Two key objectives were proposed: i) To establish a new analytical approach to the investigation and identification of visible and absorbed organic matter associated with archaeological ceramics, and ii) to explore the implications of compositional variation in complex mixtures extracted from archaeological ceramics using solvent extraction and pyrolysis techniques. These key objectives encompassed; the characterization of residue patterns, identification of the principal contents, evaluation of the re-use of vessels, the variation in resin composition caused by heating, a comparison of two different burial environments within Egypt and a comparison to sherds from Israel.
This is the first comprehensive study of the biomarkers present in a range of Canaanite amphorae imported into Egypt during the Late Bronze Age. Specifically this project has achieved the following: 1) evaluation of a novel extraction and derivatisation reagent, 2) the development of a new sampling methodology, 3) confirmation of the contents of the jars as resin and lipid matter through identification of triterpenoids and fatty acids, 4) assessment of organic preservation at different sites, detailed examination of compositional changes in triterpenoids with heating, and 6) through affiliation with a concomitant study (Serpico et al. 2003) of contents and clay composition of the jars, important verification of link between specific fabrics and commodities. This latter is especially important as it will enable archaeologists to trace trade routes for the organic products identified in this NERC project across the Mediterranean.
Results were achieved through analysis of 192 samples predominantly from two sites in Egypt, Amarna and Memphis. A number of sherds from Israel and other sites in Egypt were also analysed (see Table 1) as well as eleven sherds from Amarna which carried inscriptions relating to their original contents (incense, two types of vegetable oils and honey). In addition, almost fifty different samples of modern authentic resins, oils and oil producing seeds, nuts and tubers were obtained from Kew Gardens, Egypt, and other sources. This has been used to build a reference collection for comparison with ancient organic residues. The compositions of these samples has been examined and compared to published data. It was also possible to analyse a small selection of ancient desiccated oil seeds from Amarna to compare to ancient and modern oil compositions.
|
Amarna Canaanite |
Memphis Canaanite |
Israel Canaanite |
Unprovenanced / non-Canaanite |
Sherds |
47* |
17 |
24 |
16 |
Residues |
60* |
0 |
1 |
27 |
Table 1. The provenance of Canaanite sherds residues. Different fabric type and locations within the sites are not shown in this table. * includes 10 sherds with attached residues.
It was proposed that trimethyl(a,a, a-trifluoro-m-tolyl)ammonium hydroxide (TMTFTH) would be a rapid extraction and derivatization agent for use with gas chromatography and combined gas chromatography/mass spectrometry, for the examination of archaeological organic residues. It would reduce the number of method steps, and thus minimise potential contamination, smaller sample sizes would also be possible and lead to increases in throughput. To explore these benefits, TMTFTH was examined under a variety of different extraction and analytical conditions, with free, ceramic absorbed, modern and ancient samples containing triglycerides, free fatty acids, diterpenoids and triterpenoids. However, this reagent was found to cause analytical artefacts for the terpenoids, and therefore solvent extraction with diazomethane derivatization was established for these molecules. However, the artefacts were not significant enough to mislead the identification of Pistacia resin. TMTFTH was found to hydrolyse and methylate triglycerides into their constituent fatty acids in addition to methylating free fatty acid residues, but the recovered yields were lower than for the more conventional alkaline hydrolysis (saponification), followed by diazomethane derivatization. Saponification therefore became the method of choice for fatty acid residues, with sample size reduced to a standard 0.1g of ceramic. Considerably smaller than previously reported for ceramic residues. Comparison of solvent extraction, TMTFTH, saponification and TMTFTH after solvent extraction was carried out to examine the relative efficiency of these different extraction protocols (Stern et al. 2000).
The aims of pyrolysis were twofold: first to identify the composition of any crosslinked macromolecular lipids, second, to examine the comparative roles of conventional solvent extraction, TMTFTH and saponification in releasing such components. Pyrolysis GC-MS was carried out at the NERC Organic Mass Spectrometry Facility. In total 19 samples were analysed. The previously saponified samples yielded no peaks, indicating that no lipid residue remains, either free or as macromolecular material. In contrast, those which had been solvent and TMTFTH extracted yielded complex chromatograms, showing that lipids were still present. Sherds inscribed for Honey, and Sherds from Memphis and Saqqara associated with the transport of oils and resins yielded no pyrolysis products, confirming that the absence of organic remains is due to lack of preservation and not the generation of bound crosslinked macromolecules. Ancient Pistacia residues yielded complex chromatograms which were difficult to interpret. This suggests that pyrolysis does not yield as much information as conventional techniques, but given the numbers of factors involved, this can only be viewed as a preliminary study.
Triterpenoids were extracted from both as visible and ceramic-absorbed residues. Comparison to characteristic molecular markers in authentic samples identified this material the resin of Pistacia spp. (Figure 1). A range of other triterpenoids present in the ancient samples have also been characterised in the ancient samples (Stern et al. 2003) However, no lower terpenoids were found. This is not surprising as the monoterpenes present in Pistacia resin are particularly vulnerable to loss by evaporation.
Figure 1. Partial chromatograms of (upper figure) modern Pistacia resin, (lower figure) visible residue from Amarna sherd associated with the transport of resin. Molecular identification by GC-MS: 1) methyl moronate (see Molecules with silly or unusual names for more on molecules like moronic acid!), 2) methyl oleanoate, 3) methyl isomasticadienoate and 4) methyl masticadienoate.
Archaeological evidence indicated that Pistacia residues in Canaanite amphorae and locally produced bowls at Amarna may have been heated during ancient use. To test this, modern Pistacia resin was heated under controlled conditions and the triterpenoid markers compared to unheated resin and to ancient residues believed to have been heated. A variety of changes were observed in the heated resin, with increases in 28-norolean-17-en-3-one and the presence of additional triterpenoid peaks both of which were also observed in the archaeological heated samples (Stern et al. 2003). Fatty acids were extracted from sherds associated with the transport of oils. The fatty acid distribution was in the range of C10:0 - C24:0 with a C16:0 maximum and an even- over odd-carbon number predominance. Triglycerides were only occasionally extracted, and then in low abundance. The reduced amounts of unsaturated fatty acids and the presence of (branched and odd carbon numbers and the absence of sterols, indicates that the oil residues are too degraded to identify the oil source. Dicarboxylic acids were extracted, with saponification proving to be especially efficient at their recovery. It was initially proposed that drying, semi- and non-drying oils might be identified from the dicarboxylic acid distribution. However, and most importantly, a preliminary analysis of the distribution of dicarboxylic acids extracted from ancient desiccated seeds at Amarna from these oil groups showed similar distributions (C6 - C12, with C9 the most abundant). This similarity was extended to the sherd extracts. Therefore it is clear that the dicarboxylic fatty acid distribution cannot be used to further identify the oils, quantification may be a way forward. Given the absence of sterols undoubtedly due to their low natural abundance, the reduced amounts of unsaturated fatty acids and the presence of branched and odd carbon numbered fatty acids, we conclude that the oil residues are too degraded to identify the oil source (Stern et al. 2000). Similarly, honey was not detected in the samples, most likely due to the vulnerability of sugars to degradation. Mixtures of the above molecules, which could indicate processing of raw materials in ancient times, were not observed.
While it was possible to export a number of sherds Egypt under licence, the exportation of associated soil samples, useful for the consideration of organic ingress from the burial matrix, was not possible. In order to examine potential contamination issues, sherds were powdered in sections from the interior to the exterior walls in 2mm layers. Each section was then extracted and the recovered lipids quantified. Possible contamination could be assessed by comparison of the exterior of the sherd, which will exhibit ingress from the burial environment, to the interior, where the ingress could be combined with the original contents. In total 55 sherds had the full concentration profile measured over 3-11 sections through each sherd, making a total of 254 sherd sections analysed. Where lipids were present, a definite pattern emerged whereby the yields were generally highest in the interior surface, lower in the core, then rising somewhat at the exterior, consistent with the presence of oil as the ancient storage product in the jar (Figure 2). Pistacia residues were extracted from the interior few millimetres only. Visible and ceramic-absorbed Pistacia extracts were observed to be very similar in composition.
Figure 2. Concentration profile (mg per g of sherd) of palmitic (C16:0) (black columns) and stearic (C18:0) (white columns) fatty acids extracted from Amarna sherd in a fabric associated with the transport of oil. Saponified, diazomethane derivatization. Stern et al. (2000).
To examine the effects of different environments and to examine the contents of Canaanite amphorae from a variety of archaeological sites, a number of sherds were examined from different locations: Tell el-Amarna (Egypt) 48 sherds, Tell el-Amarna 79 visible residues, Memphis (Egypt) 19 sherds, Canaanite sherds from other locations including Israel, museum collections and other sites in Egypt (7). The site of Amarna was the only location which had the preservation of visible residues (Pistacia resin). In addition, ceramic absorbed residues, of both Pistacia and oils, were only extracted from sherds at Amarna. Differences are attributed to the different environmental conditions (Stern et al. submitted).
Stern B., Heron C., Serpico M. and Bourriau J. (2000) A comparison of methods for establishing fatty acid concentration gradients across potsherds: A case study using Late Bronze Age Canaanite amphorae. Archaeometry, 42, 399-414.
Stern B., Heron C., Corr L., Serpico M. and Bourriau J. (2003) Compositional variations in aged and heated Pistacia resin found in Late Bronze Age Canaanite amphorae and bowls from Amarna, Egypt. Archaeometry, 45, 457-469. [click here for.PDF]
Serpico M., Bourriau J., Smith L., Goren Y., Stern B. and Heron C. (2003) Commodities and Containers: A Project to Study Canaanite Amphorae Imported into Egypt during the New Kingdom. In The Synchronisation of Civilisations in the Eastern Mediterranean in the Second Millennium BC II M. (Ed. Bietak M.). Proceedings of the SCIEM2000 Euro-Conference, Haindorf, May 2001. Österreichischen Akademie der Wissenchaften, Vienna, 365-375.