The Thesprotia Expedition

2004 - 2005


Classical and
Roman Times

Medieval and
later finds














A Regional, Interdisciplinary Survey Project

in Northwestern Greece

The preliminary results of the first two seasons of the Thesprotia Expedition are as follows.

I. Archaeological survey

The Kokytos Valley in Classical and Roman Times

Apart from the site at the foot of Agios Donatos of Zervochori just mentioned (PS 17), we have found no Iron Age or Archaic-period remains. Recent finds by the 8th Ephorate of Prehistoric and Classical Archaeology do show, however, that the valley was inhabited during these periods as well, thus raising a question, to be addressed in coming field seasons, of why our surface survey cannot identify relevant finds.

Agios Donatos of Zervochori (PS 25) is the most important of the Late Classical to Early Roman sites we have studied. In 2005, we collected information concerning the site and its fortifications, studying the structure of the acropolis walls, which probably date to the fourth or third century BC, and drawing a detailed ground plan. This investigation revealed two gates and a square tower with room for catapults on the ground floor. The remains of large terrace walls and rock-cut foundations of houses as well as several architectural blocks – parts of a triglyph-metope frieze, column drums, orthostate blocks, etc. – were recorded, together with a terracotta antefix. Large amounts of Italian terra sigillata indicate that the site continued in use even after the Roman conquest of Thesprotia in 167 B.C.

Thus far, we have studied three further sites dating to the Classical-Hellenistic period. The first one is part of the same small, sprawling, unfortified hamlet that had been partly excavated by the 8th Ephorate to reveal houses, graves, and a small sanctuary. This hamlet is located at the very foot of Liminari hill, exactly where the valley floor meets the slope of the hill. Two fields (PS 5 & PS 7) adjacent to the excavated areas were surveyed by intensive fieldwalking, revealing what seems to be the location of an additional house. Another site of Late Classical to Early Hellenistic date (PS 15), where illegal digging had taken place, was detected some way up on the lower slopes at the village of Sevasto. In order to save the site from further destruction, the central part of it was immediately excavated by the local Ephorate. These excavations revealed a large house dating to the fourth to the third centuries BC, but it is unclear whether other buildings existed in the vicinity of the excavated house or whether it should be considered an isolated farmstead.

The only signs of human use on the valley floor during the Classical and Hellenistic periods are connected with graves. Typical for all of Thesprotia are rich grave heroa: the most conspicuous in the Kokytos valley is the monumental hypaethral heroon at Marmara that can be dated to ca. 250 - 200 BC. Large orthostate blocks and cover slabs of graves that seem to belong to a similar grave heroon (PS 13) were found in a location akin to that of the heroon of Marmara, i.e. in the plain about half a kilometre east of the Kokytos river. The existence of these types of grave heroa, in combination with the apparent lack of large monumental temples, may reveal the existence of different social and political structures in Thesprotia from those in Greece proper.

In the Late Roman period, the settlement pattern changes drastically. Many new small sites appear on the lowest-lying areas of the valley floor. We have studied five such sites. One seems to be a sprawling hamlet or village (PS 27) next to a previously known and excavated Early Christian basilica. The remaining four are isolated farmsteads or villas, the richest one being PS 14. This site produced imported amphorae and much fine glass; according to the plot’s he had previously found sections of terracotta water piping. Another one of these small farmsteads was excavated by the local Ephorate after we finished surveying it (PS 16). It revealed only the remnants of a single, robbed Late Roman grave, but no remains of the farmstead itself.

Several other Early Christian basilicas and smaller Late Roman settlements on the valley floor have been identified through excavations, presenting the picture of a rather densely settled landscape. On the other hand, it seems that, for instance, the settlement at Agios Donatos of Zervochori had been abandoned at the time when the sites on the valley floor emerged, indicating that we may be dealing with people moving from the mountain slopes down to the plain. Thus far, the reasons for this shift in settlement are unclear; the same goes for the relationship between the sites on the plain and the large Roman colony of Photike located just to the north of our survey area.  

The Prehistory of the Kokytos Valley

Medieval and later finds