The study was carried out on the original fossil at the Georgian National Museum (Tbilisi). The D2600 specimen consists of a mandibular bone that presents fractures principally at the mandibular rami. The mandible preserves in situ all teeth except the right second premolar (RP4), left first molar (LM1) and second premolar (LP4) (Fig. 1).

The degree of wear (or wear category) was determined following the eight-score system proposed by Molnar (1971), based on the amount of dentine exposure. The plane (or direction of the wear) was assessed from mesial and buccal views and following the eight-grade category of Molnar (1971). The occlusal morphology (occlusal surface form) was classified following Molnar (1971). The degree of dentine exposure on mesial and distal surfaces (approximal attrition facets) was scored according to the four-degree system developed by Hillson (2008).

We employed the definition and classification of Dias and Tyles (2007) for the assessment of periodontal cysts.

We measured the amount of root exposure, calculated as the distance between the AC and cemento–enamel junction (CEJ). It was wrongly assumed that the increased distance between the AC and the CEJ was exclusively reflecting the occurrence of periodontal disease. On the contrary, when the distance accompanies severe dental wear it can be a sign of continuous eruption, especially when there is lack of pathological signs (Clarke and Hirsch, 1991, Ogden, 2008a, Ogden, 2008b and Ten Cate, 1998). Continuous eruption is one of the three physiological tooth movements recognised to compensate for severe dental wear (e.g., Begg, 1954, d’Incau et al., 2012, Hylander, 1977 and Kaifu et al., 2003).

The identification of periodontal disease in osteological collections is justified by the appearance of morphological changes in the alveolar bone crest (i.e., porosity, blunt-flat rim, ragged margin) and formation of a periodontal pocket of funnel shape (>5 mm) around the root. We applied Ogden (2008a) score system to assess the evidence (scores 2–4) or lack of signs (score 1) of periodontal disease in the alveolar bone.

Davies’ definition of fenestration and dehiscence (Davies et al., 1974) was used to identify and classify the openings observed in the alveolar process. Dehiscence is characterised by “v” shape and the involvement of the alveolar margin. On the contrary, fenestration displays a quadrangular morphology and does not affect the alveolar margin (Clarke and Hirsch, 1991, Hillson, 2008, Nimigean et al., 2009 and Rupprecht et al., 2001).

We used Pattersons’ (1984) classification to distinguish between two forms of enamel breakage: (a) minor chipping and (b) major segments of enamel loss or fracture.

To determine the moment of occurrence – pre-mortem vs. peri-mortem – it is important to analyse the morphology of the borders and the colouration of the fracture (Milner and Larsen, 1991, Scott and Winn, 2011 and Smith, 1984); for this we followed Smiths’ (1984) criteria.

Tooth rotation is defined as the tooth movement around its long axis (Harris and Corruccini, 2008, Natsume et al., 2006 and Primožič et al., 2012). To assess the degree of tooth rotation (normal versus abnormal), we followed Natsume et al. (2006). The angle of rotation was calculated between: (a) the line perpendicular to the lingual surface of the tooth if this was in its correct position and (b) the line perpendicular to the current lingual surface of the tooth (see Fig. 2).

Finally, and for comparative purposes, we performed a paleopathological analysis of several original specimens, belonging to different species that displayed severe dental wear and dentognathic pathologies (Table 1).

In this individual, wear has proceeded to such an extreme degree that no enamel is left on the occlusal surfaces of the majority of teeth. The pulp chamber was exposed and secondary dentine was then deposited to seal the pulp chamber and allow chewing function. This sealing process also helps to avoid pulpal infection but, at least in two cases, the bacterial infection occurred. Bacteria spread along the root canals, pulp undergoes necrosis and the tooth becomes non-vital. The infective products remain in the bone, creating a space-occupying lesion that eventually drained the debris through the buccal surface of the alveolar bone as it is thinner. This infective process resulted in the cysts observed on the buccal surface of D2600 mandibular bone at the location of the RM1 and RI2. Infective agents would have also caused the periodontal disease observed at some locations of the alveolar bone. At the same time, teeth would have undergone continuous eruption to compensate for the extreme dental wear. Another consequence of the severe dental wear is the pre-mortem enamel fractures as thinning of the enamel tissue makes it more brittle and susceptible of fracture. In turn all these processes would have enhanced the development of the left condyle TMJ.

The enamel morphology resultant from the wear process can reflect dietary preferences and behavioural patterns of hominin species (e.g., Fiorenza et al., 2011, Hinton, 1981, Teaford and Ungar, 2000, Ulhaas et al., 2007, Ungar, 1994 and Ungar and Grine, 1991). Extreme dental wear can jeopardize the individual's survival as it might lead to poor nutritional intake, disease and eventually to death (Elgart, 2010 and Lovell, 1991). The degree and rate of enamel wear highly depends on several factors such as occlusal variation, environment or behavioural activities – pre- and para-masticatory functions (Addy and Shellis, 2006).

Although Pontzer et al. (2011) concluded that there were no significant differences between the dietary habits of the Dmanisi individuals and the African predecessors, we have not found any specimen displaying a similar wear pattern as that exhibited by D2600. All the individuals of the comparative sample – Skull 5, Gibraltar 1, La Ferrassie 1, La Chapelle – exhibit severe although similar degree of dental wear among the dentition (Table 3, Table 4, Table 5, Table 6, Table 7 and Table 8 and Fig. 4). We have found that the extreme and rounded wear exhibited by D2600 anterior teeth would be similar to that found in chimpanzees and humans in relation to food preparation activities such as gripping and stripping (e.g., Elgart, 2010, Ungar and Grine, 1991 and Wolpoff, 1999). Anterior teeth from chimpanzees exhibit significant dental wear compared with posterior teeth, most likely as a consequence of food preparation activity (Dean et al., 1992, Elgart, 2010, Lovell, 1991 and Morgan and Sanz, 2006). We believe that the unbalanced degree of wear between the anterior and posterior dentition is similar to that found in D2600. When severe wear reduces teeth to stumps, these are unable to process food and longer chewing process is required (Elgart, 2010). Thus, the polished appearance of the anterior dentition is more likely due to food preparation activities and prolonged mulling. The posterior wear pattern in D2600 is characterised by dentine basin and enamel rim, except for both M3s. The unusual wear gradient between the M3 with regard to the M1 and M2 cannot solely be attributed to the difference in eruption times as it has not been recorded in any of the individuals of the comparative sample, which include also individuals of younger age than D2600. We suggest that the wear pattern of the posterior dentition would be similar to that of folivorous species, in particular to gorillas, and resemble that of Paranthropus. This type of food is more abrasive because of the phytolits contained and the dust or soil particles adhered to the food (Elgart, 2010). Due to this type of food, gorillas exhibit extreme posterior dental wear. Gorillas, as ruminants, exhibit a rapid molar wear that creates a central dentine basin and the remaining enamel would act as cutting blades (Fig. 5). This pattern is thought to benefit the ingestion of folivorous diets (Dean et al., 1992, Elgart, 2010, Lovell, 1991, Lucas, 2004 and Morgan and Sanz, 2006). According to Ulhaas et al. (2007) the pattern of molar wear recorded in Paranthropus – rapid enamel reduction and sharp edges – would result from a long chewing process of tough foods, such as tubers, that would include soil particles.

In great apes, hominins and modern humans severe wear has been shown to lead to dental migration and disease (e.g., Cousins, 1988, Crovella and Ardito, 1994, Dean et al., 1992 and Kaifu et al., 2003). Continuous eruption is an adaptive mechanism associated to enamel wear that maintains and enables efficient masticatory function through life. It is evinced by the increased distance between the AC and the CEJ, particularly in absence of pathological signs, and the preservation of the alveolar crest height (e.g., d’Incau et al., 2012, Kaifu et al., 2003 and Levers and Darling, 1983). In their study, Dean et al. (1992) recorded lack of generalised periodontal disease and an AC-CEJ distance between 3 and 8 mm in the mandibular bones of great apes. The authors concluded that continuous eruption would be the most likely cause for the recorded distance. Similar results have been reported for hominins and modern humans (e.g., Hylander, 1977, Martinón-Torres et al., 2011 and Whittaker et al., 1990). Thus, our results would also suggest continuous eruption as the likely cause of the tooth extrusion in those locations – RP3, RC, RI2, LI2, LC – where no signs of infection such as periodontal disease or periodontal cysts were recorded. However, despite the lack of generalised periodontal disease, this individual would have probably suffered recurrent inflammation of the gums (gingivitis) as it normally occurs together with periodontal disease. Approximal attrition is a process resulting from the forces generated during mastication; individuals with stronger or more frequent bites forces show higher rates of approximal wear (Galbany et al., 2010 and Smith, 2011). In turn, mesial drift compensates for this enamel loss (e.g., d’Incau et al., 2012, Kaifu et al., 2003 and Ten Cate, 1998). Fayad et al. (2004) concluded that M1 and M2 are more mesially inclined in those individuals with an erupting or fully erupted M3. This anterior migration has been recorded in baboons (Galbany et al., 2010) and humans [e.g., Homo floresiensis (Jungers and Kaifu, 2011)]. Compared to our Homo heidelbergensis – Sima de los Huesos – and Neanderthal – Tabun 1, La Quina and Ferrassie 1 – samples (Fig. 4), and the two younger specimens of Dmanisi – D2735 and D211 – Margvelashvili et al. (2013) Figs. 2, S3 and S4], D2600 displays a minimum or even null mesial drift in relation to the severe dental wear. Thus, we suggest that this lack of anterior migration of posterior teeth cannot be related to eruption timing; otherwise it could indicate that the wear process was too rapid to enable the mesial drift response.

Fenestrations and dehiscence can be the consequence of excessive occlusal forces, thin alveolar bone or tooth misalignment (Nimigean et al., 2009 and Rupprecht et al., 2001). This condition is fairly common in maxillary bones of gorillas, although less common in chimpanzees (Dean et al., 1992). The bulky appearance of D2600 roots together with the thin alveolar and advanced wear could be probable causes for the fenestrated and dehiscenced roots. In the case of RI1, infectious process is the most likely cause since a major cyst is affecting the RI2 vicinity. However, future studies with CT images of the specimen can help to support or disregard this claim.

Dental wear is the primary cause for the development of periodontal abscess, although tooth fracture can also be a likely cause (e.g., Gracia-Téllez et al., 2013, Lebel and Trinkaus, 2002, Legge, 2012, Lovell, 1991 and Tappen, 1985). Pulp infection produces necrosis and the tooth becomes non vital. If it persists, this will result in bone destruction and a probable ante-mortem tooth loss (ATL) (e.g., Dias and Tayles, 1997, Ogden, 2008a and Ogden, 2008b). Our results for RM1 and RI2 would suggest dental wear in relation to a fibrous and abrasive diet for the abscess formation. The bone resorption at the vicinity of the LP4-LM1 was more likely caused by an infectious episode such as an abscess. However, we cannot ascertain if the bone resorption would have caused ATL, since the posterior table preserves its height and the teeth could be yet in place held by soft tissues. In disagreement with Margvelashvili et al. (2013), we believe that D2600 bone resorption differs from that of the D3900 since edentulous individuals exhibit bone resorption all the way down to the mental foramen on both the buccal and the lingual surfaces (Cawood and Howell, 1988) as seen in La Chapelle (personal observation). However, in D2600 the height of the lingual crest is preserved. In their study, Lordkipanidze et al., 2005 do not suggest an aetiology for the ATL and the extensive bone resorption of D3900 and D3444. According to Dias and Tyles (1997), chronic periodontal disease produces the loss of bone from the alveolar margin, in contrast to the apical resorption in periodontal cysts. Thus, chronic periodontal disease can be suggested as a probable cause of the D3900/D3444 ATL.

Two factors are considered to cause enamel fractures: severe wear and the use of teeth in para- and/or pre-masticatory activities (Constantino et al., 2010 and Scott and Winn, 2011). The thinning of the tissue would be more prone to breakage during the mastication of abrasive and fibrous foods with adhered soil particles and during para- and/or pre-masticatory activities.

Higher tooth rotation than 30° is considered to be abnormal; authors refer to genetic and environmental factors – missing teeth, infection or food impaction – as the causes (Harris and Corruccini, 2008, Natsume et al., 2006 and Primožič et al., 2012). Contrary to the genetic cause suggested for the maxillary rotation of LP4 from D2700 individual (Martinón-Torres et al., 2008), LP4 from Aragó 21 (Smith, 1977) and bilateral P4s of LB1 (Brown et al., 2004), we believe that D2600 bilateral tooth rotation would be the consequence of environmental factors in order to enable masticatory function.

Severe malocclusion, related to severe wear and/or continuous eruption, and/or ATL are factors that damage the stable “centric” occlusal position of the mandible causing irregular loop-shaped chewing movements. This masticatory malfunction with time will develop in osteoarthritis of the mandibular condyles (Langsjoen, 1998 and LeResche, 1997). The left mandibular condyle of D2600 presented the characteristic signs of TMJ. There is a likely correlation between extreme dental wear, continuous eruption, tooth rotation and possibly ATL and the development of TMJ in D2600 individual. Factors also observed in other fossil individuals such as Skull 5 (Gracia-Téllez et al., 2013), La Chapelle (Tappen, 1985) and La Ferrassie (Wallace, 1975).

The hole seen at the right ramus is located at the pterygoid fovea. It bears the part of the lateral pterygoid muscle, a guiding muscle of the mandibular joint (Platzer, 2004). Studies found an association between the lateral pterygoid muscle and TMJ disorder (McNeill et al., 1990 and Murray et al., 2001). We believe that the loss of bone at D2600 right ramus is more likely due to diagenetic processes, since there is no sign of TMJ in the right condyle, and muscle disorders (hypertrophy, atrophy or contracture) do not enhance osteoclastic activity, otherwise blastic activity (Yang, 2002).

Despite all these pathological processes, there is evidence of the chronicity of the lesions, thus implying the survival of the individual for a certain period of time. Other hominins with pervasive dento-gnathic lesions such as ATE9-1 from Atapuerca-Sima de Elefante site (Martinón-Torres et al., 2011), La Chapelle-aux-Saints (Tappen, 1985 and Trinkaus, 1985) and D3900/D3444 (Lordkipanidze et al., 2005) have also shown signs of endurance. The question arising is if collaboration within the group was necessary for the survival of these individuals (e.g., DeGusta, 2003 and Gracia et al., 2009).

Interestingly, despite the severity of the pathological processes observed in D2600, the morphology of the mandible has not been significantly altered because the alveolar crest was minimally remodelled. According to Margvelashvili et al. (2013), in keeping with compensatory eruption of the teeth, both the corpus and the symphysis of D2600 would have increased their height. However, compensatory eruption produces the protrusion of teeth through the axial movement of the dental pieces, hyperproduction of cementum and/or remodellation of the socket, but not by increasing the alveolar crest height (e.g., Kaifu et al., 2003, Levers and Darling, 1983, Ogden, 2008a, Ogden, 2008b, Ten Cate, 1998 and Whittaker et al., 1990). On the contrary, most of the compensatory mechanisms related with occlusal adaptation provoke remodellation of the alveolar rim by reducing (but not increasing) its height, such as compensatory receding bone (Clarke and Hirsch, 1991 and Hillson, 2008). Finally, it is generally accepted that alveolar bone heights show no significant change (or a slight increase if any) with wear or age during adulthood (e.g., Hylander, 1977, Kaifu et al., 2003, Levers and Darling, 1983, Newman and Levers, 1979, Tallgren and Solow, 1991 and Whittaker et al., 1990), and never at such degree as observed within the Dmanisi mandibles. Thus, our study would not be in support of attributing the mandibular shape differences to the dental pathologies and/or their compensatory mechanisms as stated by Margvelashvili et al. (2013). In addition, and despite the greater degree of wear, we have not observed mesial drift in D2600 compared to the younger D2735 and D211. Finally, we do not agree that the ATL and bone resorption recorded in D3900/D3444 is the same as the recorded in D2600 since the lingual wall is preserved. As pointed out by Cawood and Howell (1988), in edentulous individuals we should expect alveolar resorption on both surfaces – buccal and lingual – down to the mental foramen and unlike D2600.

The study of the mandibular bone D2600 belonging to H. georgicus from Dmanisi (Georgia) revealed severe and extensive wear that enhanced other pathological processes. Thus, in relation to the wear process we identify signs of dental compensation (axial movement), tooth rotation, enamel fractures, and degenerative and infectious processes. We hypothesize that the principal and primary cause of the type and degree of dental wear of the D2600, similar to that of chimpanzees and gorillas, would be the intake of abrasive and fibrous foods such as plants and fruits. These dietary habits require pre-/para-masticatory preparation, such as gripping and stripping that is reflected in the wear pattern of the anterior dentition. Moreover, this food would also be the origin of the wear pattern observed in the posterior dentition. It is difficult to ascertain the chronology of the events, since some of the identified signs are cause and consequence of the same pathological process. This study contributes to increase the knowledge about pathological record and the state of hominins health during Pleistocene. Finally, we believe that despite the lesions, the morphology of the mandible was not significantly altered. So pathological changes cannot be invoked as the mechanism explaining the particular morphology of D2600.