Over six weeks spanning late May to early July, the Stélida Naxos Archaeological Project [SNAP] undertook its third excavation season of the early prehistoric stone tool manufacturing site and chert source on the island’s northwest coast. Under the joint directorship of Dr. Tristan Carter (Associate Professor, McMaster University), and Dr. Demetris Athanasoulis (Director, Cycladic Ephorate of Antiquities), the excavation and apotheke-based studies were undertaken by a team of 30+ young scholars, graduate students and undergraduates from Canada, Greece, England, France, Germany, Serbia and the US.

The site, first documented in 1981 during a survey the École Française D’Athènes, had initially represented something of an archaeological anomaly. While the mass of stone tools indicated that Stélida was clearly prehistoric, the material bore no resemblance to lithic tool kits that had hitherto been found in the Cyclades of later Neolithic and Early Bronze Age date. The site was tentatively dated to the Epi-Palaeolthic, but such a claim was highly contentious given that until quite recently the received wisdom was that no-one was meant to be living in the Mediterranean islands prior to the Neolithic and the advent of farming societies. The aim of our excavation has been to locate undisturbed, stratified archaeological deposits whose contents could be scientifically dated to help resolve this debate, and to contribute more generally to the rewriting of early Aegean prehistory in the context of proclaimed Palaeolithic discoveries in other insular locations, not least Crete and the Ionian islands.

This summer’s work involved the re-opening of trenches previously established on the hill’s western slopes, together with initiating a number of new sondages on the eastern flanks of Stélida. The logic as to where we excavated drew jointly upon an understanding of the local landscape, and the results of our 2013-14 survey. Concerning the former issue, two of our team’s geo-archaeologists - Takis Karkanas and Dan Contreras – had suggested areas where we were likely to find deep deposits (much of the hill is bare from downslope erosion), as for example the upper reaches of Plot DG-A, where an exposed lip of bedrock acted as a natural terrace, helping to keep soil in place. Here in 2015 we initiated trenches 1-4, two of which we had continued into this season. Our colleagues’ insight paid off handsomely, with Trench DG-A/01 eventually hitting natural after some 3.6m of stratified deposits, while Trench DG-A/03 remained unfinished by the end of the season at almost 4m deep!

These sondages provide us with a slice through deep-time, the different strata comprising an alternating sequence of landscape stability (palaeosoils) and major erosional events (colluvium). Given that the associated archaeological finds are of Palaeolithic date (see below), then by extent our different deposits likely relate to, and reflect the major climatic fluctuations of the Pleistocene, or ‘Ice Age’. One working hypothesis is that periods of major erosion may coincide with the onset of warmer times (‘interglacials’), the idea being that during the colder periods the vegetation on the hill would have been low, scrubby plants (think Canadian tundra) whose shallow roots would have provided little anchorage for the soil, whereby an onset of warmer climate greater precipitation would have resulted in significant downslope hillwash.

The archaeology of these deep sondages is challenging for a number of reasons. Firstly, given the depth of some trenches we face health and safety issues of hard-hats and shoring, and the use of long ladders to get into them, and ropes and pulleys to hoist the buckets of soil out of them. Secondly, they can be extremely productive, with some colluvial layers producing 20-30 large bags of finds, tens of thousands of artefacts that can overwhelm excavator and lithics’ specialists alike. Thirdly, the base geology of Stélida is highly alkaline which alas means that organics are rarely if ever preserved, which hiders significantly our ability to reconstruct the Palaeolithic environment and/or the subsistence practices of those who visited the site. Finally, and arguably the biggest challenge we face, the vast majority of the artefacts that we recover from the excavations come from secondary contexts, i.e. they have been redeposited from their original place of manufacture/use, usually during these episodes of hillwash erosion. Thus, while it is possible to produce good scientific dates from these deposits using the OSL technique (see below), we are not dating the artefacts themselves, but the date of the soil deposit within which they are found in, a layer that might have been produced much later in the Holocene. As such, we could have a context whose artefacts we believe to be Upper Palaeolithic on the basis of their form and manufacture, but the deposit itself was created only 3000 years ago during a period of erosion; it would be the latter date that our scientific methods would provide us.

There are however a few trenches that are providing us with some genuine in situ deposits that should allow us to date actual periods of activity, rather than generating a series of terminus ante quem determinations (‘these tools are at least date X’). As noted above, many of our sondages were established where we believed there to be deep soil deposits; a second strategy for excavation, was to target specific areas of interest, which include digging in front of a rock shelter, and opening areas where we found concentrations of interesting artefacts in the survey. Trench AK/18 was situated in front of a small rock shelter, the hope being that here we might actually find traces of seasonal habitation, rather than the ‘factory floor’ deposits we usually encounter, replete with the manufacturing debris of millennia. Having dug through over a metre of hillwash deposits we indeed came across evidence for in situ activity in the form of large ashy deposits that represent the remains of numerous fireplaces that would have been used for cooking and warmth by those visiting the chert source to make their tools before returning to their home camps elsewhere on Naxos or beyond. Preliminary studies of soil samples from these features by Charlotte Diffey of Oxford University have produced small quantities of carbonised botanical remains that should allow us to (a) reconstruct the environment of that period, and (b) provide us with samples for radiocarbon dating.

Concerning the finds themselves, 2017 was an important year for us, providing us with some key assemblages, not least a major group of Middle Palaeolithic finds from Trench DG-A/021, typical products of Neanderthal populations of the larger region. Significant quantities of Upper Palaeolithic material were found in every sondage, including what might be some blade cores of possible Aurignacian (early Upper Palaeolithic date), a period that many associate with the first appearance of Homo sapiens in Europe. We also have earlier material – albeit very eroded – that is conceivably early Middle Palaeolithic, if not Lower Palaeolithic, while the latest material continues to be Mesolithic, aside from a handful of post-Bronze Age sherds from close to the rock shelter.

Next year should be quieter, but no less engaging, as we segue from excavation mode to study season. This and last year we benefitted enormously from the input of some important visitors, not least Prof. Catherine Perlès (Paris X), Prof. Alan Simmons (UN Las Vegas), and Prof. Georgia Kourtessi-Philippakis (Athens), while conversations continue with the Municipality and Ephoreia with regard to hosting a small exhibition in Chora on Stélida in the coming years.

Tristan Carter
Associate Professor, McMaster University; co-director, Stélida Naxos Archaeological Project