Dmanisi hominins

Excavation site at Dmanisi in 2007

Dmanisi is located in southern Georgia, about 85 kilometres (52.8 miles) from the country's capital, Tbilisi. It was founded as a city in the Middle Ages and has thus been a site of archaeological interest for some time, with a prominent archaelogical excavation site being located within the ruins of the old city on a promontory overlooking the Mashavera and Pinazauri rivers.[14] Archaeological excavations began in 1936 on the initiative of historian Ivane Javakhishvili, who directed several expeditions. In 1982, archaeologists at Dmanisi discovered 3 metre (1 ft) deep pits, cut in compact sandy clay. The archaeologists believed the pits to have been made for some economic purpose in the Middle Ages and after cleaning them out, fossilized animal bones were discovered on the walls and bottom of the pits. The Georgian Paleobiological Institute of the Academy of Sciences was informed immediately and systematic palaeontological excavations began in 1983, but ended in 1991 on account of financial issues.[15]

Dmanisi

Location of Dmanisi in southern Georgia

During the 1983–1991 excavations, a large amount of animal fossils were collected, alongside some stone tools. The stone tools were quickly noted as highly archaic, far more primitive than other tools found in Eastern Europe. Biostratigraphically (dating through comparisons with fauna at other well-dated sites), the tools and fossils were dated to have been deposited at some point from the Late Pliocene to the Early Pleistocene.[15] New excavations at Dmanisi have been undertaken since 1991, wherein Georgian palaeontologists have been joined by German specialists from the Romano-Germanic Museum in Cologne, Germany.[15] From 1991 to 1999, expeditions were completely funded by the Romano-Germanic Museum and carried out annually.[15]

Georgian archaeologist and anthropologist David Lordkipanidze (left) showing a hominin jaw discovered at Dmanisi to American ambassador John R. Bass in 2010

The expedition in 1991 was highly productive, uncovering abundant animal fossils and a considerable quantity of stone tools. On the morning of 25 September, a group of young archaeologists, led by Medea Nioradze and Antje Justus uncovered a bow-shaped bone.[15] Justus was immediately intrigued by the unusual shape of the bone and intuitively guessed that it was a mandible, which was confirmed once it was unearthed more clearly and its teeth were revealed.[16]

As the heads of the expedition, Georgian archaeologists and anthropologist Abesalom Vekua and David Lordkipanidze (then in Tbilisi) were summoned to the site and on the next morning, the mandible was freed from the rock around it, a complicated process that took nearly an entire day. Once freed, the mandible was unmistakeably the jaw of a primate and importantly, it preserved a complete row of teeth with little sign of wear. The lack of wear suggested that the primate would have been young, about 20–24 years old, though its classification was as of yet unknown. After they returned to Tbilisi, the mandible was studied in detail by Vekua, Lordkipanidze and archaeologist Leo Gabunia. It was quickly determined to represent a hominid, though its precise position within the family was unclear. Although a number of primitive features were observed, it was clear that the fossil (now given the designation D211) was the most similar to fossils of Homo, not earlier australopithecines. After prolonged discussion, Vekua and Gabunia came to the conclusion that the Dmanisi hominin was probably an early Homo erectus, and that it represented the earliest Homo outside Africa. This suggestion, and the significance of the fossils, was confirmed once the basalts lying directly below the Pleistocene sediments were determined to be about 1.8 million years old.[16]

Excavations continued at the site, though hominin remains proved to be rare. In 1997, the right metatarsal bone of a hominin was discovered in the same layer as the jaw, convincing the scientists that there were more hominin remains to be found at Dmanisi. Further discoveries were made in May 1999. Because of long-lasting periods of rainfall, the site of the excavation was damaged. Archaeologist and expedition member Gocha Kiladze found a thin, coin-sized bone and since it appeared to be a skull fragment, brough it to Tbilisi, where it was determined to be the fragment of a primate skull. Kiladze, Vekua, Lordkipanidze, alongside archaeologist Kakha Kakhiani and the head of the 1999 expedition, archaeologist Giorgi Kopaliani, then visited the site and discovered further fragments. With these fragments, they were able to piece together the skull of an archaic human, with broken off teeth and a broken off upper jaw. That same year, a more well-preserved skull was discovered and together, the two skulls allowed for inferences as to the nature and classification of the fossil hominins.[5] The first skull, dubbed Skull 2 and given the designation D2282, had an estimated cranial capacity of 650 cc and the second skull, Skull 1 (D2280) had an estimated capacity of 775 cc.[3] After studying the fossils for almost a year, it was determined that they somewhat differed from H. erectus in their jaws and skulls and were closer to the earlier African species H. ergaster (now considered an early African representative of H. erectus by some). The discovery of the two skulls was highly publicised in international media and the Georgian fossils were for the first time widely acknowledged as the earliest known hominins outside of Africa.[5]

Dmanisi Skull 5 (D4500) at the time of discovery in 2005

More discoveries followed. In 2000, another hominin jaw (D2600) was discovered,[3] this time at a slightly lower layer (i.e. older) than the rest of the fossils.[17] This jaw was very large and had highly developed posterior molar teeth. The following year, Skull 3 (D2700) and its corresponding jaw (D2735) was discovered, almost perfectly preserved.[3] On account of its erupting wisdom teeth, Skull 3 was determined to be the skull of a subadult.[4] In 2002, the toothless skull of an old individual, Skull 4 (D3444, the associated jaw, D3900, was discovered in 2003) was discovered. Both Skull 3 and Skull 4 were noted as preserving a series of very primitive characteristics. The final skull, Skull 5 (D4500), was discovered in 2005. The skull matched the jaw found in 2000 and the two were concluded as having come from the same individual. The cranial capacity of Skull 5, at 546 cc, is the lowest of all the skulls and the skull is also the most robust out of all the specimens.[3] The skulls were significant not only in their set of unique features. Skull 5 was the first found completely preserved adult hominin skull from the Early Pleistocene,[4] and Skull 4 is the only toothless hominin discovered in such early sediments.[18]

In addition to the skulls, about a hundred postcranial remains have been discovered.[4] The first postcranial fossil discovered was a third metatarsal bone, recovered in 1997.[19] Postcranial fossils compromise bones from all parts of the body and include parts of the arms,[20] legs,[21] axial skeleton (vertebrae and ribs)[21] and feet.[22] The bones, some of them confidently associated with Skull 3, are from both adolescent and adult individuals.[20]

Together, the fossils at Dmanisi represent the most complete and richest collection of early Homo fossils at a single site with a comparable temporal context. The variability in age (i.e. Skull 3 being subadult and Skull 4 being significantly older) and presumably sex also gives unique insight into the variability in early populations of Homo.[4]

The Dmanisi Skulls

Image	Skull & specimen number(s)	Cranial capacity	Discovered	Published	Notes
	Dmanisi Skull 1 D2280	775 cc[3]	1999[5]	2000[23]	Skullcap of an adult individual.[24] Interpreted as male on account of the thick brow ridges and other minor skull features.[25]
	Dmanisi Skull 2 D2282 (mandible D211)	650 cc[3]	1999[5] (mandible in 1991)[16]	2000[23] (mandible in 1995)[26]	Gracile features, interpreted as the skull of an adolescent female.[24][25]
	Dmanisi Skull 3 D2700 (mandible D2735)	600 cc[27]	2001[3]	2002[28]	Skull of a young individual. Generally gracile morphology, but the upper canine teeth have large crowns and massive roots; making assessing its sex difficult.[29] A handful of traits suggests an interpretation as male to be appropriate.[25]
	Dmanisi Skull 4 D3444 (mandible D3900)	625 cc[30]	2002[3] (mandible in 2003)[3]	2005[31]/2006[32]	Skull of an elderly individual that had lost all but one tooth.[33] Interpreted as male.[34]
	Dmanisi Skull 5 D4500 (mandible D2600)	546 cc[3]	2005[3] (mandible in 2000)[3]	2013[35] (mandible in 2002)[36]	Skull of an adult individual. Skull 5 is the first completely preserved Early Pleistocene adult hominid skull found. Interpreted as male on account of its massive and prominent cranial features.[8][25]

The classification of the Dmanisi hominins is disputed and a discussion on whether they represent an early form of H. erectus, a distinct species of their own dubbed H. georgicus or something else entirely are ongoing.[2][13]

Replicas of the D211 (right, discovered in 1991) and D2735 (left, discovered in 2001) mandibles

The D211 mandible was described in 1995 by Gabunia and Vekua, who compared it to the mandibles of other early Homo; H. ergaster, H. habilis and H. rudolfensis in Africa, H. erectus in Africa and Asia and archaic H. sapiens in Europe. Out of all the other Homo, the Dmanisi jaw was determined to be the most similar to fossils attributed to H. erectus in Africa, such as OH 22, with fewer similarities to H. erectus fossils in Asia. Closer relation to early Homo such as H. habilis and H. rudolfensis was ruled out on account of the jaw clearly being smaller than those of these species, especially under the teeth in the cheeks. Tooth size was noted as small and in terms of morphology the teeth were found to be especially narrow, though the teeth were considered to be more or less comparable to teeth in H. ergaster/H. erectus, with the most comparable fossils in terms of tooth size being OH 22, the Zhoukoudian skulls, Mauer 1, Tighenif 3, KNM ER 992 and KNM WT 15000.[37] Compared to these other specimens, D211 was found to be most distinctive in its smaller size and in the teeth reducing in size distally (away from the main mass). With the many found similarities considered, Gabunia and Vekua concluded that "the most reasonable interpretation of this jaw is that it belonged to a population of H. erectus".[38]

The mixture of primitive and derived traits was also noted in a 1996 analysis by palaeoanthropologists Günter Bräuer and Michael Shultz, who compared the mandible to fossil mandibles of early Homo and H. ergaster/H. erectus. In contrast to the 1995 description, Bräuer and Schultz concluded that based on the presence of several derived traits (finding ten major traits deemed similar to derived H. erectus), the Dmanisi hominin probably was a "progressive" (i.e. derived, late) representative of H. erectus. Though they noted that such a classification would be doubtful since the fossil site was 1.8 million years old, they stated that the jaw could still represent a very early progressive form of H. erectus.[39] In 1998, palaeoanthropologists Antonio Rosas and José Bermúdez De Castro examined the jaw and also found it to exhibit a unique combination of features, some of which, in isolation, could be seen as morphological extremes within Homo. Whereas some traits, such as the talonid (crushing heel) of the premolar tooth P4 being reduced, were interpreted as more derived, many of the features exhibited, including the overall morphology and dimensions of the teeth and jaw, were considered primitive. Because a similar mosaic of jaw traits have been found in African fossils referred to H. ergaster, Rosas and De Castro suggested that the jaw best be classified as Homo sp. indet. (aff. ergaster).[40]

Dmanisi Skull 2 (D2282)

Gabunia and colleagues described Skulls 1 and 2 in 2000 and noted that though the facial skeleton of Skull 2 was fragmentary, its estimated proportions and reconstructed morphology were very similar to specimens of H. ergaster recovered at Koobi Fora in Kenya and that the dentition in particular was reminiscent in size and morphology to the dentition known from H. ergaster specimens such as KNM-WT 15000 and KNM ER 3733. Gabunia and colleagues noted that although the cranial capacities of the Dmanisi skulls was lower than the average of H. ergaster, the skulls differed from early Homo such as H. habilis and H. rudolfensis in a number of traits that more closely aligned the fossils with H. ergaster. Traits differentiating the Dmanisi hominins from early Homo were noted as including the well-developed brow ridge, the lack of cresting (otherwise present in early Homo and in other great apes), large orbits, the premolar teeth in the upper jaw having single roots and the angulation of the cranial vault. Numerous traits were noted as suggesting a close relation to H. ergaster, including the presence and morphology of the brow ridge, the overall proportions of the facial skeleton, the relative narrowness of the skull beyond the face (post-orbital constriction) as well as a comparable height of the cranial vault and the thickness of the cranial vault bones. The same features typically used to distinguish H. ergaster from Asian specimens of H. erectus were found to distinguish the Dmanisi fossils from Asian H. erectus; notably the lower cranial vault and somewhat thinner cranial vault bones in H. erectus and the smaller cranial capacity of the Dmanisi fossils. A handful of features were noted as present in the Dmanisi fossils and Asian H. erectus, but not H. ergaster, such as the presence of a supramastoid crest). Since these features also appeared in some African fossils, such as Olduvai hominids 9 and 12, they were deemed to not hold "any special phylogenetic significance".[41] Gabunia and colleagues concluded by referring the Dmanisi fossils to Homo ex. gr. ergaster ("ex. gr. ergaster" meaning "of the group including ergaster") on account of the combination of features which they perceived to align the fossils closer to H. ergaster than to earlier Homo or Asian H. erectus.[42] Gabunia and colleagues stated that the combination of features made it a possibility that the Dmanisi hominins were forerunners of both later H. erectus in Asia and hominins ancestral to H. sapiens.[43]

Replica of Dmanisi Skull 3 (D2700) and its associated mandible (D2735)

In 2002, Vekua and colleagues described Skull 3 (D2700), including its associated mandible (D2735). Through comparisons with other hominin fossils from around Lake Turkana in Kenya, it was determined that the skull shared a number of particular anatomical features with the skulls of H. ergaster/H. erectus, such as the shape of the squamous part of temporal bone, the zygomatic (cheekbone) root being very thick and there being keeling along the sagittal midline. This means that despite the cranium of Skull 3 being exceptionally small (with its cranial capacity of 600 cc being near the mean of H. habilis), and its face closely resembling that of KNM ER 1813 (a specimen of H. habilis), especially in profile, Skull 3 looks more like a small H. erectus than a H. habilis. Despite some differences, D2700 was deemed to overall be similar to Skulls 1 and 2, with its mandible resembling the D211 mandible and Vekua and colleagues viewed all fossils as belonging to one and the same taxon, not seeing sufficient ground for assigning them to multiple species. Skull 3 was provisionally referred to Homo erectus (=ergaster). The researchers also briefly mentioned the D2600 mandible (at this point not yet described) as underscoring the fact that some of the hominin fossils found at Dmanisi departed from the typical morphology of H. erectus, but that Skull 3 could nevertheless be considered as an "extremely small-brained representative of this species".[27]

The D2600 mandible was also described in 2002 by Gabunia, Vekua and Lordkipanidze, together with French archaeologists and palaeoanthropologists Henry de Lumley and Marie-Antionette de Lumley.[24] Before the mandible's discovery, the classification of the Dmanisi hominins as H. ergaster or H. erectus had seemed relatively clear, but D2600 differed in its large size, morphological features and teeth proportions not only from the previously discovered jaw at Dmanisi but also from all other hominin jaws found to date, blending primitive features otherwise seen in Australopithecus and early Homo with derived features otherwise seen in H. erectus.[18] Gabunia and colleagues deemed the differences in size and proportion to not only H. erectus and H. ergaster, but also to H. habilis and H. rudolfensis to be sufficient for the creation of a new species, which they dubbed Homo georgicus. D2600 was designated as the type specimen of H. georgicus, and all the previously discovered hominin remains were referred to this new species.[24] Although Skulls 1 and 2 were far more gracile than the new, robust mandible, the researchers determined that this was an example of marked sexual dimorphism within only one species, believing the gracile fossils to represent female individuals. Gabunia and colleagues interpreted H. georgicus as a descendant of H. habilis or H. rudolfensis and an early species "near the roots of the Homo branch", "foretelling the emergence of Homo ergaster".[44] Palaeoanthropologist Sang-Hee Lee supported the classification of all the Dmanisi hominin fossils as belonging to the same species (though made no comment on if that species should be H. erectus or H. georgicus) in 2005, noting that despite the differences in brain capacity between the skulls, they were not more morphologically distinct from each other than individuals of different sexes in modern great apes.[45]

The D2600 mandible (later associated with Skull 5), designated as the type specimen of Homo georgicus in 2002

Lordkipanidze and colleagues described Skull 4 and its mandible in 2006, noting that it was similar to the fossils discovered previously and stating that with the possible exception of the D2600 mandible, all of the Dmanisi fossils were assignable to a single species. The researchers noted that the hominin population shared several traits with more primitive hominins, such as Australopithecus and earlier Homo, notably the low cranial capacities, but that many other aspects of their morphology aligned them with H. erectus, particularly African H. erectus (H. ergaster), though they do possess some traits otherwise only seen in Asian H. erectus. On an anatomical ground, they argued that the Dmanisi hominins were close to the "stem" of H. erectus, with other H. erectus groups being more derived, and that the Georgian population could potentially be ancestral to later branches of the species in Asia.[46] That same year, a comparative analysis of Skulls 1 to 4 and the D2600 mandible by palaeoanthropologist G. Philip Rightmire, Lordkipanidze and Vekua again concluded that Skulls 1 through 4 could be assigned to the same species, but that the status of D2600 was more questionable.[47] They noted that though the fossils were similar to H. habilis in some respects, especially in size and (for some) cranial capacity, they shared far more features with H. erectus. In this respect, many of the primitive features could simply be interpreted as primitive retentions. Rightmire, Lordkipanidze and Vekua concluded that if some of the H. habilis-like traits, such as the size, cranial capacity and parts of the facial morphology, were considered plesiomorphic and primitive retentions, there would be no reason to exclude Skulls 1 to 4 from H. erectus. Because of some unique traits and these primitive retentions, Rightmire, Lordkipanidze and Vekua thus proposed that the Dmanisi hominins represented an early subspecies of H. erectus.[10] Though the others were unsure, Vekua supported the classification of D2600 as representing a distinct species separate from the rest of the fossils, preferring to keep its designation as H. georgicus.[48] They noted that if future analyses suggested that D2600 belonged to the same hominin population as the other fossils, the subspecies designation would appropriately be Homo erectus georgicus, but that if it was distinct (as H. georgicus), a new subspecies name would have to be selected for the other fossils.[10]

Dmanisi Skull 3 compared to skulls of H. naledi, H. habilis and H. floresiensis.

A 2006 comparative analysis of D211 and D2600 by palaeoanthropologists Matthew M. Skinner, Adam D. Gordon and Nicole J. Collard found that the degree of dimorphism expressed between the two mandibles was greater than expected in modern great apes and human, as well as in other extinct hominin species. With the dimorphism found to be significantly greater than what is expected in genera such as Pan, Gorilla and Homo, Skinner, Gordon and Collard suggested two alternative hypotheses; either that the fossils represented a single taxon with unusually high sexual dimorphism whose inclusion in Homo was thus doubtful, or that D2600 should be considered as a representative of a separate, second species of hominins (i.e. H. georgicus).[49] A more detailed 2008 comparative analysis of the mandibles, taking more anatomical features into account, by Rightmire, Lordkipanidze and palaeoanthropologist Adam Van Arsdale concluded that while the dimorphism between the mandibles was excessive when compared to modern humans, and to some chimpanzees, it was comparable to (or in cases, less than) the dimorphism between gorillas. They concluded that "in our view, there are currently no compelling anatomical grounds for sorting any of the Dmanisi fossils to other than a single species", but noted that this species would have possessed sexual dimorphism greater than more recent Homo.[50] Preferring the designation of H. erectus, the researchers noted that although H. erectus is generally held to not be this diamorphic, some fossils, such as smaller skulls recovered at Ileret and Olorgesailie in Kenya and larger skulls recovered at Olduvai Gorge, Tanzania and Bouri, Ethiopia, could disprove this notion.[51]

A 2008 analysis of the teeth of Skulls 2 and 3 and the D2600 mandible by Lordkipanidze, Vekua and palaeoanthropologists María Martinón-Torres, José María Bermúdez de Castro, Aida Gómez-Robles, Ann Mergvelashvili and Leyre Prado found that like other parts of the fossils, the teeth too showed a combination of primitive Australopithecus- and H. habilis-type traits and more derived H. erectus-type traits. The teeth of Skulls 2 and 3 were found to be similar, whereas D2600 somewhat diverged in the size of the teeth and in the morphology of its roots. However, this does not necessarily on its own indicate that D2600 represents another population or species since the same size differences in the teeth are seen in fossils referred to H. habilis.[52] Palaeoanthropologist P. James Macaluso Jr. came to the same conclusion in a 2010 analysis, concluding that Skulls 2 and 3 could comfortably be referred to the same species, but whether D2600 could also be referred to the same species as the rest was less clear.[53]

Another view of Dmanisi Skull 3 (D2700 & D2735)

Skull 5, recovered in 2005 and described in 2013 by Lordkipanidze and colleagues, was upon its description determined to be from the same individual as the D2600 mandible and together, the two fossils significantly expanded the morphological range of the Dmanisi hominin fossils.[8] Lordkipanidze and colleagues interpreted Skull 5 as part of the same population as the rest of the Dmanisi fossils. This conclusion was reached since all the Dmanisi hominin fossils come from the same spatial and temporal context and because new analyses of the fossil material showed that the Dmanisi skulls were within the range of variation shown in chimpanzee (Pan troglodytes troglodytes, P. t. verus, and P. t. schweinfurthii), bonobo (P. paniscus) and modern human samples. Interspecific variation in all of these samples was mostly due to variation in size and in the orientation of the face relative to the braincase, both differences also expressed within the Dmanisi sample. Lordkipanidze and colleagues interpreted that the small-faced and more orthognathic skulls represented females and/or subadults and that the more prognathic and large-faced skulls represented males.[54] The large degree of variation expressed in the Dmanisi fossils led Lordkipanidze and colleagues to suggest that the variation seen in other Pliocene and Pleistocene hominid fossils, typically used to justify several distinct fossil species, might have been misinterpreted as species diversity. Thus, the morphological diversity in contemporary African hominins, typically used to justify H. ergaster as a species distinct from H. erectus, might thus instead be due to regional variation in a single evolving lineage of hominins (H. erectus). With this in mind, the classification of the African material as H. erectus ergaster (a chronosubspecies rather than a distinct species) was suggested and since the Dmanisi hominins are believed to have originated from an early migration by the H. erectus lineage out of Africa, it was determined that they be best placed within H. e. ergaster with a quadrinomial (4-part) name; H. e. e. georgicus. The researchers considered it possible that earlier Homo, such as H. habilis and H. rudolfensis also belonged to the same single evolving lineage of Homo, though no morphological comparisons were made to test this theory.[55]

Palaeoanthropologists Jeffrey H. Schwartz, Ian Tattersall and Zhang Chi responded to the 2013 study in 2014, disagreeing with the idea that all five skulls were from the same species. Schwartz, Tattersall and Chi also suggested that the use of a sub-subspecies such as H. e. e. georgicus, with a quadrinomial name, was invalid in zoological nomenclature. Most importantly, Schwartz, Tattersall and Chi questioned if the morphological comparisons were detailed enough to come to this conclusion and questioned the methods which Lordkipanidze and colleagues had used to determine what is and is not interspecific variation. The researchers did not see the fact that the fossils were from the same site and a relatively short time period as enough to determine that they all came from the same species and that the previous claims of Gorilla-type mandibular variation but H. sapiens/Pan-type cranial variation could not both be correct at the same time. They also questioned if all morphological differences could truly be attributed to age, wear and pathology. Several traits within the skulls and teeth of all the Dmanisi skulls were put forward as "potentially species-distinguishing features" and Schwartz, Tattersall and Chi concluded that at least the D2600 mandible, and thus Skull 5 as a whole, should remain classified as a distinct species, H. georgicus, writing that "to deny this hominin a distinct identity is effectively to deny the utility of morphology in systematics, a radical proposition to which few would subscribe".[12]

Comparison of the mandibles of various species of Homo; D2600 (belonging to Skull 5) and D211 (belonging to Skull 2) are featured

The Dmanisi research team responded to Schwartz, Tattersall and Chi in the same year, maintaining that the fossils represented a single species. They noted that the distinction of H. georgicus, and the further suggestion that some of the other skulls might represent distinct taxa as well, would mean that Dmanisi would have been home to at least four different hominid taxa and thus "hold the world record in hominid paleospecies diversity documented at a single site that extends over a mere 40 m², and probably over a mere couple of centuries". The Dmanisi team wrote that Schwartz, Tattersall and Chi had deliberately ignored previous morphological analyses and also noted that character state variation in Asian and African Homo specimens, and the Dmanisi fossils, suggest that the fossil cannot be assigned to different species, accusing Schwartz, Ttaersall and Chi of effectively denying the morphological evidence from the Dmanisi fossils that did not fit with their hypothesis. One of the primary distinguishing features noted by Schwartz, Tattersall and Chi, the number of premolar tooth roots, was pointed out as not actually carrying taxonomical significance since modern Sub-Saharan humans exhibit significant variation in this specific trait. The name Homo erectus ergaster georgicus was also defended in that it was used to denote a local population of a subspecies, similar to how quadrinomials are used in botany. The researchers pointed out that although the use of quadrinomials is not regulated by the International Code of Zoological Nomenclature, it is not considered invalid.[11]

A 2017 analysis of Skull 5 specifically, with comparisons to the other skulls and to skulls of H. sapiens, Paranthropus boisei and other archaic hominins, by the team reaffirmed that the variation between the Dmanisi fossils was not excessive relative to the variation in most other hominins, with some features, such as certain midfacial measurements, even being more variable in modern humans.[56] Although certain traits were noted as setting Skull 5 "toward the periphery of the Dmanisi shape distribution", they concluded that "neither these differences, nor the proportions of the D2600 mandible, offer sufficient grounds for labeling Skull 5 as the 'holotype of the morphologically very distinctive species H. georgicus'". The results of the analysis, which compared the skulls to many specimens of both H. erectus and H. habilis somewhat questioned the current recognition of species-level diversity in early Homo in so far that the Dmanisi hominins were found to broadly share many similarities with both species. The researchers found that the Dmanisi hominins "cannot unequivocally be referred either to H. habilis or to H. erectus" and that there, in regards to early Homo, was a "continuum of forms"; Skull 5 appears to share many primitive features with H. habilis whereas Skull 1, with the largest brain, is more similar to African H. ergaster/H. erectus.[57] This led the researchers to hypothesize that H. erectus and H. habilis constitute a single evolutionary lineage which emerged in Africa and later spread throughout Eurasia. Phylogenetically, the Dmanisi population was suggested to represent a part of an anagenetic sequence, descended from H. habilis and ancestral to later H. erectus, placed near the base of the H. erectus lineage and already differentiated from H. habilis.[58] Although represented by only a few fossils, the morphology of H. rudolfensis was determined to be unique, suggesting that H. rudolfensis was a distinct species, part of a separate evolutionary lineage.[59]

Replica of Skull 5 (top) and a forensic reconstruction based on the skull (bottom)

The cranial capacity of the Dmanisi hominins ranges from 546 to 775 cc,[4] with an average of 631 cc.[64] As such, their brain size overlaps with that of H. habilis (c. 548–680 cc) and falls below the standard cranial capacity otherwise ascribed to H. erectus (800–1000 cc[4]). The encephalization quotient (a measure of brain size relative to body size) of the Dmanisi hominins (based on Skulls 1 to 4) is in the range of 2.6–3.1, at the lower end of estimates for H. ergaster/H. erectus and more similar to H. habilis and australopithecines.[22] The encephalization quotient of Skull 5 was estimated at 2.4, within the range of variation for Australopithecus.[65]

The only fully complete skull found at Dmanisi is Skull 5, which can be distinguished from all other known fossil Homo specimens by its large prognathic face and small braincase.[65] The combination of large teeth and large face with a small braincase is otherwise unknown in early Homo,[2] and the two features have previously separately been used to define different species. Had the braincase and face of Skull 5 been found as separate fossils at different localities, it is likely that they would have been attributed to different species.[66] Despite the exterior morphological similarities to earlier Homo, the anatomy of its braincase is considerably more similar to later species, such as H. erectus.[2]

Skull 5 is somewhat different from the other Dmanisi individuals, with more prominent and massive cranial structures (i. e. the face and jaw).[8] This does not necessarily mean that Skull 5 represents a different hominin population; the two other more or less complete skulls, Skulls 3 and 4, are from an adolescent (and thus not fully developed) and an old individual, respectively. Skull 5 indicates that small brains, large faces and a generally prognathic and robust morphology was all within the range of variation of the Dmanisi hominin population. Based on the skulls and the postcranial material, the Dmanisi hominins appears to have been small-brained individuals with stature, body mass and limb proportions at the lower range limit of modern human variation.[65]

Despite the apparent differences between the skulls, many features are shared between them. They all share a relatively small brain size and brow ridges are prominent in all of the adult specimens, only being less prominent in the subadult Skull 3. Although it is most massive in Skull 5, the face is relatively prognathic and massive in all of the skulls.[67]

Full-body reconstruction based on postcranial fossils and Skull 3, by Élisabeth Daynès, Museo de la Evolución Humana, Burgos, Spain.

Prior to the discovery of the Dmanisi fossils, knowledge of postcranial morphology in early Homo had been very limited. Well-preserved fossils of earlier hominins, such as Australopithecus and later Homo, such as the well-preserved skeleton of KNM WT 15000 ("Turkana Boy"; a 1.55 million year old H. ergaster/H. erectus), gave little insight into early transitions in body proportions and stature. Australopithecus were small, about 105 cm (3.4 ft) tall, and had limb proportions intermediate between those of modern humans and those of other great apes, whereas the body proportions and stature of Turkana Boy were more or less modern. Postcranial fossils attributed to H. habilis and H. rudolfensis are fragmentary, and so the time and means of transition from hominins capable of bipedalism (Australopithecus) to hominins that were obligately bipedal (H. ergaster) remained unclear. In these respects, the Dmanisi fossils fill in a number of gaps.[22]

Through calculations based on the size of their bones (no complete skeleton has yet been recovered), the Dmanisi individuals were approximately 145–166 cm (4.8–5.4 ft) tall and weighed about 40–50 kg (88–110 lbs). They were smaller than H. ergaster in Africa, possibly either due to being more primitive (H. habilis was also smaller than H. ergaster) or due to having adapted to a different palaeoecological context. Limb proportions (measured through the length of the femur relative to the tibia) in the Dmanisi fossils are comparable to those of modern humans, but are also comparable to some of the earliest Homo and fossils referred to Australopithecus garhi, dated to 2.5 million years old. In terms of the absolute length of the legs, the Dmanisi hominins were more similar to later Homo (including modern humans) than to australopithecines,[22] though the length of legs and the morphology of the metatarsals in the Dmanisi hominins was not as derived as later H. ergaster/H. erectus (such as KNM WT 15000). This might indicate that the evolution of improved walking and running performance was not a sudden change, but a continual process throughout the Early and Middle Pleistocene.[68]

Humeral torsion (the angle formed between the proximal and distal articular axis of the humerus) influences the range of movement and the orientation of the arms relative to the torso. In modern humans, the scapula (which might otherwise restrict movement) is placed dorsally, which is compensated by a high degree of humeral torsion. Comparably, the torsion in the Dmanisi fossils is quite low, which indicates differing arm movement and orientation. It might mean that the arms would have been habitually oriented more supinely (horizontally) and that the shoulder girdle might have been positioned more laterally. Athletes that require high levels of mobility in their arms tend to have reduced humeral torsion, and the Dmanisi hominins might thus have been capable of a diverse range of arm movement. Humeral torsion is also low (or entirely absent) in H. floresiensis, which means that this might be a basal trait in Homo. Either way, the functionality and morphology of the arms in the Dmanisi hominins appears to have been more similar to the arms of earlier Homo or australopithecines than to modern humans.[69]

Overall, the spine in the Dmanisi hominins appears to have been more similar to the spines of modern humans and early H. erectus than to the spines of australopithecines. The fossil vertebrae recovered at Dmanisi show lumbar lordosis, the orientation of the facet joints suggests that the range of spinal flexion in the Dmanisi hominins was comparable to modern humans and the relatively large cross-sectional areas of the vertebrae indicates resistance to increased compressive loads, suggesting that the hominins were capable of running and long-range walking. Because fossils of the lower leg, from the femur down to the metatarsals, have been found, it is possible to reconstruct the orientation and positioning of the feet of the Dmanisi hominins relative to their walking direction. In the Dmanisi hominins, the feet would have been oriented more medially (closer together) and load would have been distributed more evenly over the rays (the grooves separating the thicker areas of the foot) than in modern humans. Despite these differences, the bones recovered suggest that the feet were overall similar to the feet of modern humans.[69] In 2008, palaeoanthropologists Ian J. Wallace, Brigitte Demes, William L. Jungers, Martin Alvero and Anne Su stated that they believed that the Dmanisi fossils were too fragmentary to infer the position of the feet (as medially positioned) with this much certainty, believing that more fossils, particularly of the pelvis and additional foot bones, were required.[70]

A Oldowan stone tool from Dmanisi (right, replica), compared to a later Acheulean stone tool (left)

Over 10,000 stone tools have been recovered at Dmanisi and their stratigraphic and spatial concentrations suggests a complex record of several reoccupations at the site. Most of the tools recovered are flake tools, but a smaller number of lithic cores and choppers have also been recovered. The raw materials to make these stone tools probably came from the rivers and outcrops near the fossil site.[34][77] The presence of cores, flakes and chunks in addition to finished tools show that all the stages of knapping (shaping of stone to create tools) took place at Dmanisi. Although the technique was not very elaborate, quality rocks (such as volcanic, magmatic and sedimentary stones as well as silicified tuff) were used. The precise technique used differed from stone to stone, influenced by the shape of the initial stone. No new angles appear to have been created through the process.[77]

Flakes and choppers are generally considered to be quite simple tools and the tools found at Dmanisi are much the same as the tools of the Oldowan tradition created by hominins in Africa at least nearly a million years earlier. In addition to the tools found at the site, many unmodified stones that must have originated elsewhere on account of their mineralogical composition (meaning they had not arrived there naturally, but had been brought by hominins) have also been recovered. Larger unmodified stones may have been used as tools for smashing bones, cutting meat and pounding flesh whereas smaller stones would have served other purposes, such as throwing.[34]

The large collections of manuports (unmodified stones moved from their natural context) recovered at Dmanisi are generally interpreted as stone reserves created by the hominins to avoid repeated visits to stone collection sites.[77]

The small Dmanisi hominins may have employed aggressive scavenging, throwing small rocks to pilfer food from local carnivores. It is possible that this power-scavenging was done in groups for protection, and it may have led to the development of kinship-dependent social cooperation.[34]

There is also indirect evidence of social cooperation in Skull 4, which is from an individual that had lost all but a single tooth by the time of his death.[34] The old individual would have lived for a relatively long time after losing the teeth, indicated by the sockets of the teeth roots having been filled with bone tissue, something that is only possible if the individual in question is alive.[18] Without fire to cook food, it would have been difficult for a toothless individual to survive for several years in a periodically cold environment. Though it is possible, through the use of pounding tools, that he would have survived on his own through consuming soft animal tissues, such as brains and marrow, a more compelling possibility is that he might have been cared for by other members of his species.[34]