Purchase Details US

 

Purchase Details UK

 

EARTH AND TIMBER CASTLES

 

 

 

The most difficult area in the study of Irish castles is the recognition in the field of the earthworks of the timber castles. The country is full of earthworks of all shapes and sizes and recent work in the Archaeological Survey of Ireland has shown that it is extremely difficult to date and classify them. When we look at medieval earthworks we have to remember that we are seeing only part of the defences of the castle and that we are missing all the woodwork connected with the defences. We have two types of earth and timber castles in Ireland, namely the motte and the ringwork. However, there are a number of sites that could fit into either category while others, which are obviously of medieval origin, will not fit neatly into one or the other. For instance, Sheeaunbeg Motte, Barrinagh, Co. Roscommon was recorded as a motte, as a motte and bailey using a natural esker for part of its defences, as a ringfort converted into a platform ringfort or motte, and by the Archaeological Survey as a possible inauguration site. None of the recent work on medieval earthwork castles has come to grips with the morphology of the ringwork.' The ringwork has been described as, at its simplest, being shaped like an empty flan case.' It can therefore be easily confused in Ireland and possibly Wales with the ringfort. Indeed, in Wales, many sites which have been identified as ringworks look remarkably like large ringforts. If it is difficult to recognise a ringwork in the field based on its morphology alone, then other factors can come into play which might help us identify these sites, namely their siting and distribution. Higham and Barker in their book on the timber castles set out a framework for discussing these monuments by dividing them into five groups:

 

1. Ringworks without baileys

2. Ringworks with one bailey or more

3. Mottes without baileys or with no apparent baileys

4. Mottes with one bailey or more

5. Ringworks or mottes (with or without baileys) within earlier earthworks

 

The only category mentioned above which has so far not been found in Ireland in any great numbers is the ringwork with a bailey. However, recent fieldwork has revealed an example of one just off the north slope of the Loughcrew Hills, Co. Meath, and two possible sites in Roscommon at

Sheeaunbeg and Dundonnell. Other Irish site types which sometimes prove to be of Anglo-Norman origin are the raised raths, cliff-edge forts and promontory forts, whether inland or coastal. I will deal with these in the first section on ringworks followed by a discussion on mottes.

 

RINGWORKS

 

The morphology of the ringwork or its exact definition presents something of a problem in that it is often difficult to distinguish its earthwork remains from the more common ringfort or other unclassified earthwork enclosures. In England and Wales it is, in its simplest form, an area enclosed by a fosse and rampart. It has also been defined as having a minimum height of 2m above the level of the outside defences with the enclosed area disproportionately small compared to the massive enclosing elements. However, in Ireland we can expand on these definitions by saying the bank(s) are more pronounced and the fosse is wider than one would expect to find on a ringfort. The entrance to a ringwork is also distinguishable from a ringfort in that it will often have a pronounced ramp and each side of the gap in the rampart will be faced with stone. The example at Drumsawry, Loughcrew, Co. Meath, also has extensive stonework on top of its rampart. A number of ringworks have recently been identified in north Tipperary and, in most instances, they are not circular, being either almost squared-off or irregular in shape. A ringwork, test excavated by the author at Mulphedder, Clonard, Co. Meath, was of a very irregular shape, Fortunately, one of the three small cuttings on the river side of the site produced the parts of fifteen planks which were part of a wooden palisade and appear to be part of the outer defences of the ringwork. Several of the planks had dowel holes in them for securing to upright posts and cross timbers which would have been braced from the inside.

 

Recent excavations at Trim Castle, Co. Meath," produced unequivocal evidence for a ringwork castle which was undoubtedly the fortress mentioned in the near contemporary Norman-French poem 'The Song of Dermot and the Earl'. The poem states that Hugh de Lacy fortified a house at Trim, threw a fosse around it and enclosed it with a herisson in the year 1172. There was clear evidence for a palisade trench and bracing posts at the north-west section of the ringwork. In addition, excavations uncovered large post-holes of a wooden structure inside the ringwork which had been burnt down. The excavated evidence here fits nicely with the historical reference to the burning down of Hugh de Lacy's castle in 1173 by Roderick O'Connor. Also found were the remains of grain spread over a wide area as well as horseshoes and arrowheads, just what one would expect from a fortification which housed knights and their horses. Four other castles have also clear evidence of a ringwork fortification pre-dating the stone castle. At Ferns in Co. Wexford, excavation revealed the remains of an earthwork under the south-east angle of the keep.` A small fosse outside the castle at the east side may also have belonged to the pre-stone castle fortification. Recently at Carlow Castle, which is morphologically the same as Ferns, the fosse of a ringwork was found below the foundations of the remains of the keep. ` At Adare Castle, Co. Limerick, a wide deep fosse encloses the keep providing protection for the inner ward. Until 1961 it had been postulated that this fosse was the remains of an earlier ringfort. However, it was convincingly demonstrated that metal and wooden objects recovered from the fosse c.1845 were mainly of a mid- to late-medieval date and there was no reason to assume the enclosed area was a ringfort. The subcircular shape of the site with its wide deep fosse would rather suggest that it was a ringwork castle and would therefore have a direct parallel in Trim Castle. Also the siting for a ringfort on low ground beside the river would be very unusual. Another possible example of a ringwork pre-dating a stone castle is to be found at Clonmacnoise although some archaeologists claim that the earthwork is contemporary with the stone building. However, it is in a similar siting to Adare Castle beside a river on low ground and it would be unusual to build such a massive fosse and rampart contemporaneously with the erection of the stone fortress. Recent excavations at Kilkenny Castle and Limerick Castle have also revealed a substantial ringwork underneath the walls of the stone building. At Limerick Castle the earthen bank was, in places, revetted by stone and there was a substantial external fosse. There is also the possibility of a ringwork surrounding Lea Castle, Co. Laois.

It was stated, in 1987, that there were only five sites where excavation had uncovered possible ringworks: Pollardstown, Co. Kildare; Castletobin, Co. Kilkenny; Beal Boru, Co. Clare; Clonard, Co. Meath and Ferns, Co.Wexford. Four more can be added to this list, namely Trim Castle, Kilkenny Castle, Carlow Castle and Limerick Castle. The number of excavated examples is unlikely to increase in the near future unless a research programme is undertaken. None of the partially excavated sites other than Trim, Clonard, Limerick and Carlow have given us definite evidence of defensive details such as wooden palisading and revetting so we have to rely on observation and interpretation of the remains of the earthwork. If we compare the ringwork, using recently discovered examples in Kildare, Cavan, Meath, Laois, Offaly, Roscommon and Tipperary, to a typical ringfort there are quite pronounced differences. It has been maintained that Irish ringforts could be classified as ringworks if they occurred in England, and that ringworks in Ireland may have been classified as ringforts because we lacked a classification scheme which is sufficiently discerning." Unfortunately, until recently, this has been the situation. Lurking in the archaeological records of the Archaeological Survey are many ringwork castles classified as large ringforts or merely as enclosures. For instance, at Danestown and Rodanstown, Co. Meath, multivallate ringworks adjacent to medieval churches have been classifed as ringforts. Fortunately, fieldworkers have come on the scene with an interest in medieval archaeology, who are no longer predominantly interested in the prehistoric and early historic periods.

A number of cliff-edge forts in north Tipperary appear to be exactly the same as those identified in Glamorgan as ringworks. They basically consist of single or multiple ramparts and fosses in a semi-circle enclosing a D-shaped area which is bounded on its straight side by a cliff or ravine. In Glamorgan, ringworks are often situated close to church sites as are the mottes, but in Ireland sites recently identified do not appear to have this association. There are, of course, examples of this association of ringwork and church. At Castlerahan in Co. Cavan, for instance, there appears to have been a medieval church sited inside a bailey associated with a large and impressive ringwork. This ringwork is built on the summit of a drumlin and consists of a subrectangular platform 4m high, measuring 42m by 24m enclosed by two substantial banks of earth and stone. Another cliff-edge fort, Ballyprior, was identified in Co. Laois. Its entranceway was via a

causeway and its rampart consisted of an earth and stone bank. Seven ringwork castles, three of which are 'cliff-edge forts' were identified in Co.Wexford. The one at Newtown (Ferrycarrig) was fortified by FitzStephen in 1169-70 and excavations here revealed a stone wall on top of an earthen bank.` The site is located on a spit of land which extends northwards into the estuary and is naturally defended by steep cliffs to the north and east. By 1171 the site was defended by a fosse and a bank with a palisade. The excavation showed that the fosse was 5.2m wide at the top, 1.8m at the base and 1.9m deep. Unfortunately the finds from the fosse could not be closely dated. Thirteen cliff-edge forts have been recorded in Cork and many of these are likely to be Anglo-Norman in origin.

 

In Co. Kildare, excavations were carried out on the remains of an earthen enclosure at Pollardstown which was situated on top of a natural gravel ridge. Its siting, and the recovery of medieval finds from it, indicate that this was a ringwork. The finds included an iron arrowhead which can be dated to the twelfth century and two medieval stirrups. Another earthwork site was excavated at Beal Boru, Co. Clare, which is situated on a steep-sided triangular spur of gravel on the bank of the River Shannon. Because of its position overlooking the river, its siting was believed to have been chosen to dominate or control an important crossing

point. The castle and its earlier earthwork defences at Clonmacnoise, also on the banks of the River Shannon, must have had a similar function. The Beal Boru site was circular in plan, had a very large bank and was surrounded by a broad well-marked fosse. The overall diameter of the site was 70m to 75m but since the base of the bank was up to 17m wide, only an area some 20m in diameter was enclosed. The entrance to this site, like so many ringworks, was via a causeway and the bank was faced with stone. This site was built and occupied in two distinct and separate phases, the earliest being sometime in the eleventh century based on the coin evidence. The second phase of construction was thought to be an unfinished motte; however, the site could be complete and is almost certainly a ringwork castle built on a pre-existing enclosure. Its location and siting would also point to the latter interpretation.

 

Another ringwork is to be found on the west bank of the old course of the River Shannon, at Meelick, Co. Galway. In the Annals of Clonmacnoise it is mentioned that William de Burgo founded a castle here in 1203 close to the church. It was assumed that the present ecclesiastical remains had been built on the site of the early castle and had destroyed it. However, an enclosure 100m north of the friary is very similar to the ringwork castle at Clonard and has a similar siting. The site consists of an irregular-shaped enclosure defined by a bank with the enclosed area slightly higher than the surrounding land. There is quite an amount of stone on and in the bank and there is a ramped entrance at the south. Since there are no obvious signs of a fosse around the site it must have been defended by a substantial wooden palisade.

 

In Co. Offaly three ringworks were identified compared to eleven mottes. However, mottes were a well-known feature whereas the ringworks are newly identified. One of these is a low, rectangular-shaped flat-topped mound situated just inside the earthen ramparts which enclose the monastic settlement at Churchland/Clonmore, Seirkieran. The example at Ballynacarrig is an irregular-shaped, flat-topped mound with the appearance of a truncated motte. It has the remains of a bailey-type area which is defined by a low earthen bank. The third example at Dungar is a circular platform enclosed by a water-filled fosse. It has a causewayed entrance and the possible remains of an enclosing wall around the perimeter of the platform.

 

Fourteen ringwork castles have been identified in Tipperary (North Riding). Three of these are cliff-edge forts while another is sited very close to a ravine. Most of the sites are low platforms with low earth and stone banks around the perimeter, some with stone facing and others with wall footings. Where a fosse exists it is usually flat-bottomed and access is via a causeway. One site, at Greenan, is very similar to the Meelick and Clonard examples and is situated in low-lying, wet terrain. Recent fieldwork in Roscommon has revealed at least six ringworks in addition to those identified in a 1988 paper.' All of these sites are marked on the Ordnance Survey maps but the initial interpretation and classification was not precise enough. For instance, there is an earthen enclosure close to the bank of the River Suck at Creeharmore. It is a subcircular earthen platform, slightly raised above the level of the surrounding ground. The platform has a slight lip around its perimeter possibly indicating the remains of a palisade. The platform is enclosed by a wide flat-bottomed fosse and the site is cut off on the landward side by a bank and fosse. There is a break in the bank and its terminals turn markedly inwards to give access to a small platform which lies between the outer fosse and the inner defences. This platform is undoubtedly the remains of a timber gateway controlling the entrance and drawbridge. At the river side of the ringwork there is a passageway which is delimited by an earthen bank on each side and would appear to be some type of slipway for small boats. Another of the Roscommon sites is at Ballyglass and has been classified as a platform ringfort. It is a circular platform c.2m high and c.35m in diameter and is surrounded by a wide flatbottomed fosse. Its location on flat low-lying ground suggests that this is a ringwork rather than a raised rath. The site at Rathnallog is very similar in its morphology and is located in the same type of terrain.

 

Some sites in Roscommon have initially been identified as burial mounds but their siting and morphology would indicate that they are medieval earthworks. For instance, an impressive site at Gortnasillagh was identified first as a rath and then as a barrow. It is situated on the summit of a ridge and consists of an almost circular-shaped platform with a diameter of about 24m and a maximum height of 2.7m at the west side. The platform is surrounded by a wide, flat-bottomed fosse but there is no definite evidence for a causeway entrance. Lissalway, classified as a raised rath, has a large deep fosse surrounding it and could well be a medieval earthwork castle.

 

Without citing further examples of ringwork castles, as well as possible sites, it has become quite obvious to field archaeologists that this type of site is there to be found. Another way of identifying the ringwork other than its morphology is its location, siting and association with other type sites. In England and Wales medieval earthwork castles are often associated with ecclesiastical remains; however, in Ireland this link is not as pronounced. A good example of this relationship is to be seen at Castlerahan, Co. Cavan, where a church and graveyard are situated within the bailey of a large and impressive ringwork. Another impressive site is at Rathangan, Co. Kildare, and is situated only 100m away from the site of a medieval church. It consists of a circular platform with a diameter of about 60m and rises 2m above the surrounding land. It has a causewayed entrance at the east. When considering the siting and location of earthwork castles one has to keep in mind medieval nucleated settlement in England where one would expect to find a church, a castle and houses with their plots. In Ireland we seldom find such a neat association of all three major elements despite extensive fieldwork in Westmeath and elsewhere. However, keeping the political divisions of the early Anglo-Norman settlement with its sub-infeudation in mind, we realise that we are more likely to find discrete sites and dispersed settlement within the broad area of the manorlands which can be roughly equated with the parish boundaries. The historical records are therefore very important in helping to identify medieval earthwork castles within a manor. However, there still remains the problem of identifying sites in the sub-manor which would be of a lower status to those in the chief manor.

 

Because of the poor historical records in some areas and the scattered nature of the earthworks it will remain almost impossible to distinguish medieval sites from earlier ones. It is undisputed, however, that where there is a manor there must be a medieval earthwork castle. So, for instance, at Rathangan, Co. Kildare, there is no obvious early castle to be seen until one examines the large earthwork which is classified as a ringfort by the Archaeological Survey. On close inspection it can be seen that its banks and defences have been altered to make it into a ringwork castle. There is an annalistic reference to the rath here in 801 and later references to a castle. Its position close to the site of a medieval church would be a typical model for manorial settlement in England.

 

When identifying ringwork castles it is very important to keep in mind their location and siting so that one can isolate them from the more numerous ringforts and other earlier enclosures. It is generally accepted that ringforts are located off the tops of high ground on the slopes of hills and are nearly always isolated. Ringworks seldom, if ever, are sited in the same type of location as ringforts and usually occupy the high ground on top of ridges and small hills. They occupy areas where there is often a degree of natural defence such as at the edge of cliffs or the end of inland promontories. Ringworks are also found in low-lying, wet ground but on a slight rise. For example, ringworks at Meelick and Clonard are beside rivers. The ringworks recently identified through the fieldwork of the Archaeological Survey indicate that most of the sites are located in strategic positions either controlling river crossings or passes through valleys. Other isolated examples appear to have been placed close to areas of historical and prehistoric interest such as at Loughcrew. and Knowth in Co. Meath and

Rathmore, Rathcroghan, Co. Roscommon. The location of these three sites seems to be a statement to show that new rulers dominated areas of ancient importance rather than for strategic reasons alone.

 

West of the Shannon and in Ulster there are a considerable number of raised raths which consist of a circular raised platform enclosed by a fosse. In Co. Down, at Rathmullan, an excavated site ended as a motte but had four phases of occupation which pre-dated the Anglo-Norman fortifications. Another mound site in the same county at Gransha, was also excavated. Here, however, the final phase of occupation pre-dated the Anglo-Normans. Raised raths are defined as flat-topped mounds, sometimes with a slight bank around the perimeter of the summit and with a ramped or causewayed entrance?' From the evidence of those excavated in Ulster it can be seen that their height was a result of prolonged occupation with occasional deliberate dumping of clean upcast material. However, some raised raths were formed in one operation so their height was achieved at the initial stage of building?' At Gransha the extra height of the mound gave it a motte-like appearance, but this was achieved in one operation prior to the Anglo-Norman invasion. In the barony of Ikerrin, Tipperary, an examination of platform ringforts concluded that they were not of medieval date because of their siting and location.

 

A paper on the timber and earthwork fortifications in western Ireland pointed out that there was little easily recognisable physical evidence for Anglo-Norman settlement west of the Shannon and in parts of Munster. Certainly, there is very little evidence for the classical motte-type castle, so we are dependent on finding other types of medieval earthwork castles for the physical remains of Anglo-Norman settlement. There are some cliff-edge forts which may be ringworks in Cork and in north Mayo recent fieldwork has turned up some hill-top enclosures which could also be of Anglo-Norman origin. In Co. Sligo almost 90 raised raths have been identified but only four mottes and one ringwork. Going on the basis of excavation results from similar sites in Ulster it is possible that between 25% and 50% of the raised raths could have Anglo-Norman remains on them. Therefore, raised raths could fill the void created by the lack of more easily recognisable, medieval earthwork fortifications in the western half of Ireland. Few raised raths, outside the Ulster counties, have been found in the eastern part of the country In Monaghan, where there are only three mottes recorded, there are a number of raised raths similar in their morphology to those excavated in Co. Down. In Co. Cavan, eight mottes and one ringwork castle were recorded; however, a number of sites classified as ringforts in the Archaeological Inventory would appear to be ringwork castles. For instance at Drumcor, Drurnharid, Cornaslieve and Lisnafana the enclosed area is raised considerablv above the level of the surrounding land. The enclosing elements consist of a verv large bank and a wide deep fosse and invariably they have causewayed entrances. However, as demonstrated by the excavation of a small number of raised raths in Co. Down, there is no way that we can be certain that any of these sites could be of Anglo-Norman origin. Virtually all of the sites listed as ringforts in Cavan are isolated and cannot be classified by their association with ecclesiastical remains. Since the Anglo-Normans settled areas west of the Shannon, in Munster and counties such as Monaghan and Cavan, they must have built earthwork castles. But as there are little or no obvious remains of their earthwork castles in these areas we must assume that some sites classified as ringforts, especially the raised or platform type, are in fact Anglo-Norman in origin, or Gaelic sites which have been adapted by the Anglo-Normans.

MOTTES

 

The most obvious physical evidence of early Anglo-Norman settlement in Ireland is the motte castle. Its classical form takes the shape of a Christmas pudding. However, fieldwork in Ireland in the last 20 years has shown that there is a wide variation in the shape and size of this type of earthwork castle. The extent to which the Norman conquest of Ireland was effective depended on how the great tenants controlled the conquered areas. The Anglo-Norman earthwork castles reflect this control in that, when the land was sub-divided, the sub-tenants built a stronghold, namely an earth and timber castle. Goddard Orpen was the first person to identify Irish mottes as the earthwork castles of the Anglo-Norman . He forcibly argued for a Norman origin for many of our earthen mounds which up to this time had been considered pre-Norman or even pre-historic. Orpen was very well aware of the absence of mottes in the western half of Ireland. Another substantial paper on mottes and raths in Connacht observed:

 

The Normans adopted the motte and bailey fortress as their ideal, using sometimes a natural rock or hillock as the citadel, or making one by piling up a mound. But they used other works freely, and abandoned it generally in Connacht in the middle of the thirteenth century for the flat motte type [i.e. the raised rath]

 

Mottes are found almost exclusively in the eastern half of the country. Work by the Archaeological Survey of Ireland over the past 30 years has not changed this distribution to any great extent. In 1972, the total number of mottes was estimated to be c.340 for the whole of Ireland.` Since the Archaeological Survey commenced its data base for archaeological sites in the 26 counties this number has been increased to at least 456. In Leinster alone there are 275 mottes and between Antrim and Down there are at least 106, leaving only 75 in the rest of the country. Less than 16% of all mottes are found outside of Leinster, Antrim and Down. The distribution of these sites in Ulster is easily explained in terms of the Anglo-Norman settlement east of the River Bann and their lack of penetration westwards. While there are mottes in the centre of Ireland virtually none have been found in Cork, Kerry or west of the River Shannon. Since we know that the Normans conquered these regions they must have used other types of earthwork castles to hold their newly-acquired territories.

 

Mottes were constructed in Ireland into the second decade of the thirteenth century and they were used in Leinster in some instances at the end of that century. The earthwork castle at Moydow Glebe, Co. Longford, possibly a low motte, was not constructed until 1261. Since it cannot therefore be argued that the motte was a redundant type of fortification by the time the Anglo-Normans conquered the west of Ireland, we are forced to accept the possibility that the raised raths and platform ringforts are the Anglo-Norman earthwork castles of the west.

 

The classical form of motte has the shape of a plum pudding. However, as can be seen from the detailed recording of the earthwork castles of Glamorgan, some mottes can have very low mounds rising barely above the surrounding ground level. The most obvious division within the motte category is whether or not they have baileys. In the six counties of Northern Ireland there are few mottes with baileys compared to English examples where it is the norm. The Ulster examples are also smaller. 43 Mottes in the eastern counties of Ulster are similar to those of Galloway, the most Gaelic part of the Scottish Lowlands. Out of 336 mottes in the Republic of Ireland 149 or 44% have baileys. Mottes with baileys are concentrated on the borders of the Earldom of Ulster and only 23% of all the mottes in Ulster have baileys. Mottes and baileys are quite small in Antrim and Down, where they mainly occur, and they were often built within existing ringforts.

 

In England the baileys were large enough to hold various buildings, including a hall, while the motte itself had a wooden tower on it for refuge. The Ulster baileys generally appear to be too small to hold a hall. Only about fourteen mottes have been archaeologically excavated in the whole of Ireland and very few of these have been excavated on a large scale. Four excavations of mottes in Co. Down gave some evidence of their defences and internal structures. At Clough Castle, which is thought to be on the route followed by de Courcy when he invaded Ulster in 1177, excavations revealed post-holes of a palisade around the perimeter of the mound. This excavation, which concentrated on the motte, concluded that the mound had no occupation prior to its final topping-off. The earliest defence of the site was the palisade which had been set back from the edge of the mound by up to 1m. The entrance way through the palisade was not found nor any evidence for a bridge from the bailey. A number of pits cut into the mound inside the line of the palisade were discovered and at least one of them was a weapon pit.

 

At Lurgankeel, Co. Louth, an archaeological excavation in 1965 produced evidence for a timber palisade and a 'breastwork of timber posts and earth'. In the centre of the mound there was evidence for a timber tower. At Clough Castle, there was no evidence for a timber tower but a central hollow was found and these hollows on the top of the mottes can indicate the area where a wooden tower once stood and subsequently collapsed and decayed. As has been noted in a study of timber castles in England, 'the very act of demolishing a structure, a tower, a gate or a palisade will cause disturbance of the earthworks which will tend to collapse down the sides.' The authors of this study also pointed out that there is increasing evidence that ramparts and even the motte itself could have been revetted with vertical timbers. When slippage is removed from the side of the mound a virtual timber casing can be found and where palisade post holes are located they may in some instances be part of an earth and timber wall.

 

Mottes were often constructed by using the material dug from the enclosing fosse and piling it up on the perimeter of the intended mound so that if it was never completed it would look like a ringfort or ringwork. Beal Boru, Co. Clare, was originally a ringfort and the secondary fortification, which had its banks heaped up on the primary enclosures, was intended, in the excavator's view, to be a motte. Whether one accepts the secondary site as a motte or a ringwork does not detract from the fact that the Anglo-Normans were using an earlier structure on which to build their own fortification. At Downpatrick, the Anglo-Norman mound appears to be built at one end of an earlier fortification and seems to be an unfinished motte rather than a ringwork. At Lismahon, also in Co. Down, a mound had three building phases, the two earlier ones indicating pre-Norman occupation. The second phase of building consisted of heightening the mound and enclosing it with a wooden revetment. The third phase of construction consisted of raising the height by piling over a metre of the boulder clay on top of the platform of phase two.

 

Several other mounds in eastern Ulster such as Castleskreen and Rathmullan, Co. Down, Doonmore and Dunsilly, Co. Antrim, were occupied and fortified prior to the Anglo-Normans taking them over. Other motte-like mounds in Ulster which have been excavated - Gransha, Co. Down; Big Glebe, Co. Derry; Deer Park Farms, Co. Antrim - produced only evidence for pre-Norman occupation?' So the identification of low mounds as mottes without baileys is just as much a problem as distinguishing some of the ringwork castles from large ringforts. The only sure way of identifying isolated low mottes without baileys is to excavate them. A small mound at Dunsilly, Co. Antrim, was excavated in the mid 1970s. It was sited at the edge of a low bluff near the River Burn. Its siting near the river, and the fact that it had a bailey, clearly identified this site as a motte even though the mound itself was very unimpressive.

 

Recent fieldwork in the Republic of Ireland has shown that nearly 45% of mottes have baileys. Their siting and location beside rivers, in commanding positions, near medieval churches and directly associated with historical references can give us clear indications as to their true identity. Like the ringwork castle the smaller examples may be associated with the sub-division of the manor so that the smaller mottes should be found in the more isolated areas and would belong to sub-tenants. Twentyfive mottes were recorded in Co. Louth; fifteen of these have baileys, five are in close proximity to churches and thirteen are sited close to rivers. In Wexford, where there are only sixteen definite mottes, seven were found close to churches. In Cavan, Meath, Carlow, Laois, Wicklow, Offaly and Dublin this same association with churches, in at least 25% of the sites, was noted. Their siting close to rivers and on high ground with commanding views is also typical.

 

As shown from the excavated mottes mentioned above, the earthworks as we see them today can be very misleading, in that their present form conceals as much as it reveals. Many of the mounds would have had much steeper sides and would have been revetted with stone or timber which has been robbed or decayed, allowing the mound to slump into a more gentle slope. Also, when we look at the earthwork remains of these early castles we tend to forget that what we see are only the foundations for wooden and, less often, stone structures which were placed on and around the mound. Most if not all of the excavations that have taken place on mottes have concentrated on the top of the mound. This is unfortunate because if the motte had a bailey attached most of the domestic activity would have been located within the enclosure rather than on top of the mound. In other instances mottes selected for archaeological excavation can be so eroded and degraded that no evidence of their defences or occupation survived on top of the mound.

 

The activity on the top of the motte at Rathmullan, Co. Down, was dated to 1200 based on the finding of three coins and historical data.` The primary structure on the summit of this motte consisted of a rectangular wooden building which had been burnt down. No evidence was found for a palisade or any defensive element on the perimeter of the mound. The occupation of the motte phase of the mound at Lismahon, Co. Down, was tentatively dated to c. 1200 and the date depended largely on the finding of a coin of John de Courcy. Evidence for a rectangular-shaped building, regarded as an 'English-style' hall, was found on top of the mound with evidence for a wooden tower attached to its north-east angle.

 

The excavation on the top of the mound of a motte and bailey at Dromore, Co. Down, produced evidence of two flimsy palisades around the parameter. This site was dated on documentary evidence to before 1211-12 because the Irish Pipe Roll from the fourteenth year of the reign of King John contains several entries which refer to a new bridge, hall and other buildings at Dromore. But like the motte at Clough, it was probably erected soon after 1177 by John de Courcy or one of his followers. A motte with a double bailey at Ballyroney, Co. Down, was partially excavated?' This site like many of he mottes in Ireland is located close to a river on high ground. Both baileys were defended by a single bank and fosse. The motte was separated from the baileys by a fosse. Excavation at this site concentrated on the top of the mound but there was a reasonable quantity of small finds, including iron arrowheads and horseshoes, as well as a silver halfpenny of John (Lord of Ireland) and sherds of pottery. Another motte at Dunsilly, in Co. Antrim, was partially excavated but no remains of any structures were found on top of the mound. Dunsilly had been constructed on top of a ringfort which had been occupied for some time. As can be seen from the brief description of most of the excavated mottes the amount of structural evidence is very sparse and, in some instances, there was little evidence of occupation or of any substantial defences. It is unfortunate that no one has excavated any motte and bailey thoroughly so that we could see exactly what areas of these earthwork castles were defended and occupied.

 

An earthwork at Castleskreen, Co. Down, was partially excavated in the early 1950s and it showed that it was constructed in three phases. The final phase, built around 1200, was intended to be a motte, but it does not appear to have had any defensive features on the top of the mound.` Clough Castle, Co. Down, of all the excavated sites, had the most elaborate buildings and defences, and appears to have been occupied from c. 1200 to 1500. A number of archer pits, one with the remains of a stone loop built into the palisade, were found around the perimeter of the mound. By the middle of the thirteenth century a single-storey stone hall was constructed and, later in that century, a two-storey rectangular keep was built. A considerable quantity of small finds were found associated with the occupation on the top of the mound. Most of the pottery appears to have been local ware but some was imported from the Bordeaux region. There was also a large collection of pottery which was dated to the second half of the thirteenth century. A considerable number of iron objects and a few of bronze and bone were also discovered. So, compared to the other excavated mottes, it was relatively rich and must have been an important site, yet there are no historical references to it.

 

In Anglo-Norman Ulster, mottes with baileys were concentrated on the borders of the Earldom or beyond and there were none in the centre of the settlement in south Down or north Antrim. In the Liberty of Meath (comprising modern Westmeath and Meath) it was claimed that mottes with baileys often had a frontier location. However, work over the past twenty years by the Archaeological Survey shows a more random distribution. In Co. Louth, where a detailed survey has been carried out, most of the mottes and baileys have a riverine location rather than a political or military frontier siting. It has been suggested that the building of a seignorial castle (of earth and stone) might merely be a matter of prestige rather than defence. Further, the excavated mottes at Rathmullan, Co. Down, and Dunsilly, Co. Antrim, did not appear to have defensive features on their summits, and Castleskreen, Co. Down, had its bailey occupied throughout the thirteenth century and yet had nothing built on top of the motte. Some mottes were heightened with the intention of later fortifying them in the event of a likely attack. Henry 111 made an order in 1225 to the lesser Lords in the Vale of Montgomery to fortify their mottes with stockades. Similarly, King John made a command in 1200 that persons holding land in the marches of Ireland should fortify their castles. This probably means placing bretasches (wooden palisades) on top of already heightened mounds such as those at Dunsilly and Rathmullan.

 

The vast majority of earthwork castles were built in the initial stages of the conquest. De Lacy, for instance, had built most of his fortifications in Meath for himself and his lesser Lords by 1181, according to Giraldus Cambrensis. But the vast majority of Co. Meath castles which were returned to Walter de Lacy from 1215 onwards were of earth and timber. In addition, some of the important Marshal castles of Leinster in 1231 were earthworks. So we must remember that many of the earth and timber fortifications were built for long-term occupation and not merely for the initial conquest period. Despite the fact that Rohesia de Verdun had built a motte castle at Mount Bagnall, Co. Louth, and had a manor at Castletown and a stone castle at Castleroche, she was granted respite in 1242 regarding the colonisation of her land in the heart of Co. Louth. In the early stages of settlement, the baronies in Louth merely represented pre-existing boundaries. The charters given to Pipard and de Verdun reflect these divisions in Oriel. Mottes may still have been used as fortifications until the beginning of the sixteenth century especially in the outlying areas of a manor and, in Leinster, mottes with baileys tend to occur at the Capita of the principal land grants. In Co. Wicklow, a frontier region, few mottes with baileys occur on the periphery. Mottes without baileys in Leinster appear to have large summit areas, probably to compensate for the lack of space which would have been provided by the bailey.

 

Some mottes have 'squared-off' or angular summits as at Aghaboe, Co. Laois, or at Faughart Upper, Co. Louth. In some instances the 'squaring-off' of the mound may be due to refortification of the motte by building a masonry wall around its summit. Several of the mottes in Co. Louth illustrated by Wright had quite substantial stone fortifications." The remains of some of these fortifications can still be seen at Faughart Upper, Ash Big, Castlering and Stormanstown. At Aghaboe there are the remains of wall footings around the edge of the summit of the mound. Some of these stone walls around the edges of the summit may not be part of the original fortification and may also contribute to the 'squared-off' look of the motte. Other motte summits such as at Racraveen, Co. Cavan, have what appear to be low earthen banks around their edges, which may be the remains of an earth and timber palisade, rather than a stone wall, and would therefore be an original feature of the motte.

 

Since 1986 the Archaeological Survey has been publishing county by county archaeological inventories and it is hoped by the end of the millennium the fieldwork will almost be completed. All the present information is on a data-base and readily available but, unfortunately, when the field survey first commenced, there was often a lack of knowledge which would have helped in the precise classification of the medieval earthworks. Recent fieldwork in Tipperary, Offaly and Roscommon, to name but three counties, has demonstrated the diversity of the medieval earthwork and probably for the first time since Orpen published his work on mottes, we are making real progress in our identification of the Anglo-Norman earthworks and their distribution. This work needs to be followed-up by a campaign of selective excavation on the baileys as well as the tops of mottes and on the various forms of possible ringworks such as cliff-edge forts. Until such investigations are carried out it is difficult to see how we can progress our knowledge of the Anglo-Norman earthwork castle in Ireland.