Temporal patterns in Mesolithic activity at Duvensee, Germany

The Duvensee peat bog is one of the best-known prehistoric landscapes in northern Europe, with hunter-gatherer activities located on small islands on the western edge of an Early Holocene lake. Excellent organic preservation, precise excavation, and rigorous radiocarbon sample selection permit the application of Bayesian chronological modelling. Over 250 radiocarbon results date a dozen Mesolithic sites to an extended period between the 9th and 7th millennia BCE. Each site may only have been used briefly – perhaps only for a single season in some cases – but some were used repeatedly over the course of decades or even centuries. Site chronological models reveal sharp fluctuations in the overall level of recorded Mesolithic activity. A spike in the intensity and diversity of activity coincided with the rapid expansion of hazel ( Corylus avellana ) at the start of the Boreal biozone, but a steep decline in dated activity in the later 9th millennium BCE may be an artefact of changes in lake level, rather than reflecting a response to vegetation change, climate deterioration or societal reconfiguration. Although the Duvensee radiocarbon results can be construed to support arguments that rapid climate change had a significant impact on human activity, known climate events occurred when the record of human occupation at Duvensee was already discontinuous. It is therefore difficult to infer (or exclude) environmental factors as triggers of changes in the archaeological record. Our results suggest focusing research on the early Boreal phase, during which the widest range of site-types is preserved and accessible for archaeological investigation. • Bayesian models of 251 14 C dates from 12 early Mesolithic sites at Duvensee, Germany. • Dates span 2500 years, single sites were used for up to 400 years but often much less. • Site complexity varies with longevity, range of activities may not change over time. • Site survival and visibility are determined by lake-level fluctuations. • Scale of recorded activity cannot show role of vegetation change or climate events.