The recognition that warfare played a central role in the creation and maintenance of Roman hegemony in the Mediterranean basin and beyond is scarcely a modern conceit. After all, the Romans themselves attributed their imperial success to a preordained destiny to subjugate others and that this divine mandate should be achieved through prowess at arms and the relentless pursuit of military excellence. Many of the political structures of the Roman state were based upon the efficient mobilization of its military resources, and service with the army was deemed an essential component of the aspiring politician’s résumé. Indeed, there have been few empires before or since in which the concept of serving the state has been so intimately equated with the performance of military duty and where demonstrable success in the martial enterprise would be rewarded so directly by social and political advancement. Given the contemporary emphasis ascribed to the subject, it is not surprising that modern scholarship has devoted considerable effort to understanding the manner in which the Roman army performed its tasks. Much ink has been spilled in an attempt to reconstruct the composition and organization of the army, how it was equipped, how it operated on and off the battlefield, and how it presented itself to the outside world as an exemplary manifestation of the power and coercive will of the state. In this endeavor, the rich legacy of the surviving literary sources, the epigraphic and artistic record, and the archaeological evidence have been harnessed to provide the most complete reconstruction of any ancient military force and the functional and symbolic roles that it played within and on behalf of its originating society.

One particularly rewarding avenue of inquiry that allows the researcher a means of approaching many of these questions lies in the study of Roman forts and fortifications. These structures speak both to the practical application of military doctrine (in that they were constructed for explicit purposes and with the needs of specific garrisons in mind) and to more abstruse notions of power projection and imperial confidence. In this latter regard, any automatic equation that might be drawn between the act of fortification and a monolithic defensive mindset would be badly misplaced as it is clear that the Romans built these works not merely to delimit their territorial ambitions but also as prominent markers of control and as the harbingers of future expansion. Not least, the survival of hundreds of Roman fortifications across the empire provides an unrivaled database for explaining the tactical and strategic decision making that lay behind their construction and for analyzing the day-to-day realities experienced by the garrisons based within their walls.

This article reviews the main categories of Roman fortification, including the wall circuits provided for urban centers, the fixed bases for specific unit garrisons, and (more briefly) those examples falling within the ephemeral category of fieldworks. Given the constraints of space, linear networks such as Hadrian’s Wall and the Fossatum Africae will not be discussed as these are better regarded as frontier management systems (albeit that later barrier schemes such as the Claustra Alpium Iuliarum, blocking overland access to Italy from the east, were much more explicitly defensive in function). Whether these fortifications were meant to represent a permanent assertion of Roman authority or were designed for more episodic ends, their legacy speaks eloquently to the purposive determination of the Roman state and the deliberate professionalism of the army that served it.

Walled Cities.

Although the most overtly defensive of the three categories of Roman fortifications, the provision of fortifications around urban centers need not always have signified a primary concern with the physical protection of the inhabitants. Regardless of whether these urban circuits were represented by full-blown encircling walls with projecting towers and elaborate obstacle fields or much more modest “ditch and palisade” enclosures, there remains the possibility that the main motivating factor behind such construction could be found in the realm of the symbolic rather than the prosaic. After all, for the Romans one of the attributes of a “proper” city was that it should be enclosed by some form of defensive circuit which both delimited the urban area and dignified it. In a status-conscious world that operated on hierarchical principles in almost all forms of social and political expression, the manner in which a town presented itself was a major consideration, particularly as competition between urban centers intensified over time. If for the Greek polis an urban enceinte stood as stark proof of the autonomy of the city-state, in the Roman world moenia (walls) were viewed as one of the markers of civilization that conferred distinction to a town, especially if neighboring centers did not display such extravagant forms of conspicuous consumption. Nonetheless, although urban defenses may have been conceived of as trump cards in the status-display stakes, it is clear that for many Roman centers there remained a practical rationale behind the provision of town walls.

By the later years of the republic and the early years of the empire, the increasing security of the Italian peninsula (at least insofar as external threats to Roman hegemony were concerned) meant that there was a much less pressing need to provide new foundations with expensive wall circuits. Nonetheless, with questions of status now firmly to the fore, new colonies (usually settled as part of the “retirement package” for legionary veterans) still pressed ahead with the construction of elaborate freestanding urban defenses. Among the best preserved is the city of Aosta, formally established in 25 B.C.E. after the defeat of the alpine tribe of the Salassi. Even though there was no real security threat to the new city, Colonia Augusta Praetoria Salassorum was furnished with an impressive concrete-built wall faced with small stone blocks extending ca. 2,379 by 1,870 ft (ca. 725 by 570 m) and standing ca. 21.3 ft (ca. 6.5 m) high. The wall, 9 ft (2.75 m) thick at the base, was reinforced along its inner face by a series of buttresses and provided with 20 towers (including the gate towers) that projected ca. 14.8 ft (ca. 4.5 m) beyond the wall line (allowing flanking fire to be maintained along the wall). That the large capital outlay involved in the erection of these defenses was really an exercise in showmanship rather than in the performance of strict military utility can be seen in the provision of four gateways through the defenses, the largest of which included a wide central vehicular arch flanked by two pedestrian passageways.

These types of monumental gateway (that would horrify more security-conscious medieval rulers) are typical of Augustan through Hadrianic constructions and can be found at favored locales across the empire. Some noteworthy examples are the double vehicular arches and associated pedestrian passageways crowned by an extensive arcaded gallery of the Porte St. André at Autun (built as a very ostentatious “gift” to the loyal tribe of the Aedui by a grateful Augustus [r. 31 B.C.E.–14 C.E.]); the large masonry quadruple gateways provided at both the colonia (colony) of Colchester and the municipium (chartered town) of Verulamium (St. Albans), the latter probably being contemporaneous with the initial earthwork defenses of the town rather than the later rebuild in stone; the Hadrianic triple-arched gateway beneath the Damascus Gate in Jerusalem (again to benefit the new colonia of Aelia Capitolina); and the decorative but militarily impractical Hadrianic South Gate at Gerasa (Jerash), a freestanding structure that was not tied into the defenses of the city until the fourth century C.E. The foundations of similarly impressive but entirely symbolic gateways can be seen elsewhere in the east, for example, at Scythopolis (Beth-Shean) and Gadara (Umm Qeis).

The Roman rediscovery of the utility of city defenses qua defenses can be seen to stem from the crisis of the third century C.E. that demonstrated the vulnerability of the imperial core when the frontier systems failed. For example, apart from a few prestigious coloniae, the vast majority of Gallic cities had grown up untrammeled by walls and were therefore ill-equipped to repulse the barbarian attacks that overwhelmed the Rhine-Danube defenses. The response was a rapid and ad hoc process of rapid fortification whereby the aesthetic appeal of walls was subsumed by the need to provide effective protection for the inhabitants within them. These new urban circuits, at least in the first wave of construction, were more concerned with the delimitation of a defensible zone capable of being held by a relatively small garrison; and little attempt was made to provide comprehensive security for the entire urban area. Thus, the defended centers of later third-century Gaul are often characterized by hastily extemporized walls, making the most economical use of resources so that only a defined core area would receive the benefit of direct protection. This rapid construction is often marked by the use of spolia, or building materials recycled from demolished buildings (column drums and tombstones were frequently cannibalized in this manner), with the resulting walls making maximal use of advantageous terrain or running to link together existing well-built structures that could be pressed into service as part of the enceinte.

At Amiens and Tours the amphitheaters were conscripted for this purpose, and the initial scheme of works at Bavai saw the town’s forum turned into an emergency fort by the blocking off of open shop fronts. The example of Senlis is instructive. Here, only a small elevated area of the former urban core was defended (enclosing just under 16.1 acres [6.5 ha]), and the resulting ovoid enclosure was swiftly completed, the foundations being laid without any mortaring and with extensive use of architectural fragments being built into the defenses. Although this may suggest a rather ramshackle form of construction, the survival of the Roman work long into the medieval period suggests otherwise; and the town wall extended 23 ft (7 m) high to the parapet walk with a thickness of ca. 10.8 ft (ca. 3.3 m). Perhaps as many as 30 U-shaped towers lay astride this wall despite the circumference of the circuit being less than 2,789 ft (850 m). Such rapid refortification can also be seen in other areas of the empire suddenly exposed to threat for the first time in generations.

However, this sudden demonstration of the chimerical nature of imperial security meant that emergency building work alone would not suffice, and as the empire survived the initial shock of deep penetration attack, more precisely planned building projects were undertaken as part of a deliberate approach to urban defense. The most striking evidence for this comes from Rome itself, where Aurelian’s decadelong building scheme saw the construction of an 11.2 mile (18 km) enceinte up to 13.1 ft (4 m) thick and 65.6 ft (20 m) high liberally provided with powerful towers and 18 major defended gateways. This massive investment gave the capital its first comprehensive defenses in 400 years and signified the dawning realization that the empire could no longer rely on the field army alone for its protection. The provinces too saw this centrally planned response, with virtually all urban centers in Britain, for example (even small market towns located deep in the countryside), undertaking a grand scheme of fortification, usually involving a wide ditch fronting powerful, rubble-and-concrete walls faced with neat stone block work, studded with solid projecting bastions capable of housing heavy stone projectors. Similar fortification initiatives across the empire provide dramatic evidence for the perceived transformation of the security climate, and even if these later third- and fourth-century C.E. responses may have been disproportionate in view of the actual threat faced by much of the citizenry, they speak very clearly to the need to make the state’s commitment to the defense of its subjects both overt and tangible.


Similar rationales to those that informed the decision to provide towns with appropriate defenses can also be traced in the manner by which the construction of garrison posts and strongholds in the landscape proceeded. In general terms, the Roman authorities did not lay much emphasis on the need to provide sophisticated protection for their military bases. After all, the general perception was that the Roman army was technically (and morally) superior to any opponent that it might face and that the best approach to dealing with emerging threats was to engage the enemy in the open field at the earliest opportunity. Therefore, imbued with such confidence, it is not difficult to see why the Roman army made only basic provision for the point defense of the garrison posts.

Thus, the early imperial “playing card” fort, an extended rectangle with rounded corners that was the model for the vast majority of purpose-built bases in the first and second centuries C.E., had only limited defensive qualities. Even if the initial earth-and-timber perimeter may have been converted to stone for those locations retained for long-term military use, the conventional construction of the towers provided at the four angles of the forts and any interval turrets along the line of the wall saw these structures built against the back of the enceinte with no provision for any outward projection. These conferred very restricted opportunities for lateral flanking fire along the wall so that any enemy that succeeded in crossing the advanced obstacle field (usually arranged as two V-shaped ditches) and in reaching the base of the rampart would have been largely immune from any interdiction from missile troops stationed in the towers. This weakness was also replicated at the gateways (usually four in number), which again were either constructed flush with the wall circuit or with only a marginal degree of projection beyond it.

Although this mode of standardized construction across the empire may have reflected the limited capacity of Rome’s enemies to mount a serious threat to even modestly fortified positions (a doubtful proposition in many areas, particularly in the east), it is more likely to echo the prevailing doctrine that hostile forces were best confronted in the open and that reliance on passive defense was both tactically inappropriate and demeaning to the Roman conception of military virtue. Of course, even at the height of Roman confidence it was recognized that some advanced bases might require supplementary defensive measures to enable them to hold off an attack before reinforcements could come up to disperse the assailant. However, even in such cases, little attempt was made to change the basic configuration of the standard fort plan, with the provision of supplementary ditches (up to six at Birrens in southern Scotland) being the only concession made to address the enhanced security risk.

Although forts might vary considerably in size (depending upon the unit that the infrastructure was meant to house), the internal layout of these establishments also enjoyed a remarkable consistency, reflecting a commonality of approach to the provision of the type and placement of the buildings within the walls. At the center of the fort, at the junction of the two principal streets, was the headquarters building (principia), which accommodated the administrative tail. This was usually the most substantial structure in the complex, comprising an enclosed courtyard opening on to an aisled cross-hall (frequently appointed with elaborate architectural details) backed by a range of offices. In close proximity to the principia was the spacious accommodation of the unit commander (praetorium), a courtyard house of Mediterranean type that included the usual refinements that Roman gentlemen would expect of their private residences. The troops were accommodated in serried rows of barrack blocks (usually distributed one per century) with each contubernium (or section of eight men) sharing a two-roomed area for both sleeping and equipment storage. Separate quarters of a more commodious nature were attached to one end of each barrack block to provide the centurion (and his under-officers) with suitable accommodation.

A range of supporting buildings were also commonly found within the walls, including stables, granaries, workshops, and (somewhat more unusually) hospitals, with latrines and ovens being distributed around the inside perimeter of the ramparts. In most cases there was a bathhouse provided for the garrison, although the constraints of space (and the location of the water supply) might see such a structure located outside the defended area. Large forts designed as the base camps of legions might also have a wider range of facilities than smaller auxiliary forts, including enclosed exercise halls, swimming pools for soldier training, and amphitheaters for both entertainment and weapons drill. Parade grounds and associated reviewing stands located in close proximity to the fort might serve as substitutes for such amphitheaters for auxiliary units (although, unusually, the late first- to mid-second-century C.E. site of Tomen-y-Mur in north Wales was provided with both a parade ground and a small-scale arena).

Acting as the bases for mobile patrols, forts and the roads that linked them one to another and to the main provincial network were seen as an appropriate straitjacket for holding down recently conquered territory (as Tacitus points out, Agr. 1.20). Indeed, the necklace of control that resulted from extending such a system across the landscape was designed to isolate hostile areas (a scheme particularly useful in mountainous districts where the Roman web stretched out along the valley bottoms and severed connections between upland massifs where recalcitrant elements might hold out). However, far from being viewed merely as strongpoints from which regional control might be exercised, forts were also planned as avatars of power projection as well. The system of so-called glen-blocking forts in central Scotland, anchored on the brand new legionary base of Inchtuthil, is probably best interpreted as establishing the staging areas and jumping-off points for the planned resumption of the Flavian advance deep into the Caledonian mountains. Equally, the “forward” policy adopted by Septimius Severus (r. 193–211 C.E.) in the Near East and Libya is also manifested by the positioning of new forts (such as Qasr al-Azraq and Bu Njem) well in advance of the zone of direct imperial control, presumably to act as posts from which the active supervision (and manipulation) of the desert tribes might be more effectively pursued.

Roman Forts and Fortifications

Roman fort at el-Lejjun. Georg Gester/Photo Researchers, Inc.

view larger image

However, the same insecurity that gave birth to the comprehensive program of urban fortification from the middle third century C.E. onward also impelled the Roman authorities to take much greater care over defensive provisions for their military posts. Although it is not necessary to accept the idea that henceforth the Romans embraced a wholesale doctrine of “defense in depth” as some have suggested, this new emphasis on the tactical role of fortifications clearly demonstrates an acknowledgment that isolated garrisons would be expected (and able) to hold out in the face of concerted attack.

That heavily protected individual bases could therefore play a useful role in disrupting the momentum of any hostile incursion amounts to a significant reappraisal of the utility of passive point defense in Roman military thinking. This new approach is evidenced by the care taken over the manner in which new bases were configured with defensive provisions firmly to the fore. Purpose-built late legionary bases such as el-Lejjun and Udruh in southern Jordan or refurbished preexisting garrisons such as Léon in Spain featured impressive thick-walled curtains with numerous projecting U-shaped towers (extending 36.1 ft [11 m] from the enceinte at el-Lejjun and spaced some 49.2 ft [15 m] apart at Léon), giving a formidable defensive capability. Similar provisions were also made for other military bases, from the strongly bastioned circuits of the later Saxon-Shore forts of southern Britain to powerful “bridgehead” forts such as Deutz opposite Cologne on the east bank of the Rhine or Sucidava opposite Oescus on the lower Danube.

The much smaller burgus (small fort) or quadriburgium (a fort with four corner towers) installations of the later empire that provided secure accommodation for units tasked with local patrolling or defense may have varied considerably in terms of layout and size, but they nonetheless had quite sophisticated defensive provisions in common. At Zurzach on the Swiss Rhine, for example, a pair of tower-and-bastioned enclosures covered the road heading for the river crossing (with a further square-shaped fortlet securing the western end of the bridge on the north bank), while a typical example of an eastern quadriburgium with boldly projecting square-shaped angle towers and a well-built 8.2 ft (2.5 m) wide stone curtain delimiting an area of ca. 131 by 131 ft (40 by 40 m) can be seen at Yotvata in southern Israel. The fact that this latter site was also provided with a postern gate concealed next to one of its angle towers suggests that older Roman ideals of engaging the enemy in the field might still have enjoyed considerable currency where circumstances so permitted.


A very distinctive feature of Roman military practice was the emphasis ascribed to the discipline of field craft: that is, the extemporizing of temporary works to provide units with some elementary protection while on campaign. The resulting structures, built for short-term benefit, are usually better distinguished from more permanent installations by describing them as “camps” rather than “forts.” The Roman habit of castrametation (the process of laying out camps) had a long pedigree and was sufficiently unusual to receive considerable attention from contemporary sources. The most complete descriptions can be found in Polybius’s Histories (6.27–42, written in the second century B.C.E.), Vegetius’s Epitome of Military Matters (3.8, written at the end of the fourth century C.E.), and the unknown author conventionally identified as pseudo-Hyginus whose Concerning the Fortification of Camps may have appeared any time between the later first and the third centuries C.E. It is worth noting that Polybius (6.42), an astute and experienced observer of military matters, draws explicit attention to the fact that Roman field camps (unlike those of the Greeks) were set up in as uniform a manner as possible and that their engineers were willing to sacrifice advantageous ground to preserve this standardized layout. Although these camps were never intended to be employed for any prolonged period of time (often simply representing overnight protected encampments for the army on the march), surprising numbers have survived, albeit mostly in marginal environments.

A basic functional distinction can be drawn between fieldworks that were intended to accommodate troops bivouacking as they transited hostile territory (usually termed “marching camps”) and those that were intended to provide a short-term base of operations (“campaign camps”). The former category displays a broadly similar morphological pattern with a ditch and inner rampart outlining a rectangular/rhomboidal/trapezoidal trace. The rampart was built of stacked turf, earth or stone (depending upon the environmental circumstances), with the material excavated from the ditch forming a ready-made source of supply. The top of the rampart might then be reinforced by a palisade line comprised of the stakes that Roman troops carried as part of their regular impedimenta. Although the dimensions of the ditches and ramparts may fluctuate significantly between sites, the most obvious variant is that of size as camps were designed to be proportionate to the number of troops to be accommodated within the perimeter. As a result, many attempts have been made to trace the course of Roman campaigns by identifying linear networks of similarly sized marching camps stretching out across the landscape that are taken to represent the progress of Roman army groups as they advanced against the enemy.

The problem with the vast majority of marching camps is that only very limited archaeological work has ever been carried out in their interior areas, and it is therefore very difficult to come to any definitive conclusions as to the respective dating of these installations. Some attempt has been made to establish a chronological narrative based upon the various methods employed to protect the entrances to the camps. However, it is not clear that the deployment of traverses (tituli), whether in front of and recessed behind the gates, or inward- and outward-swinging extensions to the ramparts (claviculae) to mask the entrance gaps, can be so neatly isolated in a temporal sense. The problem is often compounded by the multiple uses of the same locations on various occasions, an entirely predictable outcome given that similar factors would have influenced the tactical choice of camping ground by different army columns at different times (particularly if the former defenses were still standing and merely required some refurbishment). Nonetheless, the systematic way in which the Roman army proceeded to provide elementary protection for its overnight encampments stands as a remarkable testament to the deliberate and professional manner in which it executed its duties.

The second category, “campaign camp,” exhibits greater divergence in form, mostly because these temporary bases performed a wider variety of tactical tasks than the marching camps. This might include sites that were intended to act as the operational nexus for Roman units posted in hostile territory for the duration of a seasonal campaign and, as such, might be provided with more substantial defenses than the conventional marching camp. A good example can be found at Mount Cílda in the Cantabrian Cordillera, where the most extensive hilltop position was fortified with an impressive arrangement of W-shaped ditches in advance of the rampart line in direct contrast to the other phases of (much slighter) Roman military activity at the site. A plausible interpretation is that this main encampment represents the command base for a large Roman column that was positioned here for some weeks or months as part of the Augustan campaign against the Cantabri and that the location was selected so as to control an open ridgetop that allowed faster (and more secure) penetration of the hill country than the densely wooded valleys below.

Other types of fieldworks that were thrown up with a view to longer-term occupancy might include temporary supply depots to the rear of the zone of conflict (Llansantffraid-ym-Mechain in the Welsh borderlands is a possible example) and the accommodation bases for construction crews working on substantial engineering projects (as can be found to the south of Hadrian’s Wall). However, perhaps the best understood of these types of field camp are the posts that were built as integral elements of siege operations. Clearly most sieges would require troops to be fixed in position for a protracted period of time, and accordingly, it would have been necessary to provide them with proper accommodation and appropriate security, particularly if there was any danger of a hostile sortie or any potential for a relief attempt by external forces. The defensive elements of Roman siege systems might include both camps and patrol bases for the attackers as well as works of circumvallation designed to encircle and isolate the target (and, depending upon the threat of intervention by hostile field forces, a supplementary outward-facing line might also be required). As with all other types of fieldwork, the local environment alongside the perceived level of threat had a major bearing on how these siege systems were planned and constructed.

Several examples of Roman siege operations have survived to a greater or lesser degree, with the examples of Numantia, Alesia, Machaerus, and Masada being particularly noteworthy. An especially instructive element of these particular reductive operations is that the extensive archaeological record is also complemented by literary accounts. Accordingly, the excavators of Alesia were primed to look for the extensive obstacle fields that Caesar describes as reinforcing his double lines of encirclement, the ephemeral traces of some elements of which (the bedding slots that anchored the cut-branch entanglements that served as the ancient equivalent of barbed wire, for example) might otherwise have been overlooked.

The Roman state may have preferred to view its enemies with disdain, and the common perception that the foe was best encountered in the field and defeated in a decisive battle before becoming a serious threat to imperial security was never really abandoned by Roman authorities. However, the crisis of the third century C.E. exposed the flaws in relying quite so completely on such a doctrine. As a result, the Romans took a much greater interest in point defense, and in subsequent years, far more attention was lavished on the science of fortification. The resulting proliferation of fortified centers, not merely in border districts but throughout the empire, conferred both physical and psychological reassurance to the citizenry and gave army units greater flexibility in deciding on when and where to engage the enemy. Isolated garrisons could prolong their resistance and (if they were bypassed by an invader) could continue serving a useful role either in tying up some of the assailant’s forces or in serving as nodes of logistical interdiction. But while the science of fixed-point defense may have improved considerably, it is worth noting that the former trademark skill of field fortification (whether in the form of marching or campaign camps) had become a neglected discipline. The later Roman army seems to have forgotten that one of the secrets to its past operational proficiency had been the realization that its troops had been as much engineers as frontline soldiers. The loss of these skills made it much less likely that old certainties regarding battlefield dominance and the exercise of moral and political suasion over its opponents could ever be satisfactorily restored.



  • Davies, Gwyn. Roman Siege Works. Stroud, U.K.: Tempus, 2006. Provides an overview of the different constructional elements that made up the Roman siege system.
  • Durand, Marc. La muraille antique de Senlis: Tours et détours senlisiens. Beauvais, France: GEMOB, 2005. A good example of the Gallic fortifications erected in haste during the crisis of the third century C.E.
  • Gregory, Shelagh. Roman Military Architecture on the Eastern Frontier, a.d. 200–600. 3 vols. Amsterdam: J. M. Hakkert, 1995–1997. A detailed synopsis concerning the planning and construction of later Roman forts in the eastern provinces.
  • Gros, Pierre. “Moenia: Aspects défensifs et aspects représentatifs des fortifications.” In Fortificationes Antiquae, edited by Symphorien van de Maele and John M. Fossey, pp. 211–225. McGill University Monographs in Classical Archaeology and History 12. Amsterdam: J. C. Gieben, 1992. Makes the case for the symbolic importance of urban fortifications.
  • Johnson, Anne. Roman Forts. London: Adam and Charles Black, 1983. Remains an excellent overview of early imperial forts and the buildings found within them.
  • Johnson, Stephen. Late Roman Fortifications. London: Batsford, 1983. An impressive compendium of the various types of fortifications built by the Romans (particularly during the third century C.E.).
  • Jones, Rebecca H. Roman Camps in Scotland. Edinburgh: Society of Antiquaries of Scotland, 2011. The most up-to-date review of the archaeological evidence for temporary camps anywhere in the empire.
  • Kennedy, David. The Roman Army in Jordan. 2d ed. London: Council for British Research in the Levant, 2004. Particularly useful for visualizing the appearance of Roman forts found along the southern flank of the eastern frontier.
  • Luttwak, Edward N. The Grand Strategy of the Roman Empire. Baltimore: Johns Hopkins University Press, 1979. A controversial but thought-provoking analysis of the different strategic postures adopted by the Roman state over time.
  • Maxwell, Gordon S. “Agricola’s Campaigns: The Evidence of the Temporary Camps.” In Agricola’s Campaigns in Scotland, edited by James Kenworthy, pp. 25–54. Scottish Archaeological Forum 12. Edinburgh: Edinburgh University Press, 1981. An example of how archaeological evidence is employed to flesh out the details of classical accounts.
  • Reddé, Michel, and Siegmar von Schnurbein, eds. Alésia: Fouilles et recherches Franco–Allemandes sur les travaux militaires Romains autour du Mont-Auxois (1991–1997). Vol. 1. Mémoires de l’Académie des Inscriptions et Belles-Lettres 12. Paris: de Boccard, 2001. A comprehensive publication of the large-scale excavations carried out at Alesia along the course of the Caesarian siege lines.
  • Welfare, Humphrey, and Vivian G. Swan. Roman Camps in England: The Field Archaeology. London: Royal Commission on Historical Monuments, 1996. The first detailed analysis of the archaeological evidence for Roman field camps on a regional scale.

Gwyn Davies