Journal of Human Evolution 46 (2004) 337–347

The face of Olduvai Hominid 12

Susan C. Antón*
Center for the Study of Human Origins, Department of Anthropology, New York University, 25 Waverly Place, New York,
NY 10003, USA
Received 30 April 2003; accepted 22 December 2003

Abstract

Facial remains of Homo erectus are rare and their scarcity hinders our understanding of the variability and
relationships in this taxon. Previously undescribed fragments of the peri-orbital region and unidentified matches
between fragments of Olduvai Hominid 12 (OH 12) enhance comparison of the African H. erectus hypodigm. The
newly reconstructed upper face and maxilla of OH 12 is most similar in size and shape to that of KNM-ER 3733,
despite being as much as one million years younger than the Koobi Fora hominin. However, the posterior vault and
mastoid region of OH 12 are most similar to OH 9. This combination of morphology suggests that the relationship
between the Olduvai and Koobi Fora portions of the H. erectus hypodigm requires reconsideration.
 2004 Elsevier Ltd. All rights reserved.

Keywords: Homo erectus; Homo ergaster; Interorbital region; Supraorbital torus

Introduction counterparts in Kenya, I describe how the identi-

fication of previously unidentified conjoins and the
Careful comparisons among Homo erectus faces recognition of additional fragments of the Olduvai
and facial fragments from Koobi Fora and West Hominid (OH) 12 face enables a more compre-
Turkana and their Asian counterparts have yielded hensive assessment of facial variation within the
important clues to their relationships (Wood, African hypodigm of H. erectus.
1991; Rightmire, 1998). Yet the scarcity of facial OH 12 is geologically much younger than the
fossils of various developmental and geological robust OH 9 that has been interpreted as a differ-
ages necessarily leaves many questions regarding ent morph than the hominins from Koobi Fora
the range of variation and relationships among the (e.g., Wood, 1991). Margaret Leakey discovered
hominins unresolved. Although the H. erectus OH 12 in 1962 on the west side of the VEK gully
remains from Olduvai Gorge, Tanzania yield (site 45b, geologic locality 86; Tobias, 1965;
less facial information than their more complete Leakey, 1971) where Beds III, IV, and the Masek
Beds outcrop (Hay, 1976). Fragments of OH 12
* Corresponding author. littered the surface of Bed III and additional
E-mail address: [Link]@[Link] (S.C. Antón). fragments were collected by scraping and screening
0047-2484/04/$ - see front matter  2004 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2003.12.005
338 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347

Table 1
Comparative cranial capacitya (in cc), sizeb (in mm), and thicknessb (in mm)

Cranial Interorbital Lamda-inion Mid supra-orbital Bregma Occipital torus Geological

capacity breadth chord/arc torus thickness thickness thickness agec (Ma)
H. habilis
OH 16 ?620 – – 7 6.5 11.5 1.6–1.7
OH 24 ?590 24 – 8 4 6 1.8–1.9
KNM-ER 1470 750 23 51/61 11 7 14 1.85–1.95
KNM-ER 1813 510 20 43/48 9 6 19 1.85–1.95

H. erectus
OH 12 727 23 53/57 10 10 18 R0.78,%1.2
OH 9 1067 – 54/? 18.5 – 18.5 R1.47
KNM-ER 3733 848 23 61/68 8 – 24 1.7–1.8
KNM-ER 3883 804 22 49/52 11 8 14 1.5–1.65
Dmanisi 2280 780 (28) 49/55 12 9 14 1.7
Dmanisi 2282 ?650 – – 10 8 10 1.7
Sangiran 2 813 – 45/47 – 8.5 16.5 w1.5
Sangiran IX 840 ? 42/46 14 (10) – <1.3

Sangiran 17 1004 (31) 58/63 17 – – w1.3

Zhoukoudian X 1225 – 48/50 13 7.5 15 w0.42
Zhoukoudian XI 1015 (28) 45/48 14 7.0 12 w0.42
Zhoukoudian XII 1030 (27) 56/59 16 9.7 15 w0.42
a
Cranial capacities from: Holloway (1973, 1981, 1983); Tobias (1991); Gabunia et al. (2000); Weidenreich (1943); Arif
et al. (2001).
b
Metrics from: Antón this study; Antón and Franzen (1997); Tobias (1991); Wood (1991); Gabunia et al. (2000); Weidenreich
(1943); Sartono and Tyler (1993).
c
Geological ages from: Feibel et al. (1989); Hay (1976); Tamrat et al. (1995); Manega (1993); Gabunia et al. (2000); Grün et al.
(1997); Antón and Swisher (2001).

an area of 30040 feet (Leakey, 1971). The gritty Descriptions

matrix adhering to the specimen, particularly the
palate, suggests that OH 12 originated below Tuff Olduvai Hominid 12 is a fragmentary calvaria
IVa in lower Bed IV. The most recent chronologies preserving much of the posterior cranial vault (OH
suggest that the top of Bed IV is coincident with 12a), a fragment of frontal and both parietals at
the Brunhes-Matuyama boundary (Swisher et al., bregma (OH 12b), isolated fragments of the left
unpublished data), making all of Bed IV (and OH (OH 12c) and right temporals (OH 12d), a partial
12) older than 0.78 Ma on the basis of its reversed left maxilla (OH 12e), and half of the supraorbital
geomagnetic polarity (Tamrat et al., 1995). The margin of the frontal (OH 12f). Rightmire (1979,
bottom of Bed III is dated to 1.25–1.47 Ma by 1993) described the anatomy and Holloway (1973)
40
Ar/39Ar dates on single feldspar crystals from estimated the cranial capacity at 727 cc. As of
tuff III-1 (Manega, 1993). Presuming that the bed 2001, the six principal cranial fragments of OH 12
associations are sound, OH 12 could be as much as were not individually lettered, but were individu-
700,000 years younger than OH 9, a surface find ally cut into foam shelving and housed in the fossil
attributed to upper Bed II (R1.47 Ma), and some hominin collections of the National Museums of
one million years younger than KNM-ER 3733 Kenya (KNM). I identify the six principal frag-
(w1.7 Ma; Table 1). ments here as OH 12a–f to clarify their description
 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347 339

Asian hominins. I also identified parts of the

coronoid process of the mandible (OH 12k),
the tympanic plate of the temporal (OH 12l)
including a small, ossified styloid process, and
several isolated vault fragments (OH 12m+).
The most significant of the new conjoins is
between the previously described right orbital rim
of the frontal (OH 12f) and the previously un-
identified interorbital region (OH 12g; Fig. 2;
Table 1). This conjoin allows evaluation of the
medial orbital and interorbital morphology of OH
12, and provides a means of orienting the pre-
viously isolated supraorbital fragment. Although
the interorbital fragment (OH 12g) is composed
anteriorly of portions of left and right nasals,
maxillae, and frontal, most of the sutures have
fused, leaving little visible trace. Viewed anteriorly,
the fragment is about 32 mm in height and 21 mm
wide and retains the right frontal process of the
Fig. 1. Schematic of fragments of OH 12 overlain on an outline maxilla, the superiormost portion of both nasals
of KNM-ER 3733 to indicate areas preserved in OH 12. This extending approximately 7.2 mm below nasion,
figure is not meant as a reconstruction of OH 12, but only as a
representation of the areas preserved in that specimen and the and approximately 9 mm of the frontal inferior to
new facial fragments. glabella (Fig. 2). Viewed from the right side, the
fragment is about 24 mm deep and is composed of
the right orbital wall, including portions of the
below and to distinguish them from the newly frontal and ethmoid, and superior portions of the
identified pieces (see Fig. 1 and Table 2). In posterior and anterior lacrimal crests that extend
addition to the principal cranial fragments, a plas- approximately 6 mm inferiorly from their origin.
tic box containing 29 hominin and non-hominin Viewed from the left side, the posterior and an-
bone fragments from the find site are also stored terior lacrimal crests are also preserved, but the
with the hominin collections. A survey of the outer cortex of the ethmoid and frontal are absent,
separately housed paleontology collections at exposing reddish trabecular bone. The posterior
KNM yielded no other skeletal remains from the surface of the fragment exposes matrix-filled air
OH 12 site. The cortical surfaces of the hominin cells that are unequally divided into smaller left
fragments are stained either black or white, with and larger right halves by a bony septum approxi-
some broken fragments exhibiting a reddish- mately 2 to 2.5 mm thick. The air cells extend
brown internal staining. Most non-hominin inferiorly approximately 15 mm before they are
fragments have reddish cortical surfaces. truncated by a postmortem fracture. Both aircells
In 2001, I found conjoins between the orbital are mediolaterally narrow, seemingly confined to
rim (OH 12f) and an interorbital fragment (OH the internasal region, and are consistent with the
12g) from the plastic box, as well as other conjoins position of ethmoid aircells. On the inferior surface
among the fragments in the box. These conjoins of the fragment no cortical bone is present and two
produce a relatively complete right superomedial smaller, matrix-filled air spaces are visible. These
corner of the right orbit (OH 12f/g), the inferior spaces extend 8 mm anteroposteriorly and 5 mm
lateral corner of the right orbital margin (OH mediolaterally and are separated by a 0.5 mm wide
12h/i), and a portion of the infraorbital plate of the bone plate, probably the perpendicular plate of the
left maxilla (OH 12j; Fig. 2) that enhance our ethmoid. Superiorly, exposed trabecular bone is
ability to compare OH 12 with other African and also infilled with matrix.
340 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347

Table 2
Key to fragments of OH 12

Specimen number Element(s)

OH 12a Posterior vault—parts of occipital and parietals
OH 12b Bregmatic fragment—parts of frontal and parietals
OH 12c Temporal—left mastoid and superior external auditory meatus
OH 12d Temporal—right anterior squama, glenoid fossa, and superior external auditory meatus
OH 12e Maxilla—left palate, roots P3-M2
OH 12f Frontal—right orbital rim
OH 12g Interorbital region—conjoins OH 12f
OH 12h Maxilla—inferior orbital rim
OH 12i Maxilla/zygomatic—masseter attachment, conjoins OH 12h
OH 12j Maxilla—left infraorbital region
OH 12k Mandible—coronoid
OH 12l Temporal—tympanic plate
OH 12m+ Miscellaneous vault fragments

The newly conjoined fragments (OH 12f/g) Olduvai hominins, the OH 12 frontal exhibits
form the superomedial corner of the right orbital a carnivore tooth score in its supratoral sulcus
rim. The piece preserves approximately the medial (personal observation).
half of the superior orbital rim, including the The second of the newly recognized joins is
supraorbital notch, portions of the supratoral between right maxillary and zygomatic fragments
sulcus (also known as the post-toral sulcus or at the inferolateral corner of the orbit (and an-
supratoral gutter), and the anterior endocranial terior portion of the temporal fossa; OH 12h/i; Fig.
surface adjacent to midline. An approximately 90( 2). The conjoined piece measures 28 mm supero-
angle is formed at the superomedial corner of the inferiorly and 23 mm mediolaterally and retains
orbit. The right frontal process of the maxilla is traces of the blunt right orbital rim, the right
convex and angled anterolaterally, but lies in the maxillary sinus, the right zygomatico-maxillary
same vertical plane as do the nasals, rather than suture, and the body of the right zygomatic. The
being located more posteriorly, thus enhancing the portion of the right orbital rim that would connect
squareness of the superomedial corner of the right this fragment with the interorbital fragment de-
orbit. The square appearance is also heightened by scribed above is not preserved. Traces of the
postmortem loss of a fragment of the ectocranial zygomaticomaxillary suture remain, including a
surface of the medial end of the supraorbital torus bisected foramen. The anterior surface of the frag-
that limits the extent of visible arching in the ment is flat to slightly convex. Inferiorly, the
browridge. massetter attachment on the inferior zygomatic
The supraorbital torus (SOT) is moderately body is rough, flat, and approximately 10.8 mm
developed in OH 12, with a well-developed supra- deep anteroposteriorly.
toral sulcus behind it. The morphology of the The previously described palate fragment (OH
sulcus slightly lateral to midline suggests that the 12e) preserves portions of the left alveolar process,
sulcus would have been continuous behind glabella lateral horizontal plates, maxillary sinus, and the
had it been preserved in this region. Laterally, the root of the nasal aperture of the maxilla, as well as
supratoral sulcus is anteroposteriorly wide, and the roots of LP3–M2 (Rightmire, 1979, 1993). The
deeply concave suggesting a steep rise of the fron- fragment measures approximately 42 mm antero-
tal squama posteriorly. Glabella is not preserved, posteriorly. A fossa, probably of pathological ori-
but enough of the brow below and to either side is gin, is consistent with a large periapical abscess,
retained to indicate that glabella was only slightly and incorporates the horizontal palatal shelf and
to non-projecting in OH 12. Like many other lingual roots of P4 and anterior M1. No midline
 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347 341

Fig. 2. New facial fragments of OH 12. Anterior view of reconjoined fragments of right supraorbital/interorbital (OH 12f/g) and
maxillozygomatic (OH 12h/i) specimens, and left infraorbital portion of maxilla (OH 12j) in approximate anatomical position.

structures are preserved. However, the nasal floor least level with, or may have extended slightly
was substantially thicker anteriorly (w13 mm) below, the floor of the nasal cavity. The roots of
than posteriorly and the descent from the region of LP3–M2 and the anterior alveolus of M3 suggest
the nasal sill into the nasal cavity is steep. It is a parabolic dental arcade with M3 positioned
impossible to say whether, or to what extent, this slightly lingual to M2. The palatal surface is rela-
descent plateaus (or is stepped sensu McCollum tively smooth. A previously undescribed fragment
et al., 1993; McCollum, 2000) to the region of the of the left infraorbital plate of the maxilla (OH 12j)
nasal sill. Most of the infraorbital region is broken that does not conjoin with the above is also
just above the alveoli exposing the matrix filled present. This piece is about 2417 mm supero-
maxillary sinus and the root of a bony protuber- inferiorly and includes the infraorbital foramen on
ance related to the canal for the infraorbital nerve its external surface and a bony tube (canal) for the
(see below). The floor of the maxillary sinus was at infraorbital nerve on its sinus surface. Although
342 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347

the two do not conjoin, this tube is related to the Homo” from both Olduvai and Koobi Fora (Table
bony protuberance preserved in the maxillary 1). The SOT is shorter in vertical height in the
sinus of OH 12e mentioned above. The anterior “non-erectus early Homo” group, than in OH 12.
surface of the infraorbital region is flat. The greater distinction of the SOT in OH 12 is in
part related to the presence of an anteroposteriorly
wide, concave supratoral sulcus that shows no
Comparisons evidence of weakening behind glabella. In con-
trast, the sulcus is entirely absent in KNM-ER
The existing and newly identified fragments are 1470, and in KNM-ER 1813 the sulcus is neither
compared metrically and morphologically with continuous nor anteroposteriorly wide, although
facial, especially orbital, remains of similarly aged from about midorbit to the midline the transition
fossil Homo. For the purposes of this paper, those between glabella and the frontal squama is slightly
remains from Olduvai (e.g., OH 16 and 24) and concave. In both OH 16 and OH 24 the lateral part
Koobi Fora (KNM-ER 1470 and 1813) formally of the supratoral sulcus is more concave than in
assigned to either H. habilis or H. rudolofensis are either of the Koobi Fora specimens. Tobias (1991)
considered together and hereafter referred to as the considers that both OH 16 and OH 24 possess
‘non-erectus early Homo’ group. OH 12 is also supratoral sulci. However, it should be noted that
compared with H. erectus (sensu lato) from Africa these features differ from the supratoral sulci
(OH 9, KNM-ER 3733, 3883), Asia (Dmanisi of OH 12 by exhibiting a more even transition
D2280, D2282; Zhoukoudian Skull X, XI, XII), between glabella and the frontal squama. Post-
and Southeast Asia (Sangiran 2, 17, 27, and IX). depositional crushing may influence this observed
Except for the Zhoukoudian remains and Sangiran morphology. In the “non-erectus early Homo”
IX, observations are based on the original fossils. group any supratoral concavity, wherever located,
OH 12 is compared, to the extent possible based is anteroposteriorly short.
on current publications, with the circa 1.0 Ma The interorbital region of OH 12 differs from
Daka and Buia crania from Ethiopia and Eritrea the Olduvai “non-erectus early Homo” specimens
(Abbate et al., 1998; Asfaw et al., 2002). in being broader and less inflected anteroposteri-
orly and having a squarer superomedial corner. In
OH 24, in particular, the nasals are strongly flexed
“Non-erectus early Homo” faces (concave anteriorly) below glabella and relatively
flat from medial to lateral. The nasofrontal region
The face of H. erectus differs systematically of OH 12 is less inflected at the saddle of the nose,
from that of the “non-erectus early Homo” group but more convex from medial to lateral. Although
by having a continuous and distinct supraorbital damaged in this region, the Koobi Fora ‘non-
torus that is frequently barlike and often quite erectus early Homo’ do not share the marked nasal
thick, a supratoral sulcus of variable shape that is inflection of OH 24, although KNM-ER 1470
continuous behind glabella, and a relatively broad, presents a slight concavity at the nasal saddle.
mediolaterally convex interorbital region. In ad- Both show a marked difference in elevation from
dition, the maxilla and zygomatic are more convex nasion to glabella. As mentioned previously, the
in the lateral malar region and are also taller in OH 12 glabella is unlikely to be projecting.
H. erectus. Other, probably wholly size-related The zygomaxillary region exhibits both size and
features including a blunter inferior orbital rim shape related variation. Shape related variation
also differentiate the two, but may be seen in larger suggests that OH 12 is H. erectus; the convex
members of the “non-erectus early Homo” group. lateral malar region of OH 12(h/i) and the flat
OH 12 conforms to the facial characters of infraorbital plate (OH 12j) contrasts with the flat
H. erectus. malar regions of the “non-erectus early Homo”
The supraorbital torus of OH 12 is larger and group and their more concave infraorbital plates
more distinct than in the “non-erectus early (at least in OH 24 and KNM-ER 1813). Size
 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347 343

related variation includes the blunter inferior smaller, its SOT vertically and horizontally less
orbital margin that is broader and blunter in OH well-developed, and its glabella is inferred to have
12 than in OH 24 and KNM-ER 1813, but is been absolutely less projecting. Both share, along
similar in shape to KNM-ER 1470. The inferior with KNM-ER 3733 and Dmanisi 2282, an appar-
zygomatic of OH 12 is also thicker at the masseter ently continuous, concave supratoral sulcus that
attachment than in OH 24 and KNM-ER 1813. rises relatively steeply to meet the frontal squama
Estimated malar height of OH 12 is intermediate (Fig. 3). However, the sulcus is wider antero-
between the smaller specimens (i.e., OH 24, posteriorly in OH 9 than in the other specimens,
KNM-ER 1813) and KNM-ER 1470. probably as a result of the greatly enlarged SOT of
The OH 12 palate is similar in size, based on this specimen. The frontal squama seems to rise
tooth roots, and depth at M2 to OH 24. Both are equally steeply in all the specimens. KNM-ER
slightly larger than KNM-ER 1813 and smaller 3883 and Dmanisi 2280 have shallower supratoral
than KNM-ER 1470. Only OH 24 appears to have sulci, particularly at the midline, than OH 12.
the differential palatal thickness from anterior to However, it is possible that this difference is some-
posterior similar to that of OH 12 (see McCollum what accentuated by postmortem deformation in
et al., 1993). The OH 12 dental arcade appears KNM-ER 3883.
somewhat more parabolic than any of these The superomedial corner of the right orbital
palates. margin of OH 12 is somewhat more square than
other orbits from Africa and Georgia due in large
part to the conformation of the medial orbital rim
H. erectus faces in OH 12 (Fig. 2). In the other specimens that
preserve this region, the upper part of the medial
In general appearance and size, the OH 12 facial orbital rim (usually formed by the frontal) angles
fragments are more similar to H. erectus faces superolateral to inferomedia, forming an obtuse
than to the “non-erectus early Homo” faces from angle with the superior orbital rim, whereas in OH
Olduvai and Koobi Fora, and are more similar to 12 this rim runs at nearly 90( from the superior
early African/Georgian H. erectus than to Asian orbital rim. In addition, the frontal processes of
H. erectus. Similarities to H. erectus (sensu lato) the maxillae and the nasals of OH 12 are located
include: vertical height, development, and position on nearly the same coronal plane, whereas in most
of the SOT; presence of a wide, strongly concave, other specimens, and particularly in OH 9, the
and (apparently) continuous supratoral sulcus; frontal processes of the maxilla are located further
higher inferred position of glabella; and less posteriorly than are the nasal bones. Although
inflected nasals that are broader and slightly partially reconstructed in this region, the D2280
more convex mediolaterally. Early African and specimen is most similar to OH 12 in this feature,
Georgian H. erectus faces tend to differ from those suggesting that these relationships are influenced
of Far Eastern H. erectus by being smaller overall, by absolute size and particularly by size of the
exhibiting different shapes of the supratoral sulcus glabellar portion of the SOT. The interorbital
(although recent finds may complicate this distinc- breadth of OH 12, estimated at 23 mm by doubling
tion), and by having flatter malar and infraorbital the measurement from midline to the right side, is
regions. In these ways, OH 12 conforms more similar to the values for KNM-ER 3733 and 3883
closely with an Early African/Georgian morpho- (Table 1). Overall, the differences are least between
logical pattern. OH 12 and KNM-ER 3733 and greatest with
OH 9.
Early Africa and Georgia The zygomaxillary regions of OH 12, KNM-ER
3733, 3883, and D2282 are convex laterally and the
The SOT and supratoral sulcus of OH 12 are infraorbital portions of each are apparently flat to
very similar to those of KNM-ER 3733 in size and slightly concave, although the region is incom-
shape. Compared with OH 9, OH 12 is absolutely pletely preserved in KNM-ER 3733 and is partially
344 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347

Fig. 3. Comparison of OH 12 (cast) and KNM-ER 3733 (original) supraorbital regions.

crushed in D2282. The inferolateral orbital margin Daka and Buia specimens (Abbate et al., 1998;
of OH 12 is blunter and more rounded than in Asfaw et al., 2002). It should be noted that vari-
KNM-ER 3733 and 3883 and D2282. The infra- ation in the supraorbital torus and other cranial
orbital foramen opens inferolaterally. characters has led some authors to question the
The palate of OH 12 is similar in size, based on taxonomic affinities of one or both of the speci-
tooth roots, to KNM-ER 3733 and D2282, but is mens (Abbate et al., 1998; Manzi et al., 2003; but
shallower than the latter. Both OH 12 and D2282 see Gilbert et al., 2003). Both specimens are sig-
have more parabolic arcades than does KNM-ER nificantly larger than OH 12 and this appears
3733 due, at least in part, in D2282 to a narrower especially true of their peri-orbital regions. Both
anterior arcade. have supraorbital tori that are absolutely thicker
than in OH 12 and that, unlike other H. erectus,
Daka and Buia are arched over each orbit. The superomedial
corner of the Daka orbit does not appear as
The following limited observations are possible strongly angled as is the case in OH 12. In superior
based on published accounts and photos of the view, the Daka SOT appears similar to that of
 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347 345

OH 9 in both its size and the position of its greatest XII has a rounded superomedial corner. Both of
projection (just lateral of glabella), and thus the Zhoukoudian specimens, however, have fron-
dissimilar to OH 12. The Buia palate is described tal processes of the maxilla that are positioned
as shallow, which may be similar to the condition relatively posterior to the nasals.
in OH 12. The inferolateral orbital margin can be com-
pared with that of Sangiran 17. Both margins are
China and Indonesia rounded and blunt. The apparent height of the
malar region is much smaller in OH 12 than in
The SOT and supratoral sulcus of OH 12 are Sangiran 17, but both are convex laterally. OH 12
less like those of H. erectus crania from Asia. is flatter in the infraorbital region than is Sangiran
Absolute toral size is markedly less robust in OH 17. The OH 12 palate is much less massive and less
12 than in all but the Sangiran 2 and Trinil deep than those from Sangiran (4 and 17) and
specimens, neither of which are well-enough pre- smaller than those from Zhoukoudian. OH 12
served at midline for detailed comparisons (Table shares a parabolic dental arcade shape with the
1). Supratoral sulcus morphology is stereotypi- maxilla from Zhoukoudian and both differ from
cally different between Chinese and Indonesian the Sangiran specimens, with their more divergent
H. erectus, regardless of brain size, with Chinese tooth rows.
specimens possessing a more defined supratoral
sulcus, especially posterior to glabella, and a
steeper rise in the frontal squama posterior to this Discussion
(Antón, 2002, 2003). Generally, the Chinese con-
dition differs from African morphology by being OH 12 shows some facial similarities to African
anteroposteriorly narrower and by being narrower H. erectus from Olduvai and Koobi Fora. These
behind glabella than laterally. That is, the Chinese similarities, coupled with the extensive cranial
supratoral sulcus “wraps around” the anterior- thickness, features of the vault including a swelling
most extent of the brain, whereas in the African at bregma, an occipital torus, and the shape of the
specimens it generally passes in front of the brain. mastoid process suggest, as previously concluded,
The fragmentary nature of the OH 12 specimen that OH 12 is a small-brained member of H.
precludes comparison of total sulcus morphology. erectus (Rightmire, 1979). Despite this conclusion,
However, what is preserved suggests a relatively some aspects of supraorbital shape, but not size, of
wider sulcus at midline than is present in the “non-erectus early Homo” crania from Olduvai
Chinese specimens, particularly relative to the (especially OH 24) anticipate that of OH 12.
small vertical size of the torus of OH 12 compared Size and bluntness of the inferolateral orbital
with those from Zhoukoudian. Compared with margin, as well as the thick vault bones, suggest
Indonesian supratoral sulci, the sulcus of OH 12 is OH 12 was a robust individual, but of relatively
wider, more concave, and the frontal squama small size based on cranial capacity, estimated
appears to rise more steeply from the sulcus than is malar height, and palate size. There is a marked
the case in Sangiran 2 (laterally), 17, or even differential between SOT thickness and vault bone
Sangiran IX which displays the most developed thickness in OH 12 that is unmatched in most
supratoral sulcus of the early Indonesian speci- other African H. erectus (Table 1). Differences in
mens (Arif et al., 2001). relative thickness cannot rule out the possibility
The shape of the superomedial orbital margin that the face and vault may have come from two
and interorbital region can only be satisfac- individuals. However, the disparity is similar to
torily assessed in Sangiran 17 and IX, and in that seen in KNM-ER 3733, which is also morpho-
Zhoukoudian Skull X and XII. Sangiran 17, with logically most similar to OH 12, except in terms of
midline peaking at the internasal suture, differs the temporal morphology. While it would be simplest
most from OH 12. Skull X approaches the square- to attribute the size differences of the face and
ness of the OH 12 orbital margin, whereas Skull especially the brow of these specimens relative to
346 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347

OH 9 to sexual dimorphism, differences in tem- vault that are subtly different from those seen in
poral bone morphology between OH 12 and H. erectus from the Far East (Antón, 2003). Only
KNM-ER 3733 complicate this assessment. further additions to the fossil record of H. erectus
Clearly, larger samples are necessary to sort out will help elucidate whether these variations are
size and sex related variation. better seen as local adaptations/genetic drift within
At between 0.78 Ma and 1.0 Ma, OH 12 is the a lineage or more discrete boundaries between
youngest of the African H. erectus fossils to date lineages.
and also the smallest of the adults. The small
cranial capacity and similarity of facial form with
KNM-ER 3733 are a surprise given the young Acknowledgements
geological age of OH 12. The similarities between
OH 12 and KNM-ER 3733 suggest the mor- I am grateful to Dr. Fidelis Masao for his
phology they share persisted for nearly one million invaluable assistance in studying the Olduvai ho-
years. Moreover, OH 12 differs from the recently minins in Tanzania and Kenya and to COSTECH,
announced Daka specimen (Asfaw et al., 2002), the Department of Antiquities, Tanzania, the
with which it is approximately coeval. Published National Museums of Tanzania, the National
data do not allow detailed comparisons, but the Museums of Kenya (KNM), Mary Muungu, then
Daka cranium is much larger and appears to acting director of Paleontology, and Christopher
combine neurocranial morphology similar to that Kiare of KNM for permits and access to speci-
of Koobi Fora H. erectus with supraorbital mor- mens in their care. Mr. J. Ndunda Mbithi, casting
phology more similar to OH 9. The new fragments director of KNM, A. Katula, and J. Kiptalam
of OH 12, on the other hand, show the opposite; molded the new specimens and provided casts. A.
sharing upper facial morphology with KNM-ER Macharia provided comparative cast specimens.
3733, but aspects of the posterior vault, the mas- Drs. Meave Leakey, Fred Spoor, and C.C. Swisher
toid process in particular, with OH 9. These find- III provided helpful discussion, and Drs. T.
ings suggest that interpretation of the relationships Harrison, B. Wood, and two anonymous reviewers
between Olduvai and Koobi Fora H. erectus provided essential critical advice. This work was
that stress differences between the two may need supported indirectly by NSF BCS-9804861 and by
reconsideration. the Center for Human Evolutionary Studies,
However, the different combinations of charac- Rutgers University.
ters in the two hypodigms beg the question of
whether Olduvai and Koobi Fora hominins should
be included in the same lineage and whether these
References
form a single species with Asian forms. Although
Asfaw and colleagues’ (Asfaw et al., 2002) cladistic
Abbate, E., Albianelli, A., Azzaroli, A., Benvenuti, M.,
analysis separated these two demes, it seems to Tesfamania, B., Brun, P., Cipriani, N., Clarke, R.J.,
have done so principally by separating all fossils Ficcarelli, G., Macchiarelli, R., Napoleone, G., Papini, M.,
into two main groups on the basis of size. Groups Rook, L., Sagri, M., Tecle, T.M., Torre, D., Villa, I., 1998.
with average cranial capacities below 1000 cc A one million year old Homo cranium from the Danakil
(Afar) depression of Eritrea. Nature 393, 458–460.
(Koobi Fora, Sangiran/Trinil, Dmanisi) separated
Antón, S.C., 2002. Evolutionary significance of cranial vari-
from those with average cranial capacities greater ation in Asian H. erectus. Am. J. phys. Anthrop. 118,
than 1000 cc (e.g., OH 9/Daka, Ngandong, 301–323.
Zhoukoudian). The inclusion of the small OH 12 Antón, S.C., 2003. Natural history of Homo erectus. Yearb.
might thus critically affect these results. Whatever phys. Anthrop. 37, 126–170.
Antón, S.C., Franzen, J.L., 1997. The occipital torus and
the affinities of OH 9, OH 12, and the Koobi Fora
developmental age of Sangiran-3. J. hum. Evol. 33,
hominins, they (and the “Daka” and Dmanisi 599–610.
fossils as well) arguably exhibit “African” features Antón, S.C., Swisher, C.C. III, 2001. Evolution of cranial
of vault shape, the cranial base, and posterior capacity in Asian Homo erectus. In: Indriati, E. (Ed.), A
 S.C. Antón / Journal of Human Evolution 46 (2004) 337–347 347

Scientific Life: Papers in Honor of Prof. Dr. T. Jacob. Manega, P.C., 1993. Geochronology, geochemistry, and iso-
Bigraf, Yogyakarta, pp. 25–39. topic study of Plio-Pleistocene hominid sites and the
Arif, J., Baba, H., Suparka, M.E., Zaim, Y., Setoguchi, T., Ngorongoro volcanic highlands in northern Tanzania. Ph.D
2001. Preliminary study of Homo erectus Skull IX (Tjg- Dissertation, University of Colorado.
1993.05) from Sangiran, central Java, Indonesia. Bull. Manzi, G., Bruner, E., Passarello, P., 2003. The one-million-
Natnl. Sci. Mus., Tokyo Ser. D. 27, 1–17. year-old Homo cranium from Bouri (Ethiopia): a reconsid-
Asfaw, B., Gilbert, W.H., Beyene, Y., Hart, W.K., Renne, eration of its H. erectus affinities. J. hum. Evol. 44, 731–736.
P.R., WoldeGabriel, G., Vrba, E.S., White, T.D., 2002. McCollum, M.A., Grine, F.E., Ward, S.C., Kimbel, W.H.,
Remains of Homo erectus from Bouri, Middle Awash, 1993. Subnasal morphological variation in extant homi-
Ethiopia. Nature 416, 317–320. noids and fossil hominids. J. hum. Evol. 24, 87–112.
Feibel, C.S., Brown, F.H., McDougall, I., 1989. Stratigraphic McCollum, M.A., 2000. Subnasal morphological variation in
context of fossil hominids from the Omo groups deposits: fossil hominids: a reassessment based on new observations
northern Turkana Basin, Kenya and Ethiopia. Am. J. phys. and recent developmental findings. Am. J. phys. Anthrop.
Anthrop. 78, 595–622. 112, 275–283.
Gabunia, L., Vekua, A., Lordkipanidze, D., Swisher, C.C., Rightmire, G.P., 1979. Cranial remains of Homo erectus from
Ferring, R., Justus, A., Nioradze, M., Tvalchrelidze, M., Beds II and IV, Olduvai Gorge, Tanzania. Am. J. phys.
Antón, S.C., Bosinski, G., Joris, O., de Lumley, M-A., Anthrop. 52, 99–116.
Majsuradze, G., Mouskhelishvili, A., 2000. Earliest Pleis- Rightmire, G.P., 1993. The Evolution of Homo erectus: Com-
tocene cranial remains from Dmanisi, Republic of Georgia: parative Anatomical Studies of an Extinct Human Species.
taxonomy, geological setting, and age. Science 288, Cambridge University Press, New York.
1019–1025. Rightmire, G.P., 1998. Evidence from facial morphology for
similarity of Asian and African representatives of Homo
Gilbert, W.H., White, T.D., Asfaw, B., 2003. Homo erectus,
erectus. Am. J. phys. Anthrop. 106, 61–85.
Homo ergaster, Homo “cepranensis,” and the Daka
Sartono, S., Tyler, D.E., 1993. A new Homo erectus skull from
cranium. J. hum. Evol. 45, 255–259.
Sangiran, Java: an announcement. Intl. Conf. Hum.
Grün, R., Huang, P-H., Wu, X., Stringer, C.B., Thorne, A.G.,
Paleoecology. LIPPI, Jakarta.
McCulloch, M., 1997. ESR analysis of teeth from the
Tamrat, E., Tohveny, N., Taieb, M., Opdyke, N., 1995. Re-
palaeoanthropological site of Zhoukoudian, China. J. hum.
vised magnetostratigraphy of Plio-Pleistocene sedimentary
Evol. 32, 83–91.
sequence of the Olduvai Formation (Tanzania). Palaeoge-
Hay, R.L., 1976. Geology of Olduvai Gorge. University of ogr., Palaeoclimatol., Palaeoecol. 114, 273–283.
California Press, Berkeley. Tobias, P.V., 1965. New discoveries in Tanganyika, their
Holloway, R.L., 1973. New endocranial values for the East bearing on hominid evolution. Curr. Anthrop. 6, 391–399.
African early hominids. Nature 243, 97–99. Tobias, P.V., 1991. Olduvai Gorge Volume 4: The Skulls and
Holloway, R.L., 1981. The Indonesian Homo erectus brain Endocasts of Homo habilis. Cambridge University Press,
endocasts revisited. Am. J. phys. Anthrop. 55, 503–521. Cambridge.
Holloway, R.L., 1983. Human brain evolution: a search for Weidenreich, F., 1943. The skull of Sinanthropus pekinensis; a
units, models and synthesis. Can. J. Anthrop. 3, 215–230. comparative study on a primitive hominid skull. Palaeont.
Leakey, M.D., 1971. Olduvai Gorge Volume 3: Excavations in Sin. Ser. D 10, 1–298.
Beds I and II, 1960–1963. Cambridge University Press, Wood, B., 1991. Koobi Fora Research Project Volume 4:
Cambridge. Hominid Cranial Remains. Clarendon Press, Oxford.