The plan is to wrap up a draft of the Thisvi paper by the end of today, and it looks vaguely possible. This weekend, I ran a bunch of queries on the finds data to attempt to determine the relationship between the ancient and post-ancient material on the site. As our Archaeological Institute of America panel is supposed to focus on the post-Classical levels at well known sites, then it seemed better to focus on the post-Classical material from Thisvi (and ignore, mostly, the idea that surface material, no matter what the chronology is always "first out").
The first step to my chronological analysis was simply to look at the distribution map generated by plotting the artifacts by period across the sites and known transects. I've put up versions of these maps before in a slightly modified form. The maps below include data from the more intensively collected sites (these are dots that do not appear in any survey transect) and in the circular collection units surveyed in the first year of fieldwork near the Hellenistic walls of Thisvi. Each dot represents one artifact. Their location within survey transect is arbitrary.
The first map (blue dots) represents Classical-Hellenistic material, the second (red dots) Late Roman material, and the third (gold dots) Byzantine-Medieval material. They clearly show that by the Late Roman period, a significant contraction in the distribution of material occurred around the city of Thisvi. The southern slope of the basin which were quite a busy place in the Classical to Hellenistic period appear to be used far less intensively in subsequent periods. This seems to represent an overall contraction in the intensive activity areas of the city of Thisvi and parallels to a certain extent the results of the survey at Thespiai to the east.
Unfortunately, the maps which appear above are incomplete. I have not yet been able to plot several of the transects from the western most area of the survey (Area C in the map below).
While I think that there is a good chance that I'll be able to place these survey transects in the future (with the help of aerial photographs), at present the best I could do for the purpose of analysis is to compare the distribution of material in each of these sections to determine whether the distributional maps show a contraction of activity or simply a shift in the main area of activity from one part of the region to another. These charts are based on the almost convincing assumption that the total sample of each area is roughly equivalent and thus the proportions of various types of material is roughly comparable.
A first glance it would appear that Area E and Area C produced substantial more post-ancient material than either Area A or Area D. This is large due to two significant Late Roman sites in each of these areas. In the case of area C, the significant concentration of Late Roman material at the beach at Vathy accounted for close to 7% of the overall percentage of Late Roman material from the area. In Area E, the result was even more dramatic with a single site (E1) producing close to 20% of the Late Roman (and post-Classical material). Eliminating these concentrations, however, produces a fairly even distribution of post-Classical material across the entire survey area ranging from 7% in Area A to just over 13% in Area E. As a result, I feel comfortable stating that the distribution of what an earlier generation of survey archaeologists might call "off-site scatter" is relatively consistent across the entire survey area. This is significant because at the site of Thespiai to the east, the survey team has argued that most of the material in the fields around Thespiai was deposited as the residents of the city spread manure (and broken bits of pottery discarded in trash piles) to fertilize crops. Thus the distribution of "off-site material" could reflect the intensity of agricultural activity in the basin and the density of settlement at the central site of the survey area, Thisvi itself.
For more on this research: