Obsidian from the Site of Piano dei Cardoni, Ustica (Palermo, Italy): Preliminary Results on the First Occupation of the Island

New investigations on Ustica (Palermo, Sicily) originated from the need to improve our knowledge of the island’s archaeological and environmental heritage. Through field surveys, particular attention was paid to human occupation in the Neolithic phases and focused on the less investigated southern side of the island. The systematic survey of the area of Piano dei Cardoni in 2018 brought to light a new Middle/Late Neolithic site, already partially documented in the literature. The island was settled for the first time during these phases, as also testified from the area of Punta Spalmatore. The presence of Serra d’Alto, incised dark burnished, and Diana styles suggests that the site and the archaeological assemblage dates from the mid to late 5th millennium BC, as confirmed by AMS dating. In addition to pottery, obsidian artifacts were also recovered, and a preliminary study of these materials is presented here. Portable XRF analyses on a sample of 41 obsidian artifacts, representing a high percentage of the lithic assemblage compared to chert tools, show that the provenance of the raw material is Gabellotto Gorge (Lipari) and Balata dei Turchi (Pantelleria). These results provide new insight into broader regional debates about obsidian technology and its exchange during the Neolithic and open an important consideration for sites that are far from the raw material sources.


Introduction and Aims
New investigations on Ustica originated from the need to update knowledge on archaeological and environmental heritage, starting from systematic and non-systematic surveys of the area, with particular attention to human occupation in the Neolithic phases and focusing on the less investigated southern side of the island.
Research on the island is developing in collaboration with the Section for Archaeological Heritage of the Superintendency of Cultural and Environmental Heritage Service of Palermo. They are a partner of the project Brains2islands, headed by the National Institute of Geophysics and Volcanology (INGV), Osservatorio Vesuviano. The broader project objectives are interdisciplinary and aim to understand the relationship between human settlement dynamics in the Southern Tyrrhenian Sea and geological and environmental variations from the Middle Holocene to today. Geographical, socio-economic, and cultural studies are the basis for understanding the "island landscape" of the volcanic islands (Aeolian, Ischia, Ustica, the Sulcitan archipelago) or close to volcanic areas such as Capri. The environmental risk to which they are subject is assessed using the topography and geology of the area together with the socio-cultural attributes.
Despite the rather small surface area of the island (less than 9 km 2 ), the pre-to protohistoric settlement pattern of Ustica is not well-defined, especially due to the absence of stratigraphic contexts. Besides, the geomorphological characteristics of the island, mostly flat, allow a potential occupation of almost the entire surface ( Figure 1).
Piano dei Cardoni (PDC) represents one of the most densely occupied areas in the pre-to protohistoric stages. Near the dock of Cala Santa Maria, its importance lies in its strategic position of control of the southern area of the island and the possibility of exploiting for agricultural purposes the whole marine terrace where it is located. It is the only area of the island that covers all the phases from the Middle Neolithic to the Middle Bronze Age. The archaeological context of PDC represents the first obsidian assemblage on the island from stratigraphic units, after the one of I Faraglioni Village (Middle Bronze Age). Here the first study performed on the lithic assemblage found in the excavation trench of 2018 is presented, with a preliminary typological investigation and associated analyses for obsidian provenance. This research, which is one of other papers in the same issue about Ustica (Foresta Martin, Larocca, Micheletti, Pallara, & Acquafredda, 2020a;Foresta Martin, Barca, & Posedi, 2020b;Tykot & Foresta Martin, 2020),¹ can be inserted in the wider research area linked to networks of so-called marginal islands (i.e. Dawson, 2020;  1 All the articles are published in the Special Issue "The Black Gold That Came from the Sea. Advances in the Studies of Obsidian Sources and Artifacts of the Central Mediterranean Area," edited by Franco Italiano, Franco Foresta Martin & Maria Clara Martinelli (part 1) https://www.degruyter.com/journal/key/OPAR/6/1/html Knodell et al., 2020) and trade of obsidian in Southern Thyrrenian (Freund, 2018;Martinelli, Tykot, & Vianello, 2019;Tykot, 2017Tykot, , 2019Tykot, Freund, & Vianello, 2013). The identification of this site and the results from the obsidian analysis open new and interesting questions about the role of the island of Ustica in the mid-5th millennium BC. Its position and insular territory may have played a preferential role in the settlement dynamics of the Middle/Late Neolithic and in the working and circulation of obsidian from the Aeolian islands.

Geographical and Geovolcanological Context
The island of Ustica ( Figure 2) is located 60 km north of the Sicilian coast, west of the Aeolian arc, and represents the top of a huge submerged volcanic complex that rises more than 2,000 m from the bottom of the Southern Tyrrhenian Sea.
The emerged part of the Ustica volcano covers an area of fewer than 9 km 2 and reaches a maximum elevation of 248 m a.s.l., with the peak of Mt. Guardia dei Turchi in the central part of the island. It is mainly composed of volcanic rocks and subordinately of marine and continental sedimentary deposits. Volcanic rocks were produced by both subaqueous and subaerial effusive and explosive eruptions (Cinque, Civetta, Orsi, & Peccerillo, 1988;de Vita, 1993;de Vita, Guzzetta, & Orsi, 1995;de Vita, Laurenzi, Orsi, & Voltaggio, 1998;de Vita & Foresta Martin, 2017;Romano & Sturiale, 1971). Marine sedimentary rocks consist of clayey sands, organogenic and detrital fossiliferous limestones, and carbonate concretions. They were sedimented during the Middle-Upper Pleistocene sea-level high-stands, related to glacio-eustatic movements, which also generated five orders of marine terraces (de Vita & Orsi, 1994). Carbonate speleothems, landslide deposits, loose detrital fans, and aprons, sometimes pedogenized, are sedimentary deposits of continental origin, related to the more recent geomorphological evolution of the island (de Vita, 1993). Volcanism at Ustica is considered to be extinct, as it was active during the Pleistocene (Barberi, Borsi, Ferrara, Innocenti, 1969;Barberi & Innocenti, 1980), with the presently exposed eruptive products dating between 750 and 130 ky (de Vita et al., 1998;de Vita & Foresta Martin, 2017). As volcanism was no longer active after the last eruption occurred at 130 ky, it is reasonable to assume that humans had to interact only with the exogenous morphodynamic processes, which characterized the evolution of the island in more recent times and still active at the present. No exposed evidence of the sea-level rise that occurred since the Last Glacial Maximum is present on the island, as the related level surfaces are presently below sea-level, at a depth comparable with other equivalent features in stable areas of the Mediterranean (Furlani et al., 2017).
The investigated area of this study is located in the southern part of the island, in the PDC countryside, which is a flat area that represents the evolution of a marine terrace, connected to the Mt. Guardia dei Turchi slopes through a talus deposit. This terrace occupies an ENE-WSW trending land-stripe, which develops between 80 and 120 m of elevation a.s.l., with a 5% mean slope. The terrace slightly dips eastward, due to its tectonic displacement. The formation of the terrace is related to a transgressive-regressive cycle of Crotonian age (isotopic stage 9; 350 ky BP; Shackleton & Opdyke, 1973), as testified by 230 Th radiometric age determinations on carbonate concretions (de Vita et al., 1998). The terrace surface was tectonically displaced from its original position and the detrital blanket continued its evolution in a subaerial environment, with the formation of a thick paleosol that also includes large rounded boulders, formed by in situ spheroidal weathering of Mt. Guardia dei Turchi lava blocks (de Vita, 1993). The excavated site is situated above and within this last detrital deposit.

Archaeological Investigations on the Island (Figure 3)
Ustica has been occupied since the Middle Neolithic, as documented by the finds of Punta Spalmatore andwith less probability -Grotta Azzurra, for which the materials are attributed to the Middle Copper Age (Mannino, 1979;Mannino & Ailara, 2016). The site of the tourist village in Punta Spalmatore, although its layout has largely destroyed the prehistoric settlement, allowed for the recovery of obsidian and ceramic materials clearly framed at the middle and final phase of the Sicilian Neolithic (Mannino, 1998), allowing for the first time, in 1997, to record the occupation of the island in these phases.
The occupation during the Copper Age is testified not only by the caves San Francesco and Azzurra, on the southern side of the island (Mannino & Ailara, 2014, 2016, but also by Piano dei Cardoni (Spatafora, 2009, 2016, Spatafora & Mannino, 2008, with materials of the early and middle phase. Finally, the island was occupied during the Bronze Age. Apart from the village of I Faraglioni (Holloway, 1997(Holloway, , 2003Mannino, 1982;Spatafora, 2012Spatafora, , 2016, we can assume that during the Middle Bronze Age the occupation was broadly spread on the whole island, while from the Middle Neolithic throughout the Copper Age (5th millennium-3rd millennium BC) the island was only partially occupied.
Obsidian materials from the island were examined only in one stratigraphic context (I Faraglioni Village, Middle Bronze Age) (Tykot, 1995); all the other analyses were performed on materials from private collections or random surface materials, whose chronology, unfortunately, cannot be assessed on a typological basis (Foresta Martin et al., 2016;La Monica, Rotolo, & Foresta Martin, 2019;Tykot & Foresta Martin, 2020). Obsidian from Pirozza, Spalmatore, on the south-western side of the island, likely belongs to the Neolithic phase (Foresta Martin & Tykot, 2019), although in the same site pottery from the Bronze Age phase also was found (Mannino, 1998).
New research on the island stems from the need to improve our knowledge of the island's archaeological and environmental heritage, starting from an extensive survey of Ustica, with particular attention to human occupation in the Neolithic phases, to understand the distribution and chronology of activity of this period.

Methods
The archaeological field survey carried out on Ustica allowed us to verify and quantify known and unknown prehistoric sites through a selective strategy of sampling and a systematic fieldwalking activity (Bevan & Conolly, 2013;Francovich, Patterson, & Barker, 2000). Previous archaeological explorations occurred in the late 1990s when the Archaeological Heritage Department of Superintendency of Palermo undertook surveys using traditional methodologies to locate all the archaeological evidence of the island (Mannino & Ailara, 2016). The second step was to create a GIS platform, based on open-source software, which includes gathering and relating all these dates as maps, landscape layers, and existing and new archaeological evidence of Ustica.
Such a framework led to a clearer understanding of the archaeological potential of Ustica, guiding us towards a second more extensive and detailed survey of PDC, one of the most promising areas of the southeastern part of the island. Following the varying shape and dimension of the modern fields, 78 material collection areas (MCA) were detected and each area was numbered and geotagged via a GPS smartphone on the Google Maps Android application. The distribution of archaeological materials (mostly lithics and pottery) was represented in a preliminary map of the density and distribution of materials using different intensity colours for each MCA. Subsequently, to exploit the archaeological potential of these large material scatters, one of the fields with a major presence of sherds was identified and chosen for a test trench of 1 × 1 m (area 1).

Results
After the first test trench in 2018 highlighted the presence of archaeological materials dating to the end of the Neolithic/beginning of the Copper Age and Middle Bronze Age, an area of about 36 sqm was opened in 2019 (area 2), to allow a further comprehension of the context. The sample pit was opened in a peripheral area of the concentration of archaeological materials ( Figure 4).
The area revealed the presence of a settlement of about 20,000 sqm (2 ha), according to the distribution of the potsherds and obsidian artifacts. The area has an approximately oval shape and stands in the middleupper part of the flat area derived from the marine terrace (see par. 2). The distribution of potsherds is quite similar in terms of density to the one of obsidian artifacts ( Figure 5).
Unfortunately, archaeological layers were superficial, and most cultural evidence was probably destroyed, at least for the most recent phases.
The presence of a settlement is nonetheless confirmed and its features are still under reconstruction. Even though a clear floor layer has not yet been found, the site preserves some remains of wall structures mostly made by earthen elements, some pits, and several tools for daily activities ( Figure 6). Pottery is represented by dark burnished ware, red burnished incised ware (Diana style), and painted ware (usually red-on-white, attributable to the style of Serra d'Alto). In terms of chronology, pottery is attributable to a Middle Neolithic/early phase of Late Neolithic, mid 5th millennium BC (Speciale, Larosa, Battaglia, & Vassallo, 2020a), with comparisons with north-western part of Sicily and Aeolian islands (see i.e. Cavalier, 1979;Martinelli & Quero, 2013;Tinè & Tusa, 2012;Traverso, 2012); pottery remains of a previous phase, attributable to the first half of 5th millennium, are present. The chronology of material culture is confirmed by 5 unpublished AMS datings (Beta Analytics Lab). Bioarchaeological remains (animal bones and vegetal remains) are quite preserved and under investigation.
The lithic assemblage is composed not only of obsidian tools, but also of chert (likely imported from the north-western area of Sicily) and a wide variety of lava millstones; they are mostly local volcanic rocks used as mullers, querns, mortars, hatchets, pestles, and smoothing tools. Most of these raw materials were locally collected and likely originated in Monte Guardia dei Turchi, while some of the small hatchets were obtained by lava slabs originating in the area of Tramontana, on the northern side of the island (Speciale et al., 2020b).

Methods
The functional study of lithic artifacts is still ongoing. A total of 41 obsidian artifacts were selected from the record of 2018 and 2019 investigations. Samples were taken from the most reliable stratigraphic units and also the surface, choosing both blades, bladelets, flakes, and debris. The visual analysis of macroscopic physical features included the observation of the fragments in white, reflected light. Samples were then analysed by K. Freund using a Bruker Tracer 5 g pXRF instrument (provided by R. Tykot) operating at 50 kV and 35 μA for 90 s and using a filter that enhances precision for trace elements (Rb, Sr, Zr, and Nb) already shown to be successful for central Mediterranean obsidian sourcing (see Freund, Tykot, & Vianello, 2015;Freund, Tykot, & Vianello, 2017;Tykot et al., 2013;Tykot, 2019). Source assignations were achieved by calibrating the raw analytical data against standard obsidian reference materials to determine the actual concentrations; the results then being compared with data generated from Mediterranean geological obsidian samples using the same instrumentation. More detailed information on the use of this pXRF is provided in Tykot and Foresta Martin (2020, this Special Issue).

Results: Obsidian Technology and Composition
In the PDC site, 37 obsidian items have been randomly collected from surface soil deposit and another 19 obsidian items have been found from the 2018 survey at "field n.64" located in the PDC area.
From the PDC stratigraphic excavation of 2018, a total of 102 obsidian items, including tools, fragments, and technological waste products, have been found; the excavation of 2019 added other 905 items to the record. Broken blades and bladelets, flakes and flake fragments, microliths, broken scrapers, chips, and angular waste attest that the production of the obsidian lithic industry took place at the site. Bulb percussion analysis suggests that the knapping process was by direct percussion through a hard hammer. The only residual core, a simple bidirectional one and almost completely exhausted, found at PDC shows bidirectional knapping through direct percussion. Most knapped obsidian production found from stratigraphic units at PDC reflects the need for regular blade and bladelet production.
Obsidian lithic analysis is ongoing using of soil microelement analysis and surface alteration wear to detect artifact function and life cycle(s). A preliminary analysis on some obsidian blade fragments highlights a relation to vegetal and faunal processing. From a techno-typological point of view, the 41 obsidian artifacts sampled can be classified mainly as blade fragments: eleven are proximal, eight mesial, and four are distal. Six artifacts are related to angular waste (or debitage) and nine are undetermined fragments. Seven artifacts, finally, show a simple marginal retouch that the micro-wear analysis will define as intentional or merely the result of use wear. One burin is also present (Figure 7).
The majority of the sample is grey to black in colour and shows a medium to high transparency index: 3-5 on a scale from 0 to 5; observed in white, transmitted light, this group exhibit a grey or light-grey colour ( Table 1). Only seven artifacts show phenocrysts (millimetric spherical structures). Of the three artifacts, characterized by a transparency index equal to 0, two (2018.2 and 2018.7), exposed to an intense white light, exhibit a greenish colour filtering through at the thinner edges, allowing us to hypothesize a provenance from Pantelleria, at a mere visual observation (Foresta Martin & Tykot, 2019), lately confirmed by pXRF analysis as further explained.
The obsidian findings from the PDC stratigraphic excavation are of good source quality, while from the surface findings lower quality obsidian raw material is attested. Of the 41 analysed artifacts employing the pXRF, 39 can be traced to the Gabellotto Gorge sub-source on the island of Lipari. None of these artifacts originated from Canneto Dentro. Two artifacts in the assemblage came from Pantelleria, both from the Balata dei Turchi sub-source. Analytical data are in Table 2. Nevertheless, the two obsidian items from Pantelleria are from a surface collection and cannot be attributed to a specific time period.

Obsidian and Its Significance in the Site
The presence of obsidian at the site, whose occupation spans about 400 to 500 years, is quite impressive. There is no evidence so far of seasonal occupation and all the data point to a year-round settlement. Ongoing bioarchaeological analyses testify to a varied economy, based on agricultural production (attested also by the relatively large number of querns, mortars, etc.) as well as the great exploitation of animal sources (both domestic mammals than avi-and ittio-faunas). Wear analysis will reveal the degree to which obsidian was used in these activities. Obsidian may have been selected for specific activities, due to its properties compared to chert. The occupation of the island during the Neolithic is concentrated so far to the area of PDC and Spalmatore, even if the exploitation of the resources of the whole island was probably very easy due to the reduced area. Ustica is far away from raw material sources, and as such has interesting implications for our understanding of regional patterns of exploitation of obsidian from the Neolithic to the Bronze Ages. Contexts with imported lithic raw materials are potential places to observe technological adaptation and variability/changes, as the uncertainty of access to lithic resources is one main factor that determines the technological production strategy (see e.g. Gaffney & Summerhayes, 2019).
The stratigraphical data for the two items from Pantelleria sources being absent, for the site of PDC we have a 100% provenance from Gabellotto source for the Middle/Late Neolithic phase (Figure 8). Obsidian in Ustica for all chronological phases comes from Pantelleria and Lipari, with a single exception from Palmarola (Foresta Martin et al., 2016).
Obsidian from the unexcavated area of Pirozza-Spalmatore, the Neolithic/Copper Age site on the western part of the island, records an origin from Lipari-Gabellotto (89%) and Pantelleria-Balata dei Turchi and Lago di Venere (11%) (Foresta Martin & Tykot 2019). Considering the origin of obsidian on the island during the Bronze Age, artifacts from the I Faraglioni Village (Middle Bronze Age) were analysed; also, during this late phase, obsidian on the island was imported from Lipari and Pantelleria (Tykot et al., 2013;Tykot, 1995;Tykot & Foresta Martin, 2020, this Special Issue).

Archaeological Interpretation of the Data (Figure 9)
Ustica is not close to the obsidian sources (about 150 km far away from Liparibut only 120 km from Alicudiand about 250 km from Pantelleria). It is possible that its link through maritime connections was direct to Aeolian sources. The route through the Sicilian coast (about 30 km from Lipari to Capo d'Orlando, 160 km from there to Capo Gallo/Palermo, and then 60 km to Ustica) seems unlikely in terms of effort, but not to exclude at this phase of the investigation. Furthermore, prevalent currents of the Mediterranean show a predominance of dynamics from west to east (El-Geziry & Bryden, 2010;Millot & Taupier-Letage, 2005), making the sailing routes from Ustica to the Aeolian Islands easier, which are also more clearly visible due to their altitude (up to 987 m a.s.l. of Salina), at the opposite of Ustica that is more rarely visible from the coast or the archipelago. No data on sailing are available for the Neolithic of the Mediterranean (nor archaeological data from shipwrecks or representation of sailing means), despite the fact that obsidian collection in the broader Mediterranean is documented since the 11th millennium (Broodbank, 2006). A sailing route from Ustica to Palermo is possible within 2 days of favourable weather conditions, assuming that a paddle-propelled craft was used (Broodbank, 2013). Longer and harder was the sailing from Ustica to the Aeolian islands (or even worse the opposite), considering a full day of at least 12-13 h during the good season (searoutes.com). Ustica-Pantelleria is not a sea route to consider without at least a stop, and getting to Pantelleria, given the distance from Sicily, was probably as hard as getting from Alicudi to Ustica, more or less being the same distance. More experimental data would be necessary for detailed estimations.
Obsidian assemblages and their typology, together with pottery typologies, testify that the Neolithic "package" arrived at Ustica for the first time through contact with the Aeolian islands. It is to pay particular attention to the quality (and the sources) of obsidian raw materials and to the potential tasks for which it was designed. While the existence of itinerant specialists is possible, the current data suggest that the initial procurement and subsequent redistribution of Lipari obsidian were undertaken by a select number of communities who had the capacity to organize and crew long-distance voyages to the quarry sites (Freund, 2018).
The site was compared with some selected excavated sites of the end of 6th and 5th-millennium in Sicily where data about lithic assemblage were available, as shown in Figure 9. The sites are open-air and caves. It is quite evident that chert is the favourite material for south-eastern, central, and western Sicily.
Gabellotto in Lipari is undoubtedly the main source of obsidian in Sicily and southern Italy during the 5th millennium BC. It is almost the only source of obsidian from the Neolithic site of Contrada Diana in  By "lithic" is listed the percentage of lithic artifacts such as chert; "distance" is the linear distance between two places; when number of obsidian samples is in bracket, it is an approximate number.
Lipari . Comparing the data of provenance in the sites of mainland Sicily, according to scanning electron microscopy (SEM-EDX) and inductively coupled plasma mass spectrometry (LA-ICP-MS) analyses, obsidian from Roccapalumba (in central Sicily, around 40 km south of Palermo) has a unique origin from Gabellotto, Lipari (Italiano et al., 2018); in Roccapalumba, besides the tools, many cores are present (Mannino, 2012), differently from PDC. For other sites of mountainous northern Sicily, Lipari represents the main source (Forgia, 2019). At the present moment, there are very few Neolithic sites in western Sicily with obsidian (i.e. Grotta dell'Uzzo, Francaviglia & Piperno, 1987, where a high percentage of Pantellerian obsidian is recorded). In general, in western Sicily there is a mix of provenance from Balata dei Turchi and Lipari sources, while in eastern Sicily Lipari obsidian is nearly exclusive (Freund et al., 2017;Pappalardo et al., 2013;Quero, Martinelli, & Giordano, 2019). It is possible that Ustica was a centre of import and production of obsidian tools and its percentage, quite similar to the sites close to the sources like Lipari and northern Sicily, makes it an unusual far-from-the-quarry site, identifying it as a hybrid site, with very close contacts to Aeolian Islands; basically, a direct destination of Aeolian obsidian exploitation before distributing it to other sites in Western Sicily and/or Central Sicily.
Observing the distribution of the percentages, Ustica could be a node of the "down-the-line" system for the spread of raw materials towards western Sicily (see, i.e. Sardinian obsidian towards Southern France, Freund & Batist, 2014). It is so far difficult to find direct contacts between Ustica and central Sicily which could justify some dynamics in the spread of the obsidian raw sources and it is more likely to separate the two lines, one directly through the bridge of Aeolian islands to Ustica and one from Lipari through northern Sicily towards central Sicily. Less easy to identify is the route through which Pantelleria obsidian could have been brought to Usticaif the 2 pantescan samples collected from the surface belong to the Neolithic phasealthough a route through western Sicily is probable. Occupation of Pantelleria island for the exploitation of obsidian sources is still quite unexplored (Cattani & Tusa, 2012), despite a change in the coastal level that could hide a very early occupation of the island since the Early Neolithic, maybe for the exploitation of obsidian (Abelli et al., 2016).
Obsidian in PDC is the main exotic good recorded on the site. Production of pottery with local materials is recorded and exploitation of local volcanic rocks for grinding tools is clear in the settlement (Speciale et al., 2020b). Of course, it is not possible to establish at this stage of the investigation if obsidian was the primary and/or only material or if it was a complement to other exchanged goods. At the same time, Ustica resources, such as volcanic millstones or pottery, could be traded to north-western Sicily and exchanged for chert resources or other goods. Moreover, compared to the rest of the lithic assemblage, obsidian prevails in northern Sicily, while in south-eastern Sicily, where obsidian sources are less reachable and chert sources are quite available, the ratio is the opposite.

Conclusion
The new data from the site of PDC confirm that the island was widely occupied since the end of the Middle Neolithic period. Ustica operated within a network of exchange of different raw materials and objects, and obsidian was undoubtedly one of the most important ones. Despite the wide use of local volcanic rocks for millstones and axes, easily collectible near PDC or on the northern side of the island, the conspicuous and heterogeneous lithic assemblage shows a fundamental reliance on lithic resources from Lipari. The volcanic origin of Ustica and the absence both of chert and obsidian make it a hybrid place, far from the access to all of these lithic sources that were imported through the sea from Lipari and, maybe in minor part, from Pantelleria, but probably also from the area of Palermo for the chert. Nevertheless, the prevalence of obsidian in the lithic assemblage (more than 92%) makes PDC's percentage similar to the close-to-thesource sites. Further investigations and a more detailed typological investigation are required to understand the role of Ustica in the network distribution of the raw material towards northwestern Sicily.