Scottish history and heritage online

HomeArticlesMons Graupius Identified

Mons Graupius Identified

The hunt for ancient Scotland’s great clash of arms in 83 AD

 “…which will you choose – to follow your leader into battle, or to submit to taxation, labour in the mines, and all the other tribulations of slavery? Whether you are to endure these for ever or take quick vengeance, this field must decide. On then into action; and as you go, think of those that went before you and of those that shall come after.”

“…and there would be glory too, in dying – if die we must – here where the world and all created things come to an end.”

Introduction

War was a fundamental aspect of Roman culture, especially where its expansionist policies met with local resistance. This warfare often could be – to use the modern term – low intensity. Such an undercurrent of simmering resentment – with the risk of violence threatening – appears to have been much the norm for most of Rome’s involvement with the tribes of ancient Scotland.

At times however the historical record is punctuated by sporadic outbursts of open warfare in Scotland, violent and bloody, and of these the best known is the battle of Mons Graupius.

The lead extracts above are Tacitus’s “pre-battle speeches” at Mons Graupius, taken from his work; the Agricola. Although these speeches are almost certainly pieces of literary invention they reflect the issues and concerns (as Rome’s elite saw it) for the Caledonian tribesmen beset by an invading superpower, and following that the Roman rallying call for high adventure with its attendant glory in Scotland, then commonly considered to be at “the end” of the known world.

Historical Summary

The battle of Mons Graupius – the first and indeed best “recorded” battle on Scotland’s ancient soil took place in Iulius Agricola’s last year in office as the governor of Roman Britannia – probably in 83 AD (though 84 AD is considered an acceptable alternative by many).

The extent of the Roman province of Britannia had grown gradually since Claudius’s invasion in 43 AD and by 82 AD, with Agricola’s conquest of Scotland below the Forth- Clyde line now fairly secure, he looked north to the remaining free lands most probably with a view to exploiting natural resources to be found there.

The peoples of these lands, split into many lowland tribes and clan-like septs within the highland massif were generalised by the Romans -and have since come to be recognised- as the “Caledonians”.

In 82 AD therefore Agricola marched his army north and, as in 79 AD when he marched into southern Scotland, he did so intent on conquest, in search of glory, and above all seeking assimilation.

Unable to bring these truculent tribes to early submission, Agricola undertook a wide ranging burning campaign of “frightfulness”- a fairly textbook Roman tactic when dealing with volatile and easily provoked Celts. This campaign may perhaps have extended as far north as the Mounth near Stonehaven.

This calculated intimidation was aimed to force the resolve of the tribes one way or the other. If submission could not be achieved through compliance then the standard Roman approach was to achieve it through open conflict. The aim would be to achieve this in open battle – where the disciplined, well equipped and above all superbly organised Romans usually held all the aces.

Under such circumstances Roman commanders – we must remember they were all politicians as well – could not only achieve their planned conquest but in the manner of doing it secure that most precious of commodities; glory.

In 82 AD, however, the tribes of the north neither submitted to Agricola nor hazarded open battle against his army, the sheer size of which they would have found staggering. Simply put they will never have seen its like before.

On the other hand, what the tribes did do was undertake the age-old and highly effective tactic of guerrilla warfare – probably at this stage this was limited to the activities of separate tribal elites with their fighting tail of armed retainers.

Here the dispersed units of an overwhelmingly powerful invader were targeted and engaged, usually successfully. We are told for instance that they descended on and stormed various Roman outposts to Agricola’s rear while he was (probably) operating in Strathmore.

The effect of such attacks, snapping at the fringes of Rome’s forces, while not individually particularly noteworthy did however taken as a whole have -as elsewhere in history- a witheringly corrosive effect on the morale of the numerically superior force being thus attacked. It was all the more damaging when it happened along lines of communications in the supposedly safe “rear”.

To counter this threat, Agricola – we are told – split his army into three separate battlegroups to better control the routes his opponents were using. This is a clear reference to the glens which afforded rapid movement within the highland zone and which provided the ready access from which the tribal warbands could spill out onto the lowlands where the Romans concentrated their efforts. In Scotland, Roman commanders appear to have been extremely reticent to hazard their infantry columns by entering the highland massif.

The Caledonians displayed both an ability to work in concert and considerable canniness for they then concentrated their forces, targeted the weakest Roman battlegroup- that of the IXth legion – and in a night assault almost succeeded in overrunning the beleaguered Romans before withdrawing in the face of Agricola’s timely relief column.

Following this serious reverse, which either happened at the very end of the campaigning season or perhaps was of sufficient seriousness to convince Agricola to try another approach he regrouped and retired south to winter quarters.

Next year, 83 AD Agricola – back in Scotland in strength – received word that the Caledonians – now a formally bound tribal federation – were mustering their armed strength at a location known as Mons Graupius; we are told by as many as 30,000 men. Marching hard Agricola came upon them and threw up a camp adjacent to Mons Graupius.

The Caledonians do not seem to have been the least disheartened by these provocative manoeuvres and declined to decamp through the night. Deploying from his camp the next day Agricola launched an assault on the Caledonian ranks on the hill, and by careful deployment of a Roman cavalry reserve at a critical stage he overcame the flanks (and probably the rear) of the Caledonians who at that stage were threatening to envelop the Roman line. This was the critical turning point in the battle which led to a general Caledonian breakdown and in Tacitus’s verdict a rout. From the wreckage of battle, the Romans claimed only 360 slain to an alleged 10,000 of their enemy.

Thereafter Agricola retired south through the lands of those he had conquered in his years as governor and within a short period, with his tenure as governor lapsed, left Britannia for Rome, never to return.

The heady events of these years have long inspired generations of readers and it is of little wonder that many have sought to identify the site where these events unfolded.

Indeed the task of identifying the location of Mons Graupius has become to many the “Holy Grail of Scottish Antiquity” and it is to that noble quest that we will now turn our attention.

To date, the near-universal approach has been to champion a particular site, usually on account of some personal link between the author and the land in question – such as ownership or being resident there – or for academics and archaeologists wishing to magnify the importance of their recent nearby archaeological discoveries.

Further, long-lasting and pervasive confusion is passed down to us by antiquarians who attributed fine Roman remains to Agricola simply because he was better known than any other Roman who campaigned in Scotland. The simplistic logic being that they had to be his, and to boot, what better event to link such remains with but his crowning glory at Mons Graupius?

The tangible result of this veritable bouillabaisse of contending sites – and the egos attached to them – is a map of Scotland exhibiting the effects of a blunderbuss discharge, it being peppered with locations championed by so many wildly different and in many cases distinctly partial readings of the available sources.

More recent investigation however has assisted in pulling together information on the nature of these early claimants and one recommended read by Maxwell; “A Battle Lost” has looked at a range of these contenders. The books scope though did not extend to covering all sites nor in providing a measurable review of all these locations. To do so in a fair and even manner, as methodically as possible, and widely available to all has been lacking to date.

This brief summary places the events of 82 and 83 AD and its later appreciation in some context. First we shall look at the sources available to us to further understand these events in greater detail.

The sources available to us

Much of ancient history is lost to us and at best we are forced to rely on and interpret a few short, terse and often fragmentary references.
The challenge is to correlate these primary sources against the often meagre findings of archaeology.
As for the actual event – the battle itself – even though it was a terrible, traumatic and bloody affair it lasted only a few hours. Archaeology’s ability to pick up on the fleeting traces of these calamitous hours is notoriously limited.

It is further hindered in Scotland where a damp climate and acidic soil often contrive to destroy much of the buried evidence.

That much of this long-decomposed evidence is now lost to us however does not mean that the events did not take place, simply that we must bring four important factors to bear when trying to understand any given battle.

The first and most important of these factors are the written primary sources. These are what the ancients themselves recorded of the events. Secondary sources are modern interpretations. The best secondary sources are those that work from information in the primary sources. The less reliable are those that look merely to other secondary sources for inspiration.

The second factor is the understanding of these events that archaeology can provide us with. Concerning Roman campaigns archaeology illuminates the fossilised footprints of Roman armies on the march – the remains of the overnight marching camps Roman soldiers bivouacked in when on campaign. Archaeology has enabled us to better understand the structure, form and sequence of use of these fleeting remains.

The third factor is the critical ability to read the ground at any of the many proposed locations and to understand how this would be used (or indeed be useless) at a tactical level as well as reviewing how suitable the ground would be for the recorded events we are told took place.

On the grander strategic level we must further test each site against our understanding of known – and proven – Flavian activity at strategic level, set against the backdrop of probable tribal areas and the tensions these would have created.

The fourth factor is an ability to demonstrate a link between a sites modern name with the historic name of Mons Graupius.
Clearly this is no easy task as language and names evolve over time, and in many cases old names are lost under the names imposed on a location by later conquering elites. Therefore a site not having any meaningful link does not necessarily stand heavily against it. However any sites that can demonstrate such lineage are of exceptional interest.

Summary

It can be seen therefore that grandly proposing a site for the honour of having been the site of the battlefield of Mons Graupius is not simply a matter of pulling a metaphorical historic rabbit out of a magician’s hat.

On the contrary any contender must necessarily fully exercise its credentials against the four factors noted above.
To do so this we must test each site against a series of common benchmarks based on these four factors which then allows analysis to be made on the strengths and weaknesses of all the contending sites.

Only then can we compare these findings with the aim of identifying the site which based on the evidence we currently have best merits the accolade of being recognised as the location of the battle of Mons Graupius.

Factor 1. The Primary Source: Tacitus’s The Agricola.

By 82 AD the governor Agricola had already spent three long years campaigning and subduing the tribes of what is now southern Scotland.
In these years the Imperial directive he had received – Governors followed and enforced Imperial foreign policy, they did not create it – had undergone a staccato beat of stop followed by start followed by stop caused by a rapid succession of Emperors to power.

First the old warhorse Emperors Vespasian died and he was succeeded by his son Titus who himself soon followed, dying in suspicious circumstances shortly thereafter. Upon this he was succeeded by his younger brother Domitian.

It is a matter of debate and interpretation as to exactly why Agricola cast his eyes north of the Forth Clyde line in 82 AD. Some reckon the new Emperor will have wished to see military success auguring his ascendancy to power. Others reckon that Agricola – who must have realised that his own extraordinarily lengthy tenure in the post of governor would soon be nearing its end – strove for a confrontation with the fabled Caledonians. In the provocative manner in which he did so he actively initiated the chain of events where military success in battle could be pursued and the attendant glory secured.

After his death in 93 AD the life and achievements of Agricola were recorded in a funerary eulogy by the noted orator- his son in law and future senator- Tacitus in or around 98 AD.

Luckily Tacitus work has survived substantially complete and in it he dwells in some detail on his father- in- laws Scottish campaigns and in particular on the battle he won against the Caledonians at Mons Graupius.

It is therefore a unique source of information for those keen to understand the Battle of Mons Graupius better and required reading for those who would take the well trodden path of trying to identify the location where the fateful encounter took place.

Modern revisionist historians attempt to muddy matters. Some question if the battle even took place, some others take the unsustainable view that “we should grow up and throw out Tacitus”.
Rarely has such detailed primary Roman material been so haughtily dismissed!

What is certain is this. As a funerary eulogy a degree of prudence should be exercised in many respects concerning the fine detail contained within “The Agricola”. The work was originally designed to mark and indeed praise the life and deeds of the man Iulius Agricola and as Cornelius Tacitus was “family” and actively undertaking the cursus honorium he stood to gain in no small measure through any reflected prestige the work engendered.

Tacitus is remarkably forthright in describing this work as a piece of “filial piety”, and while it boastfully extols the virtues and deeds of the great man in a way we are not used to in this cynical modern era, it should be borne firmly in mind that this is simply what all ancient eulogies did.

There was nothing extraordinary or out of the normal in what Tacitus undertook to do to the reputation of his father in law.

Rare were such works that dwelt on the failures and examples of poor judgement made by the deceased, or those that took time out to praise the deeds and achievements of those other than the subject. Those would be expected to be covered in other types of work or indeed in their own eulogies.

By not recognising this, modern scholars have criticised Tacitus’s work – a “garrulous embroiderer” to one. Interestingly Tacitus describes such attacks on eulogies in ancient times hinting that the same may be expected for his.
Insightful prophecy indeed!

Many people struggle with the dated style of Tacitus’s diction, and the quantity of stock standard phraseology and metaphor contained in it which makes the work appear alien to the modern reader.
It is a work of its time and we should not allow this to deflect us from the core story within the text, while exercising caution in fully accepting some of the grander claims made there. For example the numbers of enemies slain as well the great mans ubiquitous energies and abilities. He is unlikely for instance to have, as claimed, planned all and every fort sight that was built during his term in office.

Other parts, notably the pre battle speeches are pieces of pure theatre and most certainly literary invention designed to fill out the plot and raise the tempo of the piece in anticipation of the battle. Sadly, Calgacus, the Caledonian warlord is most probably an invention too, fabricated to create an individual to enable the Roman public to identify with the opponents leadership – it would in reality have been a council of tribal elders from the many tribes present – and as an individual counterpoint to Agricola himself.

However what has to be borne in mind is that the work is a piece of “oration”. As such it was intended for verbal public delivery to the ordinary citizens of Rome, not merely published as reading material restricted to the salons of the elite in Rome.
Its compositional style and the stock standard phraseology it commonly uses is also readily found in other ancient works and would not be thought at the time any the less of for it.

Those who criticise it for a lack of fine detail have not recognised that the piece was aimed at an audience of the ancient era which was not particularly interested in forensic levels of detail on foreign landscapes and the complex military manoeuvres that took place there.
For the time the contents were sufficient to give a context meeting its audience’s expectations, while the latter would slow the pace of a narrative which Tacitus clearly strove to keep both fast moving and interesting.

Critically, we must not lose sight of the fact that many of those who would have heard the piece at the time will have been intimately involved in the events being described.

We can reasonably expect them to tacitly wink an eye but at the same time still nod along in agreement with some of Tacitus magnified boasts as clearly this would reflect well on themselves by association.
What we cannot however expect is for them to have accepted a fiction which was a fabrication of complete fable no matter how skillfully it was woven within a web of clever wordsmithing.

Cornelius Tacitus work “The Agricola” therefore is something we dismiss out of hand only at our peril and its survival to the modern era is a rare and great boon.

The Agricola

The chapters which deal specifically with the events we are concerned with are those between 25 and 38.

  • Note; Click on the chapter heading for a separate box to come up, read the Tacitus text then revert to our analysis of key aspects of the text –

Tacitus Latin wordplay is of an old style and at times some of the words and phrases have an uncertain translation to a modern one. The translation therefore in modern times can appear a bit “flowery”.

If we strip this baggage off though, and accept that some modern words used in the translations are “best fit” approximations to words and phrases understood in common usage at the time, we can glean a pretty accurate flow of the events that took place.
We merely have to exercise caution over the slant Tacitus placed on the events when recording them.

One thing is certain, as noted above Tacitus’s aim was to set the reputation of Agricola against real events but record him for posterity in the most favourable light.
We must show therefore a little discrimination, as well as examining further some of the critical key words used in modern translations which leave the course of events les fully understood than it could be.

The Agricola Chapter 25 (link)

Key aspects;

Here Tacitus explains the reasons, or on the face of it the lack of real apparent cause which made Agricola go to war with the tribes north of the Forth and Clyde in 82 AD.

The reason Agricola went to war is couched in terms which suggest he was reacting to events that had forced his hand.
This is not in general terms how Tacitus likes Agricola to be remembered – the decisive man of action in control of and directing events. It is therefore probable that these reported “causes” are in the main inventions (as argued by Fraser) and that Agricola undertook a war of conquest and glory which required some later justification.

As justification would normally be required of the Emperor who made such decisions on military enterprises beyond frontier zones- not the Governor as this would appear to hint at – then this looks increasingly like Agricola went to war without a clear Imperial mandate and that the run of the mill happenstance in the north had to be magnified to create the impression of a greater threat in order justify his decision – as the man on the spot – to invade.

Once the flowery language is parsed off the message is quite clear; the tribes above the Forth – probably those of the Venicone and Vacomagi coastal lowlands are “enveloped”, an unsatisfactory modern translation but one meaning in modern usage “over-run” while the eventual response to this suggests it should also mean “harried”. Meanwhile the Caledonii of the highland zone turned to “armed resistance”.

Clearly the lowland tribes of Strathearn and Strathmore were caught unprepared for the storm unleashed on them – so much for their previous warlike manoeuvres – while the tribes of the highland zone had, due to the nature of the protection afforded by their mountainous highland fastness, some time to marshal their response.
And that the Romans at least had active cavalry patrols seeking out opposition in more remote, even highland territory is likely by their stock reference to overcoming “woods and ravines”.

The navy, the Classis Britannia was active harrying the eastern coast – and it is quite likely in this season that the western seaboard likewise received a similar treatment – which dismayed the tribes;

“now that the secret places of their sea (a reasonable description of the western coast) were opened up”.

This simply means the tribes were disheartened to be on the receiving end of a two prong joint arms action by land and sea.

The blitzkrieg effect of sudden assaults on seemingly safe locations in the rear can indeed be debilitating to morale, yet in return the Caledonians now actively visited this upon the Romans.

Agricola was still clearly campaigning in the north (“enveloping the (lowland ie Strathmore / Angus) tribes” – not the highland Caledonii) when the Caledonii took the opportunity to attack “forts”. Such permanent forts would be well to the rear of Agricola’s campaigning column and are unlikely to have been located far north of the Forth Clyde “limes”. (See addendum).

Seeking to confront the Caledonii warbands, and without knowing from which glens they would debouch from the highland zone – Tacitus`s “routes they were using” – Agricola splits his forces into three separate battlegroups.

This infers but in no way proves that the Caledonii were operating in a similar number of warbands. Agricola simply had to get eyes and ears on the ground in sufficient numbers to cover the possible routes the Caledonii were taking and engage them if they made contact.

The size of these battlegroups therefore is suggestive that Agricola in point of fact actually badly misjudged both Caledonian numbers and martial ability – notwithstanding illogical Tacitean gloss that the Caledonians outnumbered the Romans which in itself makes Agricola’s decision to divide his column even more questionable!

He had clearly considered at the time – not with the benefit of hindsight that Tacitus had – that his three detached battlegroups would each be strong enough to stand on its own two feet in a fight, though clearly it would be expected that they would as a matter of course support each other if possible in a co-ordinated manner in the event of action with the Caledonians.

The area meant by “routes they were using” clearly indicates the many glen mouths on the fringe of the highland massif that would continue to similarly preoccupy the Romans over the remaining years of the Flavian occupation.

When considering the areas targeted by the Caledonians (with forts built or under construction at that early date) then the glens facing Stirlingshire and certainly no further north than Strathallan appear to have been the setting for these events.

In so doing he has Agricola “moving forward” or “advancing” to counter the Caledonian threat. This of course sounds upbeat and positive but makes no real sense geographically as Agricola was clearly enveloping tribes in territory where he had not had any opportunity to build forts yet (and Tacitus would surely have mentioned the feat if he had).

The Caledonians, it seems, successfully outflanked Agricola and “roughed up” his line of communications.

Agricola thereafter certainly moved to intercept the routes the Caledonii warbands were taking but it remains almost certain this would involve an about-face to move to engage the enemy from a campaign previously aiming north east but now in a direction which must now have faced south west.

The Agricola Chapter 26 (link)

Key aspects;

So far so good but the Caledonians, correctly identifying the weakest, most exposed Roman battlegroup gather their forces together and attack and almost over-run the battlegroup of the IXth legion. This attack is fought off only with difficulty and reliance on a relief column led by Agricola.

For a fuller discussion of these events, see the addendums to this article.

The Agricola Chapter 27 (link)

Key aspects;

Without further to-do, Agricola, despite Tacitus shameless propaganda spin over what was at worst a near disaster, and at best hardly a glowing strategic performance by Agricola, retires south to his winter quarters to reconsider forthcoming operations in the north, with the events of the year leaving:

“…………angry feelings excited on both sides”.

Meantime the Caledonians, at last fully realised the enormity of the threat and need the to co-operate together through treaty, a doubtful outcome if the tribes had considered themselves already beaten in battle and physically “overrun” (enveloped) in that year as Tacitus would like us to believe. Clearly they did not.

Tacitus has to work hard in order – without telling barefaced lies – to give an impression other than “Rounds 1 and 2 to the Caledonians!”

The Agricola Chapter 28 (link)

Key aspects;

This chapter is generally considered to be an oratotory device, inserted into the narrative to relieve the pace before picking up the tempo again for the battle in the following chapters.

It is of interest mainly as it is regarded that the events of the Usipis mutiny, clearly on the west coast, most likely took place at Vindogara – probably modern Irvine in Ayrshire.

This then is corroborative evidence that the Roman navy also had units in the west in 82 AD who – while the bulk of the navy in the east co-operated with the land forces – penetrated the “secret places of their sea” *

  • Note: the Atlantic and the Irish Sea were then known as Oceanus “Duecaledonius”, a fairly accurate description of “their sea”.

It is interesting to speculate therefore that it may have been the activity of these units – probably a flotilla our two originally from Deva (Chester) and currently operating out of Vindogara – who harried the Caledonian communities around the western seaboard and the lochs that penetrate deep inland, an action that stung the Caledonians into the retaliatory action that ultimately cost the Romans so dear in 82 AD.

The Agricola Chapter 29 (link)

Key aspects;

Chapter 29 has frustrated those on the quest to find the site of the battle for generations.

This is almost always because they have been tied to a preconceived view on where they would like the battle to have taken place, particularly those supporting sites in the far north, and chapter 29 is undoubtedly particularly unhelpful to such interpretations.

Accordingly they ignore Tacitus account here and replace it with assumptions which place Agricola starting to campaign from where he had left off the previous year, i.e. where they would like him to be.

This ignores the very nature of Roman campaigning. At the end of each campaign season the bulk of the army would retire south to winter quarters. Roman surges north therefore could be compared to waves on a beach.

However it need not necessarily be assumed that each years campaign was like an incoming tide with progressively further penetration in one direction, attention could just as readily be redirected in another direction after the fruits or problems of the foregoing seasons campaigning had been digested and analysed.

Most modern published accounts of the battle of Mons Graupius like to pepper the pages with the odd choice phrase from Tacitus, the early section of chapter 29 however never appears.

This therefore makes it worth our detailed attention.

The bare bones of events are starkly recorded. Indeed the implication is there by default in what Tacitus – the “garrulous embroiderer” – clearly will not tell us. This is simply because certain things did not happen and no glory attended Agricola’s actions before the battle. Therefore there was nothing in those weeks or months that took place out of the mundane that could be suitably buffed up for posterity. Hence campaigning takes a back seat and much was made of a family bereavement.

He tells us that early in the summer – bear in mind campaigning would get physically underway in the closing stages of Spring – Agricola suffered the loss of an infant son.
In the text Tacitus then tells us that Agricola used the campaign to “distract him” from this loss, the only activity mentioned however is the navy again being sent ahead to “plunder”.

Let us be clear on this, Agricola had to be in the field already in order to come to grips with his opponents but there is not one shred of text that suggests that Agricola was actively campaigning in the far north.

He was certainly not battling again to overcome “woods and ravines” like last year in order to achieve a start line at his previous furthest penetration north for the ongoing “northwards thrust into the unknown” theory beloved by many modern historians.

Tacitus would have told us of this activity and simply doesn’t. This omission is exceedingly telling.

So the actions we do know off are the continuing depredations (plundering) wrought by the fleet – a repeat of what was probably the most effective and from the Roman perspective productive phase of campaigning in the previous year, and clearly considered worth repeating in order to stir the Caledonians to respond.

The land forces, obviously positioned somewhere are not mentioned because they clearly had not taken part, till approximately mid summer, in any sort of noteworthy operations.

It can be reasonably surmised that the sum of the land forces achievement in 82 AD was to blunder about in the wilds with no particularly clear conception of where they were going or how they were going to find and engage the enemy – “when shall we have an enemy? ” Tacitus later records.

However before we censure Agricola and his troops too hard on this point it is worth reminding ourselves that this is exactly what Cassius Dio records happening to Severus’s troops in Scotland in 209 AD and we can equally imagine this was also the lot of most other Roman forces engaged in campaigning in Scotland through the centuries.

Tacitus illustrates perfectly the dilemma of campaigning in the wild terrain of the north; when an attempt is made to confront the tribes via the “routes they were using”, the tribes craftily out-think the Romans (again) and almost succeeded in wiping out a third of Agricola’s army.

So why would Agricola be so concerned for his soldiers?

Notwithstanding that Agricola appears to have realised the futility as well as dangers involved in his large army – either singly or in groups – beating the heather for opponents, Fraser has argued convincingly that Agricola was placed under severe political pressure from the Emperor Domitian in the winter of 82 AD.

As we have mentioned already, manpower was an issue for the Romans at this time as Domitian’s Chatti campaign was greedily swallowing up resources, and indeed, Agricola had been instructed to and did send troops from his command – a theatre of ongoing operations in itself – to support the Emperors problematic campaigns on the continent during 82 AD.

As statements in later chapters would appear to confirm, the loss of experienced legionary citizen soldiers – with recruits in high demand on the continent – in action in Scotland in 82 AD was something that would have drawn, perhaps not imperial censure, but clearly a warning. Agricola must have felt he was skating on thin ice and his decision to “reinforce with some of the bravest of the Britons” – it was unusual at that time to deploy natives near to home – cannot have been taken lightly.

Clearly Domitian seems to have sought explanation from Agricola for undertaking his actions in the north and Agricola no doubt in return received warning that no further troops or replacement drafts for losses incurred would be available in the immediate future.

Further, given the text in later chapters and Agricola’s obvious reticence to deploy legionaries directly in harms way it is reasonable to speculate that he may have even been reminded – at the very least – of the imperial “dim view” that would be taken if further legionary citizen soldier losses were incurred.

Perhaps Domitian finally, like most devious political masters, then ironically instructed Agricola – who by now would be feeling like his arms had been well and truly tied behind his back – to get the matter sorted once and for all.

It seems the logical reading of chapter 29 therefore that the army, available in strength, was kept out of harms way while the ongoing depredations of navy provoked the tribes into action.
It would be held ready for the occasion when it could be deployed to best effect – that clearly being its superiority in conventional open battle – and not employed blundering about the wilds incurring mounting casualties again to no good end.

As well as being militarily sound, this course of action would lead to the least possible political fall-out for Agricola personally.

This is a lot to write to fill in the spaces Tacitus has left blank. However it is Tacitus’s very “garrulousness” that makes it clear we are not being told the entire picture.

This picture would appear to be one where Imperial reproach, or even blame – he earlier bemoans that “blame” falls on the individual alone – could not be successfully embroidered with spin into his narrative and which was omitted as it ran the risk of showing Agricola in less than the best light possible.

We continue;

Immediately we are told that “marching light” Agricola and the land forces reached Mons Graupius which he found occupied by the enemy. This is the first mention of activity by the land forces this year.

There we have it again, no long lead in of extended operations – which again would have been harried by the tribes. He comes to Mons Graupius as clearly this is where intelligence has furnished Agricola with information that the Caledonians are mustering.

This must have been what Agricola had been waiting for, and by “plundering” with his fleet had sought to provoke. His army cannot have been too far away from the Caledonians position in order for him to march directly there.

Why use the phrase “march light” ?

Tacitus original Latin is “expedito exercitu”, a curious antique phrase which has encountered unnecessary difficulties in interpretation.

“Marching light” is the commonly applied translation, as is the alternative “without heavy baggage”. The modern military term “forced march” has also been applied, it being argued that this sits well with the spirit of Tacitus’s tale, and the undoubted reality that Agricola would want to come to grips with his opponents right away especially given the wily Caledonians actions of the previous year. This is an understandable and to all intents a fairly unremarkable situation.

We do not believe however that these are the correct translations.

These translations have a problem. None in Latin are even close to Tacitus original “expedito exercitu”.

The common translations however approximate more closely to:

“marching light”- vacuus gravis paratus

“without heavy baggage” – impedimenta expiditus

“forced march” – vis proficisor

However there is a modern parallel, close to the intent in the literal translation of expedito exercitu ; to “quickly perform a military manoeuvre”.

What military manoeuvre would an army require to do? The only manoeuvre an army awaiting its marching orders can do is to – using modern military terminology – rapidly “concentrate the army”.

As noted above, apart from sounding positive and proactive a forced march does not tell us much. What would though is if Agricola rapidly “concentrated the army”, the literal interpretation of conducting a brisk military manoeuvre before marching straight to Mons Graupius.

What does this mean for the course of events in early summer 83 AD ?

An army numbering tens of thousands –as well as their livestock and cavalry mounts are a lot of mouths to feed and water.

This force, which we have argued was kept relatively inactive pending the results of the navy’s ongoing campaign above the Forth Clyde line would have been a drain on local supplies and it was commonly standard practice in such circumstances to disperse an army into several large groups. Each would have independent arrangements for provisioning and re-supply, something one central army group area on its own would not have been able to do for any great length of time.

And Agricola clearly would not know the precise timescale by which the Caledonians would be provoked into major confrontation in the field of battle.

We can reasonably speculate some smaller groups would be north of the Forth Clyde line reconstructing the permanent forts wrecked in the previous season and it is likely that it was scouts or patrols from these units that found the Caledonian mustering underway.

In summary the army was in southern Scotland below the Forth – Clyde line, in force in early summer 83 AD and was likely in several separate large battlegroups that would – and did – reunite at a predetermined army concentration point when intelligence came in pin pointing the Caledonian position before setting out north, no doubt at a challenging pace (which we would expect anyway) to catch the Caledonian host before it moved off.

The remainder of the chapter describes the Caledonian mustering and the bare statement that Agricola came to Mons Graupius. There is no mention of how he set up his camp, it therefore was set up close to the hill as a challenge to the tribes in standard Roman fashion.

The numbers of Caledonians and the Roman marching camp are matters that we will return to in the next section of this article.

Meantime we restrict ourselves to two things. Firstly, Caledonian warriors “were still flocking to the colours”, a grand way of saying that Agricola caught the Caledonians while they were still mustering.

Clearly the Roman army was leaguered in southern Scotland, and at the expense of boot leather managed to reach Mons Graupius in what can only have been a fairly short time. Practically this means in days, possibly only a few, not a timescale involving weeks. This has a great bearing on where this theatre of operations took place and where Mons Graupius itself was situated, and it would appear Tacitus is telling us by implication that this is not too far north of the Forth – Clyde line.

Secondly Tacitus also makes no mention of Agricola making any attempts at negotiation before the battle. Open battle was a desperately hazardous affair for any commander in ancient and medieval times.
It is a matter of record that much ancient warfare revolved about march and countermarch, commanders endlessly seeking that special something of advantageous circumstance before hazarding all in a veritable throw of the dice that could easily result in a knock out blow to either side.

Clearly the Romans marched themselves into a position from which they were happy to offer battle and the Caledonians likewise seem to have been happy to rise to the challenge and clearly showed no inclination to debouch elsewhere through the night.

Obviously attempts would have been made by the Romans to offer the tribes terms for their submission to Rome.

Agricola was literally hazarding the majority of the garrison of Britannia in one fell swoop and like Jellicoe at Jutland in modern times, the responsibility of being able to lose the war – and in Agricola’s case the de facto control of the province – in one day would obviously have sat heavily on his shoulders, no matter how “great” the man Agricola was.

Again, Tacitus silence on this entirely predictable episode is telling. Why would he do this?

Speculation can only go so far but clearly any Roman offers were refused or more probably simply ignored. Further, it remains not improbable that any Romans sent in embassies were simply murdered, a simple act by the tribal leaders which would safeguard tribal unity of purpose before the Roman embassies could spread the corrosive poison of divisive offer and counter offer.

This possibility would indeed explain why Agricola would have had no detail of this matter to give to Tacitus for him to later record. Stony silence reigns on the fate of the Roman embassies.

The Agricola Chapter 30 – 32 (link)

Key aspects;

Chapter 30, 31 and 32 covers the speech Calgacus – a “man of outstanding valour and nobility” – delivered to the tribal army.

As mentioned before, this speech is almost definitely entirely fictitious, and while stirring it is clearly a piece of Tacitean invention and indeed a literary device that would not be amiss in a modern “page-turner” novel.

Unfortunately, Calgacus (meaning “Swordsman”) is also almost definitely an invention too, created as both a literary and as an individual counterpoint to Agricola. His speech was seen as the necessary prelude to Agricola’s own “rousing” pre battle speech in the next chapter as was required by the literary convention of the time.

Formal negotiations which we discussed above, would – had they in fact gone ahead – have given Agricola an insight into exactly who he was dealing with in charge of the council of tribal elders that controlled the Caledonian army.

Without this knowledge Tacitus is forced to invent the character Calgacus. His part duly played in Tacitus’s little piece of theatre, he promptly exits stage left and we hear no more of him, hardly the outcome we would expect of the leader of the Caledonian army following their defeat.

These chapters also contain the memorable and most misquoted phrase from Tacitus’s Agricola; “beyond us nothing is there but waves and rocks ”.

However, the actual text runs; “ ……..but there are no more nations beyond us, nothing is there but waves and rocks”.

This is exceptionally telling, it is a description of the political circumstances – according to Tacitus’s Roman view – prevailing beyond the landmass of Scotland, i.e. no more identifiable tribes and lands to conquer.

However generations of champions for contending sites in the far north of Scotland speciously truncate the phrase, and hence the original meaning, to give the incorrect implication that the site of battle was so far north as to have nothing (geographically) worth mentioning beyond it other than waves and rocks.

The Agricola Chapter 33 – 34 (link)

Key aspects;

Chapters 33 and 34 have some interesting text before Agricola in turn addresses his men.

As an addendum to Calgacus speech in the preceding chapters Tacitus has the bravest of the excited native army rush to form their battleline, while meantime the Romans are still in their camp. This is of some interest as it clearly confirms the Caledonians were on the field first and had been in a position to choose the precise field of battle.

(See later section in this article for a discussion of the ground on each of the contending proposed battlefields).

What is also of interest here is that the bravest – probably the tribal elite and their retainers formed the front battleline while it is likely that it was the tribal levies and Caledonii septs that crowded in distinct clumps up the slope of the hill behind in what Tacitus clumsily tries to describe using the phrase; “in close packed tiers”.

The tribal front rank would have been composed of men used to fighting in the Celtic heroic style as individuals. They would therefore be in relative open order and poorly positioned to deal with the tightly packed close order Roman fighting tactic.
For the tribes the phrase “shield wall” is one that would not be heard for another 800 years and we should not imagine them so arrayed.

The speech Tacitus has Agricola giving is fairly standard stuff from antiquity. What we can comment on is the worry expressed about retreat in the event of defeat. Few English infantry for example in such circumstances managed to make their way back over the border after the battles of Stirling Bridge and Bannockburn in the 13th / 14th C AD and by association we do not have to take this concern as prime evidence of a Roman penetration far into the north beyond the Forth – Clyde line.

Also a hint at what exactly was contained within Domitian’s otherwise unrecorded chilling rebuke to Agricola was comes through when Tacitus has Agricola exhorting the troops to prove that Rome’s soldiers were;

“…….never to blame if wars have been allowed to drag on or the seeds of fresh rebellion sown”.

… an indication at the very least of Imperial displeasure over and the lack of mandate for the actions of the previous year.

The Agricola Chapter 35 (link)

Key aspects;

This chapter covers the Roman deployment and it makes it clear that Agricola only fielded part of his force.

Why would this be?

Firstly, Tacitus is leaning on a situation that evidently took place but reports it in a manner which makes it appear Agricola faced heavier odds than he did in order to magnify his victory.

He quotes 8,000 auxiliaries and 3,000 cavalry while not immediately mentioning Agricola’s cavalry reserve and continues this approach by altogether failing to quantify Agricola’s legionary component.
This all helps Tacitus create the illusion of Agricola battling away when the odds are stacked against him.
In the next section we will examine exactly how many men Agricola had with him at Mons Graupius and it will be clear that here Tacitus is simply embroidering matters for best effect.

Secondly he has the legionaries left out of the main battleline, they being left “pro vallum” – literally in front of the ramparts.

Again, why would Agricola do this?

Tacitus by way of justification states that;

“ victory would be vastly more glorious if it cost no Roman blood, while if the auxiliaries should be repulsed the legions could come to their rescue”

Clearly Agricola was careful with his legionaries, no doubt fearing Imperial wrath if he lost more to enemy action after the events of the preceding year. Auxiliaries though, being non citizen soldiers were – in the eyes of Rome’s elite at least – clearly the expendable canon fodder of the day.

We believe the cynicism goes further however.

If as we will prove there was no inordinate disparity in the size of the two armies, why would Agricola leave so many out of the front line?

We believe Agricola was concerned that if he lined his entire force up in plain view then the Caledonians may think twice about engaging in battle, and battle (with its glory) at the end of the day was what Agricola desperately sought.

Now that he had the Caledonian confederacy’s manpower in the field in front of him the last thing he would wish to do would be to dissuade them from committing to battle.

Lining up his entire force may do this while leaving him without a strategic reserve. Agricola clearly had plans as to the manner in which the battle would be fought and – with a mind to Imperial censure – he arrayed his forces accordingly.

11,000 auxiliary soldiers therefore were fielded, not to intimidate the tribes but to lull them into attacking. Tacitus is quite open in portraying a cynical Agricola, who while keen to protect his citizen soldiers was perfectly willing to hazard his expendable auxiliaries as bait in a Roman trap and in proving this point he has the Caledonians doing what Agricola hoped they would; scorning his “puny numbers”

Pro vallum. This is commonly translated as meaning the marching camps defences, and this is possible if not entirely satisfactory. The phrase however has been attributed at some sites to features in the landscape, some more probable than others. For a fuller discussion see the contending sites section.

The legionary deployment, as well as the reserve cavalry must have been critical to a pre determined battleplan and we should not allow ourselves to be hoodwinked into believing that Agricola – even though he does not appear in the end to have required to use the legionaries – would have left such a major component of his army sitting idly picking their noses somewhere outside his camp with no particularly clear idea of what their intended role was to be. Rank nonsense of course.

Clearly Agricola was egging the Caledonians on to indulge in a precipitate all out attack on his auxiliaries, at which point Agricola could intervene with his cavalry and legionary reserves in a manner that was planned to be the knock out blow.

The Caledonian deployment we have already discussed, however chariots, an archaic feature are mentioned in advance of the Caledonian battleline on the plain, evidence that the battle was fought – in its earliest stages at least – on fairly flat ground suitable for the passage of wheeled vehicles.

At this stage it seems to have become apparent that the deployed Roman auxiliary line was not long enough and in danger of having its flanks over run. Again why would this be, Tacitus has already told us the Caledonians were arrayed first?

We believe this has more to do with the nature of the hill and its extent than the nature of the original Caledonian deployment. Simply put the Caledonians could use the lie of the land to move around or threaten to outflank Agricola’s auxiliaries.

So Agricola massaged his deployment to extend his battlefront – still refusing to use his legionaries – till the line was long enough it would appear to counter the potential problem while being thought all the same as worryingly “thin”.

Clearly a Roman deployment based on defence in depth was abandoned in order to get as many of his auxiliaries into the extended front line. Equally clearly, the hammer blow reserve of the legions and reserve cavalry were left as they were, they being fundamental to Agricola’s battleplan and the “ace up his sleeve”.

Finally Tacitus has Agricola dismounting to show he would not flee the infantry and taking his position in front of the standards. This grand phrase is often used to imply he was in the front rank.
Not at all, the standards would be kept protected out of the thick of the fighting and it is difficult to reconcile Agricola leaving the elevated position of horseback from where he could see what was happening and deploy his reserve at the right moment as he did.

The Agricola Chapter 36 (link)

Key aspects;

Fighting gets underway in this chapter. Although among the last things to be mentioned the first action appears to have been by squadrons of the Roman cavalry engaging the Caledonian chariots who were probably hovering about throwing missiles at the auxiliaries and generally doing their best to unsettle the Roman infantry formation as it advanced.

This suggests that as the Romans infantry closed on the static Caledonian line this cavalry / chariot action became more peripheral on the flanks. Clearly the Caledonian hope that the chariots would upset the Roman battleline did not pay off.

Matters rapidly get very interesting and the detail Tacitus goes into here shows he clearly is working from a detailed after-action report.

The infantry fighting began, Tacitus tells us with an exchange of missiles. At this point the Romans were probably sheltering behind their large shields while the Caledonians showed skill at parrying the spears thrown by the auxiliaries, probably their heavy thrusting spears.

Things it would appear had rapidly got well and truly bogged down. The rear ranks of the Caledonian battleline were undoubtedly tasked with maintaining a barrage of missiles on the auxiliaries, and as the static Roman line stood a matter of yards away it is inconceivable that every tribesmen on the hill behind the main battleline with access to a sling or even stones from the ground would not now be taking the opportunity to bombard the Romans with a hail of missiles.

Watching this Agricola must have realised the Caledonians were not going to be tempted to charge and so Agricola has six of his cohorts ordered to “get going” and advance to close quarters.

Why six cohorts?

We believe these units (possibly in Agricola’s centre) at around some three thousand men were those facing the identifiable “formed” Caledonian battleline. What does this tell us of the Caledonian deployment?

It tells us that the tribal elite were just that, and if we for convenience equate their numbers roughly to the number of men in the cohorts ordered forward by Agricola then that would be 10% of their overall strength. We certainly would not expect the elite to number more than that in a mass call –out tribal army.

This appears to be confirmed when, inspired by the success of these units the other auxiliary units attacked the Caledonians “nearest to them”, clearly the looser formed elements of the tribes up the hill that were converging behind the front line and hovering on the flanks of the main action in the centre.

The fighting, tightly packed, favoured the Roman fighting stance, with short sword jabbing from behind a large shield that gave its bearer good protection. The Caledonians, who in battle would normally fight in open order seeking out individuals to fight, fared badly in this crush, in no small part due to their own rear ranks pushing forward. Their small shields offered little protection in this press while they had scant room to wield their swords and spears to good effect.

Tacitus account of the fighting here is pretty stock standard stuff similar to other accounts of Rome fighting Celts through the centuries, but there is little to doubt it probably records what happened at Mons Graupius.

As the Roman infantry moved forward it would appear they ran the risk of being outflanked and overwhelmed by the milling mass of tribesmen no doubt keen to get into the thick of it, and the Roman cavalry was sent in to attack the Caledonian flanks, probably intending to squeeze them in like book ends though in the process they themselves seem to have got bogged down as well – “quite unlike a cavalry action” says Tacitus.

Interestingly Tacitus admits to the solidity of the Caledonian ranks – indeed the only sure way to successfully face a cavalry charge – though he also blames the roughness of the ground as this clearly took place on the slopes of the hill up which the action was now slowly moving.
The impression is of a tightly packed scrum, the chaos increased for literary effect by riderless horses (suggesting Roman casualties) and runaway chariots colliding with the tightly packed struggling masses.

The Agricola Chapter 37 (link)

Key aspects;

The battle now reaches its crisis point and in chapter 37 there is sufficient text to explain what Agricola was waiting for and indeed had planned for all along.

With all Roman forces visible to the Caledonian leadership bogged down in this heaving mass on the field, they ordered a general advance of all the tribal levies so far uncommitted. As these advanced – both sides continuing to recognise the tactical value of the flanks – Agricola saw the opportunity he was waiting for and unleashed his cavalry reserve (who Tacitus had declined to mention when quantifying Agricola’s line of battle).

Agricola clearly anticipated and planned for this sort of general advance by the Caledonians, indeed his very first deployment on the field seems designed to entice the Caledonians to do just that. The Caledonians however cannily maintained their position on the high ground forcing Agricola to send in his under strength battleline to provoke them to do this.

Roman commanders were trained and expected to be able to handle large forces in the field, the occasion would however be unique for the Caledonians and their handling of the affair so far, even with Tacitean spin comes across as commendable.

Something therefore should be sought in the landscape that allows Agricola’s reserve, both cavalry and legionary, to be hidden pending the right moment to deploy. The legions had probably been intended to act as a weight of infantry backing up the auxiliary battleline from behind or hitting one flank or other of the Caledonians if they had charged the original Roman start line en-mass. As it was the Caledonians declined to do this and Agricola was forced to advance, the action moving bit by bit further away from the legions position.

The cavalry however were fast moving and it is likely they eschewed the flanks which had been sucking men in all day and worked their way round to the Caledonian rear before charging with cataclysmic effect. In the face of this fateful hammer blow Caledonian cohesion evaporated.

As the Caledonian formations broke the cavalry on the flanks got moving again now that things had opened up. The chaos and slaughter that ensued is fully covered by Tacitus and requires no repetition here.

The pursuit however seems to have got a bit out of hand after the Caledonians successfully turned on their pursuers, using the landscape behind the battlefield to advantage.

While Tacitus extols the lead Agricola took in organising the steps taken to clear the area, this masks the fact that, going by Tacitus statistics of Caledonian casualties (which are to be considered at best rounded up, at worst inflated) at least two thirds of the Caledonians escaped, not the result expected when around 5,000 cavalry are available for the chase – witness the carnage at Pinkie in 1547 under similar circumstances.

To summarise, some Caledonians, experienced enough to know what they were about sacrificed themselves in this manner and in so doing they successfully brought the worst of the Roman pursuit to a grinding halt, thereby saving their comrades.

Tacitus may claim the pursuit lasted till nightfall, it certainly would get going again once the difficult ground was cleared but it is clear the worst was stalled by the Caledonians actions.

The Agricola Chapter 38 (link)

Key aspects;

The start of chapter 38 is literary and its theme of post battle misery merely serves to paint a picture.

The latter part of the chapter is where useful information again is made available to us.

Scouts were sent out in “all directions” the next day to ensure the Caledonians were not regrouping.
Tacitus now tells us that the battle took place in late summer (August or September?) and too late in the season for land operations to be “extended over a wider area”.

This clearly signifies Mons Graupius was not in the far north of Scotland.

Agricola marches the army to the land of the Boresti where he links up with the fleet, gives the admiral a detachment of soldiers – clearly to continue harrying operations – and dispatches him with instructions to circumnavigate the north of Scotland, finally proving that Britain was indeed an island.

If Agricola was already in the far north of Scotland there would have been no need for him to do this. The imposition of Roman Imperium in the far north of Scotland therefore was carried out by the navy in 83 AD, not by Agricola with the army.

The Boresti, the fleets harbour at Trucculensis Portus and where Agricola met up with the fleet are all matters we shall discuss in later sections more thoroughly.

Summary

Tacitus work “The Agricola”, as we can see from the above analysis, contains an incredible amount of information. With some careful sifting and thoughtful interpretation sufficient material exists to enable us to understand fairly well the likely course of events in 83 AD as well as assisting our understanding of where events may have, or at the very least where they did not take place.

To assist better our knowledge of where Roman forces in the Flavian period campaigned we turn now to the archaeological record.

Factor 2. The Archaeological Source; Marching Camps

While exercising caution over Tacitus boastfulness we must however note that the bulk of the story enlightens an episode in time whose precise detail are otherwise lost to us. To understand the context of those times however and the movements of the great Roman armies on campaign in Scotland we find the “marching camp” of great assistance (link to marching camps).
Scotland already has a finer tally of marching camps than known anywhere else in the world, tangible evidence of ongoing military activity for over more than three hundred years.
Undoubtedly the far from complete picture we have of these camps will improve as more are located in the future however for the present sufficient exist – if we allow a degree of latitude for the various gaps in the completeness of our knowledge– to put together a tolerable “broad-brush” picture of the sequence of likely Roman campaigns.

The Flavian period (Circa 71 AD to 98 AD):

While Tacitus claims the honour for Agricola being the first Roman governor to invade the lands of “new” peoples it is modern era historians who have interpreted and identified these as Scotland. No national boundary of such existed in those days and it is almost certain that the Romans had previously explored parts of Scotland (link to timeline 77 AD) and had undertaken limited campaigning here as part of extended operations against southern Scottish tribes drawn into Venutius`s doomed Brigantian resistance in the north of England (link to timeline 71 AD)

Following Agricola’s tenure as governor (to whom we shall return to shortly), extensive fort building operations took place dateable to his successor – probably – Sallustius Lucullus.

He was in a stronger position than Agricola. By 82 AD Agricola had been forced to send the equivalent of a legion (comprised of vexillations supplied by all four resident legions in the British garrison) and we can assume these were accompanied as standard practice by a matching number of auxiliaries to assist in Domitian’s problematic Chatti campaign on the continent.

These had returned to Britain by this time. As well as the great many involved at this stage building permanent installations in over-run territory, troops were also now available in sufficient numbers to enable some to be sent in a column into the far north in the year(s) immediately following the battle. This would be done to secure formal treaties of submission from those large tribal groups located geographically beyond the limits of the previous years campaigning and fighting and again was standard Roman practice beyond frontiers following success in war.

This would attempt to ensure there would be no further organised resistance to the Roman order – the Pax Romana – in the north of Scotland.

At this period in time the area which the Romans held with fixed installations extended to Stracathro in Angus and was probably planned to extend further north to the strategically located Mounth near Stonehaven.

Following Lucullus removal and damnation by the Emperor Domitian Rome’s hold on the north waned for a variety of reasons (link to Timeline 88 – 100 AD). This period of ever southward retreating frontiers was characterised by small beleaguered garrisons ultimately abandoning their forts rather than the large grand armies of conquest of the earlier and later periods.

By the years of the early 2nd C AD large elements of the British garrison were on vexillation service on the continent, shoring up its defences following a drain on manpower there to fuel Trajan’s Dacian expansionist wars in the Balkans. In Britain however the Romans were on the defensive.
The frontier had retrenched to the Stanegate in northern England due to manpower shortages in the face of a resurgent Caledonian federation and belligerent southern Scottish tribes left increasingly to their own devices. Roman intervention was infrequent and at least once disastrous (link to the Eagle of the Ninth article).

By this period the morphology of Roman marching camps (link to Roman Marching Camps) had changed. Morphology refers to the physical proportions and characteristics of these camps.

Marching camps of the Flavian period are generally squarish or are described as sub-rectangular in form. Some also sport “Clavicular” style gateways, a feature which does not appear to have been much utilised after the Flavian period. These types of camp have been confidently attributed to the Flavian period in Scotland as some have been found to underlie features of later Flavian date -fort annexe at Stracathro for instance – and have yielded artifacts datable to the Flavian period – pottery sherds at Carey for example.

Marching camps later than the Flavian period exhibit a distinctly elongated rectangular plan and this shapes adoption for “better ventilation”- no minor consideration when every contubernium would have a wood fuelled fire – is explained by Hyginus Gromaticus in the late 2nd C AD. Whether exhibiting a 1:1.5 proportion or even a more elongated rectangular form, these are later Roman camps belonging to the many sojourns of Roman forces operating in Scotland between circa 120 AD and circa 403 AD.

This physical characteristic is fundamental to our understanding of where Roman armies securely datable to the Flavian period marched and the events highlighted above belong. These characteristics are also critical in allowing us to identify camps displaying morphology clearly later than and therefore not belonging to the Flavian period (article forthcoming).

The size of marching camps is also of critical importance in that it allows an estimate to be made of the size of the forces these camps contained.

The earlier “Polybian” system was still probably in use in Agricola’s time and it gives realistic capacities in camps campaigning in Scotland. At 25 acres the space noted for a legion of some 5,500 men is greater than the area later noted by Hyginus a century later than the events in 83 AD. Hyginus work though is contradictory in matters of detail and while the Polybian model may be considered by some as slightly generous for the Imperial forces less reliant on large baggage trains used in Polybius time it does however reflect very well indeed the need to accommodate various other factors.

These range from the large cavalry contingents – a noteworthy element of Agricola’s force-, adequate space for Governors (in many cases even Emperors) entourages, their large bodyguard units and the grain convoys that undoubtedly accompanied Roman forces in Scotland where foraging could not be relied on alone to feed such vast numbers of men and beasts.

For lesser camps the ratio still works well as smaller camps have a greater proportion of their area vis-a-vie larger ones devoted to the standardised dimensions of defensive perimeter and internal streets.

The desktop rule of thumb we apply then is a notional 4.5 acres per 1,000 troops and Roman surveyors would have has rules of thumb to calculate the capacity of a camp based on the length in paces of each side of the camp.

Some degree of discrepancy would no doubt arise between each camp due to the human factor of setting these out by physically pacing the distance, chains being impractical over such distances. A Roman regulation “pace” of five (Roman) feet was measured as two steps which equates to 1.47m metric. Many camps, especially larger ones display rhomboidal shapes and this is symptomatic where high ground contained within vast camps precluded ready triangulation between opposing corners by the surveyors on the ground.

Also, at around 3 defensive camp stakes per running foot of rampart perimeter, the capacities noted above –the Roman infantryman each carried two of these stakes- held the correct quantity of stakes to crown the perimeter rampart and allow further caltrop type arrangements to be formed outwith the rampart without recourse to carrying more stakes in the baggage train or alternatively being inundated with a gross surplus.

The importance of this evidence, coupled with the locations of probable early occupation period permanent forts (i.e. Agricolan) is of exceptional importance in demonstrating where the Romans actually were during the Flavian period and by default where they were not.

The marching camps attributable to the Flavian period excluding those camps in southern Scotland not relevant to the events of 82 and 83 AD are:

Name: Region: Tribe: Acres: Ratio & Capacity:

Ardoch 1 Perth Venicones <14 (e) 1:1.19 (e) 3,100

Ardoch 2 Perth Venicones <11.3 (e) 1:1 (e) 2,500

Auchinhove* Moray Taexali >27.2 1:1.1 >6,000

Bellie Moray Taexali >24 <1:1.33 >5,300

Burnfield Aberdeen Taexali <40 1:1.3 <8,800

Bochastle* (1) Stirling Caledonii 47.7 1:1.45 10,600

Bonnytown Fife Horesti 22.2 1:1.07 4,900

Cardean Perth Venicones 33.4 1:1.4 7,400

Carey Perth Venicones 113.8 1:1.05 25,300

Carmuirs Falkirk Votadini 11.6 1:1.17 2,500

Carpow Perth Venicones 109.5 1:1.03 24,300

Castledykes 1* Lanark Dumnonii 60 1:1.23 13,300

Castledykes 2* Lanark Dumnonii 41 1:1.18 9,100

Dalginross* Perth Caledonii 22.3 1:1.07 5,000

Dornock Perth Venicones 23.2 1:1.38 5,100
Dun  Angus  Vacomagi   8.1   1.21   1,800
Dunblane Stirling Votadini 32.7 1:1.23 7,200

Dunning Perth Venicones 116.3 1:1.04 25,800

Dunipace Falkirk Votadini 115.8 1:1.1 25,700

Finavon Angus Vacomagi 37.3 1:1.16 8,300

Inchtuthil 1 Perth Caledonii 50.2 1:1.13 11,100

Inchtuthil 2 Perth Caledonii 36.6 1:1.2 8,100

Inveresk Lothian Votadini 53.2 1:1.03 11,800

Inverquharity* Angus Vacomagi 5.7 1:1.17 1,200

Kintore 1 Aberdeen Taexali >29.7 1:1.3 6,600
Lochlands 1* Falkirk Votadini 23.7 1:1.27 5,200

Lochlands 2 Falkirk Votadini 44.5 1:1.24 9,800

Malling 1* Stirling Caledonii 26.4 1:1.29 5,800

Malling 2* Stirling Caledonii 11.5 1:1.07 2,500

Newstead 1 Borders Selgovae 50 1:1.1 11,100

Newstead 2 Borders Selgovae 40.7 1:1.05 9,000

Pathhead (2) Lothian Votadini   137.6 1:1   30,500

Stracathro* Angus Vacomagi 39.3 1:1.13 8,700

Steeds Stalls Perth Caledonii 5.2 1:1 1,100

Ythan Wells* Aberdeen Taexali 26.2 1:1.03 5,800

Notes:

(e) : Extrapolated size on basis of the space available and the relationship of the partial remains to other later features on the site.

  • : Indicates that clavicular gates are present.
    (1) : The unusually extended proportions at Bochastle have been attributed to the topographical constraints of the site.
    (2) : The enormous Flavian enclosure at Pathhead has long suffered incorrect identification as a 165 acre “series” Severan camp. At 137 acres it is too small for this association and exhibits neither Severan characteristics nor proportions.
    Areas: Areas noted are taken from various printed sources with latest RCAHMS dimensions taking precedence. All areas have been arithmetically checked.
    Acre: 1 Imperial Acre = 0.404 metric Hectare (Ha). To help visualisation, a football pitch is approximately 1 acre. This helps illustrate how vast some marching camps are.
    Multiples: Camp numbers given indicate there are multiple camps on the same site; the numbers given are not based on any particular numbering system used elsewhere.

Marching camps with ratios of length to breadth greater than the above and with different characteristics are later Roman marching camps. These non-Flavian examples properly belong to the various late Roman campaigns (article forthcoming) in Scotland undertaken by the following commanders:
Lollius Urbicus (circa 139-140 AD)
Activity during the Antonine occupation period (circa 141-165 AD)
Ulpius Marcellus (circa 180-184)
The Emperor Septimus Severus (circa 209-210 AD)
The Emperor Constantius Chlorus (circa 306 and 312 AD)
The Emperor Constans (circa 342 AD)
Lipucinus (circa 360 AD)
Theodosius -the elder- (circa 369 AD)
Maximus Magnus (circa 382 AD)
Flavius Stilicho (circa 395 AD)

Factor 3. Interpretation: A Practical Analysis
Having reviewed Tacitus’s account in “The Agricola” and the dispositions of bona-fida Flavian period marching camps in central and northern Scotland we shall now turn our attention to interpreting this information in a manner that we can then relate the various contending sites against.
The Size of the Opposing Armies.
The size of the Caledonian host mentioned by Tacitus; 30,000, seems excessive to many modern commentators, and it certainly is a very round figure for what at the time can only have been an estimation of numbers of the Caledonian host marshalled at the foot and on the slopes of Mons Graupius.
Clearly not all tribes in the north were involved or Agricola would subsequently have had no need to send his admiral – with additional troops – to “spread the terror of Rome…. before him” as due to the “lateness of the season” he had been unable to “extend (land) operations over a wider area” (Agricola Ch. 38).
General Wade’s military assessment of the strength of the Highland clans in the Jacobite period (early to mid 18th C AD) estimated a fighting strength of 33,000 men of fighting age. This estimate was restricted to manpower within the highland zone.
The forces opposing Agricola – based on the statements above by Tacitus in Chapter 38 – clearly did not consist of all the most northerly tribes. The Romans however – unlike the Jacobite period assessment of highland strength – on the other hand also faced the manpower of the organised and numerous “lowland tribes” such as the Venicones and Vacomagi – including their many smaller satellite tribes of which we no longer have any knowledge – as well as the strength of the scattered septs and embryonic “clans” of the highland zone.
In medieval Scotland, in times of major national duress- normally invasion and war with England – a general call up was expected to gather in all able bodied males between 16 and 60 years old.
That a similar if not wider aged call to arms would have applied in the late first century AD in the face of the exceptional level of threat posed by the massive invading Roman forces need not be doubted.
30,000 men, boys and greybeards, called out in defence of hearth and home, kith and kin and the tribal homeland – while a generalised number – may then not be too far off the mark.
The best comparator we have however in making this assessment is not the number of Jacobite clansmen of the 18th C AD, but very tellingly the numbers of troops the Romans felt necessary to bring to bear to deal with the tribes of Scotland.
The size of the Roman Army that Agricola brought north in 79 AD can best be gauged by the marching camp at Pathhead in the Lothians. This characteristically square Flavian marching camp is capable of accommodating over 30,500 men.
It clearly shows Agricola was prepared to ensure the Votadini were suitably cowed and left in no doubt that the best course of action would be to submit or more probably fully comply with the terms of any treaty hammered out in advance of Agricola’s advance into Scotland.
At this stage the Roman garrison of Britain comprised four legions (totaling around 22,000 legionaries) and this is likely to have been exceeded by or at the very least matched by a similar number of auxiliaries.
Marching north- at least at first till the tribes reaction could be gauged – with just under three quarters of the entire Roman garrison of Britannia may seem a bit over confident given that large tracts of northern England and Wales had only recently been brought under the Imperial heel. However the historical record makes it clear that those areas would fail to prove any further trial to Rome while Agricola would have been fully aware of the martial reputation of the northern tribes.
By the time Agricola advanced north of the Forth-Clyde line in 82 AD – either following Imperial instruction or alternatively actively pursuing glory- he had been forced to send the equivalent of a legion and probably a matching number of auxiliaries to assist the Emperor Domitian’s problematic Chatti campaign on the continent.
(For further discussion of Agricola’s 82 AD campaign see the addendum).
In 83 AD these troops were still on the continent. Casualties- both legionary and auxiliary- incurred in the previous years campaigning in Scotland coupled with an undoubted difficulty securing replacements (Imperial prerogative would ensure that drafts of new recruits would be directed to the Chatti theatre of operations) increased Agricola’s manpower problem to the stage we are told he took the unusual step of using in this campaign auxiliaries recruited from tribes within Britain itself. At this time non citizen troops normally served outwith their own lands).
Antiquarians, southern historians and others not prepared to give the matter much thought have tended to blithely attribute a force of 20,000 men to Agricola in 83 AD. This nice round figure was attractive -it appears- for no better reason than it is substantially less than the Caledonians at 30,000, a figure which ironically the same historians rarely find the need to question.
It should be remembered that partisan anglo-centric interest was and to an extent still is at stake through comparison; only 23 years earlier Suetonius Paulinus with only 10,000 troops had thrashed Boudicca’s rolling tribal behemoth of over 230,000 (including non combatants) in southern England slaying 80,000 in the process only two decades earlier.
This prevailing lightweight level of analysis of Agricola’s strength is something we can easily improve on.
We refer to Tacitus who gives us some useful figures to start working with. Clearly Tacitus account is designed to allow the reader to imagine a disparity in numbers between the opposing forces, and one that would of course reflect on Agricola in glowing terms. While not actually laying himself open to the charge of falsifying numbers, he does however decide to be economical with information that would otherwise give a more balanced reading of the situation on that fateful morning.
His deployed battleline of 8,000 auxiliary infantry was flanked by 3,000 cavalry split (we presume equally) between the wings. The four reserve cavalry squadrons that he unleashed with cataclysmic effect, possibly his own bodyguard can be equated to a further 2,000. This gives a combined auxiliary total of 13,000 troops.
It was standard Roman practice on campaign to match the numbers of legionary citizen troops to that of non – citizen auxiliary soldiers. 13,000 legionaries- in the event never employed in the battle – would give Agricola circa 26,000 troops at his disposal on the field on the day at Mons Graupius.
This is substantially smaller than his total field force in 79 AD but still stronger in terms of the percentage of garrison in Britain employed due to Domitian’s drain of manpower early in 82 AD. Hence Agricola’s noteworthy decision to employ locally recruited auxiliaries.
A figure of 13,000 legionaries should not be dismissed out of hand as a convenient round figure.
Out of the four legions in Britain Agricola still had available what amounted to the manpower of three full strength legions. At full strength this would muster circa 16,500 troops.
Each of the four under strength legions – it is unthinkable that with the prospect of a showdown looming with the Caledonians in the campaigning season of 83 AD that Agricola would not have brought all four north with him – will have left only a skeleton force at their legionary depots in the south, probably a cohort apiece (around 500 men).
This reduces the figure available for field duty by at least 2,000 to circa 14,500 men.
We then have to factor in casualties sustained through the preceding years campaigning (82 AD in particular) as well as losses through sickness (always more difficult to prevent in long periods on campaign than in barracks) and injuries and the figure of 13,000 legionaries available for campaigning looks entirely plausible.
In this light, a Roman force of around 26,000 men does not appear quite so extraordinarily outnumbered by 30,000 Caledonians as Tacitus would allow us to believe.
His decision to “omit” quantifying the legionary element of the force would have been a slip excusable in ancient times by those experienced enough in military matters to do the arithmetic for themselves, but would look mightily impressive to the masses of Rome who would have been unlikely to care enough to question or take the time to work out how many legionaries were actually there, principally as in the event they were not actually needed.

What space would Agricola’s army require on the field?
This is another aspect which has considerable bearing on a contending sites suitability.
We remind ourselves that Agricola drew up only his 8,000 auxiliary infantry with a further 3,000 cavalry deployed on the wings.
He left his legions in reserve and placed his reserve cavalry in a position that must have been hidden from Caledonian view given the cataclysmic impact on the course of events it made when suddenly unleashed upon the unsuspecting tribesmen.
Agricola, fearing envelopment of his flanks “extended” his line till deemed “dangerously thin” by some of his officers.
What does this mean?
The Roman line of battle usually took a chessboard formation (quincunx) though as a convenience we will call it by its more readily recognised name “manipular order” (though strictly speaking that actual term relates to Republican deployments).
As mentioned above manipular order saw cohorts drawn up chess board fashion, one line up, one back, with the individual cohorts in each battleline maintaining a large space between them to allow manoeuvring while the cohort behind, offset in echelon was positioned to guard this gap. The number of battlelines (acies) in this chessboard arrangement could be greater than two cohorts arranged thus, all depending on the wishes and concerns of the general as well as the manpower available.
If in action the front line cohorts required reinforcing or relieved the rear line cohorts would move forward between these gaps in the front battle line. This classic tried and tested technique critically provided great strength to the formation through defence in depth.
If each auxiliary regiment at Mons Graupius was some 500 strong (we believe Agricola recruited Britons to get existing auxiliary units up to full strength as opposed to creating new distinct units) then this gave Agricola sixteen cohort sized auxiliary regiments available for deployment in manipular order.
Traditionally Roman close order infantry formed up eight men deep (if not deeper) while a six man deep line could in extreme circumstances be adopted. Anything less was impractical, carried insufficient weight and could be easily broken through.
At eight men deep therefore each cohort sized auxiliary regiment would have around a sixty man frontage – assuming the auxiliary unit was not further broken up chess board fashion by individual centuries. However we keep matters simple and allow a notional metre per file for auxiliaries formed up in close order giving each auxiliary regiment in manipular order a 60 metres wide frontage with possibly a gap no greater than 40 metres between it and its neighbouring auxiliary regiment in each battle line.
A minimum of two battle lines (duplex acies) is required to create manipular order. This gives eight cohort sized auxiliary regiments per line. Each line, including the gaps, based on an eight man depth would have averaged around 760 m in length. As one line overlapped the other the entire auxiliary infantry formation would have extended to a width of around 810 m.
Allowing each rank in close order slightly more than 1m depth, and a space of around 40m between the two main battlelines would give an overall formation size in manipular order of 810m by some 60 metres depth.
It is a little more difficult to speculate on the space the cavalry divisions on the wings took up.
Cavalry were generally based on multiples of around 30 troopers. Two distinct blocks of 750 troopers on each flank, one behind the other would appear a sensible arrangement that allowed a degree of flexibility. If these were 150 troopers wide by five deep and if we allow each trooper 2m width each, before closing up on the final stages of a charge, then an overall width of around 300m added onto each flank can be expected.
This gives Agricola’s possible initial deployment on the fateful morning a frontage of around 1,500 m. This deployment – with infantry in correct manipular order and cavalry wings doubled into two powerful squadrons (alae) one behind the other on each flank- provides reasonable defensive depth.
Clearly this width was exceeded by the nature of the hill, the benefits of which the more loosely formed open-order Caledonians could take advantage of to swirl around the flanks of the Roman lines.
We should not envision the Caledonian deployment on the hill regimented in neat extended lines marshalled to match the Romans. A battleline at the foot of the slope was evidently formed to counter the Roman advance, this will have consisted of the tribal elite and their retainers. It was the lie of the land above however that would enable the Caledonians to move to outflank the Romans, not their numbers. 30,000 men soon become very thin on the ground in a continuous deployment over only 3 Km – equates to 10 men per lineal metre. This would hardly warrant the description of “rising in tiers” and is yet another nail in the coffin of overly large contending sites.
The poorer equipped and less frontline worthy members of the tribes will have coalesced into clumps of tribal bands on the heights behind this front line and it is these that Tacitus attempts to describe, and their ability to manoeuvre on the topography of the hill that concerned Agricola sufficient to force him to massage his deployment in response to this potential threat.
How far did Agricola have to extend his line – Tacitus’s “open out his ranks” – to satisfy himself the worst of the threat to his wings was countered?
We shall never know with certainty, however, on the basis above it would appear the maximum he could shuffle this force about would be to:
1.Extend each auxiliary regiments frontage to eighty men by reducing the depth of the files to six men.
2.Abandon manipular order and move the second line up into the front line, maintaining only minimal gaps between units and presenting to all intents a single extended battleline (known as simplex acies), six men or so deep and around 1,400m long.
3.Extend his cavalry wings by bringing up his rear squadron to line abreast with the first. Placing cavalry on a static battleline like this seems an incredible waste. The prime role of cavalry was one of shock action and pursuit but circumstances must have dictated otherwise to Agricola. This would extend each flank by a further 300m.
If Agricola used all these options to maximise his frontage to a worryingly “thin” degree then its frontage could have extended anywhere up to around some 2,700m plus some room to manoeuvre.
We should therefore search for a site that would allow a Roman battleline some 1,500m (1.5km / 0.9 mile) long to be threatened with envelopment and one anywhere up to 2,700m (2.7km / 1.7 mile) long where such a threat may be considered to a degree countered.
The Roman Marching Camp
26,000 men require a marching camp (at the most practical rate discussed above of 4.5 acres per thousand) of around 117 acres and one of Flavian proportions (i.e. a squareness / sub-rectangularity in plan) and not exhibiting the classic morphology of later Roman camps.
The camp must be located near to the hill; sufficiently close to enable the Romans to be aware of the unfolding events as recorded by Tacitus and for the legions deployed nearby to be able to intervene and come to the auxiliaries aid if required.
Burnswark near Ecclefechan in southern Scotland is the finest example anywhere of how “up close and personal” Roman camps in action could be placed next to a hill.
A camp therefore located several kilometres away is impractically remote.
Camps placed next to battlefields and siege works were almost always orientated orthogonally to its target -i.e. they addressed it by sitting square on.
The Hill
Tacitus does not make much description of the characteristics of the hill itself.
He mentions the ranks of Caledonians – or tribal warbands – who clustered together in groups- rising in “tiers”.
This has caused some to speculate that the hill will have presented a convex face to the Romans – i.e. a rounded or oval shaped hill and that the tiers would have visible like those of a wedding cake.
However it should be appreciated that even a moderately rounded face will have left much of the Caledonian ranks on the flanks orientated away from and therefore not focused on the action. This is certainly not the impression Tacitus gives, with the extended Roman wings apparently threatened with being overlapped and hence vulnerable to envelopment.
Others reckon a concave shape would better accord with the impression of the threat of the Caledonians wrapping around the over- extended Roman battleline that Agricola was clearly concerned about.
Fraser has quite cogently reasoned that as more bodies push inwards to the focal point of a fight in such a theatre-like profile then the crushing effects of mass crowding that can be deduced from Tacitus narrative would have been the likely result.
Tacitus mentions “up the hillsides” which in itself could possibly record a broken landscape of ridges or the broken slope leading to a length of escarpment, an attractive theory as it allows space for the Caledonians to marshal themselves, face the Roman battleline head-on in a fairly conventional manner and fixes the point of Roman advance in a way convex hills almost always fail to do. The nature of the land to either flank of the Caledonians position would also be telling as Tacitus mentions the cavalry’s progress impeded on the flanks, not just by the solidity of the Caledonian resistance but by the “roughness” of the ground.
Further, a hill, slope or ridges physical emplacement with regard to the likely line of approach by the Romans – in most cases this is dictated by watercourses, terrain and features usually quite remote from the immediate environs of the battlefield- have a bearing on understanding the hills tactical strengths or conversely its strategic weakness. For instance there is no point in the Caledonians adopting an impregnable position if in all probability the Romans were unlikely to obligingly ignore an easier front to assault in preference for the difficult one chosen by the Caledonians situated away from their most likely line of approach!
Nor is a hills suitability assisted where its features leaves its flanks so open that in all probability the Romans would exploit that face as well in the most basic of tactical manoeuvres.
In summary, the manner of approaching and tactically assaulting a hill will vary greatly between each contending site and is based on its physical context. It is the most realistic Roman approach to the hill and the manner of defending and assaulting this that we must carefully review for each location, not hold onto some simple “model” of the basic moves of the battle and apply it broad-brush to each site without consideration of the physical characteristics on the ground there.
Simply put there is little in Tacitus account to firmly identify the nature or profile of the hill or hilly feature which in itself directs us to conclude that there is little to suggest that the hill was the notable, picturesque or distinctly mountainous profile of common misconception, a feature which surely would have merited mention by Tacitus.
The location of many Roman marching camps show that Roman forces often headed towards large hills, and the number of hillforts at these – including the super-sized oppidiums at some – suggests that these hills were tribal landmarks; both military and social focal points. As such it is reasonable to suggest that many will have and would long continue to serve the local tribes as mustering points.
Indeed this is most likely exactly what was happening, or near complete when Agricola caught up with the Caledonians.
A site located near to the communication routes used by the tribes therefore is important. This will include the major arterial valleys, and their connections with the glen openings and watercourses the Caledonii would have naturally followed to leave the highland massif to join up with the hosting of the Venicones, Vacomagi and others.
Also of relevance is the hills location in respect of what the Caledonians would have considered any likely axis of Roman advance.
In the period before the Roman road network had extended far north beyond the Forth, Roman armies could march across country in lowland areas with a fair degree of impunity.
We also should not necessarily restrict their movements to directly upon routes used by their later roads- though that is in itself a fairly telling precedence of routes used and acceptably clear of the worst obstacles to passage in antiquity. This could be features such as marshes which are often now lost to us through modern land improvement.
Did the Caledonians- as some suggest- position themselves astride a route used by a later road to block the Roman advance or could this be done less obviously?
Any such tactically “fixed” position would appear to fly in the face of the strategically aware yet tactically elusive approach adopted by the tribes in the previous year – an approach that met with success. To simply sit and wait for the Romans to turn up seems an over simplification.
An anticipated line of Roman advance could be easily shadowed or flanked -much as it was in 82 AD- from a traditional tribal hosting point located not too far removed from any likely axis of Roman advance but located at a strategic- not tactical- location and as such not directly in the path of the Roman “steamroller”.
Both sides seemed intent on a showdown yet Tacitus we have to remember wrote his eulogy based on recollections given to him well after the event by Agricola. However both Agricola and then Tacitus will have tailored these memoirs somewhat, both had the benefits afforded by hindsight- a conventional battle at the time was not necessarily always a foregone conclusion.
The motivation and pressures on the two commanders would be different. Agricola would be fairly certain this year would be the last opportunity he had as governor to secure the personal glory that would accrue from a victory in the field against the fabled Caledonians.
The Caledonian leadership, as with any confederacy – the rarity of its creation was considered noteworthy – would have been led by a committee of tribal elders spurred on by the realisation of the need to take unified action against a common threat.
The leadership of the confederacy however would still be embryonic enough to harbour the divisions, jealousies, personal animosities and general fall out inherent after age-old feuds between these very tribes, who until now are historically likely to have taken the field usually against each other.
The ability therefore of the leaders of this polyglot force to maintain it at such strength in the field for any length of time is questionable.
If in the event that no decisive action came out of this mass call out – clearly a popular response to the outrage caused by Agricola’s harrying of 82 AD – then it is difficult to see the tribes having sufficient unanimity, at this early historical stage, to recreate this size of host in following years.
For the Caledonian leadership – to quote Burns – now was the time to “do or die”.
How does this influence our understanding of the probable geographical location for the hill?
It is perhaps reasonable to speculate that Agricola caught the Caledonian hosting at a major hill located at or within ready reach of a good communications network, near but not necessarily directly in line with the anticipated Roman axis of advance.
Tacitus makes no mention of the Caledonians employing Fabian “scorched earth tactics” while awaiting the Romans in the far north. The only logical place therefore would be at a location on the southern fringes of the territory of the mustered tribes, a location positioned before the advancing Roman columns could spoil their lands and harry their people yet again.
To choose a site located deep within or even to the north of the tribal lands will have made no sense whatsoever. To do so would have been corrosive to moral and divisive to the unity within the confederation, particularly if some tribes` warriors had to stand by and watch their lands burnt and womenfolk enslaved and not those of others.
The hosting point however was not necessarily the position from which the tribes originally intended to do battle with the Romans. In the event, given the rapidity and perhaps unexpectedness with which Agricola’s forces came up – “expedito exercitu” – the position will have been reviewed by the Caledonian leaders for its suitability to hazard open battle with the Romans (of whose reputation the Caledonians can only have been too well aware) and in this respect its suitability appears to have been considered acceptable.
In so choosing the ability of the mass call out tribal army – intrinsically less mobile in tight situations than the small mobile forces of the tribal elite fielded the preceding year- would have had a strong bearing in rejecting any option for a breakout.
To do battle consensus on both sides is required and this appears to have existed. The hill therefore must have been strong enough yet not too steep to enable Caledonian deployment on its slopes –they adopted a defensive posture – while neither so strong nor so steep that its characteristics proved an inordinately difficult and risky proposition for the Romans who early adopted an offensive posture.
The Plain
Tacitus tells us that between the foot of the hill and the marching camp lay a plain and it was across this relatively smooth area that the Caledonian chariots engaged the Romans as they deployed and started to advance.
Clearly this plain had to be fairly level and sufficiently free of obstructions to allow the passage of wheeled vehicles. It is only as the Romans start climbing the higher more difficult ground of the hill that we are told the unevenness of the ground affected the progress of their cavalry on the wings of the infantry line.
Tacitus makes no mention of the Romans being divided from the Caledonians by a noteworthy watercourse, something he will have done if it materially affected the manner in which the battle developed.
Watercourses are common features near and around Scottish hills and in general either run down the slope or run across its foot. Hills and the ground at their base not affected by watercourses like this are rare and this is a point of particular interest in the viability of contending sites.
The Land Beyond
Tacitus account also mentions the difficult nature of the terrain behind the Caledonian position across which they withdrew, and in some instances, turned on their pursuers to good effect.
The position of woods can change dramatically over the intervening centuries however the inference is of very broken and difficult ground, with limited line of sight which required the Romans to adopt a huntsman’s approach to flush out any Caledonians seeking shelter there or waiting in ambush.
This is a classic description of wooded uplands and is also a compelling feature of the landscape to help our search. The trees however may be long gone though.
Contact with the Fleet
After the battle, Agricola withdrew -“deducit” – the army to a location where he made contact with the fleet. He will have done this on the day following the battle as local water supplies will have become unavoidably fouled.
In marching to the sea, Tacitus does not suggest it was a great journey, a day or two at the most is the implication. During this march Agricola will have been burdened with both his own wounded as well as any prisoners taken.
The point of contact with the fleet would be a sheltered anchorage if this was the location where the entire fleet was stationed – Trucculensis Portus (Wilderness Haven) – or a navigable waterway cruised by small vessels detached from the main fleet and tasked with maintaining links with the Governor and the land forces.
Tacitus makes it clear that the fleet was busily engaged harassing the northern tribes, ranging up and down the east and probably also the west coasts. Either Agricola knew in advance where the fleet was using as a night-time anchorage or he relied on making contact with the small vessels patrolling estuaries.
The Boresti
Tacitus tells us that following the battle, Agricola “withdrew” to make contact with the fleet as it was too late in the year for operations on land to be extended (a clear indication that events did not take place in the far north). During this episode Agricola took hostages from a people known as the Boresti.
Identifying the location of the Borestis’ tribal homelands holds a major clue to the location of both the preceding events and those that followed.
The name Boresti is probably another corruption, this time of the Celtic name “Horesti” or the “Horreii” and Ptolemy’s map (based on information from the Flavian period) indicates a site called (H)Orrea located in Fife.
The later Ravenna Cosmography intriguingly mentions Poreoclassis and this is generally identified as the early 3rd C AD Severan legionary fortress at Carpow on the Tay. The classis element is usually interpreted as recording the sites link with the navy.
One possibility is that the “B” was introduced by Tacitus (or indeed by later medieval copyists) in lieu of the “H” or “P” to give the impression of an extreme northern location for the battle by mimicking the classical Greek terra incognito name of “Hyperboria” – meaning literally the north wind beyond the land of the “Boreas”.
Alternatively, the other major watercourse that bounds the Fife peninsula – the Boderiae Aest (Firth of Forth) may be suggestive of contemporary phonetic influences on the name of the tribe who lived there or at the very least in the pronunciation of their name.
We suggest that a softly pronounced Celtic “ph” before either Orrea or Orrestii will have been the original Celtic rendition that the Romans heard and in whose harsher Latin pronunciation was given a different spelling when subsequently written down. Therefore either the letter “H” or “P” was adopted depending on the strength of the particular inflection used in its original spoken form and – like other names in this quest – has subsequently been transmogrified further with the substitution of the “B” prefix, most probably for poetic licence.
This would place the Boresti inland firmly within the Fife peninsula and most likely centred around the Lomond Hills.
Of the general “Fife” peninsula between the Forth and Tay it is probable that the Venicones held sway over those areas approximating with the modern Perth and Kinross holdings there.
The remaining bulk of the Fife peninsula was probably Boresti territory and likely also satellite to the Votadini whose territory – or more probably hegemony – Ptolemy shows extending into Fife. A satellite dependency to a named southern tribe – Tacitus otherwise names no tribes north of the Forth Clyde line – makes sense and marks this tribe to the north of known tribal territories but to the south of the bulk of the Caledonian confederacy.
Alternative linguistic conspiracy theories that claim there was no Boresti tribe and that the text should read that Agricola “withdrew” to the “Boreas” i.e. the far north can be dismissed as illogical and contradictory to Tacitus explicit statement that Agricola was unable to extend land operations further after the battle and that he physically took hostages from a people called Boresti.
Conclusion
It is no mere speculation therefore that the battle took place near to Fife and that Agricola’s noteworthy decision to take hostages from the Boresti follows this ostensibly allied Votadini septs failure to prevent – or perhaps report on – the northern tribes mustering on the borders of their territory.
Return march through conquered lands
Agricola’s stately return back south is of interest “…marching slowly in order to overawe the recently conquered tribes by the very deliberateness of his movements, placed his infantry and cavalry in winter-quarters”.
Recently “conquered tribes” has been ignored by those seeking to justify a location well to the north of the lands of the tribes involved in the battle and speciously changed to recently “defeated” tribes.
This is incorrect. The tribes so comprehensively beaten at Mons Graupius would need no further immediate actions by Agricola to “overawe” them. If they were indeed over-awed then the charnel house on the battlefield itself would have amply served to do that.
The line of his return south should be identified as passing through the lands of the tribes Agricola conquered in his years as governor. This refers to the tribes of southern Scotland, through whose land Agricola had to pass to get his legions back to their fortresses in England. This otherwise fairly unremarkable event is explained by Tacitus’s need to bring his record of events to a conclusion; to give closure to the military campaign in his narrative.
Where the Romans campaigned in the north and in what strength.
Interpreting and understanding the extent of (currently known) Flavian campaigning in Scotland interestingly lets us identify where Agricola may have been and certainly where he wasn’t.
Permanent forts by the time of Mons Graupius and the year preceding it had not extended past Strathallan, and indeed it is probable that the first attempts to construct forts here were acts that contributed to provoke the Caledonians to assault these – or realistically the labour camps of the troops engaged building them.
The permanent forts extending up to Stracathro from Strathallan are dateable to the tenure of Agricola’s successor. This indicates that the land held by the Romans during Agricola’s governorship- land pinned down with fixed garrisons- extended only marginally past the Forth – Clyde line.
Southern Scotland has the greatest number of Flavian marching camps.
Above the Forth –Clyde line securely datable marching camps of the Flavian period are located most heavily in the Central region from Falkirk to Dunblane, on the mountainous fringes of Stirlingshire and Perthshire bounding Caledonii territory in the Trossachs and Breadalbane and ominously clustered along the River Earn – where the largest Flavian camps are located- and which was clearly a highly strategic line.
North of this is a single series of medium sized camps (circa 30 acres) passing through Angus, crossing the Mounth into Aberdeenshire and extending through Buchan to Auchinhove near Keith (or possibly as far as Bellie on the Spey). None of these northern camps are big enough to hold even the auxiliary numbers Tacitus allowed Agricola at Mons Graupius.
These camps clearly belong to an operation after the battle, maybe the following year. Conceivably this is the only time a Roman force of such a size – on their own – could make a progression into deepest Caledonian territory in anything other than a state of grave concern for their safety.
The size of later Roman armies campaigning in the far north never fell beneath numbers requiring massive 100 acre plus camps, accommodating numbers similar if not greater than Agricola fielded at Mons Graupius. Clearly such extended operations, requiring great logistical planning beforehand as well as huge resource of manpower remained the preserve of either the Emperor himself or armies specifically formed under his instruction for the purpose.
Central Scotland therefore exhibits the highest concentration of Flavian activity attributable to the events leading up to the close of the campaign season in 83 AD and it is to these camps within this area that we should identify the activity of Agricola.

Factor 4. The name: Mons Graupius

Ancient names within a society evolve naturally over time but also unnaturally due to outside influences. Naturally is the process of evolution in the manner a word is spoken. Culture can directly affect dialect and human nature itself will frequently soften words, with letters giving harsh staccato sounds being commonly dropped or burred in usage.

An example can be seen in the myriad dialects in the British Isles and the different emphasis each of these put on various parts of a word. This is more dramatically seen in the USA where a regional dialect went one step further and formalised such changes in common usage there with a spelling different from that in the original mother tongue.

This softening or dropping of emphasis in the pronunciation of letters is more extreme and relevant to the current discussion than “you say tomato, I say tomAto.”
This is illustrated by “butter” in common British usage becoming “buddur” in some American usage. Even in Britain though the process is developed, in the far south east of England “butter” has become “buttah” as the “R” – famously inconvenient to southern tongues- is dropped.

In the electronic information highway era this sort of process is carried at an accelerated rate and mediums such as “text-speak” can illustrate an extreme example of how a word can be shortened to make it easier, quicker and less bothersome. Of relevance to this are dypthongs in ancient spellings (groupings of co-joined vowels), particularly Celtic tongues which developed and saw many dropped over time in the general softening process.

Unnatural change comes about when an existing language is submerged beneath that of an incoming or conquering elite. The language of the tribes – Brythonic – survives now only rarely in some place names in southern Scotland, particularly to the west. The place name Lanark is a good example.

A not entirely irrelevant factor concerning Brythonic (Old Welsh) is that it was the language used by JRR Tolkien as a role model for “elvish” in his highly entertaining novel the Lord of the Rings. The recent movies, accurately portraying this style of language amply demonstrate just how different pronunciation in this spoken form is from Latin or Germanic (English) based languages which are extremely guttural and harsh by comparison.

In the south and east English influence erased all of the original Brythonic “P Celtic” place names. In the highlands Gaelic may sometimes yet still recall a variation of an original Brythonic place name where this was understood and part fashioned into a “Q Celtic” Gaelic equivalent name.

One common misconception of the name Mons Graupius however must be dealt with straight away.

The “Grampian” Mountains and the modern political area associated with them do not reliably point to the location of the battle.

These mountains were named after Mons Graupius by the early Aberdonian historian Hector Boece in the 16th C AD.
Boece was an early exponent of the partisan approach of championing a site close to home. In so doing he initiated the transmogrification of the traditional name for the mountainous core of old Scotland: Drumalban – literally “spine of Scotland”– via a fictitious piece of fancy “Granzeban,” eventually arriving at Grampian. The “M” is an imposition from Drumalban but necessary for the deception!

This name has subsequently stuck and is the classic example of the “cart pushing the horse” in the philological quest for the original name of the battle site.

Boece’s singularly partisan act has been the root of more common misconception on the location for the site of Mons Graupius than any other and remains to this day a pervasive influence, even among Scotland’s historical institutions.

Another point which must be clarified is that just as “Grampian” was Boece’s deliberate mis-spelling of Tacitus’s “Graupius”, then first letter “G” is certainly the Latin pronunciation of the original Celtic “C” -the Romans normally took their cue from the existing Celtic name for a site.

Graupius therefore is “Craupius”.

By peeling off another layer of Latinising accretions; the “us” or “ius” suffix we arrive at the core word; Craup(i).

By appropriately identifying the likely original Celtic spelling of this Latin pronunciation we arrive at the commonly accepted root word; Croup, Croupi or Croupii.

From here we may now start searching.

The meaning of Croup is the subject of scholarly debate. Welsh “Crwb” means bump and some suggest that this may have been a descriptive quality that was attached in antiquity on account of the visual characteristics of the hill.

Another suggestion has been that Croup could be the name for a hill, a people, or a region in which the foregoing were located.

This means that the hill the Roman identified -with the necessary Latin prefix/title for hill (not mountain!) “Mons”- could be rendered as simply as “bump shaped hill”, the “Hill of the Croupii” or “Hill at the Croup”

The Contenders

Scotland has more hills than you can metaphorically shake a stick at. Accordingly we must exercise discrimination in trimming the list from every hill down to those that exhibit characteristics which have in the past or now make them worthy of consideration.

No fewer than 29 specific sites have historically been linked to the battle or have been identified for reasons which make their inclusion here worthwhile. We have ignored generalisations such as “somewhere in Caithness or Sutherland” as lacking in specifics that we can analyse.

The benchmark testing – based on the four factors we have discussed at some length above – against which we will interrogate these contending sites are:

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Is the Flavian activity in the area likely to be Agricolan?

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Is the marching camp located close to the site?

Does the marching camps position and orientation “address” the site?

Does the marching camp display Flavian characteristics?

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around?
A realistic Roman assault up?

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Can the “Boresti” be identified locally?

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

The Contenders:

Location Region Tribal area OS Coordinates.

Ardoch Moor Perth Venicones NN836114

Bennachie Aberdeen Taexali NJ690257 (alt’ve NJ715251)

Carey Perth Venicones NO174155

Carpow Perth Venicones NO218174

Craig Rossie Perth Venicones NN994133

Culloden Inverness Taexali (possibly) NH740450

Dalginross Perth Caledonii NN773207

Dunning Perth Venicones NO042142

Fendoch Perth Caledonii NN920283

Fortingall Perth Caledonii NN736470

Gask Ridge Perth Venicones NO033180 (alt’ve NN999182)

Glen Eagles Perth Venicones NN913077

Hill of Bruxie Angus Vacomagi NO822802

Hill of Edzell Angus Vacomagi NO590700

Huntlyhill Angus Vacomagi NO620636

Inverquharity Angus Vacomagi NO364557

Kempstone Hill Angus Taexali NO873895

Knock of Crieff Perth Venicones NN874224

Lomond Hills Fife Horesti NO204084

Moncreiffe Hill Perth Venicones NO166192

Mondboddo Angus Vacomagi NO751793

Monifieth Angus Venicones NO495345

Mormond Hill Buchan Taexali NJ970560

Pass of Grange Moray Taexali NJ525525

Peterculter Aberdeen Taexali NJ825000

Stormontfield Perth Caledonii NO144474

Strathfinella Angus Vacomagi NO702762

The Caterthuns Angus Vacomagi NO560660

Tillymorgan Moray Taexali NJ652352

Note; Modern “regions” shown above are abbreviated for convenience. For example where Perth is noted read Perth and Kinross.

Ardoch Moor

Introduction

Ardoch Moor (also known as Muir of Orchil) is the rough plain (app. OD +150m) in Strathallan immediately fronting the rolling escarpment leading up to Coire Odhar (OD +357m). The place is famous for the well known group of Roman installations at Ardoch near Braco.

It was without doubt the remarkable state of preservation of the fort there which generated sufficient antiquarian interest to prompt Chalmers in 1807 to hail it as the location of the battle, a suggestion subsequently taken up by Stuart (1845) in his classic work Caledonia Romana.

The site is bounded by the River Knaik and the Machary Water and interestingly contains a farmstead whose name – “Victoria” – records the fame the site received by Chalmers claim.

The slight remains of the Roman roads “agger” or ditches hereabouts were probably the features which Chalmers explained as a rampart which the legions threw up and stationed themselves “behind” out of the way of the Caledonian chariotry.

The upland road leading to Comrie following the course of the River Knaik was the route Chalmers suggested that the retreating Caledonians took.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes

Is the marching camp located close to the site?

Yes

Does the marching camps position and orientation “address” the site?

Not very well, it appears to owe its orientation more to the adjacent Roman road.

Does the marching camp display Flavian characteristics?

No, it displays a site structural sequence indicating a later period.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, at 119 acres it could accommodate 26,400 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Width suits very well.
Depth – the long rolling profile of the escarpment running up to Coire Odhar, at over 1,500m, would give the Caledonians too deep a formation between front ranks on the plain and those located on higher ground for effective command and control to be maintained.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No. The watercourses running along the moor are generally minor.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:
The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, very broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

The sites central location would enable a march to the Tay or Forth.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes if the Tay or Forth was used.

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Ardoch is a site that has been too readily dismissed in the past.

Without doubt the main Flavian works in the area are the two-stage permanent forts and the associated labour camps housing the troops employed constructing them.

The remarkable log jam of marching camps on the site in the main belong to later periods (article forthcoming) but it is interesting to note that not only was Ardoch a pivotal strategic location in ongoing Roman campaigning north of the Forth – Clyde line but that on at least one other occasion a Roman battlegroup of very similar size to the forces Agricola deployed at Mons Graupius passed this way.

The proposed battlefield is suitable in most respects except for the depth of the Caledonian position which appears too deep. The suggested headlong retreat along the pass of the River Knaik would also have been suicidal for the tribes who would more probably have scattered into the high ground behind Coire Odhar.

Ardoch Moor achieves a respectable 15 appropriate answers out of 19, making this site a “strong contender”.

Bennachie

Introduction

Bennachie is an extremely imposing and highly visible mountainous hill in the Garrioch north west of Inverurie.

J.K.S. St Joseph – the noted exponent of 20th C aerial reconnaissance – discovered the exceptionally large Roman marching camp at Logie Durno, several kilometres from Bennachie in 1978.
St Joseph argued that the camp, the largest then known north of the Forth – Clyde line was Agricola’s on the eve of the battle of Mons Graupius.

He attempted to explain the great distance between the camp and the hill as a “precautionary” measure to lessen the likelihood of Caledonian attacks through the night.
Further, the camp at Logie Durno -St Joseph asserted – appears to have ominously veered off a direct line of march between other camps at Kintore and Glenmailen/Ythan Wells.

St Joseph therefore proposed a battle centred around Kirkton of Oyne – around +100m OD – with the Caledonians massed on Bennachie which rises to +518m OD. For many years since Bennachie has been popularly hailed as the site of the battle.

Roman Scotland visited the site in 2007 full of anticipation and revisited twice in 2008 to check our findings. We were struck by both the vastness of the site as well as the unacceptable gradients on Bennachie the Caledonians and Romans would have to contend with.

Walking the ground, an alternative site centred on Chapel of Garrioch, part of the sweeping wings on the Bennachie massif offered itself as a fine site, a good match for the description of the site by Tacitus and in many ways superior to St Josephs over-large and sprawling site above Kirkton of Oyne.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the circa 30 acre string of camps arcing through Aberdeenshire to Moray do not hold sufficient capacity and post date 83 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Durno.

Is the marching camp located close to the site?

No, it is approximately 5 km away or 3km to the alternative Chapel of Garrioch site.

Does the marching camps position and orientation “address” the site?

Traditional site; yes. Alternative site; no.

Does the marching camp display Flavian characteristics?

No, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No. At 144 acres Durno is too large and could accommodate a force of 32,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, the traditional Bennachie site is too large, over 8km wide.

Reducing this frontage to an alternative location between its outlying lower slopes around Hill of Knockollochie and Gallows Hill centred on Chapel of Garrioch gives a very credible width of just under 3 km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Yes. The camp at Durno sits directly behind the River Urie and the line of Roman advance and deployment is similarly crossed by the Gadie Burn.
Both of which would be noteworthy impediments; first to the auxiliaries forming up and secondly to the ability to rapidly commit the legionaries in the event of the auxiliaries coming under difficulties.

The alternative suggested location at Chapel of Garrioch is close on fordable reaches of the Urie and not hampered by the Gadie Burn.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

The ground on the traditional extended battlefront is extremely broken around Oyne. The reduced battlefield below Garrioch is ideal.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Bennachie has a distinctly mountainous profile and steep gradients, even at lower levels around the northern quadrant of the hill. Gradients at the suggested reduced width battlefield are appropriate.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, a broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No, a 3 day march encumbered with wounded would be necessary to reach the Aberdeenshire coast, probably following the River Don.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Bennachie is a vast site and the sheer bulk of the hill alone makes it difficult to imagine 30,000 warriors on its slopes being so thickly massed as to be able to rise in the “tiers” described by Tacitus.

The traditional site occupying much of Bennachie is simply too large for the action and numbers recorded by Tacitus. The alternative site at Chapel of Garrioch however is extremely imposing, of a practical gradient for the Caledonians to be marshalled on while its strong theatre – like profile would indeed make any commander assaulting it naturally concerned for his exposed flanks.

The ground between the battle site(s) and the camp at Logie Durno is bisected by the River Urie and the Gadie Burn. Deployment and the course of events would be both hampered by these as well as the rough broken ground at Kirkton of Oyne.

The camp itself at Logie Durno unfortunately has late Roman marching camp morphology and its great size is sufficient to accommodate 32,000 men, some 6,000 men more than we have calculated Agricola deployed. At no point has anyone suggested that the Romans outnumbered the Caledonians.

Leaving the legionaries at the camp, either in front or behind the Urie would leave them too far removed from the scene of the fighting to influence events if required.

Finally St Joseph’s ominous diversion of the line of march, is unfortunately entirely imagined.
The distinctive peak of Mither Tap is long visible from the wide lands of lowland Aberdeenshire from as far as Normandykes on the Dee and was clearly used as a waymarker by the various Roman forces who marched this way.

There is no “deviation” in line of march to Durno, it simply lies on the route a force would take through the Garrioch between Kintore and Ythan Wells while skirting the River Urie, a natural corridor that would be later followed by medieval roads and modern communication routes.

Bennachie will always remain a “must visit” site and Roman Scotland hopes that the alternative Chapel of Garrioch site we have offered at least endeavours to address some of the problems inherent with the traditional Bennachie site.

Unfortunately, with only 5 and 7 appropriate answers out of a possible 19 for the traditional and alternative sites respectively, Bennachie rates no better than “highly unlikely” and can no longer be rated as the favoured site for the location of the battle.

Carey

Introduction

Carey is located in farmland at Abernethy in Perthshire at the junction of the River Earn with the Tay. It fronts the conspicuous Castle Law which rises from the low flat plain – circa +20m OD – that the marching camp is situated on up to +249m OD at its summit.

Roman Scotland identified the viability of this site in 2008. At the time of visiting the site its greatest interest was the size of the proven Flavian marching camp there, part of the Dunning – Carpow series which are noteworthy in being the correct size and capacity to hold the size of force we have calculated Agricola fielded at Mons Graupius.

However while walking the site and closely inspecting the names of local topographical features we identified for the first time the “Croupie Craigs”, clearly the original name for the hill – latterly renamed and now termed Castle Law.

As a direct rendering of Croup or Croupi this rates as evidence of exceptional importance.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Carey.

Is the marching camp located close to the site?

Yes

Does the marching camps position and orientation “address” the site?

Yes

Does the marching camp display Flavian characteristics?

Yes, the camp has Flavian morphology which is further confirmed by pottery sherds datable to the Flavian period found there.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, at 113.8 acres it is capable of accommodating 25,300 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

The Castle Law outcrop has a frontage facing the camp of about 2km which is rather too narrow to have caused Agricola concern for his flanks.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, the Carey Stank is likely an improved modern field drainage feature.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? No, the Craigs in part have severe gradients.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, access via lighters on the lower reaches of the River Earn is immediately available, it is not however a noteworthy march to the marching camp at Carpow unless encumbered with a large number of wounded and prisoners.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, Carpow.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Yes, Croupie Craigs located on the Castle Law outcrop is the best etymology available anywhere, and linked to evidence at Dunning is extremely compelling.

Summary

Considerable interest attaches to this site and its credentials, 15 appropriate answers out of a possible 19 are sufficient to rate this site as a “strong contender”.

The size of the battlefield however is restricted, and the gradients on the hill are too steep to be practical.

The marching camps association though with the similarly sized examples at Dunning and Carpow sheds light on the movements of the force that encamped there and this is dateable to the Flavian period by the pottery recovered at Carey.

The philological evidence brought to bear by Croupie Craigs is evidence of the very highest calibre.

Carpow

Introduction

Roman Scotland identified the land around Newburgh in Fife as a contender in 2008 primarily as the site contains the Carpow marching camp, one of the series of suitably sized Flavian marching camps in the area.

The land at Carpow itself saw several phases of Roman occupation and a large Severan fortress occupied the site in the early years of the 3rd C AD.
Carpow was called “Horrea Classis” or “Porreoclassis” in the late Roman Ravena Cosmography.

Ormiston Hill (+236m OD) sits prominently above the River Tay, and chokes further easy access along the south shore of the Tay estuary here. The lower slopes extending past Easter and Wester Clunie would provide an ideal Caledonian position, with the suggested battlefield positioned on the low lying ground (circa +30m OD) between the hill and the (later) fortress site.

This land would probably be on or near the boundary between the Venicones and the tribes of Perth – most probably the Horesti.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Carpow.

Is the marching camp located close to the site?

Yes

Does the marching camps position and orientation “address” the site?

Yes

Does the marching camp display Flavian characteristics?

Yes

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, at 109.5 acres it can accommodate 24,300 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, a site centred on Clunie extending to lower slopes of Ormiston Hill and Pitcairlie Hill has a concave profile 3.5 km long.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Apart from Gillies Burn running across the face of Ormiston Hill, no.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes though with only a limited depth.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Lower slopes only.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No. Access to the fleet is immediately available on the Tay and would not require a further march.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Yes by close association with the escarpment with excellent philological links at Dunning and Carey.

Summary

Carpow benefits as a contender by its association with the suitably sized Flavian camp there. At an impressive 15.5 appropriate answers out of a possible 19, Carpow rates as a “strong contender”.
Further, Carpows association with later recurring Roman activity is also interesting.

There are some sheer falls to parts of the upper slopes of Pitcairlie Hill though, and the depth of the battlefield is quite compact. The saddle between Ormiston and Pitcairlie Hills could be singled out as a weakness in any Caledonian position there which allows us to speculate over why Agricola sent forward his Batavian and Tungrian cohorts before committing his other auxiliaries.

Sited directly on the banks of the Tay, we wonder however why Tacitus would record Agricola undertaking a march following the battle to make contact with the fleet when this would be immediately available to hand.

Carpow however is a fine site that well justifies its inclusion in the list of contenders.

Craig Rossie

Introduction

Keppie in “The Legacy of Rome” highlights the possibility of Craig Rossie as the scene of the fighting. Several authors also allude to a privately published work – Smith 1987 -supporting a site in this area though this publication remains unseen by Roman Scotland at time of writing.

The distinctive profile of Craig Rossie, close to Auchterarder in Strathallan is, like Bennachie and the Lomond Hills, attractive principally on account of its striking profile, long visible from the A9. It lies on the southern fringes of Venicone territory.

Roman Scotland visited Craig Rossie in 2008 to assess the most likely location on its slopes for the suggested clash of arms.

Craig Rossie forms part of the northern fringe of the Ochills and its prominence is in no small part due to the fact that the line of that range of hills describes a distinct right angled bend, and it is on this prominent corner that Craig Rossie is situated.
As a hill therefore it is anchored to the Ochils to its rear but presents at least two long rolling faces to its fore up which a Roman commander could choose to launch his assault.

Of these two fronts the first faces NNW at Cloan (+120m OD rising to +306m OD), the second faces north towards Blair Hill below Rossie Law (+80m OD rising to +324m OD).

The latter is superior and would allow a Roman advance from the direction of Aberuthven on ground – excepting White Moss Loch – otherwise perfect for a Roman advance and deployment as well as being of sufficient size for the Caledonians to threaten envelopment.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Innerpeffray and Dunning

Is the marching camp located close to the site?

No, Innerpeffray approximately 7 km distant, Dunning 3.5km away.

Does the marching camps position and orientation “address” the site?

Innerpeffray; yes.
Dunning; no.

Does the marching camp display Flavian characteristics?

Innerpeffray; no, it displays late Roman marching camp morphology.
Dunning; yes.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Innerpeffray; no at 130.5 acres it is too large and could accommodate a force of 29,000 men.
Dunning; yes at 116.3 acres can accommodate 25,800 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes. The most probable line of Roman assault would be from the direction of Aberuthven onto Blair Hill. The frontage here is less than 3km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? No, the Craig Rossie escarpment has a mountainous profile and the steep gradients that associate with such a profile. Suitable gradients are on lower slopes only.
A realistic Roman assault up? As above, on lower slopes only.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, the Ochil hills.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, the Tay or Forth.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes, especially on the Tay.

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, Dunning, Carey or Carpow.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Possible, see Dunning and Carey.

Summary

With a creditable rating of 15.5 out of a possible 19 points Craig Rossie admirably rates as a “strong contender”.

Of all the “prominent” contending hills; Bennachie, the Lomond Hills and Craig Rossie, it is Craig Rossie that nearest matches the overall criteria.

However as with all such mountainous profiles its gradients – in this case to the upper slopes – are severe and would hinder Caledonian deployment and a Roman force displaying any modicum of cautions willingness to advance up such an obstacle.

The camp at Dunning does not address the site, however adequate space is available near Aberuthven and it is interesting to speculate if one may some day be sought and found there, possibly around the Laigh of Rossie.

Culloden

Introduction

Many antiquaries drew, and even now some of Scotland’s current historical institutions continue to draw analogy between Mons Graupius and the fateful Jacobite battle at Culloden in 1746 close to the Moray Firth near Inverness.

Here they envisage(d) tribal / highland armies being overwhelmed under similar circumstances by more organised and disciplined lowland forces.

In so identifying a common cause with the site where the Jacobites turned at bay in 1746 reliance is usually placed on what unfortunately is the least reliable element of Tacitus’s “The Agricola”; the pre battle speeches.

These speeches, undoubted pieces of literary invention, were inserted in the work as a piece of ancient convention, all aimed to spice up the tale. They have Agricola “chasing down” the last of the free tribes, a sentiment post Jacobite commentators empathised with, it was to all intents what the Government forces had done since crossing the Spey in 1746.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Possibly, a fort has been suggested at Easter Galcantray near Cawdor.

Is the Flavian activity in the area likely to be Agricolan?

No, activity so far north is isolated. Easter Galcantray could possibly be the terminal destination of Lucullus circa 30 acre string of camps.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

No

Is the marching camp located close to the site?

N/A

Does the marching camps position and orientation “address” the site?

N/A

Does the marching camp display Flavian characteristics?

N/A

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

N/A

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, too large. It is an expansive and fairly flat moorland with no real size constraints.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The site and its immediate environs lacks a hill.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

The moorland is relatively flat.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? N/A
A realistic Roman assault up? N/A

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Marching on to the Moray Firth would be considered a widening of operations.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, similarities with the name of the Moray town of Forres have been suggested.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Unfortunately sites so far to the north suffer – based on our current knowledge – from only the most intangible of Flavian remains and these – due to their small size – post date the turbulent events of 83 AD.

No securely datable remains of Flavian marching camps of sufficient size are known extending through Moray to the Moray Firth.

Of exceptional interest in defining the probable limits of Roman Imperium in the north are the recent findings at the tribal foundry recently unearthed at Culduthel farm near Inverness (link to Iron masters of the Caledonians).
This shows that the site appears to have prospered through this period, not an outcome we would expect for what appears to have been a tribal arms production centre if the fateful show-down with the Romans took place a mere few miles away.

Persistent attempts to build a link between Rome’s attempts to conquer the tribes of Scotland and the 18th C British Government’s attempts to impose control of the Highland zone following the “45” through the creation of roads, bridges and forts is more imagined than real.

Culloden, at only 3.5 appropriate answers out of a possible 19 rates as being “not worth serious consideration”, a verdict some of the historic institutions in Scotland currently whimsically suggesting a site for Mons Graupius “somewhere near Inverness or even beyond” should now take seriously on board.

Ultimately Culloden’s historic inclusion as a contender is an example of how an academic analogy can just be taken too far. The site itself certainly bears no resemblance to the battlefields description by Tacitus.
Critically it even lacks an appropriate hill (unless we re-orientate the field to the low escarpment at Balloch below Culloden Muir), a factor which to date has been little considered.

Dalginross

Introduction

Dalginross near Comrie in Perthshire – on the edge of Venicone / Caledonii territory was championed by Gordon in 1726, who quaintly stated;

“…in fine…..to an Antiquary this (Dalginross) is a ravishing scene”.

Gordon had the auxiliaries, on the eve of the battle, squeezed into the marching camp there while the cavalry were supposed to have been billeted in the permanent fort. The Legionaries seem to have been a bit of a nuisance, and in light of Gordon’s already creakingly overstretched accommodation were simply overlooked in his account!

Gordon’s account waxes lyrical about the nature of the surrounding hills, but remains frustratingly imprecise about which hill in particular he considered the Caledonians to have occupied nor where the fighting took place.

The camp and fort, although of Flavian date can be discounted of having a place at Mons Graupius, their capacities being far too small.

The hills pose more of a problem. Roman Scotland visited the site in 2007 and 2008 and on both occasions failed to find a suitable site within the Comrie “basin” where the events related by Tacitus could take place.

The site is surrounded menacingly on all sides by hills, and on two sides by the River Earn and the Water of Ruchill. Except for a narrowing fillet of land fronting Barr Dubh the area simply does not have flat ground sufficient for the Romans to deploy nor for the opening phases of the battle to be played out.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes

Is the marching camp located close to the site?

Dalginross; on site,
Dornock; 11 km away.
Innerpeffray 1; 12.5 km away.
Innerpeffray 2; 13 km away.
Strageath; 12 km away.

Does the marching camps position and orientation “address” the site?

Dalginross is in a central location surrounded on all sides by hills.
Others; no.

Does the marching camp display Flavian characteristics?

Dalginross and Dornock; yes,
Others; Strageath and Innerpeffray camps display late Roman marching camp morphology and site structural sequence.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No
Dalginross; too small, at 22.3 acres it can accommodate only 5,000 men.
Dornock; too small, at 23.2 acres and can accommodate only 5,100 men,
Innerpeffray 1; too small, at 67.3 acres it can accommodate only 15,000 men,
Innerpeffray 2; too large, at 130.5 acres it can accommodate 29,000 men,
Strageath; too small at 32.4 acres and can accommodate only 7,200 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The site does not readily relate to the detail of Tacitus account.

None of the surrounding hills except Barr Dhub can be approached without crossing either the River Earn, Water of Ruchill or River Lednock, none of which was mentioned by Tacitus.

The undulating escarpment of the northern flank of Barr Dhub has almost sufficient space for the Romans to form up at its base however its undulating profile does not suit a Caledonian battleline with acute reverse slopes hampering any reasonable arraying of Caledonian forces there.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

No

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? No
A realistic Roman assault up? To lower slopes only.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Doubtful, with wounded it would be a minimum 2 day march along the Earn to the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Exactly where Gordon’s senses were “ravished” remains a mystery, and having walked the area several times we believe that like Fendoch and Antiquarian promoted sites elsewhere the practicalities of suitably accommodating the forces on an appropriate battlefield were simply overlooked by Antiquarians fired with zeal and a misdirected excess of enthusiasm.

We shall return to Dalginross again though, its story is not yet fully told.

Dunning

Introduction

Feacham proposed the site of Dun Knock on the fringes of the town of Dunning in Strathearn in 1970.
The extent of the marching camp at Dunning was at that time improperly understood and Feacham suggested that it was a small Stracathro gated Roman marching camp that faced the small hillock of Dun Knock near the Duncrub Burn. In so doing he hypothesised an alternative reading of Tacitus to suit the smaller scale of the action that the (alleged) small camp and small hill would allow.

The same year however St Joseph using aerial photography found that the marching camp was in fact much larger, and indeed did not have clavicular gateways. The battle here therefore – not least as St Joseph was soon actively promoting the site at Bennachie – was dismissed as Dun Knock was – rather unfairly – considered “an insignificant pimple”.

Further findings on the site proved of continuing interest. Aerial photography again demonstrated that the small hillock was the site of a multi vallate iron-age hill fort, proving therefore that the Duncrub Burn took its name from the fort. Dun Knock therefore was a later Gaelic name simply meaning Fort Hill.

One of the factors that had originally drawn Feacham`s attention was the work of Watson in 1926 who philologically linked Duncrub to the early medieval battle of Dorsum Crup.

Dorsum Crup is an excellent etymological link to Mons Graupius, Dorsum being a later Latin reference to ridgeline. Crup is to all intents the same as Croup, the original root word we are searching for.

At some time in the intervening centuries, certainly in the post medieval period, the “P” has changed to a “B”. Like Croupie Craigs at nearby Carey this stands as evidence of the highest calibre for the original place name of Mons Graupius.

Roman Scotland visited the site in 2007 and on several occasions in 2008. The hillock of Dun Knock on the southern fringe of Venicone territory is indeed diminutive and could not possibly have held 30,000 Caledonians in battle array.

The land behind Dun Knock however – mentioned only fleetingly by Fraser who was held in thrall with the Gask Ridge in the other direction – is occupied by the Clevage hills and the setting here is outstanding.

It certainly does not rival Bennachie, the Lomond Hills or nearby Craig Rossie in terms of the “picturesque”, however while walking the ground it is clear it would allow the events recorded by Tacitus to be played out with ease. Roman Scotland therefore identified the site on the Clevage Hills in 2007, revisiting the site several times in 2008 to check our findings.

The Clevage Hills are a recognised constituent part of the Ochils “Northern Hills”, a rolling ridgeline (Dorsum) that stretch from Craig Rossie, past Dunning and, fronting the River Earn, along to at least as far as Carpow on the Tay.

The proposed Caledonian position is on the slopes of the Clevage Hills which stretch for some 3km from Middle Third to Craigenroe Hill (circa +290m OD).

The Roman auxiliaries will have deployed out of the side of their camp and marshalled their battleline on the approximate line of the modern Bridge of Earn Road (circa +50m OD) centred near Garvock.
An initial Roman deployment 1.5 km wide will indeed have been menaced by the length of the Caledonians position on the Clevage Hills and a redeployment to 2.7 km width will have to an extent countered this threat.

The local feature Bogtonlea suggests that boggy ground lay to the Romans flank and rear near Nethergarvock which explains the position chosen by the Romans for their camp on the slightly higher ground fronting the hillfort on Dun Croup / Crub – fort at the Croup or as it latterly came to be known in Gaelic speaking Scots times; Dun Knock.

The legionaries and Agricola’s cavalry reserve may have deployed out of the front of the camp, and were probably positioned here behind Dun Knock – crowned with its multi vallate hill fort – where they would be hidden from the Caledonians on the Clevage Hills but located sufficiently close to be able to intervene in a matter of only a few minutes if required.

Tacitus tells us they were held in reserve and Maxwell in 1990 cogently argued that the cavalry reserve (at least) had to be held in a position where they would be hidden from the Caledonians in order for their eventual counter attack to be launched to such cataclysmic effect, no doubt on account of the surprise of their sudden appearance on the battlefield at the critical moment.
These are points which we shall return to.

This location also convincingly explains Tacitus own much misunderstood phrase, the legions were stationed “pro vallum” . This phrase is usually now – not entirely satisfactorily – translated as “in front of their marching camps defences”. A superior reading would fit at Dunning; i.e. before the (hillforts) ramparts.

The Caledonian position is bounded on its flanks to the west by the Dunning Burn and further to the east by the Water of May. Behind the Clevage Hills ridgeline lies a high plateau and this is protected by the curving course of the Water of May. This would make an excellent and readily defendable Caledonian hosting place, as well as being suitable ground for the post battle events recorded by Tacitus to unfold over.

Dunning in Strathearn is only slightly off the Romans likely axis of advance along Strathallan, but close enough for the Caledonians to have monitored the Roman’s movements and easily move to intercept them if required, or alternatively to use the sites central location to outflank the Romans in one of several possible directions.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes

Is the marching camp located close to the site?

Yes

Does the marching camps position and orientation “address” the site?

Yes

Does the marching camp display Flavian characteristics?

Yes

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, at 116.3 acres it can accommodate 25,800 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the Clevage Hills between Middle Third and Craigenroe Hill extends to 3km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No. The features of the site are an extremely good fit to Tacitus description. The multi-vallate hillfort on Duncrub hillock in front of the marching camp could explain Tacitus description of legionary deployment pro vallum – before the ramparts. The watercourses on the site are generally small or of no consequence to the events described by Tacitus.
The dip in the ground and burn near Ha’ castle is behind the eventual Roman battleline and would not have had a great influence on events.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, very appropriate.
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, an upland setting with broken ground.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be addressed locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, matching sized and shaped camp at Carey (11 miles = 1 slow days march) on the Earn, then Carpow beyond at the Earns outflow into the Tay.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Yes, Duncrub has been philologically equated to derive from the early medieval battle of Dorsum Crup.
Dorsum = spine or ridge, Crup is a recognised derivative of Croup. This is excellent etymology and is an exceptional piece of evidence.

Summary

Dunning is unique among all the contenders in achieving 19 appropriate answers out of a possible 19. This is outstanding and of exceptional importance.

The cumulative importance of;

a camp which is of correct size and Flavian morphology,
a site with the best philological links,
a site of an appropriate size and lacking features inconsistent with Tacitus account of the battle,
a tribal mustering point sensibly located on the southern reaches of the threatened tribes lands and one from which the tribes could undertake one of several strategic options depending on how the Romans advanced into their lands

….cannot be understated. 19 out of 19 (100%) rates this contending site as one with exceptional credentials.

Fendoch

Introduction

In 1778 Colonel Shand identified a putative Roman marching camp of around 113.5 acres, located between the River Almond and Fendoch Burn near Buchanty in Logie Almond, an identification Christison concurred with in 1898. Crawford however queried the identification in 1925 and the putative camp currently remains unproven.

(The site subsequently gained fame in the late 1930`s when archaeologists explored the permanent fort here, the digs findings enabling the post to be hailed as the most completely known Roman fort anywhere at that time).

Logie Almond skirts Caledonii territory. Shand proposed that the marching camp was Agricola’s base at Mons Graupius and that the fighting took place on Dun Mor, principally as it is crowned with an iron age fort and guards the narrow southern entrance to Glen Lyon – a location known as the Sma` Glen.

Roman Scotland visited the site (circa +200m OD) in 2008, and noting the extremely difficult and constricted ground fronting Dun Mor (+466m OD) identified Stroness Hill (+336m OD) to the camps immediate south as a potential alternative candidate site here.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

The permanent fort here is attributable to the earliest years of Sallustius Lucullus Governorship.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes,
Colonel Shand speculated that one was on the site of Fendoch in 1778 but this remains unproven.
Innerpeffray 1 and 2 are approximately 9 km away.

Is the marching camp located close to the site?

Proven – No

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

No, both Innerpeffray camps display late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No
Innerpeffray 1 is too small, at 67.3 acres it can accommodate only 15,000 men,
Innerpeffray 2 is too large, at 130.5 acres it can accommodate 29,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Dun Mor has an extreme highland setting and does not have adequate space to allow Roman deployment.

The frontage of Stroness Hill extends as a ridge but similarly does not have sufficient space at its foot for the Roman battleline to deploy.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The River Almond intervenes between Fendoch and the lower slopes of Dun Mor.
The Fendoch Burn crosses the foot of Stroness Hill.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Neither does (the plain at Fendoch is behind the fort as viewed from Stroness Hill).

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? No, both are quite steep.
A realistic Roman assault up? The Romans would likelier assault along the Stroness Hill escarpment from the south west where gradients are easier. To gain easier gradients Dun Mor would be assaulted on a frontage (too narrow) between Dallick House and Dunie.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, Tay at Perth.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

At only 6 appropriate answers out of a possible 19 neither site at Fendoch rates any better than “highly unlikely”.
Neither site has convincing topography to match Tacitus account, Dun Mor in particular is particularly ill suited, being too narrow, bisected by the River Almond with no plain whatsoever for the recorded chariot action.

The site is however worth visiting to view one of Sallustius Lucullus “Glen Blocker” forts, set against a truly spectacular backdrop.

Fortingall

Introduction

Fortingall is located in Glen Lyon, north of Loch Tay in Caledonii territory.

The noted antiquarian Horsley, in 1732 suggested the location as the site of the battle as he felt that the action must have taken place within the Highland massif itself.

Horsley was undoubtedly attracted to Fortingalls’ traditional Roman associations – the site is reputed to be Pontius Pilates birthplace – and he claimed that the small rectilinear earthwork there was Agricola’s headquarters – Horsley’s “Proscenium” – at the battle. In fairness to Horsley the site fooled Roy too who planned the earthwork in Military Antiquities of the Romans in North Britain published in 1793.

The earthwork however is medieval, and Fortingalls ancient association with Rome is religious and certainly no older than the early monastic community suspected there.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

No

Is the Flavian activity in the area likely to be Agricolan?

N/A

B: Marching Camps

Is there a marching camp in the vicinity of the site?

No

Is the marching camp located close to the site?

N/A

Does the marching camps position and orientation “address” the site?

N/A

Does the marching camp display Flavian characteristics?

N/A

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

N/A

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, there is less than 2km width open ground at Fortingall into which the Romans could deploy.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The site is bisected by the Aalt Odhar watercourse. The River Lyon, a considerable river, is situated immediately behind the Roman position (if facing Balnacraig) or intervening between the forces if the Caledonians were posted on Drummond Hill. Either way The River Lyon would have been noteworthy.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

No, there is insufficient space for the Romans to properly deploy a minimum width battleline let alone a stretched one.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? No, too steep.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, the land behind is mountainous highlands, pursuit would have been directly along Glen Lyon which is not how Tacitus describes the battles aftermath.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No, several days travel along the Tay would be required to reach its navigable reaches.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Fortingall is further into the highland zone than would be expected for a legionary fortress. Therefore a location sited back at Inchtuthil would be very appropriate.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Fortingall scrapes only 2 appropriate answers out of a possible 19 and “does not rate serious consideration”.

The site is too small, is supported by no known marching camps near or approaching it. Again, a contender has been chosen on the basis of an academic hunch, this time placing it in the Highlands while sitting piggyback on spurious local traditions of Roman association which unfortunately have not stood the test of time.

Fortingall is a beautiful and secluded spot but identifying it as Mons Graupius is no longer viable.

Gask Ridge

Introduction

Fraser’s excellent publication; The Roman Conquest of Scotland, The Battle of Mons Graupius identified the Gask Ridge as a contending site in 2005.

He identified the camp at Dunning as being Agricola’s base prior to advancing across the River Earn and assaulting a suggested Caledonian position on the rolling Gask Ridge escarpment immediately behind.

In justification of a site located on the Gask Ridge – a site of both proven Flavian as well as several phases of later Roman activity – Fraser borrowed the philological rendering of Moncreiffe Hill at Perth – Monid Croib – some miles distant to this escarpment.

Fraser’s site, involves a riverine assault across the Earn to the east of Innerdunning on low haughland (circa +10m OD) leading up to the concave heights of the Cairnie Braes (circa +130m OD) behind.

Fraser’s is a seminal work on the motivations and political pressures surrounding and influencing Agricola’s actions and is a highly recommended read.

Roman Scotland visited the site in 2008. Sufficiently concerned with the topography of the proposed section of the Gask Ridge we propose after close examination of the ground an alternative Gask Ridge site, close by on more practical ground to the west of Dalreoch Bridge centred on Hilton of Gask (Circa +20m OD rising to circa +100m OD).

This allows more practical gradients and sufficient space free of the course of the River Earn – even if its course has meandered over the years as Fraser suggests – for the plain at the foot of the hill mentioned by Tacitus which is conspicuously absent below the Cairnie Braes.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Forteviot (adjacent), Dunning (3.5km) and Innerpeffray (6km)

Is the marching camp located close to the site?

Dunning; No
Forteviot; Yes
Innerpeffray; No

Does the marching camps position and orientation “address” the site?

Dunning; Yes
Forteviot; Yes
Innerpeffray; No

Does the marching camp display Flavian characteristics?

Dunning; Yes
Forteviot; No, camp displays late Roman marching camp morphology.
Innerpeffray; No, camp displays late Roman marching camp morphology and structural sequence on its site.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Dunning; Yes, 116.3 acres accommodates 25,800 men.
Forteviot; No, 63.3 acres accommodates only 14,000 men.
Innerpeffray; No, 130.5 acres accommodating 29,000 men and is too large.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No. Fraser’s suggested battlefront extends beyond 3.5km and is centred on Drum of Garvock. The site however could easily accommodate a battleline considerably longer with the Roman left flank still remaining uncovered as the Gask ridge to the west is a long escarpment.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Yes. The River Earn, a considerable and noteworthy obstacle flows at the foot of the Gask Ridge. The river hereabouts has cut steep embankments that would not allow general fording simultaneously across an army’s broad frontage.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

No. The River Earn may have changed course slightly over the years but in general terms it does not allow a plain of sufficient size on the identified site. Better level ground is available on the plain west of Dalreoch Bridge either before or after the Earn is crossed.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Cairnie Braes, especially at Wester Cairnie is quite steep. More suitable gradients are available to the west of Dalreoch Bridge.
A realistic Roman assault up? As above

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, the land to the immediate rear of the Gask Ridge is fairly flat and featureless.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, the Tay

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, Carey and Carpow, though Fraser does not use these suggesting – contrary to Tacitus – further harrying to the north took place before Agricola finally withdrew south.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Fraser suggests the philological root of Moncreiffe Hill – Monid Croib – should also be applied to the Gask Ridge. Croib – possibly Gaelic for tree- has some linguistic similarity to Croup and in Gaelic form may possibly echo an older Brythonic place name.

Summary

The Gask Ridge performs very well. Fraser’s traditional site achieves 14 appropriate answers out of a possible 19, securing an “interesting” rating.
Our alternative Dalreoch site addresses some of the problems inherent with the traditional site. This site achieves a very creditable 16 appropriate answers out of a possible 19 rating this site as a very “strong contender”.

There are however problems.

The borrowing of good philological evidence that belongs instead to Moncreiffe Hill does unfortunately seem like a sleight of hand.

Fraser’s work however is robust enough to attempt to address the major failing in the site – the need to assault across the River Earn – a noteworthy event Tacitus is worryingly silent on.

In proposing an initial assault to secure a bridgehead – similar to such assaults that took place in 43 AD in southern England and by Agricola at Anglesey using Tungrian and Batavian units – he notes the precedence of using these regional troops in such riverine assaults.

In so doing Fraser is the first to attempt to explain the reason behind Agricola’s decision to initially advance only a part of his battleline into contact.

Fraser’s reasoning is good.
However, it should be pointed out that the Batavians and Tungrians previously employed in such marine bridgehead assaults were cavalry, not the infantry cohorts that Agricola advanced at Mons Graupius.

Further, and more worryingly the Earn has cut very deep banks hereabouts that do not suit an army fording it on anything other than an extremely narrow frontage. That the Earn has not changed course dramatically is proved by the position of the adjacent later marching camp at Forteviot. All of which begs the question; why Agricola did not use the ample ground available there instead of Dunning which is too far away to control the site?

Finally, and most tellingly we must query why Agricola, intent on assaulting the Caledonians openly mustered on the Gask Ridge, marched off the beaten path up Strathallan to camp at Dunning, only to face a difficult cross river assault up the ridge along which the later Roman road proves they used and were aware of as a ready means of communications?

Very simply put, unless Agricola was operating further along the Earn and caught wrong footed it makes no sense why Agricola would deploy in such a manner when he could simply advance along the line of the later road (its southern elements may even have been put in hand at this early stage) up Strathallan allowing an easy assault of the Caledonian position along the spine of the ridge itself and in so doing negating the need to cross a river and climb a hill in the face of the enemy!

The Gask Ridge is a highly recommended site to visit and while pondering the implications of such speculated manoeuvres it provides the visitor with excellent views, being located in the cock- pit of Roman involvement north of the Forth – Clyde line.

Glen Eagles

Introduction

Sibbald, as long ago as 1707 highlighted the possibility of Agricola leaving his camp (sic) at Ardoch and launching an assault on the high ground of the Ochils to the east.

Much later, Pitblado, in a privately published work in 1935 suggested Glen Eagles.
It is uncertain exactly the ground he intended, and his theory has been ridiculed both as Gleneagles is a particularly well known golf course (it is unknown if Pitblado was a keen golfer) coupled with his rather off- beat theory which inexplicably had spot inundation of sea levels 30m higher than today!

Like the Rev`d Small at Merlsford however, we cannot allow Pitblados minor lunacies to jaundice us against fairly reviewing the site and in 2008 Roman Scotland visited and identified the most probable site in this locality.

Taking rather more inspiration from Sibbald we looked at the rolling hills behind Bardrill the northern flank of which forms the Glen Eagles – nothing to do with the golf course and separated from it by the A90.

This site, on the southern fringes of Venicone territory rises from circa +150m OD to +503m OD on Wetherhill, and has a small iron-age fort and several standing stones nearby.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, at Ardoch.

Is the marching camp located close to the site?

No, 6.5 km away.

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

No, the marching camp displays a late structural sequence on its site.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, at 119 acres it could accommodate 26,400 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, a width of just over 3 km is available on the rolling escarpment of Wether Hill and Craigentaggert Hill between the Ruthven Water and the Braes of Ogilvie.
The gradient of these slopes rolls for over 1,500m, this would give the Caledonians too deep a formation between front ranks on the plain and those located on higher ground for effective control to be exercised and maintained.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No. The slope of the hill is cut by several small burns running down the slope though these would have only minor impact on a Roman deployment and advance and the recorded events that followed.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken uplands in the Ochil hills.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, Tay or Forth.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, Dunning, Carey and Carpow.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Glen Eagles, the long pilloried candidate site, clocks up an impressive 15 appropriate answers out of 19 possible answers placing this site as a “strong contender”.

Obvious failings are the lack of a correctly sized Flavian period camp in its immediate vicinity – not the late Roman one at Ardoch that the site benefits from its association with. Also the hill stretches to a considerable depth for Caledonian command and control to be effectively maintained.

However notwithstanding the above, the site, based in an area of proven Flavian activity has noteworthy credentials that at the very least merit a less dismissive approach than it has otherwise suffered from to date.

Hill of Bruxie

Introduction

After proposing Mondboddo, Roy in 1790 hedged his bets somewhat by then suggesting a site further north – “Perhaps even nearer Stonehaven than Mondboddo”.

Suitable ground becomes increasingly limited as the low lying countryside narrows towards the choke point of the Mounth at Stonehaven, however Maxwell in 1990 highlighted the Hill of Bruxie as possibly being the feature Roy had in mind.

Like Mondboddo, features such as the Hill of Bruxie will have been noted by Roy’s fellow officers during their march north following the Jacobite forces in the closing stages of the “45”.

The Hill of Bruxie sits in a commanding position over the traditional road here and the likelihood of the 18th C Jacobite army turning at bay here would have been a possibility the pursuing British Army’s staff corps would have been aware of. Roy, as at Mondboddo is likely to have considered this a potential situation which may have had precedence in antiquity at Mons Graupius.

Hill of Bruxie is a large freestanding hill of classic convex plan deep in Vacomagi territory. Its peak at +216m OD commands the plain below at around +100m OD and is visible for miles to the south as well as from north of the Mounth.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the circa 30 acre string of camps arcing through Aberdeenshire to Moray do not hold sufficient capacity and post date 83 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, at Kair House.

Is the marching camp located close to the site?

No, 4.5 km away.

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

One may possibly underlie the known camp which displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No the speculated Flavian camp at around 92 acres is not large enough and could only accommodate 20,400 men

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the lower southern girth of Hill of Bruxie is about 3km long. The hill however has a distinct convex profile and a depth which would give the Caledonians too deep a formation between front ranks on the plain and those located on higher ground for effective control to be exercised and maintained.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, only minor watercourses are present.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, Hill of Bruxie is a conspicuous solitary hill located in a fairly lowland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Bervie bay would be the closest, Stonehaven would constitute a further advance north.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Hill of Bruxie achieves 9 appropriate answers out of a possible 19, which unfortunately rates this site as “highly unlikely”.

This hill better than anywhere else demonstrates the difficulty that would be faced by Caledonians deployed on a convex hill, with the difficulties inherent in a wide – circa 3 km – deployment leaving them unavoidably facing radially outwards and in a poor position to contemplate enveloping their opposition.

On the other hand its conspicuous solitary setting would allow a predatory Roman general great latitude in choosing which direction to assault the hill from, a sensation readily apparent to the modern visitor circumnavigating the hill.

Lastly, the low rolling lowlands hereabout do not convincing match Tacitus account of the events following the battle.

Hill of Edzell

Introduction

Maxwell identified the strategic importance of the area around Edzell in Angus in 1990.

Roman Scotland visited Edzell in 2007 and 2008 and identified the Hill of Edzell, located behind the town of Edzell as a site worthy of careful consideration.

This hill is located in a pivotal position on the North Esk – a major watercourse. The hill bounds the lands of lowland Strathmore – along which the Roman advance would be made – while also plugging the gap leading through Glen Esk (ultimately) to the Cairngorms. This would therefore be an ideally located assemblage point for the Caledonians of the highland zone to meet up with their lowland confederates.

Further the hill, rising from around +70m OD to +228m OD sits commandingly facing the choke point between the West Water and North Esk. This strong position, stretching from Gannochy to Edzell Motte, with well guarded flanks is of a suitable size for the battle.

The Romans subsequently acknowledged the strategic importance of the area, building the Flavian fort at nearby Stracathro. This fort is currently the most northerly securely identified Roman permanent installation anywhere in the Empire.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Possibly, however the Flavian marching camp and fort probably date to Sallustius Lucullus term as Governor.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Stracathro and Keithock.

Is the marching camp located close to the site?

No, both are located circa 5km away.

Do the marching camps positions and orientation “address” the site?

No

Do the marching camps display Flavian characteristics?

Stracathro; yes, this camp is the clavicular gate “type site”.
Keithock; no, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Stracathro; no, at 39.3 acres this could accommodate only 8,700 men.
Keithock; no, at 63.3 acres it can accommodate only 14,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

A battleline, centred on Hill of Ezell and stretching between the West Water and The North Esk would have a frontage around 3km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, the excellent harbourage of the Montrose Basin, in itself the best candidate for the lost site of Trucculensis Portus.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes, excellent location and Flavian connections at Dun on the Montrose Basin.

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is not too far in the rear of this contending site to be still considered acceptable.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

The Hill of Edzell is an intriguing possibility and at 12 appropriate answers out of a possible 19 it rates as “interesting”.

Like other sites in Strathmore however the inescapable conclusion is that any tribal hosting here, while having good communications with the Caledonian septs of the highland zone would however have been bought at the cost of the tribes abandonment of far too much valuable lowland territory to the Roman torch– indeed all that of the Venicones and much of the Vacomagi.

The suggested site though is interesting and the battle very “fightable” here. Indeed we can imagine Agricola’s dense columns passing the funnel point where the space between the North Esk and West Water is at its narrowest to be met by a wave of Caledonian chariots and the grave concern mentioned Tacitus mentions Agricola had for his flanks as the battlefield rapidly increase in width.

No appropriately sized Roman Marching camp is known in the immediate vicinity, however further investigation of the area around Edzell wood to Edzell town itself may perhaps pay dividends as such a location would be a strong one and would directly address the suggested Caledonian position.

Huntlyhill

Introduction

Roy in 1790 expressed interest in the marching camp at Keithock near Brechin in Vacomagi territory.

Roy had formerly shown some interest in the large camp at Oathlaw – Battledykes (probably on the northern boundary of Venicone territory), but was intrigued by Keithocks implication of a further Roman onward advance beyond Oathlaw and the South Esk.
Maxwell in 1989 followed this lead up, hypothesising that the length of ridge above the Roman fort at Stracathro – Huntlyhill – could be the site of the battle.

This area, before the crossing of the North Esk is highly strategic, the Romans themselves recognising this with the fort they subsequently erected there, probably by Sallustius Lucullus in the year(s) following Mons Graupius.

In suggesting the site of Huntlyhill – itself the scene of a later medieval battle – Maxwell speculated that Stracathro may have been the Roman post subsequently called “Victoria”.

A large boulder elevated on a simple stone plinth on the crest of Huntlyhill and visible for some distance is commonly hailed to mark the scene of the medieval fighting and echoes Keithocks alternative name of Battledykes.

The suggested Roman forming up position would be on the level ground around +60m OD currently occupied by the A90 dual carriageway with the Caledonians on the slope beyond which rises to around +120m OD.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Possibly, however the known Flavian marching camp and fort probably date to Sallustius Lucullus term as Governor.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Stracathro and Keithock.

Is the marching camp located close to the site?

Yes,
Stracathro; just over 1km.
Keithock; adjacent.

Does the marching camps position and orientation “address” the site?

Stracathro; yes
Keithock; yes / short face-on.

Does the marching camp display Flavian characteristics?

Stracathro; yes, the clavicular gated marching camp “type” site.
Keithock; no, it displays a later Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Stracathro; no, at 39.3 acres it can accommodate only 8,700 men.
Keithock; no, at 63.3 acres it can accommodate only 14,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Huntlyhill is located on a lengthy running escarpment, there is nothing to define or limit the width of any battle fought there.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Stracathro is located behind the Cruik Water.
Keithock; no.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes, but this is restricted to the narrow strip occupied by the A90 or the site of the later Keithock camp site.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Very appropriate.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, it is a fairly open landscape leading to the coast.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, at the Montrose Basin.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes, excellent location with Flavian connections at Dun on the Montrose Basin.

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is behind but not too far in the rear from this contending site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Huntlyhill is a good – if not readily accessible – site to visit not least as it amply demonstrates the kind of hilly ground upon which men were prepared to venture battle in the past.
Those who are determined to support locations on hills with very steep gradients and mountainous profiles are encouraged to visit and walk this site, then return to their preferred step site and then walk that one again in comparison.

Even unencumbered without equipment and armour the relatively gentle slope of Huntlyhill is tiring to walk up and would still be a challenge to the Romans making an uphill assault.
However, as mentioned, Huntlyhill forms a part of the continuous ridge that the A90 shadows while passing through this part of Angus, and it is by no means certain that the conflict in 1452 AD involved any sort of uphill fight. It is more probable that this later fight was along the crest of the ridge, a possibility that amply demonstrates the weakness of the hills flanks to the most basic of Roman flanking attacks.

At 15 appropriate answers out of a possible 19 Huntlyhill rates as “interesting”.

Inverquharity

Introduction

Maxwell also speculated in 1990 on the battle taking place on ground lying close to the fort at Inverquharity – near Kirriemuir in Angus- which is situated to the north west of the River South Esks crossing points around Forfar.
The Forfar area “crossings” were used – based on later marching camp remains – on several occasions by Roman armies on campaign.

Inverquharity`s position however, off this apparently direct line of march on the map, was according to Maxwell, suggestive of Roman interest in this “forward” area that could have been based on the battle having taken place here.

He expressed an interest in Cat Law (circa +678m OD), an outlier of the Grampian mountains sited next to Glen Prosen and interestingly close to the area that may perhaps have been the junction between the lands of the Venicones, Vacomagi tribes and the Caledonii.

Roman Scotland visited the site in 2008 and found Cat Law too mountainous and far removed from the sort of level plain Tacitus recorded for the initial phases of the action.

Roman Scotland however identified an alternative site fronting Caddam Wood, close to the famous Kirriemuir golf course. The notable escarpment here, comprising the low rolling hills of Meams, Castle, Culhawk and Kirkton Hills (circa +230m OD) presents a very plausible Caledonian position. The size available for the suggested battlefield is very convincing and the flanks of the ridge are sufficiently strong to assist it in being used as a credible Caledonian defensive position.

The low valley (circa +150m OD) between the suggested Roman and Caledonian positions is interrupted by only minor watercourses and is suitable for Tacitus description of unfolding events.

Further, and most tellingly the proposed Roman position sits astride Caddam Wood, where Keppie – a very reliable source – places a stretch of probable Roman road. This is quite compelling evidence, and, if as Maxwell suggested Cat Law was utilised as a hosting point for the Caledonians, then the low ridge mentioned above would be a suitable location to come down to and offer battle to the Romans marching along the route the (probably only slightly) later Roman road followed.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Possibly, however the known Flavian marching camp, putative road and fort probably date to Sallustius Lucullus term as Governor.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, at Inverquharity.

Is the marching camp located close to the site?

Yes, it is no further than 2km away.

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

Yes

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No, at 5.7 acres it can accommodate only 1,200 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the low escarpment formed by Meams, Castle, Culhawk and Kirkton Hills extends to an effective 3.5 km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, there are only minor watercourses.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, very appropriate.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken uplands leading into highlands.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No, a minimum of 2 days march to either the Montrose Basin or the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, fronting the highland massif along the course of the River Isla.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Inverquharity is a smashing site. However with only 12 appropriate answers out of a possible 19, it merely attracts a rating of “interesting”.

The site suffers – as do many others – in not having a suitably sized Flavian marching camp in its immediate vicinity.

The putative stretch of Roman road here – most likely a short lived later Flavian creation – makes it clear however that much around this location still remains to be found.

Its position, at the north western edges of – if not beyond – Venicone territory however does beg the question of it being too far north to maintain a tribal confederations unity in the face of the advancing Roman column and the torch they will have put to over-run tribal homelands.

Kempstone Hill

Introduction

Maitland, impressed with the remains of the Roman Marching camp at Raedykes – north of the Mounth at Stonehaven in the southern fringes of Taexali territory – was the first to identify the site in 1757.

Roy, who generally favoured sites in Strathmore, is known to have visited Raedykes and the site was subsequently included by the publisher in his posthumous work in 1793 – Military Antiquities of the Romans in North Britain.

Stuart in 1822 was next to hail the site, identifying an enclosure (now lost) to the south at Arduthy as Agricola’s camp before the action, and Raedykes as his camp after the fight.

Finally Crawford, who appears to have hung somewhat on the coat tails of the legendary MacDonald (who drew the site up early in the 20th C) claimed, in 1943, that Raedykes (+200m OD at its highest) was Agricola’s base before the battle. Crawford was interested in the cairn field – which he incorrectly interpreted as the remains of a “Caledonian village” – on Kempstone Hill (rising from +70m to +132m OD) and which, he claimed was the scene of the fighting.

Until interest was generated in the Pass of Grange in the following two decades, Kempstone Hill was held, on account of the high academic esteem in which its various proposers were viewed, to be the generally accepted site of the battle of Mons Graupius.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the string of circa 30 acre sites that pass this location are likely to date to Sallustius Lucullus term as governor and post date Mons Graupius.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes

Is the marching camp located close to the site?

No, 3km away.

Does the marching camps position and orientation “address” the site?

No, it addresses Meikle Carewe to its NNW and further into Vacomagi territory.

Does the marching camp display Flavian characteristics?

No, the camp displays late Roman marching camp morphology, showing two distinct structural phases of occupation.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No, probable earliest camp – Severan – at around 140 acres is too large and could accommodate over 31,000 men.
Late Roman re-use of site at circa 105 acres is slightly too small and accommodates a force of 23,300 men – or less due to nature of the terrain.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, Kempstone Hill has a frontage of less than 1.5km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Only White hill and Mondboys Burn, neither of which are major obstacles.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Very easy gradient.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, the north sea coast is directly behind the site.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Immediate access is available at Stonehaven, not a noteworthy march away.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Raedykes is arguably the finest marching camp in Scotland to visit. It has a fascinating late Roman site structural sequence and is one of the key sites that assist in identifying the progress of several late Roman armies on campaign in lowland north east Scotland (article forthcoming).

It is the fine quality of its remains, its clearly strategic siting, coupled not least with the fine vistas available from the top of Garrison Hill within it that undoubtedly inspired the interest from those who proposed its involvement in the battle.

Crawford’s interpretations, however are not without problems (see his dismissal of Tillymorgan and Peterculter) while his Caledonian “village” has already been mentioned.

In the context of the search for Mons Graupius, Kempstone Hill is noteworthy perhaps best for being extremely unremarkable. It does not sit well at all with the spirit of Tacitus account, being both diminutive and lower in height than the ground on which Crawford would has us believe Agricola encamped.

Indeed, Kempstone Hill is so unimposing that from the summit of Garrison Hill it takes the assistance of a map and compass to actually identify which one of the low rolling gorse covered hillocks it actually is!

Further a Caledonian deployment on Kempstone Hill, which appears on paper to be a well sited strategic block to the (modern road) north, makes however – with the cliff lined north east coast worryingly hard behind – for uncharacteristically poor tactical judgement by the Caledonian tribal elders. And this is notwithstanding that with a marching camp situated at Raedykes, the Romans can be considered to have already breached whatever strategic impasse the Caledonians may have intended to place in their path to prevent them penetrating north of the Mounth.

Again, like all sites located north of the Mounth, a federated tribal army of the northern tribes ensconced on Kempstone Hill would have had plenty of time to watch the Romans devastate Venicone and Vacomagi territory to the south, as well as allowing an unacceptable degree of vulnerability to the glens leading into the Caledonii highland heartlands stretching north west from Stirlingshire.

Kempstone Hill secures only 7 appropriate answers out of a possible 19 and achieves a rating of “highly unlikely”. This is a quite telling outcome for a previously accepted location, albeit one which unfortunately benefitted from distinctly partial readings of the historical sources.

Knock of Crieff

Introduction

Roman Scotland identified the site of the Knock of Crieff in 2008 while following a line of philological enquiry that had previously led to quite promising results at Moncreiffe Hill.

The word Knock has been raised elsewhere (Pass of Grange) but it is in fact a fairly common and unremarkable Gaelic appellation for a hill in Scotland.

The modern name of the Perthshire town of Crieff is possibly the fairly modern anglicised spelling for the Gaelic pronunciation of Croib (the end is pronounced ff in lieu of soft gaelic phonetic b) – which in itself Fraser suggested (at the Gask) may be related to the root word Croup .

This is of great interest and here means;

Knock (noteworthy shaped hill) of Crieff (Croib).

This is perhaps not a perfect fit, but when we compare this to:

Mons (Latin for hill) Croupius (the correct root word from which Graupius is likely to have been arrived at) then we have an interesting piece of etymology worth following up.

Crieff is further of interest on account of its location. It sits on the River Earn, and known Flavian campaigning is heavy along the line of the Earn.

As to its suitability for a tribal mustering point, Crieff`s later reputation as Scotland’s “wild west” where Highland drovers brought their herds down to market amply illustrates its position as a key nodal point in the transhumance between Highland and Lowlands.

The escarpment of Knock Hill – circa +270m OD – also offers itself as a superb position for the Caledonian deployment, and while its lower outlying slopes above Tomaknock would prove to be an obstacle to immediate deployment on the proposed Roman battleline (on the general line followed by the modern A85– circa +130m OD), a force sent to secure it could be plausibly used to explain Agricola decision to dispatch his Tungrian and Batavian cohorts in advance of the rest of his battleline, a tactical decision Tacitus offers no adequate explanation for.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Yes

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Innerpeffray 1, Innerpeffray 2, Dornock and Strageath.

Is the marching camp located close to the site?

No:
Dornock; 4 km away.
Innerpeffray 1; 5 km away.
Innerpeffray 2; 5.5 km away.
Strageath; 4.5 km away.

Does the marching camps position and orientation “address” the site?

No, they are all positioned relative to the River Earn or the Roman road at this location.

Does the marching camps display Flavian characteristics?

Dornock; Yes
Innerpeffray 1; late Roman morphology, questionable site structural sequence however.
Innerpeffray 2; Late Roman marching camp morphology and site structural sequence.
Strageath; Late Roman, probably Antonine.

Is the size of the marching camps reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No:
Dornock is too small, at 23.2 acres and can accommodate only 5,100 men,
Innerpeffray 1 is too small, at 67.3 acres it can accommodate only 15,000 men,
Innerpeffray 2 is too large, at 130.5 acres it can accommodate 29,000 men,
Strageath is too small at 32.4 acres and can accommodate only 7,200 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the Knock of Crieff position extends 2.5km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The quite steep hillock above Tomaknock fronting the site would have to be negotiated by the Romans when forming up.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes except for uppermost slopes.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes long the River Earn to the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

Yes, the Dunning – Carpow series marching camps.

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, ideal.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Knock is simply Gaelic for a humped shaped hill, Crieff could be rendered from Gaelic Croib which may possibly be derived from Croup– see Gask Ridge for fuller details.

Summary

There is little negative to be said about Crieff`s credentials except that it lacks an appropriate camp situated in close proximity to the suggested battlefield. It is its failure in this respect which prevents the site attracting more than 14 appropriate answers out of a possible 19. This still rates the site as “interesting”.

Sufficient suitable ground exists before Tomaknock where a marching camp could be sought. Notwithstanding this, the site is clearly visible from the rash of multi period marching camps at Innerpeffray even though these do not address the Knock of Crieff in their orientation.

That said, Crieff`s philological link, its proximity to areas of proven Flavian period campaigning, the suitability of the size of battlefield as well as the areas proven connection with transhumance between Highland and Lowlands make this a very interesting contender indeed.

Lomond Hills

Introduction

The eye catching profile of the Lomond Hills rising spectacularly above the flat plain of agricultural Fife has long attracted attention.

Merlsford, a site on the River Eden at the foot of the Hills near Gateside was- according to local tradition – the scene of a battle in ancient times. The Rev`d Small in 1823 published an account which claimed this as the site of the battle, a claim later backed by Miller in 1829.
These claims followed the then widely held antiquarian belief that many of the events of 82 and 83 AD recorded by Tacitus in The Agricola took place in Fife.
The seriousness with which the site is now held has been weakened by some of the more extravagant claims made by Small.

However the site merits closer inspection and Roman Scotland visited it in 2008.

A possible site can be suggested. We must ignore the ford on the Eden, a fight at a ford was never mentioned by Tacitus. However the lie of the land beyond on the considerable flat plain below the Lomond Hills is extremely impressive.
Hereabouts the Eden is not a major obstacle and will not have greatly impeded the Roman deployment, nor their ability to retreat in extremis. We may then excuse Tacitus not mentioning it behind the Roman deployment.

Although the Lomond Hills have steep gradients their lower slopes tower above this plain in a most convincing manner.
By restricting a Caledonian deployment to the saddle between East and West Lomond a credible battlefield – incidentally still centred on Merlsford – is now apparent.

Intriguingly – and uniquely in this search – a Roman spearhead held in the National Museum of Scotland is attributed to an auxiliary soldiers grave at Merlsford.

However given Smalls evident willingness to attribute any old features in the landscape as well as any ancient artifacts found a Roman provenance the concern is that the spearhead may have been found locally (there is a marching camp at Auchtermuchty) or indeed may echo a conflict at Merlsford at some other time.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

No, the only currently known Flavian activity in Fife away from the River Earn area is at Bonnytown near Saint Andrews.

Is the Flavian activity in the area likely to be Agricolan?

N/A

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes at Auchtermuchty.

Is the marching camp located close to the site?

No, approximately 3 km away.

Does the marching camps position and orientation “address” the site?

Unknown at time of writing– probably not. Awaiting published plan.

Does the marching camp display Flavian characteristics?

No, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No, it is too small, at 59.6 acres it can accommodate a force of only 13,200 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

The Lomond Hills escarpment stretches for approximately 9 km and is too large. Speculatively, restricting deployment of the Caledonians to the saddle between East and West Lomond peaks centred on the ford at Corston Mill (the traditional battle site) would reduce the width to around 3.5km which is more realistic.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No. The River Eden approximates with the position a Roman Army would deploy on to face the plain at the foot of the Lomond escarpment. Thereabouts it is not a substantial watercourse nor is it an impediment to withdrawal. It is therefore not noteworthy.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, on lower slopes, the upper slopes have steep gradients.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, upland setting of Fife Regional Park.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, to either the Tay or Forth.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes, especially the Tay.

Can the “Boresti” be identified locally?

Yes, Horrea / Horesti is reasonably speculated here as being centred on the Lomond Hills.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No.

Summary

Unfortunately, like Bennachie, a picturesque mountainous setting does not guarantee a sites credentials when rigorously measured against a wide set of disciplined parameters.

With only half of the questions being answered appropriately the site – improved from Small and Millers suggested site – unfortunately only receives a category of “highly unlikely”.

Roman forces marched near this location at some time, and its conspicuous profile makes the Lomond Hills likely as a tribal hosting centre.

The local traditions of a conflict in the area have to be treated with the greatest caution but it is highly probable that the Lomonds were not the backwater location that the current limited available information suggests.

Moncreiffe Hill

Introduction

Roman Scotland identified Moncreiffe Hill in 2008 after reconsidering Fraser’s proposal at the Gask Ridge. Fraser borrowed the philological rendition of Moncreiffe – Monid Croib – and transferred it to the Gask Ridge, several miles away.

Moncreiffe Hill, the true location of Monid Croib therefore deserved further investigation.

Moncreiffe Hill – identified as Monid Croib – is topped by Moredun Top hillfort (+190m OD) and occupies the bulk of the peninsula of land (lowest +10m OD) that is formed by the meandering nature of the Rivers Earn and Tay prior to their confluence at Perth.

Perth is centrally located in the territory of the Venicones, and in a later era was the capital of the Southern Picts. The name Monid Croib is known to us from records of an early 8th C AD battle fought there between rival contestants for Pictish dynastic control.

The area around Perth has ancient credentials, and as a centre of habitation it would have acted like a magnet to Roman forces on campaign seeking to force the tribes to commit to a decisive engagement.

Previously unconsidered, Roman Scotland visited the site, aware of Fraser’s argument in favour of a cross river assault further along the River Earn at Cairnie.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes, across the River Earn from the Dunning – Carpow chain of marching camps.

Is the Flavian activity in the area likely to be Agricolan?

Yes.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Carey and Carpow.

Is the marching camp located close to the site?

Yes, immediately on the opposite bank of River Earn.

Does the marching camps position and orientation “address” the site?

Both are orientated to the river which would have to be crossed to reach Moncreiffe Hill, probably near Easter Rhynd.

Does the marching camp display Flavian characteristics?

Yes, both.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Yes, both.
Carey at 113.8 acres can accommodate 25,300 men.
Carpow at 109.5 acres can accommodate 24,300 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No. A Roman deployment between Mains of Kinmonth and Elcho would give a frontage of 3km but this focuses on a Caledonian concave position centred near Rhynd which would be incapable of enveloping the Roman flanks below.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Crossing the River Earn would be noteworthy.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes at the Rhynd frontage, not around though.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No. The peninsula on which Moncreiffe hill is situated is set between the Rivers Earn and Tay which narrows at its landbridge at Kirkton Hill. It would have been a choke point from which refugees from the battle would have had great difficulty escaping mounted pursuit.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

A noteworthy march is not required, the site is located on the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

Yes, Horrea in Fife.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is sited aggressively – but not too far – in advance of this contending battle site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Fraser borrowed the philological rendition of Moncreiffe Hill – Monid Croib – when proposing the Gask Ridge. Croib – Gaelic for tree- has some linguistic similarity to Croup and in Gaelic form may possibly echo an older Brythonic place name.

Summary

Moncreiffe Hill, a site not previously considered, rather surprisingly achieves 14 appropriate answers out of 19, rating this newcomer as “interesting”.

Clearly the site suffers from the same problems that leave some doubt over the Gask Ridge contenders – namely the requirement for a major river crossing to come to grips with the tribes – of which Tacitus makes no mention.

In this respect however it is perhaps better suited as more land is available for the troops fording to deploy on the opposite bank, something that is missing below the Cairnie Braes.

Concern must also be voiced over the “island” nature of Moncreiffe Hill in respect of the events following the battle as well as the convex hills plan at Rhynd physically hampering any Caledonian envelopment of the Roman flanks (as with any convex hill).

However the site merits from close association with no fewer than 2 appropriately sized and dated marching camps sitting across the river from it and from whose positions Moncreiffe Hill is clearly visible.
All if nothing else creating interest by adding another interpretation behind the movements of the large Flavian marching camps in that series.

Mondboddo

Introduction

Roy in 1790 proposed Mondboddo near Glenbervie in northern Strathmore, deep within Vacomagi territory as a possible contender. In so doing he was heavily influenced by his brother officers experiences as they advanced north in the latter stages of the “45”.

Mondboddo was considered by military intelligence at the time as a possible location where the retreating Jacobite forces may halt and seek to confront the Government forces that were pursuing them.

Maxwell in 1990 considered this location further, and felt that the sites location – prior to the choke point at Stonehaven – was strong. Roman Scotland visited the site in 2008.

The suggested Caledonian position is on the low rolling ridgeline spanning between Knock Hill (+218m OD) to Herscha Hill (+223m OD). The corresponding Roman position near Auchenblae extends from Bankhead to Cushnie and is centred around Mondboddo.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

Possibly at the extent of Agricolan operations in 82 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, at Kair House.

Is the marching camp located close to the site?

No, 4.5 km away.

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

Possibly underlying the known camp which displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No, the possible Flavian camp at a speculated 92 acres it is not large enough and could only accommodate 20,400 men

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, it is rather too long. A frontage extends over 4km between Knock Hill and the low lying flanks of Herscha Hill.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

The Luther Water and broken ground in the Glen of Drumtochty would have been something of an impediment to ready Roman deployment.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, ideal.
A realistic Roman assault up? As above. However the ridge has exceptionally weak flanks, particularly Herscha Hill and there is little to force the Romans to commit to a simple frontal assault.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, after Droop Hill it is fairly low lying territory. The Caledonians would also have their backs to the Bervie water which is situated immediately behind.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, at Inverbervie.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Mondboddo sits at about the very northern limit that a legionary post subsequently sited at Inchtuthil would appear acceptable.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Mondboddo achieves 8 appropriate answers out of a possible 19, and as such it acquires a rating of “highly unlikely”.

The orientation of the ridge is not one particularly well suited to block a suggested Roman advance -as Roy and Maxwell considered – along the line taken by the modern A90.

However in fairness to the theory the Romans on this occasion could have simply marched closer to the hills along the line of the modern Fettercairn road, in which situation Mondboddo does prove to be more of an obstacle.

The hill however does not inspire a great deal of confidence as a strong defensive position; and the Caledonians we remember adopted a defensive posture.

It certainly would suit the Jacobite scenario where musketry could be played out before a Highland charge down the fairly easy slope could be delivered.

Mondboddo, we consider is a thing of its time, a hang over from the Jacobite episode and yet another seemingly cross era analogy which sadly, under the light of serious scrutiny, does not deliver particularly convincingly.

Monifieth

Introduction

Roman Scotland identified the site in 2008 while following philological investigation that led to quite promising results at Moncreiffe Hill and Knock of Crieff.

The further rendition of “fieth” from “Crieff” from the original Croup may seem one change in consonant too far however phonetics may have developed further in the intervening centuries so it was considered best to visit the area and determine if anything in the lie of the land substantiated the possibility.

Rather interestingly, modern Monifieth (varies around +30m OD) on the mouth of the Tay estuary is dominated to its north by Laws Hill near Drumsturdy (+120m OD) which in itself is crowned by a small but powerfully sited hillfort and broch.

The suitability of the broad plain below is eminently apparent when viewed from the hills upper slopes and a battle on the hills southern flank was considered.

While checking references on the map and discussing these with the local farmer it became apparent that features to the west of the Laws Hill site – the “Roman Hill” and “Roman Woods” – have, locally at least, associations of interest in the search (RCAHMS have no data concerning these). Therefore an alternative orientation from Laws Hill facing east to these “Roman” sites was also considered.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

No, there is currently no evidence of Flavian campaigning along the Carse of Gowrie.

Is the Flavian activity in the area likely to be Agricolan?

N/A

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes,
Tradition of “Roman woods” and “Roman Hill” adjacent to the site,
Gagie marching camp,
Muir of Lour / Kirkbuddo marching camp,
Invergowrie marching camp.

Is the marching camp located close to the site?

No,
Roman woods and Roman Hill (speculative only) is located adjacent to the site.
Gagie; 4.5 km away
Muir of Lour / Kirkbuddo; 8 km away
Invergowrie; 15 km away.

Does the marching camps position and orientation “address” the site?

Roman Hill; Possibly
Gagie; Possibly
Muir of Lour / Kirkbuddo; yes, on its short axis
Invergowrie; no

Does the marching camp display Flavian characteristics?

Gagie; unlikely
Muir of Lour / Kirkbuddo; no, late Roman marching camp morphology.
Invergowrie; no, late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No,
Gagie; at between 3.7 to 10 acres (estimated) size, it can accommodate between only 800 and 2,200 men.
Muir of Lour / Kirkbuddo; at 60.1 acres it can only accommodate 13,300 men.
Invergowrie; at 29.7 acres it can accommodate only 6,600 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No. At its widest a frontage of almost 2km is available on the south face of Laws of Monifieth.
To the west facing Roman Hill the frontage is further reduced to around 1.5km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, low lying farmland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No, immediate access is available on the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

Possibly. The name of the site is superficially similar to Moncreiffe – see Moncreiffe Hill and Gask Ridge contenders. It is however not a strong link.

Summary

Monifieth as previously mentioned is probably a “consonant change too far” as it secures only 8 appropriate answers out of a possible 19, rating the site as “highly unlikely”.

It is an object lesson that leads should be followed up though in the quest to identify Mons Graupius, while also recognising how powerful our 19 questions (based on our four factors) are at sorting a sites credentials out.

In the past the “Roman” connection or association would have been enough – coupled with the hillfort and broch – for antiquarians to hail the site as the definitive site of the battle.

Mormond Hill

Introduction

The conspicuous bulk of Mormond Hill, rising out of low lying agricultural Buchan (circa +80m OD rising to around +230m OD) was for long an antiquarian favourite, appearing on Baldwin and Cradock’s map of 1834 (link).

In choosing this location antiquarians relied heavily on the least reliable aspect of Tacitus’s account in The Agricola; the pre battle speeches.

Here some antiquarians used the (incorrectly) truncated piece of Tacitus where his Calgacus character theatrically claims;

“……..beyond us; nothing is there but waves and rocks”

This clearly appealed to antiquarians as Mormond Hill, located very deep in Taexali territory, is close to the north-west coast and the Kinnaird Head promontory (Taezolorum Pr).

Roman Scotland visited the site in 2008; an appropriate battlefield can be suggested on the southern concave slopes of Mormond Hill centred on Hillfoot.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

No

Is the Flavian activity in the area likely to be Agricolan?

N/A

B: Marching Camps

Is there a marching camp in the vicinity of the site?

No

Is the marching camp located close to the site?

N/A

Does the marching camps position and orientation “address” the site?

N/A

Does the marching camp display Flavian characteristics?

N/A

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

N/A

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the south face of Mormond Hill is a good size at 3.5 km long.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, the watercourses on the plain below the hill are insignificant.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, however there are some steeper gradients on the upper slopes.
A realistic Roman assault up? Yes.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, the landscape behind is flat coastal lowlands.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, a two days march to Peterhead.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Mormond Hill, a site like Fortingall and Culloden chosen on geographical assumptions only – and in this case uncritical acceptance of Tacitus’s pre battle speeches – fares poorly, achieving only 6.5 appropriate answers out of a possible 19. This rates the site as “highly unlikely”.

In fairness however, the antiquarians who selected this site – unlike Fortingall and Culloden – at least picked one with a hill and features generally consistent with the description of the battle.

The archaeological findings to date have not been kind to the North West however and no Roman presence is as yet known in this part of Buchan, a factor which has weighed heavily against this site.

Like the Pass of Grange site, we could critically ask ourselves why a polyglot force of northern tribesmen would assemble so near metaphorical lands end, no doubt watching from a distance the lengthy and leisurely burning of the bulk of their homelands by the otherwise unopposed approaching Roman column?

Locating the site of the battle to the extremities of the landmass is far too simplistic a rendering of Tacitus’s “………….behind us, nothing is there but rocks and waves”.

This was a metaphor; to the Romans once above the Forth Clyde line they were in the territory which constituted the end of the known world; “where all created things come to an end”.

It could equally be argued, if we loosen pre conceptions of Roman expansion as a necessarily south to north orientated axis of advance, to one where any orientation facing the highland massif could reasonably be so described, and in light of which we can better understand contemporary Roman conceptions of the highlands as represented on Ptolemy’s map (link) .

Pass of Grange

Introduction

The Pass of Grange site overtook Kempstone Hill in the post war years as the commonly held favourite location for the battle following its identification by Burn in 1953. Burn was not alone; Henderson Stewart (1960), Ogilvie and Richmond (1967) and Wellesley (1969) all concurred with this site.

The Pass of Grange, near Keith in Moray is one of the most northerly candidate sites, and is of strategic importance as it is a choke point on the communications route next to the River Isla running north-west through Huntly aiming for the northern coastal lowlands and Moray.

Like Roy at Mondboddo, Burns theory was simple. A Roman army marching this way would be forced to navigate the constricted Pass of Grange between the Sillyearn Ridge and the River Isla.

A Caledonian host emplaced along the Sillyearn Ridge would Burns reasoned, block Roman passage north into the rich agricultural lowlands beyond before further havoc could be wrought by the invaders.

Clearly Burns was heavily influenced by the strategies employed by Wellington in a later era; first at Bussaco then at Waterloo.

Burns was among the first to question what size of battlefield the forces there would require. However, in the face of his findings he later extended the Caledonian employment impractically further in order to include the very conspicuous Knock Hill.

(He employed the space between the Sillyearn Ridge and Knock Hill as the area from which the Caledonians launched their chariot attack upon the Romans).

Knock Hill was probably included to add weight to the theory as at the time attempts (not altogether successful) were being made to philologically link Knock to the “bump” derived from Welsh Crwb.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the circa 30 acre string of marching camps arcing through Aberdeenshire towards Moray do not hold sufficient capacity and post date Mons Graupius.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes; Auchinhove, Burnfield and Muiryfold.

Is the marching camp located close to the site?

No; both Auchinhove (4 km) and Muiryfold (2km) are located behind the hill.
Burnfield is located -before the hill- some 4.5 km away.

Does the marching camps position and orientation “address” the site?

No. Auchinhove and Burnfield respectively address the Rivers Isla and Deveron near to which they are located. Muiryfold eccentrically addresses the rear of the Sillyearn Ridge near the Isla.

Does the marching camp display Flavian characteristics?

Auchinhove; yes,
Burnfield; probably,
Muiryfold; no it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No,
Auchinhove; at around 30 acres it can accommodate only some 6,000 men.
Burnfield; the maximum available space on site approaches 40 acres. The site is likely to belong to the same circa 30 acre series as Auchinhove and will probably accommodate a similar number of men.
Muiryfold; at 101.5 acres it can accommodate a force of only 22,500 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No. The Sillyearn Ridge itself extends well over 3.5 km. When including Knock Hill this further increases this frontage to over 6 km. This is excessively long.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Yes, the noteworthy gap between the Sillyearn Ridge and Knock Hill though attempts have been made to press it into service for the Caledonian chariots.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes, except it is currently cut by drainage channels suggesting the ground was probably boggy in antiquity.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, except for Knock Hill which has very steep gradients.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No. Fairly broken upland is evident only in Pass of Granges immediate vicinity, this rapidly gives way to coastal lowlands.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes. Contact could be made at Cullen, Banff or Portsoy.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No, Forres would constitute a further considerable advance.

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No, Knock is a common later Gaelic appellation to many hills in Scotland.

Summary

The Pass of Grange’s subsequent great acclaim among academics stemmed in the main from the many marching camps discovered through aerial reconnaissance at that time, which proved, if nothing else that Roman arms had penetrated further north than the earlier recognised furthest penetration at Glenmailen had suggested.

Not all these discoveries were necessarily happy findings though for Burns, but he persevered and wrestled them into his hypothesis in order to keep it alive.

No large Flavian camp is known this far north. Both the Flavian and late Roman camps here seem to indicate forces on the march peacefully settling down for the night after navigating the pass. The camp preceding this, at Burnfield is fairly far removed and sits behind the Deveron and is possibly the camp used by the same force that used Auchinhove, perhaps even on its return journey south.

Again, a site so far north, justified on the basis of a Caledonian strategic emplacement designed to block the Roman advance before they reached the territory beyond makes little sense.

Quite apart from the fact that such a northerly position will have left most of the northern lowland tribe’s extensive lands open to Roman ravaging – a situation hardly conducive to morale in a polyglot force – it also appears that the Pass of Grange sits squarely within Taexali tribal territory.

By this late stage of a Roman advance north what would the tribes believe they would be defending? The entire lands of the Venicones and Vacomagi, and much of the Taexali`s lands will have been over-run leaving most of the glens leading into the Caledonii highland zone exposed for Roman penetration at leisure.

This northern site therefore would have engendered nothing but dismay in the vast bulk of the assembled tribal host, and it is extremely improbable under such circumstances that it could have been held together as a unified body.

On the ground itself, the Sillyearn Ridge, notwithstanding Burns calculations on numbers involved, feels almost slightly too long to be convincing, a concern magnified when the Caledonian dispositions are extended to include Knock Hill at which point the suggested battlefield becomes vast.

The Pass of Grange site achieves only 7 appropriate answers out of a possible 19, rating this previously favoured site as “highly unlikely”.

Peterculter

Introduction

Roman Scotland identified Newmillhill at the edge of Peterculter in Aberdeenshire – and located opposite the large Roman marching camp of Normandykes – as a contender in 2008.

Crawford had dismissed the site while promoting Kempstone Hill in 1949 on account of the fact that the River Dee lay behind the marching camp.
The River Dee is indeed a formidable obstacle here, and while this does not automatically rule the site out, the exposed position it will have placed the Romans in prior to an action is something we would naturally have expected Tacitus to have mentioned.

Notwithstanding this, the relationship of Normandykes across a low valley to Newmillhill is salutary, the camp and hill squaring up to each other like metaphorical prize fighters in a most convincing manner.

Also, with warfare in this era being an up close and personal affair, there is no need as suggested elsewhere for the Roman camp and the Caledonians position on the hill to be far removed from each other.
Legionaries, left deployed in front of the marching camp here are ideally located to come to the auxiliaries aid on Newmillhill if required, as the activities taking place there are clearly observable from Normandykes.

Yet further interest attaches to the size of the marching camp, its capacity close to holding the size of force we have calculated Agricola fielded at Mons Graupius.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the circa 30 acre string of camps arcing through Aberdeenshire to Moray do not hold sufficient capacity and post date 82 and 83 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Normandykes.

Is the marching camp located close to the site?

Yes, adjacent.

Does the marching camps position and orientation “address” the site?

Yes, notably well.

Does the marching camp display Flavian characteristics?

No, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Close, at 106.5 acres it can accommodate 23,600 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No, a Caledonian position between Woodside and the Gormack Burn centred on Newmillhill has a frontage of only 2 km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, ideal.
A realistic Roman assault up? As above.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

No, the landscape behind is a fairly flat lowland one.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, either at Aberdeen or Stonehaven.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

At 11 appropriate answers out of a possible 19, the site achieves a rating of “interesting”.

Clearly the camp displays late Roman marching camp morphology, and while the nature of the site it sits upon could be used to explain the proportions adopted in-extremis, the recognisable group of camps carrying on north in similar proportion and size appears to suggest that this was not the terminal point of the northward thrust of the expedition that bivouacked at Peterculter (as Tacitus suggests it was for Agricola’s force).

Also, we could be forgiven for wondering why the Caledonians themselves did not occupy the ground on which Normandykes sits and defend the crossing of the River Dee?

Notwithstanding this, Peterculter is an interesting site that repays a visit.

If Mons Graupius did not take place here, then like Tillymorgan we ask ourselves; what did?

Stormontfield

Introduction

Playfair suggested Stormontfield near Coupar Angus in 1797, aware of reports of Roman stations in the area (the Inchtuthil fortress site) and convinced that the Cleaven Dyke played a critical part in the action.

The medieval deer dykes on the Garrydrums were also pressed into service as the Caledonian encampment.
Further, the cairnfields spreading north-west to Dunkeld behind the Garrydrums were supposed to be the graves of the fallen and as found it was suggested marked the passage of the rout after the battle.

Such an antiquarian approach, scooping up all the available relics of antiquity in a location and attributing it to the events surrounding the battle is not unique however the size of the resultant site Playfair suggested is vast, comfortably filling a 10 km grid!

Further persistent association with the site was continued by Macdonald early in the 20th C . While excavating the legionary fortress at Inchtuthil, MacDonald proudly pronounced that this was the site Agricola set out from and returned to after the battle, though he cannily refused to be drawn on the precise site of battle.

The somewhat draconian MacDonald’s influence was widespread in cliquish academia and the ongoing persistent belief in the battle having taken place north of Inchtuthil still prevalent today stem from this quite incorrect statement made originally by Macdonald.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the activity on the site is related to the construction of the legionary fortress – Victoria – which dates at the earliest to the year(s) following the battle.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Inchtuthil 1 and Inchtuthil 2.

Is the marching camp located close to the site?

Yes, by virtue of the expansive size of the proposed battlefield.

Does the marching camps position and orientation “address” the site?

No, they address the fortress.

Does the marching camp display Flavian characteristics?

Yes, dateable by function and association with the Flavian fortress.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

No;
Inchtuthil 1; at 50.2 acres it can accommodate only 11,100 men.
Inchtuthil 2; at 36.6 acres it can accommodate only 8,100 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

No. The expansive escarpment of the Garrydrums extends to over 11 km between the River Ericht and the Buckny Burn at Benachally.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

Yes. At a quite incredible 7 km the proposed battlefield is unrealistically deep and due to its size is unavoidably cluttered with features of the landscape inconsistent with Tacitus account. The site is crossed by both the Lunan and Lornty Burns while three small lochs (Clunie, Marlee and Rae lochs) effectively split the site.

Further “Craig Roman”, the aptly named escarpment behind Kinloch – as well as being the obvious position for the Caledonians to defend – blocks direct line of sight to the proposed battlefield on the lower slopes of the Garrydrums beyond.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

No, though it has been proposed that the chariot action took place nearer to Delvine in front of the Cleaven Dyke, an act which further increases the size of the suggested battlefield.

The Cleaven Dyke has been put forward as a suggested explanation for Tacitus pro vallum.

Legionaries deployed here would however be left unaware of unfolding events on the lower slopes of the Garrydrums and poorly placed to come quickly to the aid of the auxiliaries.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes though the rolling profile of the hill includes reverse slopes.
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, the Tay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, excellent association.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Stormontfield achieves 9.5 appropriate answers out of a possible 19. This rates it as “highly unlikely”.

The Flavian camps that Stormontfield merits by association with are however the construction camps for the fortress built the year following the battle at the earliest.

Possibly the best accolade this long straggling battlefield attracts is that it is the contending site closest to the fortress named Victoria in honour of the victory achieved at Mons Graupius.

However it is an overly simplistic approach to consider an immediate proximity to the site of battle as a pre-requisite when naming the fortress.

Strathfinella

Introduction

Sometime before 1793 Roy – influenced by the experiences of his brother officers in the “45” – suggested a location “near Fettercairn” in Angus.

Roman Scotland has long viewed the prominent mass of Strathfinella Hill (near Fettercairn) in the Howe of the Mearns with interest, and marrying the two prompts together visited the site to determine its applicability in 2008.

The proposed battleline, along the slopes of Strathfinella Hill, terminating on Black Hill is very compelling indeed.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

This is possibly the area of Agricola’s most northerly operations in 82 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Kair House and Marykirk.

Is the marching camp located close to the site?

No,
Kair House; 5 km away,
Marykirk; 9 km away.

Does the marching camps position and orientation “address” the site?

No, they address their local crossing points of the Bervie Water and North Esk.

Does the marching camp display Flavian characteristics?

Kair House; possibly underlying the known camp which displays late Roman marching camp morphology.
Marykirk; No, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Kair House; no at (a speculated) 92 acres it is not large enough and can accommodate only 20,400 men.
Marykirk; close but too large. At 126 acres it can accommodate 28,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, an impressive Caledonian position 3.5 km long is available on the concave theatre-like slopes of Strathfinella Hill, centred on East Cairnbeg and extending between the lower slopes of Black Hill and the heavily broken ground of the ridge above Odmoston.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, there are only minor watercourses on the immediate site. To get to the site the North Esk would have to be crossed – at a distance – from Marykirk while the Luther Water and Bervie water would require to be crossed from Kair House.

On the contrary however the broken ground above Odmoston admirably matches Tacitus account of difficult “rough” ground on the flank impeding the progress of the cavalry.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, the broken ground around Drumtochty leads to the Highland zone.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, a days march to Inverbervie.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes, Inchtuthil is not too far removed to the south from this site.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

Strathfinella achieves 11 appropriate answers out of a possible 19, rating this site as “interesting”.

Should earlier suggestions of a square enclosure on the site of the Kair House marching camp be proven then Strathfinellas rating will improve as will our understanding of Agricola’s most likely northern penetration – though this was probably in 82 AD.

Strathfinellas theatre-like profile and the clearly rough ground on the flanks fits well with Tacitus’s account.

It is of an appropriate size for the forces involved and bounds the modern Fettercairn road that Roy suggested was another route the Romans may have followed north.

The flat farmlands of the Howe of the Mearns have ample space where a suitable Flavian camp may be sought and one day found.

In conclusion the site is well worth a visit.

The Caterthuns

Introduction

The White and Brown Caterthuns form a conspicuous twin peaked hill, these being joined by a gentle saddle.

Each of the hills is crowned by a conspicuous hillfort, the older one on the Brown Caterthun is dateable to the mid 1st millennium BC while the earliest parts of that on the White Caterthun may be contemporary with the events discussed here.

The Caterthuns sit as the dominant outlier of the Angus Hills – which roll out from the Grampians and Cairngorms – and commands the passage north contained between the hill and the Montrose basin – of which it commands excellent views. Northwards thereafter the land widens out into the expansive Howe of the Mearns, this choke point therefore is indicative of the Caterthuns strong strategic position and explains its longevity.

Behind, the hill has excellent communications with the minor glens coming off Glen Esk and from which the Caledonian septs could be expected to arrive. Research on the layout of the Brown Caterthun makes it evident that the site had for generations been a focal point, of probably military as well as religious and civic matters.

The Caterthuns, identified by Roman Scotland as a contender was visited in 2008.

The plain below – or level platform around +130m OD rises to the twin peaks in a challenging but realistic gradient achieving the summits around circa +290 m OD.

A speculative Caledonian battleline on the lower slopes stretching between Balzeordie and Lundie Castle centred on Kildare stretches just over 3 km and allows a very convincing battlefield.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

This is the probable (though as yet unproven) area of Agricola’s operations in 82 AD.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Stracathro and Keithock.

Is the marching camp located close to the site?

No, around 5 km away.

Does the marching camps position and orientation “address” the site?

No

Does the marching camp display Flavian characteristics?

Stracathro; yes.
Keithock; no, it displays late Roman marching camp morphology.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Stracathro; no, at 39.3 acres it can accommodate only 8,700 men.
Keithock; no, at 63.3 acres it can accommodate only 14,000 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, the lower slopes of the Caterthuns extend for slightly over 3 km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, minor watercourses only. The declivity at Donaldson’s Den, a natural feature may perhaps have a connection with Tacitus pro vallum.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

The lower slopes around Forthill and Kilgarie are appropriate.

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, ideal.
A realistic Roman assault up? Yes.

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, broken upland setting leading to the highland zone.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

Yes, excellent location and Flavian connections at Montrose Basin.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes, Montrose basin is probably Trucculensis Portus – “Wilderness Haven”.

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

Yes

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

The Caterthuns achieve 12 out of a possible 19 answers, which rates the site as “interesting”.

Although no appropriately sized and dateable marching camp is located near the Caterthuns, there is plenty of space where one may perhaps have been located.

Further investigation therefore over the early 12th C AD medieval Stracathro battlefield in and around Blair Muir between Cairndrum and Lummington may perhaps pay dividends.

The Caterthuns dominant position in Strathmore, its proximity to the Montrose Basin and its proven tribal nodal importance before and after the events noted make it an intriguing site which well repays a visit.

Tillymorgan

Introduction

The Hill of Tillymorgan is located north of the Garrioch between Inverurie and Huntly and is on the route several Roman forces used when passing this far north.

While writing in support of the battle taking place on another site; Kempstone Hill, Crawford in 1949 followed Haverfield`s conclusion that no battle could be fought at the large marching camp at Glenmailen as this would involve the Roman forces “facing south”.

Crawford somewhat disingenuously concluded this reasoning by stating that this was due to the lack of any hills lying to the south of the camp at Glenmailen!

The conspicuous hill of Tillymorgan, formerly known as Slate Hill, lying south of Glenmailen is included on McRonald’s 1789 dated plan of Glenmailen and ironically this is contained in Crawford’s 1949 publication “Topography of Roman Scotland”!

Further Crawford actually noted how he had viewed the remains of ditches around the hill of Tillymorgan when visiting Glenmailen!

This is THE object example of how even well respected authorities could ignore or dismiss out of hand – on entirely spurious grounds – sites other than those they were single-mindedly determined to advocate.

As for Haverfield`s concern, few proposed sites have simplistic north / south orientated battlelines, the recent favourite “Bennachie” being a fine example (Roman facing SSW to Caledonian facing NNE).

Tillymorgan has a 30 acre Flavian camp at Ythan Wells addressing it, this camp in turn is overlain on a different orientation by the late Roman marching camp at Glenmailen. Tillymorgan clearly therefore is a site of sufficient importance to the Taexali tribe in this period to have merited being visited by Roman forces on at least two such occasions.

Roman Scotland visited the site in 2007 and identified it as a viable contending site in the north east.

The Hill of Tillymorgan is located in an area of high ground, its lower slopes are around +220m OD. They rise gently in the main though more steeply near to the summit which is around +380m OD.

Its lower north facing slopes provide an ideal field of battle – excepting the limitations imposed by a convex hill plan – as well as sufficient space for the Romans and Caledonians to deploy in line with Tacitus`s account.

A: Campaigning

Is the site located in an area of known Flavian campaigning?

Yes

Is the Flavian activity in the area likely to be Agricolan?

No, the circa 30 acre string of camps arcing through Aberdeenshire to Moray do not hold sufficient capacity and post date Mons Graupius.

B: Marching Camps

Is there a marching camp in the vicinity of the site?

Yes, Ythan Wells and Glenmailen.

Is the marching camp located close to the site?

Yes, 1.5 km.

Does the marching camps position and orientation “address” the site?

Glenmailen; no,
Ythan Wells; yes.

Does the marching camp display Flavian characteristics?

Glenmailen; no, it displays late Roman marching camp morphology.
Ythan Wells; yes.

Is the size of the marching camp reasonably close to the 117 acre criteria required to accommodate 26,000 men?

Glenmailen; close, at 111 acres it can accommodate 24,600 men,
Ythan Wells; no, at 26.2 acres it can accommodate only 5,800 men.

C: Site topography

Is the site of an appropriate size to accommodate the forces engaged?

Yes, a Caledonian position between the Black Burn at Fisherford and the River Urie at Auchintender centred on Hill of Tillymorgan has a realistic frontage just under 3.5 km.

Does the critical area of the engagement include terrain or features which are specifically inconsistent with Tacitus account?

No, the site is only crossed by minor watercourses.

Does the site have a “plain” at the foot of the hill suitable for the recorded chariot action?

Yes

Is the hill of an appropriate gradient for:

The Caledonians to be marshalled and advance down and around? Yes, though there are steeper gradients to the upper slopes.
A realistic Roman assault up? Yes

Does the countryside behind the Caledonian position have terrain consistent with the actions recorded in the immediate aftermath of the battle?

Yes, a broken upland setting in the direction of Glens of Foundland.

Can reasonably ready access to the fleet (approx a day or twos march) be achieved from the sites location?

No, it would require at least three days march through Formartine to the coast at Cruden Bay.

Is the likely point of contact with the fleet consistent with the broad strategy and location of the Roman fleet on the east coast?

Yes

Can the “Boresti” be identified locally?

No

Are there known marching camps which hint at Agricola’s movements after leaving the site of battle?

No

Is the sites location consistent with the position chosen the following year for the siting of the legionary fortress of Inchtuthil?

No, siting Inchtuthil so far to the south of the scene of victory on the field does not sit convincingly.

D: The Sites Name

Is there evidence in the locality for names from which Mons “Graupius” could be derived?

No

Summary

While the known Flavian activity on the site post dates Agricola’s operations, the site achieves 12.5 appropriate answers out of a possible 19. This attracts a well deserved rating of “interesting”.

In fairness to Haverfield the reason a Roman force (on 2 occasions) circumnavigated the hill to address it from the north is worthy of discussion. As mentioned elsewhere, convex hills allow the Romans the luxury of choosing their preferred direction from which to address or assault such hills and Tillymorgan (like Hill of Bruxie) is classic proof of this.

Tillymorgan is an intriguing site sitting in a scenic upland setting. A single direction Caledonian retreat to the west by the Glens of Foundland would however be required to meet Tacitus account of the events following the battle.

Far from being dismissed, Tillymorgan amply deserves its place among the list of contenders.

The Contenders: a Conclusion

The clearest and most straightforward – if abstract way – to review the findings of this interrogation of contending sites is to look arithmetically at the results.

The “score” given is a representation of the number of questions (out of a total of nineteen) where a satisfactory answer was forthcoming.

This unavoidably may appear to be a broad-brush approach but it is an illuminating manner of illustrating the results where all sites have been measured evenly, fairly, fully and without preconception or partisan bias.

Note: Half points have occasionally been given where a clear cut answer one way or other could not be given – for instance at hills with steep gradients only on their upper slopes.

Contending sites in alphabetical order:

Site: Score: Probability (measured as a %):

Ardoch Moor 15 79%

Bennachie Traditional 5 26%

Bennachie Alternative 7 37%

Carey 15 79%

Carpow 15.5 82%

Craig Rossie 15.5 82%

Culloden 3.5 18%

Dalginross 10 53%

Dunning 19 100%

Fendoch 6 32%

Fortingall 2 11%

Gask Ridge Traditional 14 74%

Gask Ridge Alternative 16 84%

Glen Eagles 15 79%

Hill of Bruxie 9 47%

Hill of Edzell 12 63%

Huntlyhill 11 58%

Inverquharity 12 63%

Kempstone Hill 7 37%

Knock of Crieff 14 74%

Lomond Hills 9.5 50%

Moncreiffe Hill 14 74%

Mondboddo 8 42%

Monifieth 8 42%

Mormond Hill 6.5 34%

Pass of Grange 7 37%

Peterculter 11 58%

Stormontfield 9.5 50%

Strathfinella 11 58%

The Caterthuns 12 63%

Tillymorgan 12.5 66%

Contending Sites ranked by probability:

Site: Ranking: Probability (measured as a %):

Sites above 95% probability are prime contenders. 100% probability is considered a likely positive identification.

Dunning 1 100%

Sites between 75 and 95% probability are categorised as strong contenders.

Gask Ridge Alternative 2 84%

Carpow 3 eq 82%

Craig Rossie 3 eq 82%

Ardoch Moor 5 eq 79%

Carey 5 eq 79%

Glen Eagles 5 eq 79%

Sites between 50 and 75% probability are categorised as interesting.

Gask Ridge Traditional 8 eq 74%

Knock of Crieff 8 eq 74%

Moncreiffe Hill 8 eq 74%

Tillymorgan 11 66%

Hill of Edzell 12 eq 63%

Inverquharity 12 eq 63%

The Caterthuns 12 eq 63%

Huntlyhill 15 eq 58%

Peterculter 15 eq 58%

Strathfinella 15 eq 58%

Dalginross 18 53%

Sites between 25 and 50% probability are categorised as highly unlikely.

Lomond Hills 19 eq 50%

Stormontfield 19 eq 50%

Hill of Bruxie 21 47%

Mondboddo 22 eq 42%

Monifieth 22 eq 42%

Bennachie Alternative 24 eq 37%

Kempstone Hill 24 eq 37%

Pass of Grange 24 eq 37%

Mormond Hill 27 34%

Fendoch 28 32%

Bennachie Traditional 29 26%

Sites below 25% are categorised as not worth serious consideration.

Culloden 30 18%

Fortingall 31 11%

This analysis makes for some interesting, and in some cases extremely noteworthy reading. It is particularly evident that recent favourites such as Bennachie, even with an improved location, Pass of Grange and Kempstone Hill all fare particularly poorly.

Culloden’s rating makes it clear that citation of the Moray Firth and the extreme far north as the location of the battle is no longer a tenable approach.

Considerable attraction devolves to sites around the River Earn and in Perthshire, and here at Dunning, on a site located on the Clevage Hills the remarkable and exemplary probability of 100% is achieved across a thoroughly wide ranging and tough set of criteria, the likes of which these contenders have never been put against to date.

When we set out to explore the contenders and started writing this work we hoped to be able to conclude “on the balance of probability” that one site perhaps best merited the accolade of being identified the site of Mons Graupius.

We certainly now have, the Clevage Hills superbly fit the description of events by Tacitus, they are located in a proven area of Flavian campaigning and the camp at Dunning – the correct size for the likely Roman forces involved – sits in a coherent series of the largest known and irrefutably proven Flavian camps north of the Forth.

A Caledonian mustering there makes perfect sense with the Romans likely to pass near but not directly through this location in any march north through Strathallan, and cannily away from the direction Rome saw of concern following the preceding season; the area fronting the Trossachs.

Located to the south of Venicone territory – one of the major stakeholders in the Caledonian tribal confederacy – the position on the Clevage Hills is perfect for a formidable defensive encampment and its location enables the tribes to take whatever action was deemed necessary before the Roman columns started harrying their lands to the north of this position again.

It is singularly well located to either address any Roman expansionist moves north in the campaign season of 83 AD or to take the fight further south perhaps with the aim of encouraging the tribes in southern Scotland to throw off the Imperial yoke.

The rare survival of the original sites name Croup or Croupii to this day at Dunning and Carey would suggest that the Ochil`s Northern Hills, seen even to this day as a distinct component of the Ochil Hills was the region known as the Croup or Croupi. The later medieval battle of Dorsum Crup proves this, the battle at Dunning on the “Ridge of Cr(o)up” while the association with the hill ranges name survives at nearby Carey with the Croupie Craigs.

Mons Craupius – Hill of the Croup – has at last been given a robust and tested identification.

It is now time to turn our attention to the events of 83 AD which – based on our findings – we shall now endeavour to recreate.
To set the full context of these years we will – fairly obviously – use Dunning as the location of the fateful site of the battle.

(Note; for those who prefer, or in the future propose, other contending sites then similar integration in the narrative below may help them assess their preferred site in relation to the complex situations in play and most likely encountered at the time).

Mons Graupius – A Reconstruction.

Exactly when Agricola crossed back into Scotland in 83 AD cannot be guessed with any certainty. His immediate attention, Tacitus tells us, had been drawn by the premature death of his year old son in early summer.

To distract himself from the grief he turned to operations against the northern tribes though at this stage the implications in Tacitus narrative are that this was restricted to naval operations. This is Tacitus’s “plundering”, a poor translation for what was undoubtedly a repeat of the harrying conducted by the fleet the previous year.

As we have already discussed Agricola will have had no opportunity to make good his losses of the previous season with troop drafts from units on the continent as, with the Emperors own campaigns proving problematic, new recruits would be funnelled there.

Tacitus tells us that he reinforced his army with Britons who had through long years of submission proven their loyalty, suggesting southern Britons.
The southern British tribes had been disarmed for almost two generations, sufficient time for them to loose their Celtic martial traditions and practices.

Rather then than imagining that Agricola raised the “Belgic” tribes of southern England, we should imagine that Agricola sequestered natives raised for service in the Roman army and scheduled for transfer to units on the continent.

These troops will have been raw, but induced into the Roman military system and would be clothed and equipped like auxiliary soldiers anywhere else within the empire.

As well as the services of these British troops – at this time it was considered unusual for troops raised in one province of the empire to serve in that area due to fears over previous rebellions under similar circumstances in the past – Agricola clearly had to hand the message he will have received form the Emperor Domitian.

As we have discussed, receipt of this message and the contents therein can only be inferred from other statements and acts Tacitus records Agricola making.

Clearly Agricola was censured for his citizen soldier legionary losses incurred at the end of the preceding season, was advised replacements would not be forthcoming and that Agricola should exercise greater caution and strategic cunning in the use of the forces already available to him.
Agricola perhaps construed the meaning as including auxiliary recruits not yet embarked for the continent.

Finally, while clearly not instructing a halt to campaigning Domitian must have advised Agricola to ensure that he could not be successfully prosecuted at a later date for letting a war “drag on” and thereby “sowing the seeds of future rebellion”, and by default causing embarrassment to Domitian’s still young regime.

The threat of Imperial displeasure and loss of patronage, plus the risk of being prosecuted in Rome following his tenure as governor would have been very real and worrying considerations to Agricola, a state of mind in early summer 83 AD that would not be assisted in the least by a family bereavement.

Matters therefore suggest that the campaigning season of 83 AD started slowly. By early summer only the navy were undertaking active operations in the north while Agricola’s land forces, mostly still in their winter quarter forts and fortresses were probably preparing for the forthcoming season while Agricola’s new recruits were hurriedly trained and brought up to standard.

Given that Tacitus admits that Mons Graupius took place late in summer, it was probably near mid summer before Agricola’s columns crossed the modern border before leaguering up in what plausibly appears to have been two large encampments in central southern Scotland.

The first of these at Inveresk in the Lothians (53.2 acres) marshalled together units that marched up the eastern side of England, while a similarly sized encampment at Castledykes in Lanarkshire (60 acres) would accommodate units, likely mostly legionary that came up the west side of England.
As it has been cogently speculated that the very large and unusual legionary fortress at Deva (Chester) was Agricola’s headquarters for operation in the north then it will have been here at Castledykes that Agricola himself would have been based with the western battlegroup.

This gives these two camps an aggregate capacity of 113.2 acres, an area capable of accommodating 25,100 men.

We have already mentioned how the Romans – intractable as ever- would have prioritised reconstruction at the forts attacked in the preceding with troops left in garrison in southern Scotland over the preceding winter probably deployed early in the season to the task.

It is reasonable to speculate therefore, that Agricola’s army waited in two large battlegroups to ease the strain of provisioning for an unpredictable period of time.

Meanwhile his navy attempted to stir up the Caledonians to action again, while the troops above the Forth Clyde line employed in reconstruction work would also have been tasked with active scouting, aimed to identify situations where the Caledonians, tempted into the lower reaches of their territory in numbers could be caught up with quickly and engaged.

And this appears to have been exactly what happened.

Once notified that a major tribal hosting was underway at The Croup (Mons Graupius – the Ochils northern hills) by scouts probably acting out from work at Ardoch or Strageath, Agricola would have immediately ordered an army concentration (expito exercitu) en route to this location.

Lying reasonably equidistant to both Inveresk and Castledykes (circa 35 and 26 miles as the crow flies respectively), and situated at the junction of natural corridors still in use today is the large probable Flavian marching camp at Dunipace, hard on Agricola’s Forth – Clyde frontier line created in 80 AD.

It is conceivable that the army- with a hard march behind them – could conceivably have concentrated (expedito exercitu) at Dunipace only 24 hours after Agricola received breathless word from a messenger that the Caledonian mustering had been found. This is however a matter we shall return to later.

This large camp, at 115.8 acres, is a perfect match for the 117 acre requirement we have calculated; capable of accommodating 25,700 men.

It is noteworthy that the locality between Dunipace and Camelon is peppered with Roman marching camps suggesting that the area was subsequently used on many occasions as a Roman jumping-off point in pushes north of the Forth – Clyde line.

If the troops out posted in Strathallan left only cavalry vedettes – tasked with maintaining visual contact with the Caledonians – joined Agricola’s army at Dunipace then the arithmetic would appear to suggest that there were somewhere between 6 and 900 hundred of them.

Dunning is around 32 miles (as the crow flies again) from Dunipace and it is inconceivable that Agricola would not have set out after only the briefest of rests, possibly still in the hours of darkness marching hard to reach Dunning before his foe eluded him again.

The Roman army regulation pace could in theory achieve the notional distance in some 8 hours.

If we allow a plausible 10 hours then it is possible that Agricola startled the Caledonians – who may only have had a few hours notice of unfolding events – fording the River Earn at Forteviot (thereby cutting off a line of Caledonian withdrawal north) and then advancing in battle order.

Arriving before Dun Knock it is possible he threw up his marching camp by early afternoon that day, potentially within a mere 48 hours of the Caledonian hosting first being detected.

The construction of a marching camp illustrates a combination of factors.

First, no immediate confrontation on the field took place.

Secondly Agricola would have been aware of the effect that around 60 miles marched in somewhat less than two days will have had on his troops battle readiness. Agricola at this stage will also have dispatched embassies to the tribes to seek their submission while meantime planning for their possible failure. Accordingly he will have had the local ground scouted and formulated his battleplan.

With no results from any embassy, the Caledonians took to the field early next day, probably not long after sunrise and certainly before the Romans had formed up.

Did the commanders make speeches to their troops before the battle?

We can never know. Julius Caesar followed tradition and famously addressed his troops before going into action, and as he recorded them in his widely read writings it later simply became formulaic for others to record Roman generals doing the same.

There was no open space in a marching camp to form up all his army and address them as Tacitus would have us believe. This was usually done in the battleline and as Tacitus tells us the Caledonians beat the Romans onto the field then Tacitus has evidently telescoped events for literary convenience.

Agricola, apparently taciturn in character, probably never made any such speech. The Roman deployment was split making this difficult for Agricola to do. Deploying out of Dunning marching camps east gate the auxiliary units – infantry and cavalry – will have marched in column of line a short distance eastwards till they reached a point centred on the position adopted by the Caledonians on the Clevage hills, probably at Garvock along the line now overlain by the modern Bridge of Earn road from Dunning.

The legions and cavalry reserve, exiting the camps northern gates, either remained in formation immediately in front of the camp, or lay hidden from Caledonian eyes on the lee slopes of Dun Knock, below the ramparts of the iron-age hillfort on its crest – a deployment which merits Tacitus’s phrase “pro vallum” – before the ramparts.

A unique view of the initial deployments would have been gained from the summit of Dun Knock, and it is here that Agricola probably positioned himself with the standards and his officer corps, excellently placed to direct Roman actions through the coming fray.

The Caledonian host was a loose amalgam of tribes and family groups. The deployment of the bulk of their army would reflect this, being loosely arrayed along the slopes of the Clevage Hills with “the bravest” – the tribal elite – forming a recognisable battleline at the foot of the hill. It is against this that the auxiliaries were deployed.

Indeed it is the very looseness of the Caledonian deployment that would have excited concerns in the Roman command that the tribes would – as they ultimately did – use the lie of the land on the flanks below Middle Third and Craigenroe Hill to descend and wrap around the flanks of Agricola’s auxiliaries.

This was the point at which Agricola abandoned traditional maniple order by thinning his ranks and moving his second line of maniples up, creating in effect one continuous battleline stretching between Dun Knock and Gateside.

To their front, the chariots of the Caledonians manoeuvred on the plain at the foot of the slope between Pitcairns and Easter Clevage, trying their best to unsettle the auxiliary ranks with feinted attacks and thrown missiles.

And at that, it appears an impasse was arrived at.

Agricola’s plan was straightforward though cunning and involved not inconsiderable risks. As we have earlier discussed the fear foremost in Agricola’s mind was the possibility of the Caledonians declining combat and disappearing into the high ground, an outcome Agricola had experienced the previous year. Certainly the Caledonians had exhibited particular canniness in refusing to face Agricola’s massed forces while targeting lesser, but still considerable numbers of Romans the previous year in the attack on the IX legions marching camp.

Further, as we have discussed there is sufficient evidence in Tacitus text that indirectly tells us Domitian had warned Agricola not to risk further legionary (citizen) losses while paradoxically advising him not to let the war “drag on or sow the seeds of future rebellion”.

Agricola’s deployment therefore was a direct response to these separate influences and constraints, while cleverly using the lie of the land he was presented with at Dunning.

The legions, stationed out of the immediate battleline were given a degree of security so that “the victory would be more glorious should no Roman blood be spilt”. Their deployment pro vallum was noteworthy enough for Agricola’s to remember and Tacitus to record all these years later.

In our opinion, pro vallum, in front of the ramparts of Dun Knock would keep this strategic reserve – critically on the flank – hidden from Caledonian eyes on the Clevage Hills beyond.

Not only would this enable them to come to the auxiliaries aid, it was, we believe primarily aimed to deliver a hammer blow on the flank of the anticipated Caledonian headlong assault on the apparently outnumbered Roman auxiliary battleline.

The auxiliary line was therefore kept deliberately weak, first to prevent an overwhelming show of Roman force dissuading the Caledonians from staying on the field, and secondly – “scorning our puny numbers” – to provoke the Caledonians to descend the hill in a precipitate all-out assault. Agricola’s reserve was placed to attack any such Caledonian frontal assault as it hit home in turn in its flank below Dun Knock.

Simply put, the expendable auxiliary were bait in a well thought out trap, but one still fraught with risk should the – arguably – weaker half of his army be overwhelmed.

The famous military maxim cautions that “no plan survives contact with the enemy”, suggesting that a successful general is one who can anticipate his opponents moves and have in place plans to deal with developing eventualities.

In this respect Agricola, unfairly criticized by many – did show flair and competence as a commander. This is because, with the armies deployed, and despite the nuisance caused by the cavorting of the Caledonian chariotry, the Caledonian army en-masse showed no inclination to accept the proffered bait to descend the hill pell-mell in a wild and disordered charge.

Clearly the Caledonian elders were no dafties and with an admirable control managed to keep their men, difficult to control at the best of times, in good order on the advantageous high ground of the Clevage Hills.

Forget Tacitean spin, this surely must have disturbed Agricola as the required response would be for him to advance his troops against the opposition sacrificing much of the advantage in his initial “baited trap” deployment. The greater the distance the auxiliaries advanced the greater they would be removed from the legions support while they would also suffer the handicap of fighting uphill, no task to relish.

Agricola however does not seem to have abandoned his plan. Even as he advanced his line close to the Caledonian line it is left a matter of debate whether the exchange of missiles was the reason his line did not immediately close with the opposition or whether again Agricola was merely placing his line temptingly close to the Caledonian mass, a charge of whose front line would probably excite and draw along with them the eager tribal masses crowded behind.

Again the Caledonian leadership appears to have exercised considerable command and control, holding their men in the positions already agreed.

Clearly with his hope of a disordered Caledonian charge appearing less likely, Agricola finally ordered six cohorts of his troops to advance to contact. We have already discussed that this frontage of troops were probably those who faced a formed and recognisable Caledonian battleline, those behind no doubt milling in looser large bodies.

The cavalry squadrons had already seen off the Caledonian chariots as the battle lines drew closer however it was with the first clash of infantry that the battle “proper” got underway in earnest.

Close order infantry fighting was the preferred mode of combat for Roman infantry, particularly legionaries. The men deployed here however were auxiliaries who while trained for close order fighting were more used to more open order conditions.

Unless Tacitus is regaling us with stock phraseology on the nature of such contests with Celts through the ages, all we have to go on is his narrative is that the Roman fighting stance aided the auxiliaries over the Caledonians who appear, on the reading of the text to have suffered early the effects of crowd crush.

This crushing was the likely result of the excited Caledonian rear ranks pushing forward onto the front ranks leaving them crushed against the Romans shields with scant space to wield their weapons and vulnerable with only small shields to protect themselves.

The fight seems to have gone ominously one way from the start, and while Tacitus is no doubt exaggerating the success of the auxiliary line – now joined by the rest of the auxiliary battleline – it did however seem at this stage to have taken the upper hand pushing forward up off the plain onto the lower slopes near Boghall.

The Caledonians however must have maintained cohesion, like the English centre at Flodden in 1513 any force that survives being pushed back over 200 yards is clearly not being over-run but is managing to maintain order in the face of adversity.

The Caledonian front lines were not therefore beaten comprehensively as described by Tacitus but were pushed back by the relentless close order advance of the auxiliaries.

The Roman advance however appears to have started losing impetus, no doubt as the effects of the slope took its toll on tired legs and as fresh Caledonians were fed into the battleline.

Obviously things needed shoring as at this stage the loose milling Caledonian “groups” behind appear to have broiled round the flanks of the auxiliary battleline, a move countered by Agricola moving the cavalry stationed on the wings to charge in to the Caledonians flanks like book-ends in an attempt to push them back in behind their own front line.

Tacitus is quite forthright in describing the resultant crush. These 3,000 cavalrymen were hampered by the steadfast defence of the Caledonians as well as “the roughness of the ground” – meaning by this stage they may have been on a more considerable gradient – added to the already congested struggling mass in the centre – “most unlike a cavalry action” Tacitus tells us.
It is inconceivable, with disciplined order breaking down that the Romans at this stage were not on the receiving end of notable casualties themselves as mens legs gave way through exhaustion and the crush of the press.

Seeing this key development – and it would appear Agricola, holding his nerve, had still not got his legions moving – that the Caledonian leadership, sensing the critical phase of the battle had arrived ordered forward their remaining tribesmen.
Charging forward those in the centre added their weight to the crush – which at this stage must have been truly horrendous – while many more swirled around the flanks to come to grips with their enemies.

Men would be hacking, heaving, pushing and struggling to keep both on their feet and to maintain even that precious inch or so to enable them to breathe. Maimed and wounded horses would be thrashing about with men on both sides crushed by their flailing hooves and it is probably around this time that the (infantry) cohort Prefect Aulus Atticus was killed, probably pulled off a hamstrung and gutted mount.

This then was the inglorious and horrendous reality of ancient warfare, up close and personal.

And that would appear to be what Agricola had waited for – a complete Caledonian commitment in the centre of the field. Now was the turning point of the battle as he deployed his hidden cavalry reserve to such devastating effect.

This force, probably some 2,000 strong would appear to have taken a circuitous ride round the west of Dun Knock then due south where they will have forced a passage across the Dunning Burn high up the hillside near Balquhandy.

Executed a smart wheel in line abreast they will then have charged downhill into the rear of the Caledonians, all of whom by now were busily engaged with the Roman auxiliaries to their front.

And this hammerblow was without doubt conclusive.

The final nail in the coffin of Caledonian resistance would be the sudden appearance of the legions who by now would have been ordered up by Agricola.
Even though not directly mentioned by Tacitus this would have had the same morale sapping impact suffered by the English at Bannockburn and the French at Waterloo (both of whom dissolved in panic) when faced with the imminent arrival of enemy reinforcements.

On the slopes of the Clevage Hills it most probably led to a general Caledonian collapse.

Tacitus glossed over the legions deployment and we believe he does not mention their eventual impact on Caledonian morale in order to obfuscate the legions somewhat inglorious role so far during the fighting. Smoke and mirrors.

Celtic society held a warriors brave death in battle in very high esteem indeed and Tacitus mentions many warriors choosing this over flight or captivity. For the masses though, farmers in attendance by tribal obligation, the main concern would be to escape the slaughter with kith and kin now that the “game” was clearly over.

And at that, in most cases, it will have been every man for himself, sometimes seeking safety in numbers, at other times seeking safety alone from the hacking blades of the pursuing cavalry depending upon circumstances.

Back over the ridge of the Clevage Hills, through their bivouac and down its reverse slopes, rough broken country with deep defiles they will have run pell mell. And here the leading elements of the pursuit came to grief, halted by groups of Caledonians clearly selling their lives dearly to assist their friends – and any family who attended the hosting – to escape.

How did they get so far before the enthusiastic outriders of the chase – Tacitus’s “first rash pursuers” caught up with them?

The evidence would clearly suggest some time passed before a full mounted pursuit got underway beyond the battlefield, clearly showing havoc reigned on the field for a considerable time, probably as the Caledonian front and centre, with little scope for easy withdrawal were gradually overcome.

Tacitus grandly claims 10,000 Caledonians fell. This will have included wounded and injured, of no value in the Roman slave markets and dispatched where they lay.
What of this figure of 10,000 fallen?

Without a shadow of a doubt Agricola was merely the first and certainly not the last recorded invader who, eventually successful in battle after frustrating campaigns in Scotland felt compelled to magnify the scale of their victory in order to assuage their otherwise bruised egos.

The pundits of later English Kings and generals would continue to claim similar fantastically rounded numbers of Scots and French casualties with a monotonous regularity which merely urges us to exercise caution in blithely accepting such formulaic and partial reporting.

That said and accepted there are a few statistics that that have been quoted in recent years on the casualties incurred in ancient and medieval battles which are worth considering.

Mons Graupius we must bear in mind was at the end of the day a signal defeat for the Caledonians. Precedence exists for us in the records of other decisive battles to not necessarily dismiss a loss of a third of the Caledonians at Mons Graupius as being altogether impossible.

Even the notably militarily successful English have suffered badly through the years in this way; a third of Boudicca’s behemoth of 240,000 were butchered by the Romans at Mancetter by a smaller Roman force than Agricola deployed at Dunning, a third of the Lancastrians never made it off the field at Towton, at least a third of the English were massacred at Stirling Bridge and while a greater number than this eventually forded the Bannockburn over a damn of their own drowned, few of the English (infantry) it appears subsequently made it home.

Three of these battles are all noteworthy in having a watercourse that significantly impeded the flight of the defeated – the tribes own wagons did the honours at Mancetter. The Water of May lies behind the Clevage Hills.

Further, losses in any defeat always appear to have been at their worst in a rout as cavalry’s prime role in the aftermath of battle was to pursue and harry the fugitives, principally to ensure a beaten army stayed beaten and did not regroup.

The Caledonians did not regroup – arguing for the comprehensive nature of the defeat sustained to a well led Roman army fielded with what appears to have been a great deal of thought given beforehand as to the manner in which the action would be fought. The chase, as discussed above did not start early enough – or well – for the Romans and this explains why (at least) two thirds of the Caledonians won their freedom from the fateful field of battle and were not trampled underfoot in the Water of May.

We do not believe therefore that the hills, burns and woods beyond were strewn with numerous dead fugitives, the majority of the dead would be those who lay piled like a high tide mark on the field of battle, cut down – or expired – in the press or surrounded in Agricola’s final masterstroke and unable to cut their way clear.

Butchered where they stood, or as Tacitus records willingly choosing a glorious warriors death in combat these men, the elite front ranks and the keenest stationed behind were these mass casualties. The bulk of Roman cavalry, no doubt finding it difficult to neatly extricate themselves from the mass in the centre of the field that they had been mired in probably hovered on the field to annihilate the remaining opposition while most of the Caledonians made good their escape, Tacitus’s text as good as says so much.

We simply have little way to quantify the magnitude of these Caledonian losses but can elicit a certain understanding.

Just as a “third” is a figure found often repeated for the losses in a major defeat, analysts have also found that in battle wounded casualties outnumber the outright slain, the quoted ratio is 2 out of 3 casualties are wounded, not immediately killed. In modern times this has led, with the benefit of excellent and immediate medical services to a salutary survival rate among the wounded.

While the Romans also had an excellent medical corps on hand, the Caledonians did not and of the numbers of Caledonian slain, it is fairly certain that at least two thirds will have been wounded and dispatched by a sharp thrust as the legions marched over and cleared the field of survivors. Those less badly wounded will have been unlikely to maintain the energy for a successful escape, and along with the greybeards called to arms will have been those most likely to have been cut down in the mounted pursuit that eventually followed.

In summary, Caledonian losses must remain a mystery, but if Tacitus exaggerates freely elsewhere in his work then his conveniently rounded total of 10,000 Caledonian fallen must be viewed with a healthy degree of skepticism.

What of Roman losses?

Clearly any report of heavy Roman losses would not be an acceptable end to Tacitus tale, and of losses he mentions only 360 “on our side”. Importantly this phrase does not differentiate between citizen and non citizen soldier losses.

Speculation has gone before that this was limited to citizen losses, auxiliary losses being left unmentioned and unquantified. This is unlikely as we feel the pressures on Agricola were such that citizen losses were the last thing he wished to freely bandy about.

However we have already discussed the techniques of smoke and mirrors Tacitus used in his description of the size of the Roman army, we have not one single reason therefore to believe that he would now change tact over Agricola’s losses on the field.

At the very least we can say that 360 casualties – probably outright fatalities who fell on the field – would be considerably outnumbered by Roman wounded, at least twice as many based on the statistics reviewed above.

Given that the auxiliary infantry was battered under a noteworthy barrage of missiles for some time, were then caught in an uphill struggle which degenerated into a compressed heaving mass in the field it remains improbable that they escaped as lightly as Tacitus would have us believe.

Sufficient also survives in Tacitus account recalling the fate of the Roman cavalry during the battle; riderless horses, cavalry caught in the mass (witness Aulus Atticus`s fate) and cavalry reported killed in the pursuit to indicate quite clearly that they did not come out of the affair with a clean bill of health either.

Simply put, like Caledonian losses, we cannot be precise about the butchers bill for the Romans.

What we can be sure of though is that notwithstanding the best efforts of Rome’s excellent medical corps many of the wounded will have expired later that day and in the days following from the traumatic cutting and crushing wounds sustained and that Tacitus will not have recorded these.

Ultimately Rome’s losses were undoubtedly higher than Tacitus alludes to but as we have nothing solid to work with we can speculate no further than we have done.

That night, after the battle, the Caledonians warriors will have continued to make good their escape, splashing across the Water of May before fanning out to find fordable tracts of the Earn.

Tacitus account of Roman scouts searching “in all directions” supports a battle on the Clevage Hills, from where the tribesmen would have fled in many directions, desperate to gain their freedom.

Unlike the suggestion by Tacitus the field of battle that night – a ghastly place – was probably deserted by the living except scavengers (of all sorts).

Enslavement or death would have been the lot of any tribespeople found in the area and an eerie silence no doubt lay about the dreadful place.

The Romans will have attended to their dead in a funeral pyre while their wounded will have suffered in their camp under the surgeon’s knife.
In all probability the Caledonian dead, stripped of weapons and valuables will have been left lying where they fell to moulder.

One author, pandering to modern sensibilities has suggested that Agricola may have shared Wellington’s sentiment on the night of Waterloo where, appalled at the slaughter he recorded that “Next to a battle lost, the saddest thing is a battle won”.

While this is a poignant sentiment, we have little to make us believe Agricola’s sleep would have been so troubled that fateful night!

Agricola had sought and secured the victory that he had striven long and hard for – and at no small risk of political recrimination to follow. He would have savoured his victory and thought little for the misery and suffering his actions had inflicted on the natives or indeed his own auxiliaries, deployed without the available support that was to hand.

And we need not necessarily think that Caledonian spirit, notwithstanding the misery and loss at personal and family level was particularly horrified. Celtic society prized those who achieved an honourable death in battle, and in time – the magnitude of the defeat meant this would be a generation in coming – they rallied to continue resisting the Romans in the north.

We must put aside modern views therefore as to how this slaughter was interpreted at the time. For the Romans it was a cause for unbridled celebration, just as it would be for the tribes whenever they in turn butchered the Romans. The modern “genteel” Highland sword dance owes its true lineage – we remind ourselves – to the ancient warrior’s exultant celebratory dance over the body and weapons of his vanquished foe.

The charnel house that was the field of battle will have fouled all the watercourses in the area and it will have been the need for drinking water that will have pressed Agricola to break camp at Dunning early the following morning.

Based on the evidence of marching camps the route he took led along the line of the River Earn east towards the Tay, their route shadowing the slopes of the “Croup” (the Ochils Northern Hills) – now silent and deserted Tacitus tells us – where after 11 miles the column, encumbered with their wounded – probably near a thousand – encamped at Carey.

From here, close to the Tay efforts were put in hand to communicate with ships patrolling the Tay. Once contact was made, the fleet – the majority operating out of the sheltered anchorage of Montrose Bay: “Trucculensis Portus” / “Wilderness Haven” were summoned to rendezvous.
Agricola’s army, making one last short move, set up camp at Carpow on the Tay where an embarkation of wounded and reserves for the fleet could be loaded safely given the lateness of the season and the unpredictability of the north sea.

Either here or at Carey Agricola took hostages of the local Horresti (or Boresti) tribe of Fife in guarantee for their continuing submission to Rome, suggestive they had not taken arms against Rome but had perhaps failed in treaty obligations to warn Agricola in good time of the Caledonian mustering taking place close to their territory.

The battle, as Agricola may have anticipated since the previous winter took place on the southern fringes of the northern tribes and Caledonii septs territory, yet he clearly felt the need to intimidate others, unreachable by land given the lateness of the season.

This would hint that the tribes of the far north – probably those north of the Mounth –had not been committed at Mons Graupius.
It was on these, the Taexali, Decantae, Smertae and others on whom the fleet now perpetrated (further?) acts of terror to ensure they got the message loud and clear and did not contemplate rising in opposition to Roman Imperium, now by default likely to land on their doorsteps.
Recently excavated signs of devastation with a likely Flavian horizon unearthed at Birnie near Elgin may pay silent testament to the fleets actions late that year.

The fleet we are told, as well as taking the “terror of Rome before them” were instructed to conclude the navigation of northern waters, proving (what had been a long reported fact) that Britain was indeed an island.

Much has been speculated over this voyage, it being suggested the text describes a complete circumnavigation of the British Isles after exploration of the far northern isles and a siting (only) of far off Thule (most probably Iceland).

More likely, the fleet operating in the east met up in the far northern waters with those flotillas which had been operating on the west coast before returning to over-winter at their anchorage in Trucculensis Portus (Montrose Basin) in anticipation of next years operations in the north.

Afterwards, Agricola retired south. Although Tacitus makes much on Agricola’s slow leisurely progress through the lands of the tribes he conquered during his term as governor – effectively lowland southern Scotland – we cannot help but imagine Agricola, with his prize of glory in battle secure and gained without enormous citizen soldier loss proceeding south with the pretended calm of a shoplifter insouciantly strolling with their ill-gotten gain to the shop exit eager to make good their escape.

Simply, Agricola, well aware his term of office was drawing to a close wanted nothing to jeopardise the apparent finality of his victory. He knew that the lengthy consolidation phase would be the responsibility of others – and nothing positive for Agricola personally would be gained by further action in the field where only further risk ran hand in glove with further confrontation.

In this phase (as well as others we shall return to in the addendums), and also in Tacitus decision to tell us about it we can glean how certain matters may have panned out during the stages that preceded the battle.

Agricola clearly had felt the need to intimidate the tribes of southern Scotland by a show of force and triumphalism after Mons Graupius. Tacitus tells us this plainly. What he does not do is explain why.

Murmurs of dissent over Roman rule in southern Scotland must have been the result once news spread of Roman set backs at the end of the 82 AD campaign season.

This would confirm why Agricola stationed his troops in central Scotland in summer 83 AD before advancing north of the Forth only once intelligence gave him a precise location to march to. Inveresk and Castledykes – camps perfectly sized to accommodate Agricola’s forces – tellingly hold strategic sway over the territory of the main southern tribes; the Votadini, Selgovae and Dumnonii and his massive army would smother any risk of wildfire rebellion breaking out.

These then were Agricola’s real conquests, overrun and studded with permanent garrisons.

Agricola’s standing may have been subsequently swollen by Tacitus’s grand claim that he reached the end of the known world – a phrase used by successive Romans for the lands beyond the frontiers of their time – but Agricola planted no installations in the lands of the northern tribes.

Simply put his experience there was limited to one purely of military campaign, effectively the twin prongs of risk and bloody battle.

Summary

Now we have found Mons Graupius – The Holy Grail of Scottish Antiquity- is there more to the matter than simply “ticking the box” against one of histories longer lasting and more engaging mysteries?

We believe so.

As far as we know the Caledonian tribes never again made the mistake of standing in conventional open battle against the iron clad formations of Rome.
War appeared to rage though. Later Roman biographers would continue to claim that other Roman generals and Emperors won “successes” against them.

However these hollow successes, probably more imagined or political than real failed to subdue the peoples of the north for any length of time, or for that matter to ever incorporate them within their Empire.

It is apparent that the tribes of Scotland’s greatest lesson from the Romans was the need to work together – a slow process – but one which was crucial to achieving sufficient degree of unity of purpose to resist the Romans, and then other (even more predatory) southern neighbours such as German-English once they had left.

Similarly for the Caledonians one thing was taken to heart after Mons Graupius; that in battle the Romans were an awesomely organised and equipped opponent. The Caledonians learnt, it would appear, not to underestimate them while apparently still not fearing them.
Rome never conquered these peoples, not because they were backward, poor or of an organisational culture ill suited to assimilation within the Roman state. The Romans met more extreme examples of all these classes elsewhere and successfully assimilated them into the empire.

Rather, the lit ember of resistance burned on, fanned occasionally but never extinguished. This, the recognisable embryonic seed of the pugnaciously independent nature of the peoples of Scotland was its first – and certainly not its last – manifestation in the historical record.

Cultures everywhere on contact with Rome had withered or were stamped out. The northern tribes however held on, against incredible odds following defeat at Mons Graupius through long years of Roman occupation in southern Britain and long after through to the modern day.

That spirit of “never lay down” is something we should celebrate, first apparent in the historical record after 83 AD its continued ethos allowed the country of Scotland to develop over the following centuries and to then survive the continual maelstrom of attempted English subjugation.

Ultimately however the longest lasting impact of Mons Graupius was that made by those on the losing side, who, it would appear, fired their compatriots and most importantly “those that came after (them)” as indirectly prophesised by Tacitus– not to fight on the tactical agenda of their adversaries but to carefully manage the manner in which they would thereafter (generally) successfully resist foreign invaders.

As the site of the battle has at long last been positively identified the Scottish Parliament should now erect a simple and dignified monument on the site of Mons Graupius at Dunning, in grateful memory of those souls who long ago made a stand on the slopes of the Clevage Hills and roared their defiance at the assembled might of Rome.

Addendum;

What happened in 82 and 84 AD and where?

82 AD

Now that we have identified the scene where campaigning in 83 AD and the location of the great battle itself took place, it is understandable that many will now wish to identify the location of the events of 82 AD, Agricola’s penultimate year of campaigning and the scene of the Caledonian assault on the marching camp of the Ninth Legions battlegroup, as well as understanding better the places and events of the year(s) following 83 AD.

We calculated that Agricola marshalled a force of around 26,000 men in 83 AD for service at Mons Graupius.
It is reasonable to suppose that Agricola – although faced with demands in 82 AD for troops to be dispatched to assist in Domitian’s Chatti war on the continent- will have marched north with broadly similar numbers.

Tacitus has Agricola taking the noteworthy step of recruiting British auxiliaries for his final years campaigning in 83 AD. This may have been a result of a combination of yet a further drain on units being transferred to the continent, but more probably to replace losses in action in 82 AD.

However the events of 82 AD may have persuaded Agricola that additional troops were required to complete the business; for all Tacitus gloss the year ended in a near disaster for the Romans and in 83 AD Agricola will have unlikely marched north to offer pitched battle with fewer troops.

Agricola probably entered Caledonia across the Teith at Doune and while bounding the mountains to his left will have struck north east into lowland Perthshire; Strathallan, Strathearn and Strathmore; “enveloping the tribes” as Tacitus records.

Meanwhile the main fleet harried the east coast, probably sheltering when necessary in the Montrose basin and Stonehaven further north while ships patrolled the Tay where communications could be achieved with the army while it operated in Strathearn and above the Tay.
Naval units, clearly active in the west out of Vindogara (Irvine) harried the western seaboard, penetrating deep into the landmass up the sea lochs there “opening up the secret places of their (Caledonian) sea”.

The eastern coastal lowlands was home to the Venicones and Vacomagi tribes. This was rich agricultural land that would both serve to assist in provisioning Agricola’s forces as well as underscoring his campaign of intimidation through the harrying and destruction that was undoubtedly meted out across it.

Meantime cavalry vedettes would have penetrated the nearer reaches of the glens entering the highland massif and we can well imagine the depredations that will have been visited on the inhabitants there.

Vacomagi territory lay to the north of the Venicones, their lands probably meeting at the South Esk and it was to the rich lands between the Esks that Agricola was intent on penetrating when cataclysmic word reached him that the cunning Caledonians had slipped past his extended south western flank and assaulted forts to his rear.

The bridgehead fort at Doune – on the fringes of Caledonii territory shows only one phase of occupancy and therefore appears to have weathered the storm.

By 82 AD Agricola had construction put in hand at Ardoch and Strageath, both are forts noted for having two structural phases of occupancy in the Flavian period. It was the unfinished defences of both of these forts therefore – as well as the encampments of the troops employed in building them – that came on the receiving end of Caledonian retribution.

These then are likely to have been the locations where the Caledonian’s action prompted Agricola to change tactic, split his force into 3 separate battlegroups and position them to better control the “routes the Caledonians were using”.

Three large Flavian marching camps are of great interest and their likely capacities is telling.

1.Dalginross (near Comrie), positioned to cover Glen Lednock as well as the approaches to the River Earns headwaters at Loch Earn is some 22.25 acres in area, and would accommodate 5,000 men.
2.Bochastle (near Callander) covers Strathyre and the route in and out of the Trossachs at Loch Venachar. Here the anticipated threat is countered with a larger force. The camp measured 47.75 acres and is capable of holding a hefty battlegroup of 10,500 men.
3.Malling (near Menteith) covers the gap between the Trossachs and Flanders Moss and the threat at this most southerly camp of the three appears to have been sufficient to warrant the 5,800 men bivouacked in the marching camp of 26.4 acres to have been reinforced with a second force of 2,600 men who tacked their smaller 11.6 acre marching camp onto the existing larger camp. This battlegroup now weighed in at a respectable 8,400 men.

The weighting clearly shows the south west was deemed vulnerable and hints that the watersheds from the Trossachs were the routes recently employed by the avenging Caledonians warbands when they stormed the forts in Strathallan.

The Roman force available to Agricola in these three camps therefore would appear to have totalled some 23,900 men. This equates very well with the view that Agricola will have reinforced his numbers for the campaign of 83 AD from that available to him in 82 AD.

The fairly nominal difference between forces effectively 26,000 and 24,000 strong highlights the difficulties Agricola faced raising additional manpower – while making good his losses- when Imperial prerogative was undoubtedly taking priority in drafts of legionary and auxiliary recruits on the continent. This explains his decision to employ British auxiliaries to bolster the thinned ranks of existing auxiliary units.

How did this force arrive at these locations and where were they before this?

Unfortunately, as yet there are no securely identified large Flavian marching camps known in Strathmore that can help identify exactly where Agricola’s large battlegroup ranged ever north-eastwards “enveloping the tribes beyond the Forth” nor how far they reached before developing events caused his abrupt redeployment in the opposite direction – south west.

Interest attends the large camp at Kair House near Laurencekirk. Early analysis of the outline of this camp – visible on aerial photographs- gave proportions of a sub rectangular camp of around 92 acres, an area capable of holding around 20,400 men.

Latterly however a developing understanding of this camps layout has given the unmistakably distinct rectangular proportions of a late Roman marching camp.

It is not outwith the bounds of possibility however that some of the earlier crop mark interpretations may have picked up on the ghostly remains of an earlier use of the site, underlying the later camp.

Later reuse of good encampment locations near large water courses is fairly common. Kair House’s location on high ground above the crossing point of the Bervie Water deep in Vacomagi territory would gel well with the route Agricola would in general terms be most likely to use when heading for the Mounth at Stonehaven.

Kair House is also sufficiently remote from the Trossachs for the Caledonians to strategically exploit the situation and in the process teach Agricola an object lesson on the consequences of over-reaching himself and exposing his extended flanks and supply lines.

Doubling back to address the problem(s) – Tacitean spin has him “advancing” to meet the threat – Agricola split his forces into three independent battlegroups – each a composite brigade of legionaries with auxiliary units in support – probably somewhere near Innerpeffray.

Like Kair House the large camp there is from a later period but the general location or its environs on the highly strategic River Earn would be the logical place for Agricola to split his force into three distinct battlegroups to better cover the routes “they (the Caledonian warbands) were using”, i.e. those highland glenmouths opening out onto Strathallan and Stirlingshire.

It was at this point -Tacitus tells us – that various nay-sayers in Agricola’s staff corps urged a continued headlong retreat to the Forth Clyde line, something Agricola (at this stage) was not yet prepared to be seen doing while also confirming Agricola had “advanced” nearer to this line, not further away from it.

Located near Innerpeffray is the Flavian camp at Dornock on the banks of the River Earn. At 23 acres it clearly locates the first position of the detached battlegroup of 5,000 men who would then proceed east along the river to the matching sized camp at Dalginross.

A second battlegroup, an almost matching sized contingent of 5,800 men also detached itself from the main body and leaguered up at Malling on the shores of the lake of Menteith in Stirlingshire, reaching there via as yet unknown marching camps.

Agricola’s own column, with staff corps nay-sayers no doubt buzzing around his ears, probably retreated as far south as the marching camp at Dunblane. He would likely have done this in order to assess the effect or indeed extent of Caledonian penetration into lowland Scotland – the “conquered lands” which took Agricola three long hard years to subdue.

The size of this third camp (32.75 acres, capable of holding 7,300 men) quantifies the number of men in Agricola’s force employed in his northern thrust at some 18,100 men.

This force would require a series of camps in the order of 81.5 acres in size up through Strathmore. Hopefully future investigation of sites like Kair House might bring to light evidence for Agricola’s most northerly penetration of Scotland in 82 AD.

The unaccounted remaining 5,800 men (from the subsequent total of 23,900 we shall discuss presently) was probably an under strength legion with a matching compliment of auxiliaries left guarding the Forth – Clyde isthmus – possibly loitering at Lochlands near Falkirk while Agricola and the main army ranged ever north-eastwards looking for a fight.

It would be a detachment of this force that was engaged building the forts at Ardoch and Strageath (and the roads to link them), and who were attacked by the Caledonians, incensed not least by the navies depredations on the western seaboard but by this physical intrusion on or worryingly near to their lands.

When this force of 5,800 men came up, the archaeological evidence would suggest Agricola split them into two, one group 2,600 strong (perhaps the under strength legion) being sent to reinforce Malling where there must have been threatening signs of Caledonian activity, while the remainder, some 3,200 strong were attached to Agricola’s central powerful column; now 10,500 strong which duly took up station in front of the highland massif at Bochastle (Callander).

Dalginross, Bochastle and Malling would all be locations of continuing concern to Rome, however at this stage there is nothing to suggest Agricola’s troops threw up anything as yet more formidable than marching camps while they actively patrolled looking for signs of Caledonian activity and awaited developments.

It can be seen that Agricola’s dispositions are heavily weighted to the south of this line, a fact the Caledonians seized on. The most northerly and remote force is that at Dalginross near Comrie. Here then, the battlegroup Tacitus records as being the “weakest” of the three, based around the IX Hispana legion was targeted and assaulted by the Caledonians.

There is compelling evidence to corroborate this.

It is the smallest camp of the three securely datable camps which “control the routes the Caledonians were using”, we can well imagine its force of around 5,000 would be a temptingly sized target for the Caledonians who we are told had not yet bound themselves into a larger federation by treaty; that would happen later that winter.

Dalginross sits at a strategic confluence of glens and waterways- the very reason the town of Comrie later evolved there. It is pressed in by the surrounding by hills; a veritable trap.

The “marshes” the retreating Caledonians subsequently sought refuge in can be readily explained by the effects of the confluences of not only the Ruchill Water but also of the River Lednock with the large River Earn, all in an age before land improvement, field drainage and river bank improvements.

General Roy, with a soldier’s eye suggested Dalginross as the scene of this fight and we agree with him.

A Roman fort was located here within a year or two (we do not however agree with Roy’s conclusion that this was the “Victoria” of Ptolemy’s map) and later reoccupation in the Antonine period illustrates the risks the Romans were prepared to continue to take to ensure this problematic hotspot remained closely supervised.

In targeting this weakest battlegroup, located remote from the bulk of Agricola’s remaining forces the Caledonians again displayed considerable strategic skill.

As before they pinpointed the exact position where their stronger opponent had over extended themselves.

On this occasion, either by ruse or luck they appear to have prompted their stronger opponent to strengthen one flank, then attacked and almost succeeded in overrunning the other in one daring night attack. Strictly speaking, they deserved to succeed!

Agricola’s actions on the day also lend weight to this theory.

Once alerted to the Caledonians movements he appears – understandably enough – to have attempted to concentrate his entire force.
Malling is near but not close to Bochastle and Agricola appears to have been forced to set out for Dalginross – an awkward journey around Ben Vorlich – before his column was yet joined by his troops from Malling.
Tacitus as we should expected attempts to put a positive spin on this, “sending ahead his cavalry and the fastest of his infantry”; in other words the nearest troops; the bulk of the Bochastle garrison while he appears to have waited for the hard marching Malling garrison to come up.

The real fighting skill of the northern tribes, not even yet mustered at full strength can now be gleaned.

They had tied down and were in the process of overpowering 5,000 Roman troops in the camp at Dalginross, when as many as 10,500 of the Bochastle garrison men showed up and engaged them.

Yet it was only when the remainder, Mallings 8,400 foot sore force-marched troops showed up and “let out a yell” while their standards “shone” in the rising morning sun that the tribesmen pulled off. The relieving column in other words let it be known that they were arriving. This is not the stuff of drama as Tacitus would have it, this sounds more like drastic tactics upon reaching a beleaguered force in the process of being over-run.

Even while disengaging the tribesmen managed to blockade some of the camps gates and Tacitus is forced to admit that stiff fighting took place there.

This action prevented the survivors of the IXth from following up the tribes withdrawal – indicative that the tribes retreated away from Agricola’s relieving troops and therefore northwards through Comrie and across the various watercourses there – and also prevented turning the withdrawal into the rout that Tacitus clearly wished it had been.

Clearly Tacitus cannot hide the facts, barely disguised by the spin, that the whole affair, for the IXths battlegroup at the very least was a damn near run thing.

While night operations, especially hastily conceived ones are notoriously difficult to execute well, it is clear however that only when – in all probability – all Agricola’s army of nearly 24,000 men were to hand that the Caledonians, probably numbering only a fraction of this number withdrew.

Having well and truly cocked a snook at the might of Rome it is little wonder that Tacitus subsequently recalls how the tribes:

“……..with unbroken spirit they persisted in arming their entire fighting force”

….while Tacitus’s account of the mood of overwhelming joy in the Roman camp at the end of 82 AD remains laughably unrealistic.

It is with this bitter pill swallowed that Agricola withdrew his forces south of the Forth Clyde line in the winter of 82 AD, probably already composing in his head the field report to the Emperor which would justify his actions and minimise the stature of his losses that year.

Given that his army had marched far, yet had not been able to dictate the course of events nor force the Caledonians into pitched battle on grounds and timing of the Romans choice, Agricola would have started to consider long and hard how he would pick up the challenge of successfully overcoming Caledonian resistance the following year.

84 AD

Late 83 AD saw the end of Agricola’s tenure as Governor of Britannia. Tacitus’s work tells of Agricola’s triumph and later years in Rome, but is hereafter silent on Scotland except for one cryptic remark in another of his histories.

Here he reflects that Britain (read Scotland) once conquered was “immediately forgotten” – the alternative translation is “thrown away”.

This has been correctly interpreted to record the later (not immediate) drift southwards in Roman control of Scotland over the years that separated Agricola’s governorship to the time when Tacitus actively wrote his histories during the early years of the Emperor Trajan’s reign some fifteen years later.

The bulk of this period belonged to Domitian’s reign, and Tacitus, the political weathervane, writing at the start of Trajan’s reign is eager to criticise Domitian’s tyrannical rule.

We do not therefore have Tacitus’s detailed – if partial – account of actions in Scotland to rely on for the years that followed.

Agricola’s successor as Governor was Salustius Lucullus, and our knowledge of this important Governor – for certain key activities took place during his years of tenure – is hampered by a silence in the historical record which can be explained by his “damnation” by a jealous Emperor Domitian.

This event, we are told elsewhere, came about – strangely enough – by Lucullus agreeing to have a particular type of javelin named after him, an act which would appear to have been sufficient to draw the ire of the Emperor. We know little of the matter and the episode smacks of typically quirky ancient tittle-tattle that probably merely serves to mask murkier reasons why Lucullus fell prey to Domitian’s anger.

Certainly his damnation, which effectively was an enforced erasure of his memory after execution would explain the silence that now reigns in the historical written record for the consolidation works he led in Scotland in the years following Mons Graupius.

What Sallustius Lucullus’s true crime was we shall probably never know with certainty. His memory was certainly never reclaimed for posterity and he was never venerated as a martyr following Domitian’s death, suggesting few – if any – felt much common cause with the man.

What is clear from archaeology is that after Mons Graupius, most likely the following year Lucullus put in hand building works which would encompass the lands of the tribes where Agricola had actively campaigned in 82 and 83 AD.

This would, based on current knowledge, extend no further north than Stracathro near Edzell in Angus and may mirror the extent of Agricola’s furthest penetration north in 82 AD.
It would however make strong strategic sense to continue the occupation up to the Mounth at Stonehaven and it appears likely from the vast haul of nails found in the workshop at Inchtuthil that this is exactly what was planned, if perhaps never taken to fruition.

It was 84 AD – at the earliest – when work got underway constructing the legionary fortress at Inchtuthil, quite far inland on the banks of the River Tay under Lucullus.
It is undoubtedly this post that received the honourary sobriquet of “Victoria” (shown hereabouts in Ptolemy’s later map) celebrating Rome’s victory at Mons Graupius the preceding year, and it was – interestingly – a name probably coined by Lucullus himself.

We can be forgiven therefore for taking time to wonder if Lucullus in turn wished to bask in the reflected glory of Agricola’s victory – Lucullus may have viewed it as Rome’s victory and may have considered that he was the current stakeholder of such kudos.

Perhaps therefore Lucullus’s penchant for naming places and weapons was more widespread than we now know, a habit which may have caught the eye of– or been pointed out to – a jealous and vindictive Emperor!

The fortress Victoria at Inchtuthil was not alone. It sat as the lynch pin in a network comprised of a wide range of installations, creating a frontier line with defence in depth which ran north-north-east from Doune up to Stracathro, always skirting the highland massif on its left.

This frontier line came some four years after Agricola’s range of posts were thrown up across the Forth – Clyde isthmus and was to be Roman Scotland’s second frontier, its most northerly whose extent – without exception – comfortably encapsulated all the territory where the events of Agricola’s campaigns in 82 and 83 AD were played out.

A chain of forts were built at the various glenmouths of the highlands fronting Strathallan and Stirlingshire, many on the same sites used by Agricola’s marching camps when he attempted in late summer 82 AD to control “the routes the Caledonians were using”.

Behind these forward posts a communications road running north passed beyond Agricola’s reconstructed forts at Ardoch and Strageath onto the Tay, and almost without doubt (if currently incompletely understood) further north connecting the frontier forts beyond the Tay.

Along this road, now better known by association with its mid 2nd C AD Antonine reincarnation as the Gask Ridge frontier, were located a series of signalling towers. There were almost certainly fewer of these, more widely spaced than their 2nd C AD Antonine period successors which in turn owed more to post Hadrianic concepts of fixed frontier “limes”.

These signal towers, notoriously difficult to date with any degree of accuracy, do however hint at one of several perceived needs following Agricola’s campaigns in the north.
While line of sight from these towers up to the advanced “glen blocking forts” – even those like Dalginross in proven “bandit country” – was limited, they did however provide speedy communication south from the main line of forts on the road, and undoubtedly (though still to be proven by archaeology) up to the legionary fortress “Victoria” at Inchtuthil at the very least.

These dispositions undertaken under the governorship of Lucullus clearly show the Romans learnt the lessons contained in army after-action reports dating to the two preceding years (forget Tacitean spin claiming undiminished success, that would not be written for another 15 years).

These certainly would appear to confirm our hypothesis that the forts – like Agricola’s marching camps were placed in the lower highland glenmouths to control movement and provide early warning of the movements of tribal warbands or armies.

The GHQ north of this planned network – Victoria – it would appear was planned to have a ready communication link into southern Scotland where, just as Agricola had done for much of the fateful summer of 83 AD, much of Rome’s manpower in northern Britain was posted.

We can reasonably speculate therefore that in the same way the glen blocking forts attempted to solve one problem for the Romans, the communication towers were also put in hand to counter another failing.

Had Agricola not received word from north of the Forth as quickly as he would have wished once the Caledonians were found mustering?

Did the original messengers sent from his troops reconstructing wrecked forts in Strathallan fail to get through, losing Agricola precious days in which to catch the Caledonian hosting?

We shall never know for sure but the construction of a signal network appears a salutary response to something which – just like the inability to control the glenmouths – the Romans took heed of and sought to rectify with simple expedient steps on the ground in 84 AD and the years that followed.

Lucullus therefore was the man under whose governorship Rome started pinning down the lands of the Venicones and Vacomagi and, it would appear, the highland Caledonii septs by controlling the routes they used to gain access into lowland (Roman controlled) Perthshire and Angus by constructing permanent forts and associated signal towers along the newly extended road from the south.

The agriculturally productive Venicones, Vacomagi and others would undoubtedly pay their newly imposed Roman taxes in grain. Another piece of recent archaeology in Scotland casts a spotlight on the nature of tribute imposed on the upland Caledonii. The pastoral upland tribes would however, as their descendents continued to do for centuries, have relied on cattle rearing.

Therefore it was to process a Caledonian tribute of cattle that the fort at Elginhaugh in the Lothians was converted into a livestock gathering post in the years immediately following Mons Graupius.

The troops detached from Britain’s garrison to help Domitian’s Chatti campaign had returned to the Britain by now, giving Lucullus a greater strength than Agricola had available in 82 and 83 AD.

The remains of marching camps hint at the deployment of these troops in the field in the years immediately after Mons Graupius.

Around 11,000 men appear to have been tasked with the construction of Victoria (in the first season of building work at least before being subsequently reduced to around 8,000 men).
Many others will have been involved in the other construction work for this new northern frontier. Itt would appear however that the need was felt to operate in the lands of the Taexali north of the Mounth, i.e. beyond the lands now being studded with permanent garrisons.

This clearly suggests that the Taexali and the tribes to their north west in Ross and Caithness had not been committed at Mons Graupius, and while this did not necessarily mitigate against the Romans setting garrisons in their land, it may have been deemed – Tacitus records Agricola’s verdict on the matter – as an “extension” of operations.

There must in 83 AD therefore have been little ambiguity that the tribes there, even if not committed to the Caledonian confederation in the battle were nonetheless hardly philo-Roman in their political outlook. That Agricola sent his fleet “ahead” after Mons Graupius to “carry the terror of Rome” before him proves the point.

Lucullus therefore probably concurred and deemed “immediate” annexation of lands north of the Mounth as an unnecessarily provocative “extension of operations”, at this stage at least.

The series of (circa) 30 acre sized marching camps – holding around 7,000 men – arcing up through Strathmore, beyond the Mounth to Auchinhove (at least) point however to Lucullus taking steps to secure the lands beyond his newly created northern frontier by treaty. In Maxwell’s term this column undertook a “show the flag” procession through these subdued if not entirely cowed lands.

It must be borne in mind however – and we cannot stress this strongly enough – that this was the only time Rome ever dared march so small a force in such extended operations in the far north – all later camps here are massive; rarely under 100 acres in size.

Proof indeed that this column marched north in the year(s) immediately following Caledonian defeat in battle.

In doing this they will have aimed to intimidate and coerce treaties from the tribal leaders at their customary tribal gathering places. As Mons Graupius clearly did not happen at Tillymorgan and places like it, then evidently this is exactly what did take place here, submission would have been demanded and received from what must have been a major tribal sept of the Taexalis at their traditional hosting place.

Archaeology may in time point to where this force – or forces over a few seasons -ultimately marched.
Fragmentary remains of small camps or forts – as yet unproven – dot the northern coast of Buchan along to Moray but firm identification and dating as yet remains inconclusive.

The most northerly accepted site, the marching camp at Bellie is greatly eroded by the powerful shifting River Spey and may possibly mark the terminus point of the 30 acre camp series.

However, it is just as probable that these fragmentary remains, mostly coastal, are the remains of later short lived attempts to dominate the tribes of the far north, most probably in the early 3rd C AD Severan period.

However we can not rule out the possibility that Lucullus, once construction of his limes were well underway may have made a brief attempt to create permanent posts in the far north, an attempt cut short by the sudden Roman redeployment southwards that took place sometime between 86 and 90 AD.

Notwithstanding the lack of longevity of Rome’s presence in the far north under him, we may ask one thing of Lucullus.
As he clearly was the governor under whom Roman land forces marched so far north, and if his men reached the Spey (given the small numbers we must consider it extremely unlikely that he accompanied them) then it is under his aegis that claims could reasonably be made on more geographical terms for reaching the “end of the known world” than Agricola.

Was this why Tacitus twelve years later tried so hard to attribute the honour of reaching “the end of the world” to his father in law (who in fairness had indeed won the actual battle against the northern tribes even if he did not physically set foot so far north) when he later wrote Agricola’s eulogy?

It remains entirely plausible that in Agricola’s eulogy – composed over a dozen years after Lucullus time – Tacitus took the opportunity to vent the partisan irk he probably felt as a young man when, fired with the zeal of his father in laws success, he heard of the notoriously boastful Lucullus’s claims concerning the geographical extent of his military and political endeavours in northern Scotland. Until that time this had been the sole preserve of Agricola.

Lucullus suffered damnation from a remorseless Emperor Domitian who – like most paranoid despots through history – was only too ready to listen to the malicious whispers of his immediate circle of sycophantic senatorial cronies in Rome.

Tacitus, in not going into detail on Lucullus’s fate (as he was to do on behalf of others) is entirely out of character. Indeed, Tacitus, a Senator himself during Domitian’s ill starred reign goes to no little length to justify by proxy (he used Agricola as his model) how a good man could survive the reign of a despot with his character still clean.

We suspect Tacitus is trying just that little bit too hard here, and, coupled with his almost deafening silence on both Lucullus’s actions in Scotland and ultimate fate it prompts us to quite reasonably speculate that underhand motives may lie behind all of this.

Insufficient evidence is available to lend strength to any legal pleading on the matter. However it seems not unreasonable to speculate that Tacitus – irked we must remind ourselves by Lucullus’s undoubted loud boasts over his actions in Scotland – himself may have been actively whispering poison in Domitian’s ear concerning Lucullus, an act he understandably will not have wished recorded for posterity given Lucullus’s ultimate fate.

After the event, Tacitus’s written works, as would be expected in such circumstances dissemble over the entire matter, leaving the way free for them to focus for posterity – without distraction – on Agricola alone.

Lucullus simply had no-one to make his case for him nor any remaining written record to defend his correct place in history.

Tacitus, by implying a magnified geographical kudos for his father in law, attempted to obfuscate the credit due to the later governor and until now he has allowed Agricola to successfully hog centre stage on the matter.

The greatest impact however, unknown by Tacitus at the time, would be that which his record – slanted to overshadow Lucullus – would have on history’s understanding of the location where the battle of Mons Graupius – Agricola’s crowning glory- took place.

Simply put; until now – where our “four factors” have allowed a wide ranging and rigorous analysis of all the current available evidence – simplistic and incautious readings of Tacitus has had everyone searching for Mons Graupius – Scottish antiquities Holy Grail – looking in the wrong places!

First published in 2008 on RomanScotland.org.uk

Euan Lindsay
Euan Lindsay
Euan is a former soldier, a retired architect, amateur historian and re-enactor with decades of experience.

Related articles

Latest

read more