Contextualizing Bronze Age Obsidian Use at the ‘Ritual Spring’ of Mitza Pidighi (Sardinia)

Abstract This study focuses on obsidian consumption at the ‘ritual spring’ of Mitza Pidighi in west-central Sardinia, Italy. The site dates to the late Nuragic I to Nuragic III phases of the Bronze Age (ca. 1350–850 B.C.) and is found just east of a contemporaneous residential village, Nuraghe Pidighi. While recent years have seen a surge of archaeological literature on the subject of obsidian use at residential sites throughout the island, there has been little consideration of its role in other archaeological contexts, a research bias that this presentation aims in part to redress. For this study, 142 obsidian artifacts from Mitza Pidighi were analyzed non-destructively using a Thermo Scientific ARL Quant'X EDXRF spectrometer to determine their geological origins. In addition, each artifact was analyzed techno-typologically to allow for the reconstruction of the entire chain of events leading up to an artifact's discard. The sourcing results show that obsidians from all four Sardinian subsources are represented at the site, although most come from just one outcrop; the typological analysis indicates that people were physically knapping obsidian near the well to create expedient flake tools and non-prismatic bladelets. In combination, these results have important implications for interpreting the social, economic, and symbolic function of Mitza Pidighi and in understanding the role of obsidian outside of domestic contexts.


Introduction
This study explores Bronze Age Nuragic obsidian consumption at the 'ritual spring' of Mitza Pidighi in west-central Sardinia (Figure 1). Obsidian is a volcanic glass whose circulation in Sardinia during the Neolithic is well documented ( Table 1; see Freund 2014a; Lugliè et al. 2007Lugliè et al. , 2008Tykot 1996). However, the archaeological study of these raw materials in later time periods has received far less attention in what appears to be a critical period that sees the reconfiguration of long standing island-wide exchange networks (see Freund 2014b;Freund and Tykot 2011).
For this study, a total of 142 obsidian artifacts from Mitza Pidighi were analyzed non-destructively at the McMaster Archaeological XRF Laboratory (MAX Lab) using a Thermo Scientific ARL Quant'X EDXRF spectrometer to determine their geological origins. The samples were run under two analytical conditions to generate data (in ppm) for the elements iron (Fe), zinc (Zn), rubidium (Rb), strontium (Sr), yttrium (Y), zirconium (Zr), niobium (Nb), and barium (Ba), elements already shown to be successful in distinguishing between the various West Mediterranean sources and subsources (Freund 2014b;Le Bourdonnec et al. 2005;Tykot et al. 2013). In addition, each artifact was analyzed techno-typologically to allow for the reconstruction of the entire chain of events leading up to an artifact's discard.
The sourcing results show that obsidians from all four Sardinian subsources are represented at the site, although most come from just one outcrop. The typological analysis indicates that people were physically knapping obsidian near the well to create expedient flake tools and non-prismatic bladelets, making this one of the first detailed discussions of Nuragic bladelet production to date. By integrating raw material sourcing and techno-typological analysis, this broadbased artifact characterization study has important implications in interpreting the social, economic, and symbolic function of Mitza Pidighi and in understanding the role of obsidian outside of domestic contexts.

Background Information
There are four obsidian sources in the West Mediterranean, although only obsidians from the four Sardinian subsources at Monte Arci are known to have been exploited by people on the island itself. These subsources include SA, SB1, SB2, and SC ( Figure 2). Despite the presence of Mesolithic populations on Sardinia, it is not until the beginning of the Neolithic that we see the first evidence of obsidian use by the islanders and the long-distance procurement of Sardinian raw materials by populations on Corsica and mainland Italy (Bigazzi and Radi 1998;Freund and Batist 2014;Le Bourdonnec et al. 2014;Léa 2012).  While the long-distance procurement of Sardinian obsidian continues into the Chalcolithic, used by communities on Corsica and mainland Italy (see Bigazzi and Radi 1998;Hallam et al. 1976;Randle et al. 1993), there is a sharp fall-off in the number of sites at which obsidian has been reported. The declining use of Sardinian obsidian in the Chalcolithic and into the Bronze Age mirrors the diminished exploitation of other West Mediterranean sources in the third to second millennia B.C. in that obsidian consumption becomes a more local phenomenon, largely restricted to populations within the immediate vicinity of the source areas (Freund 2014b). Nevertheless, there is a complexity to Bronze Age obsidian exploitation on Sardinia that has yet to be fully explored.

Site Background
Mitza Pidighi is found at the border between the Campidano Maggiore plain the Paulilatino basalt plateau and dates to late Nuragic I to Nuragic III phases of the Bronze Age (ca. 1350-850 B.C.). The site consists of a natural spring surrounded by an oval-shaped construction of basalt blocks approximately 15×6 meters in size (Figure 3), a common construction type found throughout the island and has been hypothesized to be related to a range of ritual activities (see Usai 2000Usai , 2013. The excavation was led by Dr. Alessandro Usai from 1999Usai from -2008 and is contemporaneous with the village of Nuraghe Pidighi found 20-30 meters west of the site. Since the spring is not a residential structure, and likely served a different social and religious function, the obsidian artifacts from Mitza Pidighi were treated separately. In total, 801 chipped stone artifacts were recovered from Mitza Pidighi, of which approximately 78% was obsidian and the remaining material black/ gray rhyolites.

Sampling
All of the lithics recovered from the excavation were bagged together according the unit and stratum from which they were recovered. In order to get a representative sample of artifacts recovered throughout the site, all of the materials were laid out in their provenience-specific bags and one-third of the bags (i.e. every third) were randomly selected for analysis. This percentage was chosen because it would result in a robust enough sample to make valid interpretations while still fitting within logistical constraints such as time and money.
Because of the visual similarities between the black/ gray rhyolites in the assemblage and the obsidian, it was not possible to distinguish between the two with any certainty. Therefore all artifacts, including the rhyolites, were elementally analyzed. In the end, a total of 142 (23%) obsidian artifacts from Mitza Pidighi were analyzed. Out of this selection, two artifacts were too small to  be run by EDXRF (cf. Davis et al. 1998), but were nevertheless included in the typological analysis as their presence in the assemblage has important implications. Since it was necessary to catalogue and photograph the entire assemblage before export, it can be confirmed that that the selected material is representative of the whole in that it is mainly composed of flakes, bladelets, angular waste, and non-cortical cores.

Analytical Procedures
All artifacts were analyzed non-destructively at the McMaster Archaeological XRF Laboratory (MAX Lab) using a Thermo Scientific ARL Quant'X EDXRF spectrometer. While EDXRF has its limitations (see Davis et al. 1998;Shackley 2011: 8-10), this technique was chosen for two main reasons. The first is that EDXRF allows for the relatively rapid, non-destructive analysis of large numbers of artifacts, in turn offering archaeologists a chance to address a wide range of relevant research questions. Moreover, EDXRF is particularly useful in the analysis of homogenous materials such as obsidian and has been used with increasing frequency in obsidian sourcing studies from around the world (Freund 2013). Since the voltage and amperage settings of XRF machines are fixed, they can only create enough energy to analyze elements within a specific range on the periodic table. It just so happens that the mid-Z elements within this range are found in minor and trace quantities in West Mediterranean obsidians and can be used to discriminate between the products of the various geological sources (Crisci et al. 1994;De Francesco et al. 2008;Freund and Tykot 2011).
Before analysis, each piece was given a unique 'Mac' number and subsequently cleaned in an ultrasonic tank with distilled water for ten minutes following the analytical protocols and methods devised by Shackley (2005, appendix) and Carter et al. (2013). The spectrometer is equipped with an end window Bremsstrahlung, air cooled, Rh target, 50 watt, X-ray tube with a ≤7.6 micron (0.3 mil) beryllium (Be) window, an X-ray generator that operates from 4 to 50 kV in 1kV increments (current range, 0 -1.98mA in 0.02mA increments), and an Edwards RV8 vacuum pump for the analysis of elements below titanium (Ti). Data is acquired with a pulse processor and analog to digital converter. In this case, the samples were run under two analytical conditions. The samples were first run under a Mid Zb analysis condition with the X-ray tube operated at 30 kv using a 0.05 mm (medium) Pd filter in an air path for 200 seconds livetime to generate X-ray intensity Kα-line data for elements iron (Fe), zinc (Zn), rubidium (Rb), strontium (Sr), yttrium (Y), zirconium (Zr), and niobium (Nb). The second was a High Zb analysis condition with the X-ray tube operated at 50 kv using a 0.63 mm (thick) Cu filter in an air path to detect the element barium (Ba).
Trace element intensities were converted to concentration estimates by employing a least-squares calibration line ratioed to the Compton scatter established for each element from the analysis of international rock standards. These comprise AGV-2 (andesite), BCR-2 (basalt), BHVO-2 (hawaiite), BIR-1a (basalt), GSP-2 (granodiorite), QLO-1 (quartz latite), RGM-2 (rhyolite), SDC-1 (mica schist), STM-2 (syenite), TLM-1 (tonalite), and W-2a (diabase) from the US Geological Service [USGS], plus JR-1 and JR-2 (both obsidian) from the Geological Survey of Japan. In turn, the USGS standard RGM-2 is analyzed during each sample run to check machine calibration and accuracy. The data was then translated directly into Excel for Windows software for manipulation and analysis. Artifacts displaying anomalous values were re-run to ensure accuracy and precision. The quantitative results were ultimately compared with data generated from geo-referenced in situ geological samples of obsidian using the same instrumentation.
In addition to elemental characterization, each artifact was analyzed techno-typologically. This included recording attributes pertaining to flaking type (platform, bulb, lip etc.), to understand how the blanks had been knapped. Artifacts were also categorized as nodules, cores, flakes, blades/bladelets, or angular waste, data that along with presence of cortex (divided into distinct percentage categories) allowed for the reconstruction of the obsidian reduction sequences. Only the percentage of dorsal cortex was recorded on the flake categories. This in turn allowed for the identification of the various forms in which obsidian entered the site prior to its reduction. Finally, any form of deliberate modification in the form of retouch was documented as a means of describing Nuragic tool types.

Results
Table 2 provides a compiled list of the sourcing results from Mitza Pidighi. Based on an examination of the bivariate plot of the elemental ratios of rubidium (Rb) and strontium (Sr) to niobium (Nb), it is clear that Sardinian SC obsidian was the predominant subsource utilized at the site, comprising approximately 80% of the assemblage (Figure 4). Nevertheless, 11 SA artifacts were also found as well as seven SB1 and 11 SB2 artifacts.
There is clear evidence for the primary reduction of obsidian at Mitza Pidighi, where the entire reduction sequence is represented by the presence of cores as well as flakes, bladelets, and angular waste ( Figure 5). The entire assemblage (n=142) weighed 0.3 kg. and contained three bladelet cores, 13 flake cores, 19 bladelets, 95 flakes, and 11 pieces of angular waste. Eighteen percent (n=26) of artifacts had at least some cortex, although 15 of these 26 artifacts only displayed between one and 20% cortex. Table 3 provides a breakdown of blank types by source. Table 2 Elemental data of Mitza Pidighi obsidians and in situ geological specimens as determined by EDXRF in parts per million (ppm). RGM-2 standard data is also included. Note that 'Mac' numbers are not consecutive because several pieces turned out not to be obsidian, but rather the rhyolites mentioned in the text, often similar to opaque SC obsidians.  At Mitza Pidighi, over half (55%) of the artifacts were intentionally modified, including nine of the 11 SA artifacts, six of the seven SB1 artifacts, six of the 11 SB2 artifacts, and 50 of the 111 SC artifacts. Despite their low total numbers, non-SC materials showed a relatively high percentage of retouch, and there is clearly something going on at the site that is not happening at other Nuragic sites. A consideration of tool categories further highlights these distinctions. Mitza Pidighi differs from contemporaneous residential sites in the lack of lunates and in the high  number of notched pieces (n=15) and pièces esquillées (n=9); three end scrapers and two denticulates were also present ( Figure 6). Lunates are best defined as small elliptical flakes with a natural or artificially backed margin (Freund 2014b, 246;. They are characteristic Nuragic artifacts that are surprisingly absent from Mitza Pidighi, but nevertheless present at the nearby site of Nuraghe Pidighi 20-30 meters west of the spring (Freund 2014a).
Bladelets comprise about 14% (n=20) of the assemblage while they are virtually absent from most known Nuragic sites. They are, however, found at Nuraghe Pidighi (Freund 2014a). Most of the bladelets are small and have parallel margins and a uniform thickness across their length (Figure 7a). They are too small to be produced through direct percussion, as evidenced by corresponding bladelet cores (Figure 7b), and it is for these reasons that I argue that bladelets at Mitza Pidighi were produced using a mode of pressure flaking (see Andrefsky 2005, 118-119;Crabtree 1972, 44). While bladelets are mainly made from SC obsidian, they were also created from the other subsources, despite the lack of cores.

Discussion and Conclusions
As expected, all of the obsidian found at Mitza Pidighi comes from the Monte Arci source located approximately 20 to 25 kilometers to the south. The predominance of SC obsidian at the site is in keeping with general patterns of obsidian exploitation at other Nuragic sites in Sardinia (Figure 8). However, while the sourcing results indicate that obsidian use at Mitza Pidighi is not unique in terms of the materials being exploited, the techno-typological analyses brought forth patterns of exploitation not previously documented in central Sardinia.
The results of the analysis indicate that people at Mitza Pidighi were physically knapping obsidian near the well to create expedient flake tools and non-prismatic bladelets. This is of particular interest because bladelet production has not been previously reported at contemporaneous Nuragic sites in central Sardinia, other than the nearby site of Nuraghe Pidighi 20-30 meters west of the site (see Freund 2014a; Freund and Tykot 2011;Locci 2005). Mitza Pidighi is also unique in that there is a surprising lack of lunates, a characteristic Nuragic artifact type, and in the high number of notched pieces and pièces esquillées found at the site. While the presence of these artifact types might be expected at residential sites, their presence at a 'ritual spring' would seem to indicate that a diverse range of activities were carried out near the well itself.
While Mitza Pidighi is not a residential site, it must certainly be expected that these materials were brought to the site by the inhabitants of the nearby village of Nuraghe Pidighi and possibly by others within the wider region. However, because of the broad similarities in both the sources represented and in the artifacts found at Mitza Pidighi (most notable bladelets), it is likely that the nearby residents of Nuraghe Pidighi were the ones knapping these materials as opposed to distant communities and peoples.
This study makes an important contribution to our understanding of obsidian consumption in Bronze Age Nuragic Sardinia and has important implications for understanding the role of obsidian outside of domestic contexts. More broadly, this research highlights the importance of an integrated approach to provenance studies in archaeology, i.e. one that combines elemental characterization data with techno-typological analysis to extrapolate the full suite of behaviors behind people's engagement with archaeological materials.
auspices of the Regione Autonoma della Sardegna. I also thank the Soprintendenza per i Beni Archeologici di Cagliari for permission to export the artifacts to Canada for analysis; all artifacts have been safely returned. My work in Sardinia was funded by an Edith M. Wightman Travel Scholarship, an Ontario Graduate Scholarship (OGS), and a McMaster Department of Anthropology Travel Grant. The elemental characterization was undertaken at the McMaster Archaeological XRF Lab (MAX Lab), which was created and is currently financed by a Canada Foundation for Innovation Leader's Opportunity Fund / Ontario Research Fund (awarded to Dr. Tristan Carter). I also thank Deanna Aubert for artifact photography and Dr. Carlo Lugliè for help in procuring geological material.

Conflicts of Interest
The author declares no conflicts of interest.

Author Biography
Dr. Freund's research and publications span a range of specializations, including artifact characterization studies, spatial statistics, and Mediterranean prehistory. Dr.   of exchange in structuring prehistoric society as well as the influence of incipient metallurgy on the stability of pre-existing social structures. Central themes in his work include: a) the dynamics, form, and pace of cultural change, b) power and inequality in small-scale societies, and c) the socio-economic significance of exchange and long-distance relations.