During 2006 research in the cave concentrated on the Holocene levels with the aim of documenting the upper level of the stratigraphic sequence, and on the other hand extending the area relating to the hearth discovered in 2005, at the top of the late-glacial sequence in the level date to circa 12.000 years ago (C14 dates un-calibrated). A better record was made of the lowest part of the late-glacial sequence, dating to between 16.000 to 14.000 years ago, on the basis of the un-calibrated C14 dates. (MiBAC)

The research programme for this field season focused on square or box trenches (G8, H8, 17. 18, 110, J7, J8, J10, L9, L10). The continuation of the excavations of trenches 17, 18, J8 was limited. An external trench was also opened in the talus of the cave, next to that of 2005, called Hx1, and continuing the alignment of areas G and H. The two square metres excavated were called respectively Gx2 and Hx2. The enlargement of the area excavated on the exterior greatly increased our understanding of this part of the excavations, which previously was limited. A series of levels relating to artefact of the Palaeolithic period were revealed to a depth of -320cm. In particular, in the lower levels (Detritical end 4 and Detritical end 5) a series of elements a _cran_ (points and blades) came to light which allows us to attribute these complexes to the early Epigravettian period. No faunal remains were conserved, but charcoal was recovered which permitted radiocarbon dating. Inside the cave, the mesolithic levels relating to the sauveterriana industry, datable to the Early Holocene, and removed during the first excavation seasons, were identified only in a remaining layer of square G8. Most of the surface of the latter was disturbed by a large pit of the Neolithic period – the latter also affected the adjacent squares, and also cut through a series of underlying Pleistocene levels. In the rest of the cave (squares H8, 17, 18, 110, J7, J8, J10, L9, L10) the excavated deposits are characterized by artefacts of the epigravettian lithic industry. The majority of the sequence is of the Final Epigravettian period, but the presence of the Early Epigravettian in the talus may suggest that this type of industry is also present in the deepest levels of the cave, with an industry which is hardly abundant, and difficult to characterize within the Epigravettian period.

Excavations were carried out inside the cave of the Holocene levels of squares G8 and 17, and in the late Pleistocene levels of squares H8, H9, H10, 18, J7, J8, J10, K8, L8, L9 L10. A new box-trench or square was also opened, the sole square of row 11, called H11, at the mouth of the cave. As regards the _talus_ of the cave, on the exterior the excavations in Hx2 and Hx3 continued, and a new square was opened in a continuation of row G: il m2 investigated is Gx2. The excavation in row 8 confirmed the excellent state of preservation of the deposit in this area of the site, far from the disturbances caused by the vegetation on the exterior of the cave, and in particular by the roots which have intruded inside the cave. The fine and undisturbed sediments of levels PS1-PS6 have the clear imprints of branches and other vegetal elements, which have been slightly burnt leaving their forms fully visible. The excavation this year has brought to light a certain number of these, relating to level PS6. As regards this level, the abundant fish remains ( _Salmo trutta_ ) should also be noted: one example was still articulated. From the same level, but in square H10, a fine bone punch was also recovered. The investigations this year have enriched the archaeological documentation for all the excavated areas both as regards the lithic industry and the faunal remains, although the latter are almost totally lacking from the earliest levels. Viceversa, charcoal samples have been taken also in the latter levels, which are being radiocarbon dated. Finally we should note that, as well as some fragments of polished bone, some decorative elements been recovered: a _Columbella rustica_ from Holocene or post glacial levels, a small marine gastropod from late pleistocene levels of the cave, and a Denatalium fossil from the late pleistocene levels of the _talus_.

Research focused on the late Pleistocene levels within the cave (squares 17, G8, H8, 18, J8, K8, L8, M8, H9, 19, J9, L9, H10, I10, L10, H11, H12), as well as in those on the exterior of the talus (squares Ex2 and Fx2). Some unstable areas which had in part collapsed also made this necessary. The excavation of square H11, the outermost trench located at the entrance to the cave, was extended to the adjacent H12, which had been affected by a great collapse of rubble, in order to investigate where the latter rested, and also examine possible effects of ancient earthquakes. Great masses of collapsed rubble are also found along the slope beneath the cave, geologists specialized in the recent tectonics of the Abruzzo Appennines have been contacted for the study of this phenomenon. The general aim of the excavation was to obtain a better understanding of the area which had been repeatedly occupied during the late glacial period, and to try and correlate the internal levels of the cave with those brought to light in the external _talus_. The discovery of a point a _cran_ from the deposits within the cave allows us to confirm that the Early Epigravettian is present not only in the _talus_, but also in this part of the site, where until now only the Epigravettiano finale had been identified, followed by the ancient Mesolithic and rare traces of the Neolithic. The excavations this year also identified some habitation structures: a small hearth immediately overlying the collapsed levels found at the base of the archaeological fill recovered so far and therefore part of one of the earliest occupation phases of the cave; and a pit, relating to one of the latest levels of the late Pleistocene sequence. This can be interpreted as linked to cooking activities, and is similar to other examples recovered previously in another part of the cave, but in slightly earlier levels. During this research campaign the presence of rock art was discovered on the walls at the back of the cave: a small relief of a vulva, a second small vulva whose image was depicted by modifying a natural crack, and finally a deep but short horizontal groove. The plan and the recording of these elements will be completed in 2010. The preliminary observations suggest, on the basis of the typology of the images and their height, that they relate to the occupation phases of the industry of the Final Epigravettian, for which we have a radiocarbon date of around 12.000-13.000 bp (=calibrated date of c. 12.500-13.500 BC).

The research was undertaken on the Pleistocene levels in squares 17 (partial), J7, K7, L7, G8, H8, M8, N8, O8, G9 (partial), J9, K9, L9, H10, I10, K10, H11, H12, as well as in Dx2, in the external talus. In particular, work was renewed in row 7, the innermost excavated to date, which had not been touched for four years, following the taking of a resin cast of the surface of PS1 which came to light in K7-L7. These two squares were characterised by traces of burning and a particularly well-preserved vidange de foyer, which in this way was documented in three dimensional form. Overall, the aim was to improve knowledge of the area occupied during the Late Glacial and to attempt to correlate the internal levels of the cavern with those brought to light in the external talus. In H11-H12 work continued on the exposure of the support of a large collapse, similar to others along the entire slope below, which could correspond to a phase of tectonic activity during the Upper Pleistocene era. At present it can be stated that this collapse occurred before the formation of the latest levels of the final Epigravettian period, that is prior to 15,000 years ago (calibrated C14 dating). Thanks to the collaboration of Bradford University and funding from the Leverhulme Foundation, sampling of the complete late Pleistocene sequence was undertaken with the aim of identifying the presence of microtephra (microscopic volcanic ash). If found, these would then provide a very precise chronological correlation with the archaeological levels from other sites where they have been identified. As mentioned, in 2009 rock-art was indentified on the back wall of the cavern. Four different elements were documented, at a constant height of circa +30-+40 above point 0, at a distance of 2-3 m one from the other. Three of these elements were discovered during the 2009 excavations, the fourth was seen later. The latter was constituted by the modification of an angular piece rock, which had been worked by percussion, abrasion and incision to create a stylized female profile 75 mm high. The type is known from publications as “Venuses of the Gönnersdorf-Lalinde type”. These female figures, existing both as incisions and figurines, are dated throughout Europe to between 15,000 and 13,000 cal BP. The arrangement of the four elements found to date (one groove, two vulvas, and now the Gönnersdorf-Lalinde type silhouette) suggests that they could belong to a single chronological horizon. The latest discovery confirms what had been hypothesised in 2009, when it was suggested that the rock-art of Grotta di Pozzo could be placed in relationship with the levels of the final Epigravettian period, with a C14 dating of around 16-15,000 cal BP. The rock-art from Grotta di Pozzo was presented at the IFRAO “L\'art pléistocène dans le monde” 6 - 11 Septembre 2010, Ariège - Pyrénées, France. See the pdf file at http://ifraoariege2010.fr/

This research examined the Holocene levels in quadrants J6, K6 and L6, the Pleistocene levels in quadrants G9, H9, I9, I10, J10, K9, K10. Row 6. In J6 the surface layer disturbed by animal burrows and natural erosion was removed together with other layers from the Holocene sequence, thus reaching the Pleistocene level PS1. In K6 the same Holocene levels were exposed, as well as evidence of accumulations of land-living gastropods from a Mesolithic deposit of mollusc shells. In L6, below the Holocene levels, and above the roof of PS1, patches of a deposit were identified which comprised, among other materials, a half-moon made from a seashell and incised with notches. The latter suggests a dating to Dryas 3 period, not otherwise attested in the grotto, and therefore a continuation of the seasonal occupation between the final Epigravettian levels known to date, and those of the early Mesolithic with Sauveterrian industry. The excavation was halted at the roof of the PS1 occupation level. Row 9. In G9, a series of patches of level PS6 were removed, together with levels PS7, PS8, PS10 and PS11, here all re-deposited, reaching the roof of “PSα lens” which appeared in situ.. In the adjacent H9, a similar sequence was excavated, also terminating at the roof of “PSα lens”. In I9, after going through the same sequence, the various lenses of PSα were excavated, down to PS13, a minimum thickness of which was excavated. In K9, the excavation crossed PSα, stopping at the roof of PS12 to the north, and that of PS 13 to the south. Overall, this zone was rich in stone finds of the Epigravettian period and bone material, some of which burnt. Row 10. In I10, excavation began at the lower part of the level of detritus PS7, finding the same stratigraphic sequence as documented in the adjacent quadrant I9, down to PS13. In J10, excavation continued of the last lens of PSα, and the sequence relating to PS13 was removed, with levels corresponding to three different occupations, each with traces of burning and a substantial amount of lithics. The underlying PS14 presented a small sub-circular structure with lithic and bone material, defined as “Structure A”. Beside this the earliest archaeological levels found to date were recognised, those excavated the previous year in L10: PS16 and PS18. A lower occupation layer, here identified for the first time, was denominated PS19. In K10, the removal was completed of what was left in 2010 of occupation levels PS16 and PS18. These were very dark charcoal-rich layers with a fine matrix. Below, level PS 19 was identified and excavated, as in quadrant J10. The excavation halted at the roof of PS100 (collapsed material and gravel). At the end of the campaign, PS19 was dated by C14 to 19 ka bp (23 cal BP). It was therefore confirmed that the evidence from Grotta di Pozzo documents a particularly early re-colonisation of the mountain area following the LGM (Last Glacial Maximum).

As planned, the late Pleistocene levels in various quadrants of rows 6, 9, 10, and 11, were excavated. Row 6. In J6, K6 and L6, the upper part of the late Pleistocene sequence constituted by occupation level PS1 was removed, which together with an ashy lens indicated the presence of a combustion structure. A number of small limestone blocks and cobblestones can also be attributed to this structure. A large amount of material was recovered. The excavation stopped at the roof of PS6, on top of which were earlier combustion structures, attesting a continuity of seasonal occupation. Row 8. In H8 the removal of level PS6C, with ashy lens, revealed another occupation level. Excavations stopped at PS6D. Row 9. In G9, residual patches of occupation levels and detritus associated with eroded layers, re-deposited and hypothetically attributable to the “PSα” levels, were removed. Underlying lenses were removed down to the roof of PS13, _in situ_ and on whose surface blocks from the collapse of the vault rested. In H9, a similar sequence was excavated, down to PS13, where a collapsed block was also identified. A modest amount of material was recovered. In J9, a larger scale intervention removed the deposits forming the earliest Pleistocene sequence (PS7, PS8, PS9, PS10, PS11, several lenses of PSα and various levels of the PS13 group, attributable to diverse periods of occupation). Excavation halted at the roof of the layer of detritus PS14. In K9, layers PS12 and PS13. In the latter, part of a probable combustion structure was uncovered. Both layers produced a substantial quantity of lithic finds. In L9, layers PS13, PS14, and PS15 were removed, and a moderate quantity of materials recovered. These layers overlay the earliest occupation levels, PS16/PS18, attested by very slight traces of charcoal, and burnt bone. The excavation stopped at the collapses-gravels relating to a period immediately after the last, Würm glaciation). Row 10. Excavation in I10 began at PS13, with traces of combustion for four occupation phases, down to PS14. The absence of levels PS15, PS16, PS17 meant that PS18 was reached directly. Row 11. In I11, opened for the first time, the removal of the heavily disturbed Holocene levels and underlying layer of detritus, revealed an accumulation of 26 small limestone blocks and cobblestones, showing signs of exposure to heat. These formed an intact, sub-circular structure, not altered by subsequent interventions, flanked by an accumulation of ash, which was part of layer PS1 at which the excavation was halted. The campaign provided further evidence of the area occupied during the Late Glacial periods. Further evidence for the various moments of early and final Epigravettian occupation was also documented. The re-colonization of the site and therefore of the mountain area after the UGM, as early as the 19ka bp (23cal BP) onwards was also confirmed.

This season, the late-Pleistocene levels in quadrants J7-K7- L7, G8- H8- I8- K8—L8, G9, H9, I9—J9- K9, I10, J10 and J11 were investigated. Row 7. In J7, K7, and L7, the upper part of level PS6 was examined, on which hearth structures were present that were earlier in date than those documented in level PS1. The upper part formed by intermittent and discontinuous ashy lenses was removed. A substantial quantity of lithics and fishbone remains were recovered. Row 8. In G8, H8, and I8, level PS6C was excavated. This layer was formed by detrital deposits alternating with ashy lenses that stopped at the roof of PS6D. Lithics and bones were found. In L8, continuing from row 7, a part of PS6A was removed that presented a series of calcified roots and ash, already documented in row 8 during previous campaigns. Lithics, bones, and fish remains were recovered. Row 9 In G9, H9, I9 and J9 layer PS13 was investigated (PS13, “black lens of PS13bis”, “PS13bis”, “black lens of PS13ter” and “PS13ter”, “Ps13 quarter”), with four different occupation layers, probably very close together in time. The excavation of PS13 was completed in G9, and adjacent H9 was investigated in order to get a precise definition of the stratigraphic sequence. In H9 the excavation halted at the roof of “PS13ter”, in I9 the detrital layer PS14 was left exposed in the N+W, as was a dark occupation layer in the rest of the quadrant. As work continues this will be correlated with the general stratigraphic sequence (this is probably PS18). IN J9 and K9 excavation also stopped at the roof of level PS14 on which a series of very dark charcoal patches, defined “PS13quater”, and areas of rubefication were visible. All were associated with a hearth denominated “structure 6”. The hearth was sub-circular and presented a series of clayey, greenish grey-brown fills. The edges and bottom of the hearth were reddened by exposure to heat. A modest quantity of lithics and bone finds were recovered. Row 10. Excavation in I10 began at level PS13/PS18. The substantial depth of the deposit suggests it is formed by two layers that were not separated by a detrital layer and cannot be clearly defined. Row 11. The removal of the badly disturbed Holocene layers continued down to PS1. The excavations provided more evidence of occupation in the area during the Late Glacial period, documenting different and repeated occupations with early and late Epigravettian lithic industry. At the same time, the early re-colonisation of the site and therefore of the mountain areas after the LGM (Late Glacial Maximum) was confirmed already in the 19ka bp (23cal BP), as given by C14 dating.

This season’s research investigated late Pleistocene levels. In K9 and J9, the removal of “PS14” exposed the grey layer “grey lens on PS15”, lying between “PS14” and “PS15”, sloping N-E. Subsequently, the roof of “PS16/PS18” was exposed. In I9, “PS14” was redefined as “grey lens on PS15”. In the N-W quadrants, two black patches, in the form of small very irregular pits, were excavated. In H9, a residual patch of “PS13”, a strip of “PS13 bis”, a small section of “PS13 ter” and a layer interpreted as “PS 13 quater” were removed. Among the E-S quadrants, during the excavation of “PS13 bis” a semicircular patch of charcoally sediment was identified. Denominated “Black lens 2 of PS13 bis”, its section was “gutter” shaped and it could be the bottom of a manmade structure. In H8, a residual patch of “PS6D” was removed and the excavation of “Inter PS6-2” begun. In I8, the “black lens of PS6C”, containing abundant materials, also present in J8 E-S quadrants, and associated ashy lenses were removed. In J8, the “black lens of PS6C”, present in the E-S quadrants, was removed, and the edges of “PS6A III lens”, already exposed in the N-W quadrants, were defined. In K8, only surface cleaning of “PS6A III lens” was undertaken. In L8, the level “Detrital 3” was removed, exposing the edges of “PS6A III lens”. In J7, the “lens II of PS6A” was removed, exposing a very dark layer across the entire quadrant, which partially occupied the E-S quadrants in K/. In K7, the “lens II of PS6A” was removed, exposing a brownish layer in the central part of the quadrant and, in the N-W quadrants, the roof of “lens III of PS6A”. In L8, the “lens II of PS6A” was removed, exposing “lens III of PS6A” and an associated ashy lens. In addition, work was undertaken in the S-W quadrants of I11 and in the S quadrant of J11, on the levels underlying “structure 5”, defining and removing several layers in close association with it (ashy area; reddish layer).