IRON AGE FIBULAE FROM CASTRO DE PRAGANÇA (PORTUGAL)

Proceedings of the 2nd International Conference Archaeometallurgy in Europe Aquileia, Italy, 17-21 JUNE 2007. Organized by Associazione Italiana di Metallurgia, Milano IRON AGE FIBULAE FROM CASTRO DE PRAGANÇA (PORTUGAL) Ana Ávila de Melo Museu Nacional de Arqueologia, Portugal Elin Figueiredo DQ, Instituto Tecnológico e Nuclear, Portugal Maria de Fátima Araújo DQ, Instituto Tecnológico e Nuclear, Portugal João C. Senna-Martinez FL, Universidade de Lisboa, Portugal ABSTRACT Castro de Pragança is located in Estremadura province, in the Portuguese western coast. The site has been identified as a Copper Age fortified settlement by most of Portuguese archaeologists. However it shows evidence of later occupation, during Bronze and Iron Ages. Although most of the metal artefacts found in the site can be dated to Copper and Bronze Ages, an unusual set of nine fibulae was recovered that covers the period from the Iron Age until the Roman conquest. Nondestructive elemental analyses showed they are mostly bronzes alloys with variable lead contents. However, four of the nine fibulae are made both of bronze and iron pointing out to very important archaeological and technological issues. Until the 5 th century BC the Portuguese Estremadura was a strategic region where both Mediterranean and Atlantic influences were visible in the material culture, but during the 2 nd Iron Age another link increasingly grows its influence the Continental one. KEYWORDS Iron Age; bronze; fibulae; Portugal; EDXRF INTRODUCTION The site of Castro de Pragança (Cadaval, Estremadura) is located on a hill top near Cadaval, a village, about 55km North of Lisbon, in the Portuguese province of Estremadura and was discovered by the end of the 19th century (1893). Known as castle by the local population at the time of its discovery, it was excavated by Dr. J. Leite de Vasconcelos the first director of the Museu Nacional de Arqueologia (Lisbon, Portugal). During his archaeological fieldwork he performed several campaigns until the beginning of the 20 th century - he recovered an enormous amount of artefacts, including ceramics and one of the largest collections of metal artefacts and metal debris found in a Portuguese prehistoric settlement. Later, about 1930, L. Trindade surveyed this site and performed some test trenches. In 1988 and 1990 more than fifty years later - new excavations were carried out by J. L. M. Gonçalves [1] who focused the fieldwork mainly in the identification of the fortification wall perimeter. 1

Iron Age Fibulae from Castro de Pragança (Portugal) Figure1 - Castro de Pragança (photo by Ana Sousa) and its location in the Portuguese territory (at top left) The Estremadura Copper Age is frequently related with the theme of fortified settlements. Castro de Pragança is no exception, and several papers published about this settlement focused mainly in this issue [1]. Although the site is often quoted as an example of the Calcolithic fortified settlement from western Portugal, it has in fact a long time span occupation from Copper Age through Bronze and Iron Ages. And above all, Castro de Pragança provided one of the largest collections of prehistoric metal artefacts of Museu Nacional de Arqueologia: more than 500 artefacts, artefacts fragments and metal remains, including slag and metal debris. Among all these metal artefacts and remains there is a group of nine fibulae dated from the Iron Age, which provide a rather unusual finding for such a settlement in Estremadura. 1. CASTRO DE PRAGANÇA FIBULAE Tools and weapons are the majority of metal artefacts recovered from Castro de Pragança, although there is a significant group of artefacts that can be classified as adornments. The majority of the artefacts can be dated from the Bronze Age, but there are two important groups that can be dated from Copper and Iron Ages. Among these there is a set of nine fibulae that is most unusual. Prehistoric fibulae are prestige items and for that reason they are usually recovered from burials. Castro de Pragança fibulae can be dated from the Iron Age through to the Roman Conquest, according to typological classification. This collection is composed of nine fibulae fragments, but none of them is complete in some cases the pin misses, or the bow is not complete. The regular absence of some of the fibula components bow, head and foot turned more difficult the typological classification of these artefacts. Therefore, the typological classification was made mainly according to one or two of the components, usually the bow and sometimes also the head. The typology of the fibulae identifies them as local productions. Furthermore, part of this fibulae collection six fibulae - has recently been classified and published [2]. 2

Ana Ávila de Melo, Elin Figueiredo, Maria de Fátima Araújo & João C. Senna-Martinez This collection may be divided in four main typological groups Acebuchal, Trasmontano, La Tène I and La Tène III types. 1.1 Acebuchal type Castro de Pragança has two fibulae fragments (2005.10.26 and 2005.10.27) that can be classified as Acebuchal type. One of the fibulae presents a bow composed of two divided thin plates with decoration longitudinal stria fillet - and part of the foot; a rivet fixes the bow and the foot. The bow and the foot fragment measures 63 mm length and 12.5 mm width and weights 3.36 g. The other fibula fragment, very corroded, also has a bow composed of two divided thin plates with decoration longitudinal stria fillet; it measures 50 mm length and 5 mm width and weights 1.68 g. In both fragments the spring is missing and in both cases the bow is composed of a divided thin plate. The Acebuchal type is very common during the 1 st Iron Age in the Iberian Peninsula. According to the same author [2], this particular type has been developed during 8 th to 7 th century BC, presenting western Mediterranean influences and may be related with other 1 st Iron Age Iberian fibulae, such as Alcores or Bencarrón types. Figure 2 Acebuchal type fibulae 1.2 Trasmontano type Among the wide group of La Tène I fibulae there is a particular one which is a local production of the Portuguese NE region. This group was identified in the beginning of the 20 th century [2] but recent research showed that it is widespread in northern Iberian Peninsula, namely associated with NW Iberia Iron Age Hill Fort culture (cultura castreja). This type was developed at the beginning of the 2 nd Iron Age (5 th century BC) and continued through the 1 st century BC, after the roman conquest; these fibulae are composed of two independent parts - the bow and the spring and the bow has a vessel shape. This collection has five items recently classified as Trasmontano type fibulae [2]. All of them have vessel shape bows, but they may be subdivided in two subtypes, according to the presence (2005.10.75, 2005.10.76 and 983.299.180) or absence (2005.10.79 and 2005.10.80) of a turned up foot, bent back towards the bow. Two of the five fibulae have a short foot, and one of them (2005.10.79) still has part of the pin and a longitudinal decorated bow with a very thin crest shaped plate; it measures 52 mm length and 36 mm width and weights 21.6 g; the other (2005.10.80) has also a vessel shape bow with a longitudinal frame decoration and a short foot with 48 mm length and 32 mm width and weights 26.15 g. The other subtype includes three fibulae (2005.10.75, 2005.10.76 and 983.299.180); neither of them being complete the spring and 3

Iron Age Fibulae from Castro de Pragança (Portugal) the pin missing. They have a vessel shape bow and a turned up foot, bent back towards the bow; the bow is decorated with incised lines on the extremities in all of them. Two items (2005.10.75 and 983.299.180) present a decorated baluster shape raised up foot and the other (2005.10.76) has horizontal incised lines decoration on the extremities of the bow and of the raised up foot; this smaller brooch measures 36 mm and 21 mm width and weights 7.24 g. One of the two items (2005.10.75) with a decorated baluster shape foot has a broken extremity of the bow opposite to the foot; it measures 41 mm length and 361 mm width and weights 15.93 g. The other brooch (983.299.180) with a decorated baluster shape foot measures 48 mm length and 32 mm width and weights 15.93 g Figure 3 - Trasmontano type fibulae. From left top to bottom right: 2005.10.75; 983.299.180; 2005.10.76; 2005.10.80; 205.10.79 1.3 La Tène I type The fibula 2005.10.18 is one of the most complete brooches of this collection since it has the complete spring, part of the pin and part of the bow. This item has a 20 loops bilateral spring with an iron axle, measures 60 mm length (bow and pin) and 27 mm (spring) width and weights 13.57 g. This particular type, a local production related by S. Ponte [2] with La Tène I fibulae was developed during the 1 st Iron Age (7 th century BC) and fully developed and widespread during the 2 nd Iron Age (through 6th century BC, till the 3 rd century BC at the beginning of Roman conquest). 4

Ana Ávila de Melo, Elin Figueiredo, Maria de Fátima Araújo & João C. Senna-Martinez Figure 4 - La Tène I fibula - 2005.10.18 1.4 La Tène III type The only fibula classified as La Tène III (983.299.164) has a vessel shape bow with decoration incised longitudinal lines; the head is composed of a tubular beam with part of the axle in its interior. The brooch measures 48 mm length and 26 mm width and weights 28.5 g. This particular type has a time span from the 3 rd century BC through to the 1 st century AD and was widespread through the Iberian Peninsula mostly from the end of the 2 nd Iron Age to Augustus period. Figure 5 - La Tène III fibula - 983.299.164 2. NON-INVASIVE EDXRF SUPERFICIAL ANALYSES OF THE FIBULAE All the nine fibulae were submitted to superficial non-invasive elemental analysis carried out by energy dispersive X-ray fluorescence spectrometry (EDXRF). It was used a Kevex 771 spectrometer with an incident X-ray beam that permits analyses of a circular area up to 3 cm diameter [3]. For this reason, some small components of the fibulae could not be analysed 5

Iron Age Fibulae from Castro de Pragança (Portugal) individually (as for the pin in 2005.10.79 brooch, and rivets in the 2005.10.26 brooch). Always as possible more than one analysis was performed in different areas of each fibulae component. Quantification was made using a standard material - Phosphor bronze SS551 from British Chemical Standards. Detailed descriptions of the equipment as well as the quantification procedures were previously described [4]. Once analyses are performed over corroded surfaces (unclean metal) the analytical data suffer high influence from the corrosion composition (more evident when corrosion layers are thick). Thus, if compared with the metal bulk composition, higher Sn and Pb values are expected from these analyses, due to decuprification [5] EDXRF analyses showed that four fibulae, 2005.10.79, 2005.10.80, 2005.10.77 and 983.299.164, have an iron axle and that the other components are made of a bronze alloy. The other five fibulae have all their components made of bronze. The EDXRF results for the bronze components are shown in Table 1. Table1. EDXRF results for the bronze components of the nine fibulae. Analyses refer to unclean surfaces (surfaces covered with corrosion layers). (n.d. not detected) % wt. normalised Brooch n.º Analysed area N.º of analyses Cu Sn Pb As Sb Fe Ni 2005.10.26 Bow 3 72.2±6.8 23.1±4.9 1.4±1.3 n.d. 0.1±0.0 2.2±1.1 n.d. Foot 1 71.1 14.7 10.3 n.d. 0.1 3.8 n.d. 2005.10.27 Bow 2 79.1±1.7 18.6±2.3 0.9±0.1 n.d. n.d. 1.4±0.5 n.d. 983.299.180 Bow 2 76.6±0.6 21.9±0.4 1.0±0.1 n.d. n.d. 0.5±0.2 n.d. 2005.10.75 Bow 2 62.8±0.7 34.1±1.3 1.5±0.1 0.5 0.7±0.0 0.6±0.2 0.2 2005.10.76 Bow 2 49.6±6.3 42.6±6.7 5.6±0.3 n.d. 0.9±0.1 1.3±0.3 n.d. 2005.10.79 bow+pin 3 75.9±1.3 21.2±1.3 2.6±0.2 n.d. 0.2±0.0 0.1±0.0 0.2 2005.10.80 Bow 2 70.1±10.7 24.7±7.8 4.6±2.6 n.d. 0.3±0.1 0.4±0.0 n.d. 2005.10.77 Bow 1 41.9 45.0 9.2 n.d. 2.6 1.0 n.d. Spring 1 57.4 33.8 7.9 n.d. 1.1 1.0 n.d. 983.299.164 Bow 1 64.5 30.0 4.1 n.d. 0.7 0.8 n.d. tubular beam 1 61.2 35.6 1.9 n.d. 1.2 0.8 n.d. For all fibulae, except the 2005.10.75 bow, no arsenic was detected. Low (<0.1%) or non (n.d.) antimony impurities were detected in most of the fibulae, except for the two La Tène types 2005.10.77 and 983.299.164, where Sb>1% was detected. Iron content (from 0.8 to 2.2%) is due to the incorporation of this element from the surrounding soil into the corrosion layers during burial. This can easily be confirmed through the higher iron values of the 2005.10.26 and 2005.10.27 fibulae that were the only ones that seem not to have been subjected to a conservation treatment (where most of the more superficial soil residues would be taken off and some protective layer applied in the surface of the metal). Nickel impurities are only detected in 2005.10.75 and 2005.10.79 fibulae. Tin contents range from 14.7 to 57.4%. As said before, high tin contents are expected for this superficial analysis due to decuprification [5]. Past work [6] covering artefacts from Castro de Pragança site has demonstrated that tin may reach values up to five times higher than its content in the metal bulk when using this analytical procedure. Due to decuprification phenomenon lead is also expected to show higher values [7], which can be 4 times higher as published in recent work 6

Ana Ávila de Melo, Elin Figueiredo, Maria de Fátima Araújo & João C. Senna-Martinez [8]. All the analysed fibulae show some lead content, varying from 1 to 10%, in the corroded surface of their bronze components. If decuprification is to be considered as the main corrosion phenomenon (susceptible of altering significantly the corrosion layers composition) Sn and Pb are expected to show similar enrichments in the results of the EDXRF superficial analyses. The Fig. 6 shows a graph with the Sn and Pb contents plotted. It can be seen that all the fibulae bronze constituents, except for the 2005.10.26 foot, can be grouped in an area that follows an Sn and Pb enrichment expected for similar metal compositions. Only 2005.10.26 foot constituent shows outstanding values: it shows the smallest Sn value (14.7%) and the highest Pb value (10.3%) of all the performed EDXRF analyses. This could mean that this bronze constituent is made with a bronze alloy having a higher Pb content than the bronze used for the manufacture of bronze constituents of the other fibulae. 12 10 8 2005.10.26 foot 2005.10.77 bow 2005.10.77 spring Sn and Pb "enrichement" due to corrosion Pb (% wt.) 6 2005.10.80 bow 2005.10.76 bow 4 983.299.164 bow 2005.10.79 bow +pin 2 0 2005.10.26 bow 983.299.180 bow 2005.10.27 bow 983.299.164 tubular beam 2005.10.75 bow 0 10 20 30 40 50 60 Sn (% wt.) Fig.6 Graph with Sn and Pb contents obtained through EDXRF analyses performed over unclean surfaces (covered with corrosion layers) of the nine fibulae. Most of the analysed bronze components (all excluding the 2005.10.26 foot) can be grouped in an area which shows Sn and Pb enrichment due to corrosion (higher values are due to thicker corrosion layers) illustrating that these brooch components are likely to been made in a similar bronze alloy 3. DISCUSSION Castro de Pragança as other sites from recent Prehistory in the Portuguese Estremadura province has received influences both from the Mediterranean and the Atlantic areas. In fact, the Portuguese Estremadura has a strategic location that provides a link between the Atlantic Western Europe and the Western Mediterranean, since very ancient times. It also links the Portuguese interior rich tin region of Beira Alta with the Atlantic coast [9]. The Phoenician early arrival at the Portuguese western coast increased those contacts [10], as the oriental contacts were established from the western coast towards the interior, following the river basins Tagus being the most significant one 7

Iron Age Fibulae from Castro de Pragança (Portugal) [11]. The indigenous settlements in this area have increased in number during the Late Bronze Age [12] and benefit from these oriental contacts before and after Phoenician colonisation. The iron technology is generally associated with the Phoenician colonisation of the Iberian Peninsula, but recent archaeological fieldwork in Portugal has shown a much more complex situation, with iron artefacts being recovered in indigenous settlements, even before the Phoenician arrival [13]. Contacts between indigenous populations and oriental traders were established before Phoenician colonisation and continued afterwards. Bronze and Iron Ages metal artefacts recovered in these settlements reflect this complex situation: a local small scale metal production in indigenous settlements and a massive metal production in Phoenician colonies [14]. The use of iron together with a copper alloy in a local production at such an early stage may indicate that indigenous populations rapidly adopted iron technology. 4. FINAL REMARKS Castro de Pragança Iron Age fibulae are probably local productions although with different foreign influences. EDXRF superficial analyses showed a regular use of bronze alloys with some Pb contents. Four fibulae show a noteworthy combination of bronze components with an iron axle, pointing out to the rapid use of iron in indigenous settlements. During the transition of the Late Bronze Age to the Iron Age there was an increase in the lead content in bronze artefacts from Western Europe. This is a characteristic of the so called Atlantic Metallurgy. The Portuguese reality seemed to be somehow different. Recent studies made on artefacts from Late Bronze Age Baiões/Santa Luzia site (Beira Alta, Portugal) [15] and Fraga dos Corvos site (Trás-os-Montes, Portugal) [16] show that the metal used was a bronze alloy, with traces of Pb in the first site, and with Pb present in small contents in the metal bulk (<2%) in the second site, probably as a result of ore impurities. The oriental influences decreased after the 5 th century BC in the interior and western Portuguese regions. As the oriental influence decreased, the material culture revealed the appearance of the Continental links during the 2 nd Iron Age. The Iron Age fibulae from Castro de Pragança testify the increasingly influence of the continental models that came through the interior regions of Iberian Peninsula. Acknowledgements This work has been carried out in the framework of the project METABRONZE Metallurgy and Society in Central Portugal Late Bronze Age, POCTI/HAR/58678/2004, founded by the Portuguese Science Foundation (FCT). The second author acknowledges the FCT for the SFRH/BD/27358/2006 grant. REFERENCES 1) J. L. M. GONÇALVES, O Arq. Port. 8/10, (1990/1992), p.33-40. 2) S. PONTE, Corpus Signorum das Fíbulas Proto-Históricas e Romanas de Portugal (2006), p. 233, 271, 322. 3) KEVEX, XRF ToolBox II Manual, Kevex, USA, 1990. 4) M.F. ARAÚJO, L.C. ALVES and J.M.P Cabral, Nucl. Instr. Meth. B 75 (1993), p.450 5) L. ROBBIOLA, J.-M. BLENGINO and C. FLAUD, Corr. Sci. 40 (1998), p.2083 6) E. FIGUEIREDO and M.F. ARAÚJO, The European Corrosion Cong. Proceedings, Lisbon, (2005), 0-481-4 cd. 7) J. TATE, Nucl. Instr. Meth. B 14 (1986), p.20 8

Ana Ávila de Melo, Elin Figueiredo, Maria de Fátima Araújo & João C. Senna-Martinez 8) E. FIGUEIREDO, P. VALÉRIO, M.F. ARAÚJO and J.C. SENNA-MARTINEZ, Nucl. Instr. Meth. A (in press) 9) J. C. SENNA-MARTINEZ, (1995) Actas dos Iº s Cursos Internacionais de Verão de Cascais:18 a 30 de Julho de 1994, p. 63. 10) A. M. ARRUDA, O Arq. Port. 23, (2005), p.12, 105-106. 11) R. VILAÇA and A. M. ARRUDA, Conímbriga XLIII (2004), p. 11 12) J. L. CARDOSO, Est. Oeiras 12, (2004), p.305. 13) J. C. SENNA-MARTINEZ, Trabalhos de Arq. Da EAM 6, (2003), p.43 14) J. C. SENNA-MARTINEZ, Anejos de AespA, 35 (2005), p.901 15)P. VALÉRIO, M.F. ARAÚJO, J.C. SENNA-MARTINEZ and J. L. INÊS VAZ, O Arq. Port. 24, (in press) 16) E. FIGUEIREDO, J.C. SENNA-MARTINEZ, R.J.C. SILVA, M.F. ARAÚJO, J.M.Q. VENTURA, H. CARVALHO, this issue. 9