Materials and Tools across Volcanoes: Exploitation of Georesources in Piano dei Cardoni (Ustica, Italy) during Prehistory

The paper aims at merging the first results from the analyses of the georesources exploited in the site of Piano dei Cardoni (Ustica island, Italy) during the Neolithic phases of its occupation (Middle-Late Neolithic, 4.7-4.2 ka cal BC). Grinding tools consist of a very varied typology of local volcanic rocks, easy to collect and available very close to the investigated site. A selection of shapes and lithology is applied to reach the best performance of the tools. The elevated number of grinders, pestles, mortars testify to an intense activity of food/plant processing in the site. The absence of chert or obsidian resources on the island pushed the human communities to import such raw materials from the Aeolian islands and probably from the north-western area of Palermo. Pumice is collected on the same island, probably due to the local availability and its good quality. Similarly, local clay resources are used for the manufacture of ceramics, mostly burnished and incised wares. Ustica was therefore almost autonomous for the exploitation of resources, with volcanic rocks readily available in abundance and with the most significant exception being chert and obsidian. This last one probably imported and worked on the island and then moved towards North-Western Sicily.


Introduction
In the definition of the cultural landscape, whatever the declination is, two factors are always present: human and environment and how they interact [Sauer, 1925;Mercuri et al., 2010;Mercuri, 2014;Clement and Cassino, 2018; production, but also for lithic assemblages, are not only about the technology itself, but they also represent cultural choices, social concerns and the record of connection routes [Lemonnier, 1993;Mentesana and Fragnoli, 2020;Dobres, 2009;Clarkson, 2008;Horowitz and McCall, 2019]. In particular, in volcanic areas, the dialogue between past communities and the local environment should be read through the interpretation of the geological dynamics [Di Vito et al., 2009;Di Renzoni et al., 2016;Riede et al., 2020]. Volcanoes present not only hazards but also a unique opportunity for georesource exploitation. For example, the role that lava querns and millstones play from the onset of agriculture until recent times is well known all over the Mediterranean and beyond [Santi et al., 2015;Renzulli et al., 2019;Kadowaki, 2021], sometimes from great distances already during prehistory [Bracchitta 2015].
Furthermore, during the Middle/Late Neolithic (6th-5th millennium BCE) in the Central Mediterranean, during a phase of growing contacts and consolidated exchange of goods, obsidian played a primary role in the establishment of short-and long-distance networks [Martinelli et al., 2019]. As a further example, soils composed of volcanic ash can have an important influence on agricultural production [Neall, 2006]. Volcanic islands are even more complex and intriguing for the analysis of cultural landscapes, because of their limited resources [McLaughlin et al., 2018;Fitzhugh et al., 2016], the possible discontinuity in the human occupation [Dawson, 2014;Horsburgh et al., 2017;Di Napoli et al., 2020] and also the capacity to develop very specific cultural landscapes [Rackham, 2012;Walter and Hamilton, 2014].
The case study analysed here is the small volcanic island of Ustica. Despite being quite remote from the Sicilian coast, the perception of its marginality might have changed throughout prehistory, according to the exchange dynamics and the cultural contacts of the Southern Tyrrhenian [Fitzpatrick et al., 2016]. This paper aims at merging the first results from the analysis of the archaeological materials of the site of Piano dei Cardoni (Ustica, Italy) during its phases of Neolithic occupation (4.7-4.2 ka BCE). After the typological classification of the artefacts (lithic tools and ceramics), the georesources and their uses are analysed: raw material identification and provenance are tackled via analyses of micro-and macro-components of pumice, obsidian, lava stones and ceramics, in order to explore the exploitation of Ustica's insular landscape and the potential network of contacts with other regions.

Geological and archaeological context
The investigated area is located in the north-eastern part of Piano dei Cardoni at Ustica, which is a flat land representing a marine terrace. It is connected to the Ustica central relief by a talus deposit. This erosional surface is related to a transgressive-regressive cycle of Crotonian age (isotopic stage 9; 350 ka BP [Shackleton and Opdyke, 1973], set on older submarine and subaerial lavas [de Vita et al., 1993;1998]. During its formation, the abrasion surface of the terrace was the site of marine sedimentation that formed a detrital blanket composed of fossiliferous carbonate concretions and silt, sands and gravels due to the reworking of pre-existing volcanic rocks at the foot of the sea-cliff that borders the terrace. Afterwards the terrace was tectonically displaced from its original position and the detrital blanket continued its evolution in a subaerial environment, with the formation of talus breccias and a thick paleosol that also includes large rounded boulders, formed by in situ spheroidal weathering of lava [de Vita, 1993]. The Piano dei Cardoni lowland occupies an ENE-WSW trending land-strip, more-or-less parallel to the central ridge of the island, which ranges between 80 and 120 m of elevation a.s.l., with a 5% mean slope. Moreover, the terrace slightly dips eastward, due to tectonic displacement. landslide deposits, loose detrital fans and aprons, sometimes pedogenized, are deposits of continental origin, related to the more recent geomorphological evolution of the island [ de Vita, 1993]. Volcanism was active since the lower/middle Pleistocene until the last eruption, which occurred at 130 ka BP [Romano and Sturiale, 1971;Cinque et al., 1988;de Vita, 1993;de Vita et al. 1995;1998;de Vita and Foresta Martin, 2017]. Therefore, it is reasonable to assume that humans had to interact only with the exogenous morphodynamic processes, active on the island in more recent times [Speciale et al., 2019] (Figure 2).

Claudia Speciale et al.
Ustica has been occupied at least since the Middle Neolithic, as evidenced by the finds of Punta Spalmatore and -with less certainty -Grotta Azzurra, for which the materials are attributed to the Middle Copper Age [Mannino, 1979;Mannino and Ailara, 2016]. The modern tourist village at Punta Spalmatore largely destroyed the prehistoric settlement, but obsidian and ceramics dated between the middle and final phase of the Sicilian Neolithic were recovered [Mannino, 1998], revealing for the first time, in 1997, the occupation of the island in these phases. The occupation during the Copper Age is attested not only in the caves San Francesco and Azzurra, on the southern side of the island [Mannino and Ailara, 2014;, but also in Piano dei Cardoni [Spatafora, 2009; with materials dated to the Early and Middle phases of the Sicilian Copper Age (4th millennium BCE) (Figure 3). The 2018-2020 fieldwork campaigns at Piano dei Cardoni revealed the presence of a settlement and a funerary area with a megalithic stone tomb [Speciale et al. submitted]. Although a clear floor layer for the settlement 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. Underneath these structures, the semi-underground megalithic funerary structure was found ( Figure 4).
Ceramics retrieved during these last excavations have been grouped into 5 main ceramic wares, based on surface appearance and treatment: • Dark burnished ware, which shows black to brown irregular spots on a burnished surface; • Red burnished ware, known as Diana style, distinguished from the previous ware by being mostly red or with irregular black spots on a red burnished surface; • Incised and burnished ware, with linear incision on a burnished surface, which might be black to brown to red; • Red painted ware, where the red paint might be applied on a white slip, making it similar to Bande Rosse/Serra d'Alto style ceramics, or directly on a smoothed surface; • Some other vessels present a solely smoothed surface ranging black to brown to red. Based on the ceramic wares encountered, the site lifespan can be mostly set between Middle Neolithic and an early phase of Late Neolithic, mid-5th millennium BC [Speciale et al., 2019], in line with the occupation of other areas in the north-western part of Sicily and Aeolian islands [Tiné and Tusa, 2012;Cavalier, 1979;Martinelli and Quero, 2013;Traverso, 2012]. Bioarchaeological remains (animal bones and vegetal remains) are also preserved and are currently under study [Speciale et al., 2021b].

Obsidian -materials and methods
The functional study of obsidian lithic artefacts is still ongoing. A total of 41 obsidian artefacts were selected from the record of 2018 and 2019 investigations. Samples were then analysed by K. Freund. More detailed information on the use of pXRF is provided in Tykot and Foresta Martin [2020].

Obsidian -results
The detailed results of the analysis are published in Speciale et al. in press 2021a. The obsidian findings from the Piano dei Cardoni stratigraphic excavation are of good source quality, while from the surface findings lower quality obsidian raw material is attested. Of the 41 analyzed artefacts employing the pXRF, 39 can be traced to the Gabellotto Gorge sub-source on the island of Lipari. The other two obsidian items are from Pantelleria, but they belong to a random collection from the surface and they cannot be attributed to a specific chronology. The only reliable source for this phase of Piano dei Cardoni is the island of Lipari.

5
Exploitation Georesources in Piano dei Cardoni  (Table 1) Based on macroscopic observation, the chert of the artefacts found at Piano dei Cardoni is mainly of the finegrained type and only a few specimens are coarse-grained. A preliminary observation with a stereoscopic microscope (8X and 100X) [Tringham et al., 1974] has been performed.

Chert -materials and methods
Products have been divided into four broad categories: laminar, flakes, fragments and angular waste. Out of a total of 81 artefacts, there are 33 laminar products, 10 flakes, 7 fragments and 31 angular waste, as detailed, with percentages (italic), in tab. 1. Among the laminar products, 4 are entire, 8 are proximal fragments, 19 are mesial and 2 are distal. Laminar products range from 6 to 18 mm in width, with a mean value of 11 mm. The surface of the artefacts is fresh for the 80,2% of the total; 11 artefacts (6 are laminar), that is the 18,5% of the total, are burned and only 1,2% shows a patina. Artefacts are generally unretouched, with the exception of 6 laminar products, among which 2 have been recognized as sickle elements. Remains of the cortex are almost totally absent on the whole assemblage. Provenance of chert is highly difficult to assess, despite new methods are developing (see i.e. [Gurova et al., 2021;Muntoni et al., 2021]. The preliminary analysis on Piano dei Cardoni tools does not allow for the moment to determine their provenance.

Chert -results
The analysis revealed the good preservation of the surface for the future planning of use-wear analysis with a high-magnification approach [Keeley 1980] and identified 4 artefacts probably exploited as sickle elements ( Figure 5.a-d). The distribution of the polish on the surface of the inserts suggests, at a preliminary observation in "low power" approach, that they were probably diagonally hafted, as described and detailed in Mazzucco et al. [2018 and likely attesting here to the harvesting tradition, characterized by a curved sickle used for cutting at a low and middle height. Two of the possible sickle elements ( Figure 5.b and 5.d) show a peculiarity: the polished surface is just one of the two faces, suggesting that the elements have been intentionally (?) removed from two original sickle inserts or, less probably, that a post-depositional surface modification (in this case resulting in a surface similar to a cereal micropolish, when observed at low magnification) affected only one of the two faces of the fragments. Final products, such as proximal, mesial and distal parts of blade(let)s are the main category of the chert finds (40%), while by-products from knapping representing a similar percentage (38%), suggest modest or sporadic knapping activity on the site. In Figure 5 we present a selection of the variety of chert reaching the site. There is a clear preference for very fine-grained varieties, but a coarse-grained variety is also present, as 5c (selected as sickle element) and 5e. Chert is absent on Ustica and Lipari and we hypothesize that it was imported from the area of Palermo or even from Southern Italy. Further analyses on possible chert sources from Sicily may help confirm this. (Table 2) 66 lithic tools were collected and divided into categories according to their morphology and macroscopic wear traces. Despite grinding tools attracting increasing attention from researchers, there is no shared agreement on the use of the terms to define the tools, probably due to the large number of shapes and local variables (see i.e. [Lucarini and Radini, 2021;Hamon, 2008;Dietrich et al., 2020]. Some cm 3 of 20 tools were then sampled, analyzed  and described according to their lithology to assess provenance and raw material selection. Among these samples, 6 belong to the pumice tools and 13 to the other volcanic tools. At least one tool was sampled for each category, except for mortars, percussors and fishing net weights for the difficulty of sampling without creating relevant damage. Pumice tools were all analyzed because of their potential provenance from the island of Lipari.

Exploitation Georesources in Piano dei Cardoni
The lithological analysis of the specimens was carried out through the observation of their mineralogical and textural macroscopic characters on rupture surfaces exposed for this purpose, supplemented, where necessary, by observations made with the aid of a 10x magnifier lens or a binocular microscope with a 7.5x to 60x continuous zoom.

Pumice -results (Figure 6g)
Six pumice tools (ns. 13, 15, 16, 17, 18 and 19) were macroscopically analyzed. All the samples are grey/greyish pumice lapilli with a variable degree of vesiculation, aphyric or with rare phenocrysts of feldspar (plagioclase and anorthoclase) and femics (pyroxene and/or amphibole); samples n. 15 and 17 have clear signs of mechanical abrasion. The provenance of the raw material is clearly traceable to the trachytic pumice-fall and -flow deposits of the Grotte del Lapillo unit (420 ka [de Vita et al., 1998]), widely exposed on the island, Breccias (de Vita, 1993), which is related to the reworking of Cala S. Maria lavas in a submarine environment during the Crotonian (isotopic stage 9 of [Shackleton and Hopdike, 1973]) transgressive-regressive cycle, responsible for the formation of the first order marine terrace of about 350 ka [de Vita and Orsi, 1994;de Vita et al., 1998]. One sample (n. 3) is a tuff fragment, consisting of a lithified, massive and poorly sorted coarse ash deposit, with subangular to sub-rounded, basaltic scoriaceous lapilli, lithic fragments and loose crystals. This sample is likely correlated to the unit of M. Costa del Fallo Tuffs (approx. 500 ka), or the Falconiera Tuff (130 ka [de Vita et al., 1998]). The last sample from the tomb (n. 4) is a fragment of a subaerial basaltic lava with scarce microphenocrysts of plagioclase and iddingsitized olivine, with altered carbonate encrustations. It very likely belongs to the M. Guardia dei Turchi unit (520 ka [de Vita, 1993;de Vita et al., 1998]), which was partially submerged by the sea during the formation of the first order marine terrace [de Vita, 1993;de Vita and Orsi, 1994].
Samples 5, 9, 10, 12 and 14, used for hatchets, upper grinders, polishers and pestles, are submarine basaltic lavas, compact or micro-vesiculated, aphanitic or microcrystalline with abundant plagioclase and subordinate olivine crystals, and probably belong to the unit of Cala S. Maria lavas or perhaps to the submarine part of the M. Guardia dei Turchi unit (de Vita, 1993). Subaerial basaltic lava fragments, well vesiculated, with plagioclase microphenocrysts from M. Guardia dei Turchi unit were used for the lower grinders n.7 and 11 and the upper grinders n.6 and 8; sample n. 8 is also characterized by the neoformation of alteration minerals within the vesicles, while sample n. 11 contains iron oxides deposited in the vesicles.
The lithologic characteristics of all the samples, widely represented in the rocks exposed at Ustica, reveal that all the raw materials in the Piano dei Cardoni site are from local georesources.

Exploitation Georesources in Piano dei Cardoni
By recognizing the sand-sized monomineralic granules and lithic fragments and by evaluating their size, morphology and relative frequency, it is possible to acquire insights on raw materials sourcing and the overall production process [Whitbread, 1995;Gliozzo, 2020;Montana, 2020].
From the petrographic analysis of the above-described Neolithic ceramic set, it appears that the majority of the studied samples (29 out of 30) belong to the same fabric and on the basis of their lithology can be considered as a local production characterized by abundant volcanic inclusions of basaltic nature. Based on textural aspects (e.g., approximate firing temperature, aplastic inclusions abundance and size distribution) Fabric 1 can be further clustered into five subgroups. The detailed petrographic descriptions of the microfabric in which the analysed samples have been classified are reported as supplementary material (Supplementary 1), while a schematic description of fabrics is given in Table 4.
Fabric 1a comprises most of the studied samples (n = 16). They were fired at a relatively higher temperature (around 800° C) as can be inferred from the low optical activity of the groundmass. This microscopic fabric is distinguished by the coarser average size and abundance (packing = 15-20% area) of the volcanic aplastic inclusions ( Figure 8A-B). On the contrary, in Fabric 1b (n = 4), aplastic inclusions are less abundant (10% area) and finer ( Figure 8C). Fabric 1c (n = 6) includes samples characterized by the same volcanic aplastic inclusions (qualitatively and quantitatively), being the firing temperature visibly lower than fabrics 1a and 1b (estimated around 700° C by the optical activity of the groundmass) ( Figure 8D). Fabric 1d is characterized by a ceramic body showing numerous cloths composed of microcrystalline calcite, mostly with irregular shape (Figure 8E-F (Figure 10)

The island of the grinding stones
Despite being a small, remote, island with limited resources, Ustica hosts a high variety of useful volcanic resources that probably made it suitable for Neolithic colonizers. The rocks used for anthropic purposes are exclusively of local lithology, easily reachable from the site. All the samples relate to units exposed close to Piano Turchi which are more vesiculated, were used for the upper and lower grinders, while the underwater ones of Cala S. Maria, more compact, were used for the pestles and the hatchets. Of relevance is the presence of easily available pebbles eventually exploitable for different knapping stages of chert and obsidian products. All of this suggests the creation of tools for grain processing, but due to still ongoing archaeobotanical analysis, we cannot exclude the use of grinding tools for the processing of other vegetal resources.
The pumices seem to have been employed as an abrasive for cleaning or similar uses, as they are very rounded but with the presence of vitreous dust inside the vesicles (the pumices naturally remodelled in the sea are instead very clean). Ustica is a key site for the interpretation of lithic assemblages in small islands because both raw resources (chert and obsidian) required transportation by sea from other territories. Additionally, there may have been a preferential channel of exchange and communication with the Aeolian archipelago during these first phases of the occupation of the island. Finally, it is likely that obsidian was imported as a raw material to the island, as testified by the outstanding number of big, medium, and small fragments (sometimes detected only by the 0.5 mm mesh) that attest to knapping activities on the site.
Obsidian in Piano dei Cardoni is the main exotic good at the site. The role of Ustica as a trading place for obsidian towards north-western Sicily can be inferred by several indicators [Speciale et al., 2021a]. At this stage, it is not possible to establish 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 or other goods. Chert was likely imported from the area of Palermo (south-eastern Mesozoic reliefs are characterized by the presence of chert lens, nodules and beds), but due to some cultural aspects, other areas such as the Southern Apennines or the Gargano subregion cannot be excluded. If an intentional modification of used tools is confirmed by future analyses, such as that inferred for two fragments hypothetically originating from sickle elements, it can be interpreted as a need to recycle a raw material naturally lacking on the Claudia Speciale et al.
14 island and probably preferred for specific purposes (functional analysis will explore this possibility). Reuse of sickle elements and double use of the same inserts (suggested by the glossy appearance of sickle elements on both the edges), for example, was a technique for prolonging the use of the flint tools seen also at Lipari during the Late Neolithic at Contrada Diana [Martinelli, 2000]. On the contrary, at a preliminary stage of study, the observation of a high percentage of by-products of obsidian from Lipari may suggest an easier access to this specific raw material.
A similar distribution in the preference of the chert grain size has been observed at Contrada Diana [Martinelli, 2000]. Chert analysis confirms that Ustica has more similarities to the close-to-the-source of obsidian sites, as previously supposed [Speciale et al., 2021a]. Further analyses will help in clarifying its possible source, that can now only hypothetically placed in North-Western Sicily.

Local materials for local ceramics
The majority of the ceramic samples examined showed inclusions (minerals and lithic fragments) that can be associated to the magmatic products characteristic of the island of Ustica [Cinque et al., 1988;de Vita et al., 1998;Peccerillo, 2005]. These ceramics can be distinguished from the coeval productions of Lipari by the absence of intermediate or acidic rock fragments that characterize the magmatism of the Aeolian Island (e.g. potassium-rich andesitic basalts, pumice, obsidian, sanidine phenocrystals) [Williams, 1980;Williams and Levi, 2008;Brunelli et al., 2013]. Furthermore, in the ceramic pastes at Ustica, the femic minerals (especially clinopyroxenes) are much less frequent than that found in pottery produced at Lipari and the plagioclase crystals are not zoned unlike those of the andesites on Lipari. Both compositional and textural aspects of the local ceramic paste provide interesting evidence about potters' choice in raw material sourcing. They selected clayey-sandy soils in the island developed over volcanic deposits, being low calcareous and generally rich in iron oxides. They formed after the pedological evolution of lavas and/or pyroclastic deposits weathered by exogenous agents. Colluvial sediment of this type forms preferably under favourable morphological conditions, like depressions in the bedrock, where they can reach a suitable thickness and collect a greater quantity of clay particles, which gives plasticity and workability. The lumps composed of microcrystalline calcium carbonate (micrite), which were found in several samples, could be ascribed to this process. Hardened CaCO 3 concretions can form in soils by chemical precipitation from infiltration waters.
They are frequently found in arid areas, where even thick layers (caliche) may form. However, in the case of the studied ceramic samples, the presence of small micrite lumps (millimetric or mostly sub millimetric) could be also interpreted as secondary (burial) calcite precipitated directly in the ceramic sherd. This hypothesis is corroborated by the presence of small plagioclase and clinopyroxene crystals incorporated in the microcrystalline calcite mass ( Figure 8F).
The textural variabilities observed in the examined ceramic fabrics could be ascribed to several reasons such as different raw materials sourcing areas or to "natural" variability within the same exploitation point. Nevertheless, there is a striking continuity in using similar raw material for the manufacture of different wares across the period examined and in the tempering procedures (if any). In fact, all the petrographically recognized local ceramic pastes (fabrics) have as a common denominator the same type of volcanic aplastic inclusions (of various size, ranging from 0.04 mm to 2 mm or more).
Only the sample UST 19/013 (Fabric 2) shows the absence of volcanic inclusions and compositional/textural characteristics (concerning both the aplastic inclusions and the groundmass). This suggests that the sample was imported from North-Western Sicily, possibly from the Palermo area. Its surface treatment (a reversed triangle painted in red out of the rim, probably attributable to "tricromica" ware) makes it stand out from the other painted ceramics found at the site.
Further sampling has been planned to better picture local and possible non-local manufacture and to locate the raw material catchment areas.

Conclusions
The first results from the analyses of the georesources in the site of Piano dei Cardoni (Ustica, Italy) shed new light on the exploitation of local and non-local resources on a small volcanic island. Volcanic rocks were easily collected and were located very close to the site (in a range of 2 kilometres). They were selected to achieve the best performance of the grinding and milling tools. The elevated number of grinders, pestles and mortars testify to intense activity revolving around food/plant processing at the site. This was apparently not accompanied by a high frequency of crops, maybe due to taphonomical processes or maybe to the exploitation of wild species.
The absence of obsidian resources on the island pushed human communities to import it from the Aeolian islands.
Chert was also imported although its provenance remains undefined. The re-use of some chert tools leads us to believe that it was scarcer than obsidian. Pumice was collected on Ustica, probably due to local availability and good quality. Most of the ceramic products sampled so far were made using local clay sources and only one sample seems to have come from the North coast of Sicily. The study of local clayey raw materials (already in an advanced state of completion through simulations and experimental firings) and the analysis of a larger number of ceramic finds, will allow a more complete interpretation of the "production chain". In conclusion, the Neolithic inhabitants of Ustica were exploiting all the raw materials available on the island, especially volcanic rocks, while at the same time procuring more distant materials, such as obsidian and chert. Despite being preliminary, this interdisciplinary study reveals that even a distant and small island such as Ustica was not excluded from the circulation of materials and objects that in the Middle/Late phase of the Neolithic occurred in the central Mediterranean.