For the first time the Ljubljana geophysics team directed by Branko Music and composed of Igor Medaric, Jurij Soklic, and Lucija Soberl, ventured out into the Sagalassian countryside, guided by H. Vanhaverbeke and her three archaeological surveyors Nele Goeminne, Lies Vercauteren, and Kim Vyncke (all KULeuven). They spent three days on the suspected sites of suburban villas identified during previous archaeological surveys south and southeast of the city by the high density of tiles or brick and pottery, together with fragments of window glass, mosaic tesserae, marble veneer (crustae), and in some cases hypocaust tiles (for floor heating). We wondered whether these surface concentrations coincided with remains of luxury constructions below the soil. The geophysicists used both a magnetometer (to identify anomalies in the earth's magnetic field caused by differences in the magnetic properties of archaeological materials, especially burnt clay, and surrounding soil) and georadar (to determine the depth and geometry of archaeological objects).

The field with the presumed villa site of Çatal Oluk was recently cultivated and the topsoil plowed, so one could expect badly preserved archaeological architectural remains. Because geophysical methods are complementary, at least two had to be used. The resistivity method, which is frequently used with magnetometry, could not be applied because of dry topsoil. Thus magnetic and georadar methods were combined. On the magnetogram many anomalies of different origin could be seen. The strongest pointed to magnetization of brick and tiles. Because of their magnitude and shape, some anomalies of this type should be interpreted as kilns. Several strong or very weak magnetic anomalies were also observed and separated into two different classes: those resulting from modern land use and those that are possibly of archaeological origin. These results did indeed yield traces of a large stone building, consisting of several rooms, which are apparently filled with tiles and brick. Adjacent to it, are a number of kilns. On the surface, this area coincides with a dense spread of tiles and bricks, as well as pottery and window glass. It seems safe to assume that a major building (villa?) was located here that also produced pottery, brick or tiles. In the area south of this concentration, the outlines of several more buildings are evident through geophysical prospection. However, very few tiles or bricks and sherds were collected on the surface. Further analysis may indicate whether these buildings are possibly earlier or later than the Roman period, or whether they represent more vernacular parts of the villa complex.

The second possible villa site of Gökpinar was apparent on the surface as both the in situ remains of a monumental tomb with two beautifully executed sarcophagus lids, as well as a ca. 50 by 50 m area full of fieldstones and depressions, which may be the only remains of a large structure. We surveyed an area 20 by 25 m on the hill where these remains are located. The results of the magnetometry were very confused, because of the high number of stone blocks on and near the surface. The processing of the georadar data is under way. Surprisingly, the first results are very promising. Linear high-amplitude echoes characteristic for a response to architectural remains can be easily traced over the whole surveyed area. In fields west of the hill, 1,200 m2 were surveyed by magnetometry and 1,600 m2 with both methods. On the magnetogram, at the western edge of surveyed area a strong north-south linear anomaly is clearly discernable. From the shape and magnitude of the magnetization, it must be interpreted as a thick limestone wall. The same wall, aside of some other linear anomalies of badly preserved architectural remains, can be observed also on the mapped view with the georadar results.

In the area of the last presumed villa site, Akyamac, which was characterized by a very high amount of pottery but only few tiles on the surface, 2,800 m2 were surveyed by magnetometry and georadar. Because of the high magnetic background of the bedrock, the results of the magnetic prospection are difficult to interpret. There seem to be a number of walls in the subsoil, at least at the western part of surveyed area, where the background is magnetically quieter. We will have to await the results of final processing of the georadar data to be absolutely sure. After the initial georadar data processing these expectation were partly confirmed. Some of depicted linear anomalies are indeed of archaeological origin.