Introduction

The arrival of agriculture and livestock to the Iberian Peninsula promoted changes in social and economic structures (Rojo Guerra et al. 2012). The new species and products were linked to a dramatic change in economy that, unlike hunting and gathering, required long-term planning and investment in the production of food, with deferred results (Vicent 1990, 1998) and gave rise to a new type of people; farmers. These new agricultural economies needed specific requirements to be able to thrive (e.g. political control over territory, agrarian capital, storage of goods, long-term management of resources). It was mastering these conditions that led to the origins of the first complex societies. The current paradigm, accepted by most scholars (see Bernabeu 2003; Delibes de Castro and Fernández Manzano 2000; Díaz del Rio 2006) is that, such socio-economic political differences began to become visible in the interior and North of the Iberian Peninsula during Chalcolithic (c. 3300–1900 BCE) (Delibes de Castro and Fernández Miranda 1993; Delibes de Castro et al. 1995). One of the consequences of the new situation was the emergence of regional differences that are observed within the archaeological record (Bernabeu 2003). For example, the size and type of settlements changed during the Chalcolithic along the Iberian Peninsula (c. 3000–2200 BCE). While small and open settlements dominated, the northern and eastern regions of the peninsula, fortified settlements appeared in the western and southern regions (Garcia Sanjuán and Murillo-Barroso 2013; Rojo Guerra et al. 2005). During the Final Chalcolithic and Early Bronze Age (c. 2200–1900 BCE), there was a diversity and richness of the Bell Beaker settlements and burial assemblages in Iberia (Garrido-Pena 2014). Later, during the Middle and Final Bronze Age (c. 1900–1000 BCE) in the North and centre of the peninsula, economy and society seem to have stabilised without significant changes, compared to the previous stages (Díaz del Rio 2001, 2006). Nevertheless, little is currently known about the diet and economy of these past populations.

To explore the economy and lifeways of the Prehistoric inhabitants of Central Northern Spain, this research is focused on sites located from the NE of the Castilian Plateau to the Cantabrian coast (Fig. 1) between c. 3000 and 1500 BCE. To date the subsistence strategies of the past populations inhabiting this region are poorly known. From the Northern plateau, Chalcolithic human remains are rare (Carmona Ballestero et al. 2013) and faunal assemblages recorded from such sites are relatively small (Riquelme Cantal 2009; Marín-Arroyo 2011). The situation is similar in the Cantabrian coast, in the northernmost part of the study region, with few zooarchaeological studies existing from this time period (Altuna 1980; Altuna and Mariezkurrena 2012; Altuna and Mariezkurrena 2011; Castaños 1984; Mariezkurrena 1990; Ontañón Peredo 2003; Peréz Ripoll and López Gila, 2012). Recent excavations at Arroyal I (Carmona Ballestero and Arnaiz Alonso in press), Fuente Celada (Alameda Cuenca-Romero et al. 2011; Carmona Ballestero 2013) and El Hornazo (Carmona Ballestero 2013; Carmona Ballestero et al. 2013), all located in the province of Burgos, and El Abrigo de la Castañera (Sierra Saínz-Aja 2014) in Cantabria have revealed assemblages of human skeletal remains from mortuary contexts, in addition to associated mammal remains. Dietary stable isotope analysis of these human and animal elements presents a valuable opportunity to understand the economic behaviour of these early farmers, during a period of emerging social complexity.

Fig. 1
figure 1

Map of the northern Spain showing the location of the sites studied from the Castilian Plateau and the Cantabrian Region

Full size image

Stable isotope analysis has commonly been used in archaeology since the mid-1980s as a technique for reconstructing palaeodietary behaviour, providing insights into human nutrition and economic strategies (Schoeninger et al. 1983; DeNiro 1985; Schoeninger and DeNiro 1984; Chisholm et al. 1982). Stable isotope analysis of bone collagen is a reflection of diet consumed by an individual over the last 10–15 years of life, informing on long-term average diet. This method is especially valuable in determining C3 and C4 plant consumption (Tieszen 1991), the use of animal protein, and the inclusion of marine foods in diet (Schulting and Richards 2002). Although there have been numerous applications of this technique in Prehistoric Europe (e.g. Richards et al. 1998; Schulting et al. 2010; Oelze et al. 2011; Tafuri et al. 2009), it is only in recent years that this subject has received attention in Spain, with studies attempting to redress this balance in the Mediterranean Coast (Salazar-García et al. 2014) and the Basque Country in the north of Iberia (Fernández-Crespo and Schulting 2017; Fernández-Crespo et al. 2018; Fontanals-Coll et al. 2015; Sarasketa-Gartzia et al. 2018).

Stable isotope analysis of animal bone collagen related to human remains can not only provide a valuable baseline for human diet but also provide valuable insights into animal management practises (e.g. Jones and Mulville 2016, 2018; Stevens et al. 2013; Müldner et al. 2014). In this study, δ13C and δ15N analysis of animal bone collagen will also be used to explore the use of the landscape by these early farming populations.

Materials and methods

The sites included in this comparative study are from Central Northern Spain, spanning the NE area of the Old Castilian Plateau to the Cantabrian coast (Fig. 1). Within this wider geographical region, the sites derive from two environmental zones. Four of the sites (Arroyal I, El Hornazo, Fuente Celada and Ferrocarril-La Dehesa) are located in the NE area of the Spanish North Plateau, an inland region with a cool and dry continental Mediterranean climate. This area is characterised by long and cold winters and short and warm summers and a strong contrast between day and night temperatures. In contrast, El Abrigo de la Castañera is located in the Cantabrian region which has an Atlantic climate and lies close to the coast. The Cantabrian Region today is characterised by year-round rainfall and limited seasonal temperature variation (Peel et al. 2007). The distinctly different environmental conditions of the two regions mean that there is potential for different models of dietary behaviour and animal management. A summary of the samples analysed from each site is presented in Table 1.

Table 1 Summary of the human and animal species sampled for δ13C and δ15N analysis from each site
Full size table

The sites

Arroyal I (Alfoz de Quintanadueñas, Burgos, Castile and León, Spain)

The site of Arroyal I (Fig. 1) was excavated between 2011 and 2012 by a research team from the University of Burgos. The site is a megalithic tomb comprising of a rectangular chamber (3 × 3.5 m), a long corridor (6 m) and a stone mound. The tomb was used as a collective burial location over a period of 400 years during the Late Neolithic (c. 3300–2900 BCE, Table 2), before it was abandoned (Carmona Ballestero and Arnaiz Alonso in press). During the Chalcolithic, the tomb came back into use when it was extensively remodelled. At that time, Neolithic layers were almost entirely removed; the corridor was filled with rocks and sediment (as a closure event) and a stone wall and a floor of limestone blocks were built inside the chamber. Following this, several consecutive and isolated burials were interred. One of these contained a young individual, dated to c. 2465–2211 BCE (Table 2), that was buried with a set of four vessels (two bell beakers bowls and two carinated ones), surrounded by human long bones and skulls from previous contemporaneous burials. In the next phase of activity, the dolmen was closed re-using materials from the site (in a secondary position) and the external mound height was increased. Finally, an isolated pit grave containing a female human body was created inside the mound, which dateds to c. 2348–2200 BCE (Table 2).

Table 2 Radiometric dates available from Arroyal I, El Hornazo Fuente Celada in Burgos and Abrigo de la Castañera in Cantabria
Full size table

El Hornazo (Villímar, Burgos, Castile and León, Spain)

El Hornazo is a ‘field of pits’ located in Villímar, Burgos (Fig. 1). In 2004, a research team excavated an area of almost 1 ha where 179 different structures were discovered, including pits, ditch pits, pit-graves and post holes (Carmona Ballestero 2013). This site represents a Chalcolithic settlement divided into three different activity areas: habitation, storage and resource processing. In the storage area, two pit-graves were documented. The first grave (Pit 103) included a human subadult, buried with a bowl, infilled with sediment and few domestic remains that is dated to c. 2576–2473 BCE. In the second pit (Pit 140), dating to c. 2860–2574 BCE, another subadult was found buried in a crouched position. Deposits within Pit 140 contained domestic remains including pottery and animal bones. A single biconical bead was registered as a grave-good inside Pit 140. AMS dates of the human remains belong to the Pre-Beaker Chalcolithic (Table 2). In conjunction with the radiocarbon dates, the stylistic and typological attributes of archaeological remains determined that the entire aggregate of negative structures were formed during a single phase of occupation.

Fuente Celada (Alfoz de Quintadueñas, Burgos, Castile and León, Spain)

Fuente Celada, located in Quintanadueñas (Burgos), was excavated in 2008 by the same research team of El Hornazo. It is a settlement with many rock-cut pits (‘silos’) that was occupied initially during the Neolithic and before the main focus of occupation occurred in the Early Chalcolithic (Alameda Cuenca-Romero et al. 2011; Carmona Ballestero 2013). In addition to a more domestic area (containing ‘silos’, post holes, mud fragments used in the huts, stone mortars and other domestic remains), human remains were discovered inside several pits, dating to the Chalcolithic between c. 2900 and 1900 BCE. Pit 19 contained a young woman, buried in an unusual inverted position (almost vertically placed, with the head at the bottom and the feet at the top). This pit also contained the largest accumulation of potsherds found at the site (1018 fragments, 19.5% of the total) dated to c. 2860–2574 BCE. In Pit 5, the disarticulated human remains of three individuals (a subadult and two male adults) were recorded, also belonging to the Chalcolithic (Table 2).

Ferrocarril-La Dehesa (Alfoz de Quintadueñas, Burgos, Castile and León, Spain)

Ferrocarril-La Dehesa is also located in Quintanadueñas (Fig. 1). The site was partially excavated in 1981 by opening a small sondage, before an open excavation area was conducted by L. Villanueva Martín and M. E. Delgado Arceo in 2016. In total, 147 contexts or activity areas were registered within this settlement: most of them simple rock-cut pits (‘silos’) but with a few dwellings, kilns and a pit-grave. In all cases, the features were filled with deposits containing domestic remains. The site is dated by relative methods with the Cogotas I type pottery, either in its formative phase (Protocogotas) or advanced phase one (Cogotas I Pleno) to the Middle Bronze Age (c. 1750–1450 BCE). The tomb (Pit 700) included an adult individual, deposited in a foetal position without any grave goods.

El Abrigo de la Castañera (Obregón, Cantabria, Spain)

El Abrigo de La Castañera is situated in the south of Santander Bay, in Cantabria (Fig. 1). This is the only site in this study that is located near the Atlantic Sea, which has a distinctive climatic and environmental context in comparison to the sites located in the province of Burgos. The site is an 85-m2 deep rock shelter, that is a part of a seven-cave karst complex, oriented towards the NE. The site was surveyed in the late 1960s (Gomarín Guirado 1972) and partially excavated during the 1970s (Rincón Vila 1982, 1985). Later, it was subsequently studied by local archaeologists (Ruiz Cobo and Serna González, 1999; Ruiz Cobo 1991, 1996). Since 2011, a multidisciplinary research team led by C. Vega-Maeso has been excavating the shelter and surveying its surroundings.

The rock shelter has evidence of domestic use, including living areas, and an animal corral, during the Neolithic and Chalcolithic. A well-preserved Bronze Age stratigraphy linked to funerary events (c. 2100–1500 BCE, Table 2) was also recorded within the shelter (Vega-Maeso 2017). During recent excavations, human remains from at least six individuals were recovered associated with faunal remains (including domestic and wild species), lithic and bone industry, shells and pottery (Sierra Sainz-Aja 2014).

Sample preparation

Samples were prepared using facilities in the Institute of Biomedicine and Biotechnology at the University of Cantabria in Spain. Collagen extraction was undertaken following procedures outlined in Richards and Hedges (1999). Bone fragments between 0.6 and 0.8 g were cleaned by abrasion to remove any possible surface contamination before demineralisation in 0.5 M HCL at 6–8 °C for between 3 and 10 days. Samples were then washed using de-ionised water. Samples were gelatinised in a weak acidic solution (pH 3 HCL) at 70 °C for 48 h, then filtered with 5–8 μm Ezee® filters, prior to freeze-drying. Samples were analysed for δ13C, δ15N and two samples were additionally analysed for δ34S (samples SUC28 and SUC19) using mass spectrometry at Iso-Analytical (Crewe, UK).

The δ13C values and δ15N values are reported relative to the international standards V-PDB and AIR standards. In-house standards were used to calculate analytical error which was ± 0.1‰ (1σ) or better. All specimens discussed further had % C values above 35%, % N values above 15% and C:N values between 2.9 and 3.4 indicative of in vivo collagen (De Niro 1985; van Klinken 1999). The δ34S values are reported relative to the international standard VCDT. Quality indicators of C:S values between 600 ± 300 and N:S values between 200 ± 100 (Nehlich and Richards 2009). All sample information, raw data and quality indicators for each sample are included in the Supplementary Material.

Results and interpretation

The results obtained are presented alongside initial interpretation about human diet and animal management at each archaeological site. The subsequent discussion is focused on the wider implications of these results in terms of the Prehistoric economies and lifeways in Central Northern Spain.

Arroyal I (Alfoz de Quintanadueñas, Burgos, Castile and León, Spain)

The human stable isotope values from Arroyal I (n = 9) form a main cluster of individuals dated to the Late Neolithic, Chalcolithic and Beaker period, with δ13C values ranging between − 19.3‰ and − 19.7‰, and δ15N values ranging between 9.2 and 10.9‰ (Fig. 2; Fig. 3). One individual (SUC28) plotted away from this main cluster with a lower δ13C value of − 20.8‰, while the δ15N value of 10.3‰ is similar to the δ15N values observed in the main cluster. The humans have δ15N values 3–4‰ higher than the herbivores sampled from the site, which is a typical relationship between consumers and prey during the Holocene (Hedges and Reynard 2007) and suggests that the humans were eating a C3 diet that included animal protein (meat and potentially milk).

Fig. 2
figure 2

Human and animal bone collagen δ15N and δ13C values analysed from Arroyal I

Full size image
Fig. 3
figure 3

Human and animal bone collagen δ15N and δ13C values analysed from El Hornazo. The specimen labelled as ‘a’ is discussed in more detail in the text

Full size image

Individual SUC28 appears to have consumed a different diet to the other humans sampled based on the δ13C values observed. To explore the relationship of this individual and the main cluster, δ34S analysis was undertaken. Collagen from SUC28 and SUC19 was analysed to explore whether these individuals were derived from the same place. SUC28 yielded a δ34S value of 14.8‰ and SUC19 had a δ34S value of 10.2‰ (Table 3) indicating that these individuals derived from regions that were isotopically different in sulphur, likely representing geographically distinct locations.

Table 3 Human and animal δ15N, δ13C, and δ34S values from each of the sites analysed in this study
Full size table

El Hornazo (Villímar, Burgos, Castile and León, Spain)

The Chalcolithic humans from El Hornazo (n = 2), have δ13C values between − 19.2 and − 19.7‰ and δ15N values of 8.6‰ and 9.5‰ (Fig. 3; Table 3), indicative of terrestrial diet of animal protein in addition to plant products. One of the four dogs sampled has values indistinguishable from the humans analysed (δ13C − 19.1‰, δ15N 8.2‰). The other dog specimens plotted have δ15N values ranging between 6.3‰ and 6.4‰ and δ13C values between − 20.2‰ and − 20.7‰, demonstrating differing dietary behaviour of dogs at the site. The cattle have δ15N values ranging between 5.3‰ and 8.3‰, with δ13C values lying between − 19.3 and − 20.8‰ (Fig. 3). Two of the ovicaprids and one cow are plotted with the human specimens exhibiting higher δ13C and δ15N values. Cow ‘a’ (Fig. 3) had a lower δ13C value of − 19.3‰.

Fuente Celada (Alfoz de Quintadueñas, Burgos, Castile and León, Spain)

The Fuente Celada humans (n = 3), dating to the Chalcolithic, had δ13C values ranging between − 19.3 and − 19.5‰ and δ15N‰ values ranging between 8.3 and 9.2‰ (Fig. 4; Table 3). The tight range in values observed suggests that these individuals consumed a very similar diet, including terrestrial animal protein.

Fig. 4
figure 4

Human and animal bone collagen δ15N and δ13C values analysed from Fuente Celada plotted with the human specimen available from Ferrocarril-La Dehesa

Full size image

The faunal data show a wide dispersal of stable isotope values. Whilst the sheep and ovicaprids had a relatively small range in δ13C values of between − 20.1‰ and − 20.7‰, the δ15N values were much greater, between 4.3‰ and 6.8‰. The larger range in δ15N values could indicate a greater diversity in the diet of these animals, potentially representing different pasturing locations or differences in winter foddering within the population analysed.

The cattle from the site have a wider range in δ13C values between − 20.4‰ and − 21.3‰ and a smaller δ15N range than the ovicaprids, between 4.8‰ and 5.4‰. The wider range of δ13C values could also be indicative of differing management practises of individuals, such as the use of a variety of pastures or foddering animals on differing plants.

Ferrocarril-La Dehesa (Alfoz de Quintadueñas, Burgos, Castile and León, Spain)

Only one human specimen was available to study from this site and attributed to the Middle Bronze Age. This individual is plotted next to the Chalcolithic specimens of Fuente Celada, located in the same village (Fig. 4). The individual from Ferrocarril-La Dehesa has a δ13C value of − 19.6‰ and δ15N value of 9.1‰ implying a diet rich in animal protein as seen in neighbouring Fuente Celada.

El Abrigo de la Castañera (Obregón, Cantabria, Spain)

The humans analysed from El Abrigo de la Castañera (n = 3) belong to the Early Bronze Age and have δ13C values ranging between − 21.3‰ and − 20.8‰ and δ15N values ranging between 9.2 and 10.1‰ (Table 3; Fig. 5). This is indicative of a diet comprising of animal protein, mirroring the pattern seen at El Hornazo, Fuente Celada, Arroyal I and the individual from Ferrocarril-La Dehesa.

Fig. 5
figure 5

Human and animal bone collagen δ15N and δ13C values analysed from El Abrigo de la Castañera

Full size image

All humans and animals have lower δ13C values than seen at the other sites (Figs. 5 and 6). This difference could be due to the location of La Castañera, which is in Cantabria, whereas the other sites are analysed from Burgos. Cantabria is a distinctive region, with the Atlantic Ocean running along the northern edge and the mountain range of the Cantabrian Cordillera to the South, making it climatically different to the Northern Meseta. Few contemporary studies from Northern Spain exist, but red deer sampled from Neolithic, Chalcolithic and Bronze Age levels at El Mirón Cave, also located in Cantabria, have δ13C values ranging between − 22.2 and − 20.9‰ (Stevens et al. 2014). The environmental differences between these two regions likely explain the difference in δ13C values seen between El Abrigo de la Castañera and the other sites studied.

Of the faunal remains analysed, two of the three pigs have stable isotope values typical of consuming in an open landscape with δ13C values of −22.0‰ and − 21.2‰ (Fig. 5; Table 3). The other pig specimen has a much lower δ13C value of − 23.0‰ (Fig. 5; Table 3), which is typical of an animal feeding in a forested environment, under the influence of the canopy effect (Van der Merwe and Medina 1991).

Fig. 6
figure 6

Human bone collagen δ15N and δ13C values analysed from Arroyal I, El Hornazo, Fuente Celada, Ferrocarril-La Dehesa and Abrigo de la Castañera with mean Chalcolithic ovicaprid and cattle values. Error bars represent 1σ

Full size image

Discussion

Diet in early prehistoric northern Spain

The δ13C values of the humans and animals analysed were consistent with individuals consuming a diet dominated by terrestrial C3 resources expected for this time period (Fig. 6). At Arroyal I, Fuente Celada and El Hornazo, there was a trend towards human δ13C values being higher than − 20‰ (Fig. 6). This is not unusual for Southern Europe, where a notable difference to Northern Europe is observed, with the latter typically falling within values of − 23 to − 20‰ (Ambrose 1993; Pollard 2007) and represents different environmental and climatic conditions between European regions.

The sites represent different burial types and chronological periods, with Arroyal I being a megalithic monument, Fuente Celada containing several pit human burials, an individual burial from Ferrocarril-La Dehesa, also found in a pit grave, and burials within the rockshelter of La Castañera. Despite differences in chronology and burial type, the stable isotope evidence of diet for all individuals were remarkably similar (aside from the δ13C values at El Abrigo de la Castañera discussed previously in the results) (Fig. 6), with humans at all sites plotting c. 3–4‰ higher in δ15N than the contemporary herbivores. The signatures observed are consistent with a mixed diet, including animal protein, likely in the form of meat and potential milk products. The diet of Chalcolithic, Neolithic and Bronze Age individuals within these sites was relatively homogenous at least in terms of the quantity of protein consumed. This initial evidence contrasts with evidence from the Late Neolithic/Early Chalcolithic in La Rioja Alavesa region (Basque Country), where monumental graves were observed to have higher δ13C and δ15N values than the cave burials, consistent with potentially different agricultural strategies and access to pastures, linked to differential social status (Fernández-Crespo and Schulting 2017). Further work in the Northern Cantabrian region would help to explore potential links between burial types and social status.

The question of whether milk and dairy products were being produced and consumed at these sites remains open. The small zooarchaeological assemblages available for these sites prevent detailed mortality profiles from being generated, which can give insights into milk, meat and wool economies (Payne 1973). Cattle dominated the faunal assemblages at both Fuente Celada (Riquelme Cantal 2009) and El Hornazo (Marín-Arroyo 2011), followed by ovicaprids. Some immature cattle were also identified at El Hornazo (Marín-Arroyo 2011), which is more typical of dairying economies (Payne 1973). The limited zooarchaeological datasets available, together with other indirect evidence, such as the analysis of the ceramic assemblages, indicate that at least from the Chalcolithic period milk was consumed, mostly from bovines (Carmona Ballestero 2013: 302–306). During the Bronze Age, vessels called encellas (cheese strainer) became commonly produced, which is considered a good indication of when dairy products like cheese became a regular part of diet (Salque et al. 2013). Lipid residue analysis of Neolithic pottery from the Iberian Peninsula showed the presence of dairy fats, with dairying thought to be taking place in open air sites (Spiteri et al. 2016), showing that the capacity for milk production existed, at least, in the Mediterranean coast and it is not unreasonable to suggest that it was part of the economy at the Northern Iberian plateau sites.

Despite the evidence of riverine fish being recorded in pottery contents in Arroyal I (Ordoñez et al. 2017) and La Castañera (Vega-Maeso, pers. comm.), there is no evidence of routine freshwater fish or marine protein consumption at these sites visible in the bone collagen evidence. This suggests that, if they were being consumed at the sites, then it was not in sufficient quantities to be registered in the long-term bone collagen record and would indicate infrequent or occasional use of these resources. Similar stable isotope studies carried out in Neolithic and Chalcolithic sites across the Iberian Peninsula show similar results with no evidence of aquatic resource consumption registering in bone collagen values achieved (Arias Cabal 2005; Fernández-Crespo and Schulting 2017; Fontanals-Coll et al. 2015; Salazar-García 2011; Sarasketa-Gartzia et al. 2018).

In terms of possible plant material available, wheat and barley are the most commonly identified cereals of the Neolithic-Bronze Age in the Iberian Peninsula (Buxó and Piqué 2008) and crucial plant sources within human diet. It is possible that pulses, such as peas, lentils and fava beans, which appear in the Iberian Peninsula from the early Neolithic onwards (Zapata et al. 2004), were also a component of diet at this time. Of the terrestrial C4 plants, millet is documented in the Iberian Peninsula during the 3-2nd Millenia BCE, before becoming more widely cultivated in the 1st millennium BCE (Buxó and Piqué 2008; Moreno-Larrazabal et al. 2015). Foxtail millet (Setaria italica) was identified in Kobaederra cave during Chalcolithic (c. 3331–2906 BCE) and Arenaza during the Bronze Age (c. 2136–1746 BCE) (Zapata 2002), both sites located in Bizkaia (Basque Country). In the studied sample, if C4 plants were a component of human diet, then they were not being consumed in sufficient quantities to register in the bone collagen record. Other studies of Neolithic, Chalcolithic and Bronze Age diets in Spain show similar results with C3 signatures evidence in Neolithic and Chalcolithic Iberia (Fernández-Crespo and Schulting 2017; Fernández-Crespo et al. 2018; Fontanals-Coll et al. 2015; Sarasketa-Gartzia et al. 2018).

At the sites analysed, the carpological record for Fuente Celada was sparse with only 13 fragments of plant remains recorded (from Pits 42 and 35), showing the presence of wild strawberry (Fragaria vesca), bread/macaroni wheat (Triticum aestivum/durum) and emmer (Triticum sp.). Despite being a small assemblage, it gives an insight into the crops being produced by these populations and in the case of the wild strawberry, the plants available in the wider area. Wild resources have also been identified in assemblages from the North-East of the Iberian Peninsula including, acorns, hazelnuts and wild grapes (Antolín and Jacomet 2015), indicating that wild plants may have played an important role in diet, alongside the domestic crops. In El Hornazo, a storage pit (Pit 130) with intact contents, including a mixture of wheat (Triticum aestivum/durum) and sloes (Prunus spinosa) was preserved (López-Dóriga et al. 2011). One of the limitations of stable isotope analysis is that it is biassed towards representing protein consumption, meaning that there is a trend towards representing high protein products, such as those from animals, which can mask the contribution of plants to diet (Hedges and Reynard 2007) and, knowing the exact proportions of plant contributions to diet is challenging to establish.

Insights into animal management in northern Spain

The small number of animal remains studied for stable isotope analysis means that it is only possible to make tentative comments about the animal management strategies used at the sites, but can provide initial insights into prehistoric pastoralism in Nothern Spain. The range of δ13C values of the herbivores was all consistent with those expected for open C3 environments.

The stable isotope values of the faunal remains from Fuente Celada showed relatively large ranges in δ13C values of cattle and δ15N values of sheep, indicating that these animals were not necessarily consuming a homogenous diet. It is possible that a range of different pastures was used, resulting in isotopically dispersed values of those animals. The scale of movement of animals between pastures is a matter of debate. Sheep transhumance from the Neolithic period onwards has been identified using incremental analysis of tooth bioapatite δ13C and δ18O values (Tornero et al. 2018), which in the bone collagen record of animals could produce averaged values that differ from each other, especially if animal movements vary seasonally or annually. Transhumance over larger distances requires relatively complex infrastructure in terms of territories occupied in the wider landscape and would be logistically and politically complex to negotiate (Carmona Ballestero 2013). Instead, it is possible that animals were grazed in various pastures, representing different biotopes within the same political regions. This scenario could be indicative of smaller scale movements within individual group territories, which is consistent with models of landscape use in Later Prehistory in Central Iberia as proposed by Díaz del Río (2001, 2006). This pattern of small-scale animal movements also identified in bone collagen δ13C and δ15N analysis at the prehistoric site of Danebury (United Kingdom), where animals kept in different parts of the landscape had characteristically different δ13C and δ15N values (Stevens et al. 2013). Another explanation could be that different fodder was used, at different times, during the occupation of the site, and could suggest changing agrarian economies.

The faunal cattle and sheep with higher δ15N values may represent younger animals exhibiting a nursing effect, which typically elevates δ15N by approximately one trophic level due to the consumption of their mother’s breastmilk (e.g. Schurr 1997; Fuller et al. 2006) and could explain their elevated δ13C and δ15N values relative to other herbivores at the site. Whilst individuals did not have the porous bone structure, typically associated with infantile individuals, the lack of distal ends of bones with fusion evidence means that they could be animals exhibiting nursing signatures. The cow labelled as individual ‘a’ in Fig. 3 has an unusually high δ15N value of 8.8‰, combined with a lower δ13C value of − 21.5‰. The lower δ13C suggests that this individual may have been raised on an isotopically different region to the other individuals analysed. This may be indicative of the use of multiple pasture landscapes in the management of animals, with pastures further afield potentially being used to raise animals before bringing them to the site for slaughter. This could be due to seasonal uses of different tracts of land, which if used routinely could impact on the long-term bone collagen record of an individual. Another possibility is that this animal bone could be from an animal trade on the hoof or a cut of meat traded to the site. The trade of perishable goods is relatively complex to track archaeologically, in comparison to items such as ceramics or lithics. Zooarchaeological methods can often distinguish between producer and consumer sites, where trade and exchange were central to the economy (Crabtree 1990), but occasional or infrequent trade or exchange of animals would require individual life histories of animals to be created and is more complex to establish when only small assemblages are available. Future incremental analysis of available animal teeth for δ13C and δ18O analysis would help to establish more about Prehistoric landscape use and could provide insights into the possible exchange of animals.

The similarity in the wild red deer δ15N value from El Hornazo and the majority of the domestic herbivores at the site (and at Fuente Celada, also located in the same village) suggest that sheep and cattle were occupying similar parts of the landscape as the red deer. A further implication, based on the similarity of δ15N values of the wild and domestic species, is that the cattle and sheep analysed were not fed on crops that were enriched with manure, which would then produce elevated δ15N values in plants (Bogaard et al. 2007) and thus consumers of these plants. The δ13C values of the domestic herbivores are slightly lower than the wild red deer (which would not typically be subjected to anthropogenic management processes) and could hint at the cattle and sheep being fed on some domestic crops, potentially as winter fodder. Differences in the stable isotope values of wild and domestic species at the Neolithic site of La Draga was also attributed to the use of crops as fodder for the domestic stock (Navarrete et al. 2017), which supports this hypothesis.

The three suid specimens sampled at El Abrigo de la Castañera had different diets. One of the Sus sampled had a lower δ13C value, which is characteristic of an animal feeding in a forested environment (van der Merwe and Medina 1991). Osteologically, it was not possible to determine if these animals were domestic or wild. If these individuals were all domestic, the implication is that differential animal management was being employed. Studies of Neolithic pigs in England have shown that domestic pigs may have been herded under partially wooded environments (Hamilton et al. 2009). This may, to some extent, have been a method of protecting crops from pig rooting, which can damage plants and shoots (Ballari and Barrios-García 2014). Alternatively, the specimen from La Castañera may be a wild boar, which typically inhabit forested environments and would explain the lower δ13C value. If true, this would suggest that wild animals were exploited, either through targeted hunting or chance encounter. The presence of red deer in the faunal assemblages at Fuente Celada (Riquelme Cantal 2009) and El Hornazo (Marín-Arroyo 2011) also shows that wild animals were used, as an occasional supplementary resource for these farming populations, and wild boar may also have been an additional resource exploited.

The relationship between humans and dogs

Dogs have been used as a proxy for human dietary behaviour (e.g. Clutton-Brock and Noe-Nygaard 1990; Fischer et al. 2007; Schulting and Richards 2002b) and how far they accurately represent human diet in archaeological populations has been debated (Guiry 2012). It is generally accepted that the diet of dogs generally tracks that of the human population they live with, irrespective of temporal and geographical situation or the culture of populations (Guiry and Grimes 2013). The Chalcolithic dog from Arroyal I was indistinguishable from the human δ13C and δ15N values, indicating that this animal had a similar diet to the humans, either through intentionally being fed scraps from the human table or scavenging refuse from human waste. At El Hornazo, the situation is more complex also during Chalcolithic, with one dog reflecting the human diet exactly plotting alongside the humans, and three other dogs plotted with the herbivores. This suggests that dogs are not necessarily a good proxy for human diet in this region. The difference in the diet of the dogs may represent the role of dogs at the site, in terms of whether they are household pets, being fed scraps from the human table (including meat) or whether they were working animals, that may have consumed a more regimented diet comprising predominantly of vegetal matter.

Prehistoric population movements in the northern Iberian Peninsula

The movement of human populations during Neolithic, Chalcolithic and Bronze Age period in Europe is currently a subject of great interest and genetic evidence has demonstrated links between Neolithic Iberian populations and central European Beaker populations (Olalde et al. 2018), indicating the large-scale movements involved. The large difference between the δ34S values of the two individuals analysed from Arroyal I suggests that they were not from the same location, which is also supported by SUC28 having a lower δ13C value than other individuals in the population analysed. Published relationships of δ34S values show that within living populations, an average range of 1.9‰ in sulphur values is observed, with slightly higher range of 2.4‰ between archaeological populations (Nehlich 2015: 8). The large difference between the δ34S values of both individuals indicates that these people lived in different locations before they came to be interred at the site. It could be suggested that individual SUC28 came from a more saline or coastal environment due to the higher δ34S value. The effect of salinity on δ34S values has been demonstrated, with coastal δ34S values reaching up to + 20 (Nehlich 2015). The effects of salinity on baseline δ34S values can reach distances of up to 30 km from the coast, although the effects become less potent with distance (c. 10‰) (Nehlich 2015). The δ13C of individuals can also be affected by salinity, although archaeological studies of human and animal bone show that sea spray can cause elevated δ34S of ~ + 14 to + 18 whilst still producing δ13C values of ~ − 23 to − 21‰ (Nehlich 2015, based on data from Craig et al. 2006; Fornander et al. 2008; and Linderholm et al. 2008a, b). A likely location of origin of this person could be, perhaps further north, towards the Cantabrian coast, which is geographically the closest coastline to the site. This hypothesis is supported by the individual having a δ13C that plots within the range of humans analysed at El Abrigo de la Castañera, which is also located in Cantabria (Fig. 1). Recent work at the sites of Santimamiñe and Pico Ramos, in the Basque Country, exploring late Neolithic and Chalcolithic populations using 87Sr/86Sr analysis demonstrated that there were movements of the populations between the Pyrenees to the Biscayan coast (Sarasketa-Gartzia et al. 2018), showing that there was fluidity in the population of earlier prehistoric populations in the Northern Iberian Peninsula. In Arroyal I, there is no other evidence of contact with the coast, despite its relative proximity. In contrast, El Hornazo contained one of the few maritime shells located in a non-funerary Chalcolithic inner context (Gutiérrez Zugasti et al. 2014). Until now, seashells were the only signs of contact between the inland inhabitants of the North Castilian plateau and the coast. Therefore, smaller scale intra-regional movements might have been common also during the period and region of study in these farming populations.

Studies of ceramics from the Cantabrian Region provides further evidence for the movement of populations in Prehistory (Vega Maeso 2017). The distribution of decorative motifs and shapes on pottery artefacts appear to relate to the movement of people, particularly women, rather than the exchange of vessels (Vega Maeso 2017). The tentative results from this, also indicate that the individuals buried at Arroyal I are not necessarily all local to the area and supports other evidence for population movement within the Iberian Peninsula during the Chalcolithic and Bronze Age periods. A similar hypothesis was proposed at the Chalcolithic site of La Atalayuela in La Rioja, where two individuals interred at the site with outlying δ13C values were interpreted as potentially being from different regions (Fernández-Crespo et al. 2018) again suggesting population fluidity within different regions in Prehistoric Northern Spain.

Conclusions

This study has provided valuable insights into the economy, animal management and mobility of early farmers of the Northern Spain Plateau, a region which has received little attention so far. Analysis of bone collagen δ13C and δ15N values from human remains from different mortuary sites dating between the Neolithic and Bronze Age shows that the diet of individuals analysed was remarkably similar. Humans consumed a predominantly C3 terrestrial diet, including animal protein (meat and likely milk/dairy products) and wider regional archaeobotanical evidence suggests that diet would also have included C3 plants, such as wheat and barley and vegetables. The faunal samples provided insights into animal management. The diversity in δ13C and δ15N values suggests a wider use of the landscape in the management of herbivores, possibly with the access of animals to different pastures with diverse biotopes. Further work using tooth incremental isotope analysis would be valuable in exploring this concept and to explore whether there is evidence of seasonal changes in diet related to animal movements or foddering strategies. Differences were observed between individuals from the coast and those of the interior, with lower δ13C values likely due to baseline environmental differences between the two regions and δ34S analysis of two individuals suggests a level of inland mobility during this time. Based on the similarities of diets between individuals analysed, it appears that the socioeconomic conditions experienced between the Neolithic and Bronze age in central Northern Spain remained stable or without remarkable dietary changes. This study has helped to expand our understanding of diet, economy and lifeways of an understudied location, where few skeletal remains are available and has demonstrated the need for further research in the region to provide detailed reconstructions of agricultural practises, diet and mobility of Prehistoric populations.