Farmers at the Frontier: A Pan European Perspective on Neolithisation

By Peter Rowley-Conwy

All farming in prehistoric Europe ultimately came from elsewhere in one way or another, unlike the growing numbers of primary centers of domestication and agricultural origins worldwide. This fact affects every aspect of our understanding of the start of farming on the continent because it means that ultimately, domesticated plants and animals came from somewhere else, and from someone else. In an area as vast as Europe, the process by which food production becomes the predominant subsistence strategy is of course highly variable, but in a sense the outcome is the same, and has the potential for addressing more large-scale questions regarding agricultural origins. Therefore, a detailed understanding of all aspects of farming in its absolute earliest form in various regions of Europe can potentially provide a new perspective on the mechanisms by which this monumental change comes to human societies and regions. In this volume, we aim to collect various perspectives regarding the earliest farming from across Europe. Methodological approaches, archaeological cultures, and geographic locations in Europe are variable, but all papers engage with the simple question: What was the earliest farming like? This volume opens a conversation about agriculture just after the transition in order to address the role incoming people, technologies, and adaptations have in secondary adoptions.


The book starts with an introduction by the editors which will serve to contextualize the theme of the volume. The broad arguments concerning the process of neolithisation are addressed, and the rationale for the volume discussed. Contributions are ordered geographically and chronologically, given the progression of the Neolithic across Europe. The editors conclude the volume with a short commentary paper regarding the theme of the volume.

• Language: English
• Publisher: Oxbow Books
• Release date: Feb 15, 2020
• ISBN: 9781789251418

Book preview

Introduction

Agricultural origins: where next?

Kurt J. Gron, Lasse Sørensen and Peter Rowley-Conwy

The chapters in this book examine the earliest farming across Europe – for each region, how farming spread, what exactly it was that spread and the form that the earliest farming took. For too long the search has been on for external or internal factor ‘causing’ the transition to agriculture in Europe. This perspective has ignored the human factor: individuals were engaging with unfamiliar environments, unfamiliar people, unfamiliar plants and animals, and unfamiliar technologies; but they had to eke out an existence using traditions, techniques and abilities not developed for the situation in which they found themselves.

This volume is an attempt to understand agricultural origins from the other side, the perspective that by placing primacy on the earliest farming we may better understand the individual factors, processes and actors involved in Neolithisation. How things ultimately resolved themselves may prove key for understanding this long-researched topic, and this book is meant to take the first step in that direction.

Early agriculture in Europe

The primary centres of agricultural origins now number, depending on one’s definition of a centre, in the high single-digits or low double-digits (Current Anthropology 52(S4) 2011). These are places where people took plants and animals into a close and integrated relationship. This integration of several plant and animal species into one integrated system under human control is perhaps the best definition of what agriculture is. Under this definition, Europe was not a primary centre of agricultural origins, because no integrated system arose that was based on native resources.

Mesolithic hunter-gatherers in Europe were not passive actors in their landscapes. They engaged in active niche construction, manipulating various species in ways that they considered beneficial to themselves (see e.g. Rowley-Conwy and Layton, 2011). But the idea of farming as a system was imported into prehistoric Europe. This fact affects every aspect of our understanding of the start of farming on the Continent because it means that, ultimately, domesticated plants and animals came from somewhere else and from someone else. In an area as vast as Europe, the process by which food production became the predominant subsistence strategy was, of course, highly variable, but in a sense the outcome was the same everywhere. Therefore, a detailed understanding of all aspects of farming in its absolute earliest form in various regions of Europe can potentially provide a new perspective on the mechanisms by which this monumental change comes to human societies and regions.

The European case is unique in that the aggregate evidence is substantial, and relatively well studied. It is also unique because the various factors that almost certainly influenced the spread of agriculture – climate, environment, biogeography, cultural setting etc. – varied widely across the Continent. As such, the ways that agriculture and its practitioners adapted to each individual setting demonstrates both the remarkable flexibility of farming systems, and also how farming is affected by outside influences. Europe is the perfect laboratory for understanding the practicalities and realities of farming’s flexibility.

In the primary centres of agricultural origins the participants almost certainly had no idea they were embarking on a path towards an integrated agricultural system. Therefore, at the most basic level, proposed explanations for agricultural origins in these centres can be distilled down to just exogenous, natural factors, and endogenous, cultural factors (Price and Bar-Yosef, 2011). Basically:

•something changes in the environment that permits or compels a pre-existing relationship with wild species to become a commitment;

•human population outgrows the environment’s ability to consistently provide enough food and this permits or compels a pre-existing relationship with wild species to become a genotypic commitment;

•changes in the social structure, make-up, organisation or dynamics of human societies permit or compel a pre-existing relationship with wild species to become a genotypic commitment.

In reality, causality is almost certainly a combination of these.

Secondary adoption, i.e. the adoption of agriculture outside the primary centres, is different for two fundamental reasons: 1) the potential for immigration, and 2) the possibility for choice on the part of indigenous peoples in whether or not to commit to agriculture. The explanatory frameworks proposed for the primary centres may, of course, still apply, but the key differentiating factor in secondary adoption is that the actors were aware they were bringing farming to a previously unfarmed region, or, in the case of indigenous peoples, were aware of the potential change to their traditional lifeways that agriculture was bringing.

‘A Zone of Variability’: lengthening the timeline

At a 2009 Wenner-Gren symposium attended by many of the key players researching the origins of agriculture worldwide, a remarkable consensus was reached that there is a ‘Zone of Variability’, or period during which people in the primary centres seem to negotiate, explore and audition the new way of life represented by agriculture (Price and Bar-Yosef, 2011; and see Weiss et al., 2006). In this discussion, any consensus is remarkable, and more remarkable still are the implications for the situational awareness of those starting to farm. A decade later, it is still a relevant line of inquiry.

Conventional chronologies define the start of the European Neolithic as the moment domesticated plants and animals first arrived in a landscape. However, it is starting to become clear that the transitional ‘Zone of Variability’ is also to be found in secondary contexts such as Europe. There are various lines of evidence in support of this. In Scandinavia, for example, there is a period of overlap between the last foragers and the first farmers with a duration of several centuries (Gron and Sørensen, 2018). In Portugal, a similar situation is observed, also lasting a few hundred years (Bicho et al., 2017). The same may be true in Britain, with the last foragers surviving for several hundred years into the Neolithic (Gron et al., 2018). Perhaps Europe’s best example is in the Iron Gates, where complex patterns of genetic mixing occurred (González-Fortes et al., 2017; Bonsall and Boroneanţ, 2018), which can only be described as a cultural, economic and genetic negotiation between foraging and farming.

The start of the Neolithic is, however, not the end of the Neolithisation process. Conventional timelines of secondary agricultural origins are simply too short. The earliest phase of the Neolithic, the Early Neolithic (or EN), is a unique and consistent aspect of the chronology of secondary agricultural transitions and is rightfully considered part of the transition. The EN starts at the moment domestic plants and animals arrive in a region, the point of time conventional chronologies record the end of the Mesolithic. Its subsequent duration, however, is context, culture, environment and other factor specific. Despite its end being much harder to define archaeologically, there are several tell-tale lines of evidence that indicate that the EN may be nearing completion:

A) widespread anthropogenic alteration of the landscape;

B) the final abandonment of foraging sites of a Mesolithic character;

C) the commencement of monumental or communal construction.

It takes time for unfarmed landscapes to be transformed into farming landscapes (see Redman and Foster, 2008), and this period of transition should be considered part of the Neolithisation process. This is because the economic and environmental conditions, and consequently the cultural setting, of the ‘mature’ or ‘full’ Neolithic are not yet present with the arrival of domestic plants and animals. A series of EN farms in small forest clearings would have a different adaptive subsistence strategy, and also different social relationships with each other and with any surviving indigenous communities, than agricultural settlements in later periods in a cleared agricultural landscape. This is almost certainly why the building of monumental structures is sometimes absent in the EN (e.g. Gron and Sørensen, 2018; Roberts et al., 2018), despite being characteristic of the cultures and regions from which the farmers almost certainly came (Sørensen, 2014). Until the landscape is agrarian to a considerable extent within the limits of given ecosystems and Neolithic cultural systems, it remains transitional, and therefore still within the process of Neolithisation.

The arrival of farming in a previously unfarmed landscape causes a sea change in local human-environment interactions, both for the last foragers and for the incoming farmers. Environments are fundamentally changed by the presence of farmers, and the character of farming is fundamentally influenced by the local environment (Geertz, 1963). For any remaining foragers, resource availability, mobility and all associated social processes will be disrupted. Farming’s impact on the local environment will be immediate but take some time to become apparent, and is related to the inverse relationship between increasing population and decreasing forest cover (Mather and Needle, 2000). Once widespread anthropogenic alteration of the natural landscape occurs, indigenous foraging patterns will likely have been almost completely disrupted, and the transition to agriculture complete.

The final abandonment of Mesolithic-type foraging sites is another indication that Neolithisation may be complete. A complete shift to farming indicates that the indigenous population A) is no longer present, B) has adopted farming or C) has been assimilated or otherwise subsumed into the farming population. The transition to agriculture is at that point complete.

The commencement of monument construction is another signal that Neolithisation is complete. In Britain and southern Scandinavia, for example, monument construction occurs not at the start of the EN, but lags several centuries behind (Gron and Sørensen, 2018; Roberts et al., 2018). This strongly suggests that, during the EN, A) there were not enough farmers, B) there was a lack of community organisational structures to organise construction, or C) there was a lack of surplus resources and time to build them. Once these criteria are met, construction commences, and is therefore indicative of a permanent, organised and resourced agricultural community.

These are by no means the only indicators of the completion of the Neolithisation process, but may present a starting point for understanding the unique processes occurring in the EN. The interplay between the last foragers, the first farmers, the environment and the demographic changes occurring in the new, burgeoning agrarian landscape will inevitably be complex.

A new approach

The above discussion provides a concrete roadmap for inquiry into the ‘Zone of Variability’, or ‘Negotiation Phase’ (Gron and Sørensen, 2018) of Neolithisation. This middle-range approach connects foragers and farmers, while accommodating continental-scale variability in environment, culture, setting and numerous other factors. Ultimately, by recognising that the Neolithisation process involves both Mesolithic and Neolithic peoples, and probably a lot of confusion on the part of those involved, we can move this debate forward.

Acknowledgements

The editors gratefully acknowledge the ‘Northern Worlds’ project of the National Museum of Denmark and funded by the Augustinus Foundation for supporting the colour images seen in this volume. We would also like to thank all those who refereed the chapters in this volume.

References

Bicho, N., Cascalheira, J., Gonçalves, C., Umbelino, C., Rivero, D.G. and André, L. (2017) Resilience, replacement and acculturation in the Mesolithic/Neolithic transition: the case of Muge, central Portugal. Quaternary International 446, 31–42.

Bonsall, C. and Boroneanţ, A. (2018) The Iron Gates Mesolithic – a brief review of recent developments. L’anthropologie 122, 264–280.

Geertz, C. (1963) Agricultural Involution: The Process of Ecological Change in Indonesia. Berkeley, University of California Press.

González-Fortes, G., Jones, E.R., Lightfoot, E., Bonsall, C., Lazar, C., Grandal-d’Anglade, A., Garralda, M.D., Drak, L., Siska, V., Simalcsik, A., Boroneanţ, A., Romaní, J.R.V., Rodríguez, M.V., Arias, P., Pinhasi, R., Manica, A. and Hofreiter, M. (2017) Paleogenomica evidence for multi-generational mixing between Neolithic farmers and Mesolithic hunter-gatherers in the Lower Danube basin. Current Biology 27, 1801–1810.

Gron, K.J., Rowley-Conwy, P.A., Fernandez-Dominguez, E., Gröcke, D.R., Montgomery, J., Nowell, G. and Patterson, W.P. (2018) A meeting in the forest: hunters and farmers at the Coneybury ‘Anomaly’, Wiltshire. Proceedings of the Prehistoric Society 84, 114–144.

Gron, K.J. and Sørensen, L. (2018) Cultural and economic negotiation: a new perspective on the Neolithic transition of southern Scandinavia. Antiquity 92(364), 958–974.

Mather, A.S. and Needle, C.L. (2000) The relationships of population and forest trends. The Geographical Journal 166(1), 2–13.

Price, T.D. and Bar-Yosef, O. (2011) The origins of agriculture: new data, new ideas: an introduction to Supplement 4. Current Anthropology 52(S4), S163–S174.

Redman, C.L. and Foster, D.R. (eds) (2008) Agrarian Landscapes in Transition: Comparisons of Long-Term Ecological and Cultural Change. New York, Oxford University Press.

Roberts, D., Valdez-Tullett, A., Marshall, P., Last, J., Oswalt, A., Barclay, A., Bishop, B., Dunbar, E., Forward, A., Law, M., Linford, N., Linford, P., López-Dóriga, L., Manning, A., Payne, A., Pelling, R., Powell, A., Reimer, P., Russell, M., Small, F., Soutar, S., Vallender, J., Winter, E. and Worley, F. (2018) Recent investigations at two long barrows and reflections on their context in the Stonehenge World Heritage Site and environs. Internet Archaeology 47, DOI: https://doi.org/10.11141/ia.47.7.

Rowley-Conwy, P. and Layton, R. (2011) Foraging and farming as niche construction: stable and unstable adaptations. Philosophical Transactions of the Royal Society B 366, 849–862.

Sørensen, L. (2014) From hunter to farmer in northern Europe: migration and adaptation during the Neolithic and Bronze Age. Acta Archaeologica 85, 1–794.

Weiss, E., Kislev, M.E. and Hartmann, A. (2006) Autonomous cultivation before domestication.

Science 312, 1608–1610.

Chapter 1

Growing societies: an ecological perspective on the spread of crop cultivation and animal herding in Europe

Maria Ivanova

In a good year we eat the wheat and the livestock eats the barley, in a bad year we eat the barley and the livestock. (Farmer’s testimony, Jordan [Pasternak, 1991])

Introduction: an ecology of life

Most early farmers across Europe were descendants of colonists, as were their crops and livestock (Zohary and Hopf, 2000; Larson et al., 2007; Olalde et al., 2015; Scheu et al., 2015; Hofmanová et al., 2016; Mathieson et al., 2018; Mittnik et al., 2018). The south-west Asian ancestry of domestic cereals and pulses, cattle, sheep, goats and pigs was not a trivial detail in the process of their translocation to different latitudes and climates in Europe. Some traits of these organisms were modulated by skill and technology to match the new environments, and in some cases this resulted in genetic shifts in the crop and livestock populations. Until recently, archaeologists could at most speculate about the roles of management practices and genetic change in the dispersal of farming. In the last two decades, however, rapid advances in isotopic and biomolecular archaeology have provided powerful new tools to infer the growth conditions of past crops (Bogaard et al., 2013; Araus et al., 2014), feeding, birth seasonality, movement and dairying in prehistoric herds (Balasse and Tresset, 2007; Evershed, 2008; Bogaard and Outram, 2013; Balasse et al., 2017; Miller and Makarewicz, 2017), human dietary preferences and mobility patterns (Bentley, 2013; Schulting, 2018), and the genetic ancestry, phenotypes and functional genetic traits of past domestic organisms and humans (Brown et al., 2009; Larson and Burger, 2013; Allaby et al., 2015; Brown et al., 2015; Orlando et al., 2015; MacHugh et al., 2017; Daly et al., 2018). This previously unimaginable level of detail regarding the ‘biological’ dimension of the past now urges archaeologists to reconsider the notion of human agency as detached from, and acting upon, an external domain of the non-human environment, and ultimately the dichotomy of culture and nature itself.

Table 1.1. Cultural groups and their absolute dating (abbreviations: EN = Early Neolithic, LN = Late Neolithic, EC = Early Chalcolithic, LBK = Linearbandkeramik, TRB = Trichterbecherkultur).

The allusion to Tim Ingold’s essay ‘Growing plants and raising animals: an anthropological perspective on domestication’ (1996) in the title of this chapter is not accidental. The present study draws on this and two other of Ingold’s seminal texts, The Perception of the Environment (2001) and more recently his introduction to Biosocial Becomings (2013), to come up with a more satisfactory way of describing human activity within the environment. The notion of humans inhabiting an encultured social domain outside and above the world of other living organisms, and of the latter being merely a source of raw materials, Ingold argues, is a distinctly modern western view on things. People whose livelihoods actually depend on growing plants and animals, and whose involvement with the environment is immediate and pragmatic, have a different view on their relationship to other living organisms – as coming to terms with rather than control and exploitation. By replacing the culture-nature dualism with the ‘synergy of organism and environment’, such relational thinking enables a genuine understanding of humans within the system of their ecological relationships. ‘Both humans and the animals and plants on which they depend for a livelihood’, Ingold asserts, ‘must be regarded as fellow participants in the same world, a world that is at once social and natural’ (Ingold, 1996, 22). The present chapter examines how the earliest crops were grown, how livestock was raised and how the first farmers fashioned their livelihoods at different latitudes in Europe from the perspective of an ‘ecology of life’. Because of the south–north delay in the arrival of farming, the absolute dates for the period of interest are regionally variable, ranging from 6700–5500 cal BC in the Mediterranean to 4000–3000 cal BC in north-western Europe (Table 1.1).

The ecology of crop and livestock translocations

In the temperate zone, solar radiation and rainfall fluctuate over the year and living organisms need to prepare in advance for these seasonal changes. Many plants and animals use day length (or photoperiod), which changes regularly through the year, to synchronise their life cycles with the turn of seasons (Bronson, 1988). Photoperiod responsive organisms need a certain duration of exposure to light in a 24-hour period to initiate different stages in their development, and most importantly to enter the reproductive phase (transition from vegetative stage to flowering and setting seeds in plants, or to sexual receptivity and fertility in animals). The critical photoperiod shifts with latitude (Fig. 1.1), and the flowering/ breeding season occurs later in the year as the distance from the equator increases (Lee, 1970; Bronson, 1988). Since photoperiod sensitivity is genetically regulated, problems may occur when plants and animals are brought to latitudes where their photoperiodic triggers are mismatched with the local conditions (Bradshaw and Holzapfel, 2010). In a classical example of a failed translocation, breeding ibex stock from Austria (with a similar latitude), Turkey and Sinai (more southern latitudes) were introduced to Czechoslovakia to replace an extinct ibex population. The resultant north–south hybrids rutted in September/October instead of December/January (as the native ibex did), gave birth in midwinter instead of late spring and the population died out completely (Greig, 1979). In addition to the photoperiod response, plants exhibit genetically controlled responses to thermoperiodism, described in so-called ‘cardinal temperatures’ – the base (above which development begins and accelerates with temperature), optimum (with maximum growth rate) and ceiling (at which growth stops) temperature. Progressing to the next stage in a plant’s life cycle requires the accumulation of a certain amount of ‘growing degree days’ (GDD or °D) with temperatures between the lower and upper threshold – the higher the average daily temperatures, the more rapidly a plant develops (Kelleher, 2003). Plants from warm climates, needing high total °D, may be unable to reach maturity after translocation to higher latitudes.

Fig. 1.1. Variation in day length (sunrise to sunset) at different latitudes in the Northern Hemisphere (after Bradshaw and Holzapfel, 2010).

The implications of photo- and thermoperiodism for the spread of crop cultivation and herding from south-west Asia into Europe are obvious. In the semi-arid Mediterranean climate of south-west Asia, genetically determined photoperiodic triggers allow cereals, pulses, sheep and goats to prepare for seasonal change and complete their cycles of reproduction before the summer droughts: wild barley (Hordeum spontaneum) flowers in early March (Lister et al., 2009), and modern wild sheep (Ovis orientalis gmelini) give birth between February and March (Valdez, 2008). With increasing latitude, winter cold grows in importance as the major seasonal constraint, the growing season shifts forward into summer and the Mediterranean photoperiodic response becomes progressively inadequate (Lister et al., 2009). After eight millennia of crop cultivation and livestock breeding at middle and high latitudes in Europe, many of the seasonal controls have been weakened or eliminated. However, the first translocations to higher latitudes, with various cycles of day length and seasonality, must have been quite challenging.

Growing plants – sustaining soil fertility

The first south-west Asian farmers cultivated thirteen major annual crops – the hulled emmer and einkorn wheats, the free-threshing bread and durum (macaroni) wheats, hulled and free-threshing barley, field pea, lentil, chickpea, grass pea, broad bean, bitter vetch and flax (Zohary and Hopf, 2000). This diverse spectrum of crops dispersed without major changes throughout the Mediterranean from the Levant to Portugal, with the exceptions of the chickpea, which was not transferred beyond Greece, and opium poppies, which were added in the central Mediterranean (Salavert, 2017; Salavert et al., 2018). The hulled wheats, einkorn and emmer, predominate in the earliest archaeobotanical assemblages. In the western Mediterranean, these were displaced by free-threshing varieties around 5300 cal BC (Rottoli and Castiglioni, 2009; Antolín, 2016; Perez-Jorda et al., 2017).

With the spread to northern latitudes, however, crop diversity considerably declined (Colledge et al., 2005; Kreuz et al., 2005; Colledge and Conolly, 2007; Conolly et al., 2008; Coward et al., 2008; Kreuz and Marinova, 2017). A major shift took place in the Balkans, where farmers successfully grew all cereals substantially outside of their native ranges, but drastically reduced the variety of cultivated pulses (Kreuz et al., 2005; Ivanova et al., 2018). The range of crops grown regularly by the first ‘Linearbandkeramik’ (LBK) farmers of central Europe was still narrower, including einkorn, emmer, lentils and flax (Salavert, 2011; Kreuz, 2012). Barley occurs in much lower proportions and at fewer sites than wheat; in the archaeobotanical assemblages of the north-west LBK its grains are so rare that barley is sometimes regarded as a weed rather than crop (Salavert, 2011). The reduced crop spectrum in central Europe did not necessarily focus on species that are particularly suitable for cultivation at higher latitudes. The infrequent occurrence of bread wheat and pea in the earliest archaeobotanical assemblages, for example, is very surprising. Bread wheat is the highest yielding of all cereals, it is frost-resistant (grows today at latitudes of up to 68°N) and, having a long growing period, is best suited to take advantage of climates with summer rains (Kirleis and Fischer, 2014), while peas have been observed to develop and produce better yields under central European climatic conditions compared to lentils (Kreuz and Marinova, 2017). Yet, in central Europe it took several centuries for peas to replace lentils as the main pulse crop (Kreuz and Marinova, 2017) and bread wheat did not become established before the Rössen period in the mid-fifth millennium cal BC (Bakels, 2009; Kirleis and Fischer, 2014).

The first farmers in north-west Europe cultivated an even narrower range of crops. Free-threshing durum wheat, with barley or emmer as a second crop, were the only widely grown plants in the west Baltic region and Sweden during the early fourth millennium cal BC. Although durum wheat was the predominant cereal in central Europe at this time (Kreuz et al., 2014), its northward spread as a major crop is nevertheless surprising, since winter-sown durum does not tolerate freezing temperatures. Indeed, several centuries later, in the second half of the fourth millennium BC, durum cultivation in the western Baltic area declined and it was replaced by emmer and especially by free-threshing barley (Kirleis and Fischer, 2014). Emmer was the main crop of the first farmers in the North Atlantic, who also infrequently grew free-threshing wheat and barley; in the second half of the fourth millennium BC barley became a prominent crop here as well (Jones and Rowley-Conwy, 2007; McClatchie et al., 2014). In Scotland and the Northern Isles barley, more often free-threshing than hulled, was preferentially grown from the very start of farming, around 3500 cal BC (Bishop et al., 2009). This chronological trend hints that the transformation of free-threshing barley from a minor crop to a major staple in north-west Europe around the middle of the fourth millennium may have been triggered by the emergence of a new strain, which was better suited for cultivation in northern latitudes. If this was the case, the new strain was most likely a photoperiod non-responding type with delayed flowering time, capable of escaping late frosts and taking advantage of summer rains and a longer growing season in higher latitudes. Indeed, non-responsive genotypes have been shown to predominate today in barley landraces from central and northern Europe, while responsive varieties are mainly grown in south-west Asia, southern Europe and the Mediterranean basin (Cockram et al., 2007; Lister et al., 2009; Jones et al., 2011; 2012). In the absence of direct genetic evidence from the period of interest, however, the appearance of photoperiod non-responsive barley genotypes in the fourth millennium BC remains speculation. Throughout north-west Europe, the only pulse crop is peas (Pisum sativum), occurring sporadically as single grains (Kirleis et al., 2012).

Weeds associated with crop remains indicate that domestic plants were cultivated in ‘garden habitats’, created through improvement and maintenance of soil fertility in long-established plots. Such intensive forms of cultivation of cereals and pulses have been inferred from weed assemblages for the Mediterranean (Bogaard, 2005; Bogaard and Halstead, 2015; Antolín, 2016), the Pannonian Basin and central Europe (Bogaard, 2004; 2005, Bogaard et al., 2007; 2011; 2013), and for northern Europe (Kirleis and Fischer, 2014; McClatchie et al., 2014). The most extreme examples are possibly the ‘cultivated middens’ at early farming sites on Orkney, built up by farmers from ash-rich midden material and organic additives in order to pursue plant cultivation in small intensively-managed plots, growing in material accumulated from generation to generation (Dockrill and Bond, 2009). Intensive manuring is reflected in the stable nitrogen isotope values of charred cereal and legume grains from various parts of Europe, including the Mediterranean and sub-Mediterranean areas (Araus et al., 2007; Bogaard et al., 2013; Vaiglova et al., 2014), central Europe (Bogaard et al., 2013; Fraser et al., 2013), and northern Europe (Bogaard et al., 2013; Gron et al., 2017; Sjögren, 2017). Stable nitrogen isotope values of cereal grains from Vaihingen, for instance, are comparable with those of plants grown on plots manured for over a century receiving 20 to 35 t/ha of cattle manure every two years (Fraser et al., 2013).

Raising animals – sustaining prolificacy

Herd composition and growth

Large compilations of zooarchaeological data have compellingly demonstrated a pattern in the composition of the earliest herds across Europe: the flocks of the first Mediterranean farmers comprised mainly sheep/goats, while cattle predominated in central and north-west Europe (Manning et al., 2013a). Cattle herding took off in the Balkans, where migrant farmers first crossed the border of the Mediterranean and settled beyond the native climatic range of their domestic plants and animals (Manning et al., 2013b; Ivanova et al., 2018). Pigs remained a minority in domestic herds in most parts of the continent, though their proportions notably increased when farming became established in the warmer and more humid Atlantic zone, where they replaced sheep as the second important herd animal (Arbogast and Jeunesse, 2013).

Exchange of breeding stock among early herders has been inferred in several cases from the isotope compositions of strontium and oxygen in tooth samples (Knipper, 2012; Sjögren and Price, 2013; Gron et al., 2016). For instance, imported animals were common among cattle and sheep from sites in Falbygden, central Sweden. Cattle ‘immigration’ in this region appears to have been higher than human mobility, and the wide range of Sr values suggests the involvement of several interconnected areas. Since the Sr values of migrant humans and cattle overlapped only partially, some of the animals likely came to Falbygden with human settlers (e.g. through marriage exchanges), while others moved by various other mechanisms such as tributes, trade or gift exchanges (Sjögren and Price, 2013). The maintenance of genetic variability in cattle herds, and to a lesser degree in sheep herds, thus rested on far-flung circulation of breeding stock. Restocking with wild animals would have been possible for cattle, but genetic evidence indicates that female aurochs were rarely integrated into the domestic herds (Edwards et al., 2007; Achilli et al., 2008; 2009; Bonfiglio et al., 2012). The sequencing of an ancient aurochs genome has recently confirmed hybridisation and a particularly strong aurochs genetic contribution in modern domestic cattle in Britain and Ireland (Park et al., 2015). However, since aurochs and cattle coexisted in Britain for several millennia after the introduction of farming (Lynch et al., 2008), genomic data from early domestic cattle will be necessary to resolve the timing of the gene flow.

In contrast to sheep and cattle, the Sr isotope signatures of pigs obtained by several studies consistently showed locally raised animals (Bentley and Knipper, 2005; Knipper, 2012; Sjögren and Price, 2013). Contrary to cattle, cross-breeding of pigs with indigenous wild boar was ubiquitous in early herds across south-west Asia and Europe. Pigs were domesticated in south-east Anatolia and spread during the seventh millennium BC westwards into western Anatolia, where local wild females were readily incorporated into the first herds and rapidly displaced the introduced maternal lineages (Ottoni et al., 2013). The same rapid process of substitution reoccurred in Europe, where cross-breeding of introduced Anatolian pigs with indigenous wild sows began soon after the appearance of the first domestic herds in central Europe in the later sixth millennium, and by the early fourth millennium virtually all domestic pigs in Europe belonged to indigenous maternal lineages (Ottoni et al., 2013; Caliebe et al., 2017). Since the establishment of domestic herds in northern Europe post-dates the replacement, the continuation of cross-breeding in this region is not directly reflected in the genetic evidence. However, modelling of genetic data suggests that intensive interbreeding continued at higher latitudes in Europe (Caliebe et al., 2017). The widespread practice of restocking with indigenous sows made the exchange of breeding pigs obsolete, which might explain the low mobility of pigs. The sheer ubiquity of cross-breeding domestic pigs and wild sows signals that the practice was possibly not directed, but was rather the unintended outcome of commensalism. Attracted by waste, wild sows likely frequented the early farmers’ settlements, resulting in opportunistic mating with domestic males. The crossbred female offspring, carrying domestic traits but better adapted to local conditions through their wild ancestry, might have been preferentially treated by the farmers and encouraged to join their herds. It is striking that these early herders managed to maintain domestic behaviour in their pigs despite constant gene flow from wild boar. Recent studies of modern suid genomes suggest that farmers intensely selected pigs for domestic behaviour and morphology, sustaining veritable ‘islands of domestication’ within their genomes (Frantz et al., 2015).

At temperate latitudes animal breeding follows the seasonal changes of day length, temperature, rainfall and grazing. Sheep and goats are seasonal breeders whose period of sexual activity is synchronised by photoperiod (Rosa and Bryant, 2002). The exact timing of births is crucial for both animals and herders: when young are born too early, the lactating mother might suffer malnutrition with severe consequences for the body size and fitness of her offspring, while young that are born too late will not have time to attain optimal body weight before entering their first winter (Noddle, 1990; McCormick, 1998). Lambing in the wild progenitors of sheep in semi-arid southwest Asia is adjusted to coincide with abundant graze early in the year. Modern wild sheep (Ovis orientalis gmelini) in these regions have a short lambing period of one to two months in spring. At middle latitudes in Europe, oxygen isotope (δ¹⁸O) sequences in sheep teeth show that births in the Neolithic occurred over a longer period of three to four months, but the February onset of lambing was retained (Balasse et al., 2017). In comparison, native ungulates in these latitudes rut in December/January and give birth in May (Greig, 1979). Thus, the late winter start of the birth season in central European sheep, inherited from their wild ancestors, was precarious but the risk may have been offset to a certain extent by their extended lambing season. A shift to early summer lambing was observed only at the northernmost sites (latitude 59°N), showing a rapid genetic adjustment of photoperiodic response to the harsh winters of the Scottish Northern Isles (Balasse et al., 2017).

Because fertility cycles in cattle depend on food quality rather than photoperiod (Dahl et al., 2000; Balasse and Tresset, 2007), latitudinal adjustment in the timing of births presents fewer difficulties for cattle compared to sheep and goats. Natural selection would favour a calving peak coinciding with the local spring flush of grass, giving a clear advantage to cattle over sheep and goats for herders settling in higher latitudes. Calving in free-ranging herds raised without feeding supplements today is seasonal: for example, in free-ranging Grey cattle in Thrace and cattle of the Highland breed in Germany most births occur in March and April, and in free-range cattle in northern France between May and July (Berthon et al., 2018). Isotope evidence from south-eastern, central and northern Europe points at a restricted calving season of early domestic cows within two to three months (Balasse and Tresset, 2007; Balasse et al., 2013; Berthon et al., 2018), which closely corresponds to that of modern freerange cattle. The presence of several out-of-season births in the data-set supports the impression that herders did not employ strong control on the reproduction of their cattle (Berthon et al., 2018).

Feeding the herds

The average diet of an animal in the last years of its life can be inferred from the stable carbon (δ¹³C) and nitrogen (δ¹⁵N) isotopes in bone collagen samples. A considerable body of stable isotope data has accumulated in archaeology, providing a basis for comparisons of livestock diets across latitudes. The data-set included in the present study comprises 644 published paired δ¹³C and δ¹⁵N measurements on bone collagen of wild and domestic animals from 58 early farming sites (Table 1.2 and Fig. 1.2), excluding data with atomic C:N values outside the acceptable range of 2.9–3.6 (DeNiro, 1985).

The stable carbon isotope composition in bone collagen depends on the isotopic δ¹³C variability in graze and fodder. These change with environmental parameters such as temperature, humidity and light intensity, influencing the photosynthetic process (O’Leary, 1981). The change of these parameters along a north–south gradient across Europe is reflected in a decline in δ¹³C values of plants and their consumers with increasing latitude. This northward decline in δ¹³C, observed in previous studies (van Klinken et al., 2000; Goude and Fontugne, 2016) and also evident in our data-set, makes clear that climate effects overwrite dietary trends and preclude direct comparisons of δ¹³C values from different latitudes in Europe. To minimise the climate effect, the δ¹³C values of bone collagen in the present study were climate-corrected following van Klinken et al. (2000) (values for modern average July temperature (°C) were extracted from the WorldClim – Global climate data gridded data source, http://worldclim.org for each site based on its geographic coordinates). The remaining variation in δ¹³C is expected to reflect mainly differences in grazing and fodder. Light conditions in closed forests, for example, produce a gradient in leaf 13C values from canopy to forest bottom, with the most negative values near the ground (the canopy effect) (van der Merwe and Medina, 1991). As a consequence, δ¹³C values in grazing animals decrease with increasing forest density. Differences in the physiology of the consumed plants, as well as in water availability and salinity, also can give rise to isotopic variation in animals. Herbivore consumption of plants following the C4 photosynthetic pathway leads to higher δ¹³C values compared to those feeding on C3 plants (Farquhar et al., 1989). Similarly, grazing on plants in water-rich areas such as freshwater marshes may lead to lower δ¹³C (Lynch et al., 2008), while salt-marsh grazing may result in higher δ¹³C (Couto et al., 2013).

Table 1.2. List of sites with published δ¹³C and δ¹⁵N values and results of organic residue analysis of pottery. Regions: A = Mediterranean, B = north Balkan and Pannonian Basin, C = central Europe, D = northern Europe (including the Channel Islands, west Baltic and north-east Atlantic).

Fig. 1.2. Distribution of sites with stable isotopic data and results of organic residue analysis (sites included in Table 1.2).

The corrected data (Fig. 1.3) show no correlation of δ¹³C to latitude (wild animals R²=0.005, domestic R²=0.002). There are statistically significant differences in the δ¹³Cc values of sheep/goats, cattle and pigs (Kruskal-Wallis H=12.18, p=0.002). Post-hoc tests (Dunn-Bonferroni tests) reveal that pigs and cattle are not significantly different, though both diverge from sheep/goats (cattle z=2.969, pig z=−2.962, p=0.009). Generally, δ¹³C values of bone collagen across Europe indicate that domestic herds grazed preferably in open habitats. Even in the forested zones of central and northern Europe, most values lie above −22.5‰, considered as the cut-off for herbivores feeding on closed-canopy plants in these areas (Berthon et al., 2018). There are, however, also exceptions such as Vaihingen in south-west Germany and Bischoffsheim in Alsace, where the majority of the analysed cattle fed on forest vegetation (Bickle et al., 2013; Fraser et al., 2013). It is therefore possible that the low chronological resolution of the available isotope evidence disguises temporal trends related to the impact of human activity on the landscape around the first farming villages, a question that deserves further attention. Moreover, because δ¹³C values of bulk bone collagen reflect the average year-round diet, animals browsing in the forest (or receiving forest fodder) on a seasonal basis do not show lowered δ¹³C values and the seasonal forest contribution may be underestimated. Such seasonal variation in feeding can be detected by sequential sampling and analysis of stable carbon isotopes in teeth. However, forest grazing/fodder was observed only in a minority of the sampled cattle teeth from south-east and central Europe (Balasse et al., 2013; Berthon et al., 2018), suggesting that forest contribution to diet was infrequent even on a seasonal scale. Other types of winter fodder have been identified only occasionally, for example the sporadic feeding of sheep with summer-grown hay at the Romanian site of Magura (Balasse et al., 2013). Remarkably, cold-season seaweed consumption was detected in all ten sampled sheep teeth from Holm of Papa Westray (beginning of the third millennium BC) on Orkney, Scotland, confirming the systematic character of this practice at the site (Balasse et al., 2006; 2009).

Fig. 1.3. Distribution of δ¹³Cc and δ¹⁵N according to latitude (°N) and group of species.

Stable nitrogen isotope signatures in domestic and wild animals across Europe correlate very weakly with latitude (wild R²=0.032, domestic R²=0.017), showing opposing trends – δ¹⁵N values in wild animals slightly decline, while those in livestock increase with increasing latitude. The difference between the mean δ¹⁵N values of wild and domestic animals is statistically significant (Mann-Whitney U=46.632, p=0.000), and mean δ¹⁵N is 1.1‰ higher in livestock. These observations indicate that the δ¹⁵N values of the latter have been affected by human activity. Nitrogen isotope variation in plants and their consumers is related to soil δ¹⁵N content, which is generally lower in forest soils, and in forest animals respectively. Livestock, to the contrary, most likely grazed in anthropogenic landscapes that were modified by soil disturbance, burning and manuring, all of which have been shown to increase dietary δ¹⁵N (van Klinken et al., 2000), without affecting stable carbon isotope values. The differences in δ¹⁵N among domestic species are not statistically significant (Kruskal-Wallis H=5.14, p=0.077).

Studies of stable oxygen isotopes have attested to seasonal change of pastures and movement to different vegetation zones and altitudes for some of the earliest caprine herds in south-west Asia (Makarewicz, 2017). In Europe, however, the little available evidence points to predominantly stationary herding, with occasional cases of transhumance. Only one of the sampled cattle from Vaihingen, for example, was moved to an upland pasture in summer and subsequently returned to the settlement (Knipper, 2012). Analysis of strontium isotopes in sequential teeth samples of four cattle from Falbygden in western central Sweden also did not detect seasonal movement, at least not outside the geological province where they were born (Sjögren and Price, 2013).

Milking goats, sheep and cows

Slaughter profiles of sheep, goats and cattle, reconstructed from faunal assemblages, suggest that most early farming communities in Europe raised their animals with mixed aims in mind. Considerable variation has been observed in the Mediterranean, where age profiles from open-air sites tend towards preferential culling of ‘prime meat’ mature animals (Halstead, 1996; Bogaard and Halstead, 2015), but the high proportion of young age classes of sheep in caves and rock shelters possibly points at dairying (Debono Spiteri et al., 2016). In the Balkans, the overwhelming number of adults among goats, in contrast to sheep and cattle, at some sites has been interpreted as an indication for dedicated goat milking (Greenfield and Arnold, 2015). The slaughter profiles of cattle and caprines at early farming sites in central Europe appear surprisingly uniform and indicate mixed meat and milk use, with preference for meat in some areas (Gillis et al., 2017). Because of small sample sizes and poor preservation of the earliest bones of domestic animals from northern Europe, reliable published slaughter profiles for this region are not available at present (Gron and Rowley-Conwy, 2017).

Lipid residues absorbed in the matrix of ceramic vessels provide a complementary view on milk use among early European farmers. Low proportions of dairy residues are characteristic for open-air sites in the eastern Mediterranean (Pitter et al., 2013; Nieuwenhuyse et al., 2015; Ethier et al., 2017; Whelton et al., 2017) (Fig. 1.4A). Conversely, most of the residues retrieved from rock shelters and caves in the Adriatic, southern France and the Iberian Peninsula were of dairy origin, which has been interpreted in terms of increasing importance of dairying with the westward spread of farming along the Mediterranean coasts (Šoberl et al., 2008; Debono Spiteri et al., 2016). Alternatively, this strong dairying signal in caves and rock shelters may represent an artefact of their special function (Ethier et al., 2017), if caves were seasonally used to shelter animals during lambing and lactation (Bogaard and Halstead, 2015). Indeed, at some cave sites faunal evidence such as bones of newborn lambs and shed deciduous teeth point at use in late winter and early spring for this purpose (Rowley-Conwy, 2000; Forenbaher and Miracle, 2005; Helmer et al., 2005; Bogaard and Halstead, 2015). More lipid data from open-air sites in the central and western Mediterranean are needed to resolve this question.

Just beyond the northern borders of the Mediterranean, a marked dairying signal has been recently identified in the residue record from the northern Balkans (Fig. 1.4B) (Ethier et al., 2017). Further north-west, the proportion of dairying residues declines (Fig. 1.4C), and the distribution of commodities in central Europe is not significantly different from that in the Mediterranean (Mann-Whitney U=9779.5, p=0.513). Finally, the spread of farmers and their herds into the North Atlantic and Scandinavia was again associated with a very strong dairying signal in pottery (Fig. 1.4D). Thus, current evidence suggests a series of shifts rather than a gradual latitudinal increase in the importance of dairying. Dairying may have served to buffer risk when farmers crossed major biogeographic borders – from the coastal Mediterranean into the interior of the continent, and later into the colder north Atlantic and boreal zones of northern Europe.

In a bad year we eat the barley … and the fish

People are closely linked to their immediate surroundings through the food they consume and these linkages are reflected in the stable isotope composition of their tissues. To compare humans across Europe from this perspective, the present study uses a data-set comprised of published paired δ¹³C and δ¹⁵N measurements on bone and tooth collagen of 1100 individuals from 97 sites (Table 1.2 and Fig. 1.3). Samples with C:N values outside the acceptable ranges were not considered, and infants below the age of six were excluded to avoid the effect of nursing on observed values. The measured δ¹³C values were climate-corrected as described above for δ¹³C in animal collagen samples.

Table 1.3. Summary statistics for published faunal and human δ¹³C and δ¹⁵N values.

The combined δ¹³Cc and δ¹⁵N values of early farmers across Europe demonstrate an overwhelmingly terrestrial diet with few outliers even at coastal locations, as has been observed in many previous studies (Richards et al., 2003b; Fischer et al., 2007; Papathanasiou, 2011; Schulting, 2011; Lelli et al., 2012; Guiry et al., 2017; Salazar-García et al., 2017; 2018; Schulting, 2018). The farmers’ δ¹³Cc range is obviously narrower in comparison to the animals they consumed (standard deviation 1.06 for wild, 1.07 for domestic animals and 0.67 for humans) (Table 1.3, Figs 1.4 and 1.5). This possibly points to a high fraction of dietary protein obtained from plant rather than animal sources, i.e. from crops that grew in much more uniform habitats than the grazing areas of livestock.

In contrast to the animal isotope values, there is a slight positive correlation of both human δ¹³Cc and δ¹⁵N with latitude (δ¹³Cc R²=0.24, δ¹⁵N R²=0.12). The latitudinal trend is mainly due to the elevated ¹³C and ¹⁵N values from sites above 57°N, and especially from the Scottish Northern Isles (Figs 1.3 and 1.5). This peculiar situation at the northernmost margins of the early farming world has been the subject of intensive discussions. Since domestic animals, and in particular sheep, at Northern Isles sites also show elevated δ¹³C values, consumption of meat and milk from ¹³C enriched animals has been suggested as an explanation for the values in humans (Jones and Mulville, 2015). However, the high values of mean Δ¹³C and Δ¹⁵N (human-herbivore difference) at some sites indicate that such ‘baseline’ enrichment may be an insufficient explanation, and that occasional contribution of marine foods to human diet should be considered (Gigleux et al., 2017). Indeed, the combined analysis of collagen from bulk bone samples and from incremental teeth samples of humans buried at Sumburgh on Shetland has provided clear evidence for consumption of marine foods, albeit only occasionally and most likely as famine food (Montgomery et al., 2013).

Fig. 1.4. Δ¹³C values for archaeological animal fat residues from: (A) Mediterranean, 19 sites, open circles – cave/rock shelters; (B) north Balkan and Pannonian Basin, 7 sites; (C) central Europe, 10 sites; (D) northern Europe (Channel Islands, west Baltic and north-east Atlantic), 36 sites, open circles – west Baltic, black circles – north Atlantic.

The dietary contribution of other potential ‘wild’ foods, such as meat of hunted terrestrial animals or plant seeds and nuts, cannot be estimated from the stable isotope compositions of bone collagen. Judging from the faunal assemblages, however, significant differences existed between the Mediterranean and other parts of Europe. While in Mediterranean assemblages the bones of domestic animals are almost invariably predominant with >90% (Halstead, 1996; Vigne, 2000), sites beyond the Mediterranean yield various proportions of wild species, and are occasionally dominated by hunted animals (Rowley-Conwy, 2011; Manning et al., 2013b; Rowley-Conwy, 2013; Ivanova et al., 2018). Wild plants are widely attested to at early farming sites, but their dietary contribution is difficult to quantify (Colledge and Conolly, 2014). Several tubers, nuts and seeds of weedy plants, which are known from the historical records as supplementary foods (and as staples when harvests failed), are infrequently found in caches at early farming sites. Among them are typical ‘unsown crops’ growing in fallow fields, e.g. fat hen (Chenopodium album) and rye brome (Bromus secalinus L.) (Knörzer, 1967; Kreuz, 2007; Behre, 2008; Bakels, 2014), as well as hazelnuts (Robinson, 2000; Jones and Rowley-Conwy, 2007; Kirleis et al., 2012) and in higher latitudes starchy root tubers such as lesser celandine (Ficaria verna) (Kirleis et al., 2012; Klooss et al., 2016).

Fig. 1.5. Mean and standard deviation of δ¹³C and δ¹⁵N values of humans and domestic animals (cattle, caprine and pig).

The Balkans as an ‘experimental field’

The earliest successful introduction of farming beyond its Mediterranean homeland took place in the interior of the Balkans during the early sixth millennium BC. Within a few centuries, pioneer farmers spread northwards from 40° and 48°N into a range of unfamiliar environments. Most critically, they crossed the divide in precipitation regimes running through the interior of the peninsula, with sub-Mediterranean conditions characterised by summer rainfall minima to the south and continental conditions with summer rainfall maxima to the north of this border (Ivanova et al., 2018). In consequence, winter and spring frosts replaced summer droughts as the main limiting factor for successful agro-pastoralism during the course of this dispersal. The Balkans served as a unique ‘experimental field’ for the adaptation of south-west-Asian farming to higher latitudes in Europe. Over the past five years, a wide range of new bioarchaeological evidence has been acquired from the region to enhance our understanding of this key episode in the dispersal of farming in Europe (Balasse et al., 2013; Bogaard et al., 2013; Vaiglova et al., 2014; Balasse et al., 2017; Ethier et al., 2017; Jovanovic, 2017; Whelton et al., 2017; Ivanova et al., 2018; Cramp et al., 2019). Indeed, the Balkans provide an excellent example of how pioneer farmers’ livelihoods changed at (micro) regional scales.

The taxonomic composition of faunal and botanical assemblages show that distinct zones of farming and wild-resource procurement developed within the Balkans, with sites from similar biogeographic areas clustering closely irrespective of their cultural affiliation. In the transitional environments of the southern Balkans, where both summer drought and winter/spring frost might affect the growth of crops, a strategy of diversification was adopted with a wide range of cultivated species at the majority of sites. Across the precipitation border, however, the crop spectrum was significantly narrowed with the decline of several warmth-loving and drought-resistant pulse taxa (Ivanova et al., 2018). Herd compositions also changed beyond this major climatic border, possibly in response to issues of sustaining livestock prolificacy. Sheep and goats, which did not have wild relatives in the Balkans, likely experienced feeding problems in forested environments with long-lasting snow cover. Importantly, with increasing latitude and a later-onset spring, the genetically controlled photoperiod response of Mediterranean sheep and goats became increasingly out of place. Existing stable oxygen isotope evidence does not demonstrate adjustments in the timing of caprine births to higher latitudes in south-east Europe (Balasse et al., 2017), and it is not surprising that a shift from caprines to cattle herding occurred in the interior of the Balkans. Mediterranean cattle probably also suffered acclimation difficulties when first brought into the interior. However, cattle benefited from feeding behaviours and physiology similar to their local wild relative, the aurochs, and their reproduction was not constrained by photoperiod.

Pioneer herders in the interior of the Balkans were confronted with the