Research Article

A Complete Skull from Dmanisi, Georgia, and the Evolutionary Biology of Early Homo

See allHide authors and affiliations

Science  18 Oct 2013:
Vol. 342, Issue 6156, pp. 326-331
DOI: 10.1126/science.1238484

A Heady Find

In the past two decades, excavations at the archaeological site at Dmanisi, Georgia, have revealed hominin fossils from the earliest Pleistocene, soon after the genus Homo first dispersed beyond Africa. Lordkipanidze et al. (p. 326; see the cover) now describe a fossil cranium from the site. Combined with mandibular remains that had been found earlier, this find completes the first entire hominin skull from this period.


The site of Dmanisi, Georgia, has yielded an impressive sample of hominid cranial and postcranial remains, documenting the presence of Homo outside Africa around 1.8 million years ago. Here we report on a new cranium from Dmanisi (D4500) that, together with its mandible (D2600), represents the world's first completely preserved adult hominid skull from the early Pleistocene. D4500/D2600 combines a small braincase (546 cubic centimeters) with a large prognathic face and exhibits close morphological affinities with the earliest known Homo fossils from Africa. The Dmanisi sample, which now comprises five crania, provides direct evidence for wide morphological variation within and among early Homo paleodemes. This implies the existence of a single evolving lineage of early Homo, with phylogeographic continuity across continents.

Excavations at Dmanisi, Georgia, have yielded hominid fossils, a rich vertebrate fauna, and Mode I (Oldowan) stone artifacts co-occurring in well-understood geological and taphonomic contexts. Occupation of the site began shortly after 1.85 million years ago (Ma) and is documented until about 1.77 Ma (16). The temporal and geographic setting and the preservation of at least five hominid individuals make Dmanisi a site that is crucial in understanding patterns of variation, biogeography, and evolution within emergent Homo. The generalized (primitive) craniomandibular morphology, small cranial capacity, moderate body size, and a mosaic of primitive and derived postcranial features link the Dmanisi sample to largely contemporaneous fossils from Africa attributed to H. habilis and/or H. erectus. Various features of the cranial vault and base also indicate affinities to H. erectus from East Asia (4, 5, 7, 8).

Here we present a complete cranium (Fig. 1) that substantially expands the range of variation seen in the Dmanisi sample (Fig. 2), with implications as to the early evolutionary history of the genus Homo. Specimen D4500 was recovered in 2005 from the base of layer B1y in Block 2 of the Dmanisi excavation area (fig. S1) (9). D4500 represents the same individual as mandible D2600, found 5 years earlier (10). The anatomical association of D4500 and D2600 (hereafter “skull 5”) constitutes the most complete adult skull known from Early Pleistocene Homo (hereafter “early Homo”; the definition of Pleistocene follows the new terminology). Skull 5 is probably associated with the postcranial elements of an adult individual with nearly modern human body proportions [(5); supplementary text S1 and fig. S1]. Because skull 5 is intact and free from taphonomic deformation, it is now possible to document the morphology of an entire adult cranium with its associated mandible and teeth and to study anatomy not previously known for early Homo.

Fig. 1 Dmanisi cranium D4500 and associated mandible D2600.

(A and B) Anterior and posterior views; (C and D) superior and inferior views; (E and F) left and right views; (G) cranium with articulated mandible. Scale bar, 5 cm.

Fig. 2 The Dmanisi paleodeme.

CT-based visualizations, from left to right: D2280 (skull 1), D2282/D211 (skull 2), D2700/D2735 (skull 3), D3444/D3900 (skull 4), D4500/D2600 (skull 5). Scale bar, 5 cm. See fig. S3 for a full set of standard views.

Description of Cranium D4500

The morphology of skull 5 stands apart from that of any other known fossil Homo specimen through its combination of a small braincase with a large prognathic face (Fig. 1, Table 1, supplementary text S2 and S3, figs. S2 and S3, and tables S1 to S3).

Table 1

Craniodental dimensions of skull 5. Linear dimensions are in millimeters, angles in degrees, areas in square millimeters, and volumes in cubic centimeters.

View this table:


The endocranial volume, (ECV) evaluated from computed tomography (CT) data, is 546 ± 5 cm3. Coupled with its small ECV, the braincase is low but comparatively wide and basally elongate. The frontal squama is moderately sloping. The parietals are bossed and exhibit only slight midline keeling in their posterior half. The zygomatic arches are massive, and spacious temporal fossae indicate voluminous masticatory muscles. The temporal lines are well separated from each other (minimum distance from cranial midplane 12 mm behind bregma: 20 mm). The temporal bone exhibits prominent supramastoid crests and laterally wide porial saddles, which give the neurocranium a squat, bell-like shape when viewed from the rear.

The upper squama of the occipital bone has a marked protuberance immediately below its lambdoid margin, which extends across the midplane by ~30 mm on either side. The prominence of this “lambdoid hump,” together with a massive protruding inion, lends the upper occipital plane a nearly vertical orientation, whereas the nuchal plane is only moderately inclined (21°) relative to the Frankfurt horizontal (FH) plane. The occipital transverse torus is ruggedly built, and the nuchal region is deeply sculpted. A bilaminar crest linking inion with opisthion suggests a strong nuchal ligament. The mastoid processes are large and steeply inclined medially. Their inferior portion is compressed mediolaterally to form a distinctive flange-like structure that extends posteriorly. Overall, cranial superstructures of D4500 are massive and more prominent than in the other Dmanisi individuals, suggesting that it represents a male.

Temporal Bones and Cranial Base

The right zygomatic arch exhibits in vivo deformation indicating a healed but displaced multiple fracture behind the masseteric origin. The left temporomandibular joint (TMJ) as well as the left mandibular condyle bear evidence of degenerative arthritic deformation. The right TMJ and left zygomatic arch are unaffected by pathology. There is marked lateral extension of the TMJ, with 61% of the glenoid fossa beyond the external wall of the vault, below the zygomatic shelf. The fossa itself is shallow, transversely oriented (tympanomedian angle 3°), and has a low anterior wall that merges smoothly with the preglenoid planum. The tympanic plate is massive and exhibits a prominent vaginal process. Medially, the tympanic is expanded to form a robust supratubarius process. The petrous pyramid closely approaches the wall of the basioccipital, thus reducing the foramen lacerum to a crevice. The main axis of the pyramid forms an angle of 34° (petromedian angle) with the midsagittal plane. The foramen magnum is oval in shape and level (inclination 0°) with respect to the FH. Basicranial flexion, as measured by the orientation of the clival plane relative to the sphenoid plane, is low (128°).


Contrasting with the small braincase, the face of skull 5 is among the largest and most prognathic known from early Homo. The glabellar region is massive and substantially overhangs the midface. The orbits are overarched by thick bar-like supraorbital tori; a shallow post-toral shelf (rather than a sulcus) extends posteriorly toward the frontal squama. Postorbital constriction is marked. The zygomatic processes of the maxillae are massive and laterally flaring. Their anterior surface is vertically oriented and coincides with the orbital plane, thus forming a zygomatico-orbital plane (fig. S2). As a consequence of the tall and wide midface, the zygomatico-alveolar crests are markedly angled. The roots of the zygomatic processes originate between the first and second molars. The nasal floor is delimited anteriorly by a smooth (rather than sharply angled) sill. The inclination of the naso-alveolar clivus relative to the alveolar plane is 42°. The subnasal portion of the clivus is slightly concave, whereas its alveolar portion is convex, covering the large incisor roots. The palate is long and arched; the incisive foramen is located at the position of the second premolar (P4), indicating marked prognathism of the maxilla and anterior dentition.


The upper and lower dentition of skull 5 is heavily worn, and crown morphologies are only partially preserved. Estimated molar dimensions are comparatively large, and maxillary as well as mandibular molar sizes increase from M1 to M3 (Table 1 and table S3A). Judging from root orientation, the maxillary incisors were strikingly procumbent. The open bite (structural non-occlusion of the incisors), rounded labial wear surfaces on upper incisors and rounded lingual wear surfaces on lower incisors, respectively, and labiolingually oriented striations on these surfaces suggest that the anterior dentition was used in paramasticatory activities such as gripping (11). The first maxillary premolars (P3s) are situated posterolaterally to Cs, such that massive C/P3 juga square off the front of the maxilla. Similarly massive C/P3 juga are present in the mandible D2600 (12). P3s have three roots; P4s are double-rooted, with the buccal root exhibiting two canals (three-rooted P3s and double-rooted P4s have also been reported for other early Homo specimens; see table S3B). The P3s of mandible D2600 have two roots (the distal root with two canals) (12). The P4s are not preserved.

Comparative Analysis

Skull 5 is the first complete specimen to provide evidence of how the face (including the mandible) of adult early Homo was oriented and positioned relative to the braincase. Hitherto, complete craniomandibular morphology of early Homo was represented by two adolescents, whose facial skeletons were not yet fully developed (KNM-WT 15000 and D2700/D2735) (2, 13), and the senile Dmanisi individual D3444/D3900, whose gnathic morphology was strongly modified by alveolar bone resorption after tooth loss (3, 4) (Fig. 2). Among other adult early Homo specimens from Africa and East Asia, anatomical connections between preserved neurocranial and maxillofacial parts are typically incomplete, and facial-to-neurocranial orientation remains a matter of debate (14, 15). The current perception of skull morphology of early Homo is thus strongly influenced by adolescent representatives (and the less complete and taphonomically distorted adult cranium KNM-ER 1813) that exhibit comparatively orthognathic faces with lightly built superstructures. The architecture of skull 5 now indicates that small-brained, large-faced, remarkably prognathic and robust morphologies all belonged to the normal range of variation of early Homo.

The ECV of skull 5 (546 cm3) represents the smallest of the Dmanisi sample (skulls 1 to 4: 601 to 730 cm3), and it is at the lower end of variation of the H. habilis hypodigm [509 to 687 cm3 (16, 17)]. Stature and body mass estimates obtained from the postcranial elements that are probably associated with skull 5 [146 to 166 cm, 47 to 50 kg (5)] place this individual within the range of variation estimated for early Homo and at the lower end of variation of African H. erectus and modern humans (18). The skull 5 individual thus provides the first evidence that early Homo comprised adult individuals with small brains but body mass, stature, and limb proportions reaching the lower range limit of modern human variation (5, 18). Combining ECV and body mass data, the encephalization quotient (EQ) of the skull 5 individual is estimated as ~2.4, which is within the range of variation for Australopithecus (16). The larger ECV of early Homo as compared to Australopithecus africanus [340 to 515 cm3 (16)] and Au. sediba [420 cm3 (19)] might thus primarily reflect an evolutionary increase in body size rather than increased encephalization.

Although similarities between Dmanisi and eastern African fossils attributed to H. ergaster and H. erectus have previously been documented (4, 7, 8, 20), skull 5 permits us to widen the scope of comparisons (supplementary text S4 and table S4). In midfacial architecture, skull 5 exhibits morphological affinities with the broadly contemporaneous African specimens SK 847 (fig. S4) and OH 65 (21), as well as with early representatives of Homo in eastern Africa such as KNM-ER 1470 and KNM-ER 62000, dated to ~2.0 to 1.9 Ma (15). In all these specimens, the naso-alveolar clivus is moderately sloping and laterally wide, curving around the buccal roots of the P3s.

The Chemeron temporal fragment, dated to 2.4 Ma, was proposed as the earliest fossil evidence of Homo (22, 23), but given the wide variation in TMJ position and morphology in the Dmanisi and early Homo samples (table S4), this assignment remains inconclusive. However, the maxilla of skull 5 is similar in overall form and dento-alveolar features to the well-preserved early Homo maxilla A.L. 666-1 from Hadar, Ethiopia, with a geological age of 2.33 Ma (24) (fig. S4 and table S4). The generic and chronological status of A.L. 666-1 has recently been questioned, and A. sediba (1.977 Ma) has been proposed instead as a potential ancestor of the Homo lineage (25). The close morphological affinities between A.L. 666-1, additional early Homo specimens from eastern Africa such as OH 65 and KNM-ER 1470 (supplementary text S4), and now skull 5, however, effectively falsify such a scenario (26).

Variation now established by the enlarged Dmanisi sample is crucial to identifying patterns of variation among and between paleospecies of early Homo. Several studies provide evidence for variation along and across a single evolving lineage of early Homo (27, 28), with differences between fossil specimens primarily reflecting intraspecific variation at the population level and evolutionary change within a species over time (Fig. 3A). Other studies, however, interpret variation in small samples of early African Homo as reflecting taxonomic species diversity (15, 2931) and posit three or more largely contemporary paleospecies (H. rudolfensis, H. habilis, and H. erectus). According to this interpretation (Fig. 3B), morphological differences among individual fossil specimens indicate interspecific differences and concomitant phyletic diversification.

Fig. 3

Two contrasting hypotheses about the evolution of early Homo. Hypothesis (A) posits that the fossil record samples paleodemes (horizontal bands, d1 to d4) of a phyletically evolving early Homo lineage, each with a separate geographic focus but overlapping in time and morphology. Selection and drift result in an overall morphocline from early generalized to late derived forms (diagonal arrow). This hypothesis is compatible with a wide range of variation within paleodemes of a lineage (dashed double arrows) and is most compatible with the evidence of morphological variation in the Dmanisi deme (white dots). Hypothesis (B) posits that the fossil record [same dots as in (A)] samples distinct paleospecies (vertical bands, s1 to s4) overlapping in time and space. This hypothesis predicts that morphological variation within each lineage is restricted (dashed double arrows), which is incompatible with the combined evidence of multiple contemporary but morphologically disparate crania from Dmanisi (white dots).

Testing these contrasting hypotheses depends on information about variation within paleodemes; that is, within populations of a fossil species at a given point in space and time (32) (Fig. 3). A fossil sample is likely to represent a paleodeme if two conditions are met. It must come from a spatially and temporally constrained stratigraphic and taphonomic setting [which is the case for Dmanisi (4, 5)], and within-sample variation must be similar in range and mode to within-deme variation in closely related extant species.

Geometric morphometric analysis and resampling statistics show that craniomandibular shape variation among the Dmanisi hominids is congruent with patterns and ranges of variation in chimpanzee and bonobo demes (Pan troglodytes troglodytes, P. t. verus, P. t. schweinfurthii, and P. paniscus) and in a global sample of H. sapiens (Fig. 4, supplementary text S5 to S7, and figs. S7 to S9). Within all groups, variation in cranial shape is mainly due to interindividual differences in size and orientation of the face relative to the braincase. The Dmanisi sample, including skull 5, thus represents normal within-deme variation, ranging from small-faced relatively orthognathic (typically female and/or subadult) individuals to large-faced relatively prognathic (typically male) individuals.

Fig. 4

Shape variation of the Dmanisi crania in comparative context. Dmanisi crania [2:D2282 (skull 2), 3:D2700 (skull 3 and picture), 4:D3444 (skull 4), 5: D4500 (skull 5 and picture)]; African early Homo (triangles with specimen names); A. africanus (asterisks); H. erectus Java (diamond); Kabwe and Steinheim (crosses); H. neanderthalensis (plus signs); H. sapiens (dots and pictures; populations from Africa, Australia, and America); P. troglodytes: P. t. troglodytes (solid squares and pictures), P. t. verus (vertical rectangles), P. t. schweinfurthii (horizontal rectangles); P. paniscus (open squares). Black symbols indicate adult individuals; gray symbols indicate subadult individuals. Large symbols indicate male and female averages. Shape component SC1 captures within-group cranial variation from large-faced/prognathic to small-faced/orthognathic individuals; SC2 captures shape change associated with grade shifts in neurocranial size between taxa.

Comparative analysis of nonmetric cranial characters also indicates that within-deme variation is substantial and wider than typically recognized (supplementary text S2 to S4 and table S4). The Dmanisi paleodeme shows variation in maxillofacial features, which have been interpreted as evidence for taxonomic diversity in eastern African fossils (15, 31).

The Dmanisi sample thus represents the best currently available example of cranial form and form variation within the early Homo lineage. It can tentatively be characterized as follows (figs. S2 and S3): face with flaring, vertically oriented zygomatic processes that are coplanar with the orbital plane; similar upper and midfacial widths; markedly angled zygomaticoalveolar crests; marked alveolar prognathism; C roots exhibiting parallel orientation relative to the cranial midplane; naso-alveolar clivus moderately sloping and laterally delimited by C and/or P3 juga; and ECV greater than ~500 cm3. These features also discriminate early Homo from nonrobust Australopithecus species, which exhibit a more protruding but less sloping naso-alveolar clivus; inclined maxillary zygomatic processes; moderately curved zygomaticoalveolar crests; and superomedially tapering canine juga, which laterally delimit the anterior surface of the maxilla. Other features of early Homo, such as three-rooted premolars, marked postorbital narrowing, a shallow glenoid fossa, and a moderate degree of basicranial flexion, represent pleisomorphic (primitive) character states variably shared with the Australopithecus species.


As already noted by Darwin, recognizing species diversity comes “at the expense of admitting much variation” within species [(33), p. 51]. Together with data from Au. afarensis paleodemes such as A.L. 333 (3436), Dmanisi adds to the growing evidence that intrademic and intraspecific variation in Plio-Pleistocene fossil hominids tends to be misinterpreted as species diversity, especially when single fossil specimens from different localities are compared with each other (37). Evidence from skull 5 and the other four Dmanisi specimens indicates that cranial shape variation within early Homo paleodemes was similar in mode and range to that seen within modern Pan demes. Furthermore, Dmanisi indicates that an important proportion of character state variation in nonmetric features reflects intrademic variation rather than real species diversity. These findings have several implications for the interpretation of morphological diversity in the fossil record of early Homo.

When seen from the Dmanisi perspective, morphological diversity in the African fossil Homo record around 1.8 Ma probably reflects variation between demes of a single evolving lineage, which is appropriately named H. erectus. The hypothesis of multiple independent lineages (paleospecies) (15, 31) appears less parsimonious, especially in the absence of empirical evidence for adaptation to separate ecological niches. The hypothesis of phyletic evolution within a single but polymorphic lineage raises a classificatory but not evolutionary dilemma, and it is premature to describe the rate(s) of evolution in this lineage, given the small available samples. Specimens previously attributed to H. ergaster are thus sensibly classified as a chronosubspecies, H. erectus ergaster. The Dmanisi population probably originated from an Early Pleistocene expansion of the H. erectus lineage from Africa, so it is sensibly placed within H. e. ergaster and formally designated as H. e. e. georgicus to denote the geographic location of this deme [thus retracting the species status given earlier to mandible D2600 (12)].

Given the scattered and fragmentary fossil record in Africa that predates Dmanisi, questions of earliest African Homo phylogenetics and classification remain unresolved. It remains to be tested whether all of the fossils currently allocated to the taxa H. habilis and H. rudolfensis belong to a single evolving Homo lineage. Although we regard this null hypothesis as parsimonious and fully compatible with new evidence from Dmanisi, alternative scenarios exist. Given the range of variation seen in the Dmanisi paleodeme, there is no convincing signature, at present, of early Homo cladogenesis. The African fossils that postdate the Dmanisi ensemble show brain size increase and correlated change in craniofacial morphology within the evolving lineage of H. erectus. Moreover, it is likely that both the underpinning of the East Asian dispersal of H. erectus, as well as the roots of subsequent H. erectus evolution in Africa (for example, OH 9, Daka), shared greater craniofacial robusticity.

The new evidence from the ancient Dmanisi deme of early Homo reinforces the strong African affinities previously recognized for this early Eurasian outpost of our genus (7). Variation between continents may thus provide insight into the evolutionary population dynamics of early Homo. It is well known from the upper Pleistocene dispersal of modern humans from Africa that genetic and phenetic variation within demes decreases with geographic distance from Africa, due to serial founder effects and population bottlenecks. For example, modern human phenetic variation in western and eastern Asia is 95 and 85% of African variation, respectively (38). The expansion of early Homo from Africa might have occurred at longer time scales, so direct comparisons must remain tentative. Nevertheless, the observation that Dmanisi conserves a substantial proportion of cranial shape and its variation among early African Homo speaks for genetic continuity between Africa and Eurasia. Because intrademic variation at Dmanisi is similar to that found in extant and extinct near relatives, the effective population size of early Homo in western Asia might have been similar [Ne > 10,000 (39)]. Furthermore, the remarkably large and robust dentognathic remains of early H. erectus from Java (Trinil/Sangiran) (40, 41) exhibit close affinities with skull 5 (tables S2 to S4). This provides evidence for morphological (and presumably underpinning genetic) continuity across large geographic distances, and for the preservation in East Asia of an appreciable proportion of the variation originally present within paleodemes of early Homo in Africa (20, 4244) (fig. S8).

Contrasting with early Homo population continuity across continents, current paleontological data indicate a low degree of similarity between contemporaneous mammalian genera in Africa and at Dmanisi (5), implying a generally low rate of faunal exchange between Africa and Asia before and around 1.8 Ma (45). As suggested earlier (46), the populational, ecological, and evolutionary dynamics of early Homo probably differed significantly from those of coeval large mammals, including other hominid lineages. Theoretical considerations indicate that the long-range dispersal rate of a population mainly depends on rates of reproduction and local habitat expansion (47) and that intragroup cooperation can play an important role in population persistence (48). Various features such as derived lower limb and foot morphologies (5, 49), tool-mediated widening of the dietary niche toward meat eating (4, 6), and increased levels of intragroup cooperation (3) might have led to increased rates of reproduction, survival, and mobility in early Homo and the consequent establishment of stable populations outside Africa. Skull 5 and the other members of the Dmanisi paleodeme now indicate that early Homo expanded from Africa to ultimately establish substantial populations in western Asia. This earliest hominid dispersal pre-dated any significant increase in brain size. Further analyses are required to test and modify the attendant hypotheses. Ultimately, identifying hominid paleodemes and assessing within-deme variation (32) will be key to understanding mechanisms of evolution and geographic dispersal of early Homo.

Supplementary Materials

Supplementary Text S1 to S6

Figs. S1 to S9

Tables S1 to S7

References (5096)

References and Notes

  1. Supplementary materials are available on Science Online.
  2. Acknowledgments: This research was supported by the Rustaveli Georgian National Science Foundation, the Swiss National Science Foundation, the U.S. National Science Foundation, the National Geographic Society, the L.S.B. Leakey Foundation, the American Philosophical Society, the American School for Prehistoric Research, a Rolex Award for Enterprise, BP Georgia, the Fundación Duques de Soria, the A.H. Schultz Foundation, and the Foundation for Scientific Research (University of Zurich). We thank the Dmanisi excavation team for excellent work; special thanks go to G. Kiladze. G. Nioradze was the first H. sapiens to encounter D4500. We gratefully acknowledge the work of the late G. Bumbiashvili, who photographed skull 5 and the other Dmanisi hominids. We thank G. Nemsadze (N. Kipshidze Central Clinical Hospital of Tbilisi State Medical University) and I. Tatishvili (A. Gvamichava Oncology National Center, Tbilisi) for CT data acquisition. We also thank the reviewers for their valuable comments, which helped to improve the manuscript. Dmanisi skull 5 is deposited at the Georgian National Museum, Tbilisi, Georgia.
View Abstract

Stay Connected to Science

Navigate This Article