Proc Soc Antiq Scot, 113 (1983), 373-378

Caledonia and its occupation under the Flavians J G F Hind*

Tacitus mentions Caledonia inhabitants it d an s (incolentes populi; civitates), who callme h s Britons, four times in his De Vita Agricolae. Each time it is the area north of the line between Forte th Clydd han e estuaries whic designateo s s hi d (Agricola ; 27) numbe10 A . f separato r e peoples inhabitant e madth p eu s (Agricola take29.4; e b 25 o n,t together with 31.5) picture Th . s ei clea internalld an r y consistent t inconsistenno d an , t wit othee hth r sources, given that thee yar either very general and poetic in tone, or topographical and list-like in form. The writers of the 1st century AD seem to have thought of Caledonia vaguely in connection with forests (Pliny, NH 4.102; Silius Italicus, Punica 3, 597-8); wild seas (Lucan, Pharsalia 6.37), and wild bears (Martial, Sped. 7), and with a very remote people (Martial 10.44.1). s Agricolahi n I notoriously offerf topographyo y s e doelittlwa h y t seb ye giv, e three place-names relatin thigo t s area Forte Th . h (Bodotria, correc e thaf i th s ti t form Ptolemr ,fo y gives Boderia, Ravenne th d an a Cosmographer ,Clyde Bdora)th d e (Clotd)an th e saib e ar o dt delimiting estuarie soute th f Caledonia.o ho st Mons Graupius, e r Craupius, o th site f th eo s wa culminating victor f Agricolo y a ove e peopleth r f Caledonia.o s Wherever s preciselythiwa s , it presumably was in the hands of one of the constituent tribes of Caledonia, which, we learn from Tacitus, include Boresti,e th d infen ca and re frow , m Ptolemy, include e Vakomagoi,dth Taizaloi, Venikones and perhaps the Dumnonioi (fig 1). Tacitusn I r mosou , t circumstantial source, Caledonia a regio s i f ,no almost another island, and its peoples, though Britanni, are physically somewhat different from the rest (rutilae Caledoniam habitantium comae, Agr. 11). They appea havo rt e forme dconfederatioa f no tribes, at least in time of war (civitates trans Bodotriam sitas, Agr. 25; Caledoniam incolentes populi, Agr. 25; legationibus foederibust e omnium civitatium vires exciverant, - Agr.'people ) 29 s situated beyond the Forth'; 'peoples inhabiting Caledonia'; 'they had brought out all the peoples in strength by sending envoys and invoking treaties'. Tacitus estimated their total strength in battl fightin0 t soma e00 0 e3 g men f whico , h they los thira t casualtien di t Monsa s Graupius. There is no way of checking this figure, but it tallies with the view that several tribes were involved. Later, Florus know silvasd saltusa an f . s (i.12o c . c , 45); Cassiu o o speaktw Di s f o s peoples, Maiatai and Kaledonioi (the Greek form), of which the former lived nearer to the Wall (ie Hadrian's Wall), and the latter further to the north (Cassius Dio, 75.5; 76.12; 15). The emperor, , campaigned against the Kaledonioi, probably in AD 209 and against both peoples the following year. Late in the next century Ammianus Marcellinus knew of two branches of the , one of whom he calls Verturiones and the other Dicaledones (A.M. 27.8.5; Mann 1974, passim). These latter are clearly descendants of the Kaledonioi of the early 3rd century, and the former may be descendants of the Maiatai. A similar prefix, Due-Caledones, had appeared * School of History, University of Leeds, LS2 9JT 25 37 | 4SOCIET Y OF ANTIQUARIES OF , 1983

I VI A N ERTAE

X A LI)

FI G1 Suggeste d locatio f Caledoniino othed an r tribes

in Ptolemy's Geographia, where the northern ocean is called Okeanos Duekaledonios (Ptol. Geog, 2.3.1). Other writers of the later Roman Empire perpetuate the vagueness of the 1st- century poets. Ausonius speak f Caledoniao s n pearls (Mosella 68); Claudia f Caledoniano n faunae cos.V th frost (I d Honoriian s / cos. ; 26 Stilichonis Caledoniane 2.247)th f welo s a ,s a l s generally (Laus Serenae 45). Sidonius Apollinaris likewise d probablan , y anachronistically, speaks of (Carm. 1.89). The forest of Caledonia was still a by-word in the 5th and eve centurieh n9t (MartianuD A s s Capella 6.666; Nennius Hist. Britt. e 56)CaledonesTh . e ar mentioned alongside 'other Picts' in the Panegyric of AD 310 (Pan. Lat. Vet. VI (VII) 7.2), and as Calidoni alongside Verone Pictid th an n i a Lis Af to D 312/14 (Jones 1954, 21-9). Our most useful, because most detailed, information comes from Ptolemy, who mentions a Kaledonios drumos - 'Caledonian Forest' (Geogr. 2.3.8). The evidence for the location and extent of the Caledonian Forest in ancient times and the Dark Age down to medieval times has been HIND: CALEDONIA AN OCCUPATIOS DIT N UNDER THE FLAVIAN5 37 | S studied by B Clark (1963, 191-5). He came to the conclusion that it comprised, above all, the area from Stirling to , and from in the north-west as far as towards the north-east. Ptolemy regarded the 'Forest' as being 'above' (ie to the north or north-west of) people th calle eh s Kaledonioi. people th f I e took their name fro mountainoue mth forested san d area, then they probably dwelt in the main glens and straths to the south of the massif (Hanson Macinned an s come clos thio et s view (1981, 103)). Ptolem facn i s tyi more specific aboue th t Kaledonioi; their lands stretched from Lemannonios Kolpos, 'The Elm-Tree River Gulf, on the Atlantic coast and Varar Eiskhusis, 'Varar Estuary' (Ptol. Geogr. 2.3.12), on the North Sea coast. certais i t I n that Beaule Vararth s i y River (also know Farrars Morane a th n o y) e Firtth t ha norther f Loco d h nen Nes s (the Grea te identit Glen)th t bu y, of Lemannonios Kolpos bees ha n treated with some ambivalence waywardnessy sa o t t ,no t ough .LocI e b o htt Lon Locr go h Fyne fro positionins mit Ptolemy gb y sout Kintyref ho , spitbutn i ,thisf eo , Richmond place furthet di r north, and Rivet and Smith suggested that the text is in error, so that the Kaledonioi of Ptolemy could be said to stretch from sea to sea along the Great Glen, with their most southerly point being Loch Linnhe (Firt Lomef Atlantihe o th n coasa )o cSe t (Richmon 55n d , ;1944 195840 , , 135-6; Rivet & Smith 1979, 107, 133, 140-1, 189-91). Rivet and Smith posit some error which has cause geographee dth brino t r Caledoniie gth wese dowth r southt o nfa t coas o ,to tnea r Lennox (na Leamnaich), near which they rightly Lemannonios e takb o et Kolpos - presen t day, Loch Fyne or Loch Long (Rivet & Smith 1979,133, 387). All subsequent modern writers see locatmo t e Ptolemy's Kaledonioi, implicationy b d an , maie th n tribe Tacitusf so ' Caledonian confederacy eitheGreae th op f n rRomani t Ma Gle S n(O Britain, ed 4, 1978, North Sheet and p 15; Breeze 1982, 30), or in the Highland massif to its south (Clark 1958,49; Hanson & Maxwell 1983,4). Neither is in the least plausible, since the land could t supporno populatioa t n that woul able d b lea o e t powerfu o ds confederacya l doubtfus i t I . l that any tribe dwelt high up in the Grampians at any time in antiquity in sufficient numbers to become politically dominant ove lowlane th r d tribe f Caledonia,o s e samth probabls ei d an y true also of the Great Glen. What is more, if centred in the Glen, it is certain that the people would have had little direct contact with the Romans during their occupation of , Mearnse Anguth d san . The alternative, which seems to me much more likely, is that the Kaledonioi of Ptolemy stretcd di h from Loch Fyn Locr eo h BeaulLone th go t y Firth arount ,bu southere dth easterd nan n edges of the Highland massif, and not the western and northern edges, as modern scholars have assumed. Lik e Caledoniameth incolentes f Tacituso , they embrace wide dth e strath lowed san r glens from Loch Lomond (near Loch Long), and included the upper Forth, Earn, Tay, Isla Mearnse coastae th th d d lan an , plains aroun Farrare Grampiane dth th s a termn I r . fa f s o s a Scottish counties the confederacy would control Stirlingshire, Perthshire, , Angus, Aberdeenshire, Mora d Nairnan y . This supposes that they controlle e goodth l ,al d well-populated, land of central and north-eastern Scotland, with a probable centre in Perthshire. The fact that Ptolemy assigns to them no populated centres (poleis) could be explained if the Vakomagoi, TaizaJoi, Venikones, d mayban e Dumnonioi,eth were Caledonia l th par al ef o t n grouping of tribes mentioned above. These had poleis, according to Ptolemy, which may have been, in the main, native centres, some with Roman forts quartered among them (Rivet 1977, 45-64). Bannatia, Tamia, Pinnata Castra Tuessis,d an assigne Vakomagoi,e th o dt Devana assigned to the Taizaloi or Taixaloi, Horrea, to the Venikones, were all within the Caledonian area. The situation of the Dumnonioi is more problematical, since two of their six poleis lay probably on Forth-Clyde th s soutit eo ht line e (Rive ,on nort e threSmitth d & t ho an et h 1979, 139-40). If this tribe straddled the Forth-Clyde line, they could conceivably have contributed to the Caledon- | SOCIET 6 37 ANTIQUARIEF YO SCOTLANDF SO , 1983 ian hostilit Romanse th o yt ,fourte sincth y heb campaign's activities, Agricol beed aha n seeking to cut off the more southerly peoples from contacts with the north by building forts at certain point than so t line across Britain. Strong support viee th w r tha fo ,centre Caledoniie th tth f eo valley Ta prope ye is nth wa r and Strathmore, comefore th f place-namem o n s i Uppee th valle y n si Ta r y which reflect their presence precisely in this region - Dunkeld (once Dunchaillean, Dmchailden (Jackson 1954, 14-16)). Their name appears likewise in the nearby mountains and hillforts, Rohallion, Schiehal- lion (Watson 1926, 21). If further confirmation were needed of the importance of Perthshire, it is to hand in the proliferation of monuments on Tayside, in the Iron Age and Dark Age (Childe 1953, Map IV; Coutts 1970, 27-68). In the Dark Age too, the Picts (whom Ammianus divided into Verturiones Dicaledones)d an inhabited this same rang territoryf eo , whic have hw e supposed Ptolemy intended for his Kaledonioi. The seven provinces of the Picts were Strathearn and Men- teith, Atholl and Gower, , Angus and the Mearns, and , and Ross, (Wainwright 1955, 46-7). At least the first six were, if we can so interpret Ptolemy, withi aree nKaledoniaf th ao n control (see also Alcock 1980, 61-2) mighe W . t assume thae th t Vakomagoi wer majoea r majoe elementth t r no element f i , Caledoniae th n ,i n groupin tribef go s in the earlier period of Roman expansion. A similar conclusion was reached by Maxwell (1980, 7). Only late thepopulid rdi Caledoniam incolentes com regionae takth o e t n eo l nam peoplea s ea , perhaps in the later 2nd or early 3rd century, somewhat like the Franci, Saxones and Alamanni in Germany (Mann 1974, 40-1). As to the other more northerly peoples of Scotland who are presente Ptolemy db y without town centre populatiof so n (Dekantai, Lougoi, Smertai, Komavioi, Kairenoi, Karnonakai, Kreones, Epidioi), these were mainly coastal populationy groupsma o wh , have been recorded by Agricola's fleet in the fifth campaign, and when it rounded Caledonia and Orkneye th finae th ln si yea r (seventh campaigns hi f )o s (Agricola , the 38)so yf I .wer e probably smal t cove no interioe lo d rth p group f di Roman mucMa o o rs S sappearwh s O ha e Britainth n si elsewherd an Maxwelg e(e l 1980 ; Breez3 , e 1982 ; Hanso30 , Maxweln& moss i l t 1983I t . 4) , unlikely that they formed any significant part of the Caledonian civitates who fought Agricola at Mons Graupius; those were probabl tribee yth s settled soutMounthe easth d f han to . Ther interestinn a s ei g consequenc f thieo s re-deploymene Caledoniie th th f o e d th f an o t recognition tha Uppee tth r Glens, includin Greae gth t Glen, hel greao dn t power base which could challeng Romanse eth t become.I Romane sth cleay lase nexrwh e th tfort th decado t f e so th f eo 1st century AD were so heavily concentrated in Stirlingshire, Kinross-shire, Perthshire and Angus - 12 forts north of the isthmus, some having two periods. Ardoch, Strageath, Cardean and Stracathro were intende holo dt d dow Caledoniae nth n peoples (Dumnonioi r perhapo s Boresti and VakomagoiT) in their main areas of , Strathearn, Strathmore and . The line of so-called 'glen-blocking' forts, located further to the west and north, at Drumquassle, Bochastle, Dalginross, Fendoch, and the legionary fortress itself at Inchtuthil, have been inter- preted variously. Richmond, followe mosy db t modern scholars, thought that they were intended (perhaps by Agricola) to prevent the Caledonii from beyond the occupied area attacking down the glens (Richmond 1955, 45; 1958 50-1; Frere 1981, 89-91). More recently doubts have been expressed about this reason for their disposition. D J Breeze, for instance, following a hint by J Mann, thinks that the forts may have been intended to act as springboards for expeditions glene th sp u (Mann 1968, 308; Breeze 1982take w ef ,i int t 55)oBu . accoun likelihooe th t d that significant centre populationf so thud an ,substantiaa s l threat, apart from some fugitivesd di , not exist in the upper glens, still less on the mountains, we may seek a different solution to the problem of the purpose of the forts. The forts indeed appear to be 'glen-blocking', but rather opposite inth e sens Richmone e froon minde n mi th d d.ha Surely, their main purpose woulo t e db HIND: CALEDONIA AND ITS OCCUPATION UNDER THE FLAVIANS | 377 maintain surveillance ovee defeatedth r d temporarilan , y occupied inhabitants e fertilth f eo area f Caledonia.o s forte th sf I alon maie gth n lin f Stratheareo Strathmord nan thid b edi sjo centrareae e in'glen-blockingth e th f th , eo ' forts were intende holo dt people dth e back from leaving their field decampind an s glense th p situatioa ,gu n that lasted only some seve r eighno t years from A D(se0 83/9 e o 4t Hanso n Fendoch) n 1980o acciden, o n 30 ,s i t I . t that these forts were also fringing the probable extent of the 'Caledonian high forest', the Kaledonios Drumos. If this argumen 'glen-blockinge th tru s r i t efo ' auxiliary fort s- tha t they were enforcinga polic deductiof yo planitiem,n i keeping Caledoniie th t civiliaa n pursuit strathse th n si - the s i t ni tru legionare e th als r ofo y bas t Inchtuthilea , jus Rivee t ben e (Frery th nort th n Ta rdi f ho e 1981, 92-4). It lay a few miles downstream from Dunkeld, maybe even then the stronghold maie oth f n grouCaledonii.e th f po Most sites likelwa d t yi wit intentioe hth f cuttinne o th f gof farmer ordinard san y tribesmen from acces theio st r hillfort rallying centre frod san m e flighth p u t y Gorge Ta e advanceTh . d seemingldan y exposed situatio e fortressth f no , whic s oftehha n been remarked upon, coul explainee db frontiea t no ds r thusposwa t I .t designe keeo dt p back untamed raiders garrisona t bu , , planted deliberatel preciseld yan keeo yt p recentle watcth n ho y defeated people f Strathmorso Strathtayd ean .poino Thern placins n i t ewa g auxiliary fortn i s advance of it, ie higher up the Tay Gorge, since the main threat actually lay in the reverse direction soute north-eastth d ho an t , wher line auxiliarf eth eo y forts actually was. Lik commandee eth r oRhine earle legioa f th th yn n e i no year f Augustusso e legiolegate th th , f t Inchtuthina o e l was looking back into recently conquered territory, as much as, if not more than, beyond into untouched barbarian land. More close in time and space, Agricola had himself only recently surrounded defeated peoples with forts in the Pennines and North England (Agr. 20.3, civitates praesidiis castellisque circumdatae), preventing population groups escaping int highlandse oth . If Strathmore and the Mearns were as important as is here supposed, then Tacitus' view that Britai beed nha n 'conquered, the t go'n le eves ,i n more understandable than unde hypoe rth - thesis normally held (turn primum perdomita est, Agr. 10; perdomita Britannia t statime missa, Hist. 1.2.1). The Northern peoples had been subdued between AD 83/4 and 90; the wedge-shaped e unproductivmasth f o s e land, interspersed wit e interminablh th firth d an s e shore-line (Agr. 10.4-6), had been left alone after the exploration by the fleet. The Romans were not interested in ruling rocks. After the battle at Mons Graupius they clearly controlled the major centres of the Caledonii in Strathmore, Strathtay and Glen Isla. The only remaining opposition to them would be relatively mino cowede th r- , buoccupiedt tno , tribesme Aberdeenshiref no , Bucha Morayd nan . The upper glens, I suggest, contained nothing but a few irreconcilables and refugees, not whole tribes ready to threaten the briefly held extension to the Roman pro vincia in Caledonia. The Roman retreat from north of the Forth would, on this theory, have been occasioned not by British threats from beyon provincia,e d th genera y b t bu l strategic consideration Empire th wholea n i ss ea , especiall neee r Danube troopth dth fo y yb n o s e (Agr. t 41)thiBu s. conclusio s commonlni y arrive specialisty b t arede a period th a an n si d (Breeze 1982, 60-1; Hanso Maxweln& l 1983, 43-4). It was only when Caledonia had been evacuated that a significant population group was fre becomo et e hostil aliee th no e t powe r again, late emergo rt Maiatai e th s Kaledonioi.ed a an

ACKNOWLEDGEMENTS gratefu colleaguem y a m I o t l , Harold Mattingly readinr fo , gdrafa thif o t r s papefo d an r some suggestions as to presentation, and to the Editor and Readers of this journal for some references and other useful observations. Leo Rivet read a typescript copy at a late stage. None of the above is to be taken as agreeing necessarily with either, or both, of my main points. 378 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

REFERENCES Alcock L , 1980 Topuli Bestiales Pictorum Feroci Animo surve:a Pictisf yo h settlement archaeology', in Hanson, W J & Keppie, L J F (eds), Roman Frontier Studies, 1979, Oxford, 61-95. (=Brit. Archaeol Intp Re Ser, 71.) Breeze J northern e 198,D Th 2 frontiers of . London. Childe, V G 1935 The prehistory of Scotland. London. Clark ,B 196 9 'Calido Caledoniae th d nan n Forest', Bull Board Celtic Stud, (1969)3 2 , 191-5. Clarke, J 1958 'Roman and Native A.D. 80-122', in Richmond, I A (ed), Roman and Native in North Britain, Edinburgh, 28-59. Coutts, H 1970 Ancient monuments ofTayside. Dundee. Frere, S S 1981 The Flavian frontier in Scotland', Scot Archaeol Forum, 12 (1981), 89-97. Hanson, W S & Macinnes, L 1981 'Forests, forts and fields', Scot Archaeol Forum, 12 (1981), 98-113. Hanson Maxwell& S W , G ,198 3 Rome's north-west frontier: Antoninee th Wall. Edinburgh. OS 1978 Ordnance Survey, Map of Roman Britain (fourth edition). Jackson H 195,K 4 'Two early Scottish names', Scot Hist Rev, (1954)3 3 , 14-19. Jones195M H 4 ,A 'The Veron datvalue d th ean f eo a List', / Roman Stud, (1954)4 4 , 21-9. Mann, J 1968 Review of Ogilvie-Richmond edition of Tacitus' Agricola, Archaeol Aeliana, 4 ser, 46 (1968), 308. Mann, J 1974 'The northern frontier after AD 369', Glasgow Archaeol J, 3 (1974), 34-42. Maxwell, G S 1980 'The native background to the Roman occupation of Scotland', in Hanson, W S & Keppie, L J F (eds), Roman Frontier Studies, 1979, Oxford, 1-13. (=Brit Archaeol Rep Int Ser, 71.) Richmond, IA 1944 'Cn. Julius Agricola', J Roman Stud, 34(1944), 37-45. Richmond A 195 I , 8 'Ancient geographical source r Britaifo s n nort f Cheviot'o h Richmondn i , , IA (ed), Roman Natived an North n i Britain, Edinburgh, 131-55. RivetF 197L 7A , 'Ptolemy's geograph Flaviae th d nyan invasio Scotland'f no Hauptn i ,Horn & D , , StudienMilitargrenzenG n de H u z Roms, , Cologne/BonnII , 45-64. Smith& RivetF ,C L 197,A 9 The place-names of Roman Britain. London. Wainwright, F T (ed) 1955 The problem of the Picts. Edinburgh. Watson, W 1926 The Celtic place-names of Scotland.