Finished 1993
Aim: Explore regional exchange networks and their change through time / change of political landscape
Method: Multivariate analysis (main factor / cluster) of stratified ceramic samples used as fill in subsequent building phases of the main religious space in Kaunos, SW Turkey.

PhD project by Hans-Dieter Bader,
Phillipps University Marburg, Germany

The investigation was part of the long running Turkish excavation of Kaunos with team members from Ankara and Erzurum Universities at the time.

Kaunos is a Greco-Roman settlement straight opposite Rhodes on the Turkish coast occupied from about 800 BC until about 800 AD when both harbours finally silted up. Parallel to an architectural investigation of the so-called temple terrace the ceramic of the fill of the terrace was looked at to explore regional trade.

The trade is composed of local trade between city and hinterland, long distance trade throughout the Eastern Mediterreanean and trade between regional centres. Little is known about the later and it was the focus of the research.

Four main layers were observed in various test trenches throughout the terrace. The lowest level seem to have accumulated as offerings around a large stone erected on a small spur. It is likely that the cult of the mythical founder of the city sprung up in the fourth or third century BC.

The area around this stone seem to have been extended and built up at least three more times to the current layer. Some of the layers are separated by gravel layers. The dryness of the soil/stone/ceramic fill made observation of the layers difficult.

Stamped amphora handles were within all layers and allowed absolute dating of the layers. The first layer around the stone was possibly built up as an offerings layer for over 200 years. The other layers seem to have been built up within the last century BC.

The level of this terrace wasn't change until late antiquity, but the buildings on it seem to have changed substantially over time. The last phase is characterised by a large church building.

Only the fill layers with ceramic in the mix could be used for the investigation. Within all test trenches approximately 120,000 sherds were found, of which only a small sample was analysed.

A first test sample from the various trenches and layers was used to built up comparative collections of clay types, temper types, decoration, colours in the break and colours on the surface.

These collections were used to describe a large sample from each of the various levels recovered from different test trenches.

At the same time vessel forms were recorded when possible. Due to the fragmented nature of the assemblage this vessel form compendium is most likely incomplete.

Main factor analysis was used to analyse which categories form a main factor for further analysis.

The main factor analysis showed that two main factors seem to influence the assemblage:

• choice of raw materials
• production steps

In the eastern Mediterranean common vessel forms and decoration systems were used during the Hellenistic period, thus production seem to have been similar for e.g. cooking ware or table ware with slip throughout the area.

But the combination of choice of raw material and type of production - as analysed using cluster analysis - shows specific groups. These groups are a mixture of different regional production locations, specialised production type and long distance trading.

A regional trading model is suggested whereby Rhodian production on the island together with local cooking ware and a small percentage of imported possibly Athenian table ware dominate the earlier phase. The later phase sees a change in the Rhodian production from the island as exported to Kaunos plus Rhodian production from the parea - their Mainland possession - dominate the market.

With the trading influence of Rome around 50 BC arriving in the area, long distance trading and a slipped table ware production in the region gain a larger foothold. This is also indicative of high quality table ware becoming more affordable for the wider public at the time.