As in 2011, this year's activities concentrated on the renewed excavation of the Natsargora mound (Fig. 1). The aims of the season were: 1) to reach the virgin soil in the Western part of the excavation by excavating the earliest of the Kura-Araxes sequence of surfaces discovered in 2011; 2) to reach the EBA layers in its E part, where a sounding dug in 2011 had shown that the virgin soil was lying more than 1.50 m deeper, but where last year's excavation had not proceeded beyond the LBA levels; and 3) to better clarify the nature of the earliest LBA occupation at the Northern limit of the excavation, where, in quadrant 100.100c, the presence of a series of well preserved superimposed Late Bronze surfaces had been observed just under the present top soil.
    For the latter aim, a new 5 x 5 m quadrant (100.100a) was opened at the Northern limit of the area excavated in 2011. Here, we excavated a sequence of four different surfaces of yellowish compacted silt material, which extended over an area of ca 1.5 x 2.5 m oriented in NE-SW direction (Fig. 2). This was flanked on both sides by an alignment of pits (4 of them were excavated on each side), which roughly respected the limits of the yellowish surface. Other pits were however cut inside of the surface, and were covered by its later remakings. At the E limit of the quadrant, a badly preserved mud-brick wall of yellowish colour (locus 0466) appeared to mark the limit of the just described area (most probably an external space) with what might have been a roofed building. Excavation in quadrant 100.100a stopped after dismantling the earliest of this series of surfaces.
    The rest of the investigated area was the same as in the 2011 season (a ca 200 square meters area, corresponding to eight 5 x 5 m quadrants, located on the present top of the mound, South and East of Ramishvili’s EBA soundings), where we continued to follow the sequence of anthropic layers from the level reached last year. As we had already noticed last year, the EBA layers, originally ca 0.5 m thick, had been intensely reworked and disrupted ab antiquo by the Late Bronze Age people, who had excavated a very large number of subcylindrical pits, 0.5 to 1.5 m wide and up to 1 m deep; these often cut other previously refilled pits, so that the underlying sediments were almost completely destroyed throughout wide areas, where only few remains of the older features are preserved. Contrary to our hopes, this process had not only affected the Western part of the excavation, but also its Eastern part, where we had hoped to find better preserved remains of the EBA occupation. As a matter of fact, it appears that the top of the natural hill on which the Natsargora settlement had been founded at the end of the 4th millennium B.C. was heavily sloping not only toward the East, but also toward the South. As a consequence of this, the Southern part of our excavation (quadrants 099.099a, b, 099.100a, b) was lying at the settlement's periphery, in an area which was marked from the beginning by a series of irregular steps oriented roughly in NE-SW direction. These steps had been repeatedly re-shaped during the Early Bronze period, as shown by the fact that the yellowish silty surfaces followed their slope. At a later time, these steps were cut into by the Late Bronze inhabitants. These dug a number of large pits which approximately followed their outline, and marked the outer limits of the contemporary settlement.
    As for the Early Bronze occupation in this part of the settlement, it exclusively consisted of a sequence of open-air surfaces: similar to those discovered last year, most of these were plastered with a fine layer of yellowish silty soil, but some also showed a coarser paving of small sparse stones and pottery sherds. They were occupied by different types of installations, most of which were obviously connected with fire (hearths and fireplaces of different type, ash-pits, burnt areas, etc.), had been frequently renewed, and often cut into one another (Fig. 3). At the western limit of the excavated area, in quadrant 099.099a, the earliest of these open-air surfaces, which was lying directly on the virgin soil, was partially exposed at the end of the excavation. It was coated with a fine whitish plaster, and was covered by a layer of black burnt soil which contained a large amount of seeds and burnt vegetal matter. In the Eastern part of the excavation, we exposed the latest of a sequence of surfaces which showed comparable features, but were heavily damaged by the Late Bronze pits.
    From the point of view of chronology, it is important to stress that no Bedeni occupation layers were met with in this part of the excavation, where, as we already supposed last year, the Early Bronze layers were of exclusively early Kura-Araxes (Kura-Araxes II) date, and were overlain by ca 50 cm of almost sterile soil, which accumulated during the period of abandonment preceding the re-occupation of the site at the beginning of the Late Bronze Age. However, in quadrant 099.100a we discovered another pit which contained a complete Bedeni pot (Fig. 4) and fragments of other contemporary vessels. This confirms our hypothesis that the abundant Bedeni material uncovered by Ramishvili in the 1980s mainly derived from similar pits.
    Proper EBA dwelling structures appear to have been limited to the Northern part of our excavation area, where last year's excavation had unearthed the remains of a few more substantial structures, tentatively interpreted as huts. This year's excavation in quadrants 100.099b, d, 100.100c, d confirmed this impression, but also showed that the Early Bronze layers here had been very deeply affected by Late Bronze disturbances, to the point that material found in these areas was still mixed by later intrusions. For this reason, the date of the hut discovered in 2011 in quadrants 100.099b and d, which we had tentatively attributed to the Kura-Araxes period, is now less certain: the presence of a few Bedeni sherds in the area might suggest an Early Kurgan period date for this structure, but it must be admitted that at least one Late Bronze sherd was also found in the same context.
    The underlying layer, however, appeared to be of definitely Kura-Araxes date. In spite of its poor preservation due to the Late Bronze pits which disturbed it, it yielded the remains of a small rectangular room (locus 0469) with yellowish-greyish mud-brick walls, ca 35 cm wide, of rather poor quality. The room measured 140 x 150 cm, and was provided with two successive floors, both consisting of a smooth surface of yellowish silt. Each of the floors was associated with a fireplace, located along the Southern limit of the room. The later fireplace had partially cut into the earlier one, and had damaged it.     A mud-brick wall, or platform (locus 0735) was also discovered in adjacent quadrant 100.100c, along with a pisè structure of rounded shape (locus 0736) (Fig. 5) No floor was associated with these structures, and their function remains unclear. The area between them was heavily disturbed by the Late Bronze pits, and only small parts of the original EBA layer were preserved. Excavation in the area stopped when we hit the top of a thick layer of ashes, which was clearly earlier than the layer to which these structures belonged.
     The 2012 excavation season yielded abundant finds of both the Early and the Late Bronze periods, unfortunately most of them in rather fragmentary conditions and often from mixed contexts (as we could ascertain, Late Bronze pits contained up to 70% of Kura-Araxes material). This makes an evaluation of those materials (e.g. animal bones, but also informal lithic tools), which cannot be dated on typological grounds especially problematic. As for pottery, it is clear that by far the largest component of the EBA pottery assemblage is rather homogeneous, and consists of KA II types. This suggests that the site's Kura-Araxes occupation spanned a rather short period of time. The high number of successive floor levels (at least six) encountered in the Southern sector of the excavation does not contradict this hypothesis, since all uncovered structures appear to have been rather ephemeral in nature, and to have been repeatedly re-built every few years.
    The Kura-Araxes II pottery are accompanied by only a handful of Late Kura Araxes, Martqhopi (?) and Bedeni sherds, mainly coming from Late Bronze pits. This points to a very sporadic presence of these later EBA cultures at the site. As for the Late Bronze, ceramic material from the pits confirms the presence a long occupational sequence of this period, as already showed by Ramishvili's investigations. In fact, this material includes both early Late Bronze types (e.g. closed vessels with black surfaces provided with pattern-burnished decoration), which appear to mark the very beginning of the period, and late Late Bronze types (e.g. closed vessels with incised/impressed and combed decoration) which continue into the Early Iron Age.
     An important find from the 2012 campaign is represented by three additional fragments of cultic reliefs similar to those discovered in Ramishvili's excavation (Fig. 6): one of them was found in a Late Bronze pit, but two of them came from good Early Bronze contexts. It is certain that these fragments belong to more than one (probably at least three) reliefs, which showed almost identical features (obsidian-inlaid eyes, spiked top, central raised ridge, etc.). The fact that comparable reliefs have not been hitherto found at any other contemporary sites raises intriguing questions about the religious beliefs of the early inhabitants of Natsargora.
     To conclude, the 2012 excavation season has reached the aims to clarify the nature of the Early Bronze settlement on the Natsargora mound and its chronology, especially as far as the relation between the Kura-Araxes and the Bedeni occupation at the site is concerned. Considering the general homogeneity of the uncovered structures all over the excavated area, the overall limited depth of the Early Bronze levels, and their rather poor state of preservation due to later disturbances and previous excavations, it is not considered worth to continue excavating those parts of the site in which virgin soil has not been reached.