by Chris Fowler
Neolithic livestock species have complementary ecological preferences and productive potential. Sheep, as specialist grazers, traditionally converted crop stubble and fallow weeds into manure, whilst goats, cattle, and pigs also browsed or rooted in woodland and scrub. Pigs produce most offspring, followed by goats, sheep, and finally cattle. Sheep and goats were milked more than cattle in southern Europe, whilst pigs especially provided non-dairy fat for cooking and preserving meat. Cattle provided draught: oxen (castrated males) for big landowners, and cows for smallholders. Finally, sheep wool and goat hair were woven into clothing, bedding, and sacks, though the ‘hairy’ coat of early sheep was less useful. A mixture of livestock species thus favours herd security and self-sufficiency in a range of products, but conflicting feeding requirements limit the number of animals that can be kept. Conversely, a single species reduces security and self-sufficiency, but facilitates maintenance of large herds and specialization in particular products.
The products offered by an animal also depend on age and sex. Culling patterns cannot demonstrate actual use for milk, meat, or wool/traction, but clarify potential intensity of use and can be tested against life history evidence: for example, stress-related pathologies in cattle limb bones may reflect use as draught animals. Species composition, mortality, and life history thus shed partial but complementary light on the methods and outcomes of animal management. Similarly, food residues in ceramics may demonstrate processing of milk and non-dairy adipose fat, but shed no light on intensity of use. Preservation (mainly waterlogged) of textile fibres and wooden yokes, wheels, or ploughs provides welcome additional detail, but is unusual in southern Europe. Dental microwear sheds broad light on diet, including perhaps the degree of grazing pressure, in the weeks before death. Diachronic changes in biometry and morphology reflect longer-term effects of management: for example, poor nutrition favours smaller body size and competition between adult males for mates the reverse.
Stable carbon and nitrogen isotope ratios in bone collagen and tooth enamel provide direct evidence of human diet, albeit at coarse resolution (e.g. marine versus terrestrial). Attempts to identify terrestrial Neolithic diets as crop or animal-based are problematic since they depend on local isotopic signatures in animals and plants, and the latter are usually unknown. On-site traces of dung imply availability of ‘stall-manure’ for distribution and reveal where livestock were sheltered, whilst associated plant remains may shed more detailed light on animal diet than stable isotopes or dental microwear.
SOUTH-EAST EUROPE
Plant use and husbandry in south-east Europe
Most archaeobotanical evidence derives from open-air sites and is relatively extensive from Greece, the former Yugoslav Republic of Macedonia (FYROM), and Bulgaria. Despite variable sampling and recovery, crops clearly dominate most assemblages from the EN onwards; edible nuts, fruits, and other wild plants occur frequently, but usually at low levels, and evidence of storage is rare. Crops include several types of wheat (einkorn, emmer, free-threshing), barley (hulled, naked), and pulses (lentil, pea, grass pea, bitter vetch, chickpea), all represented by ‘storage’ finds in Bulgaria (Marinova 2007) and most likewise in Greece (Halstead 1994). In recent multi-site programmes of intensive sampling and recovery in Greece (Valamoti 2004, 2005) and Bulgaria (Marinova 2006), ‘flat’ sites with relatively thin occupation layers such as Kovacevo yielded lower densities of charred crop remains than tells with deep deposits. Differences in preservation conditions rather than commitment to agriculture have thus shaped much of the observed variability in archaeobotanical representation of crops.
Burnt houses with in situ stores suggest household production and consumption of a range of cereals and pulses. Bulgarian house stores such as those recovered at Slatina (Marinova 2006) suggest that mixed harvests of einkorn and emmer wheat were the dominant staple, stored as ears or spikelets (grains enclosed by chaff) requiring piecemeal dehusking. Harvesting of ears only has been inferred from the heights of weeds (Kreuz et al. 2005; Marinova and Thiébault 2008). In such stores, pulses tend to account for 20–30% by volume relative to cereals—a high proportion compared with recent extensive agriculture.
Early Neolithic (Starčevo-Criş-Körös, sixth millennium BC) settlements from Serbia, south-east Hungary, and southern Romania are predominantly ‘flat-extended’. Published archaeobotanical data are sparse but include multiple cereals and pulses. For example, einkorn, emmer, barley (mostly hulled), lentil and pea, as well as wild Cornelian cherry, are known at several Starčevo sites (Borojevič 2006, table 2.5). In Hungary, systematic flotation at Ecsegfalva 23 has yielded einkorn, emmer, and barley, with traces of other cereals and lentil; collected wild plants included water chestnut, strawberry, and hazelnut (Bogaard et al. 2007). Elsewhere, evidence of Körös use of wild plants includes a layer of hazelnut shell in a pit at Méhtelek-Nádas (Gyulai 2007). In Romania, large-scale recovery at ‘flat-extended’ Măgura-Buduiasca yielded remains of a range of cereals and pulses (Walker and Bogaard 2011). Crop diversity at these north Balkan sites, however, is less than in the southern Balkans and Greece, as typically ‘Mediterranean’ pulses (grass pea, chickpea) are absent in the earlier Neolithic. The formation of ‘tells’ in the north Balkan LN/Chalcolithic coincides with a major increase in available data, including burnt house assemblages of diverse crop ‘stores’ (Gyulai 2007), confirming that the low density of plant remains on ‘flat’ sites is a function of preservation.
Pollen analyses (Willis and Bennett 1994) and on-site charcoal (Ntinou and Badal 2000; Marinova and Thiébault 2008) suggest very limited clearance of woodland. Arable weed assemblages from the southern Balkans, especially rich in Bulgaria (Marinova 2006), suggest permanent cultivation plots: few woodland taxa but many of disturbed habitats imply plots established for 5–10 years at least (cf. Bogaard 2002). Ecological analysis further suggests autumn sowing, excluding spring sowing of floodplains. Bulgarian assemblages contain the mixture of ‘root-/row-crop weeds’ and ‘cereal weeds’ characteristic of small-scale and intensive cultivation (Jones et al. 1999). Sheep/goat dung implies herding near settlements, compatible with small-scale animal husbandry. In the northern Balkans, potential arable weeds at EN Ecsegfalva 23 and Măgura-Buduiasca also suggest long-established, intensively managed plots, which at the former site could be accommodated within areas of dry ground above seasonal floods (Bogaard et al. 2007).
Animal exploitation in south-east Europe
Despite variation in preservation and recovery, faunal assemblages exhibit some recurrent trends, especially when small samples (less than 400 identified specimens) are excluded. Domesticates usually make up more than 95% of the mammals on Neolithic open sites in Greece (Cantuel et al. 2008) and similarly dominate EN assemblages from Anza in FYROM (Bökönyi 1976), through Starčevo (following Legge 1990) and Divostin (Craig et al. 2005) in Serbia, to Ecsegfalva 23 in the Hungarian plain (Bartosiewicz 2007a). In central Greece, evidence for hunting, especially of large game (red deer, boar), increases modestly in the LN (fifth to fourth millennium BC) (von den Driesch 1987) and sharply (to 10–50%) on some Bronze Age open sites. A more rapid increase occurs in the LN (fifth to fourth millennium BC) north Balkans: at Vinča (Dimitrijevič 2008), Selevac (Legge 1990), and Polgár-Csöszhalom (Bartosiewicz 2005). Both large (especially red deer and boar) and small mammals are represented in late Mesolithic levels at Franchthi cave in southern Greece and sites such as Lepenski Vir in the Iron Gates gorge between Serbia and Romania. The contrasting scarcity of EN evidence for hunting recurs from the relatively arid and lightly wooded south of Greece, through the better-watered and more densely wooded valleys of Serbia, to the seasonally inundated Hungarian plain, and so, excluding the Greek islands, is unlikely to reflect availability of game. Interpreting this apparent avoidance of hunting (Bartosiewicz 2005, 60; 2007a, 298–299) in terms of the ‘domestic’ symbolic concerns of colonist farmers (Hodder 1990) is favoured by EN avoidance of antler for tools or ornaments and the contrasting LN mortuary
deposition of ornaments made from boar and red deer teeth in Hungary (Bartosiewicz 2005, 58). Small game, however, is not avoided (von den Driesch 1987): EN sites in Greece (Cantuel et al. 2008) and Anza in FYROM (Bökönyi 1976) have yielded mammals such as hare, fox, cat, marten, and roe deer and wetland sites on the Hungarian plain also an impressive diversity of birds and fish (e.g. Gál 2007; Bartosiewicz 2007b). Similarly, comparison of worked and unworked bone at LN (sixth to fifth millennium BC) Makriyalos in northern Greece reveals that domesticates and small game were selected, and large game avoided, as raw material for artefacts (Isaakidou 2003). Large game, therefore, normally subject to greater obligations of sharing than small game or domesticates (cf. Barnard and Woodburn 1991), might have been hunted by early farmers, but consumed (collectively?) away from excavated open settlements.
Sheep are the predominant domesticate, especially in the earlier Neolithic, at open sites in Greece and FYROM and on the Hungarian plain (Bökönyi 1976; Bartosiewicz 2007a; Cantuel et al. 2008; Halstead 1996), although cattle are equally frequent at Divostin in Serbia (Craig et al. 2005). In the LN, sheep give way to cattle or pigs, most rapidly and sharply in the north Balkans (Bartosiewicz 2005; Dimitrijevič 2008; Greenfield 1999; Legge 1990). Cantuel et al. (2008, 287) attribute the EN dominance of sheep (grassland animals) in a more or less wooded landscape to early farmers’ lack of expertise, and Whittle and Bartosiewicz (2007, 741) to the cultural conservatism of colonists. Bökönyi (1973, 168) argued that livestock reproduced too slowly for early farmers to adjust herd composition to local environments, but goats and especially pigs are more prolific than sheep and could rapidly have outnumbered them if farmers wished. The dominance of sheep would be unsurprising, however, if livestock were few and often confined to cleared plots (stubble, fallow, field margins, sprouting cereals) rather than being numerous and ranging widely across the landscape (Halstead 2006). Several lines of evidence are consistent with initially small-scale animal husbandry. First, pollen and charcoal fail to register the impact of early farmers on vegetation. Second, biometric distinction between large wild and smaller domestic cattle and pigs becomes increasingly clear through the Neolithic (von den Driesch 1987; Legge 1990), implying limited interbreeding (as do DNA and aDNA) between domestic and wild populations. The difficulty of isolating modern free-range pigs from wild boar suggests early domesticates were few enough to be herded closely or corralled. Third, dental microwear in sheep and goats from EN Ecsegfalva 23 in Hungary (Mainland 2007) and LN Makriyalos in northern Greece (Mainland and Halstead 2005) reveals a very abrasive diet, implying restriction to heavily overgrazed or freshly cultivated pasture. Early livestock in south-east Europe, therefore, were probably few in number and often enclosed on cleared land—an anthropogenic niche ideal for sheep. Conversely, increasing proportions of cattle and/or pigs in the LN may reflect larger numbers of livestock exploiting the landscape more extensively and, in Greece, possibly leaving their imprint in the palynological and geoarchaeological records (Willis 1994; van Andel et al. 1990).
Sheep exhibit a ‘meat’ culling strategy (slaughter of juvenile–sub-adult males, retention of adult females) from Greece (Halstead 1987, 1996; Helmer 2000; Isaakidou 2006) to the north Balkans (Bökönyi 1971, 650; Greenfield 2005; Legge 1990; Bartosiewicz 2007a, 300; Dimitrijevič 2008). Data are sparser for goats and cattle, but ‘meat’ mortality is evident for both at EN–FN Knossos on Crete (Isaakidou 2006) and for cattle at EN Blagotin (Greenfield 2005) and LN Selevac (Legge 1990) and Vinča in Serbia (Dimitrijevič 2008); exceptions (e.g. EN Ecsegfalva 23—Bartosiewicz 2007a) may be due to small sample size. A ‘meat’ strategy does not preclude modest exploitation for secondary products, however, and organic residues in ceramics indicate milking at least in the sixth millennium BC at LN Stavroupoli in northern Greece (Evershed et al. 2008) and EN Ecsegfalva 23 and Schela Cladovei in the north Balkans (Craig et al. 2005). Likewise, at Neolithic Knossos, numerous ‘pathological’ traces in adult cows suggest use for traction, albeit on a smaller scale than is possible with oxen (Isaakidou 2006, 2008). Similar traces are reported in smaller numbers elsewhere in Neolithic south-east Europe (e.g. Poduri, Romania—Balasescu et al. 2006).
‘Meat’ mortality precludes specialized dairying, however, and this, coupled with initially small-scale animal husbandry, means that the dietary staples were grain crops, although animal produce doubtless improved the nutritional balance, security, and variety of the food supply and was perhaps doubly important because of the social contexts in which it was consumed (see ‘Synthesis: subsistence and society in Neolithic southern Europe’). Livestock were also probably integral to early crop husbandry: manure of animals confined on arable land would have contributed to soil fertility, sheep perhaps controlled the resulting risk of ‘lodging’ (stem collapse) by light grazing of sprouting cereals, whilst cows pulling an ard (scratch-plough) or a sledge loaded with stall-manure could have enabled intensive cultivation on a larger scale. Ploughing also aids timely sowing and so reduces the risk of crop failure, particularly in southern Greece where severe summer drought places sowing under acute time stress (Isaakidou 2008). This underlines the importance of early draught cattle at Knossos.
The open settlements in fertile lowlands that dominate the Neolithic record of south-east Europe were probably occupied year-round (e.g. Bartosiewicz 2007a, 2007b; Gál 2007; Halstead 2005), but early farmers must also have ranged regularly beyond their homes and gardens for medicinal plants, raw materials, pasture or game, and social contacts. Such forays are perhaps reflected in the abundance of game (much higher than on the nearby Hungarian plain) at EN open sites in the Iron Gates gorge (Bartosiewicz 2007a). Stable isotopic analysis of human remains here and at Theopetra cave in central Greece suggests a ‘Neolithic’ diet (Bonsall et al. 2004; Papathanasiou 2003), perhaps reflecting links to nearby settlements with greater arable potential. There is no evidence that EN farmers regularly moved long distances or established distant ‘satellite’ sites in the context of seasonal herding or hunting.
In southern Greece, sparse EN settlement in fertile valleys expanded in the LN to areas less favourable to grain crops because of low rainfall or poor soils. Alongside established ‘villages’, small, short-lived sites proliferated and the use of caves increased dramatically. The agriculturally marginal location of many new sites has been interpreted in terms of seasonally mobile pastoralism (e.g. Sampson 1992) and traces of dung indicate penning of animals in Kitsos (Brochier et al. 1992, 48) and Kouveleiki A (Karkanas 2006) caves. At the small open site of Kefala and caves of Kalythies, Kastria, Skoteini, and Zas, goats are more frequent and cattle and pigs less so than at contemporary villages (Halstead 1996, 31, fig. 2), consistent with herding on a scale large enough to require adjustment to local landscape. ‘Meat’ mortality for sheep and goats again precludes specialized dairying (Halstead 1996), however, whilst isotopic and pathological evidence from human skeletons implies a crop-based diet at both inland (Kouveleiki, Skoteini) and coastal (Alepotrypa) caves and the Kefala hamlet (Papathanasiou 2003). Some caves have yielded remains of cereals and pulses, although it is not clear whether crops were grown locally, whilst others were used for burial; increasing evidence for use of caves perhaps reflects changes in social practices as much as subsistence routines. LN–FN marginal colonization by small open sites and perhaps caves thus seemingly replicated the mixed farming of earlier Neolithic villages. Any expansion in the scale of herding apparently did not weaken dietary dependence on grain crops, although more frequent crop failures in marginal areas probably made livestock more important as an emergency food source.
SOUTH-WEST EUROPE
Plant use and husbandry in south-west Europe
Archaeobotanical evidence from Dalmatia has been limited by a research focus on caves in karstic terrain unsuitable for agriculture (Moore et al. 2007a). Crops are therefore lacking even for periods when cultivation is beyond doubt (Forenbaher and Miracle 2005). However, recent large-scale sampling has recovered EN cereals (barl
ey, emmer, einkorn), pulses (lentils, grass pea), flax, and wild fruits from the open settlement of Prokovnik (Moore et al. 2007b) and a similar spectrum from the MN settlement of Danilo (Moore et al. 2007a).
Data from Italy derive mostly from open-air sites, but two caves are noteworthy. At Uzzo in Sicily (Costantini 1989), Mesolithic levels yielded sparse wild legumes, fruits and nuts (strawberry tree, acorn, grape), whilst the earliest Neolithic yielded a range of cereals (einkorn, emmer, barley) and pulses (lentil, grass pea, or vetchling), as well as olive and figs. In north-west Italy, EN remains from Arene Candide include a range of cereal types (Binder and Maggi 2001). The rarity of cereal chaff and weed seeds from central-western Mediterranean cave sites arguably reflects processing at habitation sites elsewhere (Zapata et al. 2004; Peña-Chocarro 2007).
Sparse data from open-air sites in central and southern Italy suggest cultivation of multiple cereals (barley, einkorn, emmer, free-threshing wheat) and pulses (pea, lentil, vetches, broad bean) (Rottoli and Pessina 2007). Waterlogged preservation at sixth millennium BC La Marmotta confirms a similarly broad spectrum, alongside oil-seed crops (flax, opium poppy) and a range of wild plants. Grape remains may suggest viticulture. Together with evidence from Iberia (see below, this section), abundant poppy remains from La Marmotta indicate cultivation of this species within its natural central-west Mediterranean distribution area by the mid-sixth millennium BC. Opium poppy currently provides the clearest botanical case of local domestication in Neolithic Europe.