High-resolution isotopic evidence of specialised cattle herding in the European Neolithic

Reconstructing stock herding strategies and land use is key to comprehending past human social organization and economy. We present laser-ablation strontium and carbon isotope data from 25 cattle (Bos taurus) to reconstruct mobility and infer herding management at the Swiss lakeside settlement of Arbon Bleiche 3, occupied for only 15 years (3384–3370 BC). Our results reveal three distinct isotopic patterns that likely reflect different herding strategies: 1) localized cattle herding, 2) seasonal movement, and 3) herding away from the site year-round. Different strategies of herding are not uniformly represented in various areas of the settlement, which indicates specialist modes of cattle management. The pressure on local fodder capacities and the need for alternative herding regimes must have involved diverse access to grazing resources. Consequently, the increasing importance of cattle in the local landscape was likely to have contributed to the progress of socio-economic differentiation in early agricultural societies in Europe.

with the introduction of flax, emmer, clay spindle whorls, and the advent of draft-animal use [11]. The first (pioneer) house was built in 3384 BC, followed by the construction of two additional houses in the subsequent year. By 3381 BC the settlement had grown by ten more buildings. One to three additional houses per year were erected until 3376 BC (S5 Fig). In 3370 BC, the settlement was abandoned after a fire event [1]. The excavated houses were rather uniform, with an average size of 4 m x 8 m. Two square buildings of 2 m x 2 m (houses 17 and 25, cf. Fig 2) were exceptions in this regard. Rows of houses were separated by narrow lanes.
Find attribution was performed according to Doppler [12], assuming that archaeological finds of a house reflect the activities and dietary habits of its inhabitants. The analysis of about 70,000 animal bone fragments revealed a bimodal distribution of animal remains within the settlement, suggesting differential dietary habits (mainly beef and littoral fish species in the northern part of the settlement, predominantly pork and limnetic fish in the south [13,14]. The partitioned animal bone distribution also suggests a certain degree of economic and/or cultural differentiation. Another indication for an economic division of the settlement is given by the uneven distribution of wild and domestic plant species [15,16]. Different cultural traditions, or even the presence of culturally distinct groups within the settlement [14], is suggested by the fact that besides local ceramics there is evidence for foreign-style pottery (Baden and Altheim culture), indicating increasing influence from more eastern European regions [17]. It is assumed that people from these regions introduced their technological know-how and integrated with local communities [18].

Estimating land requirements for cattle husbandry
The estimated need for pasture land of 10-20 km 2 for 60-120 cattle, respectively, is based on ethnographic data [19]: mobile cattle husbandry requires grazing areas of 4-14 ha per year for animals in deciduous forests and 8-20 ha in coniferous woodlands depending on the vertical and horizontal structure of the landscape. In the mosaic-like landscape with both deciduous and coniferous woodland as evidenced for Arbon Bleiche 3 [11], it can be assumed that one cattle needs 6-17 ha per year. We consider the mean upper end of estimates for both forest types (17 ha) to be more realistic since the Neolithic landscape was rather densely forested with only small patches of grassland suited for grazing. Although the consideration of these maxima is a hypothetical approximation, it stands in good agreement with observations in subalpine regions [20], where one cattle needed a minimum feeding ground of 5 ha during a season (3-4 months) to avoid damage to trees and allow the animal to select the best locations and plants. This considerable amount of land requirement may explain why people in Arbon Bleiche 3 had to exploit diverse grazing grounds in order to respect the carrying capacity of their environment.

Local 87 Sr/ 86 Sr signatures
Since the choice of the most suitable baseline sample for characterizing the local biologically available Sr is a matter of discussion [21][22][23], we use a mixture of different materials: the enamel of locally living animals (pigs), the water of potential drinking water sources, and (diagenetically altered) dentine. After burial, dentine (in contrast to tooth enamel) is highly susceptible to Sr isotope exchange with water in the soils. It thus reflects a mixture of biogenic and diagenetic strontium due to diagenesis [24,25] and can therefore be used as a proxy of the local 87 Sr/ 86 Sr baseline signature. The average 87 Sr/ 86 Sr of biologically available Sr at Arbon Bleiche 3 based on dentine from cattle (Bos taurus) and red deer (Cervus elaphus), as well as on tooth enamel from pigs (Sus domesticus), were in good agreement with modern Sr isotopic water signatures determined for nearby water sources, such as the river Aach and Lake Constance (S2 Table). The cattle dentine yielded a mean 87 Sr/ 86 Sr of 0.70880 ± 0.00037 (2σ, n = 12), while the mean 87 Sr/ 86 Sr of the red deer dentine was 0.70859

Identifying seasonal migration based on strontium isotope analysis
Davies recorded movements of livestock from seven Alpine settlements, and while the movement regimes observed in this study are somewhat different for each livestock, they all share a common pattern of ~7 months of stasis during the cold season [27]. These periods of movement and stasis can be identified using strontium isotope analyses by high-resolution LA-MC-ICP-MS and linked to published rates of enamel formation [28,29]. Seven months represent about 60% of the mineralization time of cattle M2, and about 50% of M3. Although we acknowledge that deviation from this may result from a non-constant rate of enamel growth [30], or from herding practices that significantly deviate from Davies' observations [27], we find that the majority (80%) of cattle that display mobility pattern 2 (ARB 10, 14, 22, 119) and 50% of the cattle representing mobility pattern 3 (ARB 16, 23, 33, 24, 114) show relatively invariant (and low) 87 Sr/ 86 Sr values over a mineralization time that is consistent with about this length of stasis. While we cannot be completely certain with regards to the exact timing of these periods of stasis, it is reasonable to assume that they relate to foddering practices, which were likely to be highly seasonal in this sub-alpine region.