Dental calculus reveals Mesolithic foragers in the Balkans consumed domesticated plant foods

Emanuela Cristiani, Anita Radini, Marija Edinborough, and Dušan Borić
PNAS September 13, 2016 113 (37) 10298-10303; published ahead of print August 29, 2016 https://doi.org/10.1073/pnas.1603477113

  1. Edited by Ofer Bar-Yosef, Harvard University, Cambridge, MA, and approved July 5, 2016 (received for review March 2, 2016)

Significance

The starch record entrapped in dental calculus of Mesolithic human teeth from the site of Vlasac in the central Balkans provides direct evidence that complex Late Mesolithic foragers of this region consumed domesticated cereal grains. Our results challenge the established view of the Neolithization in Europe that domestic cereals were introduced to the Balkans around ∼6200 calibrated (cal.) BC as a part of a “package” that also included domesticated animals and artifacts, which accompanied the arrival of Neolithic communities. We infer that Neolithic domesticated plants were transmitted independently from the rest of Neolithic novelties from ∼6600 cal. BC onwards, reaching inland foragers deep in the Balkan hinterland through established social networks that linked forager and farmer groups.

Abstract

Researchers agree that domesticated plants were introduced into southeast Europe from southwest Asia as a part of a Neolithic “package,” which included domesticated animals and artifacts typical of farming communities. It is commonly believed that this package reached inland areas of the Balkans by ∼6200 calibrated (cal.) BC or later. Our analysis of the starch record entrapped in dental calculus of Mesolithic human teeth at the site of Vlasac in the Danube Gorges of the central Balkans provides direct evidence that already by ∼6600 cal. BC, if not earlier, Late Mesolithic foragers of this region consumed domestic cereals, such as Triticum monococcum, Triticum dicoccum, and Hordeum distichon, which were also the main crops found among Early Neolithic communities of southeast Europe. We infer that “exotic” Neolithic domesticated plants were introduced to southern Europe independently almost half a millennium earlier than previously thought, through networks that enabled exchanges between inland Mesolithic foragers and early farming groups found along the Aegean coast of Turkey.

It is commonly assumed that the spread into Europe of domesticated plants, cereals, and pulses (1, 2) was part of a Neolithic “package,” which included domesticated animals and artifacts typical of farming communities (e.g., ceramics, polished stone axes). Researchers agree that such a package was introduced into southeast Europe from southwest Asia and reached inland areas of the Balkans by ∼6200 calibrated (cal.) BC or later. Genetic (3) and isotope data (4) have revealed that demographic movements brought these Neolithic novelties from the Near East into Europe through the process by which scarce Mesolithic foragers were either quickly replaced or assimilated into Neolithic lifeways. Isotope analyses have also suggested that Mesolithic diets were largely based on terrestrial, marine, or riverine protein-rich resources (5⇓–7) with scanty evidence for the consumption of plants (8). Here we show that Late Mesolithic foragers of the central Balkans consumed domesticated specimens of wheat and barley (namely grass species of Triticeae tribe) at least from ∼6600 cal. BC, almost half a millennium earlier than previously thought. We found that starch granules entrapped in dental calculus of Mesolithic (∼6600–5900 cal. BC) human teeth at the site of Vlasac (the Danube Gorges) correspond to those found in the teeth of Neolithic (∼5900–5700 cal. BC) individuals at the neighboring site of Lepenski Vir. Our results show that these starch granules originate from the consumption of main domestic crops found among Early Neolithic communities of southeast Europe (9)—that is, Triticum monococcum (einkorn wheat), Triticum dicoccum (emmer wheat), and/or Hordeum distichon (barley). As genetic and botanical studies have ruled out European ancestry for domestic wheat/barley and confirmed the Near East as the source of these crops (2, 10, 11), our assay provides the earliest evidence that domesticated plants were transmitted independently from other components of the Neolithic package through existing networks that enabled exchanges between inland Mesolithic foragers and early farming groups from ∼6600 cal. BC onwards.

The Sites

The Danube Gorges area is split by the River Danube between the territories of present-day Romania and Serbia (Fig. 1). In this region, more than 20 sites with Mesolithic deposits were discovered in the mid-1960s and yielded unprecedented data for other areas with known Mesolithic presence in Europe (12⇓⇓–15).

Fig. 1.
Fig. 1.

Map of southeast Europe showing the locations of Vlasac, Lepenski Vir, and the other sites mentioned in the text.

The site of Vlasac is situated in the Lady Whirlpool’s Gorge of the Danube on the Serbian side of the river (Fig. 1). The first excavations at the site were carried out in 1970–1971 as part of a rescue project (16). New excavations at Vlasac began in 2006, and the investigations of the site are ongoing (17). The resumed work at Vlasac has covered an area of 326 m2 (Fig. S1) and takes place upslope from the excavation area investigated in 1970–1971. Radiocarbon dates from both old and new excavations suggest that the site was more or less continuously occupied from the Early Mesolithic, from ∼9500 cal. BC, but the intensity of occupation increases from the mid-eighth millennium BC (18, 19). The Late Mesolithic occupation/use of the site covers the period between ∼7400 and ∼6200 cal. BC. New research at Vlasac has indicated that the site was continuously used throughout the period of the Mesolithic–Neolithic transition—that is, ∼6200–5900 cal. BC. Finally, there is also evidence for the use of this site in the course of the regional Early/Middle Neolithic (∼6000/5950–5500 cal. BC). During the Early Neolithic phase, the first chaff-tempered ceramic finds appeared at Vlasac (17, 18). Late Mesolithic domestic features, such as trapezoidal dwellings and numerous and overlapping rectangular stone-lined hearths, testify to the fact that this was a relatively large and likely sedentary complex forager settlement (16, 18). The total number of formal burials at Vlasac excavated in 1970–1971 comprises 87 graves, containing 164 individuals (20). Another 17 burials were excavated in 2006–2009, with the minimum number of 16 individuals. Numerous ground stones were also found at the site (17) (Fig. S2).

Fig. S1.
Fig. S1.

Site plan of Vlasac, reprinted from ref. 70. Late Mesolithic pits and burials (indicated with ‘‘F.’’ and ‘‘H.’’ prefixes, respectively) in the central part of the area excavated in 2006–2009 in the southernmost part of the site with trenches 1–3/2006, 2–3/2007, and 1/2008 (indicated with ‘‘T.’’ prefixes). Context numbers are in brackets.

Fig. S2.
Fig. S2.

A sample of Late Mesolithic ground stones from Vlasac. (Scale bar, 3 cm.)

The site of Lepenski Vir is located on the right bank of the Danube River, ∼3 km upstream from Vlasac. A surface of 2,500 m2 and 3.5 m-thick cultural layers were excavated at this site from 1965 to 1970 (14, 15). Radiocarbon measurements indicate three main phases of human occupation: (a) Early/Middle Mesolithic occupation starting in the second half of the tenth millennium BC and continuing until around 7300 cal. BC (Proto-Lepenski Vir phase); (b) Mesolithic–Neolithic transition phase Lepenski Vir I–II from ∼6150–5950 cal. BC; and (c) Early Neolithic Lepenski Vir III phase from ∼5950–5700 cal. BC (13, 21⇓–23) (Fig. S3). At the site, the transition period is broadly contemporaneous with the arrival of farming groups in the central Balkans between ∼6200 and ∼6000 cal. BC. During this phase, no evidence for farming or stock-herding has been identified to date (22), although different strands of evidence document contact/interaction with early farming groups (4, 13). The site furnished evidence of cultural complexity with numerous dwelling structures with trapezoidal floors and central stone-lined rectangular hearths, sculpted sandstone boulders, ornamented and plane pestles and mortars, and other stone and osseous artifacts and ornaments. A total of 134 graves containing the remains of 190 individuals were also reported (20).

Fig. S3.
Fig. S3.

Site plan of Lepenski Vir with the distribution of Early Neolithic Phase III burials.

Archaeobotanical investigations were not conducted at the time Vlasac and Lepenski Vir were excavated. New fieldwork at Vlasac included for the first time flotation and analysis of macrobotanical remains (17). On the Romanian side of the Danube, Schela Cladovei is the only other site of the same Mesolithic–Neolithic duration that has yielded macrobotanical remains (24). Available wood charcoals, macrobotanical remains, as well as regional palynological data provide limited information about ecological zones around the sites and wild plant resources potentially available to the Late Mesolithic population. The beginning of the Holocene in the Balkans is marked by a shift to warmer and more humid conditions, which led to the expansion of mixed deciduous forests composed of temperate arboreal species such as Quercus, Ulmus, Tilia, Carpinus, and Corylus (25⇓⇓–28). Forest undergrowth taxa such as Cornus, Cotinus, Corylus, Sambucus, and Rosaceae suggest the existence of patches of open woodland around the Danube Gorges sites (25). There are also indications of the presence of temporarily (and permanently) wet areas and flooded land close to the settlements. Riparian vegetation has been an important resource for the Late Mesolithic and for Early Neolithic groups as both wood and fruit of Ulmus and Salix/Populus were collected at sites along the Lower Danube region in association with seed/fruit of Hyosciamus, Physalis, Vitis, and Trapa (25). In addition, analyzed pollen found in coprolites from upper Mesolithic horizons at Vlasac documented the presence of Gramineae grasses around the site as well as the consumption of domesticated Cerealia (29).

Materials

Nine individuals recently excavated at the site of Vlasac were analyzed with microfossils preserved in the dental calculus of seven individuals (Fig. 2 A–E). Based on 14C accelerator mass spectrometry (AMS) dates, some directly dating the analyzed burials, five of these seven individuals are dated to the Late Mesolithic (∼6600–6450 cal. BC), and two individuals are dated to the Mesolithic–Neolithic transition phase (∼6200–5900 cal. BC) (19) (Figs. S4 and S5 and Table 1 and Table S1). Samples were also taken from the dental calculus of three crouched female burials (individuals 8, 20, and 32a) (Fig. 2F and Fig. S6) from the site of Lepenski Vir. Direct 14C AMS measurements on two of these three individuals put them in the Early Neolithic timespan (∼5900–5700 cal. BC) (Table 1 and Table S1). All three individuals are gracile and nonlocal based on their strontium isotope ratios (4) and likely originated in the first farming communities. Results from calculus analysis were integrated with the study of dentoalveolar pathologies (caries and periapical lesions) on permanent teeth.

Fig. 2.
Fig. 2.

Calculus preserved on teeth from Vlasac and Lepenski Vir: (A) Vlasac U64. × 11; (B) Vlasac U24. × 30; (C) Vlasac U44; (D) Vlasac H232; (E) Vlasac H53; (F) sampling Lepenski Vir individual 32a. (Scale bar, 1 cm.)

View this table:
Table 1.

Context, provenance, chronology, sex, and age of individuals analyzed for dental calculus from Vlasac and Lepenski Vir and summary data on the quantity of starch granules, descriptions of their features, and proposed identifications

Fig. S4.
Fig. S4.

Composite drawing with the sequence of overlapping burials excavated in trench 3/2006 at Vlasac and associated AMS dates.

Fig. S5.
Fig. S5.

Human remains from Vlasac. Adult female burial H63 (A) with disarticulated (B and C) cranial and postcranial bones of adult male H81 that was several centuries older.

View this table:
Table S1.

AMS dates from new excavations at Vlasac (2006–2009), calibrated using OxCal v. 4.0

Fig. S6.
Fig. S6.

Early Neolithic burials at Lepenski Vir analyzed for dental calculus in this paper: (A) LV8; (B) LV20; (C) LV32a.

Results

Four different typologies of starch granules were positively identified in dental calculus from Vlasac and Lepenski Vir and assigned to four important tribes of the Poaceae (grass family) and the Fabaceae (legume family), all well-described in literature (30) (Table 1). Some of the starch granules were found to be in an extraordinary state of preservation, especially considering the age of the samples, likely due to the low level of processing as a result of inhalation or accidental ingestion during the processing of starchy foods (see SI Results).

Type I.

The first type of starch granules found possess a strong “bimodal distribution” of simple granules characteristic of most grasses of the Triticeae tribe (barley and wheat tribe). Such distribution is very frequent in our samples and involves the presence of both large granules oval to subround in 2D shape, lenticular in 3D, with central hilum and few lamellae (known as A-type) and small almost spherical granules with a central hilum (≤10 μm, known as B-type) (30) (Table 1). Exceptionally, archaeological grains are still lodged close together, indicating a small amount of processing (SI Results). Bimodal distribution shows a very high number of B-type grains, and granules are also attached to the calculus matrix (Fig. 3 A and B). Bimodal distribution is known for Triticeae tribe (e.g., Triticum spp., Hordeum spp.) (Fig. 4 and Fig. S7), but none of these species had wild progenitors in the Balkans (9, 10). Thanks to the modality of preservation, it was possible to exclude wild species of the Triticeae and Bromideae tribes (known to have bimodal starch granules) that could have been eaten at the time in the region (31, 32) on the basis of their morphology (Figs. S7 A–H and S8A) as well as most recent literature on phylogenetic evaluation of Poaceae species (33). It was also noticed that recovering good quality seeds and starch granules from some of the mentioned wild wheat species was very difficult. Furthermore, Aegilops spp. are absent from assemblages with analyzed macrobotanical remains found at Mesolithic and Early Neolithic sites in the central Balkans (10). Finally, none of the species of the genus Bromus growing in the region are bimodal (SI Results). Starch granules of the Triticeae tribe with morphologies remarkably similar to those found on Late Mesolithic and Mesolithic–Neolithic phase individuals from Vlasac were found in Early Neolithic individuals from Lepenski Vir (Fig. 3C). A further confirmation that archaeological starch granules belong to domesticated species of the Triticeae tribe is provided by the pollen record, as very large grains of Cerealia grasses were recovered in human coprolites from two Late Mesolithic sites in the Danube Gorges (Vlasac and Icoana) and were interpreted as coming from domesticated species (29, 34). At Vlasac, Gramineae pollens of Cerealia type with a diameter no larger than 38.5 μm are dominant in coprolites from lower Mesolithic horizons (1.6%). This diameter is considered the limit between spontaneous and domesticated Cerealia grasses. The quantity of small Cerealia pollens decreases in coprolites from upper Mesolithic horizons, where large Cerealia-type grains with diameters larger than 50 μm (sometimes up to 53 μm) appear (3.5%). Interestingly, in upper Mesolithic horizons, pollen grains of legumes have also been found (29).

Fig. 3.
Fig. 3.

Starch granules entrapped in archaeological dental calculus: (A) starch compound in Late Mesolithic calculus (Vlasac H232); (B) bimodal distribution of lenticular starch grains in calculus dated to the Late Mesolithic (Vlasac U64. × 11); (C) starch compound in Early Neolithic calculus from Lepenski Vir (Lepenski Vir 20); (D) large starch compound found in calculus (Vlasac H53) consistent with Aveneae tribe; (E) cluster of polyhedral starch grains with central hilum and fissures associated with the Paniceae tribe (Vlasac U64); (F) starch grain entrapped in Mesolithic–Neolithic calculus (Vlasac U44). Note the oval shape, kidney-like shape 3D morphology, and sunken central hilum, which are consistent with genus Vicia (Vlasac U44). (Scale bar, 20 μm.)

Fig. 4.
Fig. 4.

Experimental starches: (A) Hordeum distichon; (B) Triticum monococcum; (C) Triticum dicoccum. (Scale bar, 20 μm.)

Fig. S7.
Fig. S7.

Experimental starches. Characteristic starch grains from modern seeds of Triticeae tribe: (A) Aegilops comosa and (B) Aegilops speltoides. The arrows indicate the lamellae: (C) Hordeum murinum; (D) Hordeum distichon; (E) Triticum dicoccum; (F) Triticum monococcum; (G) Elymus caninus; (H) Agropyrum doricum.

Fig. S8.
Fig. S8.

Experimental starch grains. Characteristic starch grains from modern seeds of Bromidae, Aveneae, Paniceae, and Fabeae tribes: (A) Bromus arvensis; (B) Avena sterilis; (C) Phalaris minor; (D) Sesleria caerulea; (E) Sesleria caerulea boiled for 1 min; (F) Setaria viridis; (G) Setaria glauca; (H) Penisetum elatum; (I) Panicum (Echinochloa) crussgalli; (J) Vicia selvatica.

Type II.

Type II of starch granules was found in the form of intact large suboval aggregates of small compound granules (Fig. 3D). Compared with our reference collection and literature (30), both compound granules and the aggregate are consistent in size and shape with those of the species from the Aveneae tribe (oats tribe), both wild (e.g., Avena sterilis, Phalaris minor) (Fig. S8 B and C) and domesticated (e.g., Avena sativa). Starch granules of the Aveneae tribe were found in the Late Mesolithic samples only, and they could have been either deliberately harvested, exchanged, or ingested as weeds of crops.

Type III.

A third group of starch granules also retrieved in dense clusters, in one case of over 200 granules, is consistent with those of the Paniceae tribe. This type of granule has a 3D polyhedral shape, with central hilum and fissures radiating from it (Fig. 3E, Table 1, and Fig. S8 F–J). Although an overlap in size and shape exists among the smaller starch granules of the tribe (35), the large number of our starches and their overall morphologies suggest species of the genus Setaria spp. as potential candidates (Fig. S8 F and G). Although several species of the genus Setaria are diffused in the central Balkans (e.g., Setaria viridis, Setaria verticillata, etc.), a secure identification of archaeological starches needs to be taken with caution, whereas identification to tribe level is rather secure. Starch granules of the Paniceae tribe were retrieved from our samples dated to the Mesolithic, Mesolithic–Neolithic transition, and Early Neolithic phases.

Type IV.

The fourth type of starch granules is consistent with the Fabaeae tribe (peas and beans) and is remarkably consistent with those found in the genus Vicia (vetches) for its small oval 2D shape and its kidney-shape 3D morphology and hilum sunken in a central cleft fissure (Fig. 3F, Table 1, and Fig. S8J). These were lower in number. Species of vetches can grow as weeds of crops, and sometimes are eaten raw.

Single starch granules were found in a number of samples, but these were low in number and often damaged, and it was difficult to reach a secure identification for them. These starches were considered undiagnostic, as they could belong to groups that could not be identified because they were damaged by processing and chewing.

Finally, a pollen grain, characterized by a spheroidal and pantaporate morphology, ca. 50 μm in diameter, pores between 3 and 5 μm, and spines protruding from the surface, was found embedded in the calculus matrix (Fig. 5A) and identified as Malva sp. (36). The presence of such pollen in Late Mesolithic calculus could be suggestive of the consumption of honey or its use as herbal medicine. In addition, nondietary evidence in dental calculus is consistent with a variety of animal (e.g., feather barbules) and plant remains (wood debris and plant fibers) (Figs. S9 C and D and S10), likely to be the result of exposure to dust generated by crafts. Such remains being below 20 μm are interpreted as potential respiratory irritants (37).

Fig. 5.
Fig. 5.

(A) Large spheroidal and pantaporate pollen of Malva sp. with spines protruding from the surface (1, spines and pores; 2, remains of ancient calculus) (Vlasac U64); (B) caries found in Neolithic individual 32a from Lepenski Vir. (Scale bar: A, 20 μm; B, 1 cm.)

Fig. S9.
Fig. S9.

Evidence of dietary and nondietary remains from Early Neolithic individuals from Lepenski Vir: (A) compound of starch grains with faceted morphology and perpendicular extinction cross embedded in dental calculus (32a); (B) compound of starch grains with faceted morphology and evident central fissure (32a); (C and D) fragments of feather downy barbules with triangular-shaped nodes typical of Anseriformes (C) and asymmetric short pronged nodes typical of Anatidae (D) (32a).

Fig. S10.
Fig. S10.

Evidence of nondietary remains from individuals from Vlasac: (A) bast fibers embedded in calculus (U64); (B) lepidoptera scale (U44); (C) feather barbule with long pronged nodes attributed to Anatidae (H53); (D) lepidoptera scale (H53); (E) feather barbule with short pronged nodes attributed to Anatidae (H232). (Scale bar: A and D, 20 μm.)

The analysis of dentoalveolar pathologies has shown the absence of caries and periapical lesions in examined individuals from Vlasac. However, antemortem loss of two teeth (38 and 48) has been recorded in one case (U64. × 11). Root exposure caries were found in Neolithic individual 32a from Lepenski Vir (teeth 16, 25, and 26) (Fig. 5B). Individuals 32a (tooth 17) and 20 (teeth 45 and 46) exhibit antemortem tooth loss. Lastly, three periapical lesions were identified in individual 8 (teeth 12, 21, and 23) from Lepenski Vir.

SI Results

Supplementary information about dietary remains identified in the Lepenski Vir dental calculus as well as nondietary remains identified in dental calculus from Vlasac and Lepenski Vir are provided below.

Dietary Remains.

The analysis of dental calculus obtained from individuals 8, 20, and 32a from Lepenski Vir yielded starch grains associated with diet. In particular, three different typologies of starch granules were identified and assigned to two tribes of the Family Poaceae and to the Family Fabaceae.

The first type of starch granules show a bimodal distribution characterized by the presence of both small round granules with a central hilum, rarely larger than 10 μm, and large oval/lenticular morphology, with central hilum and lamellae (Fig. 3C). On the basis of the experimental comparison (Fig. S7 D–F), these starches were interpreted as belonging to the Triticeae tribes (barley and wheat tribe) of the family Poaceae and were attributed to the domesticated species according to the consideration made in Considerations on the Conditions of Preservation of Starch Granules in Our Samples and Potential Pathways of Inclusion.

The second group of starch granules was retrieved in dense clusters and is consistent with the grains of the tribe Paniceae. This type of starch granule shows a faceted shape, a central hilum, and fissures often radiate from it (Fig. S9 A and B). The large number of our starch granules (sometime more than 200 per cluster) and the morphology of the single grains suggest an attribution to the genus Setaria spp. grains (see Fig. S8 F–I for experimental comparison). The presence of such starch granules is also remarkable. Species of the tribe Paniceae are in fact part of an important group of food plants known as small millets with lower archaeological visibility compared with others. Further details on this topic were published by Madella et al. (62).

Finally, a small number of simple starch granules were found in samples that had lost characteristics useful for the identification, often had no extinction cross visible, and appeared somewhat gelatinized. Due to the low number of these starch-type granules and the evident loss of their original morphology, it is not possible to put forth a positive identification. Such starch granules were described as unidentified.

Nondietary Remains.

All of the nondietary remains of the debris found in analyzed dental calculus were consistent with respirable fractions of particulate matter, which being below 20 μm would also be naturally airborne, and all of them are therefore potentially respiratory irritants (37). Our finds therefore document the potential of human dental calculus to entrap environmental debris potentially harmful to people, as suggested by an increasing number of studies (46).

Vlasac.

Nondietary evidence in dental calculus includes a variety of animal and plant remains. Small fragments of conifer wood (with some of its characteristic tracheid fibers with bordered pits clearly visible) and a large amount of plant fibers (22 in total) of varied diagnostic and nondiagnostic nature were found in a number of individuals and were particularly abundant in Mesolithic individual U64. × 11 from Vlasac. Potential bast fibers of plant origin were often observed still in situ (Fig. S10A), in different forms of preservation; some of them (seven in total) showed a narrow lumen and dislocation bands when polarizer and analyzer were oriented at 90 degrees to one another, typical of plants used for fibers in cordage and textiles such as nettle, flax, and hemp (63). Considering the region and period, the most likely candidate for our finds is nettles (Urtica spp.). Bast fibers potentially resulting from the processing of plants for cordage or textile were found in dental calculus in other studies (53, 64, 65). Among other remains were two unidentified fungal spores (for problems with the identification of spores, see ref. 46), microcharcoal and nondiagnostic plant remains showing no diagnostic feature to allow for a positive identification.

Animal remains were consistent with fragments of feather barbules, which have been identified upon the shape and features at their nodes (Fig. S10 C and E) and Lepidoptera scales (butterfly and moth), the latter identified for their distinctive morphology of blade and pedicel (Fig. S10 B and D). Lepidoptera scales have been recently found also in a dental calculus sample of Homo erectus from the Qesem cave in Israel (46), showing that such evidence can survive considerable periods of time. Pathways of inclusion of such remains vary (46). Fragments of bird feathers identified in the dental calculus of some Late Mesolithic individuals from Vlasac show shapes and features at their nodes that would be consistent with various species of ducks of the Anatidae family. These identifications are supported by the presence of identified bird remains, such as mallard Anas platyrhynchos and teal Anas crecca, in the Late Mesolithic levels at Vlasac (17). Bird feather barbules could be the result of exposure to dust generated when plucking birds, whereas Lepidopetera scales are present in the airborne dust in the natural environment. Such remains therefore could represent inhaled material or material accidentally ingested as settled dust on food.

Lepenski Vir.

Nondietary evidence in the Lepenski Vir dental calculus was consistent with animal debris and particularly with feather barbules. Two fragments from Neolithic calculus show triangular-shaped nodes, which are typically located on the downy barbules of Anseriform birds inhabiting aquatic environments (Fig. S9C), and pronged nodes suggestive of the Anatidae family (Fig. S9D) (66⇓⇓–69).

Discussion and Conclusions

Our evidence suggests that starch granules found in large numbers in dental calculus of both Late Mesolithic and Mesolithic–Neolithic individuals from Vlasac are consistent with tribes of plants that became key staple domestic foods with the start of the Neolithic in this region. The status of preservation of entire “groups” of starch granules suggests that their presence in the calculus matrix could be the result of inhalation or ingestion during processing rather than ingestion after cooking, a pathway of inclusion of starch granules rarely taken into consideration in other studies (see Supporting Information). The data indicate that the earliest domesticated seeds would have reached the Danube Gorges area already by the mid-seventh millennium BC, if not earlier—that is, at least four centuries before what is commonly believed. In Greece, at the site of Franchthi Cave (Fig. 1), three carbonized seeds of T. dicoccum from mixed Final Mesolithic/Initial Neolithic aceramic units were directly AMS-dated to between 6750 and 6430 cal. BC (95% confidence), suggesting the presence of domesticated emmer wheat in southern Greece likely before ∼6500 cal. BC (38), which is contemporaneous with the analyzed Mesolithic individuals in the Danube Gorges. Although the dated T. dicoccum seeds from Franchthi were tentatively attributed to the Neolithic occupation (38), in the light of our results one may speculate that they relate to forager rather than first farming occupation of the site. These dates are also broadly contemporaneous with the Early Neolithic occupation at Ulucak (level VI) on the Turkish side of the Aegean Sea (Fig. 1) (39). A recent series of dates from the Early Neolithic sites of Dikili Tash, Mavropigi, and Paliambela in Macedonia (Fig. 1) suggest that Neolithic communities settled in the region already by ∼6500 cal. BC (40). In the central Balkans, the earliest date for an Early Neolithic occupation comes from the site of Blagotin (Serbia) (Fig. 1), where OxA-8608 (7480 ± 55) dated a red deer antler to 6440–6230 cal. BC (95% confidence) (41).

It remains difficult to evaluate the importance of carbohydrate foods in the overall subsistence of Late Mesolithic foragers at Vlasac (25), as it is likely that fish was an important staple resource (5). However, antemortem tooth loss documented in some individuals could indicate a deteriorated dental health status due to specific dietary habits (e.g., glucose/fructose consumption derived from plants and honey). The absence of caries during the Mesolithic at Vlasac does not contradict the evidence about the consumption of plants rich in fermentable carbohydrates and could directly be related to high calculus rates documented among Mesolithic individuals at Vlasac, as calculus forms during the process of mineralization whereas caries develops through the process of tooth demineralization (42). The similarities between starch grains present in the calculus of Mesolithic individuals from Vlasac and Early Neolithic individuals from Lepenski Vir provide evidence that Mesolithic foragers of the Danube Gorges must have had acquired domestic grains before farming practices became rooted in the Balkans more permanently ∼6200/6100 cal. BC. This could be similar to the process that led to the spread of domesticated maize through trade networks among forager communities of North America before evidence of its cultivation (43, 44). It has also been argued that domesticated wheat reached Mesolithic Britain already by 6000 cal. BC—that is, 2,000 y before full-scale agricultural practices were adopted (45). In the central Balkans, foragers’ familiarity with Cerealia grasses from ∼6500 cal. BC, if not earlier—that is, at least 400 y before their full cultivation in this regional context—might have eased the later quick adoption of agricultural practices. The existence in this region of large sedentary settlements of complex foragers along with numerous grinding stone tools (Fig. S2) is in agreement with our results that processing of plant foods, including domesticate species, might have been common already in the Late Mesolithic. This significantly challenges our hitherto understanding of pre-Neolithic adaptations in southeast Europe and suggests that elements of the “Neolithic package” were transmitted piecemeal and at different times using established networks of interaction between foragers and farmers.

Methods

Calculus sampling took place on aluminum foil with starch-free gloves worn at all times. The average weight for calculus samples was 0.983 g. Once removed, samples were sealed into sterile Eppendorf tubes. In the laboratory, samples underwent specifically developed decontamination procedures (46) (SI Methods). Extraction and mounting of the microfossils entombed in the calculus matrix involved a weak solution 0.06 N of HCl (SI Methods). A solution of 50:50 glycerol and ultrapure water was used to allow for the rotation of microfossils. Examination of microfossils was carried out using Zeiss and Olympus compound polarized microscopes (100–630×) at the University of York and a Leica DM2500 polarized microscope (100–1,000×) at the University of Cambridge. Starch granules were identified on the basis of their 3D morphology, presence, and shapes of features (lamellae, hilum, bumps, and depressions of their surface), characteristics of the extinction cross under polarized light microscopy. A large collection of microremains from modern plants native to the central Balkans and the Mediterranean region collected at the Botanical Garden of Belgrade and stored at the University of Cambridge was used as an experimental reference. A collection of plants of north European, Mediterranean, and north African origin as well as nondietary items (e.g., plants used for crafts and occupational dust such as wood and pottery) hosted by the University of York was also consulted and successfully used before in other studies (35, 46). Fresh and dry botanical samples were ground in agate mortar using distilled water. Different modalities of grinding were applied to record level of damage and/or change in the size of granules. Ground plant material has also been left in water to understand the swelling process. Our reference collection also comprised experimentally processed seeds (e.g., boiled, ground, chewed) (SI Materials). Criteria of identification that are current standards in the field of modern and ancient starch granules research (47) were followed.

Caries and periapical lesions have been recorded macroscopically and using a magnifying lens. Visibility location and size of caries were described following the methodology proposed by Hillson (42). Both caries and periapical lesions were recorded at the individual tooth level (42, 48) and the Fédération Dentaire Internationale system was used for tooth numbering (49).

SI Methods

Calculus Decontamination Procedure.

The fleck of calculus was first observed under incident light microscopy, and the contaminant soil adhering to the surface was removed following established protocols (35, 51). The protocol consisted of the following steps: A fine acupuncture needle was wet in HCl 0.06 M. A very small area of dirt adhering to the surface was gently dislocated and scraped off. The calculus was then washed in ultrapure water and left to dry. It was then examined again, and the procedure was repeated until the fleck was clean. Once the surface was satisfactory clean, the calculus samples were removed using a sterile blade and then with a fine sterile acupuncture needle. Further, they were slid into sterile Eppendorf tubes. Ultrapure water was added, and the calculus was washed up to three times to remove any trace settled on it. Once washed, the calculus samples were placed into new sterile Eppendorf tubes and degraded in a weak solution 0.06 N of HCl, enough to cover the entire calculus. The contaminant soil removed was also placed on a slide and examined, as an extra safety measure, and no starch was recovered from the contaminant soil. A portion of the dissolving calculus was siphoned out with its natant and mounted on clean slides. A solution of 50:50 glycerol and ultrapure water was added to the drop on the slide to allow for the rotation of microfossils, a step necessary during the identification. It has been documented how such protocols allow the retrieval of all debris undamaged and often calculus fragments still dissolving can be observed under the microscope, providing further evidence of the integrity of the analysis (51).

Experimental Collection.

Reference collection of modern starch granules included plants native to the central Balkan region as well as many others from the wider Mediterranean region stored at the George Pitt-Rivers laboratory at the University of Cambridge. Furthermore, a reference collection hosted at the Department of Archaeology at the University of York that contains plants of North European, Mediterranean, and Northern African origin, including nondietary items, as well as processed and chewed food specimens (e.g., plants used for cordage and containers and occupational dust such as wood and pottery) was also consulted. Such reference collection has been successfully used before in other studies (35, 51). The reference collection was prepared in a clean laboratory, with handlers wearing dust-free gloves. Starch granules were extracted by gently grinding/squashing seeds in a clean jade mortar/pestle, first with and then without a drop of ultrapure water and as gentle as possible to try to obtain starch granules still in situ in the amyloplast. Starch granules were mounted using the same mounting media used for the calculus samples to secure a comparison as close as possible to the remains observed in ancient samples.

Details on the Identification of the Tribe Triticeae and Aveneae.

The identification of starch granules with bimodal distribution consisting of A- and B-type starch granules as domesticated species of the Triticeae tribe was made based on morphological and archaeological considerations mentioned in the main text of this article as well as distributions of wild species in the region (32). Following Svoboda (52), we considered the presence of wild taxa of the Triticeae tribe, and the following species are well-known: Aegilops cylindrica (A-type: min., 10.3; max., 28.5; mean, 22.5), Aegilops geniculata (A-type: min., 10.04; max., 31.12; mean, 22.34), Aegilops neglecta (A-type: min., 10.2; max., 19.28; mean, 12), and Aegilops triuncialis subsp. Triuncialis (A-type: min., 10.23; max., 34.2; mean, 21.5) (53). In our reference collection, we also have Aegilops speltoides (A-type: min., 10.2; max., 34.6; mean, 19.4) and Aegilops comosa (A-type: min., 10.3; max., 38.5; mean, 23) (see Fig. S7). Aegilops peregrina and Aegilops crassa are not bimodal (54). The possibility that archaeological starch granules were consistent with one of such Aegilops species was excluded through experimental comparison as well as on the basis of recent literature. First of all, published data show that very often starch granules of the Aegilops are large and round, reaching often 40 μm in size, whereas species of domesticated Triticeae are often below this size (55). However, in our reference collection of plant specimens from the area, this is not often the case (as shown earlier in this paragraph).

Second, although the majority of the archaeological starch granules assigned to the Triticeae tribe and proposed as domesticates are still lodged together and therefore difficult to fully measure, many are clearly within the range of domesticated species published by others (56). This being said, some of the starch granules can reach up to 40 μm, and in our reference collection, granules of domesticates can easily reach that size too. Such differences with published data can be explained by different growing conditions, as a number of environmental factors can affect the size of starch granules. Using specimens from the Balkan region was therefore very helpful. Due to the fact that archaeological starch granules are still lodged together, in our comparison we also focused on their overall morphology, the presence/absence/shape/location of lamellae, as well as the way the starch granules appear as a “cluster.”

Third, in the archaeological samples, a large number of starch granules were very clearly still lodged together as they would be in the amyloplast. This aspect allowed us to establish their strong and remarkable bimodal distribution where the large lenticular granules, A-type, are seen emerging from the mass of a high number of B-type starch granules (Fig. 3 A and B). In the light of such observations, we compared starch granules in their amyloplast in wild species of Triticeae, using both reference material and published data as detailed below. In wild species of wheat (Aegilops) as well as in genera of the Triticeae tribe present in the Balkans, such as Elymus and Agropyrum, in our reference collection, the large A-type granules are predominant in number, an aspect that becomes even clearer when these are viewed in situ (57). A-type starch granules of the Aegilops species have clearly a round 2D shape, sometimes irregular and with lumps, and a large number of lamellae across the entire granules, denser in the central area of the granules near the hilum, have been observed (see examples of A. cylindrica and A. speltoides in our reference collection; Fig. S7 A and B), and this has also been pointed out in earlier published work (30). Our archaeological samples are clearly not as round but more oval in 2D shape and they lack the number of lamellae across the entire granule as seen in the Aegilops species. Also, the distribution of lamellae in archaeological granules is different from various species of Aegilops we have analyzed and we exclude this aspect as being related to cooking on the basis of experimental comparison.

Our observations are supported by a number of studies on modern starch granules. A recent phylogenetic reevaluation of morphological variations of starch grains among Poaceae species has stressed that clear bimodal simple grain types characterize species belonging to the Triticum and Hordeum genera but is lacking in a number of wild wheat species of Aegilops, including some of the species known to be present in the Balkans (32, 33). It has also been known for some time now that species of Aegilops with bimodal distribution have low B-type starch (57), confirming observations from our reference collection. Traits on B-type starch are thought to be under genetic control (31). Thanks to the extraordinary modality of preservation of our archaeological starch granules and to the morphological and genetic aspects described above, we could exclude species of Aegilops. Within the Triticeae tribe, we also considered Hordeum murinum, as this wild species is documented in the region. However, we excluded this species on morphological grounds, as large A-type granules were consistently smaller than ours (Fig. S7C).

Fourth, we examined species of Elymus spp. (e.g., Elymus caninus, A-type: min., 10.3; max., 29.03; mean, 18.34) (Fig. S7G) and Agropyrum spp. (e.g., Agropyrum doricum, A-type: min., 10.5; max., 19.35; median, 12.4; e.g., Agropyrum farctus, A-type: min., 10.25; max., 31.5; median, 16.2) (Fig. S7H), which also look more elongated growing in the Balkans, and excluded them on morphological grounds and appearance of starch granules in the amyloplasts. In our reference collection, the starch granules have clearly different overall morphology, and the A-type granules are more oval and smaller than ours (see Fig. S7 for an example of such starch granules in their amyloplast). We believe our results to be reliable, as we evaluated the morphology of starch granules of interest in this study by using plant material from the region under study as well as other regions. We would also like to stress that the extraction of starch granules from species of Elymus and Agropyrum required several attempts to be successful despite the fact that the plants were mature, suggesting a lower production of starch granules of these species compared with other species of the Triticeae tribe.

Lastly, although the bimodal distribution is known for at least one species in the Bromideae tribe, Bromus catharticus (57), the latter is not native to the Eurasian region. However, we also excluded other species of the genus Bromus (e.g., Bromus secalinus, Bromus alopecuros, Bromus hordeaceus, Bromus arvensis) on morphological grounds derived from both our reference collection and previously published data (Fig. S8A) (57). Such species have starch granules that do not appear bimodal, and starch granules are around or below 11 μm. As we excluded wild species of the Triticeae and Bromideae tribes to be the origin of our starch granules, this leaves us with the domesticated species of the Triticeae tribe available at the time. Finally, the characteristics used in our comparison/identification related to 3D morphology and lamellae have already been used as criteria for separation among Aegilops spp. Triticum spp. and Hordeum spp. when a large number of well-preserved starch granules as well as archaeobotanical data were available (30).

The identification of the tribe Aveneae was achieved by comparison with published data as well as through the evaluation of experimental starch granules in our reference collection, as mentioned in the main text of the article (Fig. S8 B and C). Compound granules are known in different species of plants including the Poaceae (33). In our reference collection, we noticed morphological similarities of Aveneae with compound granules of Sesleria caerulea, a species growing wild in many areas of Europe (Fig. S8D). Additionally, some species of such genus are still harvested today as wild food in the central Balkans (e.g., Sesleria serbica) (31). However, in S. caerulea, the single starch granules forming the compound lump and the lump itself were smaller than the archaeological ones. A number of processes can cause starch granules to swell, including water, cooking, and even grinding. To exclude that the larger starch granules observed in our ancient samples were those of Sesleria species enlarged by the processes listed above, we carried out experimental work to see if starch granules of these species could overlap with those of the Aveneae species. We were able to exclude this possibility as our experiments show how starches of the genus Sesleria lose their morphology when ground or boiled for even just a minute (Fig. S8E). On the other hand, variations caused by cooking/processing in species of the Triticeae tribe were found to be consistent with those of other published studies (54).

Considerations on the Conditions of Preservation of Starch Granules in Our Samples and Potential Pathways of Inclusion.

We have found that our starch granules (found in large numbers and still lodged together in the amyloplast) exhibit a good overall state of preservation. Alterations of starch granules due to diagenetic processes, processing, and cooking have been the subject of a number of studies and remain an important area of research in the field of ancient starch research (44). Although it is assumed that starch granules in the mouth are by and large the result of deliberate consumption of processed and cooked plants, and therefore morphological changes are to be expected, an increasing number of papers show that a variety of pathways to the inclusion of such remains are equally possible, including accidental inhalation (58). The presence of well-preserved starch granules in our ancient samples, including those still apparently lodged in their amyloplast, could be the results of ingestion during the processing of the seeds before cooking. Plant processing for fibers as well as for starchy food generates a high amount of debris, especially when carried out for prolonged periods of time. This is a well-studied aspect of “occupational” health for the consequences that such activities have on the respiratory system when inhaled (59). Although technology and scale of processing changed over time, the processing of plant material is inevitably a “dusty craft” that results in large quantities of particles to be released in the process. Such particles can settle on hands, cloths, and hair as well as become airborne and enter the human mouth by transfer from hand to mouth or by inhalation. These processes increase the chances of uncooked and low-processed remains to become incorporated into the calculus matrix. Such a pathway of inclusion is very likely and can easily explain the good state of preservation in our remains and should always be taken into account. We also stress that many species of starchy food are consumed raw, even bitter tubers such as tigernut (Cyperus esculentus) (60). Legumes are consumed raw even today in many countries, including some European countries, such as Italy, and could result in well-preserved starch granules. Although we cannot be sure that no processing occurred in our remains, prolonged cooking and high processing result in the release of starch granules out of their amyloplat, something we can exclude in this unique case. Some of our starch granules have cracks that are very likely the result of grinding, as documented also by stone tools found at both of the examined sites (15⇓–17, 61) (Fig. S2), and this could clearly generate the release in the air of starch masses that would therefore not be grinded for prolonged periods of time or cooked, hence their still being lodged in the amyloplast in great conditions of preservation. This could also explain some of the largest starch granules, as grinding can enlarge the starch granules (56).

SI Materials

Vlasac.

Radiocarbon dates from the most recent work at Vlasac are given in Table S1. Bayesian statistical modeling of the dates is available for the vertical sequence of burials in trench 3/2006 (Fig. S4), and these results are cited in italics to separate them from individually calibrated dates (19). Radiocarbon ages of human burials have been corrected for the aquatic reservoir effect due to the consumption of fish based on the δ15N and δ13C values obtained as part of the AMS radiocarbon dating process. Age offsets are calculated using method 1, as outlined by Cook et al. in 2002 (50). The dates from the most recent excavation phase at Vlasac suggest that the excavated zone of the site was occupied from the end of the eighth millennium BC until the beginning of the sixth millennium BC. This time span is representative of the duration of Late Mesolithic mortuary activities in trench 3/2006 with a concentration of burials with available dental material, which was analyzed for starch preservation. More detailed chronological, stratigraphic, and contextual information is provided below for each of the samples on which dental calculus yielded results:

  • i) Unit 115 (Fig. S4) represents a cremation zone found in the vertical sequence of Mesolithic burials in trench 3/2006. This zone is located below the level of primary disturbed burial H81, which is directly dated by OxA-20762 to 6600–6400 cal. BC (95% probability), and it cuts primary disturbed burial H136 dated by OxA-18865 to 6630–6480 cal. BC (95% probability). The analyzed disarticulated dental remains are likely from the individual H136.

  • ii) Primary female adult burial H232 (Fig. S4) is indirectly dated by OxA-20702 to 6620–6480 cal. BC (95% probability), which dates a carbonized Cornelian cherry Cornus mas pit found in a cremation pit directly beneath burial H232 (Fig. S4) and acts as the terminus post quem for the burial. However, the archaeological context suggests that the cremation event and subsequent inhumation burial must have taken place as part of the same burial event and are thus contemporaneous (17, 19).

  • iii) Primary adult male burial H326 is found in trench 3/2008 as an undisturbed single burial in the vicinity of several other burials (Fig. S1) and has not been directly AMS-dated. On the basis of the dating of a bone tool in the nearby context 320 by OxA-24769– of 6640–6490 cal. BC (95% confidence) (Table S1) as well as on the basis of the burial’s extended supine inhumation body posture and orientation being parallel with the Danube with the head pointing in the downstream direction, this burial is dated to the Late Mesolithic phase of the site’s use, around the mid-seventh millennium BC.

  • iv) The disarticulated mandible marked as U64. × 11 (Fig. S4) was found in the fill of primary disturbed adult female burial H63, AMS-dated by OxA-16542 to 6240–6010 cal. BC (95% probability). This mandible (Fig. S5 A and C) comes from the individual found in primary disturbed, chronologically older adult male burial H81 dated by OxA-20762 to 6600–6400 cal. BC (95% probability). There were several other skeletal elements from this disturbed burial found in the fill of primary disturbed burial H63 (Fig. S5B).

  • v) Dental remains in unit 222. × 18 are found in the layer of paleosol at the bottom of the stratigraphic sequence in trench 3/2006. There is no radiocarbon date for this layer, but based on the stratigraphic sequence in trench 3/2006, the remains in this layer are contemporaneous with or earlier in age than the dated remains in layer unit 282 in the same trench, which gave two dates on two different bone tools: OxA-24809 and OxA-24810 gave ranges of 6990–6650 and 7000–6600 cal. BC (95% probability), respectively.

  • vi) Primary female adult burial H53 (Fig. S4) is indirectly dated with OxA-16544 to 6010–5840 cal. BC (95% probability), which dates a red deer skull deposited directly over the stone slab that covered the pelvic area of the deceased woman found in burial H53. Hence the date represents a terminus ante quem for the individual in burial H53. The terminus post quem for burial H53 is OxA-16542 that directly dates burial H63, found in a stratigraphic superposition directly underneath burial H53, to 6240–6020 cal. BC (95% probability). However, it is very likely that the date obtained for the red deer skull and the chronological age of burial H53 are by and large contemporaneous.

  • vii) Unit 44 relates to the layer that was found on top of stone plaques and slabs directly covering burial H53 (Fig. S4). This layer contained numerous finds of disarticulated human remains, including the dental remains analyzed here. These remains come from an adult individual and most likely can be associated with the primary disturbed adult female individual in burial H63, the skull of which was missing and which was dated by OxA-16542 to 6240–6020 cal. BC (95% probability), but it cannot be excluded at this stage that these remains belong to an older adult disturbed burial found in this particular burial location.

Lepenski Vir.

More detailed chronological, stratigraphic, and contextual information is provided below for each of the samples from Lepenski Vir on which dental calculus yielded results:

  • i) Primary adult female burial 8 lay in the fill of trapezoidal building 24, some 20 cm above the limestone floor of the building and must have been interred in this space after the abandonment and complete or partial backfilling of the semisubterranean space of this building structure. It was found on the lateral right side in a crouched position and was symmetrically facing another crouched burial in the same building space (burial 9) (Fig. S6A). These two interments are very likely broadly contemporaneous or took place within the existing memory of the first burial event. OxA-25207 and AA-58319 directly date these burials (13, 21) (see Table 1).

  • ii) Primary adult female burial 20 is found in quad. b/10 as a single inhumation placed on the lateral left side in a crouched position within a stone construction (Fig. S6B). There is no direct AMS date for this individual, and it is dated to the Lepenski Vir phase III based on its stratigraphic position and its body posture.

  • iii) Primary adult female burial 32a (Fig. S6C) is found on the lateral right side in a crouched posture at the bottom of the multiple burial place in quad. d/1. There was another crouched burial marked as 32b placed directly on top of the individual marked as burial 32a. The burial is directly dated by OxA-5828 (see Table 1).

Acknowledgments

The authors thank Snežana Vukojičić for providing access to a number of modern plant species and samples of wild wheat grasses from the Botanical Garden in Belgrade. They also thank Nicky Milner, T. D. Price, and Alfredo Coppa for providing useful comments on earlier versions of the paper. Sampling of calculus was carried out in 2012 on the osteological collection of human remains stored at the University of Belgrade. This sampling was facilitated with the help of Sofija Stefanović. Funding for excavations at Vlasac in 2006–2009 was provided through the British Academy (Grants SG-42170 and LRG-45589) and a grant from the McDonald Institute for Archaeological Research, University of Cambridge, UK (to D.B.). Analysis of dental calculus was carried out in the framework of the ERC Starting Grant Project HIDDEN FOODS (639286) awarded to E.C.

Footnotes

  • ↵1To whom correspondence should be addressed. Email: ec484{at}cam.ac.uk.

  • Author contributions: E.C. designed research; E.C. and A.R. performed research; M.E. performed anthropological analysis of dental remains from the sites presented in the paper; D.B. excavated the site of Vlasac from 2006–2009 and provided chronological framework; E.C. and A.R. analyzed data; and E.C., A.R., M.E., and D.B. wrote the paper.

  • The authors declare no conflict of interest.

  • This article is a PNAS Direct Submission.

  • This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1603477113/-/DCSupplemental.

References

  1. ↵
    1. Asouti E,
    2. Fuller D
    (2013) A contextual approach to the emergence of agriculture in southwest Asia: Reconstructing Early Neolithic plant-food production. Curr Anthropol 54:299–345.
    .
    OpenUrlCrossRef
  2. ↵
    1. Zeder MA
    (2008) Domestication and early agriculture in the Mediterranean Basin: Origins, diffusion, and impact. Proc Natl Acad Sci USA 105(33):11597–11604.
    .
    OpenUrlAbstract/FREE Full Text
  3. ↵
    1. Mathieson I, et al.
    (2015) Eight thousand years of natural selection in Europe. Nature doi:10.1101/016477.
    .
    OpenUrlCrossRef
  4. ↵
    1. Borić D,
    2. Price TD
    (2013) Strontium isotopes document greater human mobility at the start of the Balkan Neolithic. Proc Natl Acad Sci USA 110(9):3298–3303.
    .
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Bonsall C, et al.
    (1997) Mesolithic and Early Neolithic in the Iron Gates: A palaeodietary perspective. J Eur Archaeol 5:50–92.
    .
    OpenUrlCrossRef
  6. ↵
    1. Mannino MA,
    2. Thomas KD
    (2009) Mesolithic Horizons. Paper presented at the Seventh International Conference on the Mesolithic in Europe, eds McCartan S, Woodman P, Schulting R, Warren G (Oxbow Books, Oxford), pp 140–145.
    .
  7. ↵
    1. Shulting R
    (2015) in The 150th Anniversary of the Discovery of Mesolithic Shellmiddens—Vol. 2, eds Bicho N, Detry C, Price TD, Cunha E (Cambridge Scholars Publishing, Cambridge, UK), pp 153–172.
    .
  8. ↵
    1. Hardy K
    (2007) Food for thought: Starch in Mesolithic diet. Mesol Miscell 18(2):2–11.
    .
    OpenUrl
  9. ↵
    1. Filipović D
    (2013) Southwest Asian founder- and other crops at Neolithic sites in Serbia. Eur J Archaeol 4:195–215.
    .
    OpenUrl
  10. ↵
    1. Weiss E
    (2012) in Domestication of Plants in the Old World, eds Zohary D, Hopf M, Weiss E (Oxford Univ Press, Oxford), 4th Ed, pp 1–8.
    .
  11. ↵
    1. Harris DR
    (1996) The Origin and Spread of Agriculture and Pastoralism in Eurasia (UCL Press, Oxford).
    .
  12. ↵
    1. Bonsall C
    (2008) in Mesolithic Europe, eds Bailey G, Spikins P (Cambridge Univ Press, Cambridge, UK), pp 238–279.
    .
  13. ↵
    1. Borić D
    (2011) in Beginnings – New Research in the Appearance of the Neolithic Between Northwest Anatolia and the Carpathian Basin, ed Krauß R (Verlag Marie Leidorf GmbH, Rahden, Germany), pp 157–203.
    .
  14. ↵
    1. Radovanović I
    (1996) The Iron Gates Mesolithic (International Monographs in Prehistory, Ann Arbor, MI).
    .
  15. ↵
    1. Srejović D
    (1972) Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir (Thames and Hudson, London).
    .
  16. ↵
    1. Srejović D,
    2. Letica Z
    (1978) Vlasac: Mezolitsko naselje u Djerdapu (I Arheologija) (Srpska akademija nauka i umetnosti, Belgrade).
    .
  17. ↵
    1. Borić D, et al.
    (2014) Late Mesolithic lifeways and deathways at Vlasac (Serbia). J Field Archaeol 39:4–31.
    .
    OpenUrlCrossRef
  18. ↵
    1. Borić D,
    2. French C,
    3. Dimitrijević V
    (2008) Vlasac revisited: Formation processes, stratigraphy and dating. Doc Praehist 35:261–287.
    .
    OpenUrlCrossRef
  19. ↵
    1. Borić D,
    2. Griffiths S
    (2015) The living and the dead, memory and transition: Bayesian modelling of Mesolithic and Neolithic deposits from Vlasac, the Danube Gorges. Oxf J Archaeol 34(4):343–364.
    .
    OpenUrlCrossRef
  20. ↵
    1. Roksandić M
    (2000) Between foragers and farmers in the Iron Gates Gorge: Physical anthropology perspective. Djerdap population in transition from Mesolithic to Neolithic. Doc Praehist 27:1–100.
    .
    OpenUrl
  21. ↵
    1. Bonsall C, et al.
    (2015) New AMS 14C dates for human remains from Stone Age sites in the Iron Gates reach of the Danube, southeast Europe. Radiocarbon 57(1):33–46.
    .
    OpenUrlCrossRef
  22. ↵
    1. Borić D,
    2. Dimitrijević V
    (2007) When did the ‘Neolithic package’ reach Lepenski Vir? Radiocarbon and faunal evidence. Doc Praehist 35:53–72.
    .
    OpenUrl
  23. ↵
    1. Borić D,
    2. Dimitrijević V
    (2009) Apsolutna hronologija i stratigrafija Lepenskog Vira. Starinar 57/2007:9–55.
    .
    OpenUrl
  24. ↵
    1. Mason S,
    2. Boroneanţ V,
    3. Bonsall C
    (1996) Plant remains from Schela Cladovei, Romania. Mesolithic Miscellany 17:11–14.
    .
    OpenUrl
  25. ↵
    1. Marinova E, et al.
    (2013) Wild plant resources and land use in Mesolithic and Early Neolithic south-east Europe: Archaeobotanical evidence from the Danube catchment of Bulgaria and Serbia. Offa 69/70:467–478.
    .
    OpenUrl
  26. ↵
    1. Willis KJ
    (1994) The vegetation history of the Balkans. Quat Sci Rev 13:769–788.
    .
    OpenUrlCrossRef
  27. ↵
    1. Magyari G, et al.
    (2010) Braun, rapid vegetation response to Late Glacial and Early Holocene climatic fluctuation in the South Carpathian Mountains (Romania). Quat Sci Rev 35:116–130.
    .
    OpenUrl
  28. ↵
    1. Gigov A
    (1972) Pollen analysis. Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir, ed Srejović D (Thames and Hudson, London), pp 185–186.
    .
  29. ↵
    1. Cârciumaru M
    (1978) in Vlasac. A Mesolithic settlement in the Iron Gates. 2 Geology–Biology–Anthropology, eds Srejović D, Letica Z (Srpska Akad. Nauka Umetnosti, Belgrade), pp 31–34.
    .
  30. ↵
    1. Piperno DR,
    2. Weiss E,
    3. Holst I,
    4. Nadel D
    (2004) Processing of wild cereal grains in the Upper Palaeolithic revealed by starch grain analysis. Nature 430(7000):670–673.
    .
    OpenUrlCrossRefPubMed
  31. ↵
    1. Kuzmanović N,
    2. Vukojičić S,
    3. Barina Z,
    4. Lakušić D
    (2013) Sesleria serbica (Poaceae), a neglected species of the Balkan Peninsula. Botanica Serbica 37:113–120.
    .
    OpenUrl
  32. ↵
    1. Kilian B, et al.
    (2011) in Aegilops. Wild crop relatives: Genomic and breeding resources. Cereals, ed Kole C (Springer, Heidelberg), pp 1–76.
    .
  33. ↵
    1. Matsushima R,
    2. Yamashita J,
    3. Kariyama S,
    4. Enomoto T,
    5. Sakamoto W
    (2013) A phylogenetic re-evaluation of morphological variations of starch grains among Poaceae species. J Appl Glycosci 60:37–44.
    .
    OpenUrlCrossRef
  34. ↵
    1. Cârciumaru M
    (1973) Analiza poliniča a coprolitelor din statiunea arheologiča de la Icoana (Defileul Dunārii). Studii si Cercetari de Istoria Veche 24:5–13.
    .
    OpenUrl
  35. ↵
    1. Lucarini G,
    2. Radini A,
    3. Barton H,
    4. Barker G
    (2016) The exploitation of wild plants in Neolithic North Africa. Use-wear and residue analysis on non-knapped stone tools from the Haua Fteah cave, Cyrenaica, Libya. Quat Int 410:77–92.
    .
    OpenUrlCrossRef
  36. ↵
    1. Sawyer R
    (2006) Pollen Identification for Beekeepers (CLE Print Ltd, Cardiff, United Kingdom).
    .
  37. ↵
    1. Pope CA 3rd,
    2. Dockery DW
    (2006) Health effects of fine particulate air pollution: Lines that connect. J Air Waste Manag Assoc 56(6):709–742.
    .
    OpenUrlCrossRefPubMed
  38. ↵
    1. Perlès C,
    2. Quiles A,
    3. Valladas H
    (2013) Early seventh-millennium AMS dates from domestic seeds in the Initial Neolithic at Franchthi Cave (Argolid, Greece). Antiquity 87:1001–1015.
    .
    OpenUrlCrossRef
  39. ↵
    1. Çilingiroğlu C,
    2. Çakırlar C
    (2013) Towards configuring the Neolithisation of Aegean Turkey. Doc Praehist 40:21–29.
    .
    OpenUrlCrossRef
  40. ↵
    1. Karamitrou-Mentessidi G, et al.
    (2013) New evidence on the beginning of farming in Greece: The Early Neolithic settlement of Mavropigi in western Macedonia (Greece). Antiquity Project Gallery 87. Available at antiquity.ac.uk/projgall/mentessidi336/. Accessed August 17, 2016.
    .
  41. ↵
    1. Whittle A, et al.
    (2002) In the beginning: New radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antaeus 25:63–117.
    .
    OpenUrl
  42. ↵
    1. Hillson S
    (2001) Recording dental caries in archaeological human remains. Int J Osteoarchaeol 11:249–289.
    .
    OpenUrlCrossRef
  43. ↵
    1. Dickau R,
    2. Ranere AJ,
    3. Cooke RG
    (2007) Starch grain evidence for the preceramic dispersals of maize and root crops into tropical dry and humid forests of Panama. Proc Natl Acad Sci USA 104(9):3651–3656.
    .
    OpenUrlAbstract/FREE Full Text
  44. ↵
    1. Barton H,
    2. Torrence R
    (2015) Cooking up recipes for ancient starch: Assessing current methodologies and looking to the future. J Arch Sci 56:194–201.
    .
    OpenUrlCrossRef
  45. ↵
    1. Smith O, et al.
    (2015) Archaeology. Sedimentary DNA from a submerged site reveals wheat in the British Isles 8000 years ago. Science 347(6225):998–1001.
    .
    OpenUrlAbstract/FREE Full Text
  46. ↵
    1. Hardy K, et al.
    (2015) Dental calculus reveals potential respiratory irritants and ingestion of essential plant-based nutrients at Lower Palaeolithic Qesem Cave Israel. Quat Int 398:129–135.
    .
    OpenUrl
  47. ↵
    1. Torrance R,
    2. Barton H
    (2007) Ancient Starch Research (Left Coast Press, San Francisco).
    .
  48. ↵
    1. Buikstra JE,
    2. Ubelaker DH
    (1994) Standards for Data Collection from Human Skeletal Remains (Arkansas Archeological Survey, Fayetteville, AR).
    .
  49. ↵
    1. Fédération Dentaire Internationale (FDI)
    (1971) Two-digit system of designating teeth. Int Dent J 21:104–106.
    .
    OpenUrl
  50. ↵
    1. Cook G, et al.
    (2002) Problems of dating human bones from the Iron Gates. Antiquity 76:77–85.
    .
    OpenUrlCrossRef
  51. ↵
    1. Warinner C, et al.
    (2014) Pathogens and host immunity in the ancient human oral cavity. Nat Genet 46(4):336–344.
    .
    OpenUrlCrossRefPubMed
  52. ↵
    1. Svoboda C
    (1970) Grasses (Poaceae) from the Balkan Peninsula. Studii şi cercetări, Biology 17:19–45.
    .
  53. ↵
    1. Zaharieva M,
    2. Prosperi JM,
    3. Monneveus P
    (2004) Ecological distribution and species diversity of Aegilops L. genus in Bulgaria. Biodivers Conserv 13:2319–2337.
    .
    OpenUrlCrossRef
  54. ↵
    1. Howard T, et al.
    (2011) Identification of a major QTL controlling the content of B-type starch granules in Aegilops. J Exp Bot 62(6):2217–2228.
    .
    OpenUrlAbstract/FREE Full Text
  55. ↵
    1. Yang X,
    2. Perry L
    (2013) Identification of ancient starch grains from the tribe Triticeae in the North China Plain. J Arch Sc 40:3170–3177.
    .
    OpenUrlCrossRef
  56. ↵
    1. Henry AG,
    2. Hudson HF,
    3. Piperno DR
    (2009) Changes in starch grain morphologies from cooking. J Arch Sc 36:915–922.
    .
    OpenUrlCrossRef
  57. ↵
    1. Stoddard FL,
    2. Sarker R
    (2000) Characterization of starch in Aegilops species. Cereal Chem 77:445–447.
    .
    OpenUrlCrossRef
  58. ↵
    1. Radini A, et al.
    (2016) Neanderthals, trees and dental calculus: New evidence from El Sidrón. Antiquity 90:290–301.
    .
    OpenUrlCrossRef
  59. ↵
    1. Flodin U,
    2. Ziegler J,
    3. Jönsson P,
    4. Axelson O
    (1996) Bronchial asthma and air pollution at workplaces. Scand J Work Environ Health 22(6):451–456.
    .
    OpenUrlCrossRefPubMed
  60. ↵
    1. Oladele AK,
    2. Aina JO
    (2007) Chemical composition and functional properties of flour produced from two varieties of tigernut (Cyperus esculentus). Afr J Biotechnol 6(21):2473–2476.
    .
    OpenUrlCrossRef
  61. ↵
    1. Antonović D
    (2006) Stone Tools from Lepenski Vir (Arheološki Institute, Belgrade).
    .
  62. ↵
    1. Madella M,
    2. Lancelotti C,
    3. García-Granero J
    (2013) Millet microremains-An alternative approach to understand cultivation and use of critical crops in Prehistory. J Arch Anthrop Sci 8:17–28.
    .
    OpenUrl
  63. ↵
    1. Bergfjord C, et al.
    (2010) Comment on “30,000-year-old wild flax fibers” Science 328(5986):1634–1634, author reply 1634.
    .
    OpenUrlFREE Full Text
  64. ↵
    1. Blatt SH, et al.
    (2011) Dirty teeth and ancient trade: Evidence of cotton fibres in human dental calculus from Late Woodland, Ohio. Int J Osteoarchaeol 21:669–678.
    .
    OpenUrlCrossRef
  65. ↵
    1. Buckley S,
    2. Usai D,
    3. Jakob T,
    4. Radini A,
    5. Hardy K
    (2014) Dental calculus reveals unique insights into food items, cooking and plant processing in prehistoric central Sudan. PLoS One 9(7):e100808.
    .
    OpenUrlCrossRefPubMed
  66. ↵
    1. Dove CJ,
    2. Agreda A
    (2007) Differences in plumulaceous feather characters of dabbling and diving ducks. Condor 109:192–199.
    .
    OpenUrl
  67. ↵
    1. Dove CJ
    (1998) Feather evidence helps clarify locality of anthropological artifacts in the Museum of Mankind. Pac Stud 21:73–85.
    .
    OpenUrl
  68. ↵
    1. Harwood HP
    (2011) Identification and description of feathers in Te Papa’s Maori cloaks. Tuhinga 22:125–147.
    .
    OpenUrl
  69. ↵
    1. Dove CJ,
    2. Koch S
    (2011) Microscopy of feathers: A practical guide for forensic feather identification. Microscope 59:51–71.
    .
    OpenUrl
  70. ↵
    Borić D, et al. (2014) Late Mesolithic lifeways and deathways at Vlasac (Serbia). J Field Archaeol 39(1):4–31.
    .
View Abstract
Back to top
Article Alerts
Email Article
Citation Tools
Request Permissions
Mesolithic foragers consumed domesticated cereals
Emanuela Cristiani, Anita Radini, Marija Edinborough, Dušan Borić
Proceedings of the National Academy of Sciences Sep 2016, 113 (37) 10298-10303; DOI: 10.1073/pnas.1603477113
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
Proceedings of the National Academy of Sciences: 116 (13)
Current Issue

Submit

Jump to section

You May Also be Interested in

PNAS QnAs with NAS foreign associate and physicist Anne L’Huillier
Featured QnAs
PNAS QnAs with NAS foreign associate and physicist Anne L’Huillier.
Image courtesy of Erika Weiland (photographer).
Brain. Image courtesy of Pixabay/geralt.
Sex differences in metabolic brain aging
Brain metabolism scans in individuals aged 20–82 years revealed that compared with male brains, female brains appeared three to four years younger on average, suggesting potential links between sex and human brain aging.
Image courtesy of Pixabay/geralt.