For two days Patrick Degryse (archaeometrist) and Nathalie Kellens and Tijl Vereenooghe (both KULeuven doctoral researchers, the first of whom studies all our metal objects) went back to Tekeli Tepe, an ancient ore-extraction site on the territory of Sagalassos. In the past, a detailed geochemical prospection had isolated this location as an iron-production area. There, the olistostrome deposits of the Bey Daglari Mountains between the valleys of Canakli and Aglasun, five km south of the ancient city of Sagalassos, show a mineralization of magnetite. Many traces of iron production have been found in the past. Ceramic finds date these activities to the early Byzantine period. This time the geophysics team of Branko Mui accompanied our KULeuven team.

Last year, samples of the ore deposit and waste products of ore smelting and processing activities were sampled. Through chemical analysis of lead and strontium isotopes, a possible relation between the different types of ore, waste from ironworking sampled at Sagalassos and in its territory, and iron objects from the city excavations was investigated. Moreover, sampling of soil sediments around the Bey Daglari massif was performed, as a placer deposit of iron minerals was identified.

During previous campaigns, an idea of the nature of the activities and their chronology was established on the basis of surface material, but the actual spatial organization of the metallurgical activities in the Bey Daglari area (i.e. the subsurface structures) remained the subject of broader investigation. Therefore, one of the aims of the 2005 survey season was to introduce geophysical research in the area to gain spatial information on the presumed metal furnace installations and slag dumps.

At the sites associated with iron-production activities, the magnetic method is the most efficient. Thus we surveyed an area of 2,500 m2 for archaeological features related to craftwork activities. Some areas were resurveyed by georadar. This technique is not the most obvious geophysical method for detecting features such as furnaces or forges. But we used it alongside the magnetic method to provide additional data for interpretation of magnetic anomalies. The magnetometry are very clear and enable an archaeological interpretation of the site. In area 1, we discovered two very strong anomalies one weaker between them. The larger features may be furnaces and the smaller a forge. The wide east-west band of strong magnetic anomalies is perhaps a dumping area. The clear bipolar shape of the magnetic anomaly at the western edge of this area (area 2) may result from an archaeological object still in situ. Perhaps the east-west band was a blacksmith's dumping and the object itself a forge. Farther north, other evidence of the presence of furnaces or forges was obtained. So far, the iron-production area is clearly delimited only toward the south. More measurements are needed to guage the whole area's potential.

To study the technology of ironworking, all waste products from the geophysically surveyed grids were morphologically classified by means of visual inspection. In a second stage, the quantification of smithing activities was completed by means of counting and weighing fragments per survey grid. In general, as is also the case for Sagalassos, archaeometallurgical structures (workshops, hearths, working floors, anvil pits, storage pits, etc.) are poorly preserved. Frequently, the blacksmith himself reorganized his workshop, destroying the hearth to rebuild a new one in the same place. Also, smithing tools (hammers, anvils, tongs, punches, pliers, etc) are rarely found in direct connection with a workshop. On the other hand, smithy and smelting debris are common on archaeological sites, especially slag, which is very well preserved because it is difficult to recycle and little altered by burial. Therefore, the study of slag and other waste products provides a reasonable, if incomplete, idea of the activity at a smithy or a smelting site. Research (Vincent Serneels, University of Fribourg, Switzerland) has shown that morphological characteristics of smithing slags are related to the design of the hearth (position of the tuyère in relation to the bottom floor) and the airflow rate (stronger/hotter; weaker/cooler). Also shape, aspect (color, porosity), physical properties (magnetism, apparent specific gravity), and mineralogical and chemical compositions are linked to a certain extent. The nature of the material is mainly the result of the way according to which the work was done in the fire. The proportions of the different materials reflect the type of work done by the smith. On the other hand, the dimensions (size and weight) seem to be independent of the other characteristics. The weight of the pieces is not related to the type of work, but to the amount of work. Using a classification based on materials combined with the distribution of weight, it is possible to characterize the activity of a given workshop through its slag assemblage.

The classification and quantification of the Tekeli Tepe smelting and smithing debris boil down to a more detailed picture of the organisation and nature of the activities in the area. Here, we are dealing with ore extraction and smelting activities and, likely, primary smithing of blooms to produce bars and billets. These were to be transported to blacksmith workshops in the broader area. Isotopic analysis of the ore in this area and a smithing slag fragment excavated in a late Roman-early Byzantine blacksmith workshop in the Sagalassos city center has shown a close correlation. Therefore, we know that at least during the sixth and seventh centuries A.D., some economical transactions took place between this site and the city centre. To check the hypothesis for the bloomery smithing at Tekeli Tepe, we sampled fully preserved smithing cakes from this area to be sent to Belgium for detailed morphological study of the slag profile.