Military deployment along the Gask Ridge and the Highland fringe. David John Woolliscroft
The decades after the collapse of the Flavian dynasty are sometimes called the period of the adoptive Emperors, a time of great prosperity for the Roman Empire, when warfare (still present) takes second place to great building activity and the Emperors appear to be more interested in philosophy than madcap exploits in the dining or bed room. While it is hard to deny that the archaeological record in Rome and the provinces suggests a period of increased prosperity, our perception of a very ‘grown up’ Roman Empire is probably also due to the fact that for better or for worse we are missing critical details. Until the death of Hadrian (138), the epitome of Cassius Dio is still available, but the beginning of the reign of Antoninus Pius (138–161) starts with the stark statement: ‘It should be noted that the account of Antoninus Pius is not found in the copies of Dio, probably because the books have met with some accident, so that the history of his reign is almost wholly unknown.’ (Cassius Dio, Epitome of book LXX, 1)
Even when Britain does get a mention in the historical or biographical writings, such as in the Scriptores Historiae Augustae (SHA), a fourth century series of biographies with numerous question marks over its authorship, date, intention and origin, or in the letters of Fronto (the Latin professor and close friend of the later Emperors M. Aurelius (161–180) and Lucius Verus (161–169)), the material is often without clear dates and context. As a result, scholars rely mostly on the archaeological records, inscriptions and sometimes coins to reconstruct the history of the second century in Britain. As already mentioned archaeology’s strength is in detecting long term processes rather than short term events, so it is hardly surprising that we are able to document the long-term military developments in Britain, but short intervals of warfare may escape us.
One, and possibly the best documented, example of this is the War of Hadrian in Britain. We have two very short mentions of this war.
In the SHA we are told that ‘Britanni teneri sub Romana dicione non poterant’ – The Britons could not be kept under Roman control (SHA Hadrian 5, 2). The statement comes as part of a list of ‘subject’ people that caused uprisings during Hadrian’s rule: the Mauri (in Africa), the Sarmatians on the Danube, as well as trouble in Britain, Egypt, Libya (probably the Cyrenaica) and finally Judaea.
From what we can tell from other sources covering these events, the sequence appears to be roughly chronological: the Mauri uprising in western North Africa dates to Hadrian’s accession, the Danube war was resolved by Hadrian on his way from the East to Rome (so before 119/120), Egypt and Libya refer to what we know as the Jewish uprising, which also dates to the beginning of his reign, and finally Judaea can be linked to the Second Jewish War in the 130s. This would suggest that the British War may have also been part of the uprisings at the beginning of his reign.
However, it is possible that there is also a geographical sequence from the furthest southwest (Mauri) to the north (Sarmatians), even further north (Britain), to the south (Libya and Egypt) and finally the east (Judaea). The pattern is not fully correct and modern writers would probably have put Britain before the Sarmatians and progressed from west to east, so there is some debate as to which of the two criteria might have been the more important.
The British War is also mentioned in Fronto’s letters to Marcus Aurelius during the Parthian War in 162: ‘Quid? Avo vestro Hadriano imperium obtinente quantum militum ab Iudaeis, quantum ab Britannis caesum?’ – What? Under the rule of your grandfather Hadrian, how many soldiers were killed by the Jews, how many by the Britons? (Fronto, P.218, Loeb transl. vol. II, 23) Fronto was Marcus Aurelius’ Rhetoric and Latin teacher, but at the time of writing, M. Aurelius had long since finished being a student, but the exchange of letters continued, often creating little impromptu gems with considerations of style more than once trumping content. Unfortunately, this letter is very fragmentary and so we know nothing more on the subject, nor do we learn more about the losses under Antoninus Pius, apparently the subject of the next sentence. Again there is no chronology implied, but as in English, Latin cadences work better when you finish with the longest phrase, thus the three syllable ‘Britannis’ has to come second for stylistic reasons.
In addition to these two throw away sentences, there is also an inscription from Vindolanda, found reused in the praetorium (the commanding officer’s residence) in 1998, which was reconstructed by Tony Birley as follows:
D(is) [M(anibus)]
T(itus). Ann[ius …f …]
Centur[io.leg……praepositus. coh. I]
Tungr[orum Ann………stipend]
Diorum [……cecidi]
T. in bell[o Britt? ab hostibus. inter]
Fectus [ T. Annius (or Annia)...)
Fil et Arc[……]
H(eres) e[x testatmeno fec(erunt)].
To the spirits of the departed: Titus Annius…centurion of legion…acting commander of the 1st Cohort of Tungrians, a thousand strong, aged…with… years of service] died in the… War, killed by the enemy. Titus Annius… His son and Arc…his heirs had this set up, in accordance with his testament (the italicized text marks the translation of the surviving parts of the inscription (Birley 1998a and 1998b)).
Birley argues that the fighting must have taken place in the vicinity of the later Wall, but this does not follow, as the inscription comes from a gravestone, not a battlefield memorial. Graves could be built as cenotaphs, as the Caelius stone shows, set up twice, in the Roman province of Germania Inferior and outside the city of Rome, to commemorate the death of a centurio of one of Varus’ legions, although no body was available for burial.
Scholars over the last hundred years have linked the two texts with coins struck early in Hadrian’s reign (117/138) suggesting an early date for the above war and more recently for the inscription (in itself undated).
The coinage of Hadrian’s reign has attracted some attention from Romano-British archaeologists. There is frequent reference to issues regarding the mentioning of Victoria on the coinage of 119/120, as well as Mars the Avenger and Jupiter Victor, and also the now famous Britannia aes issues (RIC 577a and 577b = BMCRE 1175 and RIC II 845), which show Britannia seated on a rock, facing the viewers, holding a long sceptre (or spear) with a large shield visible to the right. The inscription in the exergue reads: BRITANNIA. In addition the mint of Alexandria issued a NIKE (Victory) issue to Hadrian in the year beginning August 119.
Much has been written about these coin issues in the context of Romano-British archaeology, suggesting that they may show Britannia defeated and thus in combination with the victory issues, a military victory over Britain. Birley (1997, 104) has raised some doubts about this reading. Problems arise on several levels.
Firstly, the Alexandrian issue commemorates a victory of the Emperor, not the Roman army, implying active participation of Hadrian, which could not be possible for a British campaign. However, Hadrian had just won a victory by resolving the Sarmatian problems in Moesia in person, it may thus be easier to ascribe the Alexandrinian issue to the Sarmatian war.
Secondly, the question of the Roman victory issues: first of all, it needs to be stated that none of the Hadrianic coins of that period appear to celebrate victories in the same way as the ‘Judaea Capta’ issues of the Flavian period, or the ‘Nemausus’ series celebrating the victory over Egypt with a crocodile chained to a palm tree. This means that either Hadrian had developed a new visual language or preferred not to commemorate the various victories in coinage.
A third problem is the context of the coin within the other coin issues of Hadrian in the period 119/122. This consists of two thematic series of coins, in addition to several special commemorative one-off issues. The first is a series of deities: Hercules, Minerva, Jupiter, Neptunus, Oceanus, Mars and Roma. The second series presents a catalogue of virtues, which includes Victoria, but also Fortuna (fate), Genius Aug/Bonus Eventus (the guardian spirit of the Emperor/the spirit of good endings), Felicitas (luck), Aeternitas (eternity), Concordia (harmony), Pax (peace), Pietas (duty), Salus (wellbeing), Spes (hope), Providentia Deorum (divine providence), Securitas (security), Libertas (freedom) and Virtus Augusti (the courage of the Emperor).
In combination these coin issues appear to present a generic catalogue of Imperial virtues/principles of the reign and Rome’s protecting deities rather than commemorating a specific victory. It certainly differs little from similar series of coins known from earlier Emperors. The only unusual feature on these coins is the frequency with which Oceanus and Neptunus occur, as well as the fact that the Felicitas Augusta coins tend to display a galley.
With regard to the Britannia coins, two features stand out: unlike other Roman issues commemorating defeated people, Britannia is not bound; quite the opposite, the long implement suggests that she is armed. By the early second century the Romans have developed a clearly defined iconography on how to portray defeated enemies, usually bound, often tied to trees or victory trophies; the Hadrianic Britannia just does not comply with any of these expectations. Furthermore, very similar designs (including the armour) appear in the provincial series at the end of Hadrian’s reign, when he honoured most of the provinces he visited with a special coin issue (this series incidentally includes another issue with a ‘dejected’ Britannia, holding her face in her right hand – still with the weapons next to her (RIC 845 = BMCRE 1723) and discussed by Jarrett (1976)). Most of the frontier provinces feature armed allegories and many of the island provinces are shown sitting on a rock, possibly to differentiate them from their landlocked counterparts. There is, in fact, a very similar coin issue for Germania (RIC 302), which shows Germania standing frontal, head turned right, holding a spear and balancing a shield with the left hand. As far as the author is aware this coin has never been claimed as evidence for a Hadrianic war in Germany; but it should be remembered that the Romans thought the only way to distinguish between the two heavily armed border provinces was to portray one standing (Germany), and the other sitting on a rock (Britain). In addition to the late series of provincial issues of the 130s, there are also a number of coins with provincial allegories that were struck contemporary to the Imperial visits to the provinces, usually accompanying coins showing Hadrian meeting the provincial armies or being greeted by the province on her knees. Seen in this context it looks more likely that instead of a victory issue, these Britannia coins were actually part of the coins issued in the context of Hadrian’s visit to Britain.
This image of Britain as seated and armed remained in use for some time beyond the reign of Hadrian. In 154 a similar design is chosen for Antoninus Pius’ Britannia issues. In this context the conclusions of D.R. Walker, on studying the coins from the Sacred Spring in Bath, are important, as he points out that while the late Hadrianic provincial series (those struck between134/138) can be found evenly distributed in all parts of the Empire, the early Britannia issues (119) (RIC 577) are near exclusively found on British sites, suggesting that they represent a special batch of small change, created for circulation in the province. This unusual distribution pattern can also be seen in the near identical Britannia asses of Antoninus Pius (Walker 1988, 290 and 294). The suggestion that a potential Hadrianic victory issue should only be available in the ‘defeated province’, where people would have known of the war anyway, seems highly unlikely and certainly without parallels in the Imperial mints. But without these coins, there is no compelling evidence for a war in Britain in 117/118, leaving us with the undated wars mentioned in the SHA and Fronto.
In addition to this early war, other scholars have suggested that a further war may have been fought in Britain in the 130s, this time citing the second Britannia issue of the 130s as evidence, and quoting in support the fact that Britain at this point was run by Julius Severus, according to the SHA, Rome’s best general at the time, although as Jarrett has pointed out, just because somebody was a good general, it did not even in Rome imply that there had to be a war wherever he went .The second argument adduced for a war in the 130s is the second ‘dejected’ Britannia issue, which as we have already seen above forms part of a wider series commemorating the Imperial travels to the various provinces earlier in the reign, and which as Jarrett pointed out, is accompanied by other issues celebrating the British army (EXERC BRIT) and his arrival in the island (ADVENTUI AUG BRITANNIAE) (RIC882) (Jarrett 1976, 145–152).
Further evidence usually quoted in support that the situation in Britain was apparently far from peaceful comes from two inscriptions by two different officers, T. Pontius Sabinus and M. Maenius Agrippa mentioning an ‘expeditio Britannica’ by Hadrian, as part of their careers. The former is sent to Britain after having been primus pilus (senior centurion) in the legio III Augusta in Africa, the latter was sent to take a force of 3,000 legionaries from Spain and Germany to Britain by the Emperor.
Maenius Agrippa, according to the inscription, was ‘chosen by Hadrian and sent on the expeditio Britannica’ (CIL XI 5632 = ILS 2735). The same Maenius Agrippa, who later became praefect of the classis Britannica and then procurator of the province, is also known from a series of inscriptions from the Roman fort of Maryport in Cumbria, where he commanded the local auxiliary unit (Birley 2005, 307–309). Unfortunately, once again none of these altars can be independently dated, so they offer no help with the reconstruction of when this expeditio Britannica may have taken place. Various reconstructions have tried to fit these inscriptions into the period of the apparently non-existent war of 117/118, with critics pointing out that Pontius Sabinus’ earlier jobs would have left him with an extremely crowded career. Frere (2000) has argued that this expedition should be moved to the governorship of Julius Severus in the 130s, to allow for the career to progress at a more manageable pace, perhaps unaware that Jarrett (1976) had already argued against the historical reality of this event.
Frere’s main argument for such a late date is that it would be impossible to call Hadrian’s visit to Britain in 122 an expeditio as this particular word has overtly military connotations of active campaigning, which would be inappropriate for an Imperial visit in the British context. He also points out that the inscription mentioned the term missus (sent) instead of using a term that implied he had accompanied Hadrian (Frere 2000, 25). Frere gives no parallels for somebody as junior as Maenius Agrippa being described as a companion (comes) of an Emperor. The term missus, which according to Lewis and Short’s Latin Dictionary can mean ‘sent’, also has two more appropriate connotations: in phrases such as ad cenam missus, it can mean ‘invited’ (in this case to dinner), while it has in other contexts the meaning of ‘guarding’ and ‘escorting’ – both meanings would be only too appropriate in the context of an Imperial visit.
More importantly, while Frere is probably right in most cases about the meaning of the term expeditio, in this particular case, the Emperor Hadrian appears to disagree with him. In 122 during his visit to Britain a series of coins are struck in Rome, showing the Emperor in armour on horseback with the legend: EXPED(ITIO) AUG(USTA) (RIC 613a and 613b =BMCRE 1261). As we know where Hadrian was during 122, first on the German frontier, then in Britain, and then in Gaul, it can be assumed that the term expeditio was indeed used by Hadrian for his trip to Britain, and thus copied by his military men. This may not have been approved by Latinists, but as Favorinus said, after being faced with a very similar misuse of Latin by Hadrian: ‘You are urging the wrong course, my friends, when you do not suffer me to regard as the most learned of men the one who has thirty legions.’ (SHA, Hadrian X V, 12–13)
It, therefore, appears that the two officers and the accompanying legionaries could have arrived at the same time as Hadrian in 122, but this still does not help to establish when the ‘war in Britain’ occurred. It has been argued that bringing 3,000 soldiers suggests the existing army had been depleted and that the legionaries were used to bring the other legions up to strength. However, most vexillationes of the first and second centuries are expected to return to their original units sooner or later and it may be just as likely that Hadrian brought the extra troops to have a sizeable field army for possible campaigning.
On the other hand Hadrian stayed only a few weeks, making it unlikely that he could have inaugurated or inspected Hadrian’s Wall in addition to leading a significant campaign. We have, however, in addition to the Alexandrian coin issue of 119, another Nike issue in the years 124/125 and 125/126 and John Casey (1987, 65–72) has argued that the likeliest, and in many ways the most convincing, candidate for these victory celebrations would be the War in Britain.
David Breeze (2003, 13–16) has drawn further attention to a series of construction changes in the building of Hadrian’s Wall, including a layer of soil which accumulated over the foundations of the fort of Birdoswald, a change in the standards of workmanship in the construction of the Housesteads fort and a narrowing of the Wall when work eventually resumed. Breeze’s proposition has not received universal acclaim, especially as the time span envisaged seems very short (Hodgson 2009, 16f). If it were to be accepted, it would provide a new scenario, where Hadrian decided to visit Britain in 122, and which he chose to have described as an expeditio on his coinage. We are not sure what prompted this visit, but there is no evidence that there was warfare in 117/118. There may have been military concerns though, as he was accompanied by a substantial number of additional troops, as well as spending some time inspecting or initiating the building of Hadrian’s Wall. It is not clear whether warfare was undertaken in advance of the Wall construction or as the building progressed, or alternatively if the building of the Wall itself destabilized the situation dramatically, leading to further campaigns in 124 and 125, which eventually required drawing more troops from the Wall itself, thus interrupting construction. Two successive victory issues in Alexandria in 124/125 may suggest protracted fighting and the casualties from the campaign, which might have included T. Ann(ius) from Vindolanda, were clearly remembered by Fronto, who would have been a very young man then, as one of the defining moments of Hadrian’s reign.
While the above might be valid and may even appear to be a convincing (one hates to use the term attractive in this context) argument to some, it should be stressed that other scenarios over the last sixty years describing Hadrian’s British War have been deemed equally compelling, only to be taken apart within a very short time. The evidence for any reconstruction of this and similarly badly recorded events of the second century is just too thin to allow for substantial confidence in any proposed scenario.
While trying to place a war within the reign of Hadrian, Hadrian’s Wall has been mentioned several times. The Wall is without doubt the military structure with which Hadrian is most associated, both in Britain and beyond. It is certainly Rome’s most architecturally elaborate land frontier and it is hard not to be impressed by its remains. In its current form, it represents the result of strategic repositioning by the Roman Empire, as well as the result of a long series of tactical decisions of how these new ideas were to be implemented.
When we first reviewed the deployment of troops in the province, two main considerations dominated the decision-making process for the Romans; policing of the conquered areas and positioning the army in a suitable spring board position for the next conquest. We have seen how this could lead to forts being placed close to former power centres (such as Colchester, Saham Toney or Hod Hill) or at the end of valleys suitable as routes of march into enemy territory, but well connected to the hinterland to allow for ease of supply (the so-called spring board sites such as Gloucester, Wroxeter, or the vexillation fortresses of the 50s in Lincolnshire).
Military deployment along the Gask Ridge and the Highland fringe. David John Woolliscroft
None of these structures was inherently designed to defend a marked out area or pre-defined territory: the spirit of the Julio-Claudian age (to avoid a problematic term such as propaganda) was to see the world as an area which was there for the taking. Drawing a line between what is Rome, and what is not Rome, is unlikely to have occurred to anybody. This does not mean that there were no parts of Britain where Rome’s control did not reach, but we have seen that client kingdoms were not necessarily deemed to be independent and other border territories could be perceived as ‘not yet Roman’, especially in view of the rapid growth of Rome in the hundred and fifty years between the fall of Carthage and death of Augustus. By the end of the first century, a different, and to modern readers, probably more realistic sense of Rome’s relationship with its neighbours had begun to develop. Conquest may still be preferable, but for example in Sala in Mauretania (modern Morocco) and also in the Delta area of Holland, towers were added that allowed forts to keep an eye on areas where access needed to be controlled, if only to prevent raiding of local livestock by populations on the opposite side. In Britain this is the state of play seen in Rome’s occupation of Scotland, south of the Highland fringe. Access in and out of the area was mainly along the wide valleys of Strathearn, Strathalmond and Strathmore, which provided fertile ground for agriculture. All the forts are carefully placed to allow control over river crossings or exits from a series of narrower interlocking valleys leaving the Highlands.
The backbone of the system was a road that came from the south. In its current form this is a carefully engineered all-weather road with cuttings and substantial underpinning to allow for the use of wheeled traffic. Recent work at Innerpeffray has, however, suggested that this road may not have been constructed in the Flavian period, but is a later addition, probably of the mid second century. If that is the case, then we are probably looking in the Flavian period, with a series of patrol tracks, suitable for mules and horses but not for heavy wagons, and road stubs leading out of the forts, which were ‘in the process of being upgraded’ when the order for the withdrawal came.
The main track along the central sector of this system crosses a small ridge, which kept traffic out of the waterlogged valley bottom, but also obstructed the view for the forts across from the Highlands on the north and the Ochills on the south side. At this point an additional feature is added in the form of a series of towers, which line the road. All of these towers have excellent views, either north or south, but in some cases both (although much of these are today obstructed by dense woodland). As they are also visible from the forts, it is possible to keep control over any traffic passing out of the Highlands or the Ochills. Further north, two towers are so far known, and they too are placed in such a way as to link the forts of Cargill, Cardean and the fortress of Inchtuthil.
The result is an area of tight control for the Romans, where they can police any disturbance with relative ease, while keeping the province further to the south and east free of problems. What it lacks is an artificial boundary, perhaps because the Highland fringe is used like an over-dimensional wall with limited exit and entrance points.
There is some discussion as to whether this represents an early Roman frontier. In some respects, particularly its depth, it differs from later Roman frontiers. Later frontier systems are often very narrow; all necessary installations are concentrated in a thin strip of land, with only the occasional outpost fort to document that Roman control was not meant to end at the barrier, while the back up of the nearest legion could be more than a day’s march away.
At the Gask and the Highland fringe, there are no outpost forts, and no hinterland legions; instead the military is deployed in almost a twenty mile wide corridor with a forward screen of forts on the exits of the Highland and a second line of forts along the tower line in the central sector. The legion at Inchtuthil is thus integrated into the frontier zone proper. All elements are clearly designed to work together. Woolliscroft and Hoffmann (2006, 234) called these elements ‘frontels’, elements that together make the frontier: the geographical barrier of the Highland fringe, towers that provide a supervized area, a supply road that keeps the forts supplied while being simultaneously guarded by forts and fortlets; some forts such as Bertha that are close enough to navigable rivers to have additional logistical functions. None of these ‘frontels’ are designed to work on their own. There is currently no evidence that there may have been a phase one with the Highland forts and a phase two with the forts along the road, but at the moment the evidence suggests that when the order to withdraw came in 86, the entire system was envisaged to operate together as a unified whole.
The Stanegate in its final phase. David John Woolliscroft
As described, this is clearly not a Roman frontier like the ones that can be dated to the second century. It still has some of the elements of the spring board deployment that we have seen in the mid first century AD in Wales, with one legion moved as far as possible to the front of the line and good logistical support for the forts in the form of supervized tracks, although the logistical roads in the Midlands were never guarded by towers as far as we can tell. However, the troop concentration is higher than before, and all later frontier elements are already there: road, barrier, tower, and fortlets. The system is then a hybrid, in-between the two types: this is the first time we see the Romans deploying a spring board for the armies and combining it with the development of a guarded zone at the margins of the province, and thus probably creating Rome’s first frontier.
In the years following the withdrawal from Scotland the military situation appears never to have stabilized to such an extent that another frontier line could develop. By 105, twenty years after the heyday of the Gask system, the forts had been withdrawn to the Stanegate. The debate continues whether there is another frontier system operating on the Stanegate before Hadrian’s Wall is built in 122. It is becoming increasingly clear that the earliest forts between the Tyne and the Eden were not designed or built to be a frontier line, but instead to have been deployed to facilitate the usual policing duties. By 105, when the area became the northernmost line of deployment of Roman troops, this situation appears to have changed. The withdrawal appears to have resulted in a much higher number of Roman forts to the west of the Pennines, and it is possible that after 105 we can see the building of some fortlets between the forts, as well as isolated towers, that extend the field of vision of the forts north beyond the high ground of the Whin Sill. However, there seems little evidence that this incipient system continued east of Corbridge, although Barri Jones first raised the possibility that it may have stretched westward beyond Carlisle to Kirkbride on the Cumbrian Coast (see Breeze & Woolliscroft 2009, 70–71 for a recent review of the evidence). If so, this system would again be a frontier without a continuous artificial barrier, but appears to be using the natural barriers of the Whin Sill and the Eden and Cumbrian coastal wetlands. Barri Jones believed that he could detect short stretches of palisade ditch along the route to Kirkbride, linking the low-lying areas and thus closing the areas to north-south traffic. If one accepts his findings, it would suggest that by the Trajanic period the Romans in Britain were already experimenting with artificial frontier obstacles. In the end the early phases of Hadrian’s Wall appear to have overtaken and replaced this system.
Western Stanegate and Hadrian’s Wall’s coastal defences. David Woolliscroft on the basis of a drawing by GDB Jones
In 122, around the time of Hadrian’s visit to Britain, a new frontier system was being constructed along the Tyne-Solway line. Originally starting at Newcastle in the west and running into the east as far as Bowness-on-Solway, it provided at its heart an 80 mile long wall (built of turf in the west and of stone east of the Irthing) which featured every mile a milecastle and two towers. In addition there were two major gateways: the Portgate in the East took Dere Street through the Wall to the North, the other must have been in Carlisle, but has not yet been identified. As originally designed, the wall and the towers along it were built to be visible to the forts along the Stanegate, such as Vindolanda and Haltwhistle Burn.
This system suggests a major change from the Gask: now the forts are kept along the road, while the towers are moved ahead, usually as far north as the field of vision allows. The result, however, is the same; the forts, fortlets and towers together create a large zone of supervision with controlled entrances at the gates and milecastles. The major differences are the advance line of forts at the feet of the hills and the proximity of the legion at Inchtuthil, in comparison to their location at Chester and York, well in the rear of the Wall. Finally, the controlled zone (nearly 20 miles in Scotland) is initially narrower in the first phase of Hadrian’s Wall; in some parts it is under a mile and a half. As a result, the first phase of Hadrian’s Wall in many ways appears to be a slimmed down version of the situation at the Highland fringe; and without the close proximity of the legion, a system that does not have the ability to be turned into an offensive spring board.
Plans for Hadrian’s Wall were, however, not static. Within a very short time, perhaps as little as a year, a new chain of forts were built up to the line of the Wall. They appear to be in many cases in addition to those on the Stanegate, many of which like Vindolanda and Corbridge stay in use. A further addition was the creation of the Vallum; a misnomer, as this is not a city wall, but the name appears to date back to the Venerable Bede and is hallowed by long usage. It consists of a deep flat-bottomed ditch which was flanked on both sides by substantial mounds set back from the lip of the ditch, as well as smaller, but contemporary mounds close to its edge. The whole obstacle is up to 30 metres in width and even in its current dilapidated state forms an impressive obstacle over substantial parts of the Central Sector of the Wall. Discussions about the function of this feature abound, and while it certainly presented an obstacle to the traffic across the area, both animals and, as Wilmott stresses, people, it is not clear if this was its main intended function. We have already seen above that David Breeze has mentioned the disruption of the wall building by the possible war in the reign of Hadrian, and it is possible that the Vallum may have been provided as an obstacle, precisely because Hadrian’s Wall itself at this point was barely higher than foundation level in many places; but there are problems linking this scenario with some of the stratigraphic evidence from the Wall itself, and the search for a possible function will be occupying researchers for many years to come.
When Hadrian died in 138, Hadrian’s Wall had on paper become a formidable obstacle: a central wall wide enough to take a wall walk, fortified with a chain of alternating milecastles and two towers, with fourteen forts along its line, which had by now been extended to Wallsend on the Tyne. Behind the Wall ran the Vallum with crossing points at the forts and the gates. In the East the system was augmented by a series of free-standing towers and milefortlets along the Cumbrian coast from Bowness-on-Solway to Maryport. The strategic reserves in case of attack were probably the two closest legions at Chester and York, a substantial distance to the south, but in between, a series of auxiliary forts to the south of the Wall continued to be occupied. There were also a series of forts in advance of the Wall along the main roads leaving the line. These outpost forts were Netherby and Birrens in the West, and High Rochester in the East; a further fort at Learchild may have had a similar function as well, as did the later additions of Bewcastle and Risingham. The wall line itself was in many areas still not finished, but in advance of the Wall there was in most areas a ditch (sensibly dispensed with along the Whin Sill crags in the Central sector), as well as a series of further approach obstacles between the ditch and the Wall itself.
Hadrian’s Wall. David John Breeze
However, with the death of Hadrian the political landscape changed once again. A new Emperor, Antoninus Pius, had to prove his credentials as a military leader and it appears that Britain was chosen as a convenient location. Hadrian’s Wall was abandoned, and a new wall, the Antonine Wall was built between the Clyde and Forth.
For most of its course, the new wall ran along the southern edge of the central valley of Scotland, often overlooking both the rivers Kelvin and Carron and their surrounding wetlands. There are, however, a number of differences, beginning with the geography. While Hadrian’s Wall stayed south of the Solway estuary, the narrower upper end of the Clyde was taken in by the Antonine Wall, which ends at Old Kilpatrick on the north side of the Clyde bridge, 37.5 miles in length from the other (possible) end at Carriden: it is thus substantially shorter than Hadrian’s Wall and lacks the elaborate defences on its western flank. To date there are only a number of small fortlets known, such as Lurg Moor, despite the fact that the Clyde is much narrower than the Solway estuary and more easily crossed by boat. In the east there were a number of forts along the southern shore of the Forth such as Cramond and Inveresk.
The biggest difference is, however, the Wall itself, which is throughout built of turf, usually on a stone base. Again we can detect a change in plan while construction was still in progress. Initially, in the first phase, six large forts and fortlets had been planned along its length, to which at a later stage a number of smaller forts were added. While Hadrian’s Wall forts on the whole were large enough to house entire units, this is demonstrably impossible with many of the Antonine Wall forts, which are much closer together than the Hadrian’s Wall forts.
There is also so far no indication of the same dense cordon of milecastles and towers that we see on Hadrian’s Wall. There certainly are a number of fortlets known from the Antonine Wall, and over the years several series of possible mile fortlets have been proposed, but very little work has been undertaken to confirm or rule out their existence. The most remarkable absence is, however, the lack of integrated towers. While these clearly have been identified in the Turf Wall sections on Hadrian’s Wall, their absence further north is striking, so much so, that over the years other features have been proposed, such as expansions, of which seven are known, some of which have produced evidence of burning. The other candidates are small enclosures, but without any remains of structures within their ramparts their function remains problematic.
A further candidate along the line of the wall is a free-standing tower with an unfinished ditch at Garnhall near Castle Cary (Woolliscroft 2009). Stratigraphically the structure is earlier than the Turf Wall, but the only finds were two unstratified glass bangles, which do not allow a decision as to whether this is ‘earlier’ during the construction process (in which case the Wall may only be later by a year or so) or ‘much earlier’ (such as the Flavian period).
Two further differences are the additional approach obstacles. While the Antonine Wall, at least east of Garnhall appears to have cippi (pits probably obscured by brushwood and possibly central stakes), similar to the ones found at Hadrian’s Wall, there is no evidence for the Vallum or a comparable structure behind the Wall. However, the ditch in front of the Antonine Wall is substantially larger than the one in front of Hadrian’s Wall (to the extent that at Watling Lodge it can accommodate a small house), and it has been suggested in the past, that this may have been influenced by the substantial size of the Vallum further south. Finally, in addition to the forts along the line there were once again forts in advance of the Wall, following the road north past Camelon. So far Antonine occupation has been proved for Ardoch, Strageath and Bertha, and Antonine finds are also known from Dalginross and Cargill, although the latter two appear not to be lying along the road.
Antonine Scotland. David John Breeze
This road building is a further feature of Antonine Scotland: the best known example, the Military Road, followed the Antonine Wall, linking the forts and fortlets together. But recent excavations on the road at Innerpeffray just east of Strageath have shown that the road appears to be later than the 63 acre camps, which would suggest that the carefully engineered all-weather road, which included a substantial road cutting and grading in Strageath and Innerpeffray, is more likely to be an Antonine feature than a Flavian one.
In addition to the forts on the Antonine Wall and behind it, a series of new forts were built in Lowland Scotland, including the large fort at Newstead, which re-occupied an earlier Flavian fort site (as did many other sites).
The manpower needed to garrison these forts came from the North of England and Wales. The latter had gradually been deprived of its garrison since the 80s and by the Antonine Period only a handful of military sites remained. The North of England fared little better. The garrisons on Hadrian’s Wall were withdrawn as were most of the ones from the Stanegate and Cumbria. South of the Wall, by current understanding, barely a dozen sites remained: Derby, Rocester, Templeborough, Manchester, Ribchester, possibly Low Borough Bridge, Brougham, Watercrook and South Shields, definitely Ambleside and Ravenglass, Maryport, Bowness, Carlisle and Corbridge. The legionary fortresses at Caerleon, Chester and York appear to have stayed in use, but appear to have largely stood empty, their garrisons since the Hadrianic period involved with building Hadrian’s Wall and then the Antonine Wall.
Antoninus’ reign (138/161) is frequently referred to as the most peaceful period of the Roman Empire, but a reference in Pausanias (who included this as an aside in his Description of Greece (VIII.43.3–4)), mentioned that ‘… in Britain… appropriated most of the territory of the Brigantes, because they too had begun a war, invading the Genounian district, which is subject to the Romans’.
Needless to say by now, this source is once again problematic. First of all we still have not identified a Genounian district in Britain (there is one in Raetia), and there are at least two Brigantian tribes in the Roman Empire (one in Britain, one in Raetia on the eastern end of Lake Constance). So Plutarch, who was anything but familiar with Britain or Raetia, may have confused which Brigantes were meant.
Originally this Plutarch reference was connected by Haverfield (1904, 142–4) to an inscription on the Tyne (RIB 1322) mentioning reinforcements travelling between Britain and Germany dating to the governorship of Julius Verus (158), the rebuilding of the fort of Brough-on-Noe under the same governor, a further possible mention of this governor at the outpost fort of Netherby, a passing comment in the SHA mentioning trouble in Britain in 161 (see below) and a coin issue of 154/155 ‘indicating a victory in Britain’ (cited after Breeze 2011, 118). Since the 1970s this Brigantian revolt has received continued criticism, not least as the coin issue in question is a reissue of the Britannia issues of Hadrian; and as we have seen above, both are only available in Britain and do not use the established victory imagery of the Roman Empire, which has led Walker (1988, 290 and 294) to suggest that they merely represent an issue of coins for use in the province of Britain, rather than a coin commemorating any specific historical event. The latest and extremely convincing refutation of any trouble in the Brigantian period came from Breeze in 2011 (pp. 118–119), who also refuted any involvement of this spurious uprising as a contributing factor in the abandonment of the Antonine Wall (or by extension, of a possible intensification of Roman control in the area north of Carlisle in the later part of the second century).
Reassessment of the evidence from the Antonine Wall in the 1990s (Hodgson 1995, 33–42) has shown that the Wall does not have a second period of occupation as once thought and that from 158 onwards, Hadrian’s Wall was increasingly brought back into operation (RIB 1389).
For the following few years, both the Antonine Wall and Hadrian’s Wall appear to have existed side by side (Breeze 2011, 121), one being gradually run down, while the southern frontier was being put back into operation and possibly in some places finished for the first time. Many of the re-occupied forts both on Hadrian’s Wall and in its hinterland have produced inscriptions naming the two governors most involved in this second withdrawal: Virius Lupus and Calpurnius Agricola.
The re-occupation was, however, far from being a restoration of the Hadrian’s Wall system of the 130s. Some elements of the system, such as the towers and fortlets along the Cumbrian Coast were not revived. The Vallum was modified by adding further crossings, thus making the Wall area more accessible, and a major change was introduced when Hadrian’s Wall acquired its own Military Way linking the forts on the Wall. In addition the Northwest and Northeast saw their hinterland forts either refurbished, or in the case of Dere Street, newly built, providing a chain of large forts along the main routes that could provide a strongly defensive force, but also potentially a substantial force for a sudden move north of the Wall (Breeze 2011, 211). Hodgson (2009, 31 HW) argues also that this is the time that the Hadrian’s Wall fort of Brough by Sands (II) finally replaced the earlier detached fort, but it is symptomatic of the gaps in our understanding that the latest report on the fort (published at the same time as Hodgson’s overview), is much more restrained in its dating and appears to favour, on the basis of the archaeological evidence, a Hadrianic date for the fort on the Wall (Breeze & Woolliscroft 2009, 77).
Another big change happened at Corbridge, where the earlier fort site was replaced with a civilian site, possibly a planned forum or market centre (which in the event was never finished). The military meanwhile moved to a new irregular walled compound south of the Stanegate, which was to house detachments of legio VI Victrix from York and legio XX Valeria Victrix from Chester (Bishop and Dore 1988).
As in the Hadrianic Period, the area under Roman control reached far beyond the line of the Wall, with ‘outpost’ forts (now including Bewcastle and Risingham close to the Wall) stretching at least as far as Newstead in the northeast and Cappuck and Birrens in the northwest. In addition it appears that even further north, occupation around Roman forts such as Castlecary and Inveresk may have continued, although the exact nature of this occupation remains a matter for debate (Leslie 2002, 17–28). We have already seen that at the time the withdrawal from the Antonine Wall began, a number of military sites in the southwest of Scotland appear to have been refurbished. This may have been roughly contemporary to the siege of Burnswark, which appears to have included an Antonine fortlet. The debate on whether this reflects a real or practice siege is still ongoing, with convincing arguments on either side. Certainly, Hodgson (2009, 189) argued in favour of the siege, citing the layer of burning and rebuilding at Birrens in 158, as well as the widespread rebuilding as part of the same reoccupation of the southwest, and proposed the tentative linkage with the Pausanias quote mentioned above; we have already seen that this is strongly refuted by Breeze (2011, 118–119). It is one of the problems of the second century that in the absence of clearly chronologically tied-in historical events, the same quotes can be used for a wide range of different incidents, and so the Pausanias quote is as easily useable for the initial conquest of the Lowlands at the beginning of Antoninus’ reign, as for this particular event. Although given the Genounian district in Raetia it may be wiser to use this passage as little as possible.
That there were at least fears of warfare at the time can be deduced by a quote in the SHA dating to the beginning of the reign of Marcus Aurelius (SHA M. Aurel 8,7): ‘War was threatening in Britain and the Chatti had invaded Germany and Raetia. Against the Britons Calpurnius Agricola was sent, against the Chatti Aufidius Victorinus.’
Calpurnius Agricola was a governor whose activities in the North of Britain are widely attested through numerous inscriptions: we have already mentioned his involvement in the refurbishment of Hadrian’s Wall and its hinterland (examples survive from Ribchester, Hardknott, Corbridge, Vindolanda, Carvoran), which demonstrate that he must have been in Britain in 163. Given that the refurbishments started by 158, it seems at first tempting to suggest that this prolonged building may reflect a period of military build-up in response to some perceived threat. However, few of these forts had seen any maintenance for nearly 25 years, and thus a 5 year programme of refurbishment is hardly excessive.
The next clear reference to warfare in Britain comes from the reign of Commodus (180/192). Xiphilinus, abbreviating Cassius Dio (72.8.1–6) reports as follows that Commodus spent some time fighting in Dacia (modern Romania) ‘but the greatest war was the British one. For when the peoples in that island, having crossed the wall, which divides them from the Roman legions, and done a great deal of damage, and cut down a certain general together with the soldiers which he had, Commodus, being alarmed, sent Ulpius Marcellus against them… such a man was Marcellus and he inflicted terrible damage on the barbarians in Britain’.
Diplomas or discharge certificates for Roman auxiliaries date Ulpius Marcellus’ presence in Britain to 178, while the conclusion of the British campaign has been linked by Tony Birley (2005, 162–170) with the seventh Imperial acclamation of Commodus of 184 and the issue of coins naming him Britannicus and celebrating Victoria Britannica (The British Victory (RIC III, Commodus 451)).
The wall that was crossed by the barbarians, on our current understanding of the withdrawal from the Scotland would have to have been Hadrian’s Wall, and Birley argues that the legate killed was not a governor (legatus pro praetore), but more likely to be one of the three legionary legates (legatus legionis), probably from York.
Identifying this incursion across the Wall archaeologically is a matter of much debate. Hodgson (2009 H W, 31–32) is happy to accept the burnt layers at the forts of Haltonchesters, Rudchester and possibly Corbridge as evidence for this incursion. Birley (2005, 167–8) on the other hand sounds a much more cautious note.
Archaeologically, as we have seen before, burnt layers are exactly that, layers of burning, and only rarely is the evidence good enough to establish if it was caused by deliberate fire-setting or accidental fires. Given the widespread use of combustible materials, accidents were frequent throughout Romano-British history, sometimes leading to widespread damage. On the other hand, three destructions of apparently similar date in close geographical proximity are likely to raise eyebrows. More evidence is clearly needed to resolve the issue.
Possibly contemporary with this trouble are a number of inscriptions from Hadrian’s Wall and its hinterland. An altar a few miles northwest of Carlisle was set up ‘in thanks for things achieved beyond the Wall’ by the Legate of legio VI Victrix (RIB 2034), but this may just as easily refer to earlier problems north of the Wall. A prefect in Corbridge set up an altar ‘after the slaughter of the Corionototae’, which has been interpreted as a possible early reference to the Picts (RIB 1140). A third monument, part of an arch, was set up in Carlisle to commemorate the defeat or slaughter of barbarians (RIB 946).
In addition to this trouble from beyond the Wall, the governor’s behaviour appears to have also generated a mutiny according to a fragment of Cassius Dio (72 (73), 9.2a) as the soldiers tried to appoint a legionary legate named Priscus as Emperor, which the latter declined. The SHA (Commodus 8.4) appears to refer to the same episode, when it suggests that Commodus was called by some Britannicus (indicating a victory in Britain), even though the Britons wanted a different Emperor.
The mutinous undercurrent appears to have lasted beyond the governorship of Ulpius Marcellus, as his successor Pertinax (governor 185/187), who would be Emperor in 193, also had to deal with continually insurgent troops (Cass. Dio 72 (73), 9.2a) and 73 (74), 4.1. – SHA Pertinax 3.5–4.1). This suggests that the situation in Britain under Commodus was anything but quiet for a variety of reasons, but once again there is very little detail available to judge the circumstances competently.
The situation appears to have come to a head on 31 December 192, with the murder of Commodus in Rome. As no male heir from Marcus Aurelius’ or Commodus’ line was available, the new Emperor had once again to be found outside the immediate Imperial family, leading as in 69 to multiple contenders, and in this case to a ‘year of the five Emperors’. Commodus’ immediate successor as Emperor was Pertinax, who lasted only three months, before he was replaced in the city of Rome by Didius Julianus. At the same time the largest armies in the Empire offered their own candidates for the throne; the armies of Illyricum and Pannonia supporting Septimius Severus, the armies of Syria and the Eastern provinces favouring Pescennius Niger, while the British legions cooperating with the German legions finally got their chance of nominating their own candidate; their governor Clodius Albinus.
Unlike in 69 the ensuing civil war took more than a year to resolve. The Roman candidate Didius Julianus was relatively quickly eliminated. Severus at first offered Albinus cooperation (by naming him Caesar). Albinus accepted, leaving Severus free to deal with the Eastern provinces and Pescennius Niger. Only after Niger’s death was Severus prepared to attack Albinus, who was eventually defeated on 19 February 197 at the Battle of Lugdunum/Lyon, a battle in which according to the SHA, 150,000 soldiers took part on both sides. These huge numbers are almost definitely inflated in keeping with Roman historiographic tradition. They suggest that British troops were likely to have been used in the battle, but there are no units named in the historical records. As it were, the involvement of Britain as the leading province in an usurpation was going to have repercussions for the province as a whole.