Why the Romans failed to Conquer Scotland?

D. J. Woolliscroft


Although the Flavian withdrawal from Scotland was probably largely caused by the withdrawal of Legio II Adiutrix and other units, following disasters on the Danube, other factors must have seriously exacerbated the effects of these force reductions, making withdrawal more inevitable and later re-occupation less vital. In particular, the relatively low state of native political development must have made the area both more difficult to govern and less of a threat if relinquished.

For centuries, considerable academic effort has been expended in attempts to identify the site of the Battle of Mons Graupius (St.Joseph 1978, 271-288, Keppie 1980, 79-88 and Maxwell, 1990). This is, of course, a worthwhile and fascinating exercise. Archaeology has been described as the handmaiden of history (although it is now unfashionable to say so) and in Mons Graupius it has a rare opportunity to look at a great historical event; a real, some would say decisive, battle. The temptation is simply too hard to resist. But, if we are honest, it has to be admitted that, in the great historical scheme of things, the exact location, or indeed course, of the battle, whilst interesting to the tactical historian, lies in its impact on the future development of Scotland, Britain and the entire Roman world and this is, frankly, minimal. Looked at simply as an incident in history, therefore, it is probably quite enough to know that Mons Graupius was fought somewhere in eastern Scotland, to the north of the Tay (although probably not in the far north) and that the home side lost.

More important, from the historian’s point of view, are the battle’s after effects and what both it and its aftermath can tell us about Rome and contemporary Iron Age Scotland. This means that, as well as asking where exactly the battle took place, we should also be looking more closely at issues such as: what exactly the Romans achieved by winning and who, exactly, was defeated. Because we have, in Mons Graupius, a most peculiar situation. Tacitus, in his Agricola (35-39), portrays the action as an historic victory for Rome and a cataclysmic defeat for the Britons, from which there should have been no real chance of recovery. Elsewhere (Tacitus Histories 1,2), he at least purported to be shocked by subsequent events, saying: “All Britain was conquered and then immediately thrown away”. In line with his theme of portraying Domitian as jealous of Agricola’s successes, he at least implies that the blame for the ensuing withdrawal rests with the Emperor, seeing it, probably somewhat disingenuously, as a virtual act of betrayal. Nevertheless, he does seem to have cause for astonishment. For here we have a supposedly great Roman victory in northern Scotland, at a time (83 or 84 AD (Birley 1976, 11-14, cf Hanson 1987, 40-45)) when many Romans still saw world conquest as their destiny, and yet only three or four years later (Hobley 1989) the Romans began a retreat, which was eventually to take them all the way back to the Tyne-Solway line: why?


Most of the energy put into studying the Flavian pull-back from Scotland has necessarily dealt with the detail and chronology of exactly how the withdrawal was phased (i.a. Robertson 1975, 364-428, Pitts and St. Joseph 1985, 279f and Hanson 1987, Ch 7), rather than with why it took place at all. But a range of possible explanations, both local and external, has been raised in the past and, although individually these have sometimes been somewhat simplistic, in that they look for a single cause for what must have been a complex process, they have recently been subjected to a more wide ranging review by Breeze (1988). For the most part, one can only agree with his conclusions. For example, there is no particular evidence that the Caledonians were exceptionally warlike or militarily effective and, even if they had been, the Roman army was used to conquering warlike peoples. Indeed, whenever the Caledonians met Rome in battle, from the time of Agricola onwards, they seem to have been beaten and so can have stood little chance of expelling the Romans by force. The Highlands too were not especially difficult terrain when compared to other areas that Rome had successfully assimilated and their potential as an environment for guerilla warfare can likewise be easily exaggerated. Besides, the southern uplands, and Pennines, which were retained, would both have been better guerilla country than the lowland areas and glen mouths which seem to represent the northern limit of Roman activity and, although guerilla warfare was practised in the ancient world, with occasional spectacular successes, its effectiveness was weakened in an age without firearms because, even with archery, its exponents still had to come to quite close physical quarters with the enemy.

The lack of a developed market economy in Iron Age Scotland, which has been suggested as a possible cause of Roman logistical difficulties (Breeze 1988, 12-15 and Groenman van Waateringe 1980, 1041), can also be exaggerated as a problem. The theory has been that the lack of local market mechanisms for managing surplus agricultural production would have made the army difficult to supply, but numerous societies, from Minoan Crete to the USSR have found alternative methods of organising, concentrating and/or distributing surpluses, albeit with varying degrees of efficiency and, as Millet points out (1990, Ch 7), it is probably anachronistic to look for a full market economy anywhere in Roman Britain.

Scotland may not have had the easily available precious metal resources that might have helped to attract the Romans into Wales (Jones and Mattingly 1990, 179-192), but the probability of low potential tax revenues from the area could also have been less of a problem than might at first be thought, at least so long as Scotland could be held by the existing army of Britain, because it should then also have involved little additional expenditure (Woolliscroft and Woolliscroft 1993, 56f). Indeed, given a sufficient degree of assimilation, money might even have been saved since, with no land borders with external powers to defend, overall force reductions might eventually have been possible in Britain.

Dr Breeze comes down in favour of an external factor as the primary cause of the pull-out: the dangerous crisis on the Danube, from 85 onwards, where major Roman defeats seem to have led to significant troop withdrawals from Britain, including the loss of one entire legion: Legio II Adiutrix. He argues that Britain would have been seen as a peripheral province and, one might add, perhaps ultimately expendable, from which forces could be drawn with the least injury to Rome’s vital interests, even if this might cause large scale policy dislocations within the province itself. The troop withdrawals would thus have forced Rome to fall back into areas which could still be held by the lesser forces now available and, although this may have been intended as a purely temporary expedient, thereafter the political will never again existed in a sustained enough form to permit a serious attempt at renewed total conquest.


But is this explanation sufficient in itself ? For, whilst one would not dispute for a moment that the loss of these forces was important, and quite probably does represent the key factor, Britain was still left with the single largest provincial army in the entire Roman Empire, with three legions and an abnormally large auxiliary force. Moreover, the Romans had had at least 2-3 years in the occupied parts of Scotland to consolidate their hold before the withdrawals began and in many other parts of the Empire, including parts of southern Britain, Roman control over regions, and even whole provinces, had been able to survive the running down of their garrisons remarkably soon after conquest. Indeed, had the conquest been completed, Britain might have been expected to have presented a prime candidate to be easily treated in this way, because again, as an island, it should no longer have required a great deal of defending; it simply needed to be kept under control. So, given the overall trigger of force reductions, were there other, internal, factors that made the withdrawal more necessary ? Or, to put it another way, were there additional considerations which made the force reductions more than usually devastating?

Although the Breeze analysis dismisses the terrain, economy and people of Iron Age Scotland as the principle causes of Roman failure, he would not deny that they did present difficulties which, although not insuperable in themselves, cannot but have worsened the army’s position. But the present writer would also like to lay a little additional emphasis on one other possible contributory factor: the state of political cohesion of the native population.

An interesting idea, first put forward by W. Groenman van Waateringe (1980) with relation to the Rhineland, and since briefly mentioned in a British context by both Breeze and Millett (Breeze 1988, 13-15 and Millett 1990, 54f and 99f), may have some bearing here and can be summarised as follows: It is well known that the Romans generally sought to avoid the need to leave large numbers of administrators in conquered provinces, at least for any length of time, and the military presence was also kept to a minimum, especially in areas away from the frontiers. Rome tended instead to take the local administrative, legal and law enforcement systems largely as she found them and simply turned them to her own ends, under the overall direction of the Governor and his minimal staff. This meant that, even after quite bloody wars of conquest, at least local elites were often left with their power and wealth substantially intact, and simply acquired responsibilities to the Roman state in taxation, judicial and order concerns which, in any case, they might already have been discharging under their pre-conquest government. In other words, Rome ruled most of her captive states not by the direct supervision or coercion of their populations, but merely by bending their governing classes to the imperial will and, even here, so long as all went well, the Empire frequently seems to have had a fairly light touch, at least until the late period. The deal was also reciprocal, often to the point of symbiosis. The local magnates not only survived; their position necessarily acquired imperial backing. They continued to run their communities, albeit on Rome’s behalf, and they could profit from doing so well, not only by preserving their local status and the semblance, and to a large extent the reality, of local power but, with time, the more able and ambitious might also now hope for still more profitable careers on the Empire wide stage and, through Rome’s almost unprecedented generosity with her own citizenship, they could aspire to become legally one of the conquering people, rather than one of the conquered. They thus acquired a vested interest in preserving the imperial power, rather than, as might otherwise have been expected, becoming natural focuses for resistance.

It was a brilliant system in many ways and, on the whole, it worked superbly well. It made efficient use of (and indeed eventually expanded) the meagre pool of scarce Roman manpower. It was relatively cheap to operate. It made local peoples feel less under the imperial thumb and the provincial elites, being small and readily identifiable, were more easily encouraged, suborned, communicated with and, if needs be, intimidated than entire populations. In short, it was a system that allowed Rome to govern an empire by, at least tacit, consent that she could never have ruled by force alone. The only problem was that it had grown up on the assumption of finding the sort of centralised, Mediterranean city state or kingdom style societies which possessed the necessary political infrastructure to make them capable of being left to run themselves in this way. Unfortunately (for Rome), as Millet (1990, Ch 2-4) points out, although at least the beginnings of such centralisation had already formed in southern Britain at the time of the conquest, they had probably not in the north, and especially in northern Scotland.


Scotland certainly shows little sign of the outward trappings of such development. For example, with the exception of a very few, poorly dated hill forts (1) in the south, (mainly in the territory of the supposedly pro-Roman Votadini, and all south of the Forth-Clyde line) the settlement pattern is largely one of dispersed individual farms and small, possibly extended family, communities (brochs and the like) with little in the way of a settlement hierarchy (Hingley 1992, 7-53). The recent discovery from the air of one or two only slightly larger defended sites in the Moray area (Jones, Keller and Maude 1993, 52ff) may make the situation a little more complex, but these are undated and even if they do prove to have been occupied in the relevant period, they may only show the existence of minor local lordlings, able to command building labour beyond their immediate entourage, or perhaps just the presence of slightly more organised communities. There also seem to have been rather fewer “high status goods” in circulation than in the south and we see no signs of the development of centralised institutions such as coinage and formalised monarchies. In other words, although communal activity and so, presumably, communal feeling did exist, there is little to suggest the existence of all but the most embryonic central authorities within the individual tribes or clans. The suggestion has been made, therefore, that the Romans may simply have found the socio/political state of Scotland to be too incompatible with their established system of provincial governance to be either workable or, at least, worthwhile, given a situation in which serious force reductions were already compounded by the other difficulties mentioned above. For it is extremely difficult to coerce a society made up essentially of individuals or, at least, small, near independent, extended family groupings. How, for example, does one even communicate one’s wishes to such people (Barrett and Foster 1991, 46), let alone enforce them ? Indeed, it is not too much of an exaggeration to say that an invading army could pass through such a society almost without being noticed. For, as Maxwell (1975, 31-49 and 1986, 60-63) has pointed out (2), how does one strike against an enemy who has no easily identifiable seat of power ? In extreme cases, even basic socio/political concepts such as power, authority (at least as wielded from afar) and, perhaps especially, taxation might have been difficult for the natives of such a society to comprehend and still harder for them to accept. For, unlike the populations of more “advanced” societies, to whom Roman conquest might often have meant little more than the replacement of one ruler by another (whose direct representatives they would, anyway, rarely if ever have seen), a decentralised society would have been asked to accept rule and complex administrative procedures where none had existed before; a very much more difficult transition to make.

If we, therefore, assume the absence of any pre-installed local apparatus to do their job for them and, perhaps more importantly, of the habit of obedience to such authorities in the minds of the natives, the Romans would have been faced with just three possible courses of action. Firstly, they could have allowed Scotland to swallow the huge investments in manpower and matériel that would have been necessary to maintain a sufficiently close level of supervision over the scattered population of an entire country and to administer any form of taxation and judicial system. This would have been a difficult enough task given the most docile of inhabitants, but even if the Caledonians proved incapable of sustained and co-ordinated guerrilla resistance, smaller scale obstructionism (and simple incomprehension) could have been expected, perhaps running to the level of localised Partisanenkrieg or Franktireur as individuals and local communities tried to protect their own interests. Quite small bands of such malcontents would have been able to exert a disproportionately disruptive influence because the possibility of ambushes, or the risk of providing parties of native warriors with a purely local superiority, would have made it difficult for the Romans to detach small groups. This would have hampered numerous essential military and administrative activities such as the taking of a census, communications, procurement, and tax collecting, all of which would, thus, have been turned into unusually manpower intensive operations, requiring forces which, in the absence of Legio II Adiutrix (et al) the Romans may no longer have had. The situation would also have demanded large numbers of small, and potentially vulnerable, towers and fortlets to act as policing posts, and a considerable program of social engineering, which the Romans may have been unwilling and ill equipped to take on. Alternatively, they could have taken Tacitus’ words to heart and quite literally created a desert and called it peace, something that was likely to be costly in Roman as well as native lives, regardless of the moral issues involved and again certainly, if more temporarily, manpower intensive. Or, thirdly, they could simply have decided to cut their losses and pull out.


These ideas can be expanded upon and taken back to Mons Graupius and even slightly before, although it has to be said that it is always dangerous to put too much faith in the Agricola and the writer’s own deductions should be seen in that light. It is very noteworthy that Tacitus does not refer to any other specific battles in Scotland before this action except for a night raid on the camp of Legio IX Hispana in Agricola’s sixth season. He does say that, in the fifth season, Agricola advanced through “repeated and successful battles” (Agricola, 24), but apparently did not consider any of these to be worthy of individual mention. This could, of course, have been simply a literary nicety, as constant battle descriptions might have detracted from the work’s current clean build up to the climax of Mons Graupius, especially if some of them had not been quite the successes claimed. Yet letting Legio IX get caught as it was hardly reflects well on Agricola’s generalship and Tacitus does refer to this, albeit whilst trying to present it in as good a light as possible, and it is difficult to believe that Tacitus’ apparent hero worship of his father-in-law would have allowed him to let any other significant engagements, let alone real successes, go unmentioned. Instead we have a rather curious situation. Military operations are conducted against people, not simply terrain and yet, as things are, Tacitus and, following him, many modern accounts read as though Agricola was merely marching through an empty landscape with the discussion centring almost entirely upon which landmarks he reached in which season. The attack on Legio IX is interesting, however. Tacitus specifically says that the Britons attacked the sleeping camp in full force. They had previously been operating in three groups, which is why Agricola had divided his own forces into three (although, to judge from the story of the Roman rescue action the columns must have been in contact), but the natives had then regrouped (Agricola, 26) and Agricola had not yet responded. The result was a situation in which the full force of the Britons attacked a single legion (no doubt plus some auxiliaries) in its camp. They seem to have achieved total surprise and actually got into the camp whilst most of the legion was still asleep. We are told, without explanation, that Legio IX had been chosen as the target because it was the weakest in Agricola’s army (and for once we might be able to corroborate Tacitus, since there is epigraphic evidence (ILS 1025) that the legion may have had a vexillation operating in Germany at about this time, further weakening the army of Britain) and yet it survived. Certainly, there was a rescue operation, which no doubt helped and may well have saved the day, but the fact remains that, when the relief force arrived, Tacitus tells us that the IXth was already up and fighting and seems to have been driving the Britons out of its camp unassisted. This begs a question: what kind of force gets itself beaten with so many advantages in its favour ? If this really was the full British force, it was either hopelessly incompetent, or remarkably small. Yet Tacitus specifically says (Agricola, 26) that once it was retreating from the camp “If it hadn’t been for the forests and marshes that covered the fugitives, that victory would have ended the war”, as if this force was all that the natives could have been expected to be able to muster. Chapter 27 of the Agricola does then go on to refer to further native recruiting and (only now) to a confederacy of action between the native tribes, but this is in preparation for Agricola’s 7th season, i.e. shortly before Mons Graupius, whereas the Romans had been operating in Scotland for several years. Assuming that Tacitus is giving us a true account, therefore, were the natives only now taking Agricola seriously, or had they only now developed, under stress, the political ability to do anything, of any moment, about him ?


We now come to Mons Graupius itself. As we have already seen, this is the first and only pitched battle of the entire Scottish campaigning sequence, or at least the only one that Tacitus thought worthy of description and, again, this in itself is interesting. In the South, Caesar had literally had to fight on the beaches, whilst the Claudian invasion had faced full scale battles soon after advancing inland. Yet, if Tacitus is to be accepted, Agricola may have operated in Scotland for three or four years before he faced a major pitched battle. Why had there not been such an action before and why was there one now ? No doubt, as in Gaul and southern Britain, some tribes would have allied themselves to Rome without a fight, but others would not. Had the Britons been operating some sort of careful delaying strategy to wear the Romans down, knowing they would be defeated in a full battle ? If so, there is no real sign that such a policy had ever been used before in Britain. Previously, the natives always seem to have struck, and been beaten. Indeed, the northern barbarians as a whole were generally famed for reckless bravery, for leaping impetuously into action and, of course, for running away just as quickly if they didn’t like the look of the way things were going. Furthermore, if this highly original, and culturally out of character, strategy had been adopted, who had formed and implemented it ? Was there some sort of Celtic Fabius Cunctator in Scotland, carefully restraining his people (whoever they were) and biding his time ? If so, why did he pick Agricola’s seventh season to finally permit a battle ? What had made him, or indeed (in a Celtic context) her, think that a battle might now work or, at least, to make him think that the delaying strategy had failed and/or that battle was no longer avoidable ? As with Fabius, had his own side decided no longer to tolerate delay ? Or are these completely the wrong sort of questions to be asking ? Was it simply that this was the first time that the northern tribes/clans had managed to get themselves into a sufficiently organised condition to put a combined army of any real power into the field ?

Tacitus says (Agricola, 29) of the period immediately before Mons Graupius “For the Britons unbroken by the outcome of the previous battle (singular) and seeing before them vengeance or slavery and learning at last (again only now) that a common danger must be repelled by union, had brought into the field, by means of envoys and treaties, the flower of all their states. Already more than 30,000 armed men were on view (at Mons Graupius) and still the stream flowed in”. This does not sound like the same army we have seen before attacking Legio IX. Even given the tendency of ancient writers to exaggerate enemy numbers, in order to stress the magnitude of their own side’s achievements (no doubt amplified in this case by Tacitus’ desire to portray Agricola to the full as a hero), it is hard to imagine even a fairly poor quality army of even approaching this size being thrown out of a camp by one half asleep legion. We are told that the Britons were already occupying Mons Graupius before Agricola advanced to the site to give battle. Furthermore, British warriors were still “streaming in” when he arrived. It would seem, therefore, that somehow, and we can only guess at the political and diplomatic processes involved, Mons Graupius had been chosen as a rallying point at which the natives had assembled and, indeed, were still assembling when the Romans appeared. In other words there had not previously been any form of single native army as such. For the army now gathering at the battle site had never marched together as such, had never fought together as such, indeed had never operated together in any way whatsoever, as such, because it had simply never existed before, as such.

If we accept this as, at least, a reasonably realistic appraisal of the the state of the native army, the next question must inevitably concern its command structure. For who, if anyone, was in command of this impromptu assembly ? The answer that has always been assumed is Calgacus, but was he ? Tacitus never says so. Calgacus is never recorded making dispositions, as Agricola was, nor is any action of his during the battle recorded. Indeed, we do not even learn his fate. The only thing he is ever said to have done is to make a (no doubt fictitious) pre-battle speech and after that we hear no more of him. Not only is he never said to be in command, he is never referred to by any term, such as king or general, that would have allowed us to take this as read. He is simply introduced as “distinguished by birth and valour amongst many chieftains was one called Calgacus”. Tacitus goes on “To the gathering host demanding battle he is reported to have spoken along the following lines”, then comes the speech, then nothing. This is unusually coy. If there had been a single Caledonian commander, the Romans would surely have found out who he was, even if only from prisoners after the battle. Even in an event such as the Varrine disaster, where hardly a Roman lived to tell the tale, Rome had no doubt who she had been fighting, so why such uncertainty here ? Furthermore, it was quite usual for Roman writers to take an interest in the actions and eventual fates of barbarian commanders, for example Vercingetorix, Boudicca, Caratacus and the like, so why not Calgacus ? True, it was normally regarded by the Romans as both the prerogative and the duty of a commander to make a rousing speech before an action, so that the entire speech episode may be Tacitus’ way of pointing out Calgacus’ position, but this need not be conclusive. So was Calgacus in command, either officially (if such a term can be said to apply here) or just by virtue of some sort of informal, for want of a better word, auctoritas or strength of character ? Or, was he just one of many petty “clan” or group leaders present, perhaps just a name that happened to come to the Romans’ attention as someone important. Perhaps, simply a man whose renown and character made him someone the other chieftains, and possibly their men, would listen to and whose ideas would carry above average weight and possibly, since Tacitus is so coy, a man who also eluded Roman capture.


What all this can be reduced to is that we have at Mons Graupius a British army formed ad hoc solely for the event, which was merely a gathering at a particular, presumably well known, landmark, which may have had no overall commander and whose survivors, after the battle, simply melted away, making no attempt (according to Tacitus) to re-group to fight again, despite facing a Roman army at the end of both the campaigning season and its supply lines. None of this looks like a military force drawn from an organised, structured society and, if we are not to dismiss Tacitus’ account as mere hyperbole, this force had even been beaten by the Roman auxillia alone without the legions becoming engaged. This may, of course, be an indication that the native army was not as large as Tacitus would have us believe, and so the battle, and with it the victory, was rather less impressive. But, if so, this too reflects badly on the natives’ standard of organisation although, to be fair, we have no idea of the size of area from which their forces were drawn.

This brings us back to where we started. After the battle the Romans began building and consolidating, obviously with every intention of staying and the withdrawal, when it came, does seem to have been both sudden and unexpected. For we find military installations, such as the legionary fortress at Inchtuthil, abandoned whilst still being built (Pitts and St Joseph 1985) and plentiful signs of hurried abandonment including again Inchtuthil and (although the dating here is uncertain) the Gask Ridge fortlet of Midgate, (Woolliscroft 1994, 307) which was abandoned whilst its ditch was in the process of being cleaned out. At the same time, the evidence does suggest that the Romans pulled out of their own accord, rather than being expelled, for the archaeological signs are mainly of careful demolition and the hiding of surplus matériel, and not of enemy action. There is also evidence that initially the Romans had planned to advance further. Tacitus believed, or at least proclaimed, that Agricola had finished the conquest of Britain, but this is manifestly not the case. However great the victory at Mons Graupius, so far as we know, the Highlands had barely been touched and, as later English, and indeed Scottish, armies were to discover, no one can truly claim to hold Scotland who does not hold the Highlands. For example, Edward I reached as far as Agricola in a single campaign, with just as little lasting effect. But the so called Glen blocking forts have never been entirely convincing as a defensive line (Mann 1968, 308), especially as they seem to have made no provision for obtaining observation cover in the glens themselves (Woolliscroft 1994, footnote 6). Likewise the presence of a legionary fortress (Inchtuthil) right on the frontier line is unusual, except in other areas, such as the Rhine and Danube, where expected advances also never materialised (3). The entire system is thus at least as well explained as the base line for planned future advances into the Highlands than as an outward, or even inward facing (Hind 1983, 377), containment system and might also be seen as the truncated fringe of what was to have been an integrated occupation deployment pattern which, if events had turned out differently, would later have been extended throughout the rest of the country.

What then caused the crash ? This paper began by fully accepting Dr Breeze’s view that the withdrawal was triggered when the loss of Legio II Adiutrix (and whatever auxiliaries accompanied it) left the Romans with insufficient forces to hold and extend the new conquests. But, the writer would argue that the fact that what was still the largest provincial army in the Empire (4) was found to be insufficient, when elsewhere geographically more difficult areas were being held by much smaller forces, may be at least in part due to the state of the indigenous society.

During the post Agricolan consolidation, the Romans would have been finding out in much more detail than Agricola himself can have had time to, just what sort of people they had conquered. They had won themselves a decentralised society with little or no experience of the sort of governance that Rome was geared up to expect and, indeed, depended on to operate through. The region was, therefore, going to require an abnormally large investment in manpower just to keep quiet and to collect what taxes it might have been able to raise. Suddenly, this manpower was no longer available.


There are even signs of such difficulties in the archaeological record, in the form of a rather abnormal deployment approach. It is often suggested that the fact that a considerable number of forts remained in use on Brigantian territory, to the south of Hadrian’s Wall, means that the Brigantes always had to be held down by force, throughout the Roman period. But, in fact, the dispositions involved in Brigantia are not what one would expect for an army attempting to control a reluctant, even mutinous population. The men were mostly stationed in relatively few quite widely dispersed forts, whereas one would have expected a real army of occupation to have needed a lot more smaller policing posts, largely fortlets or even watch towers. These are singularly lacking in Northern England, except over Stainmore, and the forts are simply sited in places where they could be supplied fairly easily, and probably existed largely because the frontier army could not, and strategically should not, all have been put on the Wall (5). In Scotland, however, things are different. Much more use is made of isolated fortlets and towers, as if areas were being watched in detail, and if this is true of the Flavian period, it is even more so of the Antonine re-occupation, in which the lessons of Flavian times may well have been learnt and taken to heart.

The scenario envisaged here, therefore, would be that just as the Romans were coming to understand the true situation confronting them, in what was always going to be a difficult area, their available forces were cut, as units were withdrawn for service elsewhere. Under normal circumstances, other newly conquered, but better developed, areas had been able to deal with such demilitarisation successfully. But, these were not normal circumstances for Rome, because there were no local institutions able to shoulder the burden, or which could be persuaded that they had a vested interest in trying. At the same time the lack of centralised authorities may have made an unoccupied Scotland seem less of a threat to the rest of the province, making withdrawal a more acceptable risk to take. Under these conditions, the Romans may simply have decided that they were no longer in a position to stay or, at least, that any likely benefits were not worth the additional dangers and, therefore, that there was not enough to be gained by staying. A more politically developed Scotland, on the other hand, might have been both easier to hold and more dangerous to set free, so that whether or not the military situation on the Danube may still have compelled a temporary withdrawal, the Romans might later have felt far more strongly motivated to return.

The Roman Gask Project
University of Manchester



BARRETT, J .C. and FOSTER S. M. 1991 ‘Passing the Time in Iron Age Scotland’ in HANSON W.S. and SLATER E.A. (ed), Scottish Archaeology, New Perceptions, Aberdeen, 44-56

BIRLEY A.R. 1976 ‘The Date of Mons Graupius’, LCM, 1.2, 11-14.

BREEZE D.J. 1988 ‘Why Did the Romans Fail to Conquer Scotland ?’, Proc Soc Antiq Scot, 118, 3-22.

GROENMAN van WAATERINGE W. 1980 ‘Urbanisation and the North-West Frontier of the Roman Empire’, in HANSON W.S. and KEPPIE L.J.F. (ed) Roman Frontier Studies 1979, Oxford (Brit Arch Rep Int Ser, 71, vol 3), 1037-44.

HANSON W.S. 1987 Agricola and the Conquest of the North, London.

HIND J.G.F. 1983 ‘Caledonia and its Occupation Under the Flavians’, Proc Soc Antiq Scot, 113, 373-378.

HINGLEY R.C. 1992 ‘Society in Scotland From 700BC to AD200’, Proc Soc Antiq Scot, 122, 7-53.

HOBLEY A.S. 1989 ‘The Numismatic Evidence for the Post-Agricolan Abandonment of the Roman Frontier in Northern Scotland’, Britannia, 20, 69-74.

JONES G. D. B., KELLER I. and MAUDE K. 1993 ‘The Moray Air Survey: Discovering the Prehistoric and Proto-Historic Landscape’, in SELLAR W.D.H. (ed) Moray:Province and People, Edinburgh, 47-74.

JONES G.D.B. and MATTINGLY D. 1990 An Atlas of Roman Britain, London.

KEPPIE L.J.F. 1980 ‘Mons Graupius: The Search for a Battlefield’, SAF, 12, 79-88

MANN J.C. 1968 ‘Cornelii Taciti: De Vita Agricolae by R.M.Ogilvie and the Late Sir Ian Richmond’, Arch Ael Ser 4, 46, 306-309.

MAXWELL G.S. 1975 ‘Casus Belli: Native Pressures and Roman Policy’, SAF, 7, 31-49.

MAXWELL G.S. 1986 ‘Sidelight on the Roman Military Campaigns in North Britain’, Studien zu den Milit_rgrenzen Roms III (Proceedings of the 13th International Congress of Roman frontier Studies), Stuttgart, 60-63.

MAXWELL G.S. 1990 A Battle Lost, Edinburgh.

MILLETT M. 1990 The Romanization of Britain, Cambridge.

PITTS L.F. and ST JOSEPH J.K. 1985 Inchtuthil, London.

ROBERTSON A.S. 1975 ‘The Romans in North Britain: The Coin Evidence’, ANRW, II.3, Berlin, 364-428.

ST JOSEPH J.K. 1978 ‘The Camp at Durno and Mons Graupius’, Britannia, 9, 271-288.

WOOLLISCROFT D. J . 1994 ‘Signalling and the Design of the Gask Ridge System’, Proc Soc Anitiq Scot, 124, 291-313.

WOLLISCROFT D .J. and WOOLLISCROFT J. 1993 ‘The Profits of a Loss Making Roman Province’, LCM, 18.4, 56-57.


1. It could be argued that the absence of large hill forts in northern Scotland might reflect the nature of the country as much as the degree of advancement of its people. There are, after all, few large cities north of the Forth-Clyde line even today, but no one would suggest that the area is either uncivilised or politically primitive. It merely reflects a lower population density and a lower economic imperative for such settlements.

2. It is true that disunity between groups can (and in Scotland frequently has) be exploited by an invader to facilitate a divide and rule policy, but even this requires reasonably large and monolithic groupings with which the invader can deal.

3. It is interesting to note that the Rhine line legionary fortress at Mainz did not move forward as a result of the more or less contemporary advance to the German Limes.

4. It should, be remembered that, as an island, Britain would have been difficult to reinforce.

5. Brigantia does seem to have been rather more centralised. It did, after all have a queen in Cartamandua who does seem to have been able to control her tribe, at least temporarily. It also contained hill forts and other signs of proto-urbanisation. Brigantia was also a former client and, to judge from the recent early datings at Carlisle, it may have been relatively easily overrun.


A long term research project to study the Romans north of the Antonine Wall