The campaign’s objectives were: – complete the open area excavation of the Pleistocene layer US 104 and the transition to the underlying US 105 in trench B1. – continuation of the excavation in order to reach the base of US 318 in trench B2. – investigation of the final patches if the Holocene deposit in trenches C-D.
In trench B1 the open area excavation of US 104 was completed and underlying US 317 was investigated; the lower level was slightly thicker and its inherent characteristics showed it to be the interface between US 104 and underlying US 105, which was much thicker.
The excavation of US 104 produced numerous micro-faunal remains; this characteristic seemed constant throughout the level that was obviously a natural formation, with birds of prey (owls) leaving the remains of their prey (owl pellets) in the cavities. It was also seen that the underlying level (US 105) was not clearly distinct from US 104, rather it presented an interface (US 317) with varied patches and tones. These can be interpreted as the residue of US 104 (loose soil and micro-fauna) and as the first patches of US 105 (more compact soil rich in clasts > 10 cm, and macro-fauna). Consequently, is seems that in antiquity the palaeo-surface was very irregular, characterised by depressions, some of a certain diameter and depth, as is usual in the cavities in natural deposits.
Once the investigation of US 104 and 307 was completed, the underlying level US 105 was investigated. This level, investigated in preceding trenches B2 and B3, was characterised by macro-fauna and large clasts. Trench B1 produced various finds, but no anatomical articulations. Given the decreasing number of finds and taking into account of the data obtained it was decided to close the excavation at the depth of -140/-150 cm below present ground level.

The last patches of US 109 were investigated in quadrants d-e 7-8-9-10 in trench B1. The excavation revealed the presence of abundant macro-fauna, occasional flint objects, and charcoal fragments. There was a clear transition from US 109 (Epigravettian phase with the presence of lithic industry) to the underlying US 104 (absence of occupation evidence and abundant macro-fauna). It was also seen that in some points in this innermost area of trench B1 US 109 seemed to deepen, following the natural change in height of the cave floor, sloping towards the inner chambers. This US was also affected in antiquity by substantial Karsitic activity. The concretion was particularly developed beyond the sector of quadrants “10” as far as trench “D” (sector of the “13” quadrants), where it was not possible to excavate. In this sector, the faunal remains incorporated in the concretion were difficult to excavate and were not always extracted in the most appropriate way. Due to the layer’s difficult taphonomy it was decided not to extend the excavation any further north in this sector of the trench.
It is hoped that in the future it will be possible to apply a more accurate and specific methodology to the identification and extraction of the faunal remains in this sector of the deposit.

The excavations in trench B2 focused on US 318, characterised by Pleistocene fauna and datable to just before 40 Kyr BP. A deep level of silty soil (US 318) was uncovered that presented sporadic faunal remains and a total absence of occupation evidence. The natural base of this level was reached, which rested on a concretion formed by the Karstic activity and formed a natural container to the matrix of the soil of this level. Given the scarce faunal material, the trench was closed.

The investigation of US 325, uncovered in 2014-16, continued in trench B3. The level was heavily concretionary and practically impossible to investigate in a way that preserved the faunal remains. Therefore, the trench was closed.
In trench D the investigation continued in the north sector adjacent to trench C. The residual patch of the proto-historic level (US 165) and underlying Neolithic level (US 283) constituted by medium sized clasts were excavated.
An area of c. 2 m2 of the Neolithic level was exposed; only a few fragments of plain impasto pottery were found that are difficult to date or attribute to a particular culture. The investigation of the remaining patch of this deposit will continue in the future.

Sampling for pollen analysis
As part of the collaboration with the Laboratories of Biological Sciences in the Faculty of Natural Sciences at Rome “Tor Vergata” University it was decided to begin sampling of the various levels of the deposit for detailed pollen analysis.

Eight samples were taken:
Sample 1. Late Pleistocene with macro-fauna (B2, q. c3 at level -235)
Sample 2. Late Pleistocene with macro-fauna (B2, q. c3 at level -204)
Sample 3. Late Pleistocene with macro-fauna (B1, q. b7 at level -114)
Sample 4. Late Pleistocene with Epigravettian lithic industry (B1, q. b7 at level -105)
Sample 5. Early Holocene-Mesolithic (B1, q. e10 at level -129)
Sample 6. Late Neolithic horizon (B1, q. e10 at level -101)
Sample 7. Early Neolithic horizon (B1, q. e10 at level -90)
Sample 8. Bronze Age/Eneolithic layer (B1, q. e10 at level -68)