We employ a two-phase process that is designed to provide both a broad overview of the study area (Phase 1) and more detailed information about areas with higher densities of cultural material or locations where we have found artefacts of particular interest (Phase 2).
The initial phase of our survey was developed to provide a systematic sample of a large contiguous block of the survey area (c. 24 ha), centred on the chert outcrops at the top of Stélida Hill. Working in pairs we ran our survey transect lines along the cardinal directions following the site grid, with all artefacts collected from a 1m2 area every 10m along those lines (Figure 1).
Figure 1: Plan of transect survey lines and grid units surveyed in 2013.
Transects were initially established at 40m intervals, and in some cases leapfrogged out further to roughly locate the outer limits of the area(s) containing dense scatters of lithic material. In 2013 we were able to walk 20 survey lines that ranged from 170 – 530m long (Figure 1). These lines and collection points were established with tape and compass (Figure 2), while collection points were documented with recreation-grade GPS (Figure 3).
Figures 2 & 3: Kyle and Katie stringing out a 10m transect line interval; Katie geo-referencing her transect collection point with a GPS unit.
As we walk our transects and undertake our 1m2 collection units we systematically record a range of information that will help us critically review our survey’s artefact distribution data (Figure 4). This includes the degree of ground slope and the percentage of vegetation cover. If the ground we are surveying is very steep and is exposed to the elements, then we are likely to see a downslope movement of artefacts as gravity and erosion serves to displace the material from its original location. We have noted how terrace walls can often act as ‘artefact traps’, with accumulations of stone tools that have come to rest behind them. Conversely, dense vegetation usually renders the underlying artefacts invisible, whereby our transects might record a series of ‘zero finds’, when in fact there maybe workshop debris beneath.
Figure 4 & 5: Dora filling out a transect collection point recording sheet; Transect collection point in middle of Stélida bush (with historic inevitability).
These and other factors have to be taken into consideration when we come to look at our artefact distribution maps; our data cannot be taken at face value. Some ‘hot spots’ might be nothing more than a mass of artefacts that have accumulated there through processes of gravity and erosion, while some of our low-density areas might have more to do with the intensity of undergrowth than an absence of archaeology.
We also write a running commentary on the ground and vegetation between collection points, and take photographs of each transect stage and each collection point, the latter producing the largest visual database of Naxian bushes in existence... (Figure 4).
The second phase of our survey was designed to generate larger samples of diagnostic artefacts from areas identified as of particular interest. These loci were systematically sampled by either undertaking targeted 1m2 total-artefact collections, or by laying a grid over the area (Figure 5-6), within which we took a number of standardised 1m2 collections.
We established ten of these grids, ranging from 3×3m - 75×80m, the latter located in front of Rockshelter A (Figure 6 or 7). Grids were established with tape-and-compass or a TOPCON total station (Figure 7 or 8), while collection points were recorded with both the total station and GPS.
Figure 6 & 7: Corner of a grid unit laid over artifact rich locus; Sean laying out a survey grid with TOPCON total station.