Terrae Antiqvae Imágenes

Artistic Expression and Material Culture in Celtic Gallaecia

Alfredo González-Ruibal

Abstract

The Iron Age art of northwestern Iberia (the ancient Gallaecia) is scarcely known by scholars outside Spain and Portugal, due to the local character of most publications dealing with this topic. This has deprived archaeologists of the knowledge of an original culture, whose artistic expressions owe as much to the Mediterranean world as to the Atlantic one. An overview of the main characteristics of this distinctive material culture tradition, within the social context of its production and use, is offered along with a catalogue of the most outstanding pieces of Gallaecian art (sculpture, jewellery, bronzework, pottery).

Keywords

Sculpture, jewellery, bronzework, oppida, hillforts, Late Iron Age, northwest Iberia.

Two caveats have to be made on the topics addressed in this article. One refers to the term "Celtic" and the second to the concept of "art". As we will see in the next paragraphs, there are very few elements that characterise "early Celtic" art par excellence, i.e. the La Tène style of the Late Iron Age. Although this style cannot simply be equated with "Celts", it can be found over most of temperate Europe, from the north of Italy to Ireland and from Brittany to eastern Europe and it was obviously used as a form of elite expression among aristocrats in the Late Iron Age in Europe (Megaw and Megaw 1989; Wells 1998). However, no clear traces of Celtic art have been discovered in northwestern Iberia to date. Even if some elements do show links with central Europe, such as the Montefortino helmets or the têtes coupées, how they arrived in this area is not easy to guess. The Montefortino helmets, for example, or at least the concept of this helmet style, may have come from southern Iberia since the artifacts themselves were apparently produced in the northwest. Personal ornament was clearly inspired by continental or Atlantic Europe, especially in the case of torcs, but technology and style (filigree, granulation), as well as decoration are, in many ways, akin to the Mediterranean culture. Many of these mixed or hybrid features are due to the doubly marginal character of the region, with regard to the Mediterranean and to other Atlantic or Continental regions (González-Ruibal 2004). If we use the label "Celtic art" to mean the "artistic" expressions of certain "Celtic" groups - those supposedly inhabiting the northwestern Iberian Peninsula in the Late Iron Age - then its use might be admissible, although I remain suspicious about the historical reality of such a construct. Finally, I will not deal with elements from the Later Bronze Age or the Early Iron Age. Only the art of the Middle and, especially, the Late Iron Age - roughly equivalent to the La Tène period (450-50 BC) in temperate Europe - will be taken into account. Otherwise, an anachronistic and confusing image of the local cultures and their evolution would be presented, a shortcoming of many histories of "Celtic art".

Secondly, the term "art" is not something that can be extrapolated from contemporary to pre-modern cultures without exploring the meaning and the role of aesthetics in those cultures. Archaeologists and art historians tend to forget the deep social roots behind what they label "art". Lévi-Strauss (1995), for example, has stressed the "semi-religious" character of many non-Western artistic expressions. I agree with Shanks (1996: 148) when he says that there can be no purely decorative surfaces, devoid of meaning, since a design always implies, at least, the conditions of its production. Besides, the clear division between art and other forms of material culture in modern Western societies is rarely shared by traditional groups. Should we include fibulae, torcs and pottery or not? Are we not adopting a male-centered view when refusing to consider pottery an art, or giving it a secondary place in aesthetics? Is architectural decoration always perceived as something imbued with aesthetic meanings or are these secondary as opposed to, say, magical principles? On the other hand, "art" cannot be reduced to its sociological background: a sense and even a theory of aesthetics is present in every society - see, for example, Glassie (1999). Although these questions are beyond the scope of this paper, it is necessary to keep them in mind when looking at the works that will be discussed below.

If we adopt a classic approach to the art of Iron Age Gallaecia, we find the following expressions:

1. Stone sculpture. Two traditions can be proposed: the "art of the oppida", in the southwest of the region (Conventus Bracarum), and representations of human heads, which are mainly found in Galicia (the largest part of the ancient and largest Roman administrative region known as Gallaecia).
   
   
2. Jewellery. This, and the following manifestations, would be labelled as "minor arts" by traditional art historians. All material expressions in the northwestern Iberian Peninsula, however, were tightly linked: the same motifs and a similar grammar can be found in stone, gold and pottery, thus allowing a "pervasive style" (de Boer 1991) to be defined. Three main types of artifacts can be identified here: torcs, belts, and earrings. A marked regional diversity, with important socio-political implications, can be seen.
   
   
3. Decorated bronzes. Only a handful of pieces have been recovered. Four types of artifacts will be taken into account, three of them related to ritual and one associated with war: cauldrons; sacrificial axes; ceremonial cart models and helmets. Unlike other "Celtic" areas in Europe, scabbards, hilts, swords and daggers lack almost any decoration in Gallaecia. Except for the ceremonial cart model, the rest of the decorated bronzes probably belong to the Late Iron Age.
   
   
4. Pottery. In the Middle Iron Age (400-100 BC) an outstanding decorated pottery tradition appeared in southwestern Gallaecia. Some baroquely decorated containers were still produced in the Late Iron Age. Local pottery has little in common with the La Tène style or other continental traditions, either in shape, technology or decoration. Parallels with other Atlantic regions are not seen until well into the Late Iron Age and only in the northernmost area.
   
   

A brief description of each artifact category will be given, followed by a catalogue of some representative pieces. The references for each work do not represent a comprehensive bibliography, but only the most important or recent titles. Older sources are mentioned only if they contain relevant context-related data. Calo Louridos' Ph.D. thesis, published as a book (Calo 1994) represents the most comprehensive and updated account of the Iron Age sculpture in northwestern Iberia, and includes a complete bibliography up to 1991 as well as a good summary of each site. A recent number of the journal Madrider Mitteilungen (Vol. 44, 2003) is devoted to warrior statues in Gallaecia. For jewellery, two catalogues were consulted: Balseiro (1994) and Álvarez Núñez (1996). Bronzes and pottery lack good monographs. Silva's (1986) Ph.D. thesis contains a significant amount of information, drawings and photographs of sculptures, jewellery, bronzes and pottery in northwestern Portugal. The place where each piece is now located is indicated following the relevant sources in the catalogue below.

The area referred to in this article is the territory called Gallaecia by the Romans, basically northern Portugal and Galicia and adjacent territories of Spain (Fig. 1A). Their inhabitants were called Callaeci. The region was divided into a northern territory (Conventus Lucensis) and a southern one (Conventus Bracarum) (Fig. 1B). The division was made on ethnic grounds: Roman boundaries seemingly reflect Iron Age traditions. The neighbouring area, called Asturia by the Romans, is less well known and its material culture is not as rich as that of Gallaecia. It stretches through the modern provinces of León and Oviedo (Asturias). The southern part of Asturia (Asturia Cismontana or Augustana) was strongly influenced by the eastern Celtiberian traditions and it will not be taken into account here.

Figure 1A. The northwestern Iberian Peninsula: the land of the Callaeci.	Figure 1B. The northwestern Iberian Peninsula. Roman territories (conventus) and modern capitals.

The following chronology will be used:

Late Bronze Age 1200-800 BC

Early Iron Age 800-400 BC

Middle Iron Age 400-100 BC

Late Iron Age 100 BC - 50 AD

Roman Iron Age 30 BC - 50 AD

Figure 2. The area of the southern Gallaecian art style, which corresponds almost exactly with that of the oppida.

Sculpture

The Art of the Oppida

The oppida built in southwestern Gallaecia from the second half of the second century BC on reflect a deep transformation in the structure of the society and material culture of the area (Fig. 2). The elites increased their power: centralization, inequality and hierarchization characterize the Late Iron Age in northwestern Iberia, as seems to be occurring at that time in most parts of temperate Europe. Also, as in the rest of Europe (Kruta 1993: 438-443), the creation of these centralized polities lead to a particular style. Complex chiefdoms, as Earle (2002: 162) reminds us, are usually associated with "great styles". Significantly, one of the most astonishing novelties in our region is the emergence of a tradition of rich stone statuary that has no equal in other regions of "Celtic" Europe, although Mediterranean influence may be responsible for the development of stone sculpture in Gallaecia. Southern Iberia enjoyed a superb sculptural tradition from the sixth century BC on, following Greek and Oriental models (Chapa Brunet 1990), and the coastal areas of Gallaecia were in touch with Semitic merchants from Andalusia (González-Ruibal 2004). Nonetheless, most of the motifs and elements are native or have wider European connections. Four main groups of sculptures can be established:

1. Warriors.
2. Anthropomorphic statues, probably representing gods.
   1. Female goddesses.
   2. Seated gods.
3. Architectural decoration.
4. Pedras formosas (lit. "beautiful stones"), i.e. decorated stones from ritual saunas.
5. Sacred pillars. Omphaloi?

Chronology

Before describing the characteristics of these sculptures, a note on chronology is required. The origins of Gallaecian sculpture are much debated. Most of the archaeologists working in the area postulate a late chronology, starting from the time of Augustus and ending in the late first century AD (see Calo Lourido 1993-94, 1994), a theory, however, that is flawed for a number of reasons.

Some decorated stones have been recovered in hillforts abandoned before the time of Augustus. The best examples are Sabroso (Braga, Portugal) and A Forca (Pontevedra, Galicia). Despite Calo Lourido's (1994: 250-256; 421-422) objections to both sites, the evidence is overwhelming: no LIA pottery has been recovered in A Forca, but MIA indigenous pottery was found along with Punic and Greek sherds - the scant Roman sherds appeared in disturbed contexts and most are likely Republican in date (Carballo Arceo 1987, 1996). In Sabroso, although later in date, only a fragment of a Roman amphora has been reported (Soeiro et al. 1982: 345), while the typical materials of "romanized" hillforts are absent - Samian ware, glass, Roman coins, tegulae - at least from stratified contexts. A Forca represents the origin of sculptural decoration during the second century BC and Sabroso its height during the first century BC, before Augustan times.

Most of the statues and friezes were not recovered during archaeological excavations and, even when they were, they were rarely discovered in their original context of use. The fact that the oppida and hillforts where sculptures have been discovered show strong traces of Roman occupation (basically from the time of Augustus to the mid-first century AD), does not imply that the scultptures were created and displayed in this period. As a matter of fact, they often appear re-used in streets and houses that were then being built or rebuilt (e.g. Mergelina 1944-45: láms. XXVIII-XXX; Viana 1963: 175; Silva 1986: est. XXVII, nº I-IA, 2; est. CXXXV, 3). That implies that they had been produced earlier. In one of the very few cases in which a carved stone was discovered in situ, inside a domestic compound, the stone was hidden by a wall some time in the first half of the first century AD (Silva 1986: 49; est. XXVII, nº I-IA).

It is difficult to understand why a clearly indigenous manifestation began after the Roman conquest only to come to an end 75 to 100 years afterwards. The huge number of sculptures known (around 600 pieces) makes this hard to believe. No clear Roman influences can be distinguished in most of these sculptural works. A local Roman art style developed from the beginning of the first century AD on that is quite similar to other local artistic expressions in the Empire, such as those from Britain or Gaul.

In the specific case of warrior statues, the weapons and dress depicted are clearly pre-Roman and the Latin inscriptions that appear on some statues are obviously of later date, one of them from the 17th century AD (Koch 2003). Some authors have argued that warrior representations are the indigenous interpretatio of the Roman thoracatae statues in the fora (Almeida 1983). This can hardly be the case, if only for chronological reasons: the first monumentalization of the Roman towns in northwestern Iberia did not take place before the Flavian era, while the hillforts start to fade away by the mid-first century AD. On the other hand, no formal or structural similarities can be found between Gallaecian warriors and Roman statues.

Thus, I propose a chronology of the second and first century BC, parallel to the emergence and development of oppida, for the art of southern Gallaecia (González-Ruibal 2003a: 326-328). The art came to an end in the aftermath of the Roman conquest of the region, as part of a wider re-negotiation of identities (González-Ruibal 2003b).

Warriors

Warrior statues are among the most outstanding artistic achievements of pre-Roman Gallaecia. They represent outsized males with a short sword or dagger, a small round shield (caetra), torc, bracelets (viriae), decorated belt and decorated dress. In some cases, they have a beard (Lezenho) and in one example, the head is covered with a helmet (Sanfins). The weapons, specially the famous caetra, have been also described in Greek and Roman texts (Strabo 3, 3, 5). The decorated belts are especially interesting: the existence of magic belts has been noted amongst Phoenicians and Punics (Blázquez 1991: 90-97), and they could have reached Gallaecia through the Semitic merchants that regularly visited the northwest coasts in search of tin and gold. The best warrior statues come from Lezenho, an unexplored hillfort near the Támega River, in the northwest of Portugal. All the examples were made of granite, as were virtually all other sculptural works in the northwest of the Iberian Peninsula. Unfortunately, as is the case with most of the statuary, only a couple of pieces have been found in an archaeological context. These are the remains from the Sanfins' acropole and the feet - wedged in an outcrop - found at the same oppidum (Silva 1999:16). The problem with the former is that they were found in a secondary position and purposefully demolished in what seems to have been a sacred area, used in the first half of the first century AD. The feet have no archaeological materials associated with them; however, it is interesting to note their location. The statue was situated in a prominent place among huge rocks beside the main gate of the second enclosure of the oppidum. The figures probably represent idealized warrior aristocrats (heroes) that symbolically protected the entrance to the settlements, while at the same time making explicit the power of the ruling elite. Parallels have been proposed with the late Hallstatt statues from Germany, such as those from the Glauberg and Hirschlanden (Höck 1999). However, the Gallaecian warrior statues were probably produced in the late second and first centuries BC and always appear linked to oppida or hillforts, while the Hallstatt examples are from the sixth to fifth centuries BC and are found in funerary contexts. In addition, there is too wide a spatial gap between southern Germany and northwestern Iberia. Warrior statues have appeared in several locations in northwestern Portugal and in the province of Ourense (Galicia) in the oppida of Rubiás and Armeá.

Figure 3. Distribution of different kinds of warrior statues. After González-Ruibal (2003a).

There are also smaller representations of warriors, most of them from southern Galicia, whose date may be earlier, although we lack any contextual data. The most famous image is the warrior from Capeludos, from northeastern Portugal, who wears a cap extraordinarily similar to that of the Hirschlanden warrior. Finally, some stone heads, probably depicting warriors, have been discovered in Galicia. There are only three good examples: Castro do Río, Anllo (both in Ourense) and Ralle (Lugo). They are oversized heads wearing torcs and probably helmets, although the less realistic detail and the rougher treatment of the surfaces prevent an easy identification (Fig. 3). The emphasis on warrior representations fits well with the situation of instability and endemic warfare that characterized the Iberian Peninsula in the second and first centuries BC. Internecine conflict and wars against Rome probably encouraged inequalities within hillfort communities, especially in southern Gallaecia, the region that was in closer contact with Lusitanians and other warlike peoples, continuously menaced by the Roman expansion. Mercenaries and warlords probably made the most, in social and political terms, of this turbulent situation.

Catalogue

Two warrior statues, Lezenho hillfort (Vila Real, Portugal) (Fig. 4).
The statues from Lezenho are the finest found in Gallaecia up to now. They show an astonishing degree of realism and detail. Both statues presented here have torcs, belts, round shields and daggers. One of them (B) has a dress decorated with interlaced S-patterns, the same pattern to be found in architectural decoration. Four statues in total are known from Lezenho hillfort. Bibliography: Silva (1986: 305-307; Est. CXX, 1-2); Calo Lourido (1994: 292-302; 2003: 10-4, Pls. 10-21); Schattner (2003). Museu Nacional de Arqueologia e Etnologia (Lisboa).

Figure 4. Two warrior statues from Lezenho. After Silva (1986).

Figure 5. Warrior statue from São Julião. After Silva (1986).

Warrior statue, Citânia de São Julião (Braga, Portugal) (Fig. 5).
This statue shows the most elaborate dress of any warrior representation known from Gallaecia. The skirt has the same diamond-shaped pattern found in other statues (Lezenho, Monte Mozinho). The shirt, however, has abstract vegetal motifs of eastern Mediterranean origin that somewhat recall La Tène-style pieces. The motif on the right shoulder is a stylized palm-tree, very common in Iron Age orientalizing and post-orientalizing art (see Blanco Freijeiro 1957: 174-196). Some examples in architectural decoration can be identified in northwestern Iberia from Lansbrica and Coeliobriga, in Ourense, Galicia, (Calo Lourido 1994: 443, 1-2) and Briteiros, northern Portugal (Calo Lourido 1994: 180), among other sites. The following inscription can be read on the warrior's shield: Malceino / Dovilonis / F, i.e. "Malceinus, Dovilo's son". The front part of the belt has a multi-armed swastika, perhaps with an apotropaic meaning. The fact that those elements decorating houses and ritual saunas also appear in warriors' dress, reveals an appropiation by the ruling elite of powerful cosmological symbols. This was one of the latest warrior statues to be discovered, in the early 1980s. Bibliography: Martins and Silva (1984); Silva (1986: 305-306; Est. CXX, 4); Calo Lourido (1994: 450-455; 2003: 19-20, Pls. 30-32); Koch (2003); Schattner (2003). Unidade de Arqueologia da Universidade do Minho.

Figure 6. Warrior statue from Sanfins. After Silva (1986).

Warrior statue, Sanfins (Porto, Portugal) (Fig. 6).
As has already pointed out already, the feet were found wedged among big boulders near the oppidum's entrance gate while the other fragments were dug up in a wide rectangular building along with anepigraphic altars in the central part of the settlement. The dismemberment of the statue in early Roman times might have been part of a ritual performance. If so, it would suggest a shift in the foundation and representation of power. The style looks coarse, compared with other statues. The helmet, although perhaps not a Montefortino (Quesada 2003: 98), seemingly predates the Roman conquest. Quesada attributes it to Caesarian times (Buggenum type, see also Quesada 1997: 555-558). This is the only full warrior statue wearing a helmet. Bibliography: Paço (1968); Calo Lourido (1994: 484-487, 493; 2003: 21-22, Pls. 37-39); Silva (1999: 16, 31); Schattner (2003). Museu Arqueológico da Citânia de Sanfins (Paços de Ferreira).

Figure 7. Warrior statue from Santa Comba. After Calo (2003).

Warrior statue, Santa Comba hillfort (?). Refojos de Basto (Vila Real, Portugal) (Fig. 7).
This huge statue (206 cm) wears the characteristic belt, round shield, dagger and viriae on both arms. The belt has a small triskel as a decoration in the back, the same used in architectural decoration. As has been said, this was probably considered a magical protective device. One of the most interesting features of this work is the presence of feet, seldom preserved in warrior statues, seemingly encased in a pair of closed (leather?) boots and shin pads (knemides). In addition, this is one of the few warrior statues with a sword, which is unsheathed over the right part of the torso. It bears a Latin inscription on the lower part of the shield: "Artifices / Calubrigens /es. et. Abianis / F(aciendum). C(uraverunt)", whose interpretation remains obscure. Apparently, the "faciendum curaverunt" refers to the inscription itself and the statue could have been reused as a landmark between both populi. Bibliography: Silva (1986: 307; est. CXXII, 1); Calo Lourido (1994: 509-515; 2003: 23-24, Pls. 40-43); Koch (2003); Schattner (2003). Private collection.

Figure 8. Warrior bust from Anllo. Author's photograph.

Warrior head, Anllo (Ourense) (Fig. 8).
The lesser quality of the work, as compared to the full warrior statues (e.g. Lezenho), is evident. The face has been roughly shaped and there is no attempt to provide a naturalistic image. Besides, the surface is badly eroded, making it difficult to ascertain whether the head was wearing a helmet or not. Slight traces of the paragnatide can be glimpsed on the left cheek. The torc, on the contrary, although not realistic, is quite obvious. The sculpture is 44 cm high. It was discovered not far from an Iron Age defended settlement, although this date is far from certain in the hillfort-crowded landscape of Galicia. Bibliography: Orero Grandal (1986: 95-98); Calo Lourido (1994: 81-83; 2003: 6, Pl. 1); Schattner (2003). Museo Arqueolóxico Provincial de Ourense.

Anthropomorphic statues

Female goddesses

The number of female representations in the pre-Roman art of Gallaecia is scant and their chronology is very dubious. We will take into account only two statues here: one from Briteiros and another one from Sendim, both found in hillforts in northwestern Portugal. These are less than life-sized representations of women, with their sexual attributes conspicuously marked. The sculptural work appears less skillful when compared to the warrior statues. The piece from Sendim was discovered in a hillfort inhabited throughout the Middle and Late Iron Age, until the beginning of the first century AD. The piece from Briteiros comes from the oppidum's acropolis, in an area that was probably devoted to ritual activities, as shown by other sculptures, inscriptions and human remains found there (González-Ruibal 2003a: 469-470). Some divinities known through the Roman epigraphic evidence in northwestern Iberia are probably female (such as Navia). Nonetheless, due to the scant information given in inscriptions, it is difficult to link statues to particular divinities. Female gods wearing torcs are known in temperate Europe, such as one of the figures depicted on the Gundestrup cauldron. Apart from the two images from southern Gallaecia, three other female representations are known in the northwest (Calo Lourido 1994: 196-197; 302-304). Two of them come from the Logrosa hillfort and the third from Carabeles (A Coruña, Galicia). They probably depict women, although only one of them (Carabeles) has any sexual attribute (breasts). All female statues are shown clasping both hands together in what has been called an attitude of respect, submissiveness or devotion, sharply opposed to the hieratic and powerful bodies of male warriors. The presence of sexual attributes suggests fertility and reproduction. The mother goddess cult (Matres) is known in Gallaecia. The most explicit inscription, however, was found outside the region: it is the altar devoted to the Matribus Gallaecis ("Gallaecian Mothers") in the Roman town of Clunia in the Spanish Meseta (CIL II 2776).

Catalogue

Female statue, Oppidum of Briteiros (Braga, N. Portugal) (Fig. 9).
The most interesting element in this work is the presence of a torc, an item of personal ornament usually associated with men in Celtic Europe, at least in the second half of the first millennium BC. Women wearing torcs are commonly associated with deities. The work is rather crude, with rough surfaces, and disproportionate. The only sexual attributes visible are the breasts. The position of the arms is very characteristic. This is the only seated female statue in northwestern Iberia. It is 48 cm high. Bibliography: Cardozo (1968); Calo Lourido (1994: 170-171, 188). Museu da Sociedade Martins Sarmento (Guimarães).

Figure 9. Female statue from the oppidum of Briteiros. After González-Ruibal (2003a).		Figure 10. Female statue from Sendim. Author's photograph.

Female statue, Sendim hillfort (Porto, N. Portugal) (Fig. 10).
In this case, the breasts are not so obviously marked as in the Briteiros piece. Nevertheless, the vagina is clearly depicted. The general gesture is quite similar to the previous image. Calo Lourido considers it a Roman votive offering, without any grounds in my view. It is 57 cm high. Bibliography: the first mention of the hillfort is found in Pinto (1933), without reference to the statue; Calo Lourido (1994: 600-602). Museu da Sociedade Martins Sarmento (Guimarães).

Seated gods

Only four sculptures representing seated gods have been discovered in Gallaecia to date. However, they are extremely interesting. The motif is undoubtedly Mediterranean; the original model comes from the Near East in the Bronze Age. Several examples of seated divinities have been recorded in southern Spain, such as those from Galera (Jaén), Verdolay (Murcia) or Baza (Granada). Interestingly, the Mediterranean tradition portrays women, sometimes linked with Tanit, often with sex attributes (breasts) well marked, as a sign of fertility and wealth. By contrast, the statues from northwestern Iberia show males, in some cases with marked genitalia, such as the piece from Braga (Bettencourt and Carvalho 1993-94). Thus, a similar meaning might be transmitted, of wealth and power (both related to sex) but the sex itself has changed from female to male, which fits well with the southern Gallaecians' androcentric ethos, as shown by the aforementioned warrior statues. The divinities are sitting on thrones, an unambiguous symbol of power, which at least in one case (Pedrafita) is quite elaborate. The style clearly resembles that of the warrior statues: an element present in most of them, bracelets (viriae), can also be found in the seated gods. The context is unknown. Only one of them was excavated in a hillfort: Lanhoso (Teixeira 1940). Another piece comes from the vicinity of the Roman town of Braga (ancient Bracara Augusta, the capital of the Conventus Bracarum), although it could have been moved there (Bettencourt and Carvalho 1993-94). The two pieces discovered in Xinzo de Limia (Ourense, Galicia) had no definite context (Ferro Couselo 1972), but were reported as coming from the vicinity of a Roman villa and road. The link with a traditional road may be more eloquent than the settlement in this case. The statue from Pedrafita (Ourense, Galicia) was discovered not far from a hillfort (Luis 1997), but, as with the other examples, the proximity to an archaeological site means little. They could well have been located outside defended settlements, in pre-Roman open shrines or sanctuaries, as is seen at Vix in France (Chaume and Reinhard 2003). The origin of these statues must be sought in the Iron Age, based on their Mediterranean counterparts, even if they were later relocated in the proximity of Roman settlements.

Catalogue

Seated god, Xinzo de Limia (Ourense, Galicia) (Fig. 11).
This is one of a couple of sculptures found in the same place near the Roman vicus of Xinzo de Limia. Some hillforts are known in the vicinity of Roman settlements. The statues were located not far from the Lima on a wooden structure. No associated archaeological materials could be found. A Roman building and a road are nearby, but no direct relation with the statues can be recognized. One of the gods holds a vessel in his hands, the other might be holding a tray. This might be showing the god's lavishness. Mother goddesses in the Mediterranean are usually portrayed with pots as symbols of fertility and generosity. Although sex is not marked, we could perhaps consider one statue to be male and the other female, as at Vix (Chaume and Reinhard 2003). The statue is 69 cm high. Bibliography: Ferro Couselo (1972); Calo Lourido (1994: 290-91). Museo Arqueolóxico Provincial de Ourense.

Figure 11. Seated god from Xinzo de Limia. Author's photograph.		Figure 12. Seated god from Pedrafita. After Luis (1997).

Seated god, Pedrafita (Ourense, Galicia) (Fig. 12).
This was the last sculpture of this kind to be discovered and it is the most elaborate of all. The throne is decorated with a triskel, one of the most frequent motifs in architectural decoration, triangles and a double moulding in the back, similar to that found in Xinzo de Limia. In this case, the figure lacks viriae, but has wristlets. The way the torso is depicted clearly resembles those of the warrior statues. The god holds a vessel with both hands, in the same manner as one of the seated gods from Xinzo de Limia. Bibliography: Luis (1997). Museo Arqueolóxico Provincial de Ourense.

Architectural decoration

The majority of the sculptures known from the Iron Age in northwestern Iberia are elements of architectural decoration. The origin of this art is intrinsically linked to the conditions that fostered the development of oppida and a particular type of composite homestead. Compound-building and expansion served as an arena for social competition in which different households exhibited their power and wealth. The sometimes complex decorative programs can be explained as a statement of social position and economic achievement, a tradition well attested among the Toraja, in Indonesia, or the Zafimaniry, in Madagascar, amongst other "house societies" (Waterson 1995). However, the decoration of households is not only a reflection of social competition among elites or elites and commoners, but also reflects symbolic concerns: The huge populations concentrated inside oppida not only made socio-political differentiation necessary but also lead to the avoidance of dangers due to symbolic and real pollution: large agglomerations of people are usually accompanied by the appearance of protective devices in the form of amulets, paintings or inscriptions among neighbours frightened of disease or symbolic contamination. The most common decorative elements are the following:

• Stone discs with swastikas or triskels carved on them. These were embedded in house walls.
  
  
• Friezes with string-work or herring-bone-shaped decoration, usually a band that went around the house walls.
  
  
• Door decorations, usually abstract motifs (SS), mouldings and string-work covering jambs and lintels.
  
  
• Amarradoiros: Pieces of unknown use - once thought to be used to hitch up domestic animals (from amarrar: "to tie"). These are elbow-shaped and have a very simple decoration, usually some incised lines.
  
  
• Couzóns (hinges): These are round pieces with a sort of shaft whose use is also unknown. The decoration is more elaborate, usually herring-bone or string-work bands.

Catalogue

Triskels and swastikas (A), "hinges" (B), amarradoiro (C), wheels (D) and rosettes (E), Oppidum of Santa Trega (Pontevedra, Galicia) (Fig. 13).
As for all other decorated stones found by Cayetano de Mergelina in the 1920s and 1930s, no topographical or stratigraphical reference has been recorded for these pieces. The variety of swastikas has no equal in northwestern Iberia and their quality is extraordinary. The stone wheel - probably related to the swastika - is especially interesting because of the religious significance of this item in Celtic Europe (Green 1984). Wheels are also found carved in open sanctuaries (Parcero Oubiña et al. 1998). Santa Trega is the oppidum with the largest collection of Iron Age sculptures in northwestern Iberia, followed by Monte Mozinho and Briteiros (northern Portugal). It is also one of the largest oppida (20 ha), located in a dramatic environment: a prominent peak facing the Atlantic Ocean and the river mouth of Galicia's longest river, the Miño. It was an important and wealthy entrepôt from the late second century BC to the early first century AD. Bibliography: Mergelina (1944-45); Calo Lourido (1994: 530-593); Carballo (1994); Peña Santos (2001). Museo da Citania de Santa Trega (A Guarda).

Figure 13. Different types of architectural decoration from the oppidum of Santa Trega. After Mergelina (1944-45).

Swastika/triskel, Oppidum of Lansbrica, modern San Cibrán de Las (Ourense, Galicia) (Fig. 14).
This stone was reused as a slab in the pavement of a courtyard, during the early first century AD. The archaeological excavations, carried out in 1947, were never published. The maximum length of the stone is 42.5 cm. Since 2003, Lansbrica, an oppidum of nine hectares, has been the object of large scale excavations that may shed light on the pre-Roman sculpture of Gallaecia. Bibliography: Calo Lourido (1994: 438, 443). Museo Arqueolóxico Provincial de Ourense.

Swastika/triskel, Oppidum of Coeliobriga, modern Castromao (Ourense, Galicia) (Fig. 15).
This piece was discovered during archaeological excavations. It probably belonged to a small rectangular house of the late MIA or LIA. The maximum length of the stone is 46 cm. Coeliobriga, the capital of the Coelerni, was an important oppidum (about 20 ha), inhabited beginning in the Early Iron Age, and later an important Roman town. Recent excavations have recovered swastikas from pre-Roman layers dating from the second to the first centuries BC (Orero Grandal, pers. comm.). Bibliography: García Rollán (1971: 196, Fig. 47); Calo Lourido (1994: 208, 217). Museo Arqueolóxico Provincial de Ourense.

Figure 14. Swastika from the oppidum of Lansbrica. After Peña Santos (2003).		Figure 15. Swastika from the oppidum of Coeliobriga. After Peña Santos (2003).		Figure 16. Door frame from the oppidum of Âncora. After González-Ruibal (2003a).

Door frame, Cividade de Âncora (Viana do Castelo, northern Portugal) (Fig. 16).
This monumental door (its maximum height is 209 cm) undoubtedly belonged to an aristocratic house in the oppidum. It was discovered in the 1870s and its specific context is unknown. The right jamb was reconstructed in concrete. The site has also produced other interesting decorated stones, most of them reused in later structures (Viana 1963). The coastal situation of this oppidum and the quantities of amphorae found in it illustrate its important development in the LIA, parallel to that of Santa Terga. Bibliography: Calo Lourido (1994: 70-80). Archaeological excavations in the hillfort are described in Viana (1963) and Silva (1986: 48-51; Est. XXVII-XXX; CXVIII-CXXIX). Museu da Sociedade Martins Sarmento (Guimarães).

Door frame and friezes, Sabroso (Braga, northern Portugal) (Fig. 17).
Bibliography: Calo Lourido (1994: 421-432). Museu da Sociedade Martins Sarmento (Guimarães).

Figure 17. Door frame and friezes from Sabroso hillfort. After González-Ruibal (2003a).

Friezes, Oppidum of Briteiros (Braga, N. Portugal) (Fig. 18).
These are highly schematized representations of a palm-tree, similar to those found in the dress of some warrior statues (see above), recalling the reworking of Mediterranean vegetal motifs by La Tène artists of Central Europe (Kruta 2000: 770-771). However, unlike La Tène-style elements, these are more geometric and abstract. These friezes, along with other similar works, might have belonged to a monumental rectangular building, perhaps an aristocratic residence or a sacred place. They were discovered by Martins Sarmento in the late nineteenth century. Bibliography: a short introduction to the oppidum: Cardozo (1976); Calo Lourido (1994: 141-148; 180).

Figure 18. Stylized palm-tree decorations from the oppidum of Briteiros. After Calo Lourido (1994).		Figure 19. String-work-shaped friezes from the oppidum of Briteiros. After Calo Lourido (1994).

Friezes, Oppidum of Briteiros (Braga, northern Portugal) (Fig. 19).
These are characteristic examples of string-work decoration, as used on jambs and wall trimmings. Cardozo (1976); Calo Lourido (1994: 149-52, 181).

Lintels, Oppidum of Briteiros (Braga, northern Portugal) (Fig. 20).
These are two late and atypical examples of architectural decoration, dated shortly after the Roman conquest. The decoration is rather crude. The most important element, conveying the idea of power and knowledge, is the writing. In early Roman times, litteracy was probably used as a means of showing social distance from commoners and acquaintance with the new rulers' culture. These lintels are also important because they show the relevance of houses, both material and socially speaking, in the social organization of the LIA communities of Gallaecia. The inscriptions say CAMAL (probably Camali) and CORONERI / CAMALI / DOMUS: "The house of Camalus, the son of Coronerus". The signs over CAMAL are probably imitating letters. The name Camalus is very widespread in Gallaecia and Lusitania (Untermann 1965, Map 26). The same name appears in other lintels and rock inscriptions from Briteiros. Bibliography: Cardozo (1976); Calo Lourido (1994: 135-138; 178).

Figure 20. Lintels with Latin inscriptions from the oppidum of Briteiros. Author's photograph.

Pedras formosas

If architectural decoration is the most popular of all the sculptural expressions in the northwest, the art reached its peak with the so-called pedras formosas (beautiful stones). The first example discovered, in the oppidum of Briteiros, in the eighteenth century, is one of the nicest works of pre-Roman Gallaecian art. The use of these stones, and of the semi-subterranean structures to which they are related, was much debated until the 1970s. There is now a virtual consensus that the buildings were baths (saunas) and that the pedras formosas were monumental façades for those baths (Fig. 21). However, two irreconcilable interpretations have arisen: one defending a Roman date and a profane use (e.g. Calo Lourido 1994; Ríos González 2000) and one arguing for a pre-Roman date and a ritual use (e.g. Almagro-Gorbea and Álvarez-Sanchís 1993). Bearing in mind that the Greek geographer Strabo (3, 3, 6), quoting sources from the second century BC, mentions the habit of taking baths among the Gallaecians; that monumental Roman baths in the area are not known prior to the mid-first century AD; that a profane use cannot explain the transport and carving of such enormous slabs nor their location in the central room instead of in the most external one, nor the difficult access to the different rooms (it is necessary to crawl through a small entrance), and, finally, that the pedras formosas have complex cosmological symbols carved on them, I consider a pre-Roman date and a ritual use the most likely option. It is also probable, as Almagro-Gorbea and Álvarez-Sanchís have pointed out, that these premises were used in rites of passage, seemingly linked to warrior groups. Meaningfully, the small entrance is almost always enhanced. Only the stone from Sardoura lacks decoration, both on the entrance and the rest of the stone, therefore underlining the idea of transition.

Figure 21. Late Iron Age ritual sauna (Briteiros 2). After Cardozo (1931).

Three types of saunas have been described (Queiroga 1992; Ríos 2000). Types 1a and 1b are those found in southern Gallaecia, in the region of the oppida. They both have three rooms and are built of large slabs and blocks. Type 1a has a very simple and coarse pedra formosa, and it is probably the oldest monumentalized sauna. Type 1b is the most widespread model of sauna and it probably dates to the first century BC. Saunas belonging to this group have well-shaped pedras formosas, sometimes with profuse decoration. Type 2 is only found in the northern area. The buildings have only two rooms instead of three, they lack pedra formosa (which were probably made of wood) and they were built of small stones (schist or slate). All decorated pedras formosas known to date come from the north of Portugal. The only Type 1B sauna from Galicia (Santa Mariña de Augas Santas, Ourense) does not have a decorated stone.

Catalogue

Figure 22. Sauna stones. A. Sardoura; B. Eiras (Arcos de Valdevez); C. Eiras (Vila Nova de Famalicão); D. Briteiros 1; E. Sanfins; F. Briteiros 2; G. Galegos; H. Tongobriga (Freixo). B, C. after Queiroga (1992); the rest after González-Ruibal (2003a).

Pedra Formosa, Type 1A, Sardoura (Viseu) (Fig. 22A).
This is the only sauna stone known south of the Douro River. It was discovered during construction work more than half a century ago and it was not properly recorded. No Roman materials were found nearby. This, along with the crude shaping of the stone, leads us to suggest an earlier date, perhaps third or second century BC, when compared with Type 1B. Bibliography: Cardozo (1949).

Pedra Formosa, Type 1A, Eiras (Arcos de Valdevez, Braga) (Fig. 22B).
This is one of the latest sauna stones to be discovered and the only one belonging to Type 1A with decoration. The decoration is atypical and simple, only the two swastikas on the top recall the decoration in Type 1B saunas, such as Briteiros 2. Bibliography: Queiroga (1992: 25).

Pedra Formosa, Type 1B, Eiras (Vila Nova de Famalicão, Braga) (Fig. 22C).
The sauna of Eiras is linked to one of the biggest oppida in northwestern Iberia, Eiras-Vermoim (ca. 65 ha.), made up of two oppida united by massive ramparts. Bibliography: Queiroga (1992: 25).

Pedra Formosa, Type 1B, Briteiros 1 (Braga) (Fig. 22D).
This was the first stone of its kind to be discovered and is the most baroquely decorated of all. The structure is extraordinarily similar to that of Eiras (Vila Nova de Famalicão). It was already known in the eighteenth century. Bibliography: an extensive and largely useless bibliography is available. See Cardozo (1976) and Calo Lourido (1994: 166-168, 185-186) for the most important references.

Pedra Formosa, Type 1B, Sanfins (Porto) (Fig. 22E).
This was the first pedra formosa discovered during archaeological excavation. The archaeological items associated were extremely poor: only a few Roman and native sherds from the period when the sauna was already abandoned. The decoration of the stone is similar to that from Freixo: a simple string-work border framing the entrance. Bibliography: Almeida (1974); Calo Lourido (1994: 487, 494).

Figure 23. Briteiros 1. Author's photograph.

Pedra Formosa, Type 1B, Briteiros 2 (Braga) (Fig. 22F)
The discovery of this stone in situ, during construction of a new access road to Briteiros, provided insight into how these slabs were actually set, although the debate about the function of the monument was to last 40 more years. Bibliography: Cardozo (1931, 1932, 1976); Calo Lourido (1994: 168-170, 187).

Pedra Formosa, Type 1B, Santa Maria de Galegos (Braga) (Fig. 22G).
Although discovered in the early 1980s and properly excavated, no report has been published. The scant data come from A.C.F. da Silva's thesis, the archaeologist who conducted the excavation. Most of the pottery is indigenous, belonging to the pre-Roman LIA. However, the drainage system and water tank were made with Roman tiles and a small fragment of a Hispanic Drag. 27 vessel was also discovered. The structure was probably refurbished and reused after the Roman conquest. The decorated stone has interesting wavy motifs, perhaps representing hot water or steam, and a small, thinly inscribed wheel, which probably has the same meaning as the swastikas present in other saunas. Bibliography: Silva (1986: 56-60, Est. XXV-XXVI, CXXXVI, 2 and CXXXVII, 5); Calo Lourido (1994: 266-270).

Pedra Formosa, Type 1B, Tongobriga, modern Freixo (Porto) (Fig. 22H).
This sauna and the one at Ulaca (Ávila, central Spain) are the only ones known that were excavated in stone. The decoration is reduced to simple string-work covering the arch. It was discovered in 1984. The Iron Age levels of the oppidum are not well preserved or published, research having focused on the later period. A Roman town of some importance was built over the old settlement and Roman baths were constructed, reusing the indigenous sauna. Bibliography: Calo Lourido (1994: 262-265); Dias (1997).

Pedra Formosa, Type 1B, Detail of Briteiros 1. (Fig. 23).
Museu da Sociedade Martins Sarmento (Guimarães, Portugal).

Omphaloi(?)

There is a small group of carved stones that cannot be clearly associated with any of the groups previously discussed. They are prismatic pillars with their four faces decorated with different motifs. I propose an interpretation based on sacred stones found among other "Celtic" peoples, such as the Irish omphaloi (O'Kelly 1989: 284-289), which have stylized vegetal motifs organized in four different parts, not unlike those from Gallaecia. I do not claim that the use and the cosmology behind Irish omphaloi can be directly translated to these pillars tout court. Rather, I would argue that they have a ritual purpose roughly comparable to other carved stones and natural rocks in the "Celtic" world, traditionally linked with political power (García Quintela and Santos Estévez 2000). In our region, there are some Roman inscriptions that may throw some light on the interpretation of these stones. There is a divinity called Crougea Toudadigoe, whose meaning seems to be "The Stone of the People" (Búa 1997: 74-75). Another divinity, Trebopala, has traditionally been interpreted as "The Stone of the Village". However, the interpretation of -pala as "stone" does not have many followers today. A translation as "flat land" is preferred.

The pillars known from southwestern Gallaecia suffer the same interpretive problems as the majority of the pre-Roman statuary in the area: the absence of a well-recorded context. Their link to Iron Age hillforts, nonetheless, is well proved.

Catalogue

Figure 24. Decorated Pillars. A. Cortinhas; B. Beiro; C. Cendufe; D. S. Miguel-o-Anjo. After González-Ruibal (2003a).

Pillar, Cortinhas (Vila Real, northern Portugal) (Fig. 24A).
The interlaced SS decorations are very similar to those from São Miguel-o-Anjo. Unlike the stones from Cendufe and Beiro, roughly the same symbol seems to be reproduced on each face of the pillar. Bibliography: Calo Lourido (1994: 233-237).

Pillar, Castro de Beiro (Ourense, Galicia) (Fig. 24B).
Only three faces are clearly visible. The triangular motifs recall those from Cendufe. The size, however, is much smaller. Bibliography: Calo Lourido (1994: 126-127). Museo Arqueolóxico Provincial de Ourense.

Pillar, Cendufe (Braga, N. Portugal) (Fig. 24C).
This pillar has Grecian frets, similar to the Iron Age stelae from Brittany (Cunliffe 2000: Figs. 4-5). A fragment of a warrior statue was also discovered here. All finds were made in the late nineteenth and early twentieth centuries. Bibliography: Calo Lourido (1994: 219-226).

Pillar, São, Miguel-o-Anjo (Viana do Castelo, northern Portugal) (Fig. 24D).
The stone was reused in the foundations of a Late Iron Age round hut. Bibliography: Pereira (1895); Calo Lourido (1994: 459-462).

Heads

Unlike the art of the oppida, heads can be found from the Douro River to northern Galicia, all over Gallaecia. Two groups can be distinguished: representations of severed human heads, similar to the Gaulish têtes coupées (Rapin 2003), and other images, probably representing gods.

It is in the northern area (Conventus Lucensis), where the largest number of this kind of sculpture, and specifically severed heads, has been found. A particular style has been recorded in this area, characterized by very simple and crude representations of human heads, often in bas-relief, with "owl faces". They closely resemble the pillars with têtes coupées from southern France (Salviat 1987; Green 2001: 100-102), even if the severed heads from southern Galicia are usually more elaborate. Two pieces from Armeá (Ourense), for example, clearly show the features of a dead person: the most naturalistic element is the swollen, half-closed eyes. A couple of severed heads carved on stone pillars, in a realistic style, have been recently discovered in Amoeiro (Trasalba, Ourense) (X. Ayán Vila, pers. comm., 2003).

Only four statues have been found during archaeological excavations - which is not bad, if we take into account the substantial problems for dating and providing a context for this kind of representation in the rest of Europe (Megaw 2003): Barán (Lugo, Galicia) (Álvarez Núñez 1993, 1993-94), A Graña (A Coruña, Galicia) (Barciela and Rey 2000: 135-136), San Cibrán de Las-Lansbrica (Ourense, Galicia) (Pérez Outeiriño 1985: 246, Pl. IV, 2) and Monte Mozinho (Porto, Portugal) (Calo Lourido 1994: 346-347, 357, 703).

The head from Lansbrica was not in its original location, due to major construction work in the area in the first century AD. However, it was found very close to a monumental gate of the oppidum's inner enclosure (where rituals or social activities probably took place). The head from A Graña was dug up in the entrance also. It was undoubtedly located over the gate. Two of the heads from Barán were located on a heap of stones in the central area of the hillfort. The chronology of the stone heads is not clear, but they were most probably in use in the first century BC. Two functions can be proposed: an apotropaic one, in the entrances to hillforts and oppida, and a ritual one, to which specific areas inside settlements were probably devoted. The ancient writers told us about the Gaulish custom of putting human heads in house doors and the practice of severing the enemy's head in war among barbarians (Strabo, 5, 29, 5; Polybius, Hist. 2, 28; Livy, Hist. 10, 26 and 23, 24; Diodorus Siculus Hist. 5, 29, 4-5). The severed head, or its representation, could simultaneously attract the power of the vanquished enemy, serve as a protection against bad spirits and as an image of the power of the community. Although it has been traditionally linked with Celtic peoples, the magical and social use of heads is very widespread, both in space and time (Rosaldo 1980). I suggest that the spread of head hunting or its representation must be linked to the increasing violence provoked both by the growing hierarchization and territorialization of the Late Iron Age and the stress brought on by the Roman wars. On the other hand, not all heads can be related to severed human heads, some of them probably represent gods, perhaps a local genius (Genius Castelli), a protector of the village. Divinities of this kind, such as Bandua and perhaps Coso, are well attested in Lusitania and Gallaecia (Olivares 2002; Prósper 2002). Whether they are related to gods or humans, heads seem to play a central role in the beliefs of Celtic peoples from all over Europe (Almagro-Gorbea and Lorrio 1990; Megaw 2003).

Another variety of heads or busts are two-faced and most probably represent Janus-like gods. Their context is unknown and their chronology dubious. Five examples, all from Galicia, are known: Amorín and Seixabre (Pontevedra), Francos and Pontedeume (A Coruña) and, perhaps, Óutara (Lugo) (Calo 1994: 67-69; 259-261; 375-376; 384-388). All have two opposite faces, except the statue from Pontedeume, which has four. Unlike severed heads, these are seldom related to hillforts. In fact, only the example from Francos may come from a hillfort and, in this case, the carving is so rough that it is difficult to ascertain if there are in fact two faces. Because of the multiple faces, they might be related with entrances, points of passage or roads. Statues with two or three heads are found in other Celtic areas (Megaw 2003: 277-278), sometimes associated with important travel routes (e.g. Hayen 1987: 134-135; Szábo 1992: 99). Interestingly, the cult of the Lares Viales had special significance in northwestern Spain.

Catalogue

Severed head, Barán (Lugo, Galicia) (Fig. 25).
This is one of a group of three sculptures, one of which was found in an archaeological context. They appeared in the acropolis of a small hillfort, in what may be identified as a ritual area, devoid of houses or other ordinary structures. However, the poor standard of the publication hampers any plausible interpretation of the site. This sculpture is a good example of the Lugo style, as represented by this hillfort and the neighbouring one of Cortes. The representation is rough and extremely simple. Bibliography: Álvarez Núñez (1993; 1993-94); (Calo Lourido 1994: 120-123). Museo Provincial de Lugo.

Figure 25. Severed head from Barán. Author's photograph.		Figure 26. Severed head from Monte Güimil. Author's photograph.

Severed head, Monte Güimil (Pontevedra, Galicia) (Fig. 26).
This is a stray find from a hillfort, whose occupation spans the period from the Early Iron Age to the beginning of the first century AD. Although it is more realistic than the examples from Lugo, it is not as elaborate as those from Armeá (see below). It is a detached sculpture similar to the head found in A Graña (A Coruña). The shape and volume of the head have been reproduced with some skill, including the ears. The marked features (sunken eyes, elongated head) slightly recall a human skull. Bibliography: Calo Lourido (1994: 662); González-Ruibal (2001: 163-164). Museo Provincial de Pontevedra.

Severed head, Armeá (Ourense, Galicia) (Fig. 27).
This piece comes from an oppidum of ca. 10 ha. The context, as for the rest of the sculptures found in this site or its vicinity, is unknown. It was carved in haut-relief and it shows considerable skill, as opposed to other têtes coupées known from Galicia. It was probably embedded in the wall of the oppidum, perhaps near a gate. Three têtes coupées are known from this settlement. Bibliography: Conde-Valvís (1950-51); Calo Lourido (1994: 98-101, 108). Museo Arqueolóxico Provincial de Ourense.

Severed head, Armeá (Ourense, Galicia) (Fig. 28).
The context of discovery is the same as for the previous piece. It bears a strong resemblance to the former, particularly the dead eyes and the technique of haut-relief on a block of stone. Their function and location was certainly the same. Bibliography: see previous piece. Museo Arqueolóxico Provincial de Ourense.

Figure 27. Severed head from Armeá. Author's photograph.		Figure 28. Severed head from Armeá. After Peña Santos (2003).

Some of the most important representations of heads from Galicia and northern Portugal (Fig. 29).
1. Chaves; 2-4, Armeá (Ourense); 5. Cortes (Lugo); 6-8. Barán (Lugo); 9. San Cibrán de Las (Lansbrica, Ourense); 10. Gaxate (Pontevedra); 11. Remesar (Pontevedra); 12. A Graña (A Coruña); 13. Moreira Nova (Pontevedra); 14. Ocastro (Pontevedra); 15. Narla (Lugo); 16. Monte Mozinho (Porto). After González-Ruibal (2003a: fig. 4.117).

Figure 29. Severed heads from northwestern Iberia. After González-Ruibal (2003a).		Figure 30. Female bust (A) and herma (B) from Seixabre. After J.J. Perles Fontao (unpublished report).

Two busts, Seixabre (Pontevedra, Galicia) (Fig. 30).
The one above represents a female. The one below is two-faced and male. They were discovered seemingly unassociated with any other archaeological remains during construction work near one of the most important natural travel routes in northwestern Iberia. Bibliography: González-Ruibal (2001: 164). Museo Provincial de Pontevedra.

Jewellery

The history of jewellery in the northwestern Iberian Peninsula in the first millennium BC is quite similar to that of the British Isles and other Atlantic regions (Northover 1994). A period of expansion in the Late Bronze Age is followed by a period of restriction (Early Iron Age), with a gradual new expansion from the Middle Iron Age that ends with an outstanding quantity of gold jewellery in the second and first centuries BC, most from hoards. The majority of the jewellery described here probably dates to the Late Iron Age. Coincidences are not restricted to the evolution and involution of jewellery, but also affect the raw material used: as opposed to the Mediterranean use of silver, gold is the prevailing metal of Gallaecian jewellery, as in the rest of "Celtic" Europe.

Torcs

Undoubtedly, torcs are the most outstanding artifacts that the indigenous communities of northwestern Iberia ever produced. Their number (around 150) is striking, especially if we consider that most of them cluster in the north of Galicia (see catalogues in Pingel [1992] and Prieto Molina [1996]). As has been said, most of them are made of gold, as in the rest of the Atlantic regions, in stark contrast with the silver jewellery typical of the Mediterranean, including the rest of the Iberian Peninsula. Silver torcs, such as the group from Bagunte (Ladra Fernandes 2001), are the product of exchanges with foreign elites or raids conducted in distant lands. Some of them weigh more than one kilogram. Although the Mediterranean influence is quite obvious (Armbruster and Perea 2000: 104; García Vuelta 2002), there are marked differences with southern traditions, one of them being the raw material (gold) employed. However, even if the Gallaecian torcs are not completely similar to those known in the British Isles or France, the very idea of the torc most probably came through the Atlantic routes. These items of personal ornament, as is also true in the rest of Europe, were worn mainly by men like the aristocratic individuals represented by the warrior statues. The links between torcs and the divine realm are demonstrated by the aforementioned statues of goddesses and seated gods. Although they were primarily used as an adornment aimed at enhancing the bodies of the ruling elite, other uses can be pointed out. On the one hand, the fact that some of them have tiny stones inside their hollow ends, producing a sound when shaken (Xanceda, Vilas Boas, etc.), proves their use as rattles. The same use, placed in the context of a religious activity, is probably attested in the Cernunnos of the Gundestrup cauldron. On the other hand, it has been demonstrated that Gallaecian torcs follow a Punic weight system, based on the shekel (Ladra Fernandes 1999b). Thus, they were probably also used as ingots, or rather, as a way of accumulating wealth in a measurable form.

With regard to the context of deposition, they seldom appear in clear domestic contexts in the northern area. By comparison, they are usually located inside the inhabited area of the oppida, in the south: fragments of torcs have been dug up in Sanfins, in the courtyard of a domestic compound (Jalhay 1950); in Troña, underneath the pavement of a first century BC round house (Ladra Fernandes 1999a); and in unknown domestic compounds in Santa Trega (Mergelina 1944-45: 37-38) and Bagunte (Ladra Fernandes 2001). Unfortunately, these are the only torcs recovered during archaeological excavations that have been properly published, although exact information on the context and stratigraphical position are only given in the Troña example. Some northern torcs are reported to have been found inside hillforts, but their context is unknown (e.g. two examples in Viladonga: Chamoso Lamas 1977). They more often appear in the surroundings of hillforts and frequently as part of hoards, along with other torcs, ingots and other jewellery or fragments of jewellery. The hoards from Foxados (A Coruña) and A Madorra (Lugo) are good examples (Carro García 1943; Ladra Fernandes 1997-98). At least one torc was recovered from a watery context (Bouza Brey 1965). The torcs from Chaves very frequently appear in hoards far from hillforts. A votive activity may be behind most of these deposits.

Figure 31. Distribution of the main groups of torcs of the Late Iron Age. After González-Ruibal (2003a).

The torcs of the Late Iron Age show a considerable diversity. A number of stylistic categories that were identified in the 1950s can, despite some problems, still be used (López Cuevillas 1951; Monteagudo 1952). I have proposed three main groups based on the old typologies (González-Ruibal 2003a: 346-347, Fig. 4.95) (Fig. 31):

Torcs of the oppida. These can be found in northwestern Portugal and southwestern Galicia. They appear in the area occupied by oppida, are always discovered inside these large settlements, and their main features are their bell- or urn-shaped terminals. The terminal profiles sometimes recall a certain type of vessel from the Late Iron Age, which was used for storing jewellery. These torcs often have a very baroque decoration of undoubtedly Mediterranean inspiration. The bar has a round cross-section. The best examples come from the oppidum of Lanhoso (see below). Other examples were located in the oppida of Sanfins (Jalhay 1950) and Santa Trega (Carballo 1994).

Torcs of Chaves. These appear in northeastern Portugal, and the best examples come from the upper Tamega basin (Chaves). They also exhibit profuse decoration. The terminals have double mouldings, called scotiae. The bar often has a square cross-section. The best example is a piece fortuitously discovered near Vilas Boas, not far from Chaves.

Northern torcs. They cluster round the northern part of Lugo and eastern A Coruña (Galicia). Within this wide group several types are known. The most characteristic types are:

• Artabrian type. Named after a pre-Roman people that occupied part of the province of A Coruña (Greek Artabroi, Latin Artabri, also called Arrotrebae) (Monteagudo 1952). The main characteristics are the wire that covers two thirds of the bar and the spirals (two or four). The ends can be pear-shaped or have double mouldings (scotiae). They mainly appear in A Coruña, but some examples are also known from Lugo. Good examples are those from Orbellido and Viladonga. The bar has a round cross-section.
  
  
• Asturian-northern Gallaecian (Astur norgalaico). In this type two thirds of the bar is covered with wire but instead of spirals the central part of the torc is decorated with filigree or simply left devoid of decoration. The ends have double mouldings (scotiae). The best examples come from northern Lugo: Burela, Marzán and Recarea. The bar has a round cross-section.
  
  
• Torcs with stamped decoration. The ends can be pear-shaped or have double mouldings (scotiae). The bar has a square cross-section. They are probably older than the rest, perhaps from the Middle Iron Age (fourth to third centuries BC), given the parallels with MIA pottery decoration.

There are also a large number of undecorated torcs, with square or round bar and pear-shaped or moulded ends, whose distribution coincides with the other types already mentioned. This can be due to different reasons: they may be older or they may belong to different social groups (age or class groups). It would be necessary to find this type of jewellery in a datable context if we want to interpret the variety of types and their geographical distribution. Despite these regional groups, torcs from a certain area are sometimes found far away from it. One of the best examples is the fragment of an Artabrian torc found in Troña (Pontevedra), around 100 km away from the production area in eastern A Coruña (Ladra Fernandes 1999a). This may be explained by the documented practice of gift exchange and plundering carried out by war leaders. Torc-giving is well attested in the early Irish litterature (Castro Pérez 1984-85).

Catalogue

Gold torc, Northern group, Undecorate, Melide (A Coruña, Galicia) (Fig. 32).
The region of Melide, in central Galicia, has yielded a striking number of torcs and torc fragments, most of them undecorated. This contrasts sharply with the local hillforts, which are usually of very modest size (less than one hectare) and with very few huts in perishable material. This might be because the communities despised immovable properties, while using gold and cattle as a means of accumulating wealth. This torc weighs 678 grams. Bibliography: López Cuevillas (1951: 24); Balseiro García (1994: 104-107). Museo Provincial de Lugo.

Figure 32. Torc from Melide. After Balseiro García (1994).		Figure 33. Torc from Burela. After Balseiro García (1994).	Figure 34. Torc of unknown provenance. After Queiroga (1987).

Gold torc, Northern group (Asturian-northern Gallaecian) Burela (Lugo, Galicia) (Fig. 33).
This is the heaviest piece of Iron Age jewellery found in northwestern Iberia. It weighs 1,812 kgs. The quality of the gold is