Rasmussen, L.H.Bredwa-Mensah, Y.Borggaard, O.K.Koch, C.B.Breuning-Madsen, H.2019-04-152019-04-152009-01Volume 109, Issue 1,Pages 69-79https://doi.org/10.1080/00167223.2009.10649596http://ugspace.ug.edu.gh/handle/123456789/29236The suitability of using traditional soil chemical and mineralogical methods combined with chemometrics to trace provenance of archaeological samples was tested on potsherds from Frederiksgave, a former Danish plantation in southern Ghana, in use from 1830-1850. Soil and six potsherds from Frederiksgave, together with potsherds from two likely production sites at Ga and Dangme Shai, were investigated by visual inspection, total element analysis and X-ray diffraction and the results analyzed by principal component analysis (PCA) and cluster analysis. The investigation clearly showed that the Frederiksgave pots were not locally produced but came from other production sites indicating trading interaction between the plantation and surrounding settlements. One sample undoubtedly originated from Dangme Shai and another one must have been produced at a third (unknown) site. The provenance of the remaining samples is uncertain, but possibly three of them came from Ga and maybe one from Dangme Shai. Ill-defined compositional variation at Dangme Shai and Ga sites is thought to be the main reason for the observed difficulties in tracing the precise provenance of the Frederiksgave samples.enChemometricsCluster analysisGhanaPotteryTotal elemental analysisX-ray diffractionArticle