In the year 120 Hadrian visited Britain in person, when he appears to have put down any attempt at insurrection; and, having adopted, or rather originated, the policy of defending the frontiers of the Roman empire by great ramparts, he fixed the limits of the province in Britain at a line drawn from the Solway Firth on the west to the mouth of the river Tyne on the east, and constructed a great barrier designed to protect it equally against the incursions of the Barbarians or independent tribes to the north of it, and the revolt of those included within the province. It consists of ‘three parts—a stone wall strengthened by a ditch on its northern side; an earthen wall or vallum to the south of the stone wall; and stations, castles, watch-towers, and roads, for the accommodation of the soldiery who manned the wall, and for the transmission of military stores. These lie, for the most part, between the stone wall and earthen rampart.’ The stone wall extends from Wallsend on the Tyne to Bowness on the Solway, a distance of seventy-three and a half English miles. The earth wall falls short of this distance by about three miles at each end, not extending beyond Newcastle on the east, and terminating at Dykesfield on the west. The result of the most recent examination of the wall is that the whole is undoubtedly the work of Hadrian.[54]
Hadrian thus made no attempt to retain any part of the country conquered by Agricola in his last campaigns, but withdrew the frontier in one part even from where it had extended prior to Agricola’s government, in order to obtain a more advantageous line for his favourite mode of defence.
19. Orphei Argonaut. v. 1171, ἢν νήσοισιν Ἰέρνισιν ἆσσον ἵκωμαι.
20. Herodot. iii. 115.
21. Aristot. De Mundo, iii.
22. Polyb. iii. 87.
23.
24. Strabo, Geog. Lib. ii.
25. Cæs. De Bello Gall. v. 12-14.
26. Strab. Geog. Lib. iv.
27. Diod. Sic. v. 21.
28. Consularium primus Aulus Plautius præpositus ac subinde Ostorius Scapula uterque bello egregius: redactaque paulatim in formam provinciæ proxima pars Britanniæ.—Tacit. in Vit. Ag. 14.
29. The Antona has been supposed to be the Avon, and an emendation of the text to Aufona has been proposed. This has been pronounced to be a happy conjecture, but the author does not think so. Avon is derived from no word that could possibly assume the form of Aufona; and it is difficult to understand what a line of forts from the Avon to the Severn was to accomplish. The Nen, which has also been suggested, confines the province too much. It was more probably the Don, which falls into the Humber. The Don and the Severn were connected by the Fosseway and the forts along its line. That the province had reached the frontier of the Brigantes in the reign of Claudius, may be inferred from the lines of Lucius Annæus Seneca:—
31. Pomponius Mela (A.D. 45) mentions them—‘Triginta sunt Orcades, angustis inter se ductæ spatiis: septem Hæmodæ, contra Germaniam vectae’ (De s. orb. iii. 6). Eutropius has ‘quasdam insulas etiam ultra Britanniam in Oceano positas Romano imperio addidit (Claudius) quae appellantur Orcades’ (Hist. Rom. lib. iv. c. 13). It is difficult to reconcile this statement with that of Tacitus, that Agricola first made the Orcades known. That any conquest took place in either case is unlikely, and they were probably annexed to the Roman Empire in the sense in which an island in the Pacific, when first observed, is declared to belong to Britain, and named Victoria. The existence and position of the Orkneys may have become known under Claudius, and first actually seen under Agricola.
32. The anonymous geographer of Ravenna gives a list of the towns of Britain when the Romans left the island. Though plainly not stated in any regular order, they are still manifestly grouped according to situation, and those north and south of the walls can be clearly distinguished. Among those north of the wall, between the Solway and the Tyne, is the town called by him Venusio, and the identity of the name shows its connection with Venusius.
33. Tacit. Annal. lib. xii. c. 40. The expression ‘regnum ejus invadunt’ shows that Cartismandua’s kingdom was now distinguished from that in the interest of Venusius. ‘Acre prælium fecere cujus initio ambiguo finis lætior fuit.’
34. Tacit. Hist. lib. iii. c. 45. Tacitus, in his Life of Agricola, implies that Vettius did nothing, and was not equal to his position; but in his sketch of the previous governors it is manifest that he endeavours to enhance the fame of his hero by lessening the merits of his predecessors. The account of the war is taken from his History, where, although he does not name Vettius, it is plain that the events there narrated happened during his government, and this accords with the lines of Statius (see Note 38), which, making due allowance for a panegyrist, certainly imply a war, the result of which had reflected credit upon him. The allusion to the Rex Britannus, from whom he took the ‘thorax,’ is curious. Venusius is probably meant.
35. This narrative of the wars of the Romans with the provincials and the Brigantes is condensed from Tacitus’s account in the Annals, the History, and the Life of Agricola.
36. Lucan (A.D. 65) is the first who mentions them—
Martial (A.D. 96) says—
37. Valerius Flaccus (A.D. 70) says—
38. Statius (A.D. 96) has the following line in his panegyric upon Vettius Bolanus—
39. Triginta prope jam annis notitiam ejus Romanis armis non ultra vicinitatem Sylvæ Caledoniæ propagantibus.—(Plin. iv. 30.)
40. A Caledoniæ promontorio Thulen petentibus bidui navigatione perfecto excipiunt Hebudes insulæ quinque numero.—(Solinus, Polyhistor. c. 22.)
41. Ibid.
42. Silius Italicus (A.D. 68) says—
and implies that the inhabitants of Thule had encountered the Romans when he says, in another place—
Statius says of Vettius Bolanus—
Compare this with the line previously quoted.
43. The expression of Tacitus, ‘æstuaria ac silvas ipse prætentare,’ shows that this was the scene of his campaign. It is only with reference to the west coast, south of the Clyde, that such an expression is applicable, and the Solway could hardly have been excluded from it. It will be afterwards shown that the Selgovæ who occupied its northern shore were a Brigantian tribe.
44. The position of the Roman camps and forts illustrates in a remarkable manner the expression ‘præsidiis castellisque circumdatæ.’ It must be kept in view, in following Tacitus’ narrative, that from the peculiarity of his style every word is pregnant with meaning, and has a precision which has been much overlooked.
45. Paullatimque discessum ad delenimenta vitiorum, porticus et balnea et conviviorum elegantiam.—Tacit. in Vit. Ag., c. 21.
46. Novas gentes aperuit.
47. That in this campaign the Roman arms reached the Firth of Tay is distinctly asserted by Tacitus, and his clear statement cannot be explained away. Agricola could only reach it by two routes,—either entirely by land through Stirlingshire and Perthshire, or across the Firth of Forth through Fife. The former is most probable, as Tacitus usually mentions crossing estuaries where it takes place; and the latter route is moreover plainly excluded, as the nations on the north shore of the Firth of Forth were still new to him in the sixth campaign.
48. ‘Quod tum præsidiis firmabatur.’ These were obviously different from and farther south than the forts mentioned in the previous campaign. The expression ‘summotis velut in aliam insulam hostibus’ is the significant one used in fixing the barrier between the provincial Britons and the Barbarians (see chap. ii. note 70), and implies that Agricola’s intention was to add the conquered country south of the firths to the province. ‘Summotis’ does not here or elsewhere mean the actual driving out of the natives, but that those within the line of separation had ceased to be ‘hostes.’
49. The operations of this year have much perplexed historians. The obvious inference from the passage is that Argyllshire was the region he visited, and the author has entered thus minutely into the consideration of what Agricola had to accomplish, and his evident policy, to show that this was the natural step he would take. It has generally been supposed that he turned back upon his steps, and that Galloway was the country ‘opposite to Ireland’ that he visited; but, as we have seen, its inhabitants could not have been said to be ‘ad id tempus ignoti,’ and the language of the early geographers rather characterises Kintyre and the Hebrides as what impressed them most as overhanging Ireland. Chalmers, in order to avoid the plain inference from the passage, is driven to suppose that the Tavaus of the third campaign was the Solway, and that Agricola had advanced no farther, but this is quite inadmissible. The only alternative, that he crossed the river Clyde from north to south and entered Ayrshire, is equally inconsistent with Tacitus’s brief but precise language. Early writers speak of the Clyde as fordable as far down as Dumbarton, and his natural course would be to return by the same route as he came. Tacitus clearly states that he crossed ‘navi in proxima,’ which shows that it was the estuary, and not the river. The Roman fleet was then probably in the Firth of Forth, and the expression seems to imply that he took the first native vessel he could get. There is on an elevated moor in Cowall, between the Holy Loch and Dunoon, the remains of a small square fort which has all the appearance of a Roman exploratory station. It commands an extensive view, in one range, of the entire Firth to its mouth, the river Clyde for many miles of its straight course, and Loch Long penetrating in another direction into what was known to the Romans as the Caledonian Forest, and, if it is a Roman work, adds strength to the natural reading of the passage, and the expression, ‘copiis instruxit,’ is singularly applicable.
50. Chalmers has narrated the Roman campaigns with a strange affectation of military language. He makes the Roman troops debouch, defile, and deploy through the hills and in the glens in the most wonderful manner, so as to have rendered the cutting off of the whole army at any point of their progress no very difficult task to the natives. He involves the troops in this march, when the army was divided into three, among the remains of small camps in the hilly region of the west of Fife in a manner to render the real account of the transaction very unintelligible. General Roy, with correcter military knowledge, but without attending to the narrative with sufficient minuteness, is not more fortunate. He supposes that Agricola’s position was at the camp at Ardoch, and that, when he divided his army into three, he remained there with the main division, and sent the ninth legion to Comrie, and the other division to Strageath, at both of which places there are the remains of Roman camps; but, independently of the expression ‘incessit,’ which implies a march forward, conceive an able general sending the weakest legion into the heart of the Grampians, at a distance of nine miles from the main body, through an almost impassable country. So far from preventing the army from being surrounded, it sent its weakest division into the midst of the enemy. In what sense, too, could Agricola be said to have followed on the enemy’s track, and how could he, between night and daybreak, have received news of the attack, and have traversed what must have been, without roads, a long day’s march? It is obvious, on a careful attention to Tacitus’s expressions, that the three divisions could have been at no great distance from each other, and the main division nearest the enemy. There is a plan of the camp at Lintrose in General Roy’s Military Antiquities, Plate XIV., which will show how singularly it corresponds with the narrative.
51. Tacitus commences the campaign in which the ninth legion was attacked by stating that it was in the sixth year of Agricola’s administration; and in his speech before the battle at Mons Granpius he says it was then the eighth year, and that the attack on the ninth legion had taken place the preceding year. This apparent discrepancy has been usually solved by supposing the word eighth a mistake for seventh, but it is more probable that the previous campaign had lasted two years. Tacitus, after the fifth year, ceases to mark the separate campaigns with the same precision, and, perhaps, was not unwilling to gloss over the little real progress that had been made during the last three years. The expression, ‘Ad manus et arma conversi Caledoniam incolentes populi,’ probably marks the commencement of the second year of the campaign.
52. In a recent edition of the Life of Agricola, from two Vatican MSS., by Carolus Wex, published in 1852, he substitutes Tanaus, Mons Graupius, and Boresti, for the Taus, Mons Grampius, and Horesti of the ordinary editions as the correct reading of these MSS., and Mr. Burton has at once adopted the two former readings. The author, however, questions their accuracy. It is hardly possible to distinguish u from n in such MSS., and they are constantly interchanged. That Tauaus is the correct reading of the first, is plain from the form of the name in Ptolemy, Ταούα or Tava, and the real form of the second he cannot doubt was Granpius. The combination of a u or v with a labial is rarely met in Celtic words. That of the dental with the labial is very common, as in Banba, an old name for Ireland; Conpur, where the same combination occurs. As to the third there is fortunately an inscription on a Roman altar at Neuwied, brought from the Roman station of Nieder Biebr on the Rhine, where some British cohorts in the Roman army were stationed in the third century, in the following terms:—
which Mr. Roach Smith thus renders:—Idus Octobris Genio Horestorum numeri Brittonum. A. Ibkiomarus Obfius posuit titulum quintanensis posuerunt votum hoc monumentum (Collectanea, ii. part v. p. 133), which seems to leave no doubt as to Horesti being the correct form, and does not inspire one with much confidence in Wex’s new readings, sanctioned as they are by Mr. Burton.
53. There has been no point in the history of the Roman occupation of Scotland which has been more contested, or made the subject of more conflicting theories, than the position of this great battle. Gordon thought it was at Dealgan Ross, near Comrie. Chalmers, with, less difficulty, from the size of the camp, at Ardoch; others in Fife, and latterly a favourite theory has placed it at Urie in Kincardineshire. Mr. Burton abandons the attempt as hopeless.
The conclusion the author has come to is, that a careful examination of the narrative, compared with the physical features of the country, rightly apprehended, points to the site he has selected, and that it presents features which remarkably correspond with the description of the battle. This position was originally suggested in the Statistical Account of the Parish of Bendochy, published in 1797 (O. S. A. v. 19, p. 367), but has not received the attention it deserves.
The combined action of the fleet—præmissa classe—as well as the history of the previous campaigns, exclude any position west of the Tay; and if Dealgan Ross is evidently not the place, from the limited size of the camp, Ardoch is equally objectionable, from there being no hill near which answers the description of ‘Mons Granpius.’ The expression ‘transisse æstuaria’ in the plural, in Agricola’s speech, places it north of the Firth of Tay. The position at Urie involves the improbability that he marched for several days parallel to the range of the so-called Grampians, if his route was by Strathmore, and there are no camps to indicate a march nearer the coast before the battle was fought. The remains of this ‘vallum’ or rampart between the Isla and the Tay are still among the most remarkable Roman works in Scotland, and are known by the name of the Cleaven Dyke. It seems to have been the work of the same general who constructed the great camp at Ardoch, for, in connection with the latter, was a small work of an octagonal shape, with many ramparts, and the only other specimen the author has observed of a similar work is at the east end of the Cleaven Dyke.
54. See for an elaborate description of this wall Mr. Collingwood Bruce’s exhaustive work, The Roman Wall, a Description of the Mural Barrier of the North of England, third edition, 1867. The main authority for Hadrian’s work in Britain is Ælius Spartianus (181), who says, ‘Ergo conversis regio more militibus, Britanniam petiit: in qua multa correxit, murumque per octaginta millia passuum primus duxit, qui Barbaros Romanosque divideret.’—(De Hadr. 11.)
The Romans had now acquired more detailed information regarding the number and position of the tribes of Caledonia, their names, the situation of their towns, and the leading geographical features of the country. These are preserved to us, as they existed at this time, by the geographer Ptolemy, and his account of the north part of the island has apparently been compiled from the itineraries of the Roman soldiers, the observations made from the fleet in its circuit round the island, and the reports of those who had penetrated into the interior of the country. From these and other sources of information he lays down the position of the prominent features of the coast—the headlands, bays, estuaries, and mouths of the rivers, and the position of the towns in the interior, by giving the latitudes and longitudes of each. These degrees of longitude, however, are subject to a double correction. First, he places the island in too northern a latitude; and secondly, his degrees of longitude are less than the true degree, and therefore the number of degrees stated between two places is greater than they ought to be. Besides this, he has fallen into the extraordinary error of turning the country north of the Firths of Forth and Clyde to the east instead of to the north. This error mainly affects that part of the country between the Solway and the Clyde on the west, and the Wear and the Forth on the east—the coast on the west being unduly expanded, and that on the east proportionably contracted. Beyond the Firths of Forth and Clyde the effect of this strange error is to alter the points of the compass, and to substitute north for west, east for north, south for west, and west for south. The former error does not much affect the accuracy of the relative distances of places near each other. The latter, with the distortion of the distances and relative position of the localities which it creates, can be corrected without difficulty, and that part of the map reconstructed as if this error had not been fallen into. Where the country is unaffected by these mistakes, his accuracy is so great, when compared with the face of the country, that his localities can be laid down, with some rare exceptions, with considerable confidence.[55]
Ptolemy places the ‘Itunae Aestuarium’ on the west, and the mouth of the river ‘Vedra’ on the east, nearly opposite each other, and there is little difficulty in identifying the former with the Solway Firth, and the latter with the river Wear.[56] It is between these points and the river Tay that the distortion of the country takes place,—the north shore of the Solway Firth being continued in the same northern line with the west coast of England, instead of stretching to the west at right angles to it,—the Mull of Galloway being his northern point, and the northern part of Scotland made to extend towards the east. The effect is, that in the remaining part of his description the word east must be understood as really north, and that the east coast, from the Wear to the Forth, is too much circumscribed in distance, while the distances on the western side of the country are proportionably made too great. It is remarkable that the part of the country thus affected by this extraordinary mistake should be exactly the scene of Agricola’s campaigns; and it appears strange that the more northern part of the country, the information as to which he must have derived from report, and the observation of the coast from the Roman fleet, should surpass in accuracy that part of the country so often and so recently traversed by Agricola’s troops, with regard to which his means of correct knowledge might be supposed to be so much greater. We are almost led to attribute more simple truth and force to the remark made by Tacitus, that ‘it frequently happened that in the same camp were seen the infantry and cavalry intermixed with the marines, all indulging their joy, full of their adventures, and magnifying the history of their exploits; the soldier describing, in the usual style of military ostentation, the forests he had passed, the mountains he had climbed, and the Barbarians whom he put to the rout; while the sailor, no less important, had his storms and tempests, the wonders of the deep, and the spirit with which he conquered winds and waves,’ than we should otherwise suppose. If it could be inferred that Agricola’s soldiers had reported exaggerated itinerary distances, and magnified the country they had traversed, and the difficulties they had overcome, and, further, had believed, that in the second campaign, while the rest of the country was unknown to them, they were marching north instead of west, the mistake would be precisely accounted for. It seems almost to add force to this conjecture, that in the very scene where this emulation between the army and the navy is recorded to have taken place, and where a whole summer was spent in subjugating a comparatively small territory—the peninsula between the firths of Forth and Tay—the distances are still more greatly exaggerated, and the area of the peninsula increased beyond all proportion.
Be this as it may, let us follow Ptolemy round the coast, keeping in view that he designates a headland by the Greek term ἄκρον, and the Latin ‘promontorium;’ a firth or estuary by εἴσχυσις, and ‘aestuarium;’ a bay or sea loch by κόλπος, and ‘sinus;’ and the mouth of a river by ποταμοῦ ἐκβολαί or ‘fluvii ostia.’ By correcting Ptolemy’s mistake, and restoring the country between the Wear and Solway on the south, and the Tay on the north, to its proper proportion, we can identify the mouth of the river ‘Alaunus’ with that of the Alne, or Allan, in Northumberland; while the next point mentioned by Ptolemy in proceeding along the coast towards the north—the Boderia estuary—is obviously the ‘Bodotria’ of Tacitus, or Firth of Forth. Directly opposite to Boderia, Ptolemy places the Clota estuary, or Firth of Clyde, and the space between the two—the neck of land on which Agricola placed his line of forts—is correct in distance. Between the Ituna estuary or Solway Firth and the Clota or Clyde, Ptolemy has three of the rivers flowing into the Solway—the ‘Novius’ or Nith, the ‘Deva’ or Dee, and the ‘Iena’[57] estuary, or that of the Cree. They can be easily identified, though the intermediate distances are too great. He mentions the river Luce by the name of the ‘Abravannus,’ the promontory of the ‘Novantæ’ or Mull of Galloway, the Rerigonius Bay or Loch Ryan, and Vindogara Bay or that of Ayr.
Proceeding northwards along the east coast, we find the peninsula of Fife unduly extended in breadth; but the great feature of the Tava estuary, which bounds it on the north, it is impossible to mistake. Its identity with the ‘Tavaus’ of Tacitus and the Firth of Tay is perfectly clear. The position of the mouth of the river Tina, between the Boderia and the Tava, corresponds with the relative situation of the river Eden, which flows through the centre of Fife, and enters the German Ocean near St. Andrews.
Having now passed that part of the country affected by Ptolemy’s mistakes, as to its direction, the relative distances correspond more closely with those of the places meant. North of the Tava, or Tay, is the river ‘Leva,’[58] and farther north the promontory of the ‘Taexali.’ These correspond in distance exactly with the mouth of the North Esk and with Kinnaird’s Head—the north-east point of Aberdeenshire. Here the coast forms a bend in a direction at right angles, corresponding to the entrance of the Moray Firth; and proceeding along the south shore we have the river ‘Celnius’ or Devern, the ‘Tuessis’ or Spey, the ‘Loxa’[59] or Lossy, and the Varar estuary, or that part of the Moray Firth usually termed the Firth of Beauly, and separated from it by the narrow channel at Kessock. After this the distances, if measured in a straight line, are found to be too great, but if the windings of the coast, which is here greatly indented, are followed, they are sufficiently correct, showing that they are derived from the itineraries of coasting vessels, and that the Moray Firth had been in fact explored. Looking across the lowlands of Easter Ross, the first landmark noticed are the high hills on the north of the Dornoch Firth, and two stand prominently out, forming the two sides of Strathfleet or Little Ferry. One of these great landmarks is noted as Ὄχθη ὑψηλή, ‘Ripa alta,’ or the high bank. Beyond these to the north is the mouth of the river ‘Ila,’ corresponding in situation with the Helmsdale river, termed by the Highlanders the Ulie. We have then three promontories noticed—the ‘Veruvium,’ the ‘Vervedrum,’ and the ‘Orcas’ or ‘Tarvedrum.’ The editions of Ptolemy vary as to their relative positions, but it is impossible not to recognise the three prominent headlands of Caithness,—the Noss Head, Duncansby Head, and Dunnet Head.
On the west coast, proceeding north from the Firth of Clyde, Ptolemy notices the ‘Lemannonius’ Bay or Loch, which corresponds in situation with Loch Long,[60] although the resemblance of name would almost lead us to infer that the geographer believed that Loch Lomond opened upon the sea. He next mentions the promontory, early known to the Romans as that of Caledonia, under the name of the Epidium promontory, which is obviously Kintyre. North of the Mull of Kintyre he places exactly in Crinan Bay, which must always have been a well-known shelter for vessels, the mouth of the ‘Longus’ river, where we now find the river Add,[61] known to the Highlanders as the Avon Fhada or long river.
THE FIVE EBUDŒ
OF PTOLEMY
Compared with
THE ISLANDS
South of
ARDNAMURCHAN POINT
W. & A.K. Johnston, Edinburgh & London.
Between Scotland and Ireland Ptolemy places the five islands which he terms the ‘Ebudæ,’ and the island of ‘Monarina;’ but these islands are attached to his map of Ireland, to which country he held them to belong, and their situation is not affected by the great mistake he committed in the direction of Scotland. The most northerly of the five he terms ‘Maleus,’ which is so obviously the island of Mull that it gives us a clue to the situation of the rest, and shows that the islands meant were those south of the point of Ardnamurchan. The remaining four, placed in a line on the same degree of latitude, and lying from west to east, are termed the two ‘Ebudas,’ ‘Engaricenna’ and ‘Epidium.’ The relative situation of the western ‘Ebuda’ towards Ireland corresponds closely with that of Isla, and the two ‘Ebudas’ were probably Isla and Jura. Scarba corresponds with ‘Engaricenna,’ and the more distant Lismore with ‘Epidium.’ These islands all lie in one line from south-west to north-east. ‘Monarina’ corresponds in its position towards Ireland with the island of Arran.[62]
Beyond the point of Ardnamurchan the western islands seem to have been comparatively unknown. No islands are mentioned which correspond with the Outer Hebrides, and the island of Skye seems only to have been known by name, as it is probably meant by Ptolemy’s island of ‘Scetis,’ which however he places apparently at random near the north-east promontory of Scotland. On the mainland three points only are noticed,—the mouth of the Itys river, which is probably the river Carron flowing into Loch Carron; the Volsas Bay or Loch, which can only be the great arm of the sea termed Loch Broom; and the mouth of the river ‘Nabarus,’ obviously the Naver; but these points must have apparently been taken from report, as it is difficult otherwise to account for his ignorance of the true position of Skye, and for the absence of all mention of the great headland of Cape Wrath, forming the north-west point of Scotland.
Along the east coast he denominates the sea the Germanic Ocean, and along the west, from the Mull of Galloway to Dunnet Head, the Deucaledonian.
Such is the wonderfully accurate notice of the salient features of the coasts of Scotland given by a geographer of the second century; but his description of the tribes of the interior of the country, and the position of what he denominates towns, as compared with the physical appearance of the country, is no less so. To these tribes Ptolemy assigns definite names, and to some the possession of what he terms πόλεις in Greek, and in Latin ‘oppida.’ That these towns were not exclusively Roman stations is plain from their being mentioned in a part of the country to which the Roman arms had not yet penetrated; neither could they have been simply the rude hill-forts, or primitive shelters in the woods, such as are mentioned by Cæsar; for they are only to be found in the southern and eastern districts, and none are noticed as we approach the rude tribes of the hill country. They certainly implied a regularly fortified town, in which the habitations of the natives were collected together, and formed the great defences of their territories, as we almost invariably find them placed near the frontiers of each tribe, or the great passes from one district to another. They would naturally form the main points of attack in any assault upon the tribe; and accordingly we usually find, within the sphere of the Roman operations, a Roman camp placed in the immediate vicinity of the remains of these towns; and the Roman stations or roads are useful in assisting the accurate identification of these within the range of their campaigns.
A line drawn from the Solway Firth across the island to the eastern sea exactly separates the great nation of the Brigantes from the tribes on the north; but this is obviously an artificial line of separation, as it closely follows the course of the Roman wall shortly before constructed by the Emperor Hadrian, otherwise it would imply that the southern boundary of three Barbarian tribes was precisely on the same line where nature presents no physical line of demarcation. There is on other grounds reason to think that these tribes, though apparently separated from the Brigantes by this artificial line, in reality formed part of that great nation.[63] These tribes were the Otalini or Otadeni and Gadeni, extending along the east coast from the Roman wall to the Firth of Forth. They had three towns—on the south ‘Curia’ and ‘Bremenium,’ whose situations correspond with Carby Hill in Liddesdale, where there is a strong native fort, and opposite to it a Roman station, and High Rochester, in Redesdale. Their northern frontier was guarded by the town of Alauna, which is placed by Ptolemy in the Firth of Forth, and corresponds in situation with the island of Inchkeith.[64]
Farther to the west, the Selgovæ or Elgovæ occupied the county of Dumfries, being bounded on the north by the chain of hills of which the Lowthers formed the highest part, and extending along the shores of the Solway Firth as far as the river Nith. Their towns were ‘Trimontium,’[65] in the exact position where we find the remarkable Roman remains on the striking hill called the Birrenswark hill; ‘Uxellum’ corresponding in situation with the Wardlaw hill in the parish of Caerlaverock, where there are the remains of Roman and native works; ‘Corda’ at Sanquhar, in the upper part of the valley of the Kith, a name which implies that it was the site of an ancient Caer or native strength. The remaining town of the Selgovæ—Carbantorigum—is placed by Ptolemy on the exact position of the remains of a very remarkable stronghold termed the Moat of Urr, lying between the Nith and the Dee.
To the west of the Selgovæ lay the tribe of the Novantæ, occupying the modern counties of Kirkcudbright and Wigtown. Their towns were—Lucopibia at Whithorn, where there are the remains of Roman works, and Rerigonium[66] on the eastern shore of Loch Ryan, the fortified moat of which is still to be seen on the farm of Innermessan.
North of the Selgovæ and Novantæ, and separated from them by the chain of hills which divides the northern rivers from the waters which flow into the Solway, was the great nation of the Damnonii, extending as far north as the river Tay. They possessed south of the Firths of Forth and Clyde the modern counties of Ayr, Lanark, and Renfrew; and north of these estuaries, the counties of Dumbarton and Stirling, and the districts of Menteith, Stratherne, and Fothreve, or the western half of the peninsula of Fife. This great nation thus lay in the centre of Scotland, completely separating the tribes of the Otalini or Otadeni, Selgovæ or Elgovæ, and Novantæ, on the south, from the northern tribes beyond the Tay, and were the ‘novæ gentes,’ or new nations, whose territories Agricola ravaged to the ‘Tavaus’ or Tay in his third campaign. They possessed six towns,—three south of the firths, and three north of them. Their towns in the southern districts were ‘Colania,’ near the sources of the Clyde, a frontier but apparently unimportant post; ‘Coria,’ at Carstairs, on the Clyde near Lanark, which, from the numerous remains both Roman and native, appears to have been their principal seat; and ‘Vandogara,’[67] on the river Irvine, at Loudon Hill in Ayrshire, where there are the remains of a Roman camp, which was afterwards connected with ‘Coria,’ or Carstairs, by a Roman road. In their northern districts the geographer likewise places three towns,—‘Alauna’ at the junction of the Allan with the Forth, a position which guarded what was for many centuries the great entrance to Caledonia from the south; ‘Lindum’ at Ardoch, where the number of Roman camps, and of hill-forts which surround them, indicates an important position; and ‘Victoria,’ situated at Loch Orr, a lake in the western part of Fife, occupied by this nation, where there are the remains of a Roman station.
On the east coast, the ‘Vernicomes’ possessed the eastern half of Fife, or the ancient Fife exclusive of Fothreve, and the counties of Forfar and Kincardine. The only town mentioned is ‘Orrea,’ which must have been situated near the junction of the Earn with the Tay, perhaps at Abernethy. The nearest Roman station to it is at Ardargie. Farther north along the coast, and reaching from the mountain chain of the Mounth to the Moray Firth, were the ‘Taexali,’ who gave their name to the headland now called Kinnaird’s Head. Their town, ‘Devana,’[68] is placed by Ptolemy in the strath of the Dee, near the Pass of Ballater, and close to Loch Daven, where the remains of a native town are still to be seen, and in which the name of Devana seems yet to be preserved.
West of the two tribes of the ‘Vernicomes’ and the ‘Taexali,’ and extending from the Moray Firth to the Tay, Ptolemy places the ‘Vacomagi,’ a border people, who lay along the line separating the Highlands from the Lowlands. The remarkable promontory of Burghead on the south side of the Moray Firth, on which the ramparts of the early town are still to be seen, was one of their positions, on which they had a town termed πτερωτὸν στρατόπεδον, Alata Castra,[69] or the Winged Camp. They had another town on the Spey near Boharm, termed Tuessis. Their frontier towns at the southern termination of their territory were ‘Tamea,’ placed on the remarkable island in the Tay, termed Inchtuthil, where numerous remains exist, and ‘Banatia’ at Buchanty on the Almond, where a strong Roman station is overlooked by the commanding native strength on the Dunmore Hill.
To the north and west of these tribes no further towns are mentioned; and as the Caledonii extend on the west along the entire length of the territories of the Vacomagi, their eastern boundary formed the line of demarcation between the tribes of the more plain and fertile districts, who had advanced one step in the progress of social life in the possession, even at this early period, of settled habitations and determined limits, and the wilder tribes of the mountain region, among whom nothing deserving the name of town in its then acceptation was known to the Romans. Ptolemy states that the Caledonii extended from the ‘Lemannonius Sinus,’ or Loch Long, to the ‘Varár Aestuarium’ or Beauly Firth, thus ranging along the entire boundary of the Highland portion of Scotland. On the west they had the remarkable chain of hills termed in the early historical documents ‘Dorsum Britanniae,’ Drumalban, or the backbone of Scotland, a native term apparently presented in a Greek form in Ptolemy’s καληδόνιος δρυμός, and converted by his Latin translator, who, puzzled by the term δρυμὸς, recognised in it only an unusual Greek word signifying an oak wood, into ‘Caledonius Saltus’ or Caledonian Wood. That this range of hills was at all times a forest in the highland acceptation of the term, having its southern termination at the head of Lochs Long and Lomond, there is no doubt.
North of the Caledonii, on the other side of the Varar or Beauly Firth, lay the ‘Canteæ’ or ‘Decantæ,’ possessing the whole of Ross-shire save the districts on the west coast. Sutherland proper was possessed by the ‘Lugi’ and ‘Mertæ.’ Along the west coast, from the Firth of Clyde northwards, were the ‘Epidii’ in Kintyre and Lorn. Beyond them the ‘Creones’ or ‘Croenes,’ extending probably from the Linnhe Loch to Loch Carron. Beyond them the ‘Carnones,’ occupying probably the western districts of Ross-shire. Beyond these again, in the west of Sutherland, the ‘Caerini;’ and along the northern termination of Scotland, including Caithness and the north-west of Sutherland, were the ‘Curnavii.’ Such were the northern tribes of Britain as described by the geographer Ptolemy in the second century, and such the knowledge the Romans now possessed of their position, and of the towns they occupied.