Abstract Carbon isotope measurements of individual fatty acids (C16:0 and C18:0) recovered from archaeological pottery vessels are widely used in archaeology to investigate past culinary and economic practices. Typically, such isotope measurements are matched with reference to food sources for straightforward source identification, or simple linear models are used to investigate mixing of contents. However, in cases where multiple food sources were processed in the same vessel, these approaches result in equivocal solutions. To address this issue, we tested the use of a Bayesian mixing model to determine the proportional contribution of different food sources to a series of different mixed food compositions, using data generated both by simulation and by experiment. The model was then applied to previously published fatty acid isotope datasets from pottery from two prehistoric sites: Durrington Walls, near Stonehenge in southern Britain and Neustadt in northern Germany. We show that the Bayesian approach to the reconstruction of pottery use offers a reliable probabilistic interpretation of source contributions although the analysis also highlights the relatively low precision achievable in quantifying pottery contents from datasets of this nature. We suggest that, with some refinement, the approach outlined should become standard practice in organic residue analysis, and also has potential application to a wide range of geological and geochemical investigations.

中文翻译：

---

使用贝叶斯推理从脂肪酸碳同位素特征重建史前陶器的使用

摘要 从考古陶器中回收的单个脂肪酸（C16:0 和 C18:0）的碳同位素测量广泛用于考古学，以调查过去的烹饪和经济实践。通常，此类同位素测量与食物来源相匹配以进行直接来源识别，或者使用简单的线性模型来研究内容的混合。然而，如果在同一容器中处理多种食物来源，这些方法会导致模棱两可的解决方案。为了解决这个问题，我们测试了贝叶斯混合模型的使用，以确定不同食物来源对一系列不同混合食物成分的比例贡献，使用模拟和实验生成的数据。然后将该模型应用于先前公布的来自两个史前遗址陶器的脂肪酸同位素数据集：英国南部巨石阵附近的杜灵顿墙和德国北部的新城。我们表明，贝叶斯重建陶器使用的方法提供了对来源贡献的可靠概率解释，尽管分析还强调了从这种性质的数据集中量化陶器内容可实现的相对较低的精度。我们建议，经过一些改进，概述的方法应该成为有机残留物分析的标准做法，并且还具有广泛的地质和地球化学调查的潜在应用。我们表明，贝叶斯重建陶器使用的方法提供了对来源贡献的可靠概率解释，尽管分析还强调了从这种性质的数据集中量化陶器内容可实现的相对较低的精度。我们建议，经过一些改进，概述的方法应该成为有机残留物分析的标准做法，并且还具有广泛的地质和地球化学调查的潜在应用。我们表明，贝叶斯重建陶器使用的方法提供了对来源贡献的可靠概率解释，尽管分析还强调了从这种性质的数据集中量化陶器内容可实现的相对较低的精度。我们建议，经过一些改进，概述的方法应该成为有机残留物分析的标准做法，并且还具有广泛的地质和地球化学调查的潜在应用。