Radiocarbon and Chronology of the Iron Age Hillforts of Northwestern Iberia
J. F. Jordá Pardo1, J. Rey Castiñeira2, I. Picón Platas2, E. Abad Vidal3, C. Marín Suárez4
1 Departamento de Prehistoria y Arqueología. Universidad Nacional de Educación a Distancia. Madrid (Spain) 2 Departamento de Historia I. Universidad de Santiago de Compostela (Spain) 3 Fundación Centro Tecnológico de Supercomputación de Galicia. Santiago de Compostela (Spain) 4 Departamento de Prehistoria. Universidad Complutense de Madrid (Spain)
Zusammenfassung: Das vorgestellte Projekt hat zum ersten Mal 388 Radiokarbondatierungen von 69 eisenzeitlichen Fundstel- len in Galizien, Asturien, im nordwestlichen Kastilien und León (Spanien) und im nördlichen Portugal in einer einzelnen Datenbank zusammengefasst. In der ersten Projektphase wurde die Verlässlichkeit der Datie- rungen überprüft, um sicher zu stellen, dass der experimentelle Wert der C14-Datierungen gut mit den ar- chäologischen Datierungen vereinbar ist. Datierungen, die nicht den grundlegenden Anforderungen der Verläss- lichkeitsprüfung entsprachen, wurden verworfen. In einem zweiten Schritt wurden die als verlässlich erachteten Daten kalibriert. Diese kalibrierten Daten wurden im Hinblick auf einzelne Fundstellen und ihren regionalen und kulturell-chronologischen Kontext analysiert, um eine Chronologie der Eisenzeit dieses Raums zu erstel- len. Allgemein konnte beobachtet werden, dass die Besiedlung der Höhensiedlungen im Nordwesten der ibe- rischen Halbinsel am Anfang des ersten Jahrtausends v.Chr. beginnt. Ebenfalls konnte beobachtet werden, dass im nördlichsten Teil dieses Raums dieser Prozess erst etwas später einsetzt, mit dem Beginn der ersten Eisenzeit. Im Gegensatz dazu erscheinen befestigte Höhensiedlungen in den südlicheren Regionen des westlichen Galiziens und im westlichen Duero-Tal bereits in der späten Bronzezeit. In beiden Regionen haben die Höhensiedlungen lange Laufzeiten, durch das ganze erste Jahrtausend hindurch. Die Besiedlung setzt sich auf vielen Fundstellen auch während der römischen Kaiserzeit fort. Die kalibrierten Datierungen weisen darauf hin, dass die Besiedlung der Höhensiedlungen hauptsächlich in den Zeitraum zwischen 8. Jh. v. Chr. und 2. Jh. n. Chr. fällt. Während des 3. und 4. Jh. n. Chr. ist eine generelle Aufgabe dieser Siedlungsfundstellen zu beobachten, mit punktuellen späteren Besiedlungsphasen bis in die Neuzeit und in einzelnen Fällen sogar in die Gegenwart.
81 Abstract The present work gathers 388 radiocarbon dates from 69 Iron Age sites, located in Galicia, Asturias, NW Castilla y León (Spain) and N Portugal, in a single database for the first time. In the first phase of our work, we examined the validity of the dates in order to ensure that the experimental value of the radiocarbon dates corresponds well with the archaeological dates, rejecting all those dates that do not adjust to the basic requirements for their validation. In a second phase, we calibrated the dates that we consider reliable. These calibrated dates have been analysed according to individual sites and to regional and chronological - cultural contexts, in order to establish a chronology for the Iron Age of the this area. In general, we observe that the occupation of the hillforts of Northwestern Iberia starts at the onset of the first millennium BC. It is also observed that this process is delayed in the most northern area of the studied zone, where this process does not begin until the First Iron Age. In contrast, in the more southern areas of western Galicia and the western Duero valley, hillforts emerge in the Late Bronze Age. In both zones the hillforts have a long occupation during the Iron Age throughout the whole of the first millennium BC. This occupation continues in many of the settlements during the Roman period. The calibrated dates indicate that the occupation of hillforts takes place mainly between the 8th century BC and the 2nd century AC. Finally, a general abandonment of the settlements is observed from the 3rd-4th centuries AC, with later puctuated occupations, which in some sites appear into the Modern Age and even, in some cases, to the present day.
1. Introduction er to the east, the Atlantic Ocean to the west and the Cantabric Sea to the north. Today, this area constitutes Since the radiocarbon dates of Mohías hillfort (As- Asturias, Galicia (provinces of A Coruña, Lugo, Ou- turias, Spain) were published in 1971 by an amateur rense and Pontevedra), northern Portugal (districts of archaeologist, the doctor Jesús Martínez Fernández Viana do Castelo, Braga, Vila Real, Bragança, Aveiro (1971), archaeological research into the Iron Age hill- and Viseu) and northwest Castilla y León (provinces forts of Northwestern Iberia have provided a high of León and Zamora). number of radiocarbon dates. However, until now, all In this study we have considered all sites of the Iron these dates have not been analyzed as a whole, ex- Age with radiocarbon dates, including plain settle- cept in pioneering works of a general (Carballo, Fábre- ments, mining activities and sepulchral caves as well as gas 1991; Picon 2008) and regional type (Rey 1996; hillforts (“castros” in Spanish terminology). That is, all Cuesta et al. 1996). This work represents the beginning radiocarbon dates on material ranging from the Late of a coordinated treatment of these radiocarbon dates. Bronze Age until the contact with Roma and the first We present the results of an interdisciplinary study un- steps of Romanization have been included. In total, we dertaken by a team of different specialists: three archae- have considered 69 archaeological sites. ologists, one geographer and one geologist, from four The principal aims of this work are: i) the construc- different work centres of Spain. tion of a database containing all the radiocarbon dates The study area corresponds with the region occu- of the Iron Age of NW Iberia, ii) the accomplishment pied by the hillforts in the northwestern corner of the of a spatial data infrastructure by mean of a GIS; iii) the Iberian Peninsula (Fig. 1), one of the Atlantic Finister- analysis of all radiocarbon dates; iv) the development rae of Antiquity. Though its limits are currently debat- of strategies of action for future investigations; and v) ed, this region essentially corresponds with the Duero the dissemination of the collected information on the river at south (or perhaps the Vouga river), the Esla riv- internet for the scientific community.
82 Fig. 1. Geographical location of the Iron Age sites with 14C dates of Northwestern Iberia.
83 The methodology that we have used in our work con- The dates were obtained in 15 radiocarbon sists of a) compilation and evaluation of all radiocar- laboratories: bon dates published until 2008; b) calibration of the dates that have passed a validity test; c) integration of - Beta: Beta Analityc Inc., Miami (Florida, USA), the data within a GIS or spatial data infrastructure; - CSIC: Laboratorio de Geocronología, Institu- and d) analysis of the dates in two ways: first focusing to de Química-Física Rocasolano (CSIC), Madrid on each hillfort and second analysing all the hillforts (Spain), as a group. This analysis has been done according to - CU: Department of Hydrogeology, Charles Univer- regional and chronological-historical contexts. There- sity, Prague (Czech Republic), fore, we have established a chronology for the Iron Age - GaK: University of Gakushuin, Tokyo (Japan), of NW Iberia. - Gd: Radiocarbon Laboratory, Technology Institute of Physics, Silesian University, Gliwice (Poland), - GiF: Centre de Faibles Radioactivites, CNRS-CEA, 2. The radiocarbon dates and their calibration Gif-sur-Yvette (France), Currently, there are 388 radiocarbon dates obtained - GrN: Isotopes Physics Laboratory, University of from 69 archaeological sites in the Northwestern Groningen, Groningen (Germany), Iberian Iron Age. These dates are widely dispersed - I: Teledyne Isotopes, Inc., Huntsville (Alabama, throughout the literature. Due to this and the limit- USA), ed space which we have here, we will mention only - ICEN: Laboratorio de Isotopos Ambientais, Institu- those publications that present the basic summaries of to Tecnológico e Nuclear, Sacavém (Portugal), dates of one or of several sites namely: Carballo and Fá- - Ly: Laboratoire de Radiocarbone, Centre de Da- bregas (1991), Cuesta et al. (1996, 2000), Rubinos et tations et d’Analyses Isotopiques, Claude Bernard al. (1999), Rubinos and Alonso (2002), Alonso (2002), University, Lyon (France), Villa (2002), Arias and Fábregas (2003) and Picón - PAL: Radiocarbon Laboratory Palynosurvery Co., (2008). Unpublished dates have been provided to us Tokyo (Japan), by their researchers. For example, the Caravia hillfort - Ua: Laboratorio Ángstrom, University of Uppsala dates have been provided by Gema Adán (Cid et al., (Sweden), in press) and Neixón hillfort dates provided by Xurxo - UBAR: Laboratori de Dataciò per Radiocar- Ayán. The dates of Coaña hillfort have been obtained boni (UBAR), University of Barcelona, Barcelona from their preliminary publication in the journals of (Spain), Asturias. - UGRA: Laboratorio de Datación por Carbono 14, We have submitted all of the dates to an analysis of Centro de Instrumentación Científica, University validity to ensure that the radiocarbon dates corre- of Granada, Granada (Spain), and spond to their archaeological contexts. All dates that - UtC: Utrecht van der Graaf Laboratorium, Univer- do not fulfil the requirements have been rejected and sity of Utrecht, Utrecht (The Netherlands). are not included in the following analysis (Mestres, 1995, 2000; Mestres, Nicolás, 1997). Our criteria fall With regards to the accuracy of the laboratories, we into two categories, those associated with the radio- have rejected only the dates obtained in the labora- carbon measurement, and those related to the archaeo- tory of Gakushuin University (GaK),Tokyo (Japan). logical association of the material dated. The technical Dates from this laboratory are known to be problem- requirements include laboratory accuracy (contamina- atic (Carballo, Fábregas 1991; Castro et al. 1996). With tion elimination and radiocarbon measurement) and respect to the precision of the dates, we have reject- measurement precision (standard deviation) whilst the ed dates with a standard deviation of greater than 100 archaeological requirements include the association years. and synchrony of the material dated with the event Two further aspects that must be considered cor- of interest. respond to the ability to compare results obtained by
84 Fig. 2. Accumulated probability curve obtained from the calibration of the radiocarbon dates of A Coruña hillforts, using the calibration curve CalPal 2007 Hulu, included in the CalPal software (Version March 2007) (Weninger et al., 2007). The calibration curve shows the plateaux called “the Iron Age catastrophe”. different dating methods and results obtained from dif- bones, shells, plant species of short or long life, etc). ferent materials. Separate analysis of the dates obtained Therefore, again we will treat the dates as an homoge- by the conventional radiocarbon method and AMS neous group, though we are conscious of the variations would lengthen the analysis. We are conscious that that might be present within this dataset. such an independent treatment might lead to slightly The reliability of the archaeological association is different conclusions. However, the small percentage difficult to assess as it can only be examined from of AMS measurements (5.4 %) suggests that the results the published information about the sample context. of the analysis are unlikely to be vastly different. There- Therefore, unfortunately we are forced to accept that fore, in this work we will treat both groups of dates as a correct association exists between the dated sample an homogenous (uniform) group. and the archaeological context of interest. We can only Within the literature the detail with which radiocar- reject a date when it is clearly anomalous when com- bon dates are reported varies considerably. Unfortu- pared to its archaeological context. nately, this hampers any analysis of the potential impact After undertaking this evaluation we have obtained of different sample material (e.g. individual charcoal 331 valid dates from 61 sites. fragments, bulked charcoal, carbonaceous sediments, To place the radiocarbon dates on a calendric scale,
85 Fig. 3. Accumulated probability curves obtained from the calibration of the radiocarbon dates of the Iron Age sites of Asturias, using the calibration curve CalPal 2007 Hulu, included in the CalPal software (Version March 2007) (Weninger et al., 2007).
86 Fig. 4. Radiocarbon chronology of the Iron Age sites of Northwestern Iberia, using the calibration curve CalPal 2007 Hulu, included in the CalPal software (Version March 2007) (Weninger et al., 2007).
we calibrated the accepted dates using CalPal (2007 by International Calibration working group for the last March-June version) of the University of Cologne 24.000 years BP (Weninger et al., 2005). (www.calpal.de) (Weninger, Jöris 2004; Weninger et One of the major problems affecting calibration al., 2007). This program transforms the Gaussian dis- within the Iron Age is the plateau in the curve be- tributions of the uncalibrated date into equiprobability tween 800 and 400 cal BC, called “the Iron Age ca- curves. This allows the visualisation of the chronomet- tastrophe” (Fig. 2). This results in large error ranges ric trends and the evalutation of the synchrony and for calibrated dates within the First Iron Age (Alonso diachrony of individual and separate sites. We have 2002). However, the elimination of dates with a stand- used the calibration curve CalPal 2007 Hulu (Fig. 2), ard deviation greater to 100 years, has slightly mini- which is very similar to the IntCal04 curve proposed mised this problem.
87 Fig. 5. Chronological amplitud of the Iron Age in the different geographical areas of nothwestern Iberia, using the calibration curve CalPal 2007 Hulu, included in the CalPal software (Version March 2007) (Weninger et al., 2007).
3. Chronological analysis territorial space, occured at a later date than in adjoin- ing areas. Indeed, many syntheses suggest that in NW Traditionally, the definition of Protohistoric periods Iberia a real Iron Age did not exist, but rather, that the of NW Iberia have been primarily based on histori- region remained in a dark stadium between the Late cal events of the Roman era and typological chronol- Bronze Age and the Roman epoch (Rey 2001). ogies. According to the interpretation of this region The radiocarbon dates obtained in the NW of Ibe- as a finisterrae or periphery of peripheries, it has been ria provide sufficient information to begin to study the assumed that both technological and social develop- Iron Age groups into a calendaric timescale, though ments, for example the adoption of iron metallurgy or they are not quite as precise as we would wish. The ra- the increasing sophistication of the use of domestic or diocarbon dates clearly and consistently demonstrate
88 Fig. 6. Descriptive maps of Northwestern Iberia with the number of dates within each hillfort, the validity of dates, the 14C methods used (conventional or AMS) and the material dated.
89 Fig. 7. Chronological maps of the 14C dated hillforts of Northwestern Iberia.
90 that the hillforts of NW Iberia had clear phases of de- sult is shown in figure 5. This allows to visualize two velopment from the Final Bronze until the Roman aspects in relation with the settlement process in the epoch. considered area: We have analyzed the calibrated dates both accord- Chronologies of individual sites indicate that the ing to individual sites (Fig. 3) and jointly for region- length of occupation varied considerably reflecting al (Fig. 4) and chronological-cultural contexts (Fig. 5), short, medium and long term durations, restricted to establishing a seriation that we have compared with a single chronological phase or including several phas- the periodization realized for the Iron Age of the ge- es. At some hillforts all radiocarbon dates are concen- ographical area considered (Rey 1996; Arias 2002; trated in the First Iron Age, others are restricted to the González 2006/2007). We have marked the conven- Second Iron Age, whilst others include both. In some tional limits between the different phases in a diffuse cases radiocarbon dates extend through the Roman way, establishing transitions between every phase, as we epoch until Middle Ages. Some hillforts are located show in the following scheme: in the same place as Early, Middle or Late Bronze Age occupations. - Early and Middle Bronze Age: until 1200 cal. BC The radiocarbon chronology indicates that the stages (until 3000 BP). of settlement and abandonment occurred probably in - Late Bronze Age: between 1200 cal. BC and 800 connection with economic and social cycles. Although cal. BC (3000 – 2650 BP). This period corresponds there are continuities between the principle stages of with the phase Ia of González’s (2006/2007) perio- occupation throughout the region, in many cases these dization for the NW Iberia hillforts. stages are delimited by generalised foundations and by - 1st Iron Age: between 800 cal. BC and 500-450 cal. periods of decay and abandonment. Even though there BC (2650 – 2400 BP). This period correspond with is continuity between the principal stages of occupa- the Initial Phase of Rey (1996), the Castrexo Inicial tion – which correspond to the 1st Iron Age, transition period of Arias (2002) and with the phases Ib and 1st-2nd Iron Age, 2nd Iron Age and Roman epoch – the Ic of González’s (2006/2007) seriation. intervals are limited by widespread foundation events - 2nd Iron Age: between 500-450 cal. BC and 100- and periods of abandonment. 50 cal. BC (2400 – 2050 BP), correlated with the Middle Phase of Rey (1996), the Castrexo Pleno 4. Geographical analysis period of Arias (2002) and the phases II and IIIa of González’s (2006/2007) seriation. For the development of the spatial data infrastructure - Roman Period: between 100-50 cal. BC and 400 we have used ArcGis that allows us to geographical- cal. AD (2050 – 1600 BP), correlated with the Fi- ly analyze the information. Within the geographical nal Phase of Rey (1996), the Castrexo Final period framework of our area of study we have introduced of Arias (2002) and the phases IIIb, IIIc and IIId of much data about the hillforts and their radiocarbon González’s (2006/2007) seriation. dates, including for example, photographs of hillforts - Late Roman Period and Middle Age: 400 – 900 cal. and archaeological materials. AD (1600 - 1000 BP). In relation to the geographical context, we have made two analyses, the descriptive and the sequen- Figure 3 shows a comparison of all sites (hillforts, se- tial one: pulchral caves and mining activities) of Asturias in the The descriptive analysis includes maps with the lo- north of the study area within their geographical con- cation of the radiocarbon dated hillforts, the hillforts text. Figure 4 shows a comparison of all geographical within the Roman geography, the number of dates areas using the current administrative divisions. within each hillfort, the validity of dates, the radiocar- Once the boundaries between the cultural intervals bon methods used, conventional or AMS, and the ma- described above had been established the construction terial dated, among other aspects (Fig. 6). of the radiocarbon chronology was complete. The re- In relation with the chronological analysis we have
91 Fig. 8. Isochronal map of the first occupation and diachronic map of the 14C dated hillforts of Northwestern Iberia. made a periodization by grey scale in order to see the to Roman period, Roman Period, and Late Roman evolution by means of sequential maps that show the and Middle Age (Fig. 7), and through the use of an is- evolution of the hillforts occupation from the Ear- ochronal map of the first occupation of the hillforts ly and Middle Bronze Age until the Middle Age, in- (Fig. 8). cluding Late Bronze Age, First Iron Age and transition The isochron map of the first occupation of the hill- to Second Iron Age, Second Iron Age and transition forts (Fig. 8) shows information of certain interest: this
92 Fig. 9. Accumulated probability curves obtained from the calibration of the radiocarbon dates of San Chuis hillfort (Asturias) grouped by their chronostratigraphical position, and archaeological materials associated: 1, Examples of pre-Roman pottery; 2, Examples of Roman common pottery, with continuity of some characteristics of the pre- Roman pottery; 3, Examples of Terra Sigillata Hispanica; 4, Roman thin-walled vessel; 5 and 6, beltplates; 7, omega fibula; 8, pendant.
93 Fig. 10. Accumulated probability curves obtained from the calibration of the radiocarbon dates of A Graña hillfort (A Coruña) and chronostratigrapy of its habitat structure evolution and archaeological materials associated.
94 map demonstrates the wide distribution of the radi- in the First Iron Age, according to both the typology ocarbon points corresponding to occupation of the and the radiocarbon dates. hillforts during the Late Bronze Age. This fact dem- This contrast, comparison of the radiocarbon chro- onstrates that the process of fortification of the settle- nology and and the techno-typological evidence ments was a generalized worry in NW Iberia and is in eradicates the idea of a late incorporation of iron met- contrast to the currently supported idea that the proc- allurgy. This material appears during the radiocarbon ess had spread from the south to the north and from interval of the First Iron Age, when tools of the Late the coast to the interior, such that the forts first appear Bronze Age were still in use. The presence of import- in the south and on the coast during the Late Bronze ed products and aesthetic and technological accul- Age, whilst in the North and the interior they do not turations from Southwestern Iberia in the imported appear until the First Iron Age. products during early radiocarbon periods invalidate the idea that the culture that was isolated until the Ro- mans came to NW Iberia. 5. Archaeological analysis: contextual readings of The definition of the Second Iron Age presents two Northwestern Iberia radiocarbon main problems. First, many deposits containing the The analysis of the radiocarbon dates must be incorpo- phase were removed during the Roman period. Sec- rated into the archaeological chronologies (typological ond, its intermediate typological position means that and stratigraphical) in a systematic way. it can be difficult to identify. Its description is limited It is evident that the the results obtained by radiocar- to its nature as a mixture of some more evolved fea- bon provide very suggestive images, but it is necessary tures relating to Roman period alongside earlier Late to compare them and to integrate them with infor- Bronze Age features. Nevertheless the importance of mation obtained from other types of archaeological this phase from the radiocarbon point of view, seems to evidence, such as typological and stratigraphic inter- increasingly confirm the typological sequences and the pretations. This is still a hanging topic in the archaeol- events that have been fixed within some stratigraph- ogy of the hillforts of NW Iberia. It is rare that studies ic records. For example, the fortification of settlements involving different methodologies are connected with- becomes more complex, iron metallurgy is evident and in publications, such as archaeological reports. In order the wide use of the iron produced is demonstrated by to begin to analyse the data in this way, the advantages material finds as well as evidence of its use in con- in the change in approach must be demonstrated. struction, agriculture, carpentry, etc. In addition, many In general, we can estimate that the stratigraphic se- other crafts, such as pottery production, goldworking quences and the archaeological materials will extend or sculpture incorporate ideas and technological and the radiocarbon stages recognized for each site. We aesthetic innovations related to the manufacture and can extend the number of places with long occupation, consumption of iron. from the Early and Middle Bronze Age to the Late The radiocarbon dates of the Roman period in the Roman period and Middle Age and we can clarify the NW Iberian hillforts indicate the foundation of settle- phases of occupation corresponding to the hillforts. ments, continuity of settlements and their decline. In The application of the radiocarbon method in many many cases it is necessary to examine imported prod- cases has been centred on certain archaeological lev- ucts because their presence can provide chronologi- els and has been excluded in others where, for exam- cal information during moments without radiocarbon ple, objects that can provide precise dates are present or dates. The history of the consumption of these prod- the level is of little importance within the aims of the ucts indicates that the hillforts demanded them with excavation. The comparison of the radiocarbon and ar- great intensity during the first times of contact and chaeological chronologies also allows us to state that that this consumption languishes when the process of some hillforts are of short duration and are restricted romanization is intensified. The demand moves then to to a single stage. The use of the hillforts of Torroso, Pe- the cities and to the villas. nalba or Penarrubia, for example, begins and finishes The analysis of the radiocarbon dates from the As-
95 turian hillforts has allowed us to establish a precise of the stone houses at A Graña, indeed later than in chronology for the multitude of bronze objects that other settlements of the Northwest Iberia (Fig. 10). A traditionally have been placed in the Late Bronze Age similar situation exists for the adoption of iron met- through the use of the relative typological chronolo- allurgy and the introduction of the rotary querns in gy (derived from imprecise typological work) (Fan- this hillfort. jul, Marín 2006). Now, using radiocarbon we can In A Graña hillfort the radiocarbon dates have al- link many of these artefacts to the first hillforts of the lowed us to place on an absolute timescale some ar- Cantabrian zone (First Iron Age) and reject the clas- chaeological materials (such as ear-rings, fíbulae or sic chronologies for this zone. This allows us to better ceramics), establishing a direct relation with the dates understand the ancient phase of the Cantabrian hill- (Fig. 10). Until now the typological seriations have forts. This appears to suggest a drastic cultural change been stylistic and evolutionary and few typological of the mobile populations of the Bronze Age of this stages have been fixed by calendaric chronologies. zone. Also we could have checked that the character- Definitively, the information published from the A istic walls of modules of the western Cantabrian zone Graña hillfort illustrates well the potential that ex- arise in the transition between the First and the Sec- ists for improving the archaeological understanding ond Iron Age, and simultaneously have place a gener- of this period. The chronological milestones that pro- alization of the stone architecture. vide the objects and that give historical and cultural Until middle of the nineties it was thought that the value to the stratigraphic sequences are not current- Asturian hillforts were associated with the Roman ly developed. Conversely, the stratigraphic sequence conquest. Our recent investigations have allowed us identified in A Graña has not provided the stylistic to establish the sequence of the ceramic products for history of the materials from all of its chronological the entire chronological period of use of the hillforts, stages, which would allow us to be more certain of which in many cases is approximately a millennium. the age of the events. For example, at the paradigmatic case of San Chuis hillfort (Fig. 9) we have identified pottery of the First 7. Conclusions Iron Age for the first time in Asturias (Marín, Jordá 2007). We propose the first radiocarbon chronology of the In contrast..., many NW Iberian hillforts offer Iron Age of the Northwestern Iberia. scanty archaeological evidence. Thus, it is very dif- The calibrated dates indicate that the occupation of ficult to relate archaeological events to the radiocar- hillforts takes place mainly between the 8th century BC bondates obtained. For example, in A Graña hillfort and the 2nd century AC. (Galicia) (Meijide 1990), the radiocarbon dates estab- We also perceive that the hillforts begin in the Late lish two discontinuous calibrated intervals: one relat- Bronze Age in western Galicia and the western Due- ed to the first occupation during the First Iron Age, ro valley. and other related to the transition between the mid- On the contrary, this process is delayed in the most dle and late phases of the hillfort, situated around the northern area of the studied zone, where this process change of era. Similarly the typological chronolo- doesn’t begin until the First Iron Age. gies show a long duration of settlement originating In both zones the hillforts hold a long occupation in the First Iron Age. However, they extend its de- during the entire Iron Age. velopment until the Late Roman period. The strati- This occupation continues in many of the settle- graphic sequence of the domestic structures shows an ments during the Roman period. uninterrupted occupation, with structures from the Finally, a general abandonment of the settlements First Iron Age until the change of era, when stone is observed from the 3rd-4th centuries AC, with later houses are built for the first time. The radiocarbon punctuated occupations, which, in some sites, occur date obtained from the foundation of a stone house throughout the Middle and Modern Ages, even un- establishes a very late moment for the incorporation til the present day.
96 Acknowledgements We want to express our gratefulness to María Gómez and Marina González who helped us to prepare the manuscript in English and very specially to Rachel Wood who checked the English translation of our text and made interesting commentaries. Special thanks to Gema Adán and Xurxo Ayán who provided us un- published radiocarbon dates from Caravia and Neix- on hillforts.
References
Alonso Matthias, F. (2002), Fechas de carbono-14 en los castros en el Noroeste de la Península Ibérica (1200 a.C.-50 d.C.). asturianos. In: de Blas Cortina, M.A., Villa Valdés, A. [eds.], Los Brigantium, 18-19. poblados fortificados del noroeste de la Península Ibérica: for- Marín Suárez, C., Jordá Pardo, J.F. (2007), Las cerámicas indígenas mación y desarrollo de la cultura castreña. Navia: Ayuntamien- del Castro de San L.luis (Allande, Asturias). In: Fanjul Peraza, to de Navia – Parque Histórico del Navia: 337-344. A. [ed.], Estudios varios sobre Arqueología castreña. A propó- Arias Vilas, F. (2002), Las fases de la Cultura Castreña en Galicia: un sito de las excavaciones en los castros de Teverga (Asturias). debate abierto. In: de Blas Cortina, M.A., Villa Valdés, A. [eds.], Septentrión. Santander: Instituto de Estudios Prerromanos y Los poblados fortificados del noroeste de la Península Ibérica: de la Antigüedad - Ayuntamiento de Teverga: 135-152. formación y desarrollo de la cultura castreña. Navia: Ayunta- Martínez Fernández, J. (1971), Castro de Mohías: resultados de una miento de Navia – Parque Histórico del Navia: 127-137. investigación geocronológica. Boletín del Instituto de Estu- Arias Vilas, F., Fábregas Valcarce, R. (2003), Datacións radiocarbó- dios Asturianos, 73: 351-356. nicas do castro de Viladonga (Lugo). Gallaecia, 22: 193-210. Martínez-Moreno, J., Mora Torcal, R., Casanova i Martí, J. (2006), Carballo Arcedo, L.X., Fábregas Valcarce, R. (1991), Dataciones de El Mesolítico de los Pirineos surorientales: una reflexión sobre Carbono 14 para castros del noroeste peninsular. Archivo Es- el significado de las “facies de fortuna” del Postglaciar. In: Al- pañol de Arqueología, 64: 244-264. day Ruíz, A. [ed.], El Mesolítico de muescas-denticulados en Castro Martínez, P.V., Lull, V., Micó, R. (1996), Cronología de la Cuenca del Ebro y el Litoral Mediterráneo peninsular. Me- la Prehistoria Reciente de la Península Ibérica y Baleares (c. morias de Yacimientos Alaveses, 11. Vitoria: Diputación Foral 2800-900 cal ANE). BAR International Series, 652. Oxford: de Álava: 158-188. Archaeopress. Meijide Cameselle, G. (1990), Tres dataciones de C14 del castro Cid López, R.M., González Santana, M., Adám Álvarez, G.E. (in de A Graña (Toques, A Coruña) y su contexto arqueológico. press), El Castro de Caravia y la Edad del Hierro en la Asturias Gallaecia, 12: 111-139. Oriental: Caravia – Diadema de Moñes - Vadinienses. In: Jordá Mestres, J.S. (1995), La datació per radiocarboni i el calibratge de Pardo, J.F., Marín Suárez, C. [eds.], Arqueología castreña en les dates radiocabòniques. Objectius, problemes i aplicacions. Asturias: novedades y propuestas metodológicas. Entemu, 16. Revista d’Arqueologia de Ponent, 5: 260-275. Cuesta, F., Jordá Pardo, J.F., Maya, J.L., Mestres, J.S. (1996), Radio- Mestres, J.S. (2000), La datació per radiocarboni. Una visión actual. carbono y cronología de los castros asturianos. Zephyrvs, 49: Tribuna d’Arqueologia, 1997-1998: 195-239. 225-270. Mestres, J.S., Nicolás, J.C. (1997), Contribución de la datación por Cuesta, F., Maya. J.L., Mestres Torres, J.S. (2000), La fase prerroma- radiocarbono al establecimiento de la cronología absoluta de la na de los castros asturianos según el radiocarbono. In: Oliveira prehistoria menorquina. Caesaraugusta, 73: 237-341. Jorge, V. [Ed.], 3º Congresso de Arqueologia Peninsular, 5, Vila Picón Platas, I. (2008), Unha aproximación a través do C14 a la Real: UTAD: 477-494. cronoloxía castrexa. Gallaecia, 27: 155-177. Fanjul Peraza, A., Marín Suárez, C. (2006), La metalurgia de hierro Rey Castiñeira, J. (1996), Referencias de tempo na cultura ma- en la Asturias castreña: nuevos datos y estado de la cuestión. terial dos castros galegos. In: Hidalgo Cuñarro, J.M. [ed.], A Trabajos de Prehistoria, 63 (1): 113-131. cultura castrexa galega a debate. Tui: Instituto de Estudios Tu- Gonzalez Ruibal, A. (2006/2007), Galaicos. Poder y comunidad denses: 157-206.
97 Rey Castiñeira, J. (2001), Apuntes para un encuadre de la cultu- de la Península Ibérica: formación y desarrollo de la cultura ra castreña en el marco peninsular. In: Oliveira Jorge, V. [ed.], castreña. Navia: Ayuntamiento de Navia – Parque Histórico 3º Congresso de Arqueologia Peninsular, 5, Vila Real: UTAD: del Navia: 159-188. 359-372. Weninger, B., Danzeglocke, U., Jöris, O. (2005), Comparison of Rubinos, A., Alonso, F. (2002), Las aplicaciones del carbono-14. Dating Results achieved using Different Radiocarbon-Age In: Berrocal-Rangel, L., Martínez Seco, P., Ruiz Triviño, C. Calibration Curves and Data. www.calpal.de. Köln: Universi- [eds.], El Castiellu de Llagú. Un castro astur en los orígenes de tät zu Köln, Institut für Ur- und Frühgeschichte. Oviedo. Madrid: Real Academia de la Historia – Principado Weninger, B., Danzeglocke, U., Jöris, O. (2007), Glacial radiocar- de Asturias: 297-303. bon age conversion. Cologne radiocarbon calibration and Rubinos Pérez, A., Fábregas Valcarce, R., Alonso Matthias, F., palaeoclimate research package
Zusammenfassung der Diskussion (Beitrag von Karl) The shown sequence has parallels in Wales and western Britain e.g. in the late bronze age with about 1000 years of occupation in some regions. In Spain there are clear distinctions between coastal regions and landlocked areas. Sometimes while the hillfort-occu- pation increases, the use of lowland settlements is being reduced. Sometimes lowland settlements even seem to be abandoned, when hillfort usage starts. But any conclusion to this matter is not very strongly scientifically based, because most of the available data (e.g. radio carbon dating) is linked to hillforts (92%, lowland settle- ments only 3%).
98