COMUNE DI CIMITILE FONDAZIONE PREMIO CIMITILE SECONDA UNIVERSITÀ DI NAPOLI DIPARTIMENTO DI LETTERE E BENI CULTURALI UNIVERSITÀ DEGLI STUDI DEL MOLISE DIPARTIMENTO DI SCIENZE UMANISTICHE, SOCIALI E DELLA FORMAZIONE CENTRO DI STUDI LONGOBARDI ARISTOCRAZIE E SOCIETÀ FRA TRANSIZIONE ROMANO-GERMANICA E ALTO MEDIOEVO Atti del Convegno internazionale di studi Cimitile-Santa Maria Capua Vetere, 14-15 giugno 2012 a cura di Carlo ebanista e marCello rotili TAVOLARIO EDIZIONI 2015 Enti promotori Comune di Cimitile Fondazione Premio Cimitile Seconda Università di Napoli, Dipartimento di Lettere e Beni culturali Università degli Studi del Molise Dipartimento di Scienze umanistiche, sociali e della formazione Centro di Studi Longobardi Impaginazione: Laura Iodice In copertina: Città di Castello (Pg), Museo del Duomo: tesoro di Canoscio, piccolo piatto. A pagina 1: Garda (Vr), ibula a vortice. © 2015 by Tavolario Edizioni San Vitaliano (NA) tel. 0815198818 - info@tavolariostampa.com ISBN 978-88-906742-9-7 CHRISTOPH EGER HABITUS MILITARIS OR HABITUS BARBARUS? TOWARDS AN INTERPRETATION OF RICH MALE GRAVES OF THE MID 5TH CENTURY IN THE MEDITERRANEAN 1. Two conliciting points of view An important question in late antique archaeology of the Mediterranean is the interpretation of few rich graves of the 1st half and the middle of the 5th century AD on the Iberian Peninsula, in Northern Africa, and in Italy. The deceased, amongst them both men and women, were equipped with golden or gold-plated jewellery and corresponding dress accessories; some males also had part of their military gear with them1. This evidence is remarkable because, at this time, the inhabitants of many regions of the Mediterranean had long adopted the habit of burying their dead, regardless of their social standing, without any grave goods at all or only with singular pieces of jewellery, a coin or a vessel2. One wonders whether the deceased were, as has long been supposed, members of such communities of half tribal and half military character which are called gentes and dominated by barbarians and thus are a very concrete archaeological relexion of the large-scale migrations that swept foreign groups into the western Mediterranean from the late 4th century AD onwards. Or did they rather belong to a Late Roman military aristocracy that cannot ethnically be subdivided by means of archaeological criteria, likewise included Imperial Roman and barbarian elites, and distinguished itself by supra-regional dress fashions and similar customs of burial and grave goods?3 As far as archaeological inds are concerned, both positions are grounded on the same methods: namely typo-chronological and chorological analyses of grave goods and comparative studies of dress and burial habits in order to clarify their origin and distribution. Yet, the results could not be more antithetic, because different distribution models of material goods and divergent concepts of the nature and identity of Late Antique communities condi- 1 Attention is furthermore deserved by the fact that some of the grave goods seem unusual and almost strange in their regional context, but have parallels in partly peripheral areas of the Mediterranean. 2 Social differences were mainly visible in the construction and position of the grave. For an overview of burial customs in the western Mediterranean see bierbrauer 2003. This paper is the shortened version of my presentation at Cimitile/Santa Maria Capua Vetere which can only illustrate the main lines of argumentation. 3 von rummel 2007, especially pp. 386-400. 214 CHRISTOPH EGER Fig. 1. Map of the western Mediterranean. 1 Pax Julia/ Beja, Portugal;2 Thuburbo Maius/Henchir Kasbat, Tunisia; 3 Capraria/Capraia, Italy/Henchir Kasbat, Tunisia. tioned these analyses. It seems necessary to view inds and features in their wider contexts and to take into account the contemporaneous material culture and rituals of the different regions of the Mediterranean and its northern periphery. This will result in a more precise notion of terms relevant for the discussion, such as old, new, innovative, foreign, native, barbarian, Germanic, steppe-nomadic, and Roman or Byzantine. My paper will concentrate on three male graves playing a central role in the discussion on the interpretation of such Mediterranean graves of the Early Migration Period (ig. 1): the warrior grave of Pax Julia/Beja in southern Portugal (ig. 2), the warrior grave of Porto Capraia on the island of Capraria/Capraia in the northern Tyrrhenian Sea (ig. 3), and the grave of Arifridos at Thuburbo Maius/Henchir Kasbat, northern Tunisia (ig. 4)4. 2. Barbarian elite or Roman military aristocracy? On the basis of the personal equipment and the weapons, the three graves can be dated to the 2nd and 3rd quarter of the 5th century AD. The time slot for all three inds is the early phase of the Migration Period, and the notion that all three buried persons had “immigrated” in the context of historically attested migrations of human groups controlled by barbarians, has indeed dominated the interpretation of these graves for quite some time. The warrior from Beja has been considered a Vandal or a Visigoth5. Arifridos has been called a Vandal nobleman by researchers whose barbarian origin 4 The three graves have been analysed exemplarily by von rummel 2007, pp. 337-353. A closer description of the three graves is dispensable here thanks to the following contributions: Beja cf. König 1981, pp. 346-352, pl. 52; von rummel 2007, pp. 342-343; Palma santos 2008. Thuburbo Maius: Poinssot 1921; Poinssot-lantier 1934; König 1981, pp. 311-312 ig. 6d; von rummel 2007, pp. 337-338; aillagon (ed.) 2008, pp. 334-336 (T. Ghalia); on the location within the church cf. König 1981, p. 332 ig. 11; Das Königreich der Vandalen, p. 363 n. 306 (Ch. Eger). Capraia: DuCCi-CiamPoltrini 1991, pp. 53-59; DuCCi-CiamPoltrini-beDini 1992, pp. 369-376. 5 Cf. von rummel 2007, pp. 344-345 with an overview of earlier suggestions; Palma santos 2008, p. 364: Visigothic; Das Königreich der Vandalen, p. 145, n. 113 (A. Wenzel): grave attributed to early East Germanic horizon, connection with Vandals, Alans, and Suebi not to be excluded. HABITUS MILITARIS OR HABITUS BARBARUS? 215 additionally seemed conirmed by the mention of his Germanic name in the mosaic inscription6. For the warrior from Capraia, a Frankish or Alemannic ancestry has been deliberated by the excavators. It was thought possible that he had been a member of Avitus’s army that had embarked from southern Gaul to Rome in AD 455 and mainly included mercenaries of West Germanic origin7. However, the long favoured “ethnic”, respectively “migrationist” model of interpretation has increasingly been challenged in the past ifteen years. On the one hand, this criticism was determined by fundamental considerations of: a) the nature of ethnicity; b) the methods of ethnic interpretation in archaeology; c) the historical background. On the other hand it was also governed by a re-interpretation of archaeological remains and their cultural ‘localization’8. It is not the aim of the present contribution to deal with criticism of ethnic interpretation in general or its methodical and historical arguments, although this seems more Fig. 2. Pax Julia/Beja, Portugal. Inventory of the warrior necessary than ever, because some grave. 1 Scale ca. 1:5; 1a Scale 1:2,5; 1b Scale 1:2; 2–4 Scale points of criticism seem too undif2:3. ferentiated or even based on misunderstandings. Thus, ethnic interpretation is often equalised, by both critics and supporters, with the quest for ethnic identity. In fact, it is more an “external” description of typical features distinguishing 6 On earlier interpretations von rummel 2007, pp. 338-339; aillagon (ed.) 2008, pp. 334-336 (T. Ghalia): Vandal; Das Königreich der Vandalen, p. 363, n. 306 (Ch. Eger): Vandal. Doubts about the expressiveness of the name in merrills-miles 2010, pp. 86-87. 7 DuCCi-CiamPoltrini 1991, p. 59; DuCCi-CiamPoltrini-beDini 1992, p. 371; on this cf. von rummel 2007, p. 352. 8 With regard to German research exemplary reference is made to Brather’s studies: bratHer 2004; on the problem of dress and grave inds in particular: bratHer 2008. For critical statements also cf. the international conferences of the last years on this topic, e.g. PoHl-meHofer (eds.) 2010; Quirós Castillo 2011; ebanista-rotili (eds.) 2011. 216 CHRISTOPH EGER (groups of) people from their neighbours. There is an emphasis on cultural localization within the space-time-system but not identity as a subjective sense of “we”. This difference between “ethnographic” description and ethnic identity is underlined by the fact that the recognised borders often need to be drawn much wider and do not possess the desired clarity9. A closer approach seems possible whenever individual communities of known names move to geographically distant or culturally very different regions and are not fully acculturated or assimilated yet. But the core problem, with which I will deal below, is largely limited to aspects of archaeological interpretation of inds and features: Can cultural peculiarities and the distribution of certain archaeological phenomena of the 5th century AD be explained by a so-called Fig. 3. Capraria/Capraia, Italy. Inventory of the warrior grave. 1–3 Scale 2:3; 4 Scale ca. 1:4. ethnic approach and the migration of groups of people (gentes or mixed parts of different gentes), or do alternative concepts based on the paradigms of “spatial turn”, respectively “cultural turn” lead to a more convincing interpretation? - The sumptuous graves of the 5th century in the Mediterranean have intensively been dealt with by Ph. von Rummel10. In his opinion, the grave equipment does not relect barbarian foreign origin of the deceased, but increased desire for representation amongst the Late Roman military elite, whose dress accessories and weaponry he characterised as Roman in the widest sense11. 9 For the 5th century AD in particular, archaeologists are only able to separate major units from each other in a satisfactory way, e. g. the “(East Germanic) Danubian Culture” which includes quite a number of gentes. On this in detail bierbrauer 2008. 10 von rummel 2007, especially pp. 386-406. Whether exactly the present three graves can certainly be considered sumptuous (princely) graves is doubtful, however. On this problem see below. 11 Another, intermediate position between this interpretation and the ethnic one was taken by D. Quast in an inspiring article on Migration Period princely graves. He stressed that the burial ceremony with rich grave goods was most likely understood and promoted by barbarian military elites and their social HABITUS MILITARIS OR HABITUS BARBARUS? 217 Fig. 4. Thuburbo Maius/Henchir Kasbat, Tunisia, Inventory of the Arifridos grave. Scale 2:3. Thus, what emerges as a central topic is the cultural allocation of these elites on the basis of the archaeological material and their manners and rituals relected by it. Both can provide important hints as to the origin and composition of the elite. But what can be termed barbaric (Germanic, steppe-nomadic) and what Roman or Mediterranean in the 5th century AD? Although researchers have become very careful about a Barbarian or Germanic attribution meanwhile, I observe very frivolous handling of the terms Roman, Byzantine, and Mediterranean. Therefore, I will explicitly investigate the following questions: 1) Can all inds from the presented three male graves be characterised as commonly Roman and did they belong to the familiar equipment of Roman soldiers/ oficers of the 5th century AD? It hardly requires further explanation that no conclusive answer is possible to the second part of the question in particular12. 2) Are all graves with grave goods of the 5th century sumptuous burials of the military elite that concentrate in the western half of the Empire and thus relect a special historic constellation of the West Roman Empire in the 5th century? The methods for this investigation are well-known, but have by no means been exhausted for the present topic: What is necessary is a comparative analysis of archae- environment: Quast 2009, p. 378. 12 In order to obtain reliable results in this ield, a global study of the origin of recruits and equipment of the Late Roman army of the 5th century would be required. This would be very dificult, however, especially for the time after 430. No closer examination of this question in von rummel 2007 (short remarks: pp. 121, 155, 181, 230, 386, 392), although this would have been essential for his hypotheses. 218 CHRISTOPH EGER ological ind types as well as a dissection of the geographical distribution of inds and features, respectively the mechanisms of their distribution. More intensely than was done before, I will strive towards long-distance contextualisation of inds and features, including the contemporaneous material culture and customs in the different regions of the Mediterranean and its (northern) periphery. 3. The swords Apart from dress accessories, weapons deposited in graves play an important role for the cultural attribution and “ethnic interpretation” of buried persons. The spathae from both Beja and Capraia have been interpreted as foreign types originating in a different region of the Roman Empire or even outside the Empire. Typological characteristics favouring a foreign origin were reviewed and refuted by von Rummel. Thus, he arrived at the conclusion that both swords might ultimately be considered Mediterranean, respectively Roman weapons that do not allow statements on a foreign cultural habitus, not to speak of a foreign origin of the warrior. However, his explanations require amendment that will take us back to somewhere near the older position: The spatha from Beja (igg. 2,1a-b, 3) is characterised by its long, but comparatively narrow blade, a massive iron cross guard decorated with cloisonné, probably a solid pommel, and a magic sword pendant13. Migration period swords with a cross guard are considered eastern types and were subdivided by W. Menghin into a “Pontic type” (with cloisonné decoration) and an “Asiatic type” (without cloisonné decoration of the cross guard)14 with clearly different main distribution areas: On the one hand, examples with cloisonné are dominant in the north-eastern Black Sea Region and the Caucasus forelands. On the other hand, swords with an undecorated solid iron cross guard are almost exclusively found in the Carpathian basin15. However, M. Kazanski pointed out that at least the specimens with a cloisonné cross guard were liable to clearly Mediterranean inluence which he exempliied by the decorative elements of the cellular work and other features16. Based on this, von Rummel tried to prove that swords with a cloisonné cross guard were Roman weapons throughout and - as in Beja - “would not have stood out as foreign”17. As a seeming proof for the use of spathae with a broad cross guard in the Roman army, von Rummel quoted a silver bowl of the Valentinian period found near Geneva in the 18th century (ig. 5)18. The bowl, recently discussed in extenso by A. Arbeiter, features Emperor Valentinian II in 13 König 1981, p. 348, ig. 20; pl. 51; also cf. miKs 2007, p. 543, A43, pl. 143 (reproduction of an earlier drawing without the pommel reconstructed by König); sword also without pommel: aillagon (ed.) 2008, p. 365, ig. e. On the magic sword pendant for the irst time DannHeimer 1961, pp. 466-467, ig. 1,3a-c. 14 First and fundamentally Werner 1956, pp. 38-43; mengHin 1994-95, pp. 165-186; for the spatha from Beja: raDDatz 1959, pp. 145-146. On this also cf. von rummel 2007, p. 346; on its origin cf. anKe 1998, p. 75: independent development of the sword with elongated narrow blade and massive cross guard in the Eurasian steppes. Most recently miKs 2007, pp. 133, 197-198. 15 KazansKi 1996, p. 120, ig. 8; KazansKi 2001, p. 411, ig. 13; also cf. anKe 1998, pp. 216-217, maps 6-6a. 16 KazansKi 2001, pp. 408-409. 17 von rummel 2007, p. 350. 18 von rummel 2007, p. 349, ig. 50. HABITUS MILITARIS OR HABITUS BARBARUS? 219 Fig. 5. Silver bowl with representation of Valentinian II, found in the surroundings of Geneve, Suisse. Without scale (the black arrow marks the Emperor´s sword). parade armour amidst six bodyguards in its heavily worn relief19. In the lower part of the bowl, at the feet of the emperor and his guard, there are a shield, a helmet, and a sword sticking in its scabbard with a rectangular end and showing both a marked cross guard and moulded hilt with a pommel end. Formally, the sword can therefore be compared to the spatha from Beja. But according to all iconographic rules, the sword does not belong to the emperor´s personal equipment. The incoherent arrangement of the weapons almost “thrown down” at the feet of the emperor indicates that shield, sword, and helmet represent spolia from defeated enemies. In the case of the sword this is beyond any doubt whatsoever, because the emperor is wearing a sword 19 arbeiter 2008, pp. 56-58. 220 CHRISTOPH EGER at his left hip (although largely hidden). Unlike the sword on the ground, the pommel of this weapon consists of a small offset cone while a cross guard deserving of this name cannot be recognised. The contrast to the sword placed at the feet of Emperor Valentinian II is striking and obviously wanted: By the cross guard and the pyramidal pommel the metal worker, probably based at Milan20, exactly did not want to represent a standard weapon of the Imperial army, but a sword immediately recognisable for the contemporaneous observer as an antagonistic, barbarian weapon by a few characteristic features. The fact that the broad cross guards of the 5th century most probably had no lasting effect on the development of western Mediterranean edged weapons is indicated by their development in northern Gaul and south-west Germany: Sword types without a cross guard make quite clear that different types of edged weapons existed in the north-west of the Roman Empire and its periphery, and that types without a marked cross guard were preferred here. Although the cloisonné decoration on the short cross bar of the hilt edge and on the scabbard was taken over, the long drawn-out cross guard is absent from Frankish and Alemannic swords of the middle and 2nd half of the 5th century21. In my opinion this evidence is signiicant for the interpretation of ind lacunae in the western Mediterranean. If one assumes a lasting inluence of Roman weaponry in areas north of the Alps (and so does von Rummel, too)22, then swords without a cross guard should have been current in the western Mediterranean, too. Another argument against an eastern origin of the spatha from Beja is the existence of an iron pommel made likely by König’s reconstruction (ig. 2,1a)23, because amongst the eastern swords with a broad solid cross guard, none with a metal pommel is known, while western sword types usually possessed a marked pommel ending offset from the tang. J. Pinar and G. Ripoll rightly objected, however, that swords with or without a pommel confronted each other in the Middle Danube Area and that a mixture of both types might have occurred here at least24. Meanwhile the border between the two sword types with a massive pommel ending has shifted further east: A spatha from Kambulta in the northern Caucasus, kept at the Historical Museum in Moscow for some time now, but overlooked by researchers so far, also possesses a pommel25. Although it cannot be immediately compared to the swords discussed before as it lacks a cross guard, it urges for caution in the assessment of swords with or without a pommel. The small solid iron ending of the tang can no longer be taken for a reliable proof of western origin of an edged weapon. The spatha from Capraia (ig. 3,4) belongs to a completely different sword type than does the previously discussed weapon. G. Ciampoltrini already hinted at its close relation to a group of spathae in northern Gaul, which was termed type IIa (Samson- arbeiter 2008, p. 60. The cross guards with cloisonné of the spathae from Tournai and Flonheim grave 5 are hardly broader than the width of the blade. Only the cross guard of the spatha from Planig grave 1 is drawn out slightly more (cf. mengHin 1983, pp. 181, 215, 224). 22 von rummel 2007, p. 350. 23 KazansKi 2001, p. 411; von rummel 2007, pp. 349-350. 24 Pinar-riPoll 2008, pp. 118, 130, ig. 3-4. 25 The sword was bought long ago. For notice of this weapon I am greatly indebted to I. Akhmedov, Historical Museum Moscow. Further information was kindly provided by M. Kazanski, Paris. 20 21 HABITUS MILITARIS OR HABITUS BARBARUS? 221 Oberlörick) by W. Menghin and is identical with group A after K. Böhner and M. Martin26. Obligatory for this type are a double-edged blade some 5-5.5 cm wide and 80-85 cm long merging into a tang without a striking cross bar or a massive cross guard as well as special characteristics of the scabbard, namely a chape adorned with a mask, which is absent from the Capraia spatha, however, and a decorated locket of type Samson-Abingdon27. Amongst the lockets executed in a very similar way, closer attention is particularly deserved by the specimens from Krefeld-Gellep grave 43 on the Lower River Rhine, Germany, and Samson grave 11, Belgium. The majority of spathae of type Samson-Oberlörick come from graves in the Rhine-Meuse Area. Few specimens are also known from Kent, England, and south western Germany, meanwhile only two pieces were found south of the Alps, namely the spatha from Capraia and a fragment from Verona. Because of this distribution pattern K. Böhner already assumed a production in “indigenous Late Roman workshops of the Meuse Area” 28. Obviously we are dealing with a regional sword type that reached neighbouring regions only in limited numbers. This becomes particularly clear from swords in graves of south-east England and the Alamannia. Both regions had a pronounced custom of depositing weapons in warrior graves which, however, mainly contained other sword types of limited regional distribution29. The areas south of the Alps are dificult to consider due to the absence of weapons from graves, but it is likely that the evidence from south-east England and south-west Germany can be transferred. Therefore, the two pieces from Capraia and Verona can rather be regarded as isolated examples of a sword type produced and mainly used in the Rhine-Meuse Area, which were probably transported by warriors on their way to the south from northern Alpine regions. An alternative transmission as booty or present cannot be completely excluded either. We are not dealing, however, with a typical Roman weapon being used in the entire western half of the Empire. 4. The belt buckles Apart from a single-edged knife from Capraia, the other inds from the three male graves exclusively consist of dress accessories decorated with inlays. Special attention is deserved by the different buckles from all three graves which served for fastening waist belts, sword belts or shoe straps. For all represented types, the number of known parallels is large enough to allow statements on regional centres of distribution and thus on potential areas of origin of the items (and their wearers). The buckle with a kidney shaped itting with stone inlay from Beja (ig. 2,4) belongs to a large group of very similar gold buckles with a distribution focus in the 26 mengHin 1983, pp. 154-155; mengHin 1994-95, pp. 158-160; böHner 1987, pp. 411, 413-416; martin 1989, pp. 122, igg. 1, 124-125. On the spatha most recently miKs 2007, p. 556, A96. 27 miKs 2007, pp. 130-131: sword of type Illerup-Wyhl with a tendency towards type Wyhl; mengHin 1983, p. 138; böHner 1987, p. 413. 28 böHner 1987, p. 411; similarly mengHin 1983, p. 154: «in workshops in Late Roman tradition». 29 On this cf. the distribution maps in mengHin 1983, pp. 163-164, igg. 92-93; böHner 1987, pp. 450-451, igg. 15-16; mengHin 1994-95, p. 162, ig. 17. 222 CHRISTOPH EGER Middle Danubian Region and on the Black Sea Coast30. Characteristic of its most numerous variant are a massive golden frame of circular thickened form and a round looped belt-plate with cellular work and marginal rivet sockets. The piece from Beja deviates from this by its frame with inlay decoration, kidney-shaped itting and mounted cabochon, the last feature being considered typologically as a younger trait of these buckles of the Danubian phase D2 (AD 400/10 to 440/50). In southern Central Europe and in south western Europe this typological group is represented only by few examples. According to R. Stark, the production of the buckles involved Late Antique (East) Roman workshops31. This is not only revealed by the cellular decoration and some cell patterns, but also by the technical solution for ixing the itting: Placing rivet pins in sockets along the external edge of the plate, connects the round gold buckles to some other buckle types mostly made of bronze and predominantly found in the eastern half of the Empire32. According to graves, small precious round buckles were worn by leading barbarian warriors in the service of Rome or - in the later 2nd quarter of the 5th century - of Attila. There is disagreement amongst researchers about the question, whether such gold buckles mainly represented a particularly ostentatious temporary fashion of the Danubian military aristocracy or whether they were distributed all over the Mediterranean and belonged to the standard equipment of high-ranking oficers of the (East and West) Roman army. It may be objected, however, that general use as a Roman military buckle can neither be proven by the overall distribution of this type, not including the entire Mediterranean or all Imperial provinces, nor by taking into account the remaining contemporaneous military belts of the Roman Empire. In the north-western provinces of the Roman Empire e. g. we encounter a completely different belt fashion of the late 4th and 1st half of the 5th century AD. The ind material is well known, particularly between the River Loire and the lower River Rhine thanks to extensive studies by H.-W. Böhme33. Independently from the much debated question of the ethnic interpretation of northern Gaulish and Belgic graves34 it must be stressed that army equipment in this part of the Roman world included a belt with multipart sets of metal ittings consisting of extremely wide belt plates with chip carved and punched decoration (ig. 6). The relatively small, but massive circular buckles with a looped belt-plate form a marked contrast to these chip carved sets of ittings with regard to their very size and shape and represent a totally different belt fashion. Signiicantly enough they are rarely found in this area and, when present at all, they were exclusively used as additional inferior buckles35. Above all, there are nearly no sumptuous buckles of precious metals with a circular cloisonné plate in Gaul, the scattered specimens of which do not cross an im- sCHmauDer 2002 II, p. 121, map 14; bierbrauer 2008, p. 40, ig. 40. starK 2000, p. 194; starK 2004, pp. 30-31. On this buckle type also teJral 1987; KazansKi 1996, p. 123; KazansKi 1999, pp. 304-307. 32 On buckle plates with rivet sockets along the edges cf. sCHulze-Dörrlamm 2002, pp. 84-86; starK 2000, p. 201; starK 2004, p. 28, ig. 4,1-2.5-6.9-14.19-20. 33 böHme 1974, pp. 195-207; for the type spectrum cf. böHme 1974 pp. 53-81 text pl. A-B. Most recently on chip carved sets of ittings: böHme 2008. 34 Cf. the contrary positions of böHme 2008 and Halsall 2007, pp. 152-162. 35 Thus e.g. in a grave from Vieuxville: böHme 1974, pl. 110,14; sommer 1984, pl. 59,5. 30 31 HABITUS MILITARIS OR HABITUS BARBARUS? 223 aginary line from the Rhine-Main Area to the Lower Seine River in a northerly direction36. Another zone within the West Roman Empire for which a distinctive belt fashion of the 4th to early 5th century is attested, is the northern Meseta of the Iberian Peninsula. Its comparatively good state of source material as to belt accessories is owed to the fact that it possesses small cemeteries and grave groups with grave goods of the so-called Douro Valley Culture37. The type spectrum of these belt accessories mainly involves regional types such as buckles of type Simancas which are just as markedly distinct from eastern buckle fashions of the late 4th to mid 5th century as from chip carved sets of the north-western provinces (ig. 9 bottom)38. Another belt fashion deviant from Middle Danubian and eastern Imperial norms suggests itself Fig. 6. Top: Distribution of belt ittings with Kerbschnitt. for the North African provinces in – Bottom: Parts of a belt set with Kerbschnitt from BaselAeschenvorstadt, grave from 1971, Suisse. the time around AD 400 and the 1st half of the 5th century, although the present state of research and publication is very scarce39. Of particular signiicance for the question of North African sumptuous belts of the Latest Roman Period are the belt ittings from the treasure of Cartennae/Ténès most likely deposited around AD 420/3040. Let us conclude by considering the overall development of Late Roman belt fashions and particularly the one of the Iberian Peninsula with regional types surviving into the 5th century: It seems hardly believable that the Beja warrior was a regular Cf. the examples quoted by KazansKi 1996, p. 122, ig. 9, nn. 28-29.48. For a summary see zeiss 1934, pp. 90-91 («Early Castilian Group»); Caballero zoreDa 1974; fuentes Domínguez 1989. On belt accessories cf. aurreCoeCHea fernÁnDez 2001. 38 On Hispanic belt types of the 4th/5th centuries see Pérez roDríguez-aragón 1992, pp. 258-260, ig. 3-5; aurreCoeCHea fernÁnDez 2001, p. 27, ig. 9; on the dating of type Simancas into the late 4th and (advanced?) 5th century aurreCoeCHea fernÁnDez 2001, p. 158. 39 On the state of source material see maCKensen 2008. 40 Monographic: Heurgon 1958. I do not support the very late dating of its deposition by Heurgon (Heurgon 1958, pp. 71-73, 77) on the basis of the large medallion disc brooch with cross pendilia. Cf. eger 2012, pp. 158-159. 36 37 224 CHRISTOPH EGER Late Roman oficer buried with current accessories spread across the entire empire. His equipment neither implies a widely distributed or even uniform fashion amongst military aristocrats present in Hispania, nor a broader regional production41. Because of his belt ittings and his sword, the soldier from Beja can hardly be related to military contingents other than barbaric or barbarised units operating here in the 2nd quarter of the 5th century, be it under Suebian or Visigothic-Roman command42. In view of its (foreign) origin, the - presumptive - belt buckle from the Grave of Arifridos must be interpreted similarly to the circular buckles discussed before. The now lost specimen has most recently been classiied amongst “oval buckles with an oval to kidney-shaped plate with a single or bipartite inlay of type C2” by M. Schulze-Dörrlamm43. Parallels are mainly found in the Black Sea and Middle Danubian Areas, only few pieces were found further west, respectively in the western Mediterranean. M. Schulze-Dörrlamm and Ph. von Rummel supposed that this limited distribution did not (only) relect a problem of the present state of source material. Their hypothesis can be disproven by a comparative analysis of the distribution of different cloisonné buckles: It reveals clear differences that refute a global explanation exclusively resting on the bad state of research in the Mediterranean, for the limited distribution of the early types of the 1st half to mid 5th century. These buckles must also be considered a warning against a rashly generalising terminology when classifying buckles as Roman/Byzantine or circum-Mediterranean. In this respect the larger of the two buckles from the warrior grave at Capraia demands careful examination, too. With its oval ring and upright oblong looped beltplate with a drop-shaped “plate-inlay” decoration the larger buckle belongs to the type Komorn-Gültlingen-Bingen after Quast or type C14 after Schulze-Dörrlamm, the dating of which reaches from the mid 5th century to the time around AD 50044. Several parallels are known in southern Germany, but similar buckles are also known in the eastern Mediterranean (ig. 7). As Schulze-Dörrlamm pointed out, the Merovingian examples with their ittings mostly made of iron and with four corner rivets are distinct from Byzantine buckles with a lat box itting of non-ferrous metal and three rivets and are likely to have been locally made imitations45. The piece from Capraia must be counted amongst the (Merovingian) imitation group and thus implies northern Alpine connections as does the sword from the same grave. 5. Special dress? The shoe buckles The dress accessories of the three graves hardly allow any statements on the garments formerly worn with them. The basic element might have been a tunic with a 41 This does not exclude a production of very small numbers only for the demand of warrior elites of the gentes that lived on the Iberian Peninsula from AD 409 onwards. 42 For the changeful history and the numerous military campaigns of this period cf. garCía moreno 1998, pp. 49-72; KamPers 2008, pp. 126-129. 43 sCHulze-Dörrlamm 2002, pp. 86-89. 44 Quast 1993, p. 86; sCHulze-Dörrlamm 2002, pp. 120-123. 45 Distinction of eastern Mediterranean and western types by sCHulze-Dörrlamm 2002, p. 121. HABITUS MILITARIS OR HABITUS BARBARUS? 225 Fig. 7. Distribution of belt buckles with rectangular plate and plate-inlaying-decoration, type KomornGültlingen-Bingen/C14. belt around the hip. Only Arifridos possessed a ibula (ig. 3,1) for fastening a cloak, probably a chlamys closed on the right shoulder according to Late Roman custom. Another striking peculiarity of this grave ind is the golden pair of miniature buckles that might have served as shoe or garter buckles. They belong to the miniature buckles with a D-shaped plate and individually mounted cabochon46. Their paring and minute size of only some 2.5 cm support their use as precious shoe ittings, although their ind position in the grave remains unknown. The supposition, that such sumptuous pairs of small buckles of the 1st half of the 5th century and mid 5th century served as shoe fasteners, rests on few reliably documented inds mainly from the Middle Danubian Area, where such small buckles were encountered in situ at the feet of the dead47. Thus, the archaeological inds imply a barbarian rite. Von Rummel, in contrast, believed that the use of buckles on shoes or stockings was once also and particularly known in 46 The type has not been included by sCHulze-Dörrlamm 2002. She only considers D-shaped buckle plates with a bezel setting (buckles of type C7). KazansKi 1994, pp. 144 (group I.2.D), 178, ig. 5,9-10 attracted attention to two isolated miniature buckles from Paris and Kerč. 47 Cf. e.g. sCHmauDer 2002, II, pp. 21-27, ig. 5 (Blučina); pp. 35-37, ig. 9 (Lébény; only a small buckle at the left foot survives). Earlier specimens still dating to the 4th century are known from the Černjachov Culture: sCHmauDer 2002, I, p. 159; bierbrauer 2008, p. 42 note 145. 226 CHRISTOPH EGER the Roman Empire and is suficiently documented by written sources and pictures48. He added that Roman Emperors wore shoes richly decorated with jewels from the 3rd century onwards. However, buckles are neither mentioned in written sources nor can they be recognised in pictures. Von Rummel continued that such an accuracy of detail could not be expected in representations anyway, the missing iconographical proof therefore being of no consequence for the matter49. I disagree with this, since the oficial image program of Roman emperors deinitely included a high esteem for detailed representation of pieces of garment despite all stylisation and idealisation. Unless there is demonstrable regress to older traditional picture schemes, we may expect representations closely committed to real Imperial costumes. This is made quite clear by the famous sculpture of “The Tetrarchs” of around AD 300 in Venice representing the Augusti and Caesares in similar military service clothing (ig. 8 top): Emphasis is laid on the cloak ibula, the belt decorated with different jewel-clad plates and the elaborately adorned sword50. Even the shoes were by no means neglected in the image: Next to the different straps that ix the campagi there are round discs signifying decorative ittings. The lacing, however, could do completely without buckles which, given the pronounced representation of one of the shoelace endings, cannot be explained by deicient attention to details. In the picture on the richly decorated Missorium from Almendralejo, Theodosius the Great is wearing a grand courtly dress, which the silver smith represented with great attention to detail. Here, too, the shoes do not possess buckles (ig. 8 centre)51. At the upper bootleg there is only a circular disc that can be interpreted as a decoration or perhaps as a kind of a button, but hardly as a buckle. An important clue with regard to this is an exceptionally well preserved shoe of leather dyed purple from R. Forrer’s excavations in the Coptic cemetery of Panopolis/Akhmim in Egypt (ig. 8 bottom)52. Typologically the Egyptian shoe is very close to the pair worn by the emperor on the “Missorium of Theodosius”. On the instep there is a circular ornament of gold colour or sheet gold reminiscent of the disc-shaped feature in the picture. Thus one need not necessarily to think of precious jewel applications or even buckles, but rather of coloured decorations or sewed-on appliques. According to Frauberger’s publication neither this specimen nor the other shoes and boots found at Panopolis had shoe buckles. The 5th-century miniature buckles that most likely served as luxury shoe ittings (shoe buckles) have typological relations to the Danube and Black Sea Regions. The fashion of shoe buckles itself is likely to be of eastern origin (eastern Black Sea Region or Sassanid Empire). As far as it is possible to assert by means of archaeological sources, shoe buckles never were a general Roman phenomenon, but were mainly current amongst the warrior elite of the Danubian Area, respectively amongst the frontier society under Danubian inluence, before they were temporarily adopted by barbarian elites of the west in the late 5th century, too. The almost complete absence of archaeological rummel 2007, pp. 118, 341. rummel 2007, p. 341. DelbrueCK 1932, p. 88, ig. 33; ragona 1963. Monographic: almagro et alii (ed.) 2000. frauberger 1895-96. 48 von 49 von 50 51 52 HABITUS MILITARIS OR HABITUS BARBARUS? 227 Fig. 8. Top: Venice-San Marco, Stone sculpture of the tetrarchs, detail: shoes of one of the tetrarchs. – Mid: Missorium of Theodosius, Almendralejo, Spain, details with shoes of the emperor and a dignitary. – Bottom: Panopolis/Akhmim, Egypt; red coloured shoe of leather, with golden ornaments. 228 CHRISTOPH EGER proof of shoe buckles in the Mediterranean cannot be explained across-the-board by the absence of grave goods: I already mentioned different regions with ind material, also including graves of the relevant time, in the context of my discussion of individual buckle types. Amongst them is the Near East, but also certain zones of the Iberian Peninsula. In none of them do we get any evidence for the use of shoe buckles. 6. Towards an interpretation of burial customs and grave goods In the discussion on the ethnic appraisal of the three male graves it has often been argued that burial with dress accessories and a weapon was typically barbarian, while the opponents of a primarily ethnic interpretation explained this by a changed desire for representation amongst the Late Roman military elite of the western Empire. Therefore we will subsequently have to ask whether weapon graves of the Late Roman world were per se a foreign feature and thus a hint at the presence of barbarians or whether they may be considered proof of the quest for new means of expression of the military elite. Secondly we will have to examine whether these tombs were always very richly equipped and connected to the social elite. Thirdly the quantity and geographical distribution of these graves is signiicant: Do they come from regions demonstrably connected to treks of Migration Period gentes or can they be found wherever Late Roman armed forces were garrisoned? 6.1. Some remarks on weapons deposited in graves in the Late Roman Empire Weapons were deinitely deposited in graves in the Late Roman Empire, but their occurrence is limited to certain regions. In the Mediterranean one of these regions is - northern Gaul and the Germanic provinces not being considered here53 - the northern Meseta in Spain with its 4th- and 5th-century grave groups and cemeteries of the Douro Valley as has been said before54. While the majority of graves had no grave goods, some tombs contained a vessel, some jewellery or dress accessories (almost exclusively buckles) and also components of arms in male graves. Particularly current was the deposition of a dagger of type Simancas or a lance head (ig. 9)55. The weapon deposits imply that a warlike rural population, possibly members of a militia set up for the protection of latifundia, was buried here. Another region with Late Antique weapon graves is the Near East. Here, however, only isolated evidence has been found yet which underlines the exceptional character of weapon deposits in graves of the Palestinian and Arabian provinces. As an example I quote the grave of an archer at Be’er Sheva’ in Israel who was buried with a set of tanged arrow heads in the 2nd half of the 4th century56. It is not unlikely that he was a member of an auxiliary unit of Saracen archers also mentioned in the written sources. These two examples should sufice for making two points: 53 54 55 56 On this cf. the explications by böHme 1974; most recently böHme 2008; contrariwise: Halsall 2007. Fundamental literature Halsall 2007 note 82. For customs of depositing grave goods cf. the tables by fuentes Domínguez 1989, pp. 123-147. gorin 2003, p. 89, ig. 132. HABITUS MILITARIS OR HABITUS BARBARUS? 229 Fig. 9. San Miguel de Arroyo, Spain, grave 10. Scale ca. 1:2; lance head Scale 2:5. – Bottom: Belt buckle of type Simancas. 1) Late Antique weapon deposits in graves need not be connected to Barbarians a priori; one rather has to analyse the context of the weapon graves and the ind material very carefully. The northern Castilian cemeteries of the so-called Douro Culture were obviously a regional phenomenon of a native rural population. The still rare examples in the Near East, however, elude closer evaluation. Here it seems likely in 230 CHRISTOPH EGER the context of written traditions (Notitia Dignitatum) that weapon deposits in graves were a custom of barbarian members of the Roman army, in this case “Saracens”57. 2) Late Antique graves containing a sword are known neither in the Roman provinces of the Near East nor in the Douro Culture. The deposit of spathae in Beja and Capraia is a new custom in their surroundings which must be considered in a differentiated way as to its sociological and cultural signiicance (origin of sword types, closest contemporaneous parallels for the custom of swords as grave goods). As has been demonstrated before, the next parallels for both the sword types and the custom of depositing swords in graves imply different barbarian areas of origin for both cases. 6.2. Elite burials and princely graves? On the social position of the deceased According to von Rummel, the male graves from Beja, Capraia, and Thuburbo Maius together with sumptuous burials such as Pouan in France and Tournai (Childeric’s Grave) in Belgium and a number of magniicent female graves were proof of a new burial custom of the Late Roman military aristocracy58. However, this argumentation puts very different graves on the same sociological level. The three Mediterranean graves have clearly restricted furnishings with regard to quantity and in Capraia also to quality (only gold-plated buckles) in comparison to the sumptuous and extremely substantial equipment of the Frankish king Childeric (died 482)59 and to the one of the nobleman buried at Pouan in the 3rd quarter of the 5th century60. Moreover, Arifridos was buried without a weapon. This does not only imply considerable differences of a social nature and of rank, but also implies the doubt: to what extent Arifridos can be considered a typical representative of the army61. This, however, shakes one of von Rummel’s core hypotheses, namely that the “inhumations habillée” of the 5th century were generally richly furnished graves or even sumptuous ones, the appearance and equipment of which should be explained by Kossack’s theory of princely burials62. If it is already permissible to doubt the classiication as sumptuous graves in the case of the three male graves, there is even more evidence for plainer equipment amongst female graves of Vandal North Africa. Mention must be made of two graves excavated at Hippo Regius/Annaba, Algeria, in the mid 19th century, the inventories of which range well below the level of graves with gold jewellery and dress accessories such as Thuburbo Maius and Carthage-Koudiat Zâteur63. Isolated inds of crossbow and bow ibulae of non-fer- 57 This supposition would require further investigations of burial customs in the Limes apron, respectively on the Arabic Peninsula. 58 von rummel 2007, pp. 375, 384, 386. 59 Good overview: Périn-KazansKi 1996; KoCH-von WelCK-WieCzoreK 1996. 60 Most recently Ph. Riffaud-Longuespé in aillagon (ed.) 2008, pp. 322-323. 61 Von Rummel’s (von rummel 2007, p. 404) suggestion that persons buried with a sword should be considered members of the army, those without a sword members of the militia non armata, i. e. of the Imperial administration, is unsatisfactory because this would reduce the deposition of a sword from the very beginning to a merely functional aspect. 62 von rummel 2007, pp. 9, 377, 382-383 with a very problematic treatment of the term «sumptuous grave». View in contrast the criteria listed by KossaCK 1974, pp. 4-5. 63 On the inventory of graves 1-2/1865 from Hippo Regius/Annaba: Quast 2005, pp. 242-247; on Carthage-Koudiat Zâteur: eger 2001, pp. 353-370. HABITUS MILITARIS OR HABITUS BARBARUS? 231 Fig. 10. Belt set from “Reastan” (probably ar-Rastan), Syria. Scale 2:3. rous metals, that might come from undocumented graves because of their good state of preservation, imply the existence of further graves of persons who cannot be counted amongst the top level of Vandal society and probably not even Vandal upper class64. 6.3. A phenomenon of the West Roman Empire? Another objection refers to von Rummel’s hypothesis that graves with more or less elaborate jewellery and dress accessories were a phenomenon of the West Roman Empire in particular, unparalleled in the East Roman Empire and connected to the special political constellation of the western Empire65. This theory is plainly wrong. Graves with grave goods of Late Roman and Byzantine Times also existed on the north-eastern periphery of the East Roman Empire (Crimea, foothills of the Caucasus) and in the south-eastern parts of the Empire: both in Egypt and in the Near East the deceased continued to be buried wearing garments, jewellery and sometimes also dress accessories from Roman times until the 6th and 7th century. It is true that, up to date, there are no graves as rich in gold as the graves from Pouan and Tournai, but 64 65 Cf. eger 2008, pp. 192-193, ig. 1,4; 2,6; eger 2012, pp. 321-322. Cf. von rummel 2007, pp. 384-386, 403. 232 CHRISTOPH EGER the level of equipment of graves such as the one of the warrior from Capraia (dress accessories of bronze and gold-plated bronze with cloisonné decoration) is reached (ig. 10)66. Weapons were deposited rarely, however. Last but not least, the uneven geographical distribution of “elite graves” within the West Roman half of the Empire speaks against a connection with the Late Roman military elite: Thus it occurred to von Rummel, too, that rich graves with the aforementioned elements massively concentrate on the north-eastern periphery, in the Carpathian Basin, and in Pannonia, while they rapidly thin out by going away towards the west and south-west and they are only sporadically found there. But at least Gaul and Upper Italy, where major parts of the comitatenses and the operational headquarters of the Roman army were garrisoned in the 1st half of the 5th century, should have produced numerous graves according to von Rummel’s hypothesis. This is not the case, however67. 7. Conclusion If we sum up our previous thoughts we cannot but develop a differentiated notion of “Barbarian”, “Roman”, and “Mediterranean” for the interpretation of individual inds and the custom of depositing grave goods in rich male graves of the 1st half till the middle of the 5th century. Our most important inding with regard to the objects from the graves at Beja, Thuburbo Maius, and Capraia is the proof of a regionally differentiated material culture even within the Roman Empire, respectively the Mediterranean Area, which permits statements regarding the mobility of material goods and persons. Depending on their particular focus of distribution, the distribution patterns of certain products such as dress accessories and weapons seem to be explicable rather by personal mobility than with general diffusion, fashion, trade, or gift exchange. In this connection, we must not disregard the historically attested mobility of Barbarians, though very heterogeneous groups of people in the Mediterranean since the late 4th century. Peculiarities of dress can also be explained much more convincingly by members of such communities than by a (uniform) new apparel of the Late Roman military aristocracy of the western Empire, as was demonstrated by the example of shoe buckles. The habit of burying a deceased warrior with his sword is unparalleled in the Late Roman Mediterranean, although it was deinitely common in different regions of both the western and eastern part of the Roman Empire to inhume dead persons of certain social standing with their personal equipment and additional grave goods, e. g. in Meseta/Spain, Egypt and the Near East. Instead of ascribing the warrior graves under discussion to an uniform West Roman military aristocracy, I would like to suggest - quasi as an alternative draft - a modiied “ethnic” interpretation: The buried persons’ cultural habitus documented by their equipment, armament, and burial customs is something new and strange within Cf. the sets of ittings with cloisonné decoration from ar-Rastan («Reastan») near Homs: Quast 1999. It is true that the number of graves increases, as can be seen e.g. from graves at Sacca di Goito, Northern Italy, discovered some years ago (sannazaro 2006). But this does not compensate the existing imbalance. 66 67 HABITUS MILITARIS OR HABITUS BARBARUS? 233 5th-century western Mediterranean, but it displays certain references to different regions on the northern periphery of the Empire. For this reason the deceased of the three model graves - Beja, Capraia, and Thuburbo Maius - are considered non-locals as to their place of origin. While the habitus and individual ind types of the Beja warrior and the nobleman Arifridos from Thuburbo Maius possess afinities with the Danubian Culture and might come from the Middle Danubian too, the evidence of the Capraia warrior rather implies an origin in a Northern Alpine, perhaps the East Frankish Area. However, their burial mode, dress, and armament do not relect their ethnicity, but a cultural imprint that allows more or less a precise geographical localisation depending on the available criteria and the particular state of research. Whether the deceased originally came from the relevant region or only obtained an indirect imprint from there, cannot be precisely determined by archaeological means. Thus, we must face the possibility that a certain habitus could be adopted by persons joining a barbarian community (gens) at a later point in their lives. In the case of migrations, this presupposes a certain prestige of the particular culture and facilities for the continuous production of “foreign” objects typical of the outit in new, distant areas68. Why a particular burial mode in complete apparel with jewellery and accessories was practised in the new settlement area cannot be extensively discussed here. One of the reasons certainly was the crisis of legitimation and identity of the leading barbarian classes, a fact which von Rummel has explicitly expressed by the example of the Vandals, too69. But in contrast to his further argumentation my accent will be considerably shifted back towards the barbarian, tribal element: The grave inds must not be seen connected to members of the Late Roman military aristocracy of the western half of the Empire in general, but to certain clansmen of barbarian gentes. From the discussion of ind types and customs it becomes quite obvious that this approach is not at all based on a strict dichotomy Germanic/barbarian vs. Roman as a background image and characterising the interpretation: Sometimes it is only nuances and regional shifts that allow a differentiation. abbreviations anD bibliograPHy aillagon J.J. (ed.) 2008, Rome and the Barbarians. The birth of a new world, Venice. almagro gorbea m. et alii (ed.) 2000, El disco de Teodosio. Publicaciones del Gabinete de Antigüedades de la Real Academia de la Historia. Estudios 5, Madrid. anKe B. 1998, Studien zur reiternomadischen Kultur des 4. bis 5. Jahrhunderts 1-2, Weissbach. arbeiter A. 2008, Der Kaiser mit dem Christogramm-Nimbus. Zur silbernen Largitionsschale Valentinians in Genf, in «Helvetia Archaeologica», 39/154, pp. 42-73. aurreCoeCHea fernÁnDez J. 2001, Los cinturones romanos en la Hispania del Bajo Imperio. Monographies Instrumentum 19, Montagnac. bernDt g.m.-steinaCHer R. (eds.) 2008, Das Reich der Vandalen und seine (Vor-) Geschichten. Forsch. Gesch. Mittelalter 13, Wien. 68 Little is known, however, about the supply of mobile groups of people with consumer goods, accessories, and jewellery. 69 von rummel 2007, p. 385. It would certainly be mistaken to reduce the burial customs to this very aspect. Religion and concepts of moral and tradition are also likely to have played their role. These aspects need not exclude each other. 234 CHRISTOPH EGER bierbrauer V. 2003, RGA² 25, Berlin, New York, pp. 210-242, s.v. Romanen. bierbrauer V. 2008, Ethnos und Mobilität im 5. Jahrhundert aus archäologischer Sicht: Vom Kaukasus bis nach Niederösterreich. Bayerische Akademie der Wissenschaften, Philosophisch-historische Klasse, Abhandlungen Neue Folge 131, München. böHme H.W. 1974, Germanische Grabfunde des 4. bis 5. Jahrhunderts zwischen unterer Elbe und Loire. Münchner Beiträge zur Vor- und Frühgeschichte 19, München. böHme H.W. 1986, Bemerkungen zum spätrömischen Militärstil, in rotH H. (ed.) 1986, Zum Problem der Deutung frühmittelalterlicher Bildinhalte. Akten des 1. Internationalen Kolloquiums in Marburg a. d. Lahn, 15. bis 19. Februar 1983, Sigmaringen, pp. 25-49. böHme H.W. 2008, Gallische Höhensiedlungen und germanische Söldner im 4./5. Jahrhundert, in steuer H.-bierbrauer v. (eds.) 2008, Höhensiedlungen zwischen Antike und Mittelalter von den Ardennen bis zur Adria. Ergänzungsband Reallexikon der Germanischen Altertumskunde 58, Berlin, New York, pp. 71-103. böHner K. 1987, Germanische Schwerter des 5./6. Jahrhunderts, in «Jahrbuch des RömischGermanischen Zentralmuseums Mainz», 34/2, pp. 411-490. bratHer S. 2004, Ethnische Interpretationen in der frühgeschichtlichen Archäologie. Geschichte, Grundlagen, Alternativen. Ergänzungsband Reallexikon der Germanischen Altertumskunde 42, Berlin, New York. bratHer S. 2008, Kleidung, Bestattung, Ritual. Die Präsentation sozialer Rollen im frühen Mittelalter, in bratHer S. (ed.) 2008, Zwischen Spätantike und Frühmittelalter. Archäologie des 4. bis 7. Jahrhunderts im Westen. Ergänzungsband Reallexikon der Germanischen Altertumskunde 57, Berlin, New York pp. 237-273. Caballero zoreDa L. 1974, La necrópolis de Fuentespreadas (Zamora). Excavaciones Arqueológicas en España 80, Madrid. DannHeimer H. 1961, Zum Germanengrab von Beja - Pax Julia, in «Germania», 39, pp. 466-467. Das Königreich der Vandalen = Das Königreich der Vandalen. Erben des Imperiums in Nordafrika, ed. Badisches Landesmuseum Karlsruhe, Mainz. DelbrueCK R. 1932, Antike Porphyrwerke, Berlin. Die Franken = Die Franken Wegbereiter Europas. vor 1500 Jahren: König Chlodwig und seine Erben 1-2, Mainz 1996. DuCCi s.-CiamPoltrini G. 1991, Capraia (Livorno). Tomba di un militare tardoantico. Materiali di corredo tardoantico, in «Bollettino di Archeologia», 7, pp. 53-59. DuCCi s.-CiamPoltrini g.-beDini e. 1992, Una sepoltura tardoantica dal Porto di Capraia Isola, in «Archeologia Medievale», 19, pp. 369-377. ebanista C.-rotili m. (eds.) 2011, Archeologia e storia delle migrazioni. Europa, Italia, Mediterraneo fra tarda eta romana e alto medioevo. Atti del Convegno internazionale di studi Cimitile-Santa Maria Capua Vetere, 17-18 giugno 2010, Cimitile. eger Ch. 2001, Vandalische Grabfunde aus Karthago, in «Germania», 79/2, pp. 347-390. eger Ch. 2008, Vandalisches Trachtzubehör? Zu Herkunft, Verbreitung und Kontext ausgewählter Fibeltypen aus Nordafrika, in bernDt-steinaCHer (eds.) 2008, pp. 183-195. eger Ch. 2012, Spätantikes Kleidungszubehör aus Nordafrika I. Trägerkreis, Mobilität und Ethnos im Spiegel der Funde der spätesten römischen Kaiserzeit und der vandalischen Zeit. Münchner Beiträge zur Provinzialrömischen Archäologie 5, Wiesbaden. frauberger H. 1895-96, Antike und frühmittelalterliche Fussbekleidungen aus Achmim-Panopolis, Düsseldorf. fuentes Domínguez A. 1989, La necrópolis tardorromana de Albalate de Las Nogueras (Cuenca) y el problema de las denominadas Necrópolis del Duero, Cuenca. garCía moreno L.A. 1998, Historia de España Visigoda, Madrid. gorin Y. 2003, Be’er Sheva’, in «Hadashot Arkheologiyot», 115 (Hebrew section 90-91), pp. 65-66. Halsall G. 2007, Barbarian migrations and the Roman West 376-568, Cambridge. Heurgon J. 1958, Le trésor de Ténès, Paris. HABITUS MILITARIS OR HABITUS BARBARUS? 235 KamPers G. 2008, Geschichte der Westgoten, Paderborn. KazansKi M. 1994, Les plaques-boucles méditerranéennes des Ve-Vie siècles, in «Archéologie Médiévale», 24, pp. 137-173. KazansKi M. 1996, Les tombes «princières» de l´horizon Untersiebenbrunn, le problème de l´identiication ethnique, in L´identité des populations archéologiques. XVIe Rencontres Internationales d´archéologie et d´Histoire d´Antibes, Sophia Antipolis 1996, pp. 109-126. KazansKi M. 1999, Les tombes des chefs militaires de l’époque hunnique, in fisCHer tH.-PreCHt g.teJral J. (eds.) 1999, Germanen beiderseits des spätantiken Limes. Spisy Arch. Ústavu AV ČR Brno 14, Brno, pp. 293-316. KazansKi M. 2001, Les épées “orientales” à garde cloisonnée du 5e-6e siècle, in International connections of the Barbarians of the Carpathian Basin in the 1st-5th centuries A. D. Proceedings of the International Conference held in 1999 in Aszód and Nyíregyháza, Nyíregyháza 2001, pp. 389-418. KoCH u.-von WelCK K.-WieCzoreK a. 1996, Das Grab des Frankenkönigs Childerich, in Die Franken, pp. 879-883. König G.G. 1981, Wandalische Grabfundes des 5. und 6. Jhs., in «Madrider Mitteilungen», 22, pp. 299-360. KossaCK G. 1974, Prunkgräber. Bemerkungen zu Eigenschaften und Aussagewert, in KossaCK G.-ulbert g. (eds.) 1974, Studien zur vor- und frühgeschichtlichen Archäologie. Festschrift für Joachim Werner zum 65. Geburtstag. Münchner Beiträge zur Vor- und Frühgeschichte Ergänzungsband 1, München, pp. 3-33. maCKensen M. 2008, Tonabformung eines spätantiken kerbschnittverzierten Gürtelbeschlags aus dem zentraltunesischen Töpfereizentrum Sidi Marzouk Tounsi. Zur Mobilität comitatensischer Truppen, in «Germania», 86/1, pp. 307-322. martin M. 1989, Bemerkungen zur chronologischen Gliederung der frühen Merowingerzeit, in «Germania», 67/1, pp. 121-141. mengHin W. 1983, Das Schwert im frühen Mittelalter, Stuttgart. mengHin W. 1994-95, Schwerter des Goldgriffspanthenhorizonts im Museum für Vor- und Frühgeschichte, Berlin, in «Acta Praehistoria et Archaeologica», 26-27, pp. 140-191. merrills a.-miles R. 2010, The Vandals, Oxford. miKs Ch. 2007, Studien zur römischen Schwertbewaffnung in der Kaiserzeit. Kölner Studien zur Archäologie der Römischen Provinzen 8, Rahden. Palma santos A. I. 2008, The Visigothic tomb of Beja (Portugal), in aillagon (ed.) 2008, pp. 364-365. Pérez roDríguez-aragón F. 1992, Los cingula militiae tardorromanos de la Península Ibérica, in «Boletín del Seminario de Estudios de Arte y Arqueología», 58, pp. 238-261. Périn P.-KazansKi M. 1996, Das Grab Childerichs, in Die Franken, pp. 173-182. Pinar J.-riPoll G. 2008, The so-called Vandal objects of Hispania, in bernDt-steinaCHer (eds.) 2008, pp. 105-130. PoHl W.-meHofer M. (eds.) 2010, Archaeology of identity. Archäologie der Identität, Wien. Poinssot L. 1921, Tombeau d´Arifridos, in «Bulletin archéologique du Comité des Travaux Historiques et Scientiiques», pp. LVII-LVIII. Poinssot l.-lantier R. 1934, L’archéologie chrétienne en Tunisie (1920–32), Thuburbo Majus, in Atti del 3° Congresso Internazionale di Archeologia Cristiana, Ravenna 25-30 settembre 1932, Roma 1934, pp. 403-405. Quast D. 1993, Die merowingerzeitlichen Grabfunde aus Gültlingen (Stadt Wildberg, Kreis Calw). Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 52, Stuttgart. Quast D. 1999, Garnitures de ceintures méditerranéenes à plaques cloisonnées des Ve et début VIe siècles, in «Antiquités Nationales», 31, pp. 233-250. Quast D. 2005, Völkerwanderungszeitliche Frauengräber aus Hippo Regius (Annaba/Bône) in Algerien, in «Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz», 52, pp. 237-315. 236 CHRISTOPH EGER Quast D. 2009, Erben Roms. Völkerwanderungszeitliche Prunkgräber auf ehemaligem römischen Reichsgebiet, in 2000 Jahre Varusschlacht. Konlikt, ed. Varusschlacht im Osnabrücker Land GmbH Museum und Park Kalkriese, Stuttgart, pp. 372-378. Quirós Castillo J.A. (ed.) 2011, Dossier Archaeology and ethnicity: Reassessing the „Visigothic necropoleis“, in «Arqueología y Territorio Medieval», 18, pp. 9-53. raDDatz K. 1959, Das völkerwanderungszeitliche Kriegergrab von Beja, Südportugal, in «Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz», 6, pp. 142-150. ragona A. 1963, I tetrarchi dei gruppi porirei di S. Marco in Venecia, Caltagirone. sannazaro M. 2006, Elementi di abbigliamento e ornamentali “barbarici” da alcune sepolture della necropoli tardoantica di Sacca di Goito (MN), in buora M.-villa l. (eds.) 2006, Goti nell’arco alpino orientale, in «Archeologia di Frontiera» 5, pp. 59-73. sCHmauDer m. 2002, Oberschichtgräber und Verwahrfunde in Südosteuropa im 4. und 5. Jahrhundert. Zum Verhältnis zwischen dem spätantiken Reich und der barbarischen Oberschicht aufgrund der archäologischen Quellen. I Text. II Katalog - Beilagen Verbreitungskarten - Tafeln. Arch. Romanica 3, Bukarest. sCHulze-Dörrlamm M. 2002, Byzantinische Gürtelschnallen und Gürtelbeschläge im RömischGermanischen Zentralmuseum 1: Die Schnallen ohne Beschläg, mit Laschenbeschläg und mit festem Beschläg des 5. bis 7. Jhs., Mainz. sommer M. 1984, Die Gürtel und Gürtelbeschläge des 4. und 5. Jahrhunderts im römischen Reich. Bonner Hefte zur Vorgeschichte 22, Bonn. starK R. 2000, Studien zu den Schatzfunden von Szilágysomlyó. Beiträge zum edelsteinverzierten Goldschmuck in der Selbstdarstellung von Eliten spätantiker Gesellschaften, München. starK r. 2004, Ein Gürtelbeschlag mit Konsuldarstellung aus dem 5. Jahrhundert, in graenert g.-marti r.-motsCHi a.-WinDler r. (eds.) 2004, Hüben und drüben - Räume und Grenzen in der Archäologie des Frühmittelalters. Festschrift für Max Martin zu seinem 65. Geburtstag. Archäologie und Museum. 48, Liestal, pp. 21-32. teJral J. 1987, Zur Chronologie und Deutung der südöstlichen Kulturelemente in der frühen Völkerwanderungszeit Mitteleuropas, in «Anzeiger des Germanischen Nationalmuseums », pp. 11-46. von rummel PH. 2007, Habitus barbarus. Kleidung und Repräsentation spätantiker Eliten im 4. und 5. Jahrhundert. Ergänzungsband Reallexikon der Germanischen Altertumskunde RGA 55, Berlin, New York. Werner J. 1956, Beiträge zur Archäologie des Attila-Reiches. Bayerische Akademie der Wissenschaften, Philosophisch-historische Klasse, Kleine Abhandlungen Neue Folge 38A, München. zeiss H. 1934, Die Grabfunde aus dem spanischen Westgotenreich. Germanische Denkmäler der Völkerwanderungszeit II, Berlin, Leipzig. References to the illustrations Figg. 1, 4 (Ch. Eger) Fig. 2 (after KazansKi 2001, ig. 4F) Fig. 3 (after DuCCi-CiamPoltrini 1991, igg. 3-6) Fig. 5 (after arbeiter 2008, ig. 9) Fig. 6 (top: after böHme 1986, ig. 8; bottom: after sommer 1984, taf. 44) Fig. 7 (after Quast 1993, ig. 51 with additions) Fig. 8 (top: Foto Ch. Eger. - Mid: DAI Madrid, Negativ R189-97-7; bottom: after frauberger 189596, taf. 22) Fig. 9 (after aurreCoeCHea fernÁnDez 2001, ig. 25. – bottom: after Pérez roDríguez-aragón 1992, ig. 4) Fig. 10 (by courtesy of The British Museum, London).