The earliest appearance of domesticated plant species and their origins on the western fringes of the Eurasian Steppe

Archaeological research into cereal cultivation during the Neolithic-Chalcolithic periods in Ukraine to date has been concentrated on the western regions of the country, especially investigations into the Cucuteni-Tripolye Culture. In contrast, the study of archaeobotanical remains in eastern Ukraine has been limited. The theory most often postulated for the earliest appearance and spread of cereal cultivation in western Ukraine states that this phenomenon is connected with the west to east movement of the Linearbandkeramik Culture (hereafter LBK) peoples and Tripolye farmer expansions during the second half of the 6th–5th millennia calBC (e.g., Chernysh 1962; Zvelebil 1989; Zvelebil, Dolukhanov 1991; Anthony 1995; Zvelebil, Lillie 2000; Dolukhanov, Shilik 2007; Dolukhanov 2008). It has been suggested that agricultural expansion into the central territories of Ukraine was undertaken by farmer groups from the Cucuteni-Tripolye Culture, who followed the forest-steppe belt to the Dnieper River no earlier than the first half of the 5th millennium calBC (cf. Dolukhanov 1986; Anthony 1994; Whittle 1996; Sanzharov et al. 2000; Zvelebil, Lillie 2000; Pashkevich 2003; Telegin et al. 2003; Davison et al. 2006). According to Telegin (1968), the contacts between the Dnieper-Donets forager cultures and the Tripolye farmer populations are marked by the appearance of Tripolye pottery imports, the occurrence of cereal impressions in pottery, and some domesticated animal remains (Telegin 1968). The further eastward spread of cereal cultivation to the other half of Ukraine (eastwards from the Dnieper River) as well as to the south-east did not occur until the 4500–3000 BP (e.g., Velichko et al. 2009.7).


Introduction
Archaeological research into cereal cultivation during the Neolithic-Chalcolithic periods in Ukraine to date has been concentrated on the western regions of the country, especially investigations into the Cucuteni-Tripolye Culture.In contrast, the study of archaeobotanical remains in eastern Ukraine has been limited.The theory most often postulated for the earliest appearance and spread of cereal cultivation in western Ukraine states that this phenomenon is connected with the west to east movement of the Linearbandkeramik Culture (hereafter LBK) peoples and Tripolye farmer expansions during the second half of the 6 th -5 th millennia calBC (e.g., Chernysh 1962;Zvelebil 1989;Zvelebil, Dolukhanov 1991;Anthony 1995;Zvelebil, Lillie 2000;Dolukhanov, Shilik 2007;Dolukhanov 2008).It has been suggested that agricultural expansion into the central territories of Ukraine was undertaken by farmer groups from the Cucuteni-Tripolye Culture, who followed the forest-steppe belt to the Dnieper River no earlier than the first half of the 5 th millennium calBC (cf. Dolukhanov 1986;Anthony 1994;Whittle 1996; Sanzharov et al. 2000;Zvelebil, Lillie 2000;Pashkevich 2003;Telegin et al. 2003;Davison et al. 2006).According to Telegin (1968), the contacts between the Dnieper-Donets forager cultures and the Tripolye farmer populations are marked by the appearance of Tripolye pottery imports, the occurrence of cereal impressions in pottery, and some domesticated animal remains (Telegin 1968).The further eastward spread of cereal cultivation to the other half of Ukraine (eastwards from the Dnieper River) as well as to the south-east did not occur until the 4500-3000 BP (e.g., Velichko et al. 2009.7).Some researchers, however, have envisaged crop cultivation and the formation of domestic animal husbandries in Ukraine arriving from the opposite ABSTRACT -This paper presents the results of the first archaeobotanical investigation of Neolithic-Chalcolitich-period sites in eastern Ukraine and southwest Russia.The goal of this research is to understand the timeline of the earliest appearance and possible geographical origins of domesticated plants species in the region of study.The research conducted consists of the retrieval and study of macrobotanical remains and the analysis of plant impressions in pottery.Three possible corridors of influence upon agriculture in eastern Ukraine are postulated in this paper, originating from the Balkans, the Caucasus, and the Eurasian steppe.
KEY WORDS -Eurasian Steppe; domesticated plants; archaeobotany; Ukraine; Neolithic; Chalcolithic DOI> 10.4312\dp.39.1 direction -the Caucaso-Caspian corridor (Shnirelman 1989;1992;Jacobs 1993;1994;Kotova 2003; Levkovskaya et al. 2003;Kotova, Makhortykh 2009).Based on human dental studies from the Dnieper Rapids, Ukraine Jacobs (1993;1994) for example, suggested the possibility of an independent and pre-Danubian route of cereal cultivation in central Ukraine, arriving via the corridor between the Black and Caspian Seas.Some Ukrainian archaeologists, such as Nadezdha Kotova, have envisaged a very early Neolithic agriculture in south-eastern Ukraine (starting from end of the 7 th millennium calBC) (Kotova 2003).Kotova based her arguments on Pottery Neolithic sites in the northern Azov Sea region and Lower Don River, where domesticated animal bones, reaping knives, pestles, horn mattocks, grinding stones and cereal pollen have been reported (Belanovskaya 1995).The available pollen evidence includes '20 large grass pollen grains', presumed to be of cereal type, from the Neolithic level (attributed to 6350 calBC) at the Matveev Kurgan-I site on the northwest coast of the Azov Sea (Krizhevskaya 1992).Kotova and Tuboltsev (1992) reported the presence of domestic sheep at the Semenovka site (beginning of the 6 th millennium calBC) located on the northern coast of the Sea of Azov.However, no macrobotanical work has been done in this region.
To date, only one hulled barley impression in pottery from eastern Ukraine has been reported from the Serebryanskoe site located on the Donets River (Pashkevich 2003).A pottery shard with mollusc temper from the archaeological layer was radiocarbon dated to the 5 th millennium calBC (Sanzharov et al. 2000).However, it is not clear if the pottery with cereal impressions can be attributed to the dated layer.This is the only macrobotanical evidence from the Neolithic of eastern Ukraine and southwestern Russia available to date.
An alternative Eurasian steppe belt route for early agricultural dispersal was suggested by Jones (2004), proposing the arrival of the broomcorn millet crop (Panicum miliaceum) in Neolithic Ukraine from the Eurasian steppe.Broomcorn millet has been identified at several Neolithic sites which lie far to the north from the standard Anatolian east-west crop movement range (Jones 2004).Broomcorn millet is not known to have been cultivated in the Fertile Crescent prior to the 1 st millennium calBC (Nesbitt, Summers 1988).Therefore, the geographical origin of broomcorn millet may thus be presumed to lie elsewhere.The earliest known carbonised broomcorn millet remains are from central China, dated to around 8000 years calBC (Lu et al. 2009).Later dates show broomcorn millet cultivation at the end of the 7 th /first half of the 6 th millennia calBC in northern China (Cohen 2002;Zhao 2005;Crawford 2006;Liu et al. 2012).Fuller (2006) claims that China is without a doubt the place where millet was domesticated.So far, the earliest broomcorn millet record from the territory between China and the Urals is relatively late, coming from the Chalcolithic Sokolniki site (3200-2500 calBC) in the southern Tumen region in western Siberia (Shnirelman 1992).However, not much known about the nature of the find.
In order to understand the earliest appearance of domesticated crops in eastern Ukraine and to test the existing theories of their origins, the author conducted archaeobotanical investigations of both cereal impressions in pottery and macrobotanical remains recovered from Neolithic-Chalcolithic period sites in eastern Ukraine and south-western Russia.

Background information about the sites under investigation
Archaeobotanical investigations were conducted at 5 principle sites.Analyses of macro-plant remains were conducted at three of the five sites: Starobelsk-I, Novoselovka-III and Razdorskoe-II.The two additional sites of Rakushechny Yar and Zanovskaya were analyzed using pottery impressions only.To augment the study, an analysis of pottery for cereal impressions was conducted using samples from 12 additional secondary sites.All primary and secondary sites are located in the Don and Donets River basins in southwest Russia and east Ukraine (Fig. 1).

Razdorskoe-II site
The Razdorskoe-II site is a well-known multi-stratified site in the steppe region of southwest Russia (47°32'12.71N; 40°38'49.28"E).The site is situated on the right bank of the Don River on a 4-8m high terrace, a few kilometres downstream from the Razdorskoe-I and Rakushechny Yar settlements.The cultural layers of the Razdorskoe-II site are overlain by sterile layers of aeolian, colluvial and alluvial sediments 1m in depth.These sediments sealed the archaeological horizon, creating an anoxic calcareous environment which allowed for good preservation of bone material.The stratigraphy of the site consists of 12 cultural horizons.The anthropogenic horizons consist of thick accumulations of molluscs (mostly Unio and Viviparus spp.), animal bones and lithics.Black lenses in the stratigraphy consist of the charcoal and ash by-products of burning.Between the anthropogenic horizons, sterile layers of peaty humus, aeolian and alluvial silt are present (Tsybrii 2008).Nine radiocarbon dates from the early Neolithic layers of the Razdorskoe-II site have been reported (Aleksandrovsky et al. 2009) (Tab. 1).A total of 144 litres of sediments were collected for flotation from two hearths and a mollusc midden at depths of 195-205cm and 200-210cm dated to the early Neolithic period.

Rakushechny Yar site
Rakushechnyi Yar is one of the best-known archaeological monuments in southwest Russia from the Neolithic period, giving the name to the Rakushechny Yar culture (Belanovskaya 1995).The Rakushechy Yar settlement is located in the Rostov region, approximately 100km upstream from the present-day city of Rostov (Aleksandrovsky et al. 2009) in the steppe zone of southwest Russia (47°33' N; E 40°40').The site is situated on an island in the Ri-ver Don, almost opposite the Razdorskoe-I site and a few kilometres downstream from the Razdorskoe-II site.
23 cultural layers have been identified at the Rakushechnyi Yar site (Belanovskaya 1995), consisting of mollusc middens, ash, charcoal, and peaty deposits with alluvial sand clusters, all of which are distributed through a 6m thick stratigraphy (Aleksandrovsky et al. 2009).Tatyana Belanovskaya (1995) reports the presence of soil digging tools and grinding stones found in layers 3-5, and 7. Domestic sheep bones have been reported from the 21 st layer onwards, while the presence of domestic cattle has been reported from layer 20 onwards (Belanovskaya 1995).Pollen analysis of the site has not identified any pollen attributed to domesticated grasses, allowing Belanovskaya (1995.152) to state that "there is not enough evidence to suggest the presence of cereal cultivation by the inhabitants of the site".
The author conducted an archaeobotanical investigation of the site by analysing pottery for cereal impressions.In total, over 1000 pottery shards were analysed from the Rakushechny Yar site from layers 23-13 and 5-2.

Starobelsk-I site
The Starobelsk-I site is located in the steppe zone of the easternmost region of Ukraine, on the western edge of Starobelsk city (N 49°17'52.3, E 38°50'58.6).The site is situated on the left bank of the Aidar River, about 70-80m from the riverbank.Across the river from the Starobelsk-I settlement lies a steep chalk cliff.The Starobelsk-I site is located about 7km south of the Novoselovka-III site.The main part of the Starobelsk-I site is situated on a small elevation of the second Aidar River terrace, which consists of a narrow strip of raised land overgrown by trees.Part of the Starobelsk-I settlement is situated in an intensively ploughed area; therefore, the cultural layer at the site varies significantly in depth.The stratigraphy of the settlement consists of four clearly distinguishable lithological horizons.Parts of the cultural layer were constituted by anthropogenic mollusc clusters, consisting of Unio sp. and Viviparus sp.All the mollusc clusters were accumulated on the edge of the settlement, on the fringe of the third horizon.
The Starobelsk-I site contains one of the earliest examples of pottery in eastern Ukraine (Manko 2003).It has also been reported that the site contains domestic animal bones belonging to cattle, pig, dog, horse and sheep/goat (Gurin 1998).Judging from lithics, Gurin also inferred that the inhabitants of Starobelsk-I settlement used sickles for harvesting and processing domesticated cereals.
The excavation of the Starobelsk-I site was conducted in the summer of 2007 by the author and Sergiy Telizhenko, during which 50m 2 were investigated and 1704 litres of sediments from 12 fireplaces floated for the purpose of archaeobotanical investigation.
The chronology of the Starobelsk-I site was previously established through 14 C dating of pottery with a mollusc temper (Manko, Telizhenko 2002).One conventional and two AMS radiocarbon dates received from the site attributed it to the beginning of the 6 th millennium calBC (Tab.2).The AMS radiocarbon dates were conducted on a tree-branch charcoal and a Sus scrofa bone, which were found in the fireplace together with fragments of the one of the oldest potteries in Ukraine (Tab.2).

Novoselovka-III site
The Novoselovka site is located about 6km south of the Starobelsk settlement on the second terrace of the River Aidar (N 49°17'09.79;E 38°49'41.69).The site is situated in an open field, which is presently ploughed and irrigated, within a large loop of the Aidar; a few kilometres to the west and northwest, steep chalk cliffs surround the site valley.The total area of the settlement is not known.However, mollusc clusters ('mollusc middens' or 'kitchen middens'), representing a disturbed cultural layer, are distributed throughout the field over a few hectares.During the period of site's occupation, the entire territory was an island in the Aidar (Gurin 1998).The recovered bone remains and pottery at Novoselovka-III were mostly concentrated within the mollusc midden horizon, indicating that the basic environment (high pH) created by the mollusc remains allowed for the preservation of some artefacts.
The Late Neolithic period of the second half of the 6 th millennium calBC at the Novoselovka-III site was determined from one AMS radiocarbon date.During the summer of 2008, 1060 litres of sediments were floated for the purposes of archaeobotanical sample collection.

Zanovskoe site
The Zanovskoe site is situated in the steppe zone of eastern Ukraine in the Lugansk district, near the village of Orovskoe (or Barovskoe) (48°48'28.77"N -38°37'27.78"E).The Zanovskoe site is situated on a periodically inundated flood plain on the left bank of the Donets, and surrounded by two oxbow lakes, Zanovskoe and Matkino.Excavations revealed Chalcolithic pottery shards and flint tools (Telizhenko, Motuzaite Matuzeviciute 2007), and a range of domesticated animal species, including cattle, sheep, goat, pigs, horses and dogs (Zhuravlov, Telizhenko 2008).The site is chronologically younger than the Starobelsk-I and Novoselovka-III sites.Three existing 14 C dates from the Zanovskoe site were received from the Kiev Radiocarbon Laboratory from animal bones and pottery (Tab. 1) (Manko, Telizhenko 2002;Manko 2006).The dates attribute this site to the Chalcolithic period Sredny-Stog culture, dated by Telegin et al. (2003) to 4400-3500 calBC.No flotation was conducted from the Chalcolithic period layers; only pottery shards were analysed for cereal impressions.

Methods
Samples of macrobotanical remains were collected from a variety of archaeological features and phases encountered at the archaeological sites: fireplaces, pits, house floors, and waste dump places such as mollusc middens.The collection of sediment samples and the procedure for manual flotation followed the descriptions in Deborah Pearsall (2000).Soil samples from each archaeological feature were collected in bags, according to the stratigraphy within each feature.The quantity of sediments taken for each sample and from each feature was then recorded.A standard 12l size bucket was used in order to keep track of the sample volume taken from the feature.Water for flotation was obtained from whatever nearby water source was available, such as oxbow lakes, rivers and springs.A 300µm mesh size was used for flotation to ensure the full recovery of plant seeds.Once the botanical material was retrieved from the sediment, each mesh containing the float was collected, labelled, and dried in a location away from direct sunlight.The sorting and identification of archaeobotanical material took place at the University of Cambridge in the George Pitt-Rivers Laboratory for Bioarchaeology.Each flotation sample was sorted individually by selecting, counting and identifying all charred seeds within the sample.
The procedure used for initial analysis of cereal impressions in pottery from shard collections involved a careful visual review of each pottery shard from both sides, specifically concentrating on evaluations of the pottery's clay temper and noting any full or partial plant part or seed/grain impression through the use of a magnifying glass.
The criteria applied for the identification of plant impressions are as follows.For a satisfactory identification to species, a cereal impression had to contain not only a distinctly impressed seed or grain shape, but also other features such as an impression of the lemma/palea pattern, grain/scutellum, or the ventral furrow with hilum.Usually, cereal impressions were identified to species if chaff components were found; chaff characteristics can be more species-specific than grain impressions, especially the glume bases of Triticum spelta or spikelets or rachis segments of Hordeum vulgare, etc. Cereal impressions found to fit the identification criteria to genus or species levels were usually recorded as cf.Plant impressions that did not fully fit the identification criteria were not recorded at all.Access to pottery shard collections was obtained at the Lugansk Hi-

Razdorskoe-II site
The flotation samples contained a large quantity of charcoals and constituted mostly of woody plant species.As can be seen in  (Fig. 2) was found in the pottery with a mollusc temper.
In the same shard, the charred glume base of Triticum spelta (spelt wheat) (7.5mm long and 3.7mm wide) was found preserved in the inner wall.This impression contains clear characteristics of spelt wheat glume dorsal veins (Fig. 2).
Finally, impressions of what appeared to be Vicia sp.genus were noted.However, the impressions were not very clear, and therefore kernel components such as hilum or radicle were difficult to identify with confidence.The second layer has three radiocarbon dates ranging over the period 5787-3372 calBC (Timofeev et al. 2004).
As can be seen, the date range of layer 2 is very broad.Such a broad time range from the same archaeological horizon might be the result of dating bioturbated material (thereby including material from the upper layers), dating charcoal, the dates from which may be biased by the 'old wood effect'.Surely, such a wide date range is too wide to gain the most accurate insight into the introduction of agriculture in the region.Nevertheless, cereal impressions in pottery and the radiocarbon dates presented above are so far the only information available to the author.
In layer No 3, both heavy, thick-walled vessels with a mollusc-organic temper and porous, light vessels with a cereal chaff temper were found.The second type of pottery contained glume impressions of wheat (Triticum spp.) (Fig. 3).In this pottery, impressions of cereal parts and imprints of wild plants and seeds were noted.This cereal chaff temper-type of pottery from Rakushechny Yar is similar to that of the Tripolye culture, where cereal-processing waste was commonly used as a clay temper for pottery and daub production (Pashkevich, Videiko 2006).Pottery with a cereal chaff temper dated to the second half of the 5 th millennium calBC was also found at the Zanovskoe site, where impressions of hulled and naked barley were detected (see below).One radiocarbon date exists from the 3 rd layer of Rakushechny Yar, falling between 3357-2702 calBC (Tsybrii 2005); that this date is younger than the layer above shows the great need to re-date the stratigraphical sequence of the site.
In layer No. 4, cereal impressions were found in two types of pottery: a light and porous pottery with a chaff temper (as above), and pottery with a vegetative matter (grass) and sand-rich temper.Most of the cereal impressions were found in pottery shards with the cereal chaff temper.The vessels with the cereal chaff temper are very light and porous, containing fragments of cereal glumes which burned away during the vessel firing process, but had their shapes preserved within the pottery walls.In this  type of pottery, the imprints of hulled barley (cf. Hordeum vulgare) (4.4mm long and 3.2mm breadth) (Fig. 4), Triticum spp.chaff and probably one broomcorn millet seed (cf.Panicum miliaceum) (2.8mm long, 1.9mm breadth) were identified.The broomcorn millet impression contained a seed with glumes and preserved parts of the charred grain.
The distinct pattern of the lemma and palaea was still present on the pottery shards, which allowed the imprinting seed to be identified to the species level as broomcorn millet.Only one impression of wheat (Tritucum cf.aestivum) (3.8mm long and 2.48mm wide) was found in the pottery type with a mollusc and grass temper (Fig. 4).Four radiocarbon dates were obtained from layer No. 4, ranging from 6026-3365 calBC (Aleksandrovsky et al. 2009;Telegin et al. 2003;Timofeev et al. 2004).
In layer No. 5, the pottery styles change to a type consisting of only heavy, thick-walled vessels with a grass, sand and mollusc temper.No shards with a cereal chaff temper were found in this layer.Wild plant seed impressions were found only from shards with the grass-rich clay temper.In this type of pottery, only one impression of a seed of Thalictrum cf.minus (lesser meadow rue) (4.6mm long and 2.4mm wide) was identified (Fig. 4).Layer 5 has three radiocarbon dates, ranging from 5479-4501 calBC (Telegin et al. 2003;Timofeev et al. 2004).
Pottery recovered from lower layers did not exhibit much contact with plants, and only a few impressions of wild plant parts and seeds were noted.Owning to the absence of a plant reference collection in the museum archives of Saint Petersburg, these plant species were not identified.Most pottery vessels in layers 8-20 were very robust and heavy, made of clay or rivermarl, with a sand, river silt and mollusc temper, and fired at a low temperature.Most of the dates from these layers fall into the period of the 7 th millennium calBC (e.g., Telegin et al. 2003).

Starobelsk-I site
The macrobotanical remains recovered from flotation samples consist of wood charcoal and land snails.The abundance of modern rootlets in the samples strongly correlates with the amount of modern contaminate seeds of Chenopodium sp. in the flotation samples.Both modern rootlets and Chenopodium sp.seeds were found in areas where the cultural layer was at its shallowest.Among the charred plant seeds discovered, only a few seeds of Silene sp.(campions) genus plants were identified in fireplace No. 5. Two seeds of Galium aparine (cleaver) were found in fireplace No. 10.In the same fireplace, one seed of Stellaria sp.(stitchwort) was identified (Tabs.5-6).The small quantity of plant seeds found at the site limits any contribution to a wider discussion of human and plant interaction, or the past ecology at the site.
It is important to note a few taphonomic aspects of the Starobelsk site.Firstly, the fireplace structures were constructed directly on the ground surface, with no deepened fire-pit structure.Such a form of fireplace construction might indicate that the fires were burning in an oxygen-rich environment capable of turning any plant remains into ashes, and therefore leaving very little plant material remaining for later recovery.Furthermore, the high abundance of terrestrial molluscs, which is greater than the amount of charcoal found in the samples by many orders of magnitude, indicates that after abandonment, the fireplaces stood exposed on the ground surface for a long time, before being deposited by aeolian or alluvial sediments.Their long exposure prior to deposition could also have affected the preservation of charred plant remains at the site.Moreover, experimental work on charcoal preservation in alkaline environments has demonstrated both higher fractionation and degraded preservation quality of deposited material (Braadbaart et al. 2009).

Novoselovka-III site
During the archaeological excavation of the Novoselovka-III site, a mollusc midden zone and one fireplace were used to obtain samples for archaeobotanical investigation.Most of the flotation samples taken for archaeobotanical data collection were from the mollusc midden feature, which was selectively sampled in places with a concentration of charcoal, burned bone and burned flint pieces.Additionally, the entire content of the fireplace was sampled.In total, 1060 litres of sediments were floated from Novoselovka-III.After sorting the archaeobotanical samples from all the floated sediments, a very small quantity of 16 charred plant seeds were recovered (Tab.7).

Analysis of pottery impressions from Neolithic settlements in eastern Ukraine
15 Neolithic and Chalcolithic sites in Ukraine were analysed for the presence of cereal impressions in their pottery remains (see the site location map in Fig. 1).Some pottery shards were accessed at the Lugansk History Museum storage centre, while others were analysed during periods of archaeological excavation (Tab.8).Pottery shards were investigated from the following sites: Orekhovo-Donetskoe-III, Starobelsk-I, II and III, Olkhovaya-V, Zelena-Gornitsya-I, V and IV, Kleshnya-II, Podgorovka-I and V, Novoselovka-I and III, Tuba-II and Zanovskoe.In total, approximately 4500 pottery shards were analysed.

Tab. 5. Plant macro-remains from the Starobelsk-I site, eastern Ukraine. Sieve size -300 µm. Key: + <10 items; ++<20 items; +++ <100 items; >1000 items ++++.
Except for the Zanovskoe settlement (see below), none of the analysed pottery shards from the sites listed above contained any clearly defined cereal impressions.A few shards from the Novoselovka-I, Olkhovaya-V and Tuba-II sites had impressions of seeds, caryopses and plant parts which appeared similar to domestic cereal.However, none of those impressions could be attributed to a domestic cereal species with confidence.For example, some seed impressions from the organic-rich pottery recovered from Tuba-II were discerned, the shapes of which appeared to be similar to that of broomcorn millet seeds.These elongated seeds were impressed in small clusters together with their stems.The seed impressions in the Tuba-II pottery were not impressed to a full seed shape, and their contours were not distinct.With the absence of either clear seed shape impressions or lemmas and paleas surface patterns, such cereal impressions cannot be confidently identified.
The only pottery shards where cereal chaff and grain impressions were identified with confidence were found at the Zanovskoe site.At Zanovskoe, three main types of pots were found: a type light and rich in organic temper with a polished surface; a type rich in organics with a rough surface; and heavy clay pots with a mollusc and coarse sand temper.Cereal impressions were identified on a few pottery shards of the first pottery type with an organic temper.The pottery shards with cereal impressions were found in the Chalcolithic period pottery, at a 50-60cm depth.This layer has been dated to 4462-3525 calBC (Manko, Telizhenko 2002).More precise analysis of these shards under the microscope has shown that the pot clay contains a cereal chaff temper.The cereal impressions represent the glumes and palaea of naked barley and a grain of hulled barley (Hordeum vulgare var.nudum and Hordeum vulgare), as well as different parts of the cereal chaff, probably that of the Triticum genus (Fig. 5).

Discussion of archaeobotanical results from the lower Don and Donets river basins
Archaeobotanical investigation has been conducted on several 8-6 th millennium calBC Date of sampling 12.07.'0714.07.'0714.07.'0713.07.'0713.07.'0713.07.'0713.07.'0717.07.'0717.07.'0717.07.'0718.12.07.'0718.12.07.'07 18.12.07.'0718.07.'07 18.07.'0718.07.'07sites situated in the Lower Don and Donets River basins.This investigation has not provided any evidence for cereal cultivation in this area during this particular period of prehistory.In addition, the archaeological or zooarchaeological evidence from these sites suggests that subsistence in this region was based on the exploitation of wild food resources, rather than food production.

Chenopodium
At the Razdorskoe-II site, dated to the 8 th -first half of the 7 th millennia calBC, a large variety of wild animal species were identified (Tsybrii 2008), demonstrating highly developed hunting skills amongst the population.This site was especially abundant in mollusc and fish bones, indicating specialised human exploitation of water resources.An abundance of bone harpoons and stone weights probably used for sinking nets, along with tools made from molluscs shells, show that human activities were tightly aligned with fish consumption (Tsybrii 2005) and food-procurement activities.The same conclusions were reached from the analysis of the Neolithic layers of the Rakushechny Yar and Razdorskoe-I sites (Aleksandrovsky et al. 2009).Furthermore, previous palynological analysis of the Razdorskoe-II site was carried out by Kremenetski (1991) and Spiridononova et al. (published in Tsybrii 2008) from approximately the same period as the macrobotanical samples reported above were retrieved.The pollen analysis at Razdorskoe-II has shown no evidence of any pollen belonging to domestic cereal species.
A situation similar to the Razdorskoe-II site was revealed while investigating the Starobelsk-I and Novoselovka-III settlements of eastern Ukraine.Except for the remains of dog, the rest of the animal bones from the 2006 excavation of Starobelsk-I resulted in the discovery of only wild species (Telizhenko, Motuzaite Matuzeviciute 2007).River mollusc middens combined with fish and turtle bones indicate the exploitation of the fresh water resources at both sites.
The abundance of flint chips at Starobelsk-I was concentrated mostly in areas around the fireplaces, indicating that tool-making activities took place at the site.It has been argued that most flint tools from the Starobelsk-I settlement are linked with hunting, flaying, meat cutting and scraping (Telizhenko 2007).It has to be mentioned that Gurin (1998) regarded the flint blades at Starobelsk as cereal harvesting tools.However, during the 2007 archaeological excavation, a flint blade was found embedded in a mollusc shell, indicating that it was used to open shells and extract their contents.
It can also be pointed out that the fireplaces at the Starobelsk-I site were wide (up to 2m in diameter), and constructed directly on the ground surface.None of the fireplaces contained any permanent structure, such as a clay bedding or stone circles, indicating their temporary use.The fireplaces at Starobelsk-I often have heat epicentres in a few different places, showing reuse of the fireplace location over different periods.The absence of pits, dwelling construction, or postholes also suggests a temporary, probably seasonal, camp where people had simple ground dwellings, or no dwellings at all.The geographical location of the site also indicates the selection of the site for hunting purposes.Across from the site the River Aidar makes a bend beneath a set of steep chalk cliffs, a situation that may have benefited the hunting strategies of the site occupants.All this evidence seems to suggest that the Starobelsk-I site was a hunting campsite, the inhabitants of which hunted wild animals and exploited river resources.
According to 14 C dates the Novoselovka-III mollusc midden site represents a later chronological period than the Starobelsk-I site; however, the subsistence patterns of its inhabitants were still very similar to those at Starobelsk-I.No traces of house constructions, postholes, or pits from the Late Neolithic period were found at Novoselovka-III.The presence of wood charcoals, burned bone and flint fractions in mollusc middens, coupled with a few wild plant seeds, show that cereals were not a part of normal kitchen waste.However, one must keep in mind that very little evidence of fire-making activities were found at the site, and therefore any potential plant crops used at the site would not have survived in an un-charred state.
The investigation of Novoselovka-III provided evidence of the population's extensive involvement in mollusc gathering activities for subsistence purposes.The seasonality studies of mollusc consumption among the inhabitants of the Ertebølle culture in Denmark has shown that the molluscs were collected in early spring -March and April (Milner 2002).Mollusc meat is not an adequate protein replacement for other animal meats; it is actually highly rich in carbohydrates and can substitute as a replacement for grains (Greenfield, Jongsma 2008).Haskel J. Greenfield and Tina Jongsma (2008) note that in areas where grains were not consumed, or consumed in very small quantities, the collection of molluscs was of greater importance.
The discoveries of wild animal bones, bone awls, stone axes, a variety of flint tools and flint chips at the Novoselovka-III site indicate more diverse activities at the site than only mollusc gathering.Furthermore, the Novoselovka-III site contained a large quantity of pottery, with the largest vessels being up to 50cm in height and 34cm in diameter (pers.comm.Sergey Telizhenko), indicating the reduced mobility of the population.Moreover, a pottery ornamentation stamp made from an incised mollusc shell (Unio sp.) was found at the Novoselovka-III site in a shell midden (Fig. 6).All these facts could suggest both local pottery production adjacent to the mollusc midden site and the proximity of human settlement to the mollusc midden site.Such evidence reinforces the argument that the absence of evidence for cereal cultivation is not because the wrong areas were sampled, but rather because it is very unlikely that the people of Novoselovka-III were involved in agriculture.
It has to be mentioned that the Novoselovka-III site is unique in its abundance of pottery and variety of pottery styles.The variety of vessel ornamentation patterns, clay temper, and the amount of pots per area excavated in Novoselovka-III site constitute the greatest pottery collection from Neolithic Ukraine.From the total excavated surface area of 133m 2 , 25 complete vessels were found.It can be noted that pottery vessels very similar in ornamentation, clay admixture and shape to those from the Novoselovka-III site have been found in layer 10 at the Rakushechny Yar site (Fig. 7), showing close contacts between the Novoselovka-III and the Lower Don populations.Overall, it has to be pointed out that the similarities in Neolithic pottery-making technology between the Novoselovka I-III and Starobelsk-I sites with the Lower Don region sites are much stronger than with that of the peoples inhabiting the Dnieper River region.The westward movement of pottery technology from Russia to Ukraine has been noted by many researchers (Danilenko 1969;Belanovskaya 1995;Kotova 1998;Kuzmin, Orlova 2000;Sanzharov et al. 2000;Gronenborn 2003;Dolukhanov, Shukurov 2004;Dergachev, Dolukhanov 2007).Such interregional similarities are important for understanding the zones of influence by human interaction along connecting waterways, which later played a role in the exchange of agricultural products.
The evidence discussed above leads to the conclusion that the human populations in the Donets nor Don basins did not practice cereal cultivation during the 6 th millennium calBC.A similar situation obtains with domestic cattle.The identification of animal species at Neolithic sites in Ukraine and Russia as domestic has added to the confusion about the origins of agriculture in the region.Animal remains at sites are usually identified as domestic without following standard animal bone gender patterns and morphometric analysis of decreasing size and age.Furthermore, due to the osteological similarity between sheep/goat and steppe antelope (Saiga tatarica), especially in dentition (Hillson 2005), the identification of Near Eastern sheep and goats are probably often confused with steppe antelope species.Steppe antelope was a very frequent game animal among the prehistoric inhabitants of the steppe (Vekilova 1971;Dolukhanov 2009).It has been noted, however, that at many Neolithic sites only sheep/ goat or steppe antelope bones have been identified, probably indicating some confusion in differentiating these threeanimal species from each other (Tsybrii 2008).A specimen from Semenovka site, identified by Kotova as sheep bone, from the lower layers of the settlement, was dated at the Beijing Radiocarbon Accelerator Unit for AMS radiocarbon dating.The date ranged between 5617-5482 calBC (BA-071462; 6595±40 BP) (Fig. 8).It was on the sheep/goat remains from the Semenovka site that N. Kotova based her theory of the Caucasian-Steppe origins of agriculture in eastern Ukraine (Kotova 2009).However, according to the opinion of a zooarchaeologist colleague from the University of Cambridge (pers.comm.Katie Boyle, Ryan Rabett and Tony Legge), this bone specimen is more likely to be that of Saiga tatarica than Ovis aries orientalis.
The presence of agriculture in the Lower Don and Donets basins has been detected by the author only in the Chalcolithic period.Cereal impressions in pottery were found at the Rakushechny Yar in site layers 4-2 and Zanovskoe's Chalcolithic layers.Rakushechny Yar and Zanovskoe are attributed to a very similar period of the Sredny-Stog culture and dated to the second half of the 5 th and the 4 th millennia calBC.The archaeobotanical investigation of the Rakushechny Yar site pottery has shown that the inhabitants were familiar with cultivating spelt wheat (Triticum spelta) and other wheat species, hulled barley (cf.Hordeum vulgare), probably flax (cf.Linum usitatissimum) and possibly broomcorn millet (cf.Panicum miliaceum).The Zanovskoe inhabitants were probably cultivating hulled and naked barley (Hordeum vulgare, Hordeum vulgare var.nudum) and wheat (Triticum spp.).Unfortunately, the cereal impressions in pottery from the Zanovskoe and the Rakushechny Yar sites do not have direct dates, but only dates of the layer and, therefore, constitute a less accurate piece of information than dates directly derived from dating charred cereal grains.It also should be pointed out that the dates from Rakushechnyi Yar were received from pottery consisting of a mollusc or humus temper -material not very appropriate for dating.The dates also have a very large error band of up to ±300 years, making these dates very inaccurate (see Tab. 1).Nevertheless, neither of the Zanovskoe or Rakushechny Yar  sites are being excavated at present, which means that flotation or radiocarbon cannot be applied, and that the existing pottery impressions and dates constitute the only information about the presence of agriculture in the Donets and Don basins from around the 5-4 th millennium calBC.During this period, an increase in the use of domestic animals in the region can be noted; a large quantity of domestic animal bones has been reported from the Zanovskoe and Rakushechny Yar sites and from other sites.For example, from the Chalcolithic layer at Zanovskoe (350m 2 ), O. P. Zuravlev identified bones from twenty-two individuals of domestic cattle, eight sheep, six goats, eight sheep/goat, twenty-two 'domestic' horses and eight dogs (Zhuravlov, Telizhenko 2008).Viktor Tsybrii (2008.71)notes that at the end of the Neolithic and Chalcolithic periods, the populations of the lower Don became predominately cattle and horse breeders.It is argued that during the 5 th millennium calBC, some changes in the society of the Donets Basin can be seen, for example, an increase in population represented by the appearance of large burial grounds, such as Alexandria (Manko 2003;Rassamakin 1999); the development of metallurgy, where the first metal objects and crucibles were found at the Kleshnya site (Manko 2006;Manko, Telizhenko 2002), and others.
It can be only speculated about the geographical origins of domesticated plant species in the region.The arrival of agriculture in the territory of Ukraine could be seen as taking place not only following the branched Danubian route, but also by following the later steppe and Caucasian route.The broomcorn millet found at the Rakushechny Yar was probably introduced via the steppe corridor from the populations at the present territories of China.Unfortunately, not much is known about millet cultivation in central Asia.So far, the earliest 14 C dated (end of the 3 rd millennium calBC) Panicum miliaceum grain was found in eastern Kazakhstan (Frachetti et al. 2010).
The  2011).At the fortified Svobodnoe site, 39% of all tools discovered were attributed to cereal cultivation and processing activities (Nekhaev 1992).The Svobodnoe site was radiocarbon dated to the second half of the 5 th millennium calBC (Rassamakin 1999).Authors have reported the presence of farming village establishments along the River Kuban, where large granite cereal grinding stones and flint sickles for cereal cultivation were recovered (Formozov 1965; Korenevskii 2004).
To the north Caucasus, agriculture spread from the central and southern Caucasus, where it existed during the Shulaveri-Shomutepe Culture, dated to the 6-5 th millennia calBC (Lisitsina, Prishchepenko 1977;Lisitsina 1978;Nebieridze 1978;Lisitsina 1984).Irrigation channels were found at the Arukhlo-I, Imiris-gora, Chakh-tepe, Kiul-tepe and Alikemek-Pepesi sites (Korobkova 1999).The sites with domesticated animal and plant remains that have seen the best investigation of archaeobotanical data are the Aratashen and Aknashen Neolithic sites in Armenia, where the majority of radiocarbon dates fall into the period of the first half of the 6 th millennium calBC (Hovsepyan 2004;Hovsepyan, Willcox 2008).Early cultivars started to spread north, reaching the northern territories of the Caucasus pro-bably only during the early metal period around the 5 th millennium calBC.The discovery of a stratified Late Mesolithic-Early Neolithic-Bronze Age site, such as Tsmi (7 th to the 3 rd millennia calBC) at an elevation of 1700m in the northern Caucasus shows "the importance of the traverse across the nearby passes in the longue durée of Caucasian communication network" (Rostunov et al. 2009.73) and the possibility that the Neolithic populations of northern Caucasus had contact with the agricultural societies to the south.
According to Yuri Rassamakin (1999), in the Chalcolithic period, the route of a bi-directional network of interaction via the steppe belt stretched from the Prut River all the way to the north-west Caucasus.Imported items from the food producing pre-Maykop populations of the Kuban River region in the northern Caucasus are distributed through Sredny-Stog and Tripolye sites (Rassamakin 2004).For example, serpentinite from the Caucasus were found at the Novye-Ruseshti site in Moldova, green serpentinite axes also from the Caucasus, and pottery typical of the fortified agricultural village of Svobodnoe (north Caucasus) were found in the Sredny-Stog sites.This demonstrates the existence of interactions between the northern Caucasus region and the Ukrainian steppe already during the Early Chalcolithic period, which continued all the way through to the Maykop culture period (Rassamakin 1999;Anthony 2007).Rassamakin (1999) also emphasises the close similarity in the pottery making traditions of the northern Caucasus region with the Lower Don region during the Chalcolithic.The similarities in the material cultures of the steppe populations in Russia and Ukraine and northern Caucasus populations have also been noted by Nikolay Merpert (1994).
To conclude, the archaeobotanical investigations conducted by the author have provided evidence that cereal cultivation began in eastern Ukraine and south-west Russia around the second half of the 5 th millennium calBC, as demonstrated by the analysis of pottery impressions from the Zanovskoe and Rakushechny Yar sites.The earliest evidence of cereal cultivation in eastern Ukraine and southwest Russia comes from the Sredny-Stog culture sites at Zanovskoe and Rakushechny Yar, where cereals and their chaff impressions were identified by the author in the layers dated to the second half of the 5 th to the first half of the 4 th millennium calBC.A variety of cereal species were identified from the grain and chaff impressions, comprising hulled and naked barley, spelt wheat, probably flax, and broomcorn millet species.
The archaeobotanical and archaeological evidence has shown that the possibility cannot be excluded that the spread of cereal cultivation into the eastern regions of Ukraine and south-western Russia may have arrived from the Caucasian corridor and the Eurasian steppe, but only during the Chalcolithic period.
Fig. 5. Lemma and palea impressions of naked barley (H.vulgare var.nudum) in Zanovskoe pottery from the second half of the 5 th millennium calBCfirst half of the 4 th millennium calBC layer, 50-60cm depth horizon, square 320.

Table 3
, the samples contained a very low quantity of charred plant seeds.Only fractions of Chenopodium cf.album (fat-hen), Hedera helix (common ivy), Persicaria sp.(knotweeds), Thlaspi cf.arvense (field penny-cress) plants and one un-identified seed were found in samples tained glume impressions of mostly Triticum spp., (wheat) crops.One shard containing a seed impression probably attributable to cf.Linum usitatissimum (flax) (4.8mm long and 2.2mm wide)

. List of sites from where pottery sherds with cereal impressions were analysed. Fig. 8. The frag- ment of Ovis aries/ Saiga tatarica meta- tarsal from the lo- wer cultural layer at the Semenovka site dated at the Beijing Radiocar- bon Accelerator Unit for AMS.
wheat and barley agriculture of the sites in eastern Ukraine and southwest Russian could have arrived from the northern Caucasus region.There is substantial evidence of agriculture in the northern Caucasus region during the early stages of the Chalcolithic period.The best-known site in the northern Caucasus, the Chokh settlement in the Dagestan re- gion, is probably dated to Chalcolithic-Bronze Age (pers.comm.Nadia Kotova).Khizri Amirkhanov  (1987)reports cultivars of barley (Hordeum vulgare var.polystichum, Hordeum var.nudum), millet (Pa-nicum spp.), and wheat (Triticum monococcum, Triticum dicoccum, Triticum aestivo-compactum) from the site, retrieved by applying flotation methods to the settlement layers.Besides the Chokh site, more Early Chalcolithic sites in the northern Caucasus provide evidence for agriculture, such as Svobodnoe, Yasenovaya Polya, Meshoko, Zamok, Khadzhokh, etc. Archaeobotanical data from the Chalcolithic fortified Svobodnoe site; macro report the remains of grain and chaff of Triticum monococcum, Triticum dicoccum, and Hordeum vulgare were found from the flotation samples (Lebedeva am extremely grateful to the Cambridge Gates Trust and the Arts and Humanities Research Council for funding my fieldwork in Ukraine.I am grateful to all the archaeologists with whom I worked in Ukraine and Russia, and who helped with access to literature and research material, especially S. Telizhenko at the National Academy of Science in Crimea, M. Tovkailo, D. Gaskevich, A. Yanevich, N. Mikhailova, A. Dyachenko, N. Kotova, and G. Pashkevich, I. Patechina at the Kiev Institute of Archaeology, as well as V. Tsybrii from the Department of Archaeology, Southern Federal University at Rostov-on-Don, Russia, and A. Mazurkevich and K. Dolbunova at the State Hermitage Museum in Saint Petersburg, Russia.I would also like to thank Malcolm Lillie at the University of Hull (UK) and his NERC ORADS grant at the Oxford Radiocarbon Accelerator Unit (OxA) for providing radiocarbon dates from some studied sites and also Zhou Li-Ping at the Accelerator Mass Spectrometer Unit in Peking University, Beijing for dating a sheep/ antelope bone from the Semenovka site in Ukraine.