perhaps three (orientation of the petrous and tympanic) of Anto n, S.C., Ma
rquez, S., Mowbray, K., 2002. Sambungmacan 3 and cranial variation
in Asian Homo erectus. J. Hum. Evol. 43, 555e562.
the listed characters support the initial attribution of KNM-ER
Argue, D., 2015. Variation in the Early and Middle Pleistocene: The phylogenetic
42700 to H. erectus. Assuming that KNM-ER 42700 is truly an relationships of Ceprano, Bodo, Daka, Kabwe and Buia. In: Behie, A.M.,
older subadult or young adult, then its calvarial shape argues Oxenham, M.F. (Eds.), Taxonomic Tapestries: The Threads of Evolutionary,
against its inclusion in H. erectus. Combined with the ambiguous Behavioural and Conservation Research. ANU Press, Canberra, pp. 215e248.
Asfaw, B., Gilbert, W.H., Beyene, Y., Hart, W.K., Renne, P.R., WoldeGabriel, G.,
discrete character evidence, it remains “… preferable to assign Vrba, E.S., White, T.D., 2002. Remains of Homo erectus from Bouri, Middle
KNM-ER 42700 to Homo sp. in order to emphasize the uniqueness Awash, Ethiopia. Nature 416, 317e320.
of this fossil's morphological pattern” (Baab, 2008a: 745). Asfaw, B., Gilbert, W.H., Richards, G.D., 2008. Homo erectus cranial anatomy. In:
Gilbert, W.H., Asfaw, B.T. (Eds.), Homo erectus. Pleistocene Evidence from the
Middle Awash, Ethiopia. University of California Press, Berkeley, pp. 265e328.
6. Conclusions Athreya, S., 2009. A comparative study of frontal bone morphology among Pleis-
tocene hominin fossil groups. J. Hum. Evol. 57, 786e804.
Baab, K.L., 2007. Cranial shape variation in Homo erectus. Ph.D. Dissertation, City
H. erectus retained a characteristic cranial form for a period of University of New York.
over one million years and across a geographic range that extended Baab, K.L., 2008a. A re-evaluation of the taxonomic affinities of the early Homo
cranium KNM-ER 42700. J. Hum. Evol. 55, 741e746.
from East Africa to Eurasia and the Far East. This study has
Baab, K.L., 2008b. The taxonomic implications of cranial shape variation in Homo
demonstrated that this cranial shape is in part intermediate be- erectus. J. Hum. Evol. 54, 827e847.
tween earlier Homo species and more derived species such as Baab, K.L., 2010. Cranial shape in Asian Homo erectus: geographic, anagenetic, and
H. heidelbergensis or H. rhodesiensis but is in part also unique, size-related variation. In: Norton, C.J., Braun, D.R. (Eds.), Asian Paleoanthro-
pology: From Africa to China and Beyond. Springer, New York, pp. 57e79.
particularly in the Asian members of the species. Recent additions Baab, K.L., Freidline, S.E., Wang, S.L., Hanson, T., 2010. Relationship of cranial
to the species from Sambungmacan, Ngawi, and Dmanisi expanded robusticity to cranial form, geography and climate in Homo sapiens. Am. J. Phys.
the range of variation but generally showed the greatest affinities to Anthropol. 141.
Baab, K.L., McNulty, K.P., 2009. Size, shape, and asymmetry in fossil hominins: the
other H. erectus fossils that were geographically and temporally status of the LB1 cranium based on 3D morphometric analyses. J. Hum. Evol. 57,
proximate. Additions to H. erectus from Africa did not show 608e622.
particular affinities to H. erectus. The Daka fossil more closely re- Baab, K.L., McNulty, K.P., Harvati, K., 2013. Homo floresiensis contextualized: A
geometric morphometric comparative analysis of fossil and pathological hu-
sembles mid-Pleistocene Homo in overall vault form, and also ex- man samples. PLoS One 8, e69119.
hibits a number of more derived discrete traits with this species, Baab, K., 2014. Defining Homo erectus. In: Henke, W., Tattersall, I. (Eds.), Handbook
while KNM-ER 42700 remains a morphological outlier relative to of Paleoanthropology. Springer, Heidelberg, pp. 1e28.
Bauer, C.C., Harvati, K., 2015. A virtual reconstruction and comparative analysis of
the Plio-Pleistocene Homo fossils analyzed here.
the KNM-ER 42700 cranium. Anat. Anzeiger 72, 129e140.
Black, D., 1929. Preliminary notice of the discovery of an adult Sinanthropus skull at
Acknowledgements Chou Kou Tien. Bull. Geol. Soc. China 8, 15e32.
Black, D., 1931. On an adolescent skull of Sinanthropus pekinensis in comparison
with an adult skull of the same species and with other hominid skulls, recent
I gratefully acknowledge the curators and staff of the American and fossil. Paleont. Sin. New Ser. D 7, 1e144.
Museum of Natural History, Aristotle University of Thessaloniki, Bookstein, F.L., 2002. Creases as morphometric characters. In: MacLeod, N.,
Forey, P.L. (Eds.), Morphology, Shape, and Phylogeny. CRC Press, New York,
ARKANAS, Duckworth Collection at the University of Cambridge, pp. 139e174.
Gadja Mada University, Lembaga Ilmu Pengetahuan Indonesia Br€
auer, G., 2008. The origin of modern anatomy: by speciation or intraspecific
(LIPI), Musee de L'Homme, National Museums of Ethiopia, National evolution? Evol. Anthropol. 17, 22e37.
Br€
auer, G., Schultz, M., 1996. The morphological affinities of the Plio-Pleistocene
Museums of Kenya, National Museum of Tanzania, COSTECH, Nat-
mandible from Dmanisi, Georgia. J. Hum. Evol. 30, 445e481.
ural History Museum (London), Naturalis Museum, Museum of Brown, P., Maeda, T., 2009. Liang Bua Homo floresiensis mandibles and mandibular
Comparative Zoology, Peabody Museum at Harvard University, teeth: A contribution to the comparative morphology of a new hominin species.
J. Hum. Evol. 57, 571e596.
Soprintendenza per i Beni Archeologici del Lazio, University of Cape
Brown, F.H., McDougall, I., 1993. Geologic setting and age, The Nariokotome Homo
Town, and the Institut de Paleontologie Humaine for allowing me erectus Skeleton. Harvard University Press, Cambridge, pp. 9e20.
access to fossil and comparative samples. Three anonymous re- Bruner, E., Manzi, G., Arsuaga, J.L., 2003. Encephalization and allometric trajectories
viewers provided valuable insights that improved this manuscript. in the genus Homo: Evidence from the Neandertal and modern lineages. Proc.
Natl. Acad. Sci. USA 100, 15335e15340.
Grant support was provided by NSF (BCS 04-24262, DGE 03-33415, Clark, R.J., 1976. New cranium of Homo erectus from Lake Ndutu, Tanzania. Nature
and DBI 96-02234), the L.S.B. Leakey Foundation, and the Sigma Xi 262, 485e487.
Foundation. This is NYCEP morphometrics contribution # 94. Conroy, G.C., Jolly, C.J., Cramer, D., Kalb, J.E., 1978. Newly discovered fossil hominid
skull from the Afar depression, Ethiopia. Nature 475, 67e70.
Day, M.H., Stringer, C.B., 1982. A reconsideration of the Omo Kibish remains and the
Supplementary Online Material erectus-sapiens transition. In: de Lumley, M.-A. (Ed.), Homo erectus et la Place de
l'Homme de Tautavel parmi les Hominide s Fossiles, Premiere Congres Inter-
national de Pale ontolgie Humaine. UNESCO Colloque International du Centre
Supplementary online material related to this article can be National de la Recherche Scientifique, Vol. 2. Pre tirages, Nice, pp. 814e846.
found at http://dx.doi.org/10.1016/j.jhevol.2015.11.004. de Lumley, M.-A., Gabounia, L., Vekua, A., Lordkipandize, D., 2006. Les restes
humains du Plioce ne final et du de but du Ple istoce
ne inferieur de Dmanissi,
Ge orgie (1991-2000). I - Les cra ^nes, D 2280, D 2282, D 2700. L'Anthropologie
References 110, 1e110.
Dean, M.C., Wood, B.A., 1982. Basicranial anatomy of Plio-Pleistocene hominids
Abbate, E., Albianelli, A., Azzaroli, A., Benvenuti, M., Tesfamariam, B., Bruni, P., from East and South Africa. Am. J. Phys. Anthropol. 59, 157e174.
Cipriani, N., Clarke, R.J., Ficcarelli, G., Macchiarelli, R., Napoleone, G., Papini, M., Dean, C., Leakey, M.G., Reid, D., Schrenk, F., Schwartz, G.T., Stringer, C., Walker, A.,
Rook, L., Sagri, M., Tecle, T.M., Torre, D., Villa, I., 1998. A one-million-year-old 2001. Growth processes in teeth distinguish modern humans from Homo
Homo cranium from the Danakil (Afar) Depression of Eritrea. Nature 393, erectus and earlier hominins. Nature 414, 628e631.
458e460. Delson, E., Harvati, K., Reddy, D., Marcus, L.F., Mowbray, K., Sawyer, G.J., Jacob, T.,
Andrews, P., 1984. An alternative interpretation of the characters used to define Marquez, S., 2001. The Sambungmacan 3 Homo erectus calvaria: A comparative
Homo erectus. Cour. Foursch. Senck. 69, 167e175. morphometric and morphological analysis. Anat. Rec. 262, 380e397.
n, S.C., 2001. Cranial growth in Homo erectus. In: Minugh-Purvis, N.,
Anto Dubois, E., 1894. Pithecanthropus erectus, Eine Menschenahnliche Uebergangsfom
McNamara, K.J. (Eds.), Human Evolution Through Developmental Change. Johns aus Java. Landes Drucherei, Batavia.
Hopkins University Press, Baltimore, pp. 349e380. Dunsworth, H., Walker, A., 2002. Early genus Homo. In: Hartwig, W.C. (Ed.), The
n, S.C., 2002. Evolutionary significance of cranial variation in Asian Homo
Anto Primate Fossil Record. Cambridge University Press, Cambridge, pp. 419e456.
erectus. Am. J. Phys. Anthropol. 118, 801e828. Durband, A., 2006. Craniometric variation within the Pleistocene of Java: the Ngawi
n, S.C., 2003. Natural history of Homo erectus. Yearb. Phys. Anthropol. 46,
Anto 1 cranium. Hum. Evol. 21, 193e201.
126e170. Franciscus, R.G., Trinkaus, E., 1988. Nasal morphology and the emergence of Homo
n, S.C., 2004. The face of Olduvai Hominid 12. J. Hum. Evol. 46, 335e345.
Anto erectus. Am. J. Phys. Anthropol. 75, 517e527.
Freidline, S.E., Gunz, P., Jankovi

c, I., Harvati, K., Hublin, J.J., 2012. A comprehensive Lordkipanidze, D., Vekua, A., Ferring, R., Rightmire, G.P., Zollikofer, C.P., Ponce de
morphometric analysis of the frontal and zygomatic bone of the Zuttiyeh fossil Leo n, M.S., Agusti, J., Kiladze, G., Mouskhelishvili, A., Nioradze, M., 2006.
from Israel. J. Hum. Evol. 62, 225e241. A fourth hominin skull from Dmanisi, Georgia. Anat. Rec. 288, 1146e1157.
Friess, M., 2003. An application of the relative warps analysis to problems in human Lordkipanidze, D., Ponce de Leo n, M.S., Margvelashvili, A., Rak, Y., Rightmire, G.P.,
paleontology-with notes on raw data quality. Image Anal. Stereol. 22, 63e72. Vekua, A., Zollikofer, C.P.E., 2013. A complete skull from Dmanisi, Georgia, and
Gabounia, L., de Lumley, M.-A., Vekua, A., Lordkipandize, D., de Lumley, H., 2002. the evolutionary biology of early Homo. Science 342, 326e331.
Decouverte d'un nouvel hominide a Dmanissi (Transcaucasie, Georgie). C. R. Macchiarelli, R., Bondioli, L., Chech, M., Coppa, A., Fiore, I., Rezene, R., Vecchi, F.,
Palevol. vol. 1, 243e253. Yosief, L., Rook, L., 2004. The late Early Pleistocene human remains from Buia,
Gabunia, L., Vekua, A., Lordkipanidze, D., Swisher III, C.C., Ferring, R., Justus, A., Danakil Depression, Eritrea. Riv. Ital. Paleontol. S. 110, 133e144.
Nioradze, M., Tvalchrelidze, M., Anto n, S.C., Bosinski, G., 2000. Earliest Pleisto- MacIntosh, N.W.G., Larnach, S.L., 1972. The persistence of Homo erectus traits in
cene hominid cranial remains from Dmanisi, Republic of Georgia: taxonomy, Australian Aboriginal crania. Archaeol. Phys. Anthropol. Ocean 7, 1e7.
geological setting, and age. Science 288, 1019e1025. Mann, A., 1971. Homo erectus. In: Dolhinow, P., Sarich, V. (Eds.), Background for Man.
Gilbert, W.H., White, T.D., Asfaw, B., 2003. Homo erectus, Homo ergaster, Homo Little, Brown, Boston.
“cepranensis,” and the Daka cranium. J. Hum. Evol. 45, 255e259. Manzi, G., Bruner, E., Passarello, P., 2003. The one-million-year-old Homo cranium
Gilbert, W.H., Holloway, R.L., Kubo, D., Kono, R.T., Suwa, G., 2008. Tomographic from Bouri (Ethiopia): a reconsideration of its H. erectus affinities. J. Hum. Evol.
analysis of the Daka calvaria. In: Gilbert, W.H., Asfaw, B. (Eds.), Homo erectus: 44, 731e736.
Pleistocene Evidence from the Middle Awash, Ethiopia. University of California Marquez, S., Mowbray, K., Sawyer, G., Jacob, T., Silvers, A., 2001. New fossil hominid
Press, Berkeley, pp. 329e347. calvaria from IndonesiadSambungmacan 3. Anat. Rec. 262, 344e368.
Gomez-Robles, A., Olejniczak, A.J., Martino n-Torres, M., Prado-Simo n, L., Bermúdez Marítnez, I., Arsuaga, J.L., 1997. The temporal bones from Sima de los Huesos Middle
de Castro, J.M., 2011. Evolutionary novelties and losses in geometric morpho- Pleistocene site (Sierra de Atapuerca, Spain). A phylogenetic approach. J. Hum.
metrics: a practical approach through hominin molar morphology. Evolution Evol. 33, 283e318.
65, 1772e1790. Martino n-Torres, M., Bermúdez de Castro, J.M., Go mez-Robles, A.,
Graves, R.R., Lupo, A.C., McCarthy, R.C., Wescott, D.J., Cunningham, D.L., 2010. Just Margvelashvili, A., Prado, L., Lordkipanidze, D., Vekua, A., 2008. Dental remains
how strapping was KNM-WT 15000? J. Hum. Evol. 59, 542e554. from Dmanisi (Republic of Georgia): Morphological analysis and comparative
Gunz, P., Bookstein, F.L., Mitteroecker, P., Stadlmayr, A., Seidler, H., Weber, G.W., study. J. Hum. Evol. 55, 249e273.
2009. Early modern human diversity suggests subdivided population structure Mbua, E., Bra €uer, G., 2012. Patterns of Middle Pleistocene hominin evolution in
and a complex out-of-Africa scenario. Proc. Natl. Acad. Sci. 106, 6094e6098. Africa and the emergence of modern humans. In: Reynolds, S.C., Gallagher, A.
Holloway, R.L., 1975. Early hominid endocasts: volumes, morphology and signifi- (Eds.), African Genesis: Perspectives on Hominin Evolution. Cambridge Uni-
cance for hominid evolution. In: Tuttle, R. (Ed.), Primate Functional Morphology versity Press, New York, pp. 394e422.
and Evolution. Mouton, The Hague, pp. 393e416. McDougall, I., Brown, F.H., Fleagle, J.G., 2005. Stratigraphic placement and age of
Holloway, R.L., 1978. Problems of brain endocast interpretation and African hominid modern humans from Kibish, Ethiopia. Nature 433, 733e736.
evolution. In: Jolly, C. (Ed.), Early Hominids of Africa. Duckworth, London, Mitteroecker, P., Bookstein, F., 2011. Linear discrimination, ordination, and the
pp. 379e401. visualization of selection gradients in modern morphometrics. Evol. Biol. 38,
Holloway, R.L., 1981. The Indonesian Homo erectus brain endocasts revisited. Am. J. 100e114.
Phys. Anthropol. 55, 503e521. Mounier, A., Condemi, S., Manzi, G., 2011. The stem species of our species: A place
Howell, F.C., 1978. Hominidae. In: Maglio, V.J., Cooke, H.B.S. (Eds.), Evolution of for the archaic human cranium from Ceprano, Italy. PLoS ONE 6, e18821.
African Mammals. Harvard University Press, Cambridge, pp. 154e248. Oppenoorth, W.F.F., 1932a. Ein neuer diluvialer Urmensch von Java. Nat. u. Mus.,
Howell, F.C., 1999. Paleo-demes, species clades, and extinctions in the Pleistocene Frankfort 62, 269e279.
hominin record. J. Anthropol. Res. 55, 191e243. Oppenoorth, W.F.F., 1932b. Homo (Javanthropus) soloensis, een pleistoceene mensch
Howells, W.W., 1980. Homo erectusdwho, when and where: a survey. Yearb. Phys. van Java. Wetesch Mededee. Dienst Mijnbauw Nederl Indie 20, 49e74.
Anthropol. 23, 1e23. Oppenoorth, W.F.F., 1937. The place of Homo soloensis among fossil men. In:
Hublin, J.-J., 1985. Human fossils from the North African Middle Pleistocene and the MacCurdy, G.G. (Ed.), Early Man. J.B. Lipincott, Philadelphia, pp. 349e360.
origin of Homo sapiens. In: Delson, E. (Ed.), Ancestors: The Hard Evidence. Alan Picq, P.G., 1990. L'articulation temporo-mandibulaire des hominide s. CNRS, Paris.
R. Liss, New York, pp. 283e288. Polly, P.D., 2008. Developmental dynamics and G-matrices: Can morphometric
Hublin, J.-J., 2001. Northwestern African Middle Pleistocene hominids and their spaces be used to model phenotypic evolution? Evol. Biol. 35, 83e96.
bearing on the emergence of Homo sapiens. In: Barnham, L., Robson-Brown, K. Qiu, Z.L., Gu, Y.Y., Zhang, Y.Y., Chang, S.S., 1973. Newly discovered Sinanthropus
(Eds.), Human Roots: Africa and Asia in the Middle Pleistocene. Western Aca- remains and stone artifacts at Choukoutien. Vertebrat. Palasiatic. 11, 109e131
demic and Specialist Press Limited, Bristol, pp. 99e121. (In Chinese).
Hyodo, M., Matsu'ura, S., Kamishima, Y., Kondo, M., Takeshita, Y., Kitaba, I., Rightmire, G.P., 1984. Comparisons of Homo erectus from Africa and Southeast Asia.
Danhara, T., Aziz, F., Kurniawan, I., Kumai, H., 2011. High-resolution record of the Cour. Foursch. Senck 69, 83e98.
MatuyamaeBrunhes transition constrains the age of Javanese Homo erectus in Rightmire, G.P., 1988. Homo erectus and later Middle Pleistocene humans. Annu.
the Sangiran dome, Indonesia. Proc. Natl. Acad. Sci. 108, 19563e19568. Rev. Anthropol. 17, 239e259.
Jacob, T., 1975. Morphology and paleoecology of early man in Java. In: Tuttle, R.H. Rightmire, G.P., 1990. The Evolution of Homo erectus: Comparative Anatomical
(Ed.), Paleoanthropology: Morphology and Paleoecology. Mouton, The Hague, Studies of an Extinct Human Species. Cambridge University Press, Cambridge.
pp. 311e325. Rightmire, G.P., 1996. The human cranium from Bodo, Ethiopia: evidence for
couverte d'un cra
Jaeger, J.J., 1975. De ^ne d'hominide dans le Pleistoce ne moyen du speciation in the Middle Pleistocene? J. Hum. Evol. 31, 21e39.
Maroc. In: Sigmon, B.A., Cybulski, J.S. (Eds.), Proble mes Actuels de Rightmire, G.P., 1998. Evidence from facial morphology for similarity of Asian and
Paleontologie-Evolution des Verte bre
s. C.N.R.S, Paris, pp. 897e902. African representatives of Homo erectus. Am. J. Phys. Anthropol. 106, 61e85.
Kaifu, Y., Baba, H., Aziz, F., 2006. Indonesian Homo erectus and modern human or- Rightmire, G.P., 2007. Later Middle Pleistocene Homo. In: Henke, W., Tattersall, I.
igins in Australasia: New evidence from the Sambungmacan Region, Central (Eds.), Handbook of Paleoanthropology. Springer-Verlag, Berlin, pp. 1695e1715.
Java. Natl. Sci. Mus. Monogr. 34, 289e294. Rightmire, G.P., 2008. Homo in the Middle Pleistocene: Hypodigms, variation, and
Kaifu, Y., Aziz, F., Indriati, E., Jacob, T., Kurniawan, I., Baba, H., 2008. Cranial species recognition. Evol. Anthropol. 17, 8e21.
morphology of Javanese Homo erectus: New evidence for continuous evolution, Rightmire, G.P., 2009. Middle and later Pleistocene hominins in Africa and South-
specialization, and terminal extinction. J. Hum. Evol. 55, 551e580. west Asia. Proc. Natl. Acad. Sci. 106, 16046e16050.
Kidder, J.H., Durband, A.C., 2000. The question of speciation in Homo erectus Rightmire, G.P., 2013. Homo erectus and Middle Pleistocene hominins: Brain size,
revisited I: the metric evidence. Am. J. Phys. Anthropol. 30, 194e195. skull form, and species recognition. J. Hum. Evol. 65, 223e252.
Kidder, J.H., Durband, A.C., 2004. A re-evaluation of the metric diversity within Rightmire, G.P., Lordkipanidze, D., Vekua, A., 2006. Anatomical descriptions,
Homo erectus. J. Hum. Evol. 46, 299e315. comparative studies and evolutionary significance of the hominin skulls from
Klingenberg, C.P., 2008. Novelty and “homology-free” morphometrics: What's in a Dmanisi, Republic of Georgia. J. Hum. Evol. 50, 115e141.
name? Evol. Biol. 35, 186e190. Rohlf, F.J., Slice, D., 1990. Extensions of the Procrustes method for the optimal su-
Kramer, A., 1993. Human taxonomic diversity in the Pleistocene: does Homo erectus perimposition of landmarks. System. Zool. 39, 40e59.
represent multiple hominid species? Am. J. Phys. Anthropol. 91, 161e171. Rosas, A., Bermudez De Castro, J.M., 1998. On the taxonomic affinities of the
Lahr, M.M., 1996. The Evolution of Modern Human Diversity: A Study of Cranial Dmanisi mandible (Georgia). Am. J. Phys. Anthropol. 107, 145e162.
Variation. Cambridge University Press, Cambridge. Santa Luca, A., 1980. The Ngandong Fossil Humans: A Comparative Study of a Far
Larick, R., Ciochon, R.L., Zaim, Y., Sudijono, Suminto, Rizal, Y., Aziz, F., Reagan, M., Eastern Homo erectus Group. Yale University Press, New Haven.
Heizler, M., 2001. Early Pleistocene 40Ar/39Ar ages for Bapang Formation Sartono, S., 1991. A new Homo erectus skull from Ngawi, east Java. Indo-Pacific
hominins, Central Jawa, Indonesia. Proc. Natl. Acad. Sci. 98, 4866e4871. Prehistory Association Bulletin 11, 14e22.
Le Gros Clark, W.E., 1940. The relationship between Pithecanthropus and Sinan- Schwartz, J.H., 2004. Getting to know Homo erectus. Science 305, 53e54.
thropus. Nature 145, 70e71. Spoor, F., Leakey, M.G., Gathogo, P.N., Brown, F.H., Anto n, S.C., McDougall, I.,
Le Gros Clark, W.E., 1955. The Fossil Evidence for Human Evolution. University of Kiarie, C., Manthi, F.K., Leakey, L.N., 2007. Implications of new early Homo fossils
Chicago Press, Chicago. from Ileret, east of Lake Turkana, Kenya. Nature 448, 688e691.
Le Gros Clark, W.E., 1959. The crucial evidence for human evolution. Am. Scientist Spoor, F., Leakey, M.G., Anto n, S.C., Leakey, L.N., 2008. The taxonomic status of KNM-
47, 250Ae252A, 299-313. ER 42700: A reply to Baab (2008a). J. Hum. Evol. 55, 747e750.
Lieberman, D.E., McBratney, B.M., Krovitz, G., 2002. The evolution and development Stringer, C.B., 1984. The definition of Homo erectus and the existence of the species
of cranial form in Homo sapiens. Proc. Natl. Acad. Sci. USA 99, 1134e1139. in Africa and Europe. Cour. Foursch. Senck 69, 131e143.
Stringer, C., 2012. The status of Homo heidelbergensis (Schoetensack 1908). Evol. Weidenreich, F., 1940. The torus occipitalis and related structures and their trans-
Anthropol. 21, 101e107. formations in the course of human evolution. Bull. Geolog. Soc. China 19,
Switek, B., Oct. 17, 2013. Beautiful skull spurs debate on human history. National 479e558.
Geographic. http://news.nationalgeographic.com/news/2013/10/131017-skull- Weidenreich, F., 1943. The skull of Sinanthropus pekinensis: a comparative study on
human-origins-dmanisi-georgia-erectus/. Accessed on: Dec. 12, 2014. a primitive hominid skull. Paleont. Sin. New Ser. D 10, 1e298.
Terhune, C.E., Deane, A.S., 2008. Temporal squama shape in fossil hominins: re- Weidenreich, F., 1951. Morphology of Solo man. Anthropol. Pap Am. Mus. 43,
lationships to cranial shape and a determination of character polarity. Am. J. 222e288.
Phys. Anthropol. 137, 397e411. Widianto, H., Grimaud-Herve, D., Sartono, S., 2001. The evolutionary position of the
Terhune, C.E., Kimbel, W.H., Lockwood, C.A., 2007. Variation and diversity in Homo Ngawi calvaria. Indo-Pacific Prehistory Association Bulletin 21, 162e169.
erectus: a 3D geometric morphometric analysis of the temporal bone. J. Hum. Widianto, H., Zeitoun, V., 2003. Morphological description, biometry and phylo-
Evol. 53, 41e60. genetic position of the skull of Ngawi 1 (East Java, Indonesia). Int. J. Osteoarch.
Tobias, P.V., 1991. Olduvai Gorge: The Skulls, Endocasts and Teeth of Homo habilis, 13, 339e351.
Volume 4. Cambridge University Press, Cambridge. Wolpoff, M.H., Thorne, A.G., Jelínek, J., Zhang, Y., 1994. The case for sinking Homo
Trinkaus, E., 2006. Modern human versus Neandertal evolutionary distinctiveness. erectus. 100 years of Pithecanthropus is enough! Cour. Foursch. Senck 171,
Curr. Anthropol. 47, 597e620. 341e361.
Turner, A., Chamberlain, A., 1989. Speciation, morphological change and the status Wood, B.A., 1984. The origin of Homo erectus. Cour. Foursch. Senck 69, 111.
of African Homo erectus. J. Hum. Evol. 18, 115e130. Wood, B.A., 1991. Koobi Fora Research Project: Hominid Cranial Remains. Oxford
Vekua, A., Lordkipanidze, D., Rightmire, G.P., Agusti, J., Ferring, R., Maisuradze, G., University Press, New York.
Mouskhelishvili, A., Nioradze, M., De Leon, M.P., Tappen, M., Tvalchrelidze, M., Wood, B.A., 1994. Taxonomy and evolutionary relationships of Homo erectus. Cour.
Zollikofer, C., 2002. A new skull of early Homo from Dmanisi, Georgia. Science Foursch. Senck 171, 159e165.
297, 85e89. Wood, B., 2011. Did early Homo migrate “out of” or “in to” Africa? Proc. Natl. Acad.
Villmoare, B., 2005. Metric and non-metric randomization methods, geographic Sci. 108, 10375e10376.
variation, and the single-species hypothesis for Asian and African Homo erectus. Zeitoun, V., Detroit, F., Grimaud-Herve
, D., Widianto, H., 2010. Solo man in question:
J. Hum. Evol. 49, 680e701. Convergent views to split Indonesian Homo erectus in two categories. Quatern.
von Cramon-Taubadel, N., 2009. Congruence of individual cranial bone morphology Int. 223e224, 281e292.
and neutral molecular affinity patterns in modern humans. Am. J. Phys.
Anthropol. 140, 205e215.
View publication stats

Baab 2016 Herectusneurocranialshape

Uploaded by

Copyright:

Available Formats

Baab 2016 Herectusneurocranialshape

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Baab 2016 Herectusneurocranialshape

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Deﬁning Homo erectus

Article · December 2015

The user has requested enhancement of the downloaded file.

Contents lists available at ScienceDirect

Journal of Human Evolution

The role of neurocranial shape in deﬁning the boundaries of an

1. Introduction single H. erectus species that includes fossils from Indonesia

Inion Where superior nuchal lines merge in midsagittal plane

3.3. Estimation of missing landmarks 3.4. Corrections for distortion in fossils

Trinil Maximum Fossils Maximum Landmarks OH 9 42700 Bodo

6.2 6.4 6.6 6.8 7.0 7.2 7.4

The patterns of pairwise Procrustes distances based on the

UPGMA trees based on BG-PCA with 8-9 a priori groups

UPGMA trees based on BG-PCA with 5 a priori groups

0.16 caused by geographic isolation and/or local adaptation of the

shape in the present study despite having a (2D) midsagittal proﬁle

intermediate between fossil Homo and H. sapiens (Delson et al.,

2001) and a median frontal squama proﬁle like that of recent

humans (Bruner et al., 2013). Importantly, it was always near the

Ngandong, Sm 1 and Ngawi fossils in morphospace (see also Anto

Freidline, S.E., Gunz, P., Jankovi

View publication stats

You might also like