9. FINES. The mechanisms and politics of frontier collapse, and the afterlife of frontier installations
Wednesday, 24 August 2022, Red Room

Rob Collins

Affiliation: Newcastle University, United Kingdo

Session Abstract: The realpolitik of the Western Roman Empire saw its fragmentation and collapse occur throughout the 5th century AD, with different processes of imperial shrinkage and consolidation occurring in the Eastern Empire through the 5th-7th centuries. Historiographic tradition links the end of Roman frontiers with barbarian invasion and settlement and imperial collapse. New archaeological evidence and reassessment of older data has prompted revision of this simple and entangled narrative of barbarian invasion and collapse. Indeed, new scholarship has called for separation if discrete issues and the impact this had on the limitanei and Roman frontiers. This session will address the following questions:

  • To what extent have textual sources and national myths pre-determined archaeological interpretation?
  • What is the evidence for abandonment or destruction at individual sites?
  • How do we recognise barbarians in the archaeological record?
  • Is there a difference between Roman occupation of military sites and post-Roman occupation?
  • Does the pattern of military occupation match that of urban and rural hinterlands of the frontier?
Time Presenter (s) Presentation


13.30 Berber van der Meulen-van der Veen Who supplied the foederati? Aspects of military equipment production in the Late Roman West
13.50 Ferdinand Heimerl Apud limitem Latina iura ceciderunt – processes of continuity and collapse on the middle Rhine frontier and its hinterland
14.10 Raymond Brulet The abandonment of borders in northwestern Gaul in the middle of the fifth century : role and responsibility of the Franks
14.30 Tomasz Dziurdzik “A former military road” – the afterlife of military installations in Trebižat river valley (Bosnia and Herzegovina)


15.20 Dan Matei The latest functioning phase of the castra in the province of Dacia
15.40 Agnieszka Tomas The transformation of a legionary fortress into a late Roman town. A case of Novae in Lower Moesia
16.10 Ignacio Arce A Tetrarchic Fort underneath the Umayyad Palace of Khirbat al Mafjar at Jericho?
16.30 Paul Kucera The borderlands of Egypt’s Western Desert in late antiquity
16.50 Nick Hodgson South Shields Roman fort as a case study in transition and abandonment at the end of empire
17.10 Russell O’Riagain The Role of Foederati and Secondary States in Ireland and Britain during the Era of Late Roman Hegemony, c.AD100–700

Who supplied the foederati? Aspects of military equipment production in the Late Roman West
Berber van der Meulen-van der Veen, Cardiff University

Recent works on Germanic foederati in the Lower Rhine region in the later 4th and early 5th century have made great progress in attesting their presence in the archaeological record using a variety of data, such as pottery fabrics (Van Thienen 2016), house plans and diet (Heeren 2016), and gold coin hoards (Roymans 2016). This last study showed clearly how dependent these groups were on these often sizable imperial payments, suggesting the foederati kept close financial ties to the Roman government. This paper will attempt to investigate other areas of material culture that might show this same core-periphery relationship, notably military-associated brooches and belt sets. After all, the arrival of Germanic groups in the region is often linked to the appearance of elaborate belt sets in male graves dated to the 4th and 5th century. (Böhme 1974; Sommer 1984). These belt sets fit in the provincial-Roman tradition of military insignia (Hoss 2012, 29), but their chip-carved decorative schemes are often seen as a Germanic trait. Little is known however about how and where they were produced. For the 4th century, stylistic studies (Swift 2000) and chemical analysis (Van Thienen and Lycke 2017) have shown that crossbow brooches were likely produced by the military fabricae, as they show remarkable standardisation in style/decoration, dimensions and alloy composition. The same might be assumed for 4th-century belt sets on stylistic grounds (Böhme 1974, 93-97), but it remains uncertain whether the fabricae were still active in the 5th century. To see whether the fabricae were ever involved in providing the foederati with belt sets, this paper will present some preliminary pXRF data on the composition of both 4th- and 5th-century buckles and fittings.

Apud limitem Latina iura ceciderunt – processes of continuity and collapse on the middle Rhine frontier and its hinterland
Ferdinand Heimerl, Generaldirektion Kulturelles Erbe Rheinland-Pfalz, Direktion Landesarchäologie

Around 475/477 the bishop of Clermont-Ferrand, Sidonius Apollinaris, wrote a letter to a certain comes Arbogast, who resided in Trier (Epistula IV 17). According to Sidonius, the Roman law had ceased at the border and the Roman speech had been wiped out from the Belgian and Rhenish lands. By contrast, the bishop praised Arbogast as one of the last bulwarks of romanitas. Without any doubt, this source must be critically reviewed with respect to its historical context. But nevertheless, the letter implies that Arbogast was the sophisticated head of a Roman community who probably still had military resources at his disposal. This paper opposes the literary sources to the archaeological evidence in terms of the latest Roman military activities from Arbogast’s residence in Trier to the middle Rhine frontier. While textual sources have pre-determined a collapse of the Rhine border in 406/407 for a long time, more recent studies argue for an intact middle Rhine frontier at least until the mid-fifth century. A short summary on the current state of research will be followed by a reassessment of military activities and fortifications in the surroundings of the former Imperial residence Trier. New evidence proves a usage of certain fortifications until the second half of the fifth century. This study contributes to the debate on differentiated processes of frontier collapse and the afterlife of military installations in different parts of the Late Roman empire.

The abandonment of borders in northwestern Gaul in the middle of the fifth century : role and responsibility of the Franks
Raymond Brulet, UCLouvain

The direct association of the term limes with rivers reflects its fourth-century usage, the status of rivers as boundaries plays a role in the new border consciousness in the Late Empire and there is a tendency for the Romans to consider frontiers as territorial and not just as divisions between peoples. But borders are only preserved or protected until they are useful and until there is a military presence to defend them. In the life of St Severinus it is recorded how the military occupation of the border ended in the middle of the 5th century along the Danube. It is not the same outcome that must be expected for northwestern Gaul. We have both a maritime and a river border, a context of early abandonment of territory within the Roman Empire and especially the omnipresence of Frankish people that we find inside and outside the frontiers as enemies or as allies. Thus, researching what the change in terms of border areas could be during the fifth century in northwestern Gaul corresponds to an investigation linked to the history of the Franks. It is therefore interesting to examine the relations between the Franks and the Empire and to follow the steps of their installation and their expansion in Gaul during the 5th century to understand the role that they could have in the protection and in the abandonment of the military border. The archaeological evidence which can be securely dated to the 5th century and related to this issue are controversial. Even if the problem is complex, it is necessary to associate the data which come to us from the frontier zone and the data which coincides in the North of Gaul with the historical advance of the Franks.

“A former military road” – the afterlife of military installations in Trebižat river valley (Bosnia and Herzegovina)
Tomasz Dziurdzik, Michał Pisz, University of Warsaw, Mirko Rašić, University of Mostar

The end of Roman frontiers as a research area is often dealing with large scale processes during a particular, late period of Roman history. However, valuable insight can also be gathered by observing changes on a more local scale – and also by comparing the situation in the 5th century with earlier examples of abandonment of frontier installations. This paper aims to analyse one particular example of an early abandonment of military installations. The ancient settlement in the Trebižat river valley (modern south-western Bosnia and Herzegovina) during the Early Empire was strongly influenced by a road of strategic importance and the presence of a garrison, both belonging to the defenses of Roman Dalmatia, the so-called limes Delmaticus (the term remains heatedly debated), which was gradually demilitarised. On the level of the Trebižat river valley the afterlife of this particular frontier had more to do with changes of a very local nature than Empire-wide processes, as the early 3rd century relocation of the auxiliary unit coincides with profound changes in local settlement patterns. On the other hand, during Late Antiquity numerous defensive installations have been constructed in the wider region, providing another point for discussion of the relations between local and global factors in the collapse of frontiers.

The latest functioning phase of the castra in the province of Dacia
Dan Matei, History Museum of Turda

The official evacuation of the province of Dacia occurred sometime 271-275, most probably in 271. Our paper will tackle the life in the Dacian castra in the epoch of the “military anarchy”, so starting with 235. Different aspects as the diminution of the number of soldiers (due to the external dislocations of a part of the soldiers), the coin circulation, the epigraphical manifestation, the use of the so called opus tumultarium and other phenomenon will stand in our focus. We will discuss also some criteria of discerning the traces of life from the last “Roman” years of the castra as opposed to the life of the first post-Roman years. Due to the significant enrichment of the informational block on the Dacian castra in the last two decades, our research has good premises to be carried.

The transformation of a legionary fortress into a late Roman town. A case of Novae in Lower Moesia
Agnieszka Tomas, University of Warsaw

The results will be partly presented in the publication in Archaeologia Bulgarica in December 2020, where we show the finds from the necropolis. For almost three centuries the fortress at Novae was one of many legionary bases enclosing a “playing-card” shaped area with a civil settlement outside its fortifications. In the late Roman period this shape was changed, and the former rectangular layout became took shape of a polygon covering the legionary camp and the area to the east of it. The excavations carried out in 2016–19 within this area resulted in the discovery of the previously unknown legionary necropolis, a so far little-known biritual burial ground. A series of disastrous barbarian raids around the middle of the 3rd century ended up with desacralization of this place in the 260s, which was thereafter followed by Diocletian’s reforms of the army and the Tetrarchic investments along the Lower Danube. Annexation of the necropolis into the town’s space seems to be exceptional, and shows how the extramural space was adapted for a new use. The discoveries shed light on the reasons and the process of transformation of Roman legionary fortresses into late Roman towns.

A Tetrarchic Fort underneath the Umayyad Palace of Khirbat al Mafjar at Jericho?
Ignacio Arce, German-Jordanian University

This paper presents the hypothesis of the existence of a Tetrarchic Fort underneath the Umayyad palace of Khirbat al-Mafjar at Jericho (Palestine), based on the preliminary result of a series of remote-sensing surveys carried out as part of the Jericho Mafjar Project in 2014, and the Marie-Curie Research Project directed by the Author at the University of Copenhagen. The orientation, dimensions and shape of these structures would allow interpreting them as belonging to a Roman fort 100m square, probably from Late Roman period, similar in dimensions, size and orientation to those of Daja’aniya, Avdat, Umm al Jimal or Khirbet el Khaw. Written sources mention the potential existence of at least three different Roman installations in the Jericho oasis from the 1st century CE throughout the Tetrarchic period. Legio X Fretensis had its winter camp at Jericho in 68CE in a location which remains a mystery. Apart from this winter camp, a military detachment would have been set permanently in Jericho, to control this strategic oasis, in the crossroads which links Jerusalem to the Jordanian plains to the East, as well as to control the traffic along the Jordan Valley, and blocking the access of Bedouin raiders. These reasons certainly determined the construction of successive Roman installations in the Jericho oasis in later periods. We know from written sources that a Roman fort was established in Jericho in 130CE, which played a role in putting down the Bar Kochba revolt in 133CE. The location of this new military installation would have been in a strategic place at the edge of the oasis, but near to crossroads, and accessible water sources, a description which fits with the location of Mafjar. The confirmation of the existence of this fort under the Umayyad palace, together with the evidence of the existence in the vicinity (if not at the same premises) of a monastery (looted for the construction of the Umayyad palace), would reinforce the model of transformation and change of use of Roman forts from the Limes Arabicus put forward by the Author. According to this interpretative model, many of these forts would have been transformed in monasteries (and in some cases palatial venues by the Ghassanids), and later into Umayyad palaces (ARCE, I. 2015: “Severan Castra, Tetrarchic Quadriburgia, Justinian Coenobia, and Ghassanid Diyarat: Patterns of Transformation of Limes Arabicus Forts During Late Antiquity”, In Collins, R. Roman Military Architecture on the Frontiers. Oxford. 98-122).

The borderlands of Egypt’s Western Desert in late antiquity
Paul Kucera, Qasr Dakhleh Project

Situated along the western periphery of Rome’s Egyptian province, the Western Desert may be viewed as an extensive natural frontier containing border districts or borderlands that were centred on large, populated oases that punctuated this hyper-arid region. Towards the end of the 3rd century CE, three of the oases witnessed a fort-building programme that coincided with the deployments of Roman auxiliary units. The main camps (castra) and the units stationed in them fulfilled a multi-faceted role within the region, not least of which was security related. There is limited information concerning the longevity of the castra as functional bases and the length of time that their garrisons served, though archaeological investigations conducted over the past two decades have brought to light much information concerning the establishment and occupation of the forts. Perhaps unsurprisingly, the 4th century is predominantly represented. One document, a Coptic ostrakon found at al-Qasr in Dakhleh Oasis where one of the castra was located, is particularly intriguing with respect to a possible end point to the strictly military purpose of the fort or indeed a potential changing aspect of the auxiliary unit stationed there. This paper explores this document further in relation to what is known of the late Roman military presence in the Western Desert, with a specific focus on the southern oases, as reflected in the archaeological record and supplemented by documentary evidence. The analysis also extends to include contextual themes such as geography, State interests, security of the region, the threats posed by raiding groups during the following 5th through 7th centuries, socio-economic, religious, and political aspects of life in the oases, the environment, and the abandonment of some oasis sites/settlements and evolution of others, including the afterlife of the forts.

South Shields Roman fort as a case study in transition and abandonment at the end of empire
Nick Hodgson, Independednt Researcher

Between 1983 and 2006 the east quadrant (2,000 square metres) of the Roman fort at South Shields, near the east end of Hadrian’s Wall, was excavated, revealing the commanding officer’s house (praetorium), several barracks, and the via praetoria of the fourth-century fort. Post-excavation work on the latest levels has now established a provisional late- and post-Roman sequence. Probably in the 370s an apsidal hall-like building was built over the via praetoria. A possible church identified overlying the principia probably dates to this time. By about 400 the praetorium was much re-configured and used by craft workers. Radiocarbon dates indicate that there was still a settled community living in the fort after 400, and the whole site was re-fortified with a new ditch after the beginning of the fifth century. Parts of the praetorium and the hall building were re-occupied and re-surfaced at some remove from 400, after a period of neglect or abandonment. The sequence ends with the burial in the centre of the praetorium courtyard of two violently killed individuals, probably before 440 according to the radiocarbon dates. Following this, the east-quadrant buildings seem to have rapidly fallen into ruin. They were overlain by Anglo-Saxon burials, indicating Anglian settlement in or near the fort in the seventh to ninth centuries, when there is reason to believe that it functioned as a royal centre. The site seems to have been finally abandoned at the time of Viking incursions in the ninth century. The implications of this remarkable sequence for debates about the fifth century transition on Hadrian’s Wall are discussed.

The Role of Foederati and Secondary States in Ireland and Britain during the Era of Late Roman Hegemony, c.AD100–700
Russell Ó Ríagáin, Queens University Belfast

There has been much important work on recent years re-evaluating the role of groups from Ireland within the Roman Empire and groups in Ireland’s participation within the sphere of Roman hegemony. This work has taken place contemporary to a shift towards treating textual source material on its own terms rather than through the lens of later historiography, placing emphasis on the socio-political context of textual production and the intended use, readership and audience of that text. With these two trends in mind, this paper will outline the material and textual evidence for the existence of secondary states in Ireland dependent on late imperial Roman hegemony. It will contrast this with the evidence for returning foederati mercenary units, assess where they may have operated prior to returning, and propose a major role for these groups in the highly dynamic political and cultural situation in Britain and Ireland in Late Antiquity. The paper will focus on three such groups: the Cruithin in northeast Ireland, the Uí Néill in central Ireland, and the Dál Riata in northeast Ireland and northern Britain. Carrying out this work requires setting aside many national myths and taking a source-critical, first-principles approach to the archaeological and documentary evidence on its own terms—a difficult but rewarding process, considering the success of the descendants of some of these groups in sponsoring the construction of layers of narrative complexes reframing their origins and relationships with other contemporary groups.