Abstract:
At the beginning of the Early Iron Age in Troia, in the settlement phase VIIb2 (ca. 1100-1000 B.C.), new vessel forms (Knobbed ware) of hand-made pottery appeared. The same vessel forms, with similar decoration can be found in several archaeological localities in Turkish Thrace and the Balkans (Bulgaria, SE-Romania). In this geoarchaeological work the primary question is whether there is a possible common origin of the pottery from different regions? Was the Knobbed ware in Troia transported from the Balkans, or do we have to do it with transfer of technological solutions, and the vessels were made on the spot from local material? Moreover, which raw materials and technique (firing temperatures, etc.) were used in producing the ceramics?
In order to answer these questions samples of 90 ceramic fragments and 69 comparative geological materials have been collected from the most important archaeological sites and survey areas of the occurrence of the Knobbed ware: Troia, the Avsa Island, Menekse Catagi, north Turkish Thrace in Turkey, and Chal, Ovcarovo, Diadovo, Pshenitsevo, Kirilovo, and Sborianovo in Bulgaria. A complex analytical procedure employed in this study included optical microscopy (petrographic modal and cathodoluminescence), chemical (XRF, NAA, EMPA) radiogenic isotopes of 87Sr/86Sr and 143Nd/144Nd, and XRD (X-ray diffractometry) analyses.
In the first step the sherds have been grouped into eight petrographic groups and several subgroups based on their non-plastic inclusion composition.
Using the major and trace elements Mg, K, Rb, Ba, Sr, Nb, Ti, Y, Zr, Fe, V, Cr and Ni, bivariate plots and PAAS-normalised multi-element diagrams were edited in order to determine the local sediment patterns of the different areas based on clayey and sandy sediments (local reference materials), sampled near the archaeological sites.
Comparing the PAAS-normalised XRF data from the sherds to those of the local sediment patterns, using petrographic data as well as geological information of the different areas, I can conclude that most sherds seem to have affinities to local sediments in Troia and most archaeological sites discussed in this study. With the help of radiogenic isotopes I could chemically differentiate some more production locations, which could not be exactly identified using the XRF data. Neutron activation and electron microprobe analyses were used to analyse single rock or mineral fragments found in the sherds in order to compare their chemical properties to the results of former geochemical studies. Thus the results could be directly applied to those rocks and several effects (e.g. firing, alteration of a pottery), which might influence the results of the provenance analysis could be excluded.
The results of the provenance analysis, corresponding with several former archaeological studies, show that petrographic and chemical signals of 82 out of 90 vessel fragments fit well the natural properties of possible local raw materials at the archaeological places of this study, indicating that these potteries were, with high probabilty, locally produced. Three of six sherds from south Bulgaria show clear affinities with north Bulgarian vessel fragments and sediments, thus I assume that their place of origin was north Bulgaria and they were transported from there to the South. Complex archaeometrical analyses proved that five other fragments from north Turkish Thrace and Menekse Catagi were not locally produced. In most cases, due to the absence of the comparative chemical and petrographic data, I could not determine the exact place of origin of the analysed sherds unequivocally, but several places could be excluded or, on the contrary, taken into consideration. One of largest problems posed in this work was to filter the chemical influence of temper on chemical and isotope geochemical composition. I could show that sherds containing non-plastic inclusions or added temper rich in trace elements (e.g. volcanites, magmatites), can influence the whole chemical composition of the (clayey) raw material of the sherds, and thus complicate the interpretation of the data. Consequently, in case of coarse wares, only the combined use of petrographic and geochemical analyses can give the effective provenance.
On the basis of microscopic textural analysis I can establish that raw materials used for the ceramic production were mainly silty-sandy clays and/or river sediments. The use of very fine-grained clays or levigated materials can be excluded. By means of the clay mineral content (XRD data) analysed sherds were grouped into three groups, indicating each one different, but relatively low maximum firing temperature: in case of chlorite content: (550) 600-650 ºC, 10Å phillosilicate and calcite content: 700-750 ºC, and 10Å phillosilicate content: 800-850 ºC. Firing-cooling procedure took place mostly under reducing atmosphere, the fact reflected by the greyish-black colour of the samples.