CIPA 2003 XIX 11 ' International Symposium, 30 September — 04 October, 2003, Antalya, Turkey 
location where the present archaeological structures could be 
found. Actually, as over an inhomogeneous and/or anisotropic 
ground, the calculated resistivity will vary, data inversion will 
be necessary to get a real electric image of the ground (Ward, 
1990). In comparison to the magnetic prospecting, the 
instruments necessary for a geoelectric survey are decidedly less 
expensive. Nevertheless time required for this kind of 
investigation is much longer. 
2.1 Data acquisition 
As regards the magnetic measures, an optical pumping 
magnetometer (Geometries G858) in gradiometer configuration, 
was used. The sensors were set to a fixed distance equal to 1.2 
m and the inferior one to 0.25 m from the ground. In general 
single square areas of 30X30 m 2 were surveyed. Data were 
acquired in the bi-directional mode along profiles 0.5 m apart 
with an average sampling step of one datum/0.125 m. After 
interpolating operations regular high resolution grid were 
obtained with a 0.25X0.125 m cell. 
The electric profiles were performed by using a dipole-dipole 
array (AM=MN=NB=lm) and employing a modern computer 
assisted georesistivimeter (Net Sistemi EW 24/410). Each 
profile, 20 meters long, has a constant separation of 5 m from 
the others and yields a network of 105 experimental data 
points until a pseudo-depth of 4 m. 
2.2 Data processing and visualization 
In all the three magnetic surveys the row data were first 
visualized as 8 bit raster images to identify the presence of 
those unwelcome disturbances commonly known in literature 
with the name of spikes, stripes and zig-zag (e.g. Scollar et al., 
1990; Eder-Hinterleinter et al., 1996). 
Spikes are mainly due to the accidental failure of the sensors 
produced by thermal or mechanic shocks as well as the presence 
of ferrous modern material lost in the ground such as rods, 
wires, studs, etc.. This can cause some non systematic errors of 
measure shown in a magnetogram as isolated anomalies either 
positive or negative. For the de-spiking step a GESD method 
was applied to non interpolated data (Ciminale and Loddo, 
2001). Spikes were identified and flagged through a statistic 
comparison with other neighbouring data within opportunely 
chosen sub areas. Once having been labeled, they were removed 
from the data set. Stripes predominately occur in the 
magnetogram when data are collected in the bi-directional 
mode. To reduce striping effect each line was simply set at zero 
mean. Positioning errors, different lengths of parallel profiles 
and the technical specifications of the magnetometer often 
cause, especially in bi-directional surveys, a strong zigzag 
pattern. This displacement effect has been minimized through a 
procedure based on the computation of the cross-correlation 
function between the magnetic profiles considered three at a 
time (Ciminale and Loddo, 2001). 
After having attenuated these effects with DIP techniques, the 
data processing was completed through filtering operations in 
order to reduce the background noise present in the three 
surveys. Finally contrast stretching procedure helped in 
emphasizing the magnetic anomalies. 
Electric data were simply visualized as filled contour plot. In 
thus case it was considered not useful to apply any inversion or 
processing to the data because the depth of the searched 
structures was already known due to nearby excavations. 
Moreover the primary aim of the survey was the detection of the 
archaeological target for which a significant contrast in 
resistivity with the subsoil was expected. 
3. RESULTS AND DISCUSSIONS 
3.1 Kyme (Turkey) 
Figure 2. Part of the high-resolution magnetic mosaic obtained 
at Kyme. The light gray features are already excavated 
structures. 
Kyme, one of the most ancient cities of the Aegean coast of the 
Anatolia, was founded around the 1050 B.C. by people who 
come from the "Fricio Locrico" and assimilated the natives 
"Pelasgi" in the first city core (Strabone, XII, 4,3). Immediately 
after its foundation, Kyme became the center of trading in both 
the Aegean and the Eastern Mediterranean seas. Here Kyme 
established its most ancient colony, Cuma in Campania (Italy). 
Despite having already been discovered as far back as the 18 th 
century, Kyme became an object of study, albeit sporadic, only 
after the second half of the 19 th century. The University of 
Catania (Italy) has been systematically studying the 
archaeological site since the 1982 (Lagona, 1993). Annual 
excavations have brought to the light several ruins, among 
which the majority of the port area (IV B.C.), a theatre (I B.C. - 
180 A.D.), a defence wall and a tetrastyle temple devoted to 
Iside (IV A.D.) and a medieval citadel (XII-XIII A.D.). 
Nevertheless, the most part of the site is still completely 
unexplored while the excavations proceed only in summer 
months. The contribution that the archaeologists expected from 
the geophysical survey, was the identification of those areas 
with a more probable archaeological content. 
During two summers (1999-2000) 37000 m were surveyed. 
Because of the articulated morphology and the presence of 
natural and artificial obstacles (threes, dried walls, cultivations, 
pylons, etc...) it was impossible to investigate the site with 
continuity. The final result is the high-resolution magnetic 
mosaic (Ciminale, 2003) a part of which is shown in Figure 2. 
The data processing allowed to point out many anomalies with 
an archaeological meaning. Unfortunately the discontinuity of 
the survey affected the interpretation. Often in fact signals 
abruptly break off. Furthermore a background noise, due to the 
outstanding presence of potsherd and bricks spread into the