GELLIGAER CAER RUFEINIG

ROMAN FORT

By Richard J Brewer, Amgueddfa Cymru – National Museum Wales

Plan of the stone fort as revealed by excavations undertaken atGelligaer between 1899 and 1913.

Buildings in the fort:

Barracks 1-2 & 12-15

Stores 3, 10 & 11

Workshop and yard 4

Granaries 5 & 8

Commander's house 6

Headquarters 7

Stable 9

The annexe:

Bath-house A

Bath-house exercise yard B

Guard-chamber C

Yard D

Workshop E

Unknown F

Cover: Workmen uncovering thehot room of the bath-house duringJohn Ward's excavations in 1909.

Gelligaer: The Legacy of Rome

1

The Roman Empire was one of the greatest that the world has ever known. Thestrength and prestige of the Roman armies have never been forgotten. At itslargest, in the 100s A.D., the Roman Empire covered much of Europe, NorthAfrica and the Near East. Britain was a relatively late addition to the Empire,becoming its most north-western province. Gelligaer is part of that history.

A Frontier FortThe fort at Gelligaer, of which theRoman name is lost, occupies aposition on a long broad ridge betweenthe Taff and Rhymney valleys. Itcommanded an extensive view of thecountryside, which in Roman timeswould have been heavily wooded. Itwas one of a line of forts, generally seton the higher land between the rivervalleys, running from Cardiff to Y Gaer(Brecon). The fort (Gelligaer 2), whichwas constructed in stone, is about 117m square, 1.4 hectares in extent. It liesto the north-west of the Parish churchof St Catwg. No masonry is now visible,

but the defences can still be seen as alow broad bank, best preserved on thesouth-west side, where the ditch is alsovisible. Depressions in this bank markthe sites of the four gateways, ofwhich that on the north-east side isthe most pronounced. Threeinscriptions found near the south-eastand north-west gates date theconstruction of the fort to the periodA.D. 103–11, during the reign of theEmperor Trajan.

Adjoining the fort on the south-eastwas an annexe of which nothing now

remains to be seen. Half of the annexewas occupied by the regimental bath-house, one of the finest known at anysuch fort in Wales. Other features thathave been uncovered include a parade-ground on the north-east side of thefort and a Roman tile-kiln in what isnow the churchyard. Roman cremationurns containing burnt bones werefound in 1910, and would seem toindicate a cemetery beside the roadleading south-east from the fort.

The stone fort was not the first Romanmilitary establishment at Gelligaer. Alarge rectangular earthwork to thenorth-west, 2.2 hectares in extent, isan earlier earth and timber fort(Gelligaer 1), probably built at the timeof the Roman conquest of Wales inA.D. 74–78. This enclosure is stilloutlined in parts by a faint bank.

To prevent any interference with thesite of the fort, it has been listed as ascheduled ancient monument, and isprotected by the Ancient MonumentsActs 1979.

Location of Roman features at Gelligaer.

1. Earth and timber fort2. Stone fort3. Annexe4. Parade-ground5. Tile-kiln6. Cremation burials

Exploration of the Site

2

Archaeological interest in the sitebegan in July 1894, when the CardiffNaturalists’ Society visited the fort. Thepossibility of an excavation wasdiscussed, but it was not until October1899 that a trial exploration began.Results proved promising and large-scale excavations were carried out bythe Society for much of 1900 and1901. The main aim was to recover theplan of the internal buildings and toexamine the defences of the stone fort.To find the position of the buildings,labourers dug parallel trenchesdiagonally across the interior of thefort. When they came across a wall, itscourse could be followed by diggingalong the top of it. Not all thebuildings were planned in this manner:some were more extensively examined,in particular the two granaries whichwere uncovered completely. The wallsof all these buildings rarely survived toa greater height than one metre abovethe Roman level, and quite often onlythe foundations remained.

At this time, excavation techniqueswere in an early stage of developmentand much of the work was carried outin a haphazard fashion. The number of

labourers employed on the site variedbetween two and twelve, and theywere supervised by interested membersof the Society, working on a weeklyrota basis. These supervisors were oftenbusinessmen who were unable todevote more than a few hours to theexcavations during their allotted week,and this led to much confusion. Veryfew notes were kept, and all the findswere mixed together. Despite all this, abrilliant account of the work waswritten by John Ward, who was thenthe Curator of the Municipal Museumand Art Gallery in Cardiff. The resultsstill provide a textbook example of thelayout of a Roman fort.

From 1908–13, John Ward undertookexcavations of his own outside thedefences of the fort. His work revealedthe annexe, parade-ground and thelarge rectangular earthwork to thenorth-west of the fort, which hesuggested was a temporary camp, usedto accommodate the soldiers duringthe construction of the stone fort.However, a more recent excavationconducted in 1963 by the lateProfessor M G Jarrett of CardiffUniversity in the south angle of this

enclosure has shown it to be an earlier fort.

The excavation, in 1913, of the Romantile-kiln, which was found accidentallyduring the digging of a grave in thechurchyard, was undertaken by the Rev.T J Jones, the Rector of Gelligaer.

Geophysical survey – that can revealburied features without disturbing theground – and trial excavations wereundertaken in 2003–4 immediately tothe north-east of the earlier fort(Gelligaer 1), but the results werelargely negative. However, someevidence of Roman occupation wasuncovered and may be associated withextra-mural activity connected withthe stone fort (Gelligaer 2).

Reports on all the early excavationshave been published in TheTransactions of the Cardiff Naturalists’Society. The finds are in the care ofAmgueddfa Cymru – National MuseumWales and some objects are displayedin the Winding House, Caerphilly.

The excavations on the site of the stone fort in progress, 1900.

When the Romans invaded Britain inA.D. 43 it is possible that their plansfor conquest did not extend muchbeyond the south-east of England.However, within five years the Roman army was fighting in what is now Wales.

At this time, south-east Wales wasinhabited by a powerful and warliketribe, known as the Silures. Theirhostility towards the Romans increasedwhen the defeated British leaderCaratacus sought refuge among them.Under his leadership they carried out aseries of raids into the new province,which opened a chapter of some thirtyyears’ campaigning in the west ofBritain. The fighting was often bitter,even after the capture of Caratacus inA.D. 51. In the following year theRoman army suffered its greatestdefeat in Britain, losing a large part ofa legion in a battle with the Silures.Despite this setback, the Romans weresoon to gain the upper hand, and bythe mid-50s they had pushed thefrontier forward to the River Usk and inplaces beyond. Control of the areadevolved upon a legionary fortress atUsk and a series of forts creating afrontier stretching from Cardiff to thelower Usk Valley and north into themiddle Wye. From bases such as these,the Roman army could mountcampaigns into Wales. They did notpenetrate the heavily-wooded rivervalleys, but followed routes over themore lightly-wooded uplands. Duringthese advances the battle-groupswould have been accommodated intemporary (marching) camps, eachlarge enough to contain the wholeforce. The defences of these campswere slight, consisting of a low banksurmounted by a palisade and frontedby a small ditch. Several of thesecamps have been found in theGlamorgan uplands.

Roman Wales: Part of the Empire

3

Roman Wales: the military networksoon after the conquest in A.D. 74-78.

By A.D. 60 the Roman army was closeto victory, after the governor SuetoniusPaulinus had carried out successfulcampaigns in both south and northWales. But a devastating revolt led byBoudicca, queen of the Iceni in Norfolk,brought the conquest to a temporaryhalt. The work required in restoring theprovince prevented further campaignsfor over a decade. Although plans forthe occupation of Wales may havebeen developed over several years, itwas only on the arrival of a newgovernor, Julius Frontinus, in A.D. 74that large-scale military operationsbegan. The Second Augustan Legionmoved to a new fortress built atCaerleon (Isca) in A.D. 74/75, fromwhich the Silures were subdued.Frontinus was responsible forovercoming much of the remainder ofWales, but it was left to his successorAgricola to complete the task. Duringthese combative years, many men willhave been killed in battle, others in theaftermath. Women, children and elderswill also have suffered from thedamage that armies inflict before andafter battle.

Following the conquest a substantialnumber of soldiers, perhaps as many as25,000, were stationed across Wales. Anetwork of roads and regularly-spacedforts stopped native forces fromgathering: a policy of divide and rule.The forts were usually sited uponestuaries for ease of supply by sea, orlay at strategic valley-junctions in theinterior or on lonely upland roads tomaintain the passage of supplies. Theywere garrisoned by auxiliary troopsunder the command of one of twolegions, the Second Augustan based atCaerleon (Isca) and the Second Adiutrix– replaced by the Twentieth in aboutA.D. 87 – at Chester (Deva).

This intensive military occupation ofWales lasted for nearly fifty years ortwo generations. However, the cost ofmaintaining such a large garrison wasenormous. There were also difficulties

in supplying the army in some of themore remote forts and, furthermore,manpower was limited. The wholehistory of the military occupation ofWales is one, therefore, of reduction inthe garrisons wherever possible – eitherby abandoning forts completely or, ifthis were not possible, by introducingsmaller units in place of larger ones.The latter indeed happened at Gelligaer,where the smaller stone fort replacedthe large earth and timber fort. By A.D.130 a substantial number of units,including that based at Gelligaer, hadbeen withdrawn from south Wales,many for service on Hadrian’s Wall.

By the late second century, there hadbeen a radical reduction in the size ofthe garrison in Wales. In Glamorgan itwould appear that none of the forts,apart probably from that at Cardiff,remained in commission, with theSilures having over time adoptedRomanised ways and more peacefulconditions prevailing. It had previouslybeen thought that soldiers continued tooccupy the fort at Gelligaer until verylate in the second century. However, areview of the pottery evidence –indicating a noticeable fall off afterabout A.D. 120/30 – casts considerabledoubt on this, though alterations toseveral buildings, perhaps indicatingprolonged occupation, remain to beexplained. By the third century it waspossible to police the whole of Walesfrom half a dozen forts in addition tothe legionary bases, probably by usingmobile rather than stationary forces.

For the greater part of the Romanperiod, the territory of the Silures wasessentially a civilian zone. Followingtheir defeat, the Silures would havebeen treated as a surrendered peoplewith no rights, to be dealt with as theRomans pleased. However, as we haveseen, the Roman administration hadgood reason for showing moderation.The more successful the attempt tomollify and encourage the localpopulation to adopt Romanised ways,

the quicker their military commitmentcould be reduced. It is against thisbackground that the Silures weregranted, under close Romansupervision, a form of self-government,probably in the earlier part of thesecond century. Caerwent (VentaSilurum) served as the capital of theCivitas Silurum, and it is here wherethe tribal council would meet, lawcourts could be set up and otheradministrative duties carried out. Thestreets of the town were laid out ingrid-fashion on either side of the mainRoman road to Caerleon. In the centrewas the forum-basilica (market-placeand civic hall). No doubt Caerwent wasa busy market town, being a favouritecentre for troops on leave; the facilitiesprovided included public baths, templesand a variety of shops. The basis of itseconomy was agriculture – there wereseveral farms within the town – pettymanufacturing and the retail trade.Other civilian settlements grew upoutside the legionary fortress atCaerleon and at some of the auxiliaryforts. At Cowbridge, on the roadbetween Cardiff and Carmarthen, a‘small town’ developed, probably as amarket centre for farms in the Vale ofGlamorgan. Farming was extensive onthe southern coastal fringe, and bothcereal cultivation and grazing wereimportant parts of the economy. In thecountryside there was a diversity ofsettlement types, ranging fromsubstantial villas and Romanisedfarmsteads, such as Ely, Llantwit Major,Llandough, and Whitton Crossroads, tothose of more humble character withfew Roman trappings. The GwentLevels, which provided excellentgrazing for cattle, sheep and horses,were also utilised extensively. Theuplands of Silurian territory may havebeen used seasonally for grazinglivestock – transhumance – butevidence for any occupation is slight.The existence of auxiliary forts in theuplands, such as Gelligaer, indicates thepresence of a native population. Also,

4

at Gelligaer there is slight evidence fora civil settlement near the site of theearly fort.

To the west of the Silures lay theDemetae, the only other tribe in Walesto be granted self-government, withCarmarthen (Moridunum) probablyserving as its capital. Within their tribalarea, there appear to have been fewercountry villas than in the south-east,and some of the old Iron Age hillfortswere still inhabited. Throughout therest of Wales, under more or lesspermanent military surveillance, therewere no further towns. A few smallsettlements grew up around theimportant copper- and lead-miningdistricts, for example at Ffrith andPentre in the north-east. There werealso civil settlements outside several of

the forts, such as Caersws andSegontium. The native populationgenerally appears to have lived eitherin their old hill-top settlements, or insmall isolated hut-groups, typical ofthe north-west. Some of these nativesettlements produce a fair number ofRoman coins, pottery and otherartefacts, demonstrating that theydeveloped, to some extent at least, amoney economy trading with the army,merchants and each other.

By the late third century, there was agrowing danger, especially in thecoastal areas, from sea-borne raiders.The withdrawal of the SecondAugustan Legion from Caerleon,probably towards the end of the thirdcentury, created the need to reorganisethe defences of south Wales to combat

this threat from the sea. A new fortwas built at Cardiff, possibly the basefor a fleet patrolling the BristolChannel. Its design was a departurefrom that of the earlier forts, thedefences being much stronger, with theprovision of projecting towers for theuse of archers. There were alsodefensive roles for pre-existing forts atLoughor and Neath. An army garrisonmay also have been maintained at theRoman town of Caerwent in the fourthcentury. At Gelligaer itself, there isslight evidence in the form of latethird- and fourth-century pottery forlate activity at the fort. The growinginsecurity in the countryside appears tohave resulted in the gradual runningdown of farms along the southerncoastal belt. Doubtless, some farmersand their families sought the relativesafety of Caerwent, while others appearto have taken refuge behind thedefences of several Iron Age hillforts.

During the final years of the fourthcentury the Roman government wasfaced with a series of disasters on theContinent. Troops serving in Britainwere needed urgently elsewhere. Theeconomic consequences of theresulting chaotic conditions werespectacular. Roman coinage stoppedreaching Britain and the infrastructurethat provided access to Romanisedgoods and services collapsed. TheRoman authorities had lost control ofBritain and by 410 it was clear thatthat the Emperor could not providefurther help. The end of Roman rule inBritain had come.

In the aftermath in Wales, small Britishkingdoms emerged to fill the powervacuum. Each competed to controlpeople, land and resources. Peopleexperienced great changes in society,politics and religion. Regional loyaltiesand power dominated all aspects oftheir lives. These changes shaped aWelsh cultural identity different fromthat of the rest of the British Isles.

5

An auxiliary soldier of about A.D. 100 (Ermine Street Guard).

The Roman Army

6

The troops stationed in Wales weredrawn from the two main branches ofthe Roman army, the legions and theauxilia. The legionaries, recruited fromRoman citizens, were the crack troopsof the army - the heavy infantry, sentout for campaigns of conquest or tosuppress revolt within the Empire. Tosupplement this force, auxiliary unitsmade up of non-citizens were raisedfrom the provinces and newly-conquered areas of the Empire. The roleof the auxiliary units was to supportthe legions in the battle-line, but formuch of the time they were taskedwith defending and policing thefrontier zones.

There were three types of auxiliary unit– light infantry (a cohors or cohort),mixed cavalry and infantry (a cohorsequitata) and front-line cavalry (anala). They were organised into units ofnominally 500 (quingenaria) or 1000(milliaria). The infantry units weredivided up into centuries, 80 strongunder the command of a centurion; thecavalry units were divided into troops(turmae) of about 30–40 men,commanded by a decurion. Each unitwas commanded by a Roman officer,either a prefect (praefectus) or atribune who answered to the legionarycommanders. The stone fort atGelligaer is thought to haveaccommodated an infantry unit of‘500’ men, whilst the early fort is of asize that may have been garrisoned bya cavalry or larger infantry unit.

The annual pay of an auxiliary soldierwas probably only one-third that of alegionary. Their reward, after serving inthe army for 25 years, was to receive,on retirement, a grant of Romancitizenship which also extended totheir children. An auxiliary would havespent much of his long service far fromhis homeland. Once retired he might

return ‘home’ or remain in the areawhere he had been posted, havingpossibly developed family links througha wife from the locality.

New recruits to the auxilia would haveto pass a medical test before beingassigned to a unit, where theyunderwent basic training. They had tobe fit, so they ran, jumped, learnt toswim, as well as being taught tomarch. The main training was in theuse of weapons and this involvedpractising with dummy weaponsagainst wooden posts. Much of thistraining would have taken place on thefort’s parade-ground. Auxiliaries werealso taught the techniques of ditchdigging and turf-building, tasks theywould have to perform when oncampaign. These essential skills weregained by constructing practice camps,small earthworks which reproduce theessential features – rounded cornersand devices for the protection of thegates – of a marching camp, built toprotect an army on campaign. Fivepractice camps have been identified onthe common a mile or so to the northof Gelligaer, and such siting within ashort march of a fort is common.

Some auxiliaries wore no armour, butthose who did wore either mail or scaleshirts over their tunic perhaps withtrousers of wool or leather. Mail wasmade of alternate lines of punched ironrings joined by riveted rings, whilescale-armour was made by sewingoverlapping rows of metal scales onto alinen support. For the protection of thehead a helmet of iron or bronze wasworn, similar to but less elaborate thanthe legionary types. For lightness andagility they carried an oval shield, madeof laminated wood covered in leather –often brightly decorated – with acircular metal boss. An infantrymancarried a spear (hasta) for throwing or

thrusting, as well as a sword anddagger. A cavalryman’s fightingequipment included a spear, used forthrowing as a javelin or for thrusting asa lance, and a long slashing sword(spatha).

The soldier was expected to maintainphysical fitness and keep up basicweapon skills by daily training. Besidesthis he had to carry out routine duties,which included mounting guard on theheadquarters building and at the gates,cleaning the centurion’s kit, andcleaning out the latrines. He also hadto maintain his own weapons andarmour, and gather fuel for cookingand the baths. Much of the soldier’sleisure time would have been spent inthe bath-house which was used as arecreational club as well as for bathing.

Auxiliary cavalryman (Ermine Street Guard).

The Early Fort

7

The first fort (Gelligaer 1) was probablyestablished as a result of the Romanconquest in 74–78, to police theupland area of the road between theforts at Cardiff and Brecon. All theforts built at this time wereconstructed from earth and timber. The defences consisted of a rampart of turf and clay which was derivedfrom the broad ditch outside; thegateways, turrets and breastwork ofthe rampart were all of timber, as were the buildings of the interior.

At Gelligaer, the early fort lies to thenorth-west of its stone successor and is now outlined by only a faint bank. It is of a playing-card shape, 174 mlong by 120 m wide (2.2 hectares),considerably larger than the later fort.Trenching by John Ward carried out in 1913 revealed an earthen bank,fronted by two V-shaped ditches. At the time, Ward suggested that it was a temporary camp occupiedduring the construction of the stonefort. However, it is clear from theexcavation carried out in 1963 that it was of a more permanent nature,built to accommodate a large unit. This excavation revealed internalbuildings of timber which subsequently had to be rebuilt,probably owing to the decay of theoriginal timber-work. The size of thecamp suggests that it was intended for a garrison of 1000 infantry orpossibly 500 cavalry. Only furtherexcavations will reveal the layout of its internal buildings, and so give a clue to the type of unit it housed. A re-examination of the pottery fromthe site would support its constructionas part of the Flavian conquest (mid-70s), but an earlier date cannot betotally precluded. It is possible thatthere may have been a small civilsettlement outside the camp, on thesite subsequently occupied by the stone

Site of early fort

Stone fort

Aerial view of Gelligaer from the south. © Crown copyright: RCAHMW.

fort. It has been suggested thatGelligaer 1 was decommissioned justshortly before the construction of thestone fort. The dating is, however,inconclusive and it is possible that

there was a gap in occupation betweenthe two forts. There is evidence thatthe buildings of Gelligaer 1 may havebeen destroyed by fire in the end, toclear the site.

During the first decade of the secondcentury, probably owing to the arrivalof a new garrison, a smaller fort wasrequired. Instead of rebuilding on thesite of the earth and timber fort, theychose a new site further towards theend of the spur. The building of thisnew fort in masonry coincided with thepolicy of gradually rebuilding in stoneat the other forts in Wales.

The most important finds from the siteare the fragments of three inscriptions,set up to record the completion of thefort. Two were found in 1909 in theditch fill near the south-east gate, anda further portion of the larger stonewas discovered near by in 1957. Thethird inscription was found in 1913near the north-west gate. It seemslikely that a dedication was placed ateach of the four gates; the secondinscription found at the south-eastgate could easily have come from theheadquarters building, and been lost inthe ditch when stone was beingremoved from the dismantled building.

All three inscriptions appear to besimilar and the largest of these hasbeen restored.

IM[P(eratori).CA]ES(ari).DIVI

NER(vae)[.FIL(io).N]ER(vae).TRAIANO

[AVG(usto). GE]RM(anico).DAC(ico).PONT(ifici)

[MAX(imo).TR]IB(unicia).P(otestate).P(atri).P(atriae).CO(n)S(uli)V

[IMP(eratori)III]I

[LEG(io) II AUG(usta)]

‘For the Emperor Caesar Nerva TrajanAugustus, conqueror of Germany,conqueror of Dacia, son of the deifiedNerva, high priest, with tribunician power, father of his country, five timesconsul, four times acclaimed Imperator,the Second Legion Augusta (built this).’

8

Trajanic building inscription from the south-east gate, recording the constructionof the fort some time between A.D. 103 and 111.

Plan of the stone fort, with measurements in Roman feet (John Ward).

The Stone Fort

9

Imaginative reconstruction of the stone fort soon after its completion (drawn by Paul Hughes and updated by Jackie Chadwick).

The reference to the fifth consulship ofthe Emperor Trajan, which he held inA.D.103, is generally accepted asindicating the approximate date whenthe stone fort was built. However, itcould have been constructed at anytime between 103 and 111, for Trajandid not hold his sixth consulship untilA.D. 112.

At this time, the plans of Roman fortsboth large and small were broadlystandard, and Gelligaer was noexception. The defensive enclosure wasalmost square with rounded corners,and four gateways symmetricallyplaced, one on each side. The interiorwas organised into three divisions. Thecentral zone lay on one side of themain road (via principalis) – connectingthe south-east (porta principalis dextra)and the north-west (porta principalissinistra) gates – and was occupied bythe headquarters (principia),commander’s house (praetorium), a

work-shop yard and two granaries, one at either end of the range. Behindthis, in the rear zone (retentura), therewere two barracks, a workshop andstore building, disposed on either sideof the street (via decumana) leading to the rear gate (porta decumana). On the other side of the main streetthe front zone (praetentura) wasoccupied by two pairs of barracksdivided by a road (via praetoria) leadingfrom the front gate (porta praetoria)to the entrance of the headquarters.The remaining space was occupied bystorage buildings and what may havebeen a stable for the baggage animalsof the unit.

The layout was not perfectlyrectangular, and this was undoubtedlyowing to the faulty setting out of thework at the start. To obtain right-anglesthe surveyor would have used aninstrument known as the groma, across-staff with plumb-lines, which was

notoriously inaccurate especially inhigh winds, thus causing an error. JohnWard demonstrated that the Romansurveyors probably used a standardten-foot measuring staff in laying outthe fort. He worked out all thedimensions of the fort and buildings inRoman feet (one pes = 112/3 inches or0.296 m) and they correspond tomultiples of ten or five feet.

The size and layout provide a perfectexample of accommodation for thesmaller type of infantry unit in theauxiliary forces, a cohors quingenaria.The name and origin of the garrison arelost. In the past it has been mistakenlyclaimed that one of the inscribedfragments contained the letters PAN(Pannoniorum or Hispanorum)suggesting that the garrison was ofPannonian (a province bordered by theDanube) or Spanish origin. The identityof the units based at Gelligaer 1 and 2still remains a mystery.

The Roman army was essentially amobile force trained to fight in the open,but a fort in potentially hostile territoryhad to be provided with defences incase of an assault by the enemy. AtGelligaer, the fortifications would havepresented a formidable obstacle, andconsisted of the following elements:

• an outer ditch• earthen rampart with stone

revetments• four gateways• corner and interval towers.

In general the ditch, where excavated,was of a typical Roman form, being V-shaped, and approximately 5.8 m wideand 2.1 m deep. However, on thesouth-west side the ditch resembledmore of a W in section, but was of nogreater depth or width than elsewhere.The ditch would have been an effectiveobstacle in any attempt to scale thewall. Between the inner face of theditch and the fort wall there was aberm 1.5 m wide, to maintain thestability of the rampart.

The rampart was the principal barrier,and had to be both difficult to scaleand hard to break through. Whereverthe rampart was excavated it wasfound to be of earth faced withmasonry on both sides, and in all it wasabout 6 m wide. Earth for theconstruction of the bank was dug fromthe ditch and the foundation trenchesfor the retaining walls. The top of theearthen bank would have been flat,probably with a gravel surface, andwould have been used as a patrol-trackas well as a fighting platform fromwhich the defenders could hurl missilesin the unlikely event of coming underattack. Ward calculated that in itsoriginal state the bank would havebeen about 11 or 12 ft (4 m) high.Fronting this was a stone revetment,which constituted the fort wall, varyingin thickness from 0.9–1.2 m. The front

was faced with large well dressedblocks, but the back was rough, beingbuilt against the earth bank; and thewhole wall was founded on layers oflarge rough stones. It seems likely thatthe wall would have been about 5 mhigh, and furnished with battlements toprovide protection for soldiers on topof the earthen bank. The innerretaining wall was of a thinner androugher construction than the outerwall, being only 0.9 m wide. Behind therampart there was a road (via sagularis)running all around the fort to allowease of movement and unimpededaccess to the defences.

At the corners of the fort, and betweenthem and the gates, were found theremains of twelve towers, which werebuilt flush with the front wall andextended to the rear. These wereexternally of the same depth as therampart and approximately 4.9 m wide,with walls just over 0.6 m thick. It wasnoted that the side walls of the towerswere not bonded to the fort-wall,which was a method possibly used toallow for differential sinkage of thetwo structures. The room at groundlevel, normally used as a store or thelike, was entered by a narrow doorwayin the middle of the back wall. The

The Defences

Edge of ditch

GuardChamber

GuardChamber

Steps

20m0

drain

Reconstruction and plan of the south-west gate (drawn by Tony Daly).

10

upper story of each tower stood abovethe level of the rampart walk and mayhave been roofed. They will have servedas look-out towers and possiblyfighting platforms from which arrowsor other missiles could be discharged.

There were four entrances to the fort,symmetrically placed, one on each ofthe four sides. Each gateway containedtwo arched roadways separated by anintervening wall (spina). The archeswere built from calcareous tufa –possibly from the Merthyr area – afavourite material for this kind ofconstruction owing to its lightnesscombined with strength; the remainderof the gates were constructed inmortared rubble faced with smallblocks of the local pennant sandstone,as used in the fort wall. The outerarches of the gate were slightly setback from the rampart face, and

flanked by guard-chambers to controltraffic in and out of the fort. The frontof each passage would have beenprovided with a stout wooden door of two leaves, which swung uponpivots and were bolted when closed.These doors, in closing, stopped againsta stone sill which crossed thethreshold, and in opening they swungback into recesses in the side of eachpassage. In one of the passages of thesouth-west gate, the sill, pivot and bolt holes were all found intact, andthe sill showed signs of being worn by the wheels of carts. The flankingguard-chambers were oblong, 3.4 m by 2.9 m, the front being flush with the fort-wall, and were entered by anarrow doorway in their rear wall. Their walls would have been carried up above the height of the rampart to form towers, and they wereconnected by the rampart-walk

which was carried across the archesspanning the portals. The gate-towersprobably had gabled roofs. Access tothe upper storey and the rampart-walkwas provided by steps at the side ofeach guard-chamber. There is evidenceto suggest that one of the passages ofthe south-west gate was blocked atsome time in the later history of thefort. This may be a key indicator forsome form of occupation on the site of the fort in the late third or fourthcenturies A.D.

Immediately in front of the south-westand north-east gates the sides of theditch were stepped, presumably toreceive the timber supports of a bridge,so that a road could cross and enterthe fort. Trenches in front of the south-east and north-west gates revealed theopposite arrangement, whereby theditch was interrupted for the roadway.

The north-west gate during excavation, showing one of the guard-chambers and the wall (spina) dividing the two roadways.

11

The headquarters building (principia)was the administrative and religiousfocus of the fort. It stood in the centreof the fort (building 7), at the junctionof the two main streets, the viaprincipalis and via praetoria.Rectangular in plan it measured 24.4 mby 21 m and was divided into threemain sections: an open courtyardsurrounded by a portico, an inner highroofed cross-hall, and a row of fivesmall rooms at the back. This basic planappears on a much larger scale in thegreat legionary fortresses.

As might be expected, the headquarterswas an imposing building, entered fromthe main road (via principalis) probablyby an arched doorway in the centre ofthe front wall. This gave access to an open courtyard with a gravelledsurface, which was surrounded by aportico on three sides. The excavationsrevealed a number of stone baseswhere columns probably stood tosupport the tiled lean-to roof of this

covered walk. This part of the principiaprobably served as a meeting place andthe orders for the day and othernotices may have been posted on thewalls at the back of the portico. Acircular, stone-lined well was providedin the corner of the courtyard to ensurea supply of water in an emergency. Thewell was filled with quantities ofbroken pottery, tile and stone when thebuilding was abandoned.

Beyond the courtyard lay the roofedcross-hall, extending the full width ofthe building. It is here that thecommanding officer would have beenable to address his officers and men,issue orders and perform military andreligious ceremonies. The entrance tothis hall from the outer court wasopen, the roof being supported by arow of large stone columns. In 1909when the ditch by the south-east gatewas being excavated, two columnbases and a capital, probably from thisbuilding, were discovered.

At the back of the building there werefive rooms, each entered from thecross-hall. The central room, somewhatlarger than the others, had a specialsignificance as the regimental shrine(aedes). Its importance was emphasisedby its position opposite the entrance tothe principia, which made it the mainfocus of the building. It is here that theemperor’s statue would have stoodbetween the standards of the unit. Itwas also usual for the pay-chest andthe troops’ savings to be locked safelyin the shrine. The foundations of thisroom were thicker than the others,suggesting that it was probably higher,standing above the roof-line withwindows in the upper sidewalls lightingthe interior. On either side were tworooms, normally interpreted as officesfor the administrative staff of the unit,two for the senior clerk (cornicularius)and his staff and two for the standardbearers, who also acted as pay clerks. A huge amount of ‘paperwork’ wasinvolved in the administration of anauxiliary fort and its garrison. With theexception of the shrine these rearrooms were heated, either from anopen hearth or a brazier.

The Headquarters Building

Headquarters

Aedes

Assembly - Hall

Courtyard

Well

20m0

12

The headquarters building, lookingfrom the rear rooms across theassembly hall and courtyard.

To one side of the principia, also in thecentral range of buildings, was thecommander’s official residence(Praetorium – building 6), withamenities befitting his status. Thecommander would probably have beena young member of the upper class(equites) and this post, a short-term

commission, would have been an earlystep in his career. He was unlikely tohave seen military service before thisposting and he would have had to relyheavily on the advice of the centurions.The house would have providedaccommodation for the officer, hisfamily, guests, and domestic servants.

At Gelligaer, this building was fairlysmall, 22.6 m by 19.5 m, comparedwith those uncovered at other Welshforts. The rooms were arranged arounda small courtyard or garden with averanda on all four sides, a practice ofthe Mediterranean world to give shade;here it provided shelter. Whenexcavated only the foundationsremained, so the uses to which thedifferent parts of the building were putare uncertain. From debris found on thesite, it is known that the building had atiled roof and glazed windows.

In front of the building there was aportico with a gravelled surface,bordering the main road (viaprincipalis), from which the house wasentered. The range of rooms behind theportico seems to have comprised theentrance-vestibule and main receptionroom. A drain leading from the north-west range may indicate that this wasthe service quarters including thekitchen; the accommodation for theslaves would also be here. The rest ofthe house would have been used as theliving quarters, including the diningroom, bedrooms and probably an office.

The Commander’s House

Courtyard

0 20m

Drain

The commander’s house duringexcavation in 1901, with thesmall courtyard in the centre.

13

The GranariesThe central range of buildings wascompleted at each end by a granary(horreum), placed next to the south-east and north-west gates for the easeof access when unloading deliveries ofgrain. It was the normal Roman armypractice to keep a year’s supply of grain– an important element in the soldier’sdiet – in each fort; the two granaries atGelligaer would have held what wasnecessary. For the proper storage ofgrain and probably other perishables –such as vegetables and meat – thegranaries had to be kept dry and free ofrodents and birds.

Unlike most of the buildings within thefort, the granaries were completelyexcavated. They were found to have the

distinctive ground-plan of massivebuttressed walls, founded on stonerafts. The floors of the granaries wereraised off the ground on a series ofcross-walls. The timber floor of thenorth-west granary (5) was supportedby six cross-walls; that of the south-east (8) by eight, though at some stagein its history this was reduced to five.Air vents at the bottom of the externalside walls and a central break in thecross-walls, produced a cool through-draught which helped to maintain thebest conditions for grain storage. It wasalso noted that large quantities of redroofing tiles were recovered from bothgranaries. The weight of the heavy tiledroofs would have been carried by the

buttresses, which would also haveallowed them to be carried out wellbeyond the side walls, so that rainwater would not trickle down the wallsand spoil the grain inside. The rainwater was subsequently carried awayfrom the vicinity by a series of drains.To achieve a good circulation of airwithin the granary, there woulddoubtless have been louvred vents nearthe top of the side-walls between thebuttresses; the wooden slats preventedsmall animals and birds from entering.At each end of both granaries therewas a roofed loading-platform ontowhich sacks of grain could be unloadeddirectly from carts without the need tocarry it up steps or a ramp.

0 20m

14

Above: Reconstruction and plan of the south-east granary (drawing Paul Hughes and updated by Tony Daly).

Below: The north-west granary during excavation.

Six barrack-blocks (centuriae) – four inthe front part of the fort (12–15) andtwo in the rear (1 and 2) – providedliving accommodation for the men.These six barracks complete therequirements for the housing of the sixcenturies of eighty men which formedthe nominal complement of a cohorsquingenaria. They were long, L-shapedbuildings, 44.2 m long by 9.1 m wide atthe narrow end. From the remainsfound it appears likely that they werelargely constructed from timber withonly the foundations and lowestcourses of the walls being of stone. Anabsence of any roofing tiles probablyindicates that the roofs were either ofwooden shingles or thatch. A row ofholes each packed with stones, in frontof the narrow section of the barrack-block, would have held the woodenposts which in turn supported the roofof a veranda.

Each barrack housed a century of about80 men with its centurion. All thebarracks were divided into two unequalparts: the broader end section providedfairly comfortable accommodation forthe centurion, while the longer, narrowportion was devoted to the men. No

partitions which would have dividedthe men’s quarters were found.However, this section was nearly ahundred Roman feet in length, so it ispossible that it was divided into thestandard ten barrack-rooms; eachhousing eight men (a contubernium).Each barrack-room would have beendivided into a front room, probablyused for the storage of arms andequipment, and a larger, inner roomused for sleeping. All the front roomswere entered from the veranda. Thestone partitions recorded within thebroader end of the barrack-blocks maymark off the quarters allotted to thethree junior officers in the century,namely the second in command (optio),standard bearer (signifer) and clerk(tesserarius).

The Barracks

Little is known about the remainingbuildings except for their plans.Opposite the commander’s house therewas a long building (9), divided downthe centre, 44.2 m long by 12.5 mwide. At its south-eastern end therewas a stone-built tank, with a lead-outlet pipe which entered a stone-lineddrain. This building has beeninterpreted as a stable, which wouldhave been needed for the baggageanimals belonging to the unit as wellas the commander’s and centurions’horses. The remaining area of thepraetentura was taken up by two small

rectangular buildings (10 and 11),which were probably used as stores. Inthe rear of the fort, a narrow buildingwith a wing at either end (4) probablyserved as the unit’s workshop (fabrica).It is here that the weapons, tools andequipment needed by the unit could berepaired or perhaps even made. Theworkshop was divided into four rooms,and in one of these there were twosmall, rectangular, stone structureswhich could have been water tanks.Evidence for some industrial activity –areas of burnt clay and fragments ofwalling possibly belonging to a series

of furnaces – was found in the walledyard fronting the fabrica. Thisworkshop-yard, next to theheadquarters building, was enteredfrom the via principalis. The entrancewas flanked by a small chamber, whichmay have been the fort’s latrine.Adjacent to the workshop, on the otherside of the via decumana, there wasanother rectangular building (3) dividedinto approximately three equalsections. It has been suggested thatthis might have been a secure store forweapons and other items of militaryequipment.

Stable, Workshop and Stores

Men’s quarters Centurion’squarters

0 40m

Veranda

Plan of a barrack-block.

Glass gaming counters. The soldiers wouldhave played board games when off-duty.

15

The Bath-house

16

Adjoining the fort on the south-eastwas a walled annexe, measuringapproximately 119 m by 65 m. As seenat other Roman forts in Britain, suchan enclosure could contain buildingsproviding extra storage, facilities suchas workshops and the bath-house, aswell as affording protection for animalsand wagons with goods in transit. Thedefences of the Gelligaer annexe werenot as strong as those of the fort,consisting of a stone wall 0.9 m thickand backed with earth, beyond whichwas a V-shaped ditch 4.6 m wide. Thisenclosure was entered by a single

gateway on the south-east side. Asmall building (C) 3.7 m wide by 2.8 mjust within the north-east angle mayhave acted as a guard-chamber. Atrench cut by John Ward outside theeast corner of the fort revealed thatthe annexe-wall had been carried acrossthe fort ditch, after a considerable amountof silt and rubble had been allowed toaccumulate. It has been suggested,therefore, that the annexe-defenceswere constructed at a much later datethan those of the fort.

The road running from the annexe gateto the south-east gate of the fort

divided the enclosure in two, thenorthern half being occupied by thebath-house (A) and its exercise yard(B). To the south of the road, the mostnotable structure was a largerectangular yard (D) 27.5 m by 13.4 mwith a small building (E) attached to itssouth-east wall. An interesting featureof the yard was the large quantity ofiron slag, coal and charcoal scatteredover the site, presumably indicating itwas used in some industrial activity.The only other structure was a smallrectangular building (F) 11 m by 3.7 mthe purpose of which is unknown.

The Annexe

It was normal for the bath-house(balneum) to be sited outside the fort,for safety reasons (they were a firerisk), and also because of the shortageof space within the fort. The bath-housewas an essential feature in Romaneveryday life, providing facilities for asoldier to bathe and cleanse his body,exercise and recreation. It was a placeto meet friends and relax – gossip,gamble, play games, have a snack anddrink – when off duty. At Gelligaer the bath-house, measuring 33.5 m by19.8 m, comprised the usual suite ofhot, warm and cold rooms, a circularsweating chamber and exercise yard.The hot and warm rooms were providedwith under-floor heating (hypocausts).The building had a tiled roof, tilefragments being found in abundanceall over the site. Some of the roomswere also decorated with painted wall-plaster, and the windows were glazed.

The Roman method of bathing wastotally different from ours. The batherundressed in a changing room(apodyterium), which at Gelligaer alsoacted as the cold room (frigidarium).

He then proceeded into a room whichwas kept at a warm, humidtemperature (tepidarium). Having raisedthe body temperature the batherentered the last room where a veryhigh temperature was maintained(caldarium). The soldier could also havea hot bath in this room. Once the poreshad been opened the body would beanointed with oil which, after a massageperhaps, was scraped off along withthe sweat and dirt using a curved

scraper (strigil). After this he made hisway back to the cold room where heimmersed himself in the cold plungebath, to close the pores of the skin toprevent a chill. As an alternative to allthis, the bather might prefer to go into

Excavation of the bath-house in 1909.William Clarke of Llandaff (Cardiff) - who wasentrusted with the excavations and providedthe workforce - and the Rector of Gelligaer,Rev T J Jones, on the site of the circularsweating chamber.

the sweating room (laconicum) whichwas heated to a high temperaturewithout any water or steam.

Roman bath-houses invariably showedsigns of constant alteration andimprovement, the baths at Gelligaerbeing no exception. The first phase ofthe building presumably coincided withthe construction of the fort, but it wasnever completed, its plan being modified.When the second stage (phase 2) ofthe building was finished, the bathersentered a combined changing room andcold room (G) with a flagged floor. Asunken bath (Q) lay in the south-eastcorner. Lined with waterproof plaster, ithad three quadrant-shaped stepsleading down into the cold water. Thecentral feature of the cold room was a

shower bath or douche, which somebathers preferred to the plunge bath; itwould have taken the form of acircular, ornamentally perforatedflagstone placed over the main drain.Beyond the frigidarium was the warmroom (D), and then the hot room (C)with its hot bath. On the east side ofeach of these rooms was a furnace; theheat from their fires circulated under araised floor and was then ducted upthe walls through a series of pipes, hotair being discharged into the open fromjust below the roof level. The provisionof two furnaces was unusual (normallythe heat was provided by one boiler-housesituated at the end of the hot room)and was undoubtedly owing to achange of plan.

At some date, the floors of the warmand hot rooms and the two furnaceswere dismantled and then rebuilt at a higher level. This reconstruction(phase 3) was probably necessarybecause water was collecting under theraised floors, since the builders appearto have misjudged the water-level.

The next development (phase 4) was ageneral enlarging and completereorganisation of the building. The two

17

Top: Baths: view of the warm room (D) fromthe hot room (C): note the rows of tile pillarsto support the raised floor.

Above left: General view of the bath-house, as seen from the church tower.

Above right: Baths: the cold plunge-bath (Q) in the frigidarium.

18

Bath - House2 - 3 Drain

Drain

Drain

Drain

4

5

CD

Q

G

CC

B C DA

EG

F

T

B C DA

EG

F

J

H

L

0 40m

Phase 4

furnaces were dismantled and theopenings were blocked up. They werereplaced by a new boiler-house (A),which showed the effects of fierceheat, built at the north end of the hotroom. Arched openings were insertedbelow the floor level between the hotand warm rooms to allow the heat tocirculate. Between the new boiler-house and the hot room a new plungebath was built (B), the hot water beingsupplied from a leaden, or perhaps asheet-iron tank which was supportedabove the furnace. Another addition, inthe corner of the cold room, was theconstruction of the laconicum, acircular, hot, dry room (F) with its ownfurnace. The cold room was alsoequipped with a larger plunge bath (T),which discharged its water through alead outlet into a stone-lined drain.Also at this time, a stone-lined culvert,with a flagged bottom and stonecovers, was built around three sides ofthe baths to help drainage. This drainpassed under the annexe-gate, and

discharged through an arched openinginto the ditch.

A final reorganisation (phase 5)included a modification of the coldroom, and an even larger plunge bath(H). A heated corridor (E) was insertedto provide a warm passage between thecold room and the hot, circular room.At this time a new latrine (L) was builton the south side over the main drain.

To the west of the bath-house projecteda courtyard, which was roughly pavedhere and there, and formed theexercise yard (palaestra). In the centreof the yard there were two horse-shoeshaped ovens probably used for cooking.

The latest coin from the bath-housewas of the Emperor Hadrian (A.D.117–38), and from this Ward came tothe conclusion that the baths had afairly short life. However, the manystructural alterations would indicateotherwise, and that the baths wereused over a much longer period of

time. As it stands, the evidence isinconclusive, for most of the potteryfrom the bath-house dates no laterthan early/mid second century, withjust a handful of later sherds, some oflate third- to fourth-century date,suggesting a subsequent phase of use.

Bath - House2 - 3 Drain

Drain

Drain

Drain

4

5

CD

Q

G

CC

B C DA

EG

F

T

B C DA

EG

F

J

H

L

0 40m

Phase 5

Bath - House2 - 3 Drain

Drain

Drain

Drain

4

5

CD

Q

G

CC

B C DA

EG

F

T

B C DA

EG

F

J

H

L

0 40m

Phase 2-3

Above left: John Ward (1856–1922). Hedirected the excavation of the bath-house in1909, when Curator of the Municipal Museumand Art Gallery, Cardiff.

Above right: Baths – the circular sweatingchamber or laconicum (F), with the remains of its furnace in the foreground.

Below: Development of the bath-house.

Another military feature was theparade-ground attached to the north-east side of the fort. This was simply alevelled area, some 116 m long fromnorth-west to south-east by 49 m widebetween the fort ditch and the crest ofthe falling ground to the north-east.Excavation in 1913 revealed a uniformlygravelled surface, bounded on thesouth-east side by a V-shaped ditch,3.7 m wide and 1.5 m deep. It wasrequired not only for parades, but alsofor drill, weapon training and religiousceremonies performed by the unit onimportant dates in the military calendar.

Parade-ground

Auxiliary soldiers being taught to march in step (Ermine Street Guard).

19

This was discovered in 1913 justoutside the annexe during the diggingof a grave in the churchyard. As found,the kiln consisted of a small, square,masonry structure sunk into theground. It had a brick floor, which was supported on a centrally vaultedmain flue which had six cross-flues on either side. At the mouth of themain flue, which was on the westernside of the kiln, there would have been a stoke-hole, but it was notpossible to dig this, for there wasanother grave in the way.

When the kiln was in use, air-driedtiles would have been stacked on the kiln floor, and then the whole structure would have been coveredwith clay or turves. A fire was then lit in the stoke-hole, and hot air fromthis passed into the main flue and the smaller cross-flues and eventuallyupwards through holes in the floor to fire the load.

Many fragments of tile (tegulae andimbrices) were found near the kiln, anumber of them over-fired and

distorted or broken. This kiln wouldhave provided the many thousands of tiles used in the fort and itsenvirons. Fragments of waste mortaria– large bowls with a rough surfaceused for grinding and mixing foods –were also found near the tile-kiln. This pottery may have been fired in the tile-kiln; alternatively anotherpottery / mortarium kiln may haveexisted near by, perhaps even sharing a stoking area, the pottery waste beingdumped around the tile-kiln once itwent out of use.

Tile-kiln

B

B

AA

A - A

B - B5m0

Above: Excavation of the tile-kiln nearly completed. Below: Tile and pottery kiln.

The chronology of the fort has alwaysbeen a matter of much debate. In1903, after the main work on the sitehad finished, John Ward held theopinion that the fort was establishedduring, or soon after, the time ofFrontinus in A.D. 74–78 as one of anetwork to hold the newly-conqueredregion, and was abandoned in aboutA.D.100. This belief was based on therecovery of late first-century potteryand seven coins, five of which weredateable to the period 69–98. Therewas also at this time an apparentabsence of rebuilding within the fort,which pointed towards a shortoccupation. However, the subsequentdiscovery of the Trajanic buildinginscriptions (in 1909 and 1913) andfurther coins – the latest being that ofthe Emperor Hadrian – led to a revisionof the fort’s history. It was now

thought that the fort was built in theperiod 105–10 and that the garrisondid not depart until 130.

This view was held until the 1950s,when the late Dr Grace Simpson re-examined the evidence. This led her tosuggest a completely different historyfor the site. The construction of thestone fort in the very early secondcentury is not disputed, but it is in factthe only fixed point in the history ofGelligaer. The first-century pottery andcoins which led Ward to presume anearly foundation for the fort can nowbe associated with the occupation ofthe earth and timber fort. Dr Simpsondrew attention to some pottery whichis undoubtedly later than 130 and totraces of alterations in several of thebuildings – including the south-westgranary and bath-house – which are

indicative of a prolonged period ofoccupation. As a result, she proposedthat the fort was occupied until theend of the second century A.D., andthat alterations to the granary andbath-house may have belonged to areoccupation in the late third or earlierhalf of the fourth century A.D. It shouldbe added that other alterations werenoted to the south-west gateway whereone of the passageways was blockedand the north corner tower where aroughly built wall was inserted, butsome caution is required for not allthese works need be Roman in date.

Recent research by Dr Peter Websterhas shown that the vast majority of thepottery found at Gelligaer dates nolater than the early/mid second centuryA.D., pointing to the withdrawal of thegarrison by A.D. 130. The studyconfirmed the presence of a smallamount of later pottery, includingseveral pieces of late third- to midfourth-century date from both the fortand bath-house, which suggests areoccupation of the site some timewithin that period. However, theevidence is so scanty that it is notpossible to determine the nature orextent of this activity. No doubt thestory of Roman Gelligaer will continueto unravel and change.

Later History of the Stone Fort

20

Roman law strictly forbade the burialof the dead within an inhabited area.During the first and second centuriescremation was in fashion, the ashes

being placed in a container (a potteryurn, glass jar, or lead canister) beforeburial. Roman cinerary urns containingburnt bones were found in 1910 to the

south of the fort. This would seem toindicate the presence of a cemeterybeside the road leading south-eastfrom the fort.

The Cemetery

View of the site of the stone fort and annexetoday. The answers to many questions aboutGelligaer’s Roman past lie beneath these fields.

© Gelligaer & Pen-y-Bryn Partnership andGelligaer Community Council.

www.romangelligaer.org

FURTHER READING

Gelligaer

G C Boon, ‘Gelligaer’ in V E Nash-Williams,The Roman Frontier in Wales (second editionrevised by M G Jarrett), University of WalesPress, Cardiff, 1969 (third edition by BBurnham and J L Davies, 2010)

C H James, ‘The excavations of Gelligaer camp’,Transactions of the Cardiff Naturalists’ Society31 (1899), 80–4

M G Jarrett, ‘Excavations at Gelligaer,Glamorgan, 1963’, Morgannwg 8 (1964), 66–9

An inventory of the Ancient Monuments ofGlamorgan Vol. I, Part 2: The Iron Age and theRoman Occupation, Royal Commission onAncient and Historical Monuments in Wales,HMSO, Cardiff, 1976

G Simpson, ‘Gelligaer’ in ‘Caerleon and the Roman Forts in Wales in the secondcentury, Part 2’, Archaeologia Cambrensis 112 (1963), 49–66

J Ward, The Roman Fort of Gellygaer,Glamorgan, London, 1903

J Ward, ‘The Roman fort of Gellygaer. Thebaths’, Transactions of the Cardiff Naturalists’Society 42 (1909), 25–69

J Ward, ‘The Roman fort of Gellygaer, The annexe’, Transactions of the CardiffNaturalists’ Society 44 (1911), 65–91

J Ward, ‘The Roman fort of Gellygaer.Discoveries made in 1913’, Transactions of theCardiff Naturalists’ Society 46 (1913), 1–20

P V Webster, 2005, ‘Roman Pottery productionat Gelligaer’, Journal of Roman Pottery Studies12 (2005), 192–200

P V Webster 2005, ‘Pottery from the firstRoman fort at Gelligaer’, Studia Celtica 39(2005), 9–15

General

C J Arnold and J L Davies, Roman & EarlyMedieval Wales, Sutton Publishing, Stroud, 2000

R J Brewer, Caerleon and the Roman Army,National Museum of Wales, Cardiff, 2000

R J Brewer, Caerwent Roman Town, Cadw (4th edition), Cardiff, 2010

P Connolly, Roman Army, MacdonaldEducation, London, 1975

J L Davies and R H Jones, Roman Camps inWales and the Marches, University of WalesPress, Cardiff, 2006

A Goldsworthy, The Complete Roman Army,Thames & Hudson, London, 2003

R Jackson and R Hobbs, Roman Britain, British Museum Press, London, 2010

J K Knight, Caerleon Roman Fortress, Cadw (3rd edition), Cardiff, 2003

W H Manning, Roman Wales. A Pocket Guide,University of Wales Press, Cardiff, 2001

Roman Britain (Historical Map & Guide),Ordnance Survey (5th edition), 2001

ACKNOWLEDGEMENTS

The second edition of this booklet hasbenefited from the valuable advice andresearch of Dr Peter Webster, who hasundertaken a reappraisal of the pottery from the site. I am also grateful to mycolleagues Edward Besly, Evan Chapman and Julie Taylor for their assistance. The original reconstructions were produced as pen drawings by Mr Paul Hughes, but these have been subsequently updated using computer graphics by Tony Daly and Jackie Chadwick, Amgueddfa Cymru –National Museum Wales. All photographs(unless otherwise indicated) have beenreproduced with the permission of theNational Museum of Wales.

Designer: Maria Dispirito

PUBLISHED BY

Gelligaer & Pen-y-Bryn Partnership andGelligaer Community Council