THE TERRACOTTA ANIMAL FIGURINES FROM DESPOTIKO: THE LIFE OF HUMANS AND OBJECTS IN THE EARLY IRON AGE CYCLADES BEYOND POLARITIES

Mandra, on the uninhabited islet of Despotiko in the middle of the Aegean Sea, is well known to the archaeological community, owing to the discovery there in 2001 of an extensive sanctuary of Apollo. Twenty-two edifices have come to light so far, and the systematic excavation continues to elucidate the long history of the site. The Early Iron Age marked the earliest activity there, traces of which offer fertile grounds for reconsidering life in the Cyclades at the time. The richest evidence for this period is offered by a secondary deposition, detected near two Early Iron Age buildings, which revealed thousands of clay sherds, extending from the late ninth/early eighth to the late sixth century BC, quantities of animal bones, and more than 60 metal objects. This article focuses on a small group of Early Iron Age terracotta animal figurines from this deposition. Critically analysing both their association with ritual and the polarity of ritual and profane, an attempt is made to unravel the lifecycle of these figurines, treating them as agents of activity. Their function and meaning are interwoven with the activities operating at the site during the Early Iron Age, at least two centuries before the foundation of the Archaic temenos.

; , ). A handmade phallus from the interior of a hearth in a rock cavity below a Late Geometric structure of unclear character (Schilardi , , n. ; , -; Mazarakis Ainian , ) has been assigned to a domestic cult (Averett , ).
Two terracotta figurines have been reported from the settlement of Zagora on the island of Andros. The torso of a horse, bearing an insiced decoration of the Middle Geometric II or Late Geometric I, was discovered close to the surface near the gate of the settlement's fortification wall.  A Late Geometric fragmentary clay birdwhich could have been attached to a vessel or a lidwas found in the area of the temple (Cambitoglou , , no. , fig. ; Cambitoglou et al. , , pl. a-c). Kourou (, -) correlated this figurine with those from the South Cemetery of Naxos and assigned it to the same Naxian workshop.
The Early Iron Age terracotta figurines from the Cyclades appear in all kinds of contexts: domestic, cultic, or ritually charged, but also funerary. With the exception of the Naxian 'grave' figurines, deriving from a primary context, the rest were recovered from secondary contexts, not stratified depositions. Most of them have been the subject of only preliminary publication, and have not been treated as a distinct categorysince they were not found in any concentrationand were discussed by the excavators alongside the rest of the finds. They have been very often considered as providing criteria for reconstructing cult activity, as gender/character identificators of the venerated deity, and occasionally as bearers of symbolic or social meanings.
This corpus of the early terracotta Cycladic figurines has been recently enriched by a small assemblage, discovered at the site of Mandra on the uninhabited islet of Despotiko in the middle of the Aegean Sea. This paper contextually analyses these specimens together with the material found alongside them and treats them not simply as 'correlates' of ritual, but as agents of activity with a biography and, more interestingly, the potential to eludicate human perceptions of animals while enlightening aspects of life at the site during the Early Iron Age.

THE DESPOTIKO TERRACOTTA FIGURINES AND THEIR ODDITIES
Despite the well-established fluidity of their meanings and functions, Early Iron Age figurines are still largely treated by scholars as ritual, or at least ritually charged, objects and as material correlates of such activities, often independently from their archaeological context.  Accordingly, trapped in a circular argument, the figurines discovered in sanctuaries acquire a religious significance, and consequently, a site/area, a structure/building, or a practice is designated as cultic (Tzonou-Herbst , ).
The Early Iron Age figurines from Despotiko formed part of a secondary deposition:  an extended assemblage, whose contents were intentionally discarded sooner or later after their initial use, for use as fill in construction work. The micro-context/exact location of the figurines within this deposition, as well as their conditiondifferences in the fragmentarity and degree of wearcould illuminate their primary use, overcoming the lacunae caused by the lack of stratigraphy or evidence for episodes of disturbance (Thurston , -). Their associated finds within their contemporary structured environment are also of primary importance in this respect. Moreover, any unique characteristics, differentiating this group of figurines and its context from its contemporaries in the Cycladic milieu, are indispensable for better grasping their function.  The figurines might have been recovered beneath the Archaic temenos of Apollo, founded approximately in the middle of the sixth century, but, nevertheless, their biography, consisting of   The same model has been proposed by M. Vetters (b, ) for the study of figurines from domestic contexts. THE TERRACOTTA ANIMAL FIGURINES FROM DESPOTIKO  stages in their lifecycle, needs to be carefully reassembled, before assigning them a priori religious significance.  Considering them within a firm contextual frame, it is worth treating the Despotiko figurines as biographical things, whose life narrative is interwoven with human lives, actions, feelings, and memories. The biographical approach to artefacts, understood as integral to human behaviour and activity, has been increasingly employed in archaeological research during the last few decades.  The emphasis has been placed either on people's material lives with the objects inseparable from their personhood and identity (e.g. Hoskins ; Whitley ) or on the social lives of the objects as shaped within different contexts (e.g. Langdon ; Shanks ). Οn a secondary level, the figurinesall of them animal representations in clayshould be treated as subjects, supplied with agency, for yielding possible insights into the connections between humans and the natural world.
The eight terracotta figurines studied here include two bovids, two large birds, the tails of two more birds or bird askoi, a small bird, and a snake head. None was recovered complete, but the breaks are sharp, their surface is not worn, and the paint is well retained.
The best-preserved example is a handmade, solid bovid with a long, cylindric torso, strong thighs, and well-formed genitals and tail (Fig. ).  The head and the lower part of the legs are missing. It is made of a coarse grey to brown fabric, containing large portions of silver mica visible on its rough surface. Its body is significantly affected by fire with a large crack running along a central line over its back. The second bovid, consisting of coarse reddish fabric with a finer and less crude surface, must have been of smaller dimensions. All that is preserved is one of its carefully shaped legs with a hoof (Fig. )   This approach has been adopted for the Mycenean figurines in the past decade: Tzonou-Herbst ; ; Thurston .  The idea on the social lives of things has been developed by Appadurai () and Kopytoff ().


The hollow terracotta bird, recovered in two large parts, is made of a deep reddish brown rough fabric with a grey-black core, containing silver mica (Fig. ). The figurine has a small head, distinctly rounded on top, and a beak of slightly oblong section, whose edge is missing.  Its short, cylindrical, neck leads to a closed wing, preserved only on one side. The details are provided by black paint, thinning to brown. The eye is represented as a round convexity surrounded by two painted rings. Two more rings appear on the neck, while the plumage is indicated by two leaf-shaped borders, the inner of which bears slightly diagonal bands.
The Despotiko bird is quite unique. Although a figurine, it should be better compared with bird askoi.  In particular, it stands close to those of Desborough's () Type I and Guggisberg's (, ) Type A.  The earliest examples of this type are known from Achaia (Guggisberg   It is . m high and . m wide. Its cylindrical neck has a diameter of . m and the head is . m long.


The Eary Iron Age bird askoi have been well treated by research due to their importance for studying connectivity and exchanges between Cyprus and the Aegean (Desborough ; Guggisberg , -; Kourou ; ).  The rendering of the beak points to the horse-bird askoi, but both the head and the painted decoration clearly indicate a bird rather than a horse (Coldstream ).  (Vlachopoulos , vol. Β, chapt. , no. ) and date to the twelfth century BC. After a gap of slightly more than a century, bird askoi are attested on Crete and the Dodecanese from the ninth century onwards (Lemos , , ). The only bird vessel rather than askos from the Early Iron Age Cyclades comes from the Naxian necropolis of Tsikalario, and it has been dated to the Middle Geometric period (Charalambidou , -, fig.  A date of the second half of the eighth century is proposed here, on the basis of the figurine's painted decoration, which brings it very close to the birds depicted on many of the Parian Late Geometric vases from the same deposition. It is striking that, although of such a late date, the terracotta bird seems to strongly preserve the traditions of the twelfth and tenth centuries BC, known from Achaia and Crete, respectively. More interestingly, it is completely different from all Naxian examples, both those of the Late Helladic IIIC (Kamini: Vlachopoulos , vol. B, -, nos , , -, fig. , colour pls -, pls -) and the Middle Geometric periods (Tsikalario).
The bird might not find parallels outside Despotiko, but the wing of a contemporary example from the same deposition suggests that at least two existed at the site. The wing is similarly painted: the triangular endings of the double-outlined plumage, whose interior preserves a diagonal band, are bordered by bands along the wing's edges (Fig. ).

ALEXANDRA ALEXANDRIDOU
 tail is decorated with horizontal bands (Fig. ).  Were it not a bird askos, this banded tail could belong to a handmade bird, rather like the Naxian examples. We cannot conclude firmly whether or not the Despotiko birds date to the ninth century: the horizontal bands on the birds' bodies could point to a later date (e.g. Guggisberg , , nos , , pl. :).
The flattened snakehead of triangular shape could have been an independent figurine, if it were not attached to a vessel or another object (Fig. ).  The eyes are denoted as deep, incised, circles around two round low convexities. Remains of concentric small circles in almost completely flaked black paint can be seen around the eyes with horizontal and vertical lines on the edge of the head and body. Once again, the painted decoration seems to point to the latter part of the eighth century (cf. Sinn , nos -, pl. :,).


Mandra of Despotiko is one of the few Cycladic sites where a concentration of figurines, rather than a few examples, is attested. The different types of terracotta animals attested at the site might be consistent with the enrichment of the figurine repertoire in the Aegean after the ninth century BC (Averett , , ), but their heterogeneity is exceptional for the Cyclades. Bovids might have been common at the time, but the snake is an extremely rare type; moreover, the idiosyncratic nature of the large birds is noteworthy too.   The middle of the sixth century saw a radical transformation of the site (Fig. ): a monumental temenos protected by an almost square enclosure was erected over an area of c.  m  , as part of a conscious, costly, and ambitious construction programme, run by the thriving Archaic polis of Paros (Kourayos and Daifa ). Building A, composed of five rooms, occupied the west part of the temenos, facing towards the entrance of the harbour. Its north part -Rooms A and Ahas been recognised as the temple of the temenos. Its south part -Rooms A to Amight have served as a feasting hall. More buildings were erected, creating a densely built grid around the sacred peribolos. The north-west part was occupied by the temple-shaped Building Δ, and the eastern part by the late sixth-century Building E (Kourayos et al. ). The still ongoing excavation has brought to light many more buildings at Mandra - in totalpart of an extended establishment, which extended over to the smaller islet of Tsimintiri bordering the bay of Mandra (Alexandridou and Daifa in press).
The parts of two Early Iron Age buildings, which came to light recently (-) at a deeper level below the Archaic temenos, reveal the occupational history of the site, which can be now placed well before the sixth century BC (Fig. ). Building O with a north-south orientation was found lying just east of the porch of Building A's Rooms A and A and south of the porch of Building Δ, just inside the north-west corner of the Archaic temenos. The building preserved a north apse and a large part of the east wall, running north-south with its lower edge slightly curving. No remains of the west wall were detected, while only a few stones have fallen from the south wall. The width of the preserved walls indicates that the building was of large dimensions with the mudbrick upper structure placed on a stone foundation. The building must have been established in the late ninth or the first half of the eighth century. It remained in use for some decades until the construction of Building Ξ at the end of the eighth century. This rectangular building, retaining only a small part of its north and west walls, is in exact alignment with Building A, with which it shares the same orientation. It was found at a distance of . m from the east walls of the Archaic building's north part (Alexandridou , -). A rectangular eschara of stone slabs, . m long and . m wide, was found . m east of the building's north wall at the same depth as its foundations. The lower part of two large pithoi, both fireaffected, were detected in situ by its north-west and north-east internal corners. A kantharos and Α muddy layer of soil, . m to . m thick, mixed partly with ashes, covered a wide area extending from the northern part of Building O to the porch of Building Δ and further to the north, as well as below the north-west room of the structure in shape Γ connecting the Archaic Buildings A and Δ. The core of this deposition was detected below the porch of Building Δ, and inside Building O, especially towards its north apse. Except for the eight terracotta animal figurines, this layer contained thousands of sherds from a variety of clay shapes, chronologically extending from the ninth to the second half of the sixth century BC, as well as numerous metal objects, two Egyptian scarabs and quantities of animal bones. Clay sherds were found widely dispersed under the Archaic building, but also in its surroundings, with their concentration thinning out towards Building Ξ. On the other hand, almost all terracotta figurines and metal objects were found concentrated in a trench just west of Building's Δ porch.
Although most of the Early Iron Age components of this layer are compatible with Building Ξ's period of use, spanning the second half of the eighth century, it is not at all certain that they originated from the building's interior nor that they should be related to it. They may well represent the archaeological remains of activities, operating either in the open air or in association with another structurenot yet discoveredin both cases in close proximity to these buildings. The wide chronological spectrum of the pottery discovered, including many Archaic vessels, is an additional argument for assigning the activities to different loci.


In all cases, the clay, metal, and other objects, as well as the zooarchaeological remains, have been removed from their original context, before being transformed into construction debris for the foundation of Building Δ of the Archaic temenos in the second half of the sixth century.

The pottery
The terracotta figurines might have been few in this deposition, but the sherds of clay vases amount to almost , in number.  Their surface remained smooth with no wear or depositions, and their painted decoration is not flaking. Fine-decorated vessels largely outnumber coarseware. Although each vase is represented by a single or a few fragments, their broken edges are sharp. A few sherds have slight remains of secondary burning. The earliest vases date to the Middle Geometric period and the ninth century BC, but the largest bulk of the Geometric pottery spans the second half of the eighth (Alexandridou in press). Most of the deposited sherds are Archaic, extending chronologically to the second half of the sixth century.
The Geometric assemblage is dominated by small open vessels, mainly skyphoi followed by kantharoi (Fig. ). Large open shapes, most possibly kraters, as well as plates, are numerous. On the other hand, closed formsamphoras or hydriaiare limited. Except for a few Protocorinthian kotylai and three conical lekythoi of the 'Argive Monochrome' ware,  the rest  The clay sherds have been quantified. The results of this quantification will be discussed in the final publication of this material, which is under preparation by the author.
 Their fabric indicates an Attic rather than a Peloponnesian origin. At least one can be assigned to the Kourou's workshop of the 'Toothed Wheel' (Kourou ).

THE TERRACOTTA ANIMAL FIGURINES FROM DESPOTIKO
 of the Geometric vases are Parian products presenting a large homogeneity in their technological features.  It is necessary to juxtapose the Geometric vases from the Despotiko deposition with the contemporary assemblages of vases from Antiparos, Paros and Delos/Rheneia in order to discern any oddities or special characteristics.
Sherds of Early Iron Age skyphoi and kantharoi, an oinochoe and another closed shape (Bakalakis , -, fig. :-) are published from the cave situated at the south-eastern part of Antiparos with a view across to Despotiko. The pottery could have been used in ritual meals held in a cave whose cultic use in the historical period is securely attested by an inscription on a stalagmite (Bakalakis ; Mavridis -; Angliker ).
Koukounaries on the naturally fortified hill, south-west of the Naousa bay on the north side of Paros, provides the ideal case for attempting a comparison with the site of Mandra, since domestic structures coexisted there with a sanctuary. After the destruction of the Late Mycenaean fortified acropolis at c.  BC, the hill accommodated a settlement, occupied from the Protogeometric period until its peaceful abandonement around the middle of the seventh century BC. A sanctuary, dedicated to Athena, was located on the hill's south-eastern slope. According to the preliminary reports and articles, many skyphoi and kantharoi were recovered from the houses of the Upper Plateau, while coarse ware and relief pithoi are also mentioned. The ongoing study of the material attests to the proliferation of skyphoi with linear or bird decoration, painted solid one-handled cups, and skyphos-kraters spanning the Middle and Late Geometric periods.  According to the excavator, pottery dominated the various offerings to the deity of the site's sanctuary. Open shapes are found broken near to the altar together with burnt animal bones. A variety of Parian Geometric clay shapes, including oinochoai, skyphoi, tripods and fenestrated supports (Schilardi ), as well as a small portion of Protocorinthian sherds, are reported (Schilardi , -; , ).
An Early Iron Age clay assemblage from Paros comes from the Delion, situated on top of a low hill, c.  km north of Paroikia (Rubensohn ). The pottery was found in a very fragmentary state almost exclusively below the foundation of the temple's cella. Most vases date to the Geometric and Archaic periods.  On the basis of the published material, open shapes, mainly skyphoi, dominate, combined with one-handled cups and plates. Large open vessels, kantharoi, amphoras or hydriai are represented by a few fragments, while painted solid juglets were more common. 'Argivemonochrome' ware is present too.
The Early Iron Age Parian material from the Delion provides the closest parallel to the Despotiko assemblage, though the number of both large, closed shapes and kantharoi is greater at Mandra. More interestingly, the quality of the vases from Despotiko is higher, with a much larger number of vases decorated with birds or other animals. On the other hand, the Delion seems to have attracted more Corinthian vases of a variety of forms, including aryballoi, alabastra, kotylai, and oinochoai. The Corinthian imports at Despotiko are limited to a few kotylai, most possibly of the Middle Protocorinthian period.
A large assemblage of Early Iron Age Parian potterynot yet fully publishedwas discovered in relation to the two polyandria in the main necropolis of Paroikia (esp. Zapheiropoulou ; ; ; , -; Agelarakis ). Medium-sized and small amphoras served as urns for the cremains of the deceased, with skyphoi or cups sealing their mouths (see e.g. Zapheiropoulou  Although plastic snakes decorate the shoulder of two Late Helladic IIIC strainer hydriae from Aplomata on Naxos (Vlachopoulos , vol. B, -, nos , , pl. ), they are absent from contemporary Parian or Cycladic vases. It should, however, be noted that painted snakes decorate the vertical handles of the Late Geometric amphora from the polyandrion of Paroikia on Paros (Archaeological Museum of Paros , Zapheiropoulou , -, no. ).
 I would like to thank Thanasis Garonis, who is currently studying the material from the Upper Plateau, for this information.
 It is not clear how much of the discovered material is included by Rubensohn () in his final publication. The recent publication of the Corinthian imports from the sanctuary by Detoratou (-) implies that more vases than those published by Rubensohn were found. . Athough no other grave offerings are reported, several Protocorinthian vases have been recently published, whose exact context and use at the necropolis, however, is not mentioned (Detoratou -).
Moving outside Paros, two assemblages of Parian pottery are of interest: the material from the purification pit of Rheneia (see Dugas ; ; Dugas and Rhomaios ; Zapheiropoulou ), and that recently published from the sanctuary of Apollo on Delos (Brisart ). The clay forms from Rheneia do not differ significantly from those coming from the necropolis of Paroikia. Closed vasesmostly hydriai rather than amphorasdominate, due to their use as ash urns, followed by oinochoai. The repertoire of the open shapes includes skyphoi and kantharoi, as well as plates (Brisart , ), a shape absent from the necropolis of Paroikia, but present at both the Delion of Paros and Despotiko.
Parian pottery is attested among the ceramic finds from the 'hieron' of Apollo on Delos (Brisart , -). The recently published sherds from the sanctuary make up just a small fraction of the excavated material, which remains unpublished, but they provide some important insights into several issues concerning their context, function, and chronological range, as well as the origin of the visitors at the site from the Early Iron Age onwards. Although the earliest vases from the sanctuary date to the Protogeometric period, the second half of the eighth century marked an increase in the volume of fine-decorated pottery, as well as a diversity in the represented styles originating from various workshops. Two Late Geometric hydriai or amphoras of the Groups Aa and Ad, respectively, are of Parian origin (Brisart , , nos -, pl. :-). Their presence at the sanctuary is important, since the shape was until now exclusively known from the purification pit of Rheneia (Brisart , ). Brisart chose to examine the pottery from the sanctuary through the 'angle de la céramique funéraire', appearing quite reluctant to acknowledge any non-funerary function for many of the shapes. This scepticism has also affected two Parian amphoras, the second of which is of large dimensions, which he did not wish to interpret as possible votive dedications for the deities (Brisart , -).
The Early Iron Age ceramic assemblage from the sanctuary of Hyria on the neighbouring island of Naxos should be brought into this discussion, owing to the volume of the data and its similarities with those from Mandra. Open shapes represent  per cent of the Early Iron Age fine wheelmade ware. Small-and medium-sized drinking shapes, mostly skyphoi, dominate. One-handled cups, kotylai, and a few kantharoi complete the repertoire of small open shapes, which included also skyphoid kraters and kraters. Forty per cent of these drinking vessels are painted solid, though skyphoi with linear decoration, birds and horses are not absent. According to Simantoni-Bournia, the decoration is not as varied and ornate as in most sanctuaries. Only a few Protocorinthian imports are reported, dating after the Early Protocorinthian period (see esp. Simantoni-Bournia a). The quantity of eating, drinking and cooking vessels, together with the burnt animal bones and the ashes from the early temples (I-III), suggest that ritual dining played an important role in the sanctuary's cult life (Lambrinoudakis , ; Simantoni-Bournia , -; a, ).
Despite the lack of detailed publications of the pottery from most of the above-mentioned sites, rough comparisons with the ceramic assemblage from Mandra can be attempted. Based on the above-presented data, all shapes from the examined deposition are attested in all kinds of contexts, domestic, cultic or funerary. The dominance of closed shapes is the main distinguishing feature of the Despotiko assemblage compared to that from the necropoleis. On the other hand, the concentration of open shapes, mostly of small drinking forms, is a common feature in settlements and cult sites. The plates, well represented at Mandra, might be attested at the Delion of Paros, but they are not absent from the pit of Rheneia. 'Argive monochrome ware' is present at the Parian Delion (Paros Archaeological Museum AK ) and on Delos (Dugas , cat. no. , pl. ), but absent from Rheneia and the necropolis of Paroikia. It is worth noting that some of the specimens from the Artemision of Delos are Attic products (Brisart , ), comparably to those from Despotiko.
The large number of kantharoi and cooking vessels, as well as the high quality of all finedecorated vases, often depicting birds and horses, distinguish the Despotiko assemblage from those from Paros, Delos and Naxos. The aesthetic value of the kantharoi is characteristic. Unlike those from the purification pit of Rheneia, where they are numerous, they are not decorated with linear patterns but with animals (Alexandridou in press).

The metal objects
The Despotiko deposition contains  metal objects, with iron items (of which there are ) dominating over bronze. They are all badly eroded and fragmentary, with their condition in many cases impeding their identification.  They belong to two main categories: items of personal ornamentlimited to fibulae and ringsand tools/weapons exclusively represented by knives, which form the core of this metal assemblage.
Six very fragmentary bronze arched fibulae have been recovered, the better preserved of which have globules on the bow. Two bronze and two iron rings, all solid and closed, complete the identifiable group of the metal pieces of jewellery from this context. Five bronze and silver rings of the same type were included in the deposit, detected below the floor slabs of Room A of the Archaic 'temple' (Kourayos and Burns , ).
At least  knives, most of which preserve part of their blade, are included among the iron objects. One of them seems to have been 'ritually killed' (Petrakis , ). Knives might not have been the only iron objects in this deposition, but their high number should be related to the large assemblage coming from the deposit of the 'temple's' Room A, the largest known in the Cyclades to date.  These knives, some of which are of large dimensions, have been interpreted as votive dedications serving sacrificial practices, but more importantly revealing a strong male focus (Kourayos and Burns , -). Knives are attested in most cult sites of the Greek mainland and the islands with their numbers being particularly high during the Early Iron Age and the Archaic periods. In this context, they have been associated with animal sacrifice, necessary either for slaughtering the sacrificial victim or for cutting the meat that had to be distributed to those participating in the communal meal (Sossau , -, -; Petrakis , -).

Τhe Egyptian scarabs
The excavated deposition contained two Egyptian scarabs, which either date to the Late Bronze Age IB-IIB period or are imitations of Egyptian prototypes. In the former case, they are heirlooms and the only Bronze Age objects from Mandra so far identified. The first scarab bears the motif of the four cobras, pointing to the Second Intermediate Period. The second example shows Pharaoh smiting an enemy with a mace, possibly of Tuthmosis III (- BC).  Aegyptiaca, Egyptian and Egyptianising artefacts, were widely distributed in the Mediterranean with a special concentration in the Aegean, where they have been mosty discovered in votive depositions in coastal sanctuaries. These objects date from the eighth to the sixth centuries BC (see esp. Hölbl ; ; Kousoulis and Morenz ). The Rhodian sanctuaries might have yielded the largest portion of these objects (Apostola and Kousoulis  with further references), but they are not absent from other cultic sites, including the Parian Delion (Rubensohn , -, pl. de).

The animal bones
Animal bones together with clay sherds formed the main components of the Mandra deposition. The bones have been precisely studied by a specialist -Dr Simon Davisgiven their potential in illuminating the character of the assemblage and of the practices related to them.  According to the study of the remains, sheep and goats were the most frequently attested animals in this context, forming  per cent of the assemblage, with sheep occurring slightly more than goats.
 Two iron and four bronze objects remain unidentified, while the layer contained four iron chunks and two pieces of iron.
 A representative sample of the knives has been included in Petrakis' doctoral thesis (Petrakis , -).  I am more than grateful to Dr Nikos Lazaridis for the identification and dating of these scarabs and to Prof. Kousoulis for his further guidance on the issue.


Steele ; Ekroth and Wallensten . These preliminary conclusions belong to Dr Simon Davis (IGESPAR, Laboratory of Archaeological Science, Lisbon), who has undertaken the study of the animal bones.


Pigs represent  per cent.  A few other animals are represented, including the single tooth of a cow, five bones of hares and two dog bones. The layer revealed abundant seashells, several fish bones and crab claws.
Less than  per cent of the analysed bones from this assemblage were calcinated, and these all belonged to caprines. According to Davis, despite the small size of the sample, there might have been a preference for femora. Τhe rest of the zooarchaeological material bears no traces of burning, while a small percentage preserved cut marks.
The study revealed that almost two-thirds of the sheep and goats were four to eight months old. Moreover, almost all pigs were slaughtered at a very young age when probably still suckling. The narrow range of the species attested at the site, limited to sheep, goats and pigs, as well as their slaughter at a particularly young age, point to a specific meat-eating policy.

RITUAL OR WHAT?
'Ritual', the 'paramount archaeological safe-word' (Haysom , ), commonly follows anything that cannot be understood or functionalised in the archaeological record for three decades now (Pakkanen , -). The relevant bibliography is vast, and the notions of ritual, religion and cult still remain contested. Archaeologists largely draw from the related extensive anthropological studies,  with those of Catherine Bell (; ) being among the most used. Bell underlined the difficulty of separating clear-cut ritual and profane activities by shedding light upon a series of interemediate zones between the two ends consisting of ritualised events, that, is ritual-like activities. According to her view, ritual does not solely apply to religious institutionalised activity (Bell , ); on the contrary, it is mostly related to the process of ritualisation and the degree to which activities are ritualised. That is, 'the degree to which the participants suggest that the authoritative values and forces shaping the occasion lie beyond the immediate control or inventiveness of those involved' (Bell , ). A highly influential framework has been proposed by Clifford Geertz (), who attempted to provide an anthropological definition of religion as a system of symbols, a symbolic communication between people, with rituals defined as one of the system's elements. In Geertz's framework, rituals are interwoven in humans' everyday existence, bridging it with religious reality.  Archaeologists have largely adopted Colin Renfrew's anthropologically inspired framework for cult identification on the basis of particular archaeological correlates, often with revisions.  Over the last two decades, archaeologists studying Early Iron Age materials have used methodologies built on particular theoretical frameworks in an attempt to characterise a context or a site as cultic/religious and to identify ritual activities ( The material evidence from the extended deposition discovered at Mandra of Despotiko shares a lot in commona set of clay vases for feasting, terracotta figurines, metal implements, animal boneswith the assemblages from other early sanctuaries, ticking most of the boxes of a religious rather than domestic context. Since the Early Iron Age buildings were discovered below the level of the Archaic temenos in close association with the 'temple' and the cultic Building Δ,


Renfrew , -. The distinction between religious and non-religious ritual was central for Renfrew, who did not, however, distinguish between cult and ritual. His framework has been criticised by a number of scholars: see Pakkanen ,  n. ; Haysom ,  n. . An alternative definition and identification of ritual has been provided by Brück () and Bradley (), who moved away from the correlates for ritual, emphasising on the fact that the modern terms adopted for ritual, cult and religion cannot apply to the past conceptions and practices they tend to describe.

THE TERRACOTTA ANIMAL FIGURINES FROM DESPOTIKO
 the most convenient assumption would be to recognise the earliest structures as cultic and the associated deposition as containing the remains of the earliest ritual activities at the site traced back to the ninth century BC. Is this, however, the whole story?
The large assemblage of Early Iron Age pottery, containing a high percentage of small open shapes in combination with fewer mixing shapes and numerous cooking pots, must have served for the preparation and consumption of food and drink at the site during that time (Alexandridou , -; in press). Such activities are further supported by the zooarcheological remains. Ζooarchaeological depositions have been examined from a number of Early Iron Age cult sites.  In almost all cases, they were mixed, including both burned and unburned bones (Ekroth , -): the burnt ones are the result of sacrificial activity, related directly to the altar, while the unburnt bones represent the remains of ritual meals consumed at the site.
Feasting has been recognised as the core of ritual practices in the early Greek sanctuaries.  Nevertheless, this communal consumption of food and/or drink did not only mark ritual events, but it formed part of everyday social life, occurring at special or unusual occasions even within the household (see esp. Dietler and Hayden , ; Hayden , , ). The archaeological traces of commensality can be strong in early Greek cult sites, but they can be firmly identified in domestic and funerary contexts too. Moreover, we should not infer that all animal bones in cult sites belong to sacrificial victims nor should all meat consumption be related to ritual activities of sacrifice. Killing and consuming an animal in a sanctuary or a house may not have differed (Ekroth , esp. ; , -). Comparably, the clay equipment necessary for cooking and consuming meat and beverages did not differ in cultic and domestic environments (Morgan , -; Kenzelmann Pfyffer and Verdan , -; Eder , ).
At Mandra, the actual thysia is rather thinly evidenced without secure identification of the specific body part chosen to be burnt on the altar, even though the burnt bones exclusively of goats and sheep are compatible with the species representing the god's portion burnt on the altar. The largest percentage of the faunal remains from the site represents consumption debris, with the high fragmentation of the bones and the visible cut marks reflecting that they had been chopped up into portions and broken for the marrow to be accessed. Not sacrificed at the altar, most of the sheep, goats and pigs must have been slaughtered and subsequently boiled for alimentary purposes. This cooking method was the most commonly applied in the Greek sanctuaries (Ekroth , -; ), and the abundant fragments of cooking pots included in this deposition point to their use for boiling meat. In this frame, the iron knives could well have been used for butchering.  At the same time, the young age of the slaughtered animals, contradicting most animal husbandry strategies, the presence of clay shapes, like dishes or lekythoi of the 'Argive Monochrome ware',  in combination with the fine quality of the decorated vases could be considered as compatible with the operation of animal sacrifice and ritual meals.
The question is whether this archaeological evidence from Mandra is enough for reconstructing a site as exclusively cultic. As analysed elsewhere, the Early Iron Age evidence from the site allows the reconstruction of an extended establishment (Alexandridou ; Alexandridou and Daifa in press). Buildings Ξ and O are its best-preserved components, whose function, however, cannot be easily deciphered (Alexandridou , -). The relation of Building Ξ with the Archaic


Building A, recognised as the 'temple' of the temenos, could indicate a cultic rather than a domestic edifice. The small eschara and the two pithoi, reminiscent of the large pithos, possibly a container for sacrificial debris found in association with an apsidal building at Asine (Wells , ; Pilz , -), could form positive arguments in the same direction. Building Ξ might then have been the first hestiatorion at the site, if adapted to Wecowski's (, ) model and seen as an intermediate step between rulers' dwellings and temples, a hypothesis which seems to be supported by the Late Geometric fragments of open small and large vessels revealed from its interior, as well as the pithoi, the related bones and the knife. In constrast to Building Ξ, no special features or spatial factors confront the interpretation of Building O as a residence, which could have housed feasting activities too (Alexandridou , -).
On these grounds and solely on the basis of the available data, life at Early Iron Age Mandra should be considered as flowing into different directions: the cultic elements, including very possibly animal sacrifice, are present, but at the same time the pottery and the inferred feasting could also be placed in a domestic framework with the quality of the equipment reflecting the high status of the participants, as well as the individuality of the occasions. Such high living standards are also inferred by the fact that animals were not kept for secondary products.
The coexistence of domestic and religious structures in other sites, like Delos, where a domestic nucleus together with its burial grounds surrounded the sanctuary of Apollo in the Early Iron Age (Poulsen and Dugas , ; Gallet de Santerre , -, -; Étienne , -; Brisart , -, ), supports our reluctance to push all early evidence from the Mandra of Despotiko into an exclusive religious frame (Alexandridou , -).

BACK TO THE DESPOTIKO FIGURINES
The terracotta animal figurines from Mandra lack wear or signs of rework/re-use. Acquired on the island of Paros, where they had been produced, they were brought to the site after a sea journey of a few hours for the fulfilment of some purpose, or for playing a role which cannot be determined as their original context remains unknown.
All figurines were found in the 'heart' of the deposition, and their condition indicates that the trench where they were discovered was not far from their original location. This may suggest that they had been initially all placed together, whether placed in relation to some structure, no longer extant, or in the open air. We will never know whether they were placed alone or alongside other objects. If its condition was not an accident of the procedure of firing, the heavily burnt clay bovid may have been disposed in fire, possibly after some action, which might or might not be related to a sacrifice.
Another indication of their possible use is provided by the hole above the tail of one of the birds (Fig. ). The function of such holes is not always easy to determine (Muhly , ). It does not seem to have been dictated by the firing process, as in the case of the examples from Tsikalario (Charalambidou , -). Unlike the Naxian small birds from the South Cemetery, which have their extended wings pierced (Κοurou , -), the position of the hole on the Despotiko bird suggests that it could have been mounted on a wooden stick or suspended with a string. In the latter case, another hole would have been necessary on the top of the head for counterbalance. After being used or simply left at the site, the figurines may have acted as tangible memories of the related activities for some time before being removed and ending up as construction debris: during this last stage of their life circle, they lost any symbolic and spiritual significance they carried.
If the Despotiko clay figurines are treated 'traditionally' as indicators of the gender and character of the venerated deities in Greek sanctuaries (see e.g. Televantou , ; Schilardi ; Kourou ), then they could be associated with the cult of a female deity with the characteristics of a potnia theron (Kourayos et al. , -). Nevertheless, I find it more attractive to adopt an ontological and sensorial approach, enabling us to view these clay animals as part of the islet's landscape and physical environmentholistically examined together with THE TERRACOTTA ANIMAL FIGURINES FROM DESPOTIKO  the topography and the built structures (Knapp )and more importantly in relation with the humans, examining how the latter conceived of, experienced and interacted with both the objects and the living beings which they embodied.
Bovids are the earliest three-dimensional clay representations of animals, appearing in the Early Iron Age since the eleventh century,  being widespread both in mainland Greece and the islands (Guggisberg , -; Averett , -). They could not have acted as clay substitutes of sacrificial animals, since the burnt zooarchaeological material from sanctuaries does not include cattle bones,  an absence also noted in the case of Despotiko. Alternatively, they demonstrate their importance in agricultural and pastoral life at that time,  while acting as symbols of power and fertility too (Guggisberg , -). The presence of the clay bovids at the site is not surprising, since Mandra is the largest and most fertile peninsula of Despotiko. Bovids no longer exist on Despotiko or neighbouring Antiparos, but were used for ploughing on both islands at least until the end of the s.  The number of clay birds either in the form of figurines or as vessels at the site is worth noting due to their rarity in the Early Iron Age Aegean, where they are mostly attached to vessels or other objects rather than being free standing.  With the exception of three Late Protogeometric ducks from the cemeteries of Lefkandi (Popham, Sackett and Guggisberg , -). Another possible bird figurine has been recognised among the material from the North Sacrificial Area at Eretria (Huber , -, nos -, pl. ). Small birds with a flat body and spread wings are known particularly from Crete and Rhodes.  As already mentioned, except for a large concentration from the Naxian South Necropolis, the other examples from the Cyclades are very few.
Since they are attested in various contexts, bird figurines have been mostly associated with a symbolic meaning, either linked to the divine or the deceased (see e.g. Bevan , -; Papadopoulos , ; Guggisberg , -; Xagorari , -; Kourou , -; Charalambidou , ). The numerous Late Helladic IIIC water birds at the site of Kynos in central Greece have been associated with fecundity rituals operating in settlements or shrines,  a function also assumed for those coming from the so-called 'ritual zone' at Lefkandi. The symbolism of the examples from the Naxian graves is, on the other hand, mentioned as funerary (Charalambidou , ). Whether the Despotiko birds should also be seen as symbols of fertility, or of a female deity, particularly Artemis or her potnia predecessor (Christou , , -; Bevan , -, -, ), cannot be concluded.
On the other hand, the two larger bird figurines with their careful painted decoration, denoting details of their plumage, head and beak, must have been manufactured in imitation of specific bird species. The seasonal wetlands around the island of Antiparos, the calm waters between Despotiko


The analyses of the faunal remains from cult sites showed that bulls were only very rarely sacrificed though consumed at sanctuaries. See, e.g., Ekroth .  E.g. Heilmeyer , , -; , . The abundance of bovids in contrast to goats and sheep has been assigned to a very special symbolism of the former as representations of theriomorphic divinities, pointing to male deities; see Averett , , .
 Two bulls are shown in a ploughing scene, decorating the belly of a pithos amphora from the necropolis of Paroikia on Paros, dating to the third quarter of the seventh century BC: Paros, Archaeological Museum B; Zapheiropoulou , figs -. Ploughing bovids appear in a scene of the Greek film 'Madalena' produced by Finos Film and filmed on Antiparos in . 


and Antiparos, as well as the uninhabited islets nearby form an ideal, protective environment for various aquatic seabirds or other bird species. The better-preserved clay bird that does not find exact parallels among its contemporary clay figurines could represent either a shag (Phalacrocorax aristotelis), which tends to nest at the rocky islets, or a woodcock (Scolopax rusticola), living in the island's inland areas.
The snake is also part of the islet's ecosystem. A number of species live today on both Antiparos and Despotiko, including javelin sand boas (Eryx jaculus) and vipers (Vipera ammodytes), whose head can be compared with that of the figurine. Snake figurines might be rare in the Early Iron Age Aegean, but the clay snakehead, as well as the plastic snakes attached on the handles of a few fine and coarse vessels, suggest a creature important for Mandra. Clay snakes are reported from the sanctuary of Athena at Koukounaries on Paros (Schilardi , ; , ; , ). Outside the Cyclades, handmade clay snakes are known from the sanctuary of Artemis Limnatis at Kombothekra in western Peloponnese (Sinn , , nos , -, pl. ), where possibly part of a local cult tradition with a chthonic dimension. They are often treated as guardians of a site or of the contents of vessels, on whose handles they are attached.  As representations of different elements of Despotiko's fauna, they could be further used as media for approaching the issue of the interrelation of humans and animals at that time. Since the s, the dichotomy between nature and culture and the distinctions between humans and animals (Ingold ; ) has been rejected in the archaeological discourse of early prehistoric Britain and north-west Europe, though this discussion has not so much affected Mediterranean archaeology (Jones , , with relevant references). Despite the recent research attempts towards a different perception of animals in other cultures (see e.g. Hill ; Argent ), non-humans are still rarely treated as agents of activity with an active role in cultural processes in Greek archaeology (Jones ), especially during the 'historical periods', which still remain deeply rooted in the Western mentality of regarding animals as 'objects' rather than 'subjects'.
Taking into consideration ethnographic data combined with the principles of the archaeology of ontology, I suggest that notwithstanding any symbolic connotations related to land and fertility, the manufacture and use of these clay figurines on Despotiko reflect the assignment to these animals of a concept of personhood by the locals, and that they were treated as conscious objects, deeply embedded in their everyday life and the perception of their world. Moreover, the engagement of people with the clay figurines must have been much more intense than that with the vases from this deposition. Materiality and tangibility are two of the main features of the figurines, since they were easy to hold, handle, carry, assemble, disassemble and even break.  From a sensorial point of view, they could not only have acted as a prelude to certain activities (Peatfield and Morris , , ), but as objects activating senses, thoughts, experiences, feelings and memories of their carriers as extensions of their body (Hamilakis ).
The condition and the oddities of these types and the uniqueness of this assemblage indicate some role of the Mandra clay figurines in the frame of a ritual behaviour. Such a behaviour is better conceived if inserted into Geertz's () conception of religiona unitary phenomenon in people's daily experience and lifeand Bell's (, , ) wide spectrum of ritualisation, as not limited to clear-cut examples of ritual, solely related to formally institutionalised religions, but to gradable ritualised activities in everyday life. After all, terracotta figurines are not absent from settlements in the Early Iron Age Cyclades (Kastro Siphnos, Koukounaries Paros),  while ritual practices in a domestic environment are also evidenced (Pilz ).
This fluidity of the early contexts at Mandra and the implausibility of distinguishing clearly between sacred/religious and profane/domestic, notions and loci largely intersecting in the Early Iron Age, places the figures studied here outside the confines of such dichotomies. If such  Guggisberg , -; Rodríguez Pérez . For their relation to water when attached on hydrias rather than chthonic symbols, see Alexandridou .    fragile, contrasting notions can be indeed cross-fertilised and the too readily accepted equatationsfigurines = cult, birds/snakes = goddessesare questioned, the focus of our research can be more emphatically placed on the early occupants of Mandra themselves, elucidating aspects of their lives, their multifaceted activities and ideally their concerns, thoughts and feelings.