[NB. The place-name identifications indicated below are all explained in Chapter 16: Roman place-names in Scotland.]


Chapter 17

The Roman invasion of Scotland – what the place-names tell us


1     It is generally believed that Agricola invaded Scotland in AD79 or 80. In this chapter an attempt is made to describe, with the help of the place-names, the sequence of events which occurred in the years after the invasion. It would, however, be somewhat cumbersome to keep writing that a certain event happened in a particular year or in the year after. In the interests of brevity, then, it is simply assumed that Agricola invaded Scotland in AD79.

2     AD79 (third season of Tacitus)

Tacitus tells us that in this year Agricola discovered new peoples, continued his devastations and reached the mouth of the Tay. He also writes that the Romans even had time to secure the land which they had taken into their possession by building forts and that they spent the winter in those forts. The text of Tacitus does not necessarily mean that Agricola reached the stretch of the river Tay at or north of Perth. It may simply mean that the Romans marched through southern Scotland, apparently with little or no resistance, crossed the Forth and then fought their way up to Strageath and then along Strathearn as far as the Abernethy/Carpow area. They then built forts in which they spent the winter, those forts presumably including Ugueste at Stirling, Bograndium at Ardoch, Tagea at Strageath, Marcotaxon on the Ruthven Water, Cermium on the Water of May and Levioxava at Abernethy.

3     AD80 (fourth season of Tacitus)

Tacitus indicates that in this year Agricola built forts across the Forth-Clyde isthmus. Now, on the face of it there would not be much point in building a string of forts across the Forth-Clyde isthmus if Agricola already held a string of forts along Strathearn, so presumably in the spring of AD80 he abandoned the forts built in Strathearn the previous year. The forts across the Forth-Clyde isthmus now defined the northern frontier of the British province. Presumably Ugueste at Stirling was kept as an outpost to control the crossing of the Forth, and Leviodanum at Doune was probably built at this time to control the crossing of the Teith and thus block the route which gave access via the fords of Frew to the central section of the isthmus. The fort at Barochan would have been built in this season to complement those across the isthmus, its function no doubt being to keep a watchful eye on fords across the Clyde.

4     AD81 (fifth season of Tacitus)

Tacitus tells us that in this season Agricola was active in the southwest of Scotland.

5     AD82 (sixth season of Tacitus)

5.1     Tacitus writes that in this year Agricola was back in the north. It was presumably in this year that he annexed much, if not all, of the territory of the Damnoni and built forts round the periphery of that territory. He built Cerma at Dalginross, Lano at Bochastle, Maulion at Malling, Demerosesa at Drumquhassle and Cindocellum at Dumbuck/Dumbarton. As part of this program he also recommissioned Bograndium at Ardoch and Tagea at Strageath, this explaining why excavation has shown that these two forts had two phases of Flavian occupation (D.J.Breeze 1982, 53). Presumably Bograndium at Ardoch, as the largest fort, was the military headquarters for the new fort network.

5.2     Agricola’s end of season report to the emperor would include a map showing the new frontier of the province extending from Dumbuck/Dumbarton up to Dalginross, over to Strageath and down to Stirling. It probably did not show the forts across the Forth-Clyde isthmus since those had probably been abandoned in the course of this year. They had become superfluous and in any event Agricola needed their troops to man the new forts built further north.

6     AD83 (seventh season of Tacitus)

6.1     It is probable that in this year Agricola moved back into Strathearn and recommissioned Marcotaxon, Cermium and Levioxava. His aim in so doing would be to encircle the proto-Maeatae and the people of east Fife and if possible prevent them having contact with tribes north of the Earn and Tay. (The term Maeatae is apparently first used about a century later than the period under consideration in this chapter. However, the writer needs a label for the people who lived in Clackmannanshire and west Fife in the Flavian period, so for want of a better term those people are referred to here as proto-Maeatae). Then, towards the end of this season the battle of Mons Graupius was fought at the foot of Craig Rossie. These points lead one to wonder whether the battle was fought not actually against all of the Highland tribes but only against the proto-Maeatae and the people of east Fife. The tribes had to bear in mind the possibility that the battle might turn against them and they might then wish to flee. The proto-Maeatae and the people of east Fife could flee into the woods and hills of Fife, i.e. go home, but if warriors from tribes north of the Earn and Tay had taken part in the battle, on a north-facing slope south of the river Earn, they would not have been able to flee home after the battle – the river Earn and the Roman army would be in the way. And if the proto-Maeatae and the people of east Fife were not in a position to field an army of 30,000 men, the number given by Tacitus for the size of the Caledonian army, then this might just be a case of Tacitus seeking to glorify his father-in-law, Agricola, by making his victory seem greater than it actually was. But it was obviously considered a significant victory for it to be included in official records and appear later in the works of Ptolemy and Ravenna.

6.2     Tacitus tells us that after winning the battle of Mons Graupius Agricola took hostages from a people called the Boresti and despatched the fleet to sail round Britain. That done he then marched slowly through the territory of "new people" and at some point sent his troops to their winter quarters. By "new people" Tacitus probably means people who were new members or subjects of the Roman empire because they had just been defeated by the Roman army. The text of Tacitus probably indicates that the Boresti, whoever they were, had not actually taken part in the battle of Mons Graupius and so were still outside the empire.

6.3     Agricola’s end of season report to the emperor appears to have included a map showing the forts Levioxava to Voran and including the name Victorie/Victoria indicating the site of the battle of Mons Graupius. It is to be noted that the fort at Fendoch, the Voran of this group, is quite different in design from the forts at Dalginross, Bochastle, Malling and Drumquhassle (G.S.Maxwell 1989, 99), this suggesting that it formed part of a later building program, namely that of AD83. The frontier of the province was now the line formed by the forts along Strathearn to Strageath, up to Fendoch and then down to the Clyde at Dumbuck/Dumbarton. But Agricola’s map would also show the forts at Ardoch, Stirling, Doune and Camelon.

7     AD84

7.1     It seems probable that a decision was taken this year to annexe the territory of the Boresti – no doubt somewhat submissive after hearing of the defeat of the proto-Maeatae and the people of east Fife – and in carrying out this project the Romans built Matovion at Cargill, Ugrulentum at Cardean and Ravatonium at Stracathro. They also built Iberran on the Bervie Water, though this lay not in the territory of the Boresti, but in that of another tribe, a tribe for which we have no name. We do know, however, that they used the hill-letter s and the river-letter b, the latter being seen in Iberban (the first b), the then name of the Bervie Water.  The large fort Tuessis (Bertha) would also have been built at this time, this fort no doubt being intended to take over Ardoch’s role as military headquarters. Presumably, too, this was the time the Gask Ridge towers were built, probably to provide fast, (nearly) all-weather communication between Bertha and Strageath. The existing towers south of Strageath would then provide communication with military headquarters in Ardoch (G.S.Maxwell 1989, 121). The towers south of Strageath would have been built in AD82 when the forts from Strageath to Stirling formed part of the ring of forts round the northern part of the territory of the Damnoni. The Gask Ridge towers would have been built in AD84 and presumably by a different military unit – this may account for the differences between the two sets of towers. One could argue that the towers to the south of Strageath served a surveillance function when built, as well as functioning as a communication system, since they would have been close to the territory of the proto-Maeatae, but the Gask Ridge towers cannot have served that function. At the time they were built, in AD84, the Romans had soundly defeated the proto-Maeatae and the people of east Fife and annexed their territory to the British province, they held Fendoch and were building new forts at Cargill and Bertha. The Gask Ridge was not, therefore, a frontier, so the towers there can only have functioned as a communication system.

7.2     Presumably the headquarters staff moved from Ardoch to Bertha when the latter fort was ready for occupation. It was presumably also in this year – AD84 – that an expeditionary force was sent north of the Mounth to reconnoitre the land north of the Grampians, as evidenced by Stracathro-type marching camps in that region. One might expect this to be the season, the season after the battle of Mons Graupius, when the Romans built Decha at Inverkeithing, Memanturum at Dunfermline, Litinomago Devoni in the Yetts o’ Muckhart area and Lodone to the south of Dunning, but an alternative and preferred scenario is presented below.

7.3     This year, the new governor’s end of season report to the emperor would include a map showing the new frontier of the province stretching from Dumbuck/Dumbarton through Drumquhassle, Malling, Bochastle, Dalginross, Fendoch, Cargill, Cardean and Stracathro to Iberran on the Bervie Water. The map would also show the forts from Tuessis at Bertha down to Decha at Inverkeithing and Poreoclassis at Camelon, but it presumably did not show Marcotaxon and Cermium and perhaps not Levioxava. These forts along Strathearn had probably been abandoned in the course of this year – they had become superfluous and the governor needed their troops to man the new frontier forts built further north.

8     AD85

8.1     This year, and on the basis of the report of the reconnaissance army sent north the previous year and the report of the naval commander who had been sent round the north of Scotland after the battle of Mons Graupius, the decision was taken to annexe the rest of Britain. It was in the context of this policy that a decision was taken to build a legionary fortress at Inchtuthil, the construction perhaps even being started later this same year. This might make sense of the presence of a fort and fortlet at Cargill, assuming the fort predates the fortlet. A fort would make sense at Cargill with no fortress at Inchtuthill, but after the decision was taken to build Inchtuthil there can have been no justification for retaining a large number of troops at Cargill, so the fortlet replaced the fort, its primary function presumably being the provision of staff to man nearby signalling towers.

8.2     This year the map included in the end of season report to the emperor showed Pinnata Castra/ Pinnatis at Inchtuthil and all the forts down to Inverkeithing and Camelon, including Strageath. It would also show all the forts from the Clyde at Dumbuck/Dumbarton up to the Bervie Water.

9     AD86 or 87

At this time, following the transfer of a legion and some auxiliary units to the Danube basin, and before the construction of Pinnatis at Inchtuthil was even finished, the idea of total conquest of the north was abandoned. Within a relatively short time the fortress at Inchtuthil and the fort at Stracathro were dismantled (G.S.Maxwell 1989, 111 - unworn Roman coins datable to c. AD86 were found in the demolition layers of Inchtuthil and Stracathro). Whether the Romans then established another linear frontier further south in Scotland is not clear, but if they did then it would seem most likely that they chose the Forth-Clyde isthmus again. It may be that a few of the forts on the old frontier, including perhaps Castlecary, Cadder and Balmuildy, were brought back into use and these, together with Mollins and Barochan, formed a new frontier. But whatever the scenario it was historically shortlived, for within 15 years or so the emperor Trajan had moved the frontier down to the Tyne-Solway line.

10     Alternative scenario

10.1     It is interesting and instructive to note the survival rate of the names of forts north of the Forth-Clyde isthmus. The names of the forts which were built in AD82 have a high survival rate. Thus Cerma appears to have survived in the town name Comrie, Maulion in Malling, perhaps the demer of Demerosesa in Drumquhassle, perhaps the dan of Leviodanum in Doune, perhaps the st of Ugueste in Stirling, the bogr of Bograndium in the name, Braco, of the village adjacent the Ardoch fort, and Tagea, reversed, in the geath of Strageath. But of the names of all the forts from Fendoch up Strathmore, of those along Strathearn to the east of Strageath and of Bertha and Inchtuthil, not one single name appears to have survived, unless the cathro of Stracathro is derived from the Celtic form of Ravatonium, namely Rascatonion, later shortened to Racatonion. And yet the don of Lodone may have survived in Dunning, the lint of Lintinomago may have survived in Lendrick Hill, Memanturum has survived in Dunfermline and Decha, reversed, has survived in the keith of Inverkeithing. This may indicate that these four forts were not built after the battle of Mons Graupius, as one might have expected, but that they, too, were built in AD82. It may be that Tacitus has confused his seasons, for he writes that at the beginning of the fifth season Agricola ‘crossed in the first ship’. But Agricola was campaigning in southwest Scotland in that season, so it is not clear why he should need or wish to sail anywhere. Now, if one supposes that that statement of Tacitus really belongs to the beginning of the sixth season, AD82, then it might indeed make sense to say that Agricola crossed in the first ship. This might mean that he crossed from, say, Cramond or Inveresk, to Inverkeithing, with the intention of fighting his way up to Strathearn and building forts along the way. It may well be that these were the very forts Tacitus says were attacked in the run-up to the battle of Mons Graupius. This might also explain why at one stage Agricola split his army into three units, with perhaps one unit patrolling the Inverkeithing to Dunning line, a second patrolling Strathearn, and the third patrolling between Stirling and Strageath, simply because Agricola did not know where the proto-Maeatae and the people of east Fife might strike next. And this may also explain why, when the proto-Maeatae and the people of east Fife finally decided to risk a set-piece battle against the Roman army, Agricola was able to ride rapidly and lightly to the scene of the battle. For no matter which army group he was with he wouldn’t have far to go to get to Craig Rossie, and he could get word to the rest of his army to assemble immediately at Dunning. In this alternative scenario it is not clear whether Marcotaxon on the Ruthven Water would have been re-commissioned in AD82 – so as to complete a ring of forts round the territory of the proto-Maeatae – or whether this task was left until AD83. In this latter year presumably Cermium on the Water of May and Levioxava at Abernethy were recommissioned so as to block contact by land between the people of east Fife and tribes living north of the Earn and Tay. Presumably the navy had the task of patrolling the Firth of Tay and the east coast of Fife so as to prevent contact by sea between the people of east Fife and tribes living north of the estuary.

10.2     It is said that Agricola was recalled to Rome very soon after the battle of Mons Graupius, though precisely when he left Britain is not clear. But if the above-suggested sequence of events is correct, then Agricola was responsible for building the forts across the Forth-Clyde isthmus, the forts from Dalginross down to Dumbuck/Dumbarton and also Barochan, Doune, Stirling, Camelon, Ardoch, Strageath, Fendoch, and also the forts in Strathearn east of Strageath, as well as the forts from Inverkeithing up to Dunning. But he did not build the forts from Cargill up to the Bervie Water. Nor was he responsible for building Bertha and the Gask Ridge signal stations. Nor did he build Inchtuthil. 

11     One final point on this theme. If one looks at the names Demerosesa, Cerma, Marcotaxon and Cermium, one will note that in each case m is the hill-letter most recently added to the compound. If it is then legitimate to deduce that this means that m was the hill-letter actually used by the people who lived in that region when the Romans arrived, then this means, surely, that the hill-letter m was used by Ptolemy’s Damnoni. This means, of course, that Victoria, between the Ruthven Water and the Water of May, was indeed in the territory of the Damnoni, as Ptolemy tells us it was. But equally interesting are Ugueste at Stirling and Memanturum at Dunfermline. It was shown in Chapter 16 that the correct spelling of Memanturum is probably Melantvrum. The ant is an inversion-type element and so the m and l must have been used in an inversion-type manner, too. In other words this area had been occupied by people who used the hill-letter n and later by those people who coined-inversion-type names in the hill-letter l, the name then becoming Lantvrum. Finally, the Dunfermline area was settled by those who used the hill-letter m and Lantvrum became Melantvrum. In the case of Ugueste at Stirling it is shown in Chapter 19 that the correct spelling of the name must have been either Muguleste or Mugulesde. It thus appears that m is the hill-letter most recently added to Melantvrum and Muguleste/Mugulesde, which rather suggests that the Damnoni also occupied Stirling and Dunfermline. This in turn suggests that the Romans did not seize all Damnonian territory in AD82, but that a rump of the tribe continued to live in the territory between the line joining Stirling and Strageath and that joining Inverkeithing and Dunning, and that these people managed to resist Roman pressure right up until the battle of Mons Graupius in the late summer or autumn of AD83. In this case it was not necessary to refer to proto-Maeatae above – these people were the Damnoni, or what was left of them.



[This page was last modified on 02 January 2021]