The impact of farming on prehistoric culinary practices throughout Northern Europe

Significance How prehistoric farming became established in Northern Europe, a region that supported dense populations of hunter-gatherer-fishers, has concerned archaeologists for over a century. Through analysis of the organic residues recovered from over 1,000 vessels dating across the transition to farming, we found unexpected consistency in the use of aquatic foods at odds with prevailing narrative of large-scale demographic and economic change. We argue that the ability of farming groups to adapt to their environment by learning hunter-gatherer-fisher practices, combined with dairying, was key to their northerly expansion. We also provide evidence of dairy use by hunter-gatherers which we attribute to long-distance exchange with farmers, implying a much greater degree of interaction and cooperation than previously described.

This PDF file includes:  AverageR is a generalized additive mixed model that uses a thin plate regression spline (1,2).The dependent variable used was the percentage of dairy samples and was established by site/millenium according to the number of samples with Δ 13 C <-3.1‰ against the total number of samples analyzed (Dataset S2).Four local average models were generated according to the following millennium time slices: 6th, 5th, 4th and 3rd millennia cal BCE.

Geographical attributes
To determine whether aquatic biomarkers were more prevalent on coastal sites or near rivers, the

European
Commission's Catchment Characterisation Model (CCM2) database (data.europa.eu/89h/fe1878e8-7541-4c66-8453-afdae7469221), was used to find the minimum distance between each site and the nearest river or tributary using the v.distance tool in GRASS GIS.This was also used to calculate the geodesic distance to the modern-day coast using the 1:10m global coastline map from Natural Earth (www.naturalearthdata.com).However, neither distance to coast nor distance to river can predict the presence or absence of aquatic biomarkers (Figure S3).The same is true for dairy biomarkers (Figure S3), both in the data reported in this paper and also the prehistoric dairying phenomenon in Central and Western Europe studied by Evershed et al. (3).Finally, in contrast to sites with hunter-gatherer pottery in the forest-steppe area of Eastern Europe (4), the sites in this study are not preferentially located close to navigable rivers when compared to a random distribution of points in the landscape (Figure S3).Whether this represents a deliberate change to settlement systems or is simply reflective of differences in catchment drainage and hypsometry requires further study.
Dataset S2 (separate file).Frequency of dairy fat residues across sites from Northern Europe.
Figures S1 to S3 supporting materials for this manuscript include the following: Datasets S1 to S2 Supporting Information Text Generating the spatio-temporal distribution of prehistoric dairying across Northern Europe To generate spatial estimates of the frequency of dairying, we employed the AverageR model available as R-based Open Access apps (https://www.isomemoapp.com/)developed within the Pandora & IsoMemo initiatives (Data search and Spatiotemporal modeling version 23.4.1.10).

Fig. S1 .
Fig. S1.Map of the study area (physical) showing the regional divisions (a-e) used in this study.Archaeological sites with typical hunter-gatherer pottery (circle), early farmer pottery (square) or both (triangle) are shown.The numbers are reported in TableS1.

Fig. S2 .
Fig. S2.Standard error of the mean (1σ) of the spatio-temporal distribution of prehistoric dairying across Northern Europe during the: (a) 6th, (b) 5th; (c) 4th; and (d) 3rd millennia cal BCE.Sites with (squares) and without (circles) evidence of domesticated animals is shown.

Fig. S3 .
Fig. S3.Barplots showing frequency of sites with and without aquatic and dairy biomarkers, scatter plots of the prevalence of aquatic biomarkers versus river distance, the density distribution of sites with respect to distance from rivers, and scatter plots showing the prevalence of dairy biomarkers versus coastal distance, including comparison with the Evershed et al. (3) dataset.