Skip to content

Compound-specific radiocarbon dating of lipid residues preserved in archaeological pottery vessels

Bristol student theses: Doctoral ThesisDoctor of Philosophy (PhD)

Authors

Research units

Abstract

While pottery vessels are widely recovered at archaeological sites their absolute dating by radiocarbon is challenging. Adsorbed lipids residues preserved within the matrix of the vessels and thus, protected from contamination in the burial environment, are widespread and often recovered in concentrations sufficient for radiocarbon dating. The most common residues correspond to animal fats, distinguishable by the dominance of the C16:0 and C18:0 fatty acids (FAs), have the potential to be dated at the molecular level using preparative capillary gas chromatography (PCGC; Stott et al. 2001; Berstan et al. 2008), however, the preliminary studies did not meet the accuracy and precision required. This thesis addressed the compound specific radiocarbon analysis (CSRA) of adsorbed lipids extracted from pottery vessels for the establishment of a reliable procedure which can be used as routine. The first consideration focussed on the elimination of exogenous contaminants associated with the isolation procedure. This consisted of (i) quantifying the contamination associated with the thermal degradation of the GC column stationary phase, which was demonstrated to be negligible, (ii) the elimination of the solvent used for recovery of trapped analytes using a new trap design, and (iii) the removal of ‘memory’ in the trapping system using a heat-based cleaning method. The accuracy and precision of 14C determinations using PCGC were first established by dating bulk and isolated compounds on a wide age range of FAs from modern standards and archaeological bog butters prior to its application to 100 pottery vessels. The 14C dates on lipids extracted from pottery vessels showed excellent compatibility for dendrochronologically and radiocarbon dated materials from sites in wetland (Sweet Track, UK) and arid (Takarkori Rock Shelter, Libya) locations. The oldest dates obtained from the site of Çatalhöyük (Turkey) were successfully tested in the pre-existing Bayesian statistical model based on the stratigraphy. However, pot lipids from a coastal site (Cliffs End Farm, UK) produced older ages than their surface residues analogues probably due to a reservoir effect from aquatic product processing in the vessels. The organic residue analysis (ORA) of ca. 900 potsherds from the Alsace (France) region provided insights into sampling strategies for pottery assemblage through recognition of the potsherds with the highest potential for CSRA, namely: (i) cooking pots with high lipid preservation, (ii) refitted potsherds to avoid residual/intrusive materials, and (iii) lipid extracts without aquatic biomarkers to avoid reservoir effect. Radiocarbon dates on Alsatian potsherds were compatible with the reference dates on bones and surface residues. The dates on Middle Neolithic potsherds were successfully included in the Bayesian statistical model based on the seriation and typological study of pottery assemblages in the region.
The technique showed its unique use to answer questions relating to the exploitation of food procurement practices, such as the emergence of dairying in central Europe, where the earliest dairy residues were directly dated to the late 53rd century BC. Furthermore, the analysis of lipid residues from a coastal site (Bornais, UK) with known use of aquatic products led to the correction of 14C dates of lipid extracts using an estimate for the percentage of marine and terrestrial fats processed in potsherds together with an appropriate value for the local marine reservoir offset.
Significantly, vessels of diverse ages and burial environments can now be routinely and accurately dated from their adsorbed lipid residues (deriving from terrestrial or aquatic commodities) with equivalent precision to other commonly dated materials.

Details

Original languageEnglish
Awarding Institution
Supervisors/Advisors
Award date23 Jan 2019

Documents

View research connections