*4.2. Location, Location, Location: Considering Vessel Topography*

As we move to refine our methodologies, it is worth noting—and attempting to better understand—the nature of certain variabilities exhibited among extractions from the same vessel and their possible relationship to the location at which sampling occurred [1,14]. One of the more striking results from within the Mochlos group, for example, was the variable evidence for *Tilia*. Docosane was present in quantities of 8.64% RA in samples taken from the lowest point of the Mochlos vat (Table 1), but was absent, or appeared only in trace quantities, from other areas of the vat such as the vessel spout, despite the use of the same solvent. These distinctions are arguably meaningful reflections of the vessel use, as the greatest abundance was found in the vat base where the contents would have most likely pooled and therein had greater opportunity to permeate the clay, as opposed to the sides, which would have had more sporadic exposure to the contents. Similar disparities were noted in relation to the presence of cholesterol, which was detected only in the vertical, conical spout of the vat, and which points to the use of wool as a filter to strain macroscopic impurities [23]. The Tourloti jar, which being a complete closed form was sampled using the swish method, also showed some variability among the three extractions taken, as indeed the total ion chromatograms (TIC) demonstrate. Note, for example the absence of cinnamic acid and camphor in extractions one and three (Figure 6, Table 1). The qualitative differences between the Tourloti samples are not yet fully understood, but do not appear to be related to degradation processes. It may be that, as with the Mochlos objects, these discrepancies are related to the site of sampling, and perhaps to the variable properties and depositional densities and locations of the compounds within, which might differentially permeate the base, side, and shoulders of the jar. Such variability suggests the importance of considering the shape, use, fabric, and function of vessels when sampling organic residues, and the importance of multiple extractions that target different areas of a vessel to ensure the broadest possible coverage and therein greater accuracy.

#### *4.3. Rehabilitating Legacy Objects*

Above all, the case study above demonstrates the viability and value of legacy objects, provided proper consideration is afforded to object biography and potential contaminants, whether natural or anthropogenic. It thus offers an important corrective to past criticisms of ORA that have dismissed the utility of such artifacts. The companion volumes *Minoans and Mycenaeans: Flavours of Their Times* [54] and *Archaeology Meets Science: Biomolecular Investigations in Bronze Age Greece* [55], two of the most prominent ORA publications in Mediterranean studies, were particularly influential in conveying this message. These collaborative volumes, which assembled contributions from seven different laboratory groups utilizing wildly varying extraction methodologies and instrumentation (with no evident coordination or standardization), allowed for easy criticism of ORA, and especially of ORA results from legacy objects because of their greater risk of exposure to anthropogenic and environmental contaminants [56,57]. The volumes shed welcome light on Bronze Age organic commodities and drew necessary attention to potentially problematic contaminants, but included unqualified critiques that essentially encouraged a "baby-with-the-bathwater" dismissal of ORA, casting doubt on the viability of ORA studies in general and legacy objects in particular.

It is certainly clear that better ORA results are usually obtained from artifacts recovered from ongoing, or at least recent, excavations [5,6,13,18,58], if only because sampling conditions and object biographies can be better controlled and understood. Yet the case study presented herein showcases a legacy object that in many respects reproduces a result obtained from a freshly-excavated object, thus demonstrating the inherent value of this class. The compounds in legacy objects may appear in diminished quantities—both in number and peak strength—relative to modern standards or freshly-excavated examples, but if properly scaled the peaks are discernible and appear to offer meaningful indicators of the presence and proportions of compounds, as the RA of linalool demonstrates (see above).

Insofar as contaminants are concerned, although there is no foolproof methodology that can guarantee that results from a legacy vessel—or indeed any object—will be free from contamination, our work does point towards a few steps that can be taken to better ensure a sample's reliability and viability, even in circumstances where the full object biography is unknown. Foremost among these is an ability to predict, identify, and screen for common contaminants, such as plasticizers (e.g., phthalates, which are not naturally occurring and can reflect the storage of an object in a plastic bag), as was the case with portions of the Mochlos vat spout. In addition, it may be possible to screen for impurities that are most likely to be present on the outermost oxidized surface of the organic residues, through the use of multiple extractions [1], and by consideration of the geological profile of the clay itself [14]. This latter, too, is a process presently under study by the OpenARCHEM project at the Massachusetts Institute of Technology Center for Materials Research in Archaeology and Ethnology. Most important, it is essential that the condition, quality, and storage of any sample, whether taken from a newly excavated or legacy object, be fully described and acknowledged in any publication. With transparency and proper consideration of object biographies and degradation processes, legacy objects can indeed yield valuable and verifiable results.

Such results are most meaningful when they can be placed in conversation with new samples, or considered as constituents within a larger ecosystem of chemically comparable or contemporary samples. This is often difficult: ORA studies tend to be siloed, published singly in various specialized outlets or buried in archaeological site reports, preventing aggregate study. It is clear that the field would benefit from a common repository that might best facilitate the comparison and sharing of ORA results. The OpenARCHEM archaeometric database [59] was created specifically in response to this challenge. Using the original ARCHEM project library of chromatographic data as its core, which includes over 5000 ORA samples, the OpenARCHEM database is designed to be a platform by which scholars of antiquity across many disciplines and regions can review, compare, and integrate ORA into their own work. Each entry presents the results from a single extraction, and includes not only diagnostic compounds with their GC-MS chromatograms and links to NIST reference chromatograms, but also information about the type of artifact sampled, its quality, condition, and storage history, as well as links to ethnohistorical textual sources that discuss the botanical species identified.

The inclusion of these different categories of information offers routes by which the scientist, the social scientist, and the humanist can ask different kinds of questions of organic data—the scientist searching by chemical compound, for example, or the humanist by ancient author. It also allows for focused comparisons, whether between samples from similar jar types or by those with comparable object histories, thus aiding in further refinement of archaeometric methodologies. Through such data-sharing, chemical signatures that went undetected or unrecognized in early studies may become identifiable (as was the case with our cinnamic acid at Mochlos), and larger patterns, such as connections between contents and their containers, can become visible. Collaborative repositories and archaeometric communities such as these tap into the promise of big data [59,60], and are a resource by which ORA can be engaged in service to larger archaeological questions.
