Castle Matrix

Chapter I

Introduction

Castle Matrix lies close to the River Deel just to the south-west of the small town of Rathkeale, which itself lies sixteen miles south-west of the city of Limerick. Both places lie within the county of Limerick, which along with the counties of Clare, Cork, Kerry, Tipperary and Waterford, comprise the ancient ecclesiastical province of Munster. Most of the places mentioned in this account lie within that province.

The unusual-sounding name of the castle is an anglicisation of a Gaelic name, probably Caislean Bhun Traisce, although the original meaning has been forgotten. The place is referred to as Caslan Mattras in 1487, and Courtmatres is mentioned in 1488. Rathkeale itself appears to take its name from one of a series of raths or ringforts dating from the period before the Anglo-Normans arrived in Ireland, and lying to the south of the town.

One of many castles on the extensive lands of the earls of Desmond, Castle Matrix is a late medieval structure of a type common in Ireland, Scotland and the northern counties of England, which is now usually referred to as a tower house. This term is not itself ancient, having been first used by John Parker in 1860. Although the tower has been heavily altered over the years, and has at some point lost its original topmost storey, it retains a layout in which each of the three storeys now remaining above what was originally a store at ground level is divided into a large room in the main body of the tower and a smaller room at the east end. The latter form part of a tier of such chambers over the entrance lobby and this allows it to be placed within a group of tower houses located mostly in the west of Ireland. It also retains several other interesting features including an unusual form of gunport.

The tower at Castle Matrix would have formed part of a manorial complex with a hall and service buildings, all probably timber-framed and covered in thatch. Although it may have moved at least once within the estate, such a complex probably existed in this vicinity ever since Rathkeale was granted to Hamo de Valognes in 1199 by King John. The estate was held by the Westpayl family in the early 13th century and then passed to the Mautravers family, whilst the priory of Augustinian canons at the east end of the town is thought to have been founded by Gilbert Harvey in the early 13th century. Rathkeale passed to Maurice, 1st Earl of Desmond in the mid 14th century, but the circumstances are now forgotten.

There is no physical trace or record of the tower at Castle Matrix being accompanied by a stone-walled enclosure of the type known in Ireland as a bawn. Most bawns in Ireland date from the second half of the 16th century, and as we shall see later, that isn’t a period when we would expect major construction work to be in progress at Castle Matrix. The tower’s proximity to the River Deel does raise the question as to whether it was accompanied by a water-filled moat, although surprisingly few tower houses in Ireland are associated with moats. If such a feature did exist here it would probably have been created long before the tower itself was built.

Chapter II

Relationship with Nearby Ancient Buildings

Castle Matrix is now isolated from the town of Rathkeale by a bypass section of the N21 main road linking Limerick with the town of Newcastle West. The original route is now the R523 road, which, coming from the Limerick direction, first passes the ruins of the 13th century Augustinian priory church. This has a row of original lancet windows on the south side of the choir and remains of two later tower-like projections, one to the north of the choir and the other at the west end. The R523 then comes through the middle of the town, and crosses the River Deel to the north-east of Castle Matrix. The bridge here, although widened and displaying no sign of great age at road level, still retains medieval arches with evidence of wickerwork centring likely to be roughly contemporary with the castle. It is not unlikely that the same team of masons worked on all three structures, and that parts of the wooden shuttering used to build arches and vaults in the tower at Castle Matrix saw further use upon either the bridge or the priory church.

South of the bridge lies the site of the medieval parish church of Rathkeale, a 13th century structure now replaced by a more modern Church of Ireland place of worship. The earls of Desmond were buried much further away in friary churches having a long association with their ancestors at Youghal in Co Cork and Tralee in Co Kerry. However, at least some of the Southwells, owners of Castle Matrix from the early 17th century onwards, were certainly buried in this churchyard and one small monument to them remains within the church still in use. On the other side of the R523, between it and the N21, lies an estate called Glebe, on which lies a ruinous and smaller tower house built in the 16th century, presumably to provide a residence for the parish priest of Rathkeale.

Castle Matrix lies in a part of Ireland offering good agricultural land which is thickly studded with medieval towers. Indeed, the county of Limerick is reputed to have once had about four hundred of them, of which about seventy still stand, at least in part. A late 15th century tower and fragmentary bawn which formed the main seat of vassals of the earls of Desmond, the MacSheehies, survive at Lisnacullia, just two miles to the north of Castle Matrix. Four miles to the north-east, at Cappagh, are ruins of a tower and double bawn of the Knights of Glin, who were descended from a junior branch of the family of the Desmond Geraldines. Two miles to the south-east is the small tower of Rathnasaer, another possession of the earls of Desmond. It is said to be where James FitzThomas FitzGerald (the Sugan earl) was proclaimed as Earl of Desmond during the rebellion of 1599. Lastly, just over two miles to the south-south-west, is Ballyallinan, with a ruined but fairly well-preserved five-storey tower with some similarities to the tower at Castle Matrix.

Chapter III

The Earls of Desmond and Their Castles

The earldom of Desmond created in 1329 for Maurice FitzGerald was one of three new Irish earldoms established during the reigns of the English kings Edward II and Edward III. Surviving contemporary records concerning 14th and 15th century lords in Ireland are scanty. This is particularly so for the earldom of Desmond, where castles and monasteries containing such records were ravaged during the struggles of the 14th earl and his allies against the authority of Queen Elizabeth I. Records of activity in this part of Ireland in the 15th century are mostly based on 17th century accounts, subject to both bias and difficulties of translation from Irish to English. We don’t even know for certain where or when some of the earls were born or buried. Even the numbering system of the earls is disputed as a result of a son of the first earl being considered mentally unfit to succeed to the title, and a situation in the 16th century where possession of the earldom was disputed. The numbering (and other information) used in this account is based on that used in the first of several 20th century revised editions of The Complete Peerage, originally published in the 1880s and 90s.

It is unlikely that the earls of Desmond spent a lot of their time at Castle Matrix. Men of their rank usually moved around from one manorial centre to another during the course of each year, conducting business and dispensing justice in their own courts and consuming agricultural produce on the spot. Each generation might have its own favoured locations for longer stays according to political, financial, sporting and other interests. Also, the preferences of their spouses (who were often immobilised by the later stages of pregnancy) might have to be taken into account. By the end of the 15th century the earls had numerous castles on their widely distributed estates extending from the northern half of Co Kerry, through Co Limerick and the northern and eastern parts of Co Cork, down into Co Waterford. We know very little about their castles in Co Kerry, most of which have vanished, but there are spectacular remains of their castle of Mocollop in Co Waterford and of several of their other castles in Co Limerick such as Askeaton and Newcastle (now called Newcastle West). Both these are within 7 miles of Castle Matrix and existed before the creation of the earldom. One of the earls also began building an impressive new castle with octagonal corner towers at Glenogra, 15 miles east of Castle Matrix, but it was long left incomplete. Another four miles further east, at Lough Gur, is a 16th Desmond tower (now known as Bourchiers after later owners) which is similar in type to Castle Matrix and will be referred to later. At times, usually for fairly short periods, the earls of Desmond were also custodians of the royal castle at Limerick, and by the late 15th century they also held the former royal castle at Dungarvan on the coast of Co Waterford. Some of their wealth derived from controlling interests (and in particular duties on imports of wine) at the ports of Kinsale and Youghal on the south coast of Co Cork, and both of those places retain still-roofed late medieval houses associated with the earls of Desmond.

Chapter IV

The Murder of the 8th Earl of Desmond

Like many lords of Anglo-Norman descent, especially those with lands adjoining those of native Irish kingdoms, the earls of Desmond had begun to adopt the Irish language and customs. The process may have started as a result of the defeat and capture in 1370 of the 3rd earl, Gerald FitzMaurice, by Brian O’Brien, King of Thomond (now Co Clare). Gerald’s younger son James, who ruled as 6th earl from 1420 until 1463, was brought up at Brian’s Irish-speaking court, probably as part of the terms of his father’s release, and James, 8th earl of Desmond was married to a daughter of one of the O’Brien kings of Thomond.

In 1484 the 8th Earl of Desmond was asked by King Richard III of England to “renounce the wearing and usage of the Irish array” and to adopt English style dress, but otherwise Earl James was high in favour with the English court. At least one of his younger brothers seems to have resented this, and supposedly arranged the killing of the earl “at his house of Courtmontresse” in 1487. The word house (used in a 19th century translation of the Annals of the Four Masters, a mid 17th century account written in Irish gaelic) is ambiguous and neither proves nor disproves that the tower (as opposed to the manor) at Castle Matrix existed by then.

The Complete Peerage, probably quoting from the same source, says that the 8th earl died “by the stroke of a bullet” fired by his “servant” John Murtagh. The word bullet again might be a mistranslation from the same source but if it means what it says then that is of great interest. Firstly it would be an unusually early reference to the use of a firearm in Ireland, although there is a record of an O’Donnell killing an O’Rourke with a firearm in the same year. Pistols were unknown in Ireland until at least the 1520s so the weapon used would have been a demi-hake or arquebus. Such a weapon would have been an expensive import from Spain or the Netherlands. It was quite heavy and needed priming so its use would be premeditated rather than spontaneous, and by someone who had undergone military training, and was probably of fairly high rank. If John Murtagh killed his master by such means the word servant describing his role is more likely to mean as a military captain or an estate steward (or bailiff) rather than one who carried out more menial household tasks.

All those involved in the killing are said to have been put to death shortly afterwards by the 9th earl, Maurice FitzThomas. Curiously enough, Maurice himself was lame as a result of a gunshot wound, as was his contemporary, James 9th Earl of Ormond. Furthermore, in 1600 a younger brother of the “Sugan” Earl of Desmond was saved from assassination whilst riding out from Lough Gur when an attendant snatched away the pistol of an English agent who was accompanying them.

Chapter V

Castle Matrix During the 16th Century

There are hardly any 16th century references to Castle Matrix. The long-standing feud between the earls of Desmond and Ormond reached a climax in 1565 when Gerald, 14th Earl of Desmond, was defeated and captured at Affane by Thomas Butler, 10th Earl of Ormond (a cousin and favourite of Queen Elizabeth). Gerald was subsequently confined in the Tower of London until 1573, so it is unlikely that major building works were executed at any of his castles after 1565. The earl’s brother Sir John of Desmond was defeated near Monasternenagh Abbey in October 1579 by Sir Nicholas Malby, Lord President of Connacht. There is no specific record of anything happening to Castle Matrix in the campaign that followed, which culminated in the earl being hunted down and killed near Tralee in 1583. Even the much larger castles of Askeaton and Carrigafoyle, both with inner and outer bawns surrounded by water defences, each only held out against Malby’s artillery for a couple of days during April 1580. Towers without outer defences like Castle Matrix weren’t expected to withstand a siege by a regular army equipped with cannon. The probability is that Castle Matrix was burned or unroofed by one side or the other, most likely by Earl Gerald’s own forces to prevent English forces from occupying it, as is known to have happened at the castle of Ballygleaghan, 10 miles north-north-east of Castle Matrix. One of Gerald’s chief allies, Lord Barry, burned his own castle of Barryscourt in Co Cork so that it required extensive rebuilding in the 1590s following his submission to the English authorities.

In the late 1580s extensive lands of the earldom of Desmond were disposed of to courtiers of Queen Elizabeth, the recipients being known as undertakers, since they undertook to establish the Protestant religion and English law, language and customs on their lands. The administrator and poet Edmund Spenser got Desmond’s castle of Kilcolman in the north part of Co Cork and made his home there, although he found himself very isolated amongst hostile neighbours, as did several of the other undertakers. More extensive former Desmond lands amounting to 40,000 acres across the counties of Cork and Waterford went to Sir Walter Raleigh, who was a closer associate of the Queen. Raleigh held the castle of Lismore in Co Waterford, and served as mayor of Youghal for a year, although it appears that a deputy did the administration since Raleigh remained in England except for when indulging in sea-faring adventures. There is no evidence Walter Raleigh ever held any lands in Co Limerick, although he and Spenser may have briefly visited Castle Matrix c1580 whilst serving as captains in the English forces led by Sir Nicholas Malby.

Chapter VI

The Billingsleys and the Dowdalls

Castle Matrix was amongst several properties in Co Limerick (which also included the castles of Kilfinny and Lisamota) which were offered to Sir Henry Billingsley, a merchant and customs officer of London. His son, another Sir Henry, eventually took up the offer but does not appear to have ever set foot in Ireland. By the late 1590s the younger Sir Henry Billingsley had settled at Penhow, a small castle of 13th century origin in the county of Monmouth (now in the south-east corner of Wales), where he served as sheriff. By 1608 Castle Matrix and Kilfinny had been disposed of to Sir John Dowdall of Pilltown in Co Waterford. His father, also called John, descended from a family of lawyers from Co Louth, had held several military commands, during the course of which he captured Sir John Desmond at Castlelyons in 1582, took the castle of Enniskillen in 1593, and later commanded the newly-built fort at Duncannon in Co Wexford.

The younger Sir John Dowdall was granted the right to hold a weekly fair on Thursdays by the high cross of Rathkeale, but he remained seated at Pilltown, of which the Dowdalls had a lease from Nicholas Walsh from c1591 until 1620. Now destroyed, that castle had also originally belonged to the earls of Desmond and was considerably stronger than Castle Matrix, having three walled courts when it was blown up by Murrough O’Brien, Lord Inchiquin in 1646.

Chapter VII

The Southwells at Castle Matrix

By 1616 Castle Matrix had been transferred to another undertaker, Edmund Southwell, one of a family of merchants from the English county of Essex, near London, and who was related to Sir John Dowdall’s wife. It was she who, as a widow, held the castle of Kilfinny against a long siege by the Confederate Irish in 1641. In the following year Castle Matrix was captured by the Herbert family. Other nearby castles held by Protestant undertaker families in Co Limerick which were attacked by Confederate Irish forces at that time included Croom and Newcastle West.

Edmund Southwell’s son Thomas, being a Protestant, would not have been regarded as hostile to the Cromwellian regime, and during the 1650s he served various terms as sheriff of the counties of Kerry, Limerick and Clare. However it is possible an unrecorded attack was made upon Castle Matrix c1649–50 to recover it from Confederate hands. Castle Matrix was listed in 1662 as Thomas’s seat when he was made a baronet by King Charles II. Thomas’s son, another Thomas, who in 1709 settled Calvinist refugees from the Rhenish Palatinate on his lands at Rathkeale, was made Baron Southwell in 1717. Castle Matrix was still the family seat when the 3rd Baron Southwell was made a viscount in 1776. At their height the family owned 60,000 acres across Ireland and had a profitable flour mill on the riverbank not far from Castle Matrix.

Chapter VIII

The Later History of Castle Matrix

Most of Ireland’s tower houses (of which about 1,050 survive at least in part) gradually became ruinous sometime during the late 17th or 18th century. Castle Matrix is one of about sixty medieval Irish tower houses which remained fully in use right up to the early 20th century. All of these towers were altered in various ways, and in many cases extended too, to adapt them either for more up-to-date forms of living or for agricultural or industrial purposes.

By the early 19th century the Southwell family had transferred to a house near Dun Laoghaire (then known as Kingstown). Subsequently Castle Matrix was let out to a series of tenants, starting with Captain Brown, who acted as one of the Southwells’ land agents. The existing windows and battlements of the tower appear to date from the late 1830s. There is uncertainty as to when the existing northern extension was built. Its features appear to be of the 1830s but the outer walls may be older. The accommodation available in the tower and extension seem comparatively modest to have served as the main seat of an 18th century viscount and the extension might be a cut-down version or a replacement of a larger structure, although if so one might expect that a written record or an illustration of a larger pre-1830s house here would have survived. A new tenancy agreement from 1853 refers to the castle as having nine bedrooms (most of them in the later extension, and some of them quite small), in addition to a dining room and library in the tower and other smaller rooms.

The only certain in-situ relic of the period after the Southwells obtained the castle in 1616 is a fireplace in the north wall of the tower at third storey level. It has a joggled lintel made of several pieces with side-brackets to secure them. It could actually be of somewhat earlier date but for the brick back of the flue which is unlikely to predate 1616. Of about the same period is a wooden fireplace surround with crudely carved figures lying in the central space of the lowest storey of the extension, but because of its nature this feature could have been brought in from anywhere at any time.

Castle Matrix seems to have been dismantled in the 1930s by its then owners the Johnstones to avoid taxes being paid upon it, and was only made habitable again after being purchased by Colonel Sean O’Driscoll. For a while it was a genealogical centre noted for its library and exhibits. After Colonel O’Driscoll died in 1991 his widow was unable to maintain the building and its grounds, both of which suffered from vandalism. Deterioration gradually set in again until the castle was sold in 2025 to another American, Bob Tiernan, by whom it is currently being refitted internally.

Chapter IX

The Tower House at Castle Matrix

The tower at Castle Matrix measures 12.5 metres long by 8.9 metres wide above a battered base extending through the height of the lowest of the four main rooms. This is about the average size in this part of Ireland for a tower house. There are stone vaults over the second and fourth storey main rooms. In the restoration of the 1960s the wooden floors of these two rooms were only partly reinstated, and now take the form of suspended galleries. All of the narrow original windows of the main rooms have been replaced by mullion-and-transom windows three lights wide which appear to date from the late 1830s, so that the interior spaces of the tower are now far better lighted than they would have been originally. The window embrasures are presumably enlarged versions of the medieval ones, except that one (probably the easternmost) of the two south-facing embrasures at second storey level is probably a 17th or 19th century insertion.

The tower is now entered through a doorway from the later extension on the north side giving onto a lobby, from which there is access to the lowest main room and also to a straight staircase rising to where a spiral staircase begins at second storey level in the south-east corner. Although the existing entrance doorway appears to contain at least some original medieval material, its position would be very unusual for a tower house of its particular type (which will be discussed in the next section). The entrance lobby has a later window facing east and the general assumption is that the entrance doorway was originally set in that location, which is, however, rather closer to the north-east corner than is normal for buildings with the sort of internal layout found at Castle Matrix.

In tower houses where the entrance doorway and the foot of a spiral staircase starting higher up are connected by a straight staircase, there is usually solid walling underneath both staircases. Castle Matrix differs from the norm in that there is a rectangular chamber in this position which is lighted from the east by a pair of narrow loops which appear to be original. This implies that the chamber itself is original. A former owner claimed that there was once a lower staircase leading down to another chamber below the entrance lobby. Underground chambers are very rare in tower houses in Ireland (and also amongst those of Scotland and England) but it so happens that there are two other towers in County Limerick that do have such chambers. One is Castle Hewson (formerly called Ballyengland), and the other is Bourchiers Castle at Lough Gur. Coincidentally, both these towers have lost their original topmost rooms above an upper vault in the same way that has happened at Castle Matrix, as will be discussed later. At Castle Hewson the underground chamber is reached by a stone staircase, but that at Bourchiers was only accessible by a hatch in a timber floor. Sarah McCutcheon, the county archaeologist for Limerick, has suggested to the author that the underground room at Bourchiers may actually be a relic of an older gateway or tower house on the site of the existing tower house, which the author thinks may be as late as the 1540s.

Chapter X

The Internal Layout of the Tower

Above the entrance lobby at ground level the tower at Castle Matrix has a tier of five subsidiary rooms, two of these being set at levels corresponding to the stone vaults over the main rooms. About a hundred tower houses in Ireland have such a tier of subsidiary rooms, those here all being reached directly from the spiral staircase and each having a single window similar to those of the main rooms, but here only two lights wide. Tower houses of this type appear to have been built in Ireland over a period of roughly two hundred years from about the 1450s until the 1640s.

The subsidiary rooms at Castle Matrix all have a width of about two metres. The width of these rooms is important, being sufficient for a double bed to be either placed across the width of the far end of the room from the staircase, or set centrally at that end with narrow floor-spaces on either side of it. The combination of the width of these rooms and their placing in a vertical tier, subdivided vertically only by wooden floors instead of vaults, distinguishes this type of tower from others elsewhere in Ireland. In other towers subsidiary rooms are either lacking, or are narrower and more randomly scattered around within the thickness of the walls, or, in the case of the thinly-walled towers of the eastern parts of Ireland, required projecting wings or turrets to accommodate them. At Castle Matrix one of the upper subsidiary rooms is now furnished as a chapel, although it is orientated north-south rather than east-west and has no piscina or other feature suggesting that it was originally intended to be used as a chapel. Rooms specifically intended as chapels are actually very rare in Irish tower houses, although they do occur occasionally in Scottish towers. There are currently no signs of a floor ever having existed between the chapel and the topmost level above (which has a stone vault) although it is likely that there originally was a timber floor in that position.

Towards the end of the medieval period there was a trend towards members of a lord’s household being allowed a greater degree of privacy, although the sharing of beds (amongst those actually allocated the luxury of space in one) remained the norm. Subsidiary rooms as wide as two metres in medieval tower houses were secondary sleeping spaces, probably mostly intended for the children of the owner, custodian or tenant of the castle, or senior household officials such as a chaplain, bailiff or military captain. At Castle Matrix the head of the household occupying the manorial complex would have slept in the third storey main room, ie that above the lower of the two vaults in the main part of the tower, where there is a good early 17th century fireplace. People of the next rank within the household would have slept in the rooms above and below, ie those covered with vaults. In tower houses of this specific type in western Ireland these rooms below vaults were neither heated nor well-lighted as originally built, so were not living rooms intended for day-time use.

Chapter XI

Angle-Loops in the Tower

Although all the original narrow windows lighting the main and subsidiary rooms at Castle Matrix have been replaced by wider openings, the spiral staircase in the south-east corner is still lighted by its original loops. These are tall but narrow openings about 7cm wide with a splayed outer opening and face either east or south. Two of them are particularly interesting. One of them has a bottom roundel and will be discussed at length in the next section. The other pierces the south-east corner of the building about half up the height of the building.

Angle-loops are an Irish speciality, found nowhere else in Europe, and are particularly daring from a constructional point of view. Usually they light staircases and small chambers or passages within the wall thickness and allow people a better command of the space around the tower. However a 16th century tower of the earls of Desmond at Lough Gur, now known as Bourchiers Castle, has a pair of angle-loops opening directly off the northern corners of the third storey main room (ie the corners furthest from the staircase) so that the embrasures of the loop pierce walling of much greater thickness than usual. This is a rare feature, but there are other pairs of angle-loops functioning in this way in towers at Rockstown, also in Co Limerick, and Lisronagh in Co Tipperary, whilst there are towers containing single examples of angle-loops of this sort at Ross in Co Kerry, Cappagh and Smithstown in Co Clare, and Carrgareely in Co Limerick.

Another tower of the earls of Desmonds, that at Kinsale (originally a customs house but latterly used as a prison) goes a step further and has corner windows of two lights. Each light pierces a side of the building whilst the mullion between them forms the actual corner. A similar feature, but with narrower lights, appears high up on a wing added in the 16th century by an earl of Desmond to an older main block at Castle Barrett in Co Cork in order to provide extra private rooms.

Following a recent visit to Bourchiers Castle, the author now considers it likely that a second angle-loop at Castle Matrix, located at the south-west corner, also probably opened directly off the third storey main room, although there is currently no trace of this internally, mainly because 20th century box-frame shelving obscures the internal corner.

Chapter XII

The Gunport at Castle Matrix

A particularly interesting loop opening off the staircase at Castle Matrix is one at fourth storey level, facing south, which has a bottom roundel or oillet about 10cm in diameter. Although a weapon of any sort could have been fired out of it, this roundel is large enough for use by a small cannon like that which can be seen in the Waterford Treasures Museum. The cannon is thought to be a relic of Perkin Warbeck’s siege of that city in 1495. It would have been mounted on a wooden board and such a weapon would have obstructed the staircase so the placing of one here cannot have been a permanent arrangement. In fact it is possible that the opening itself was intended to be enough of an intimidation to potential attackers since actually firing a gun in such a confined and awkward position like two thirds of the way up a spiral staircase would have permanently deafened anyone close to it. The feature appears to be original and suggests a date for this level of the building at least not earlier than the 1490s.

Openings of this sort, wide enough for cannon, are described by the author in his books as gunports so as to distinguish them from gunloops, a term usually used to refer to slightly smaller openings dating from the second half of the 16th century and the early 17th century, which were intended for the use of pistols or muskets. In general, examples of gunloops in Ireland are not as neatly made as the gunport at Castle Matrix. Other gunports directly comparable to that at Castle Matrix are in fact rare in Ireland. That at Coolisteague in Co Clare affords the closest parallel in term of position, ie set high up in a tower house wall, although that gunport has no vertical opening associated with it. Other simple circular gunports without a vertical loop above are found at Castle Kirke (a 13th century keep much altered in the 16th century) and in the outer wall to the east of the Banqueting Hall at the Desmond castle of Newcastle West.

In a rectangular tower in another part of the castle at Newcastle West is a corner loop with a horizontal cross-loop and a bottom roundel. That loop in turn is paralleled by one at Kilkea in Co Kildare, a seat of the earls of Kildare, and several similar loops at Granagh, near Waterford. The latter are thought to be the work of Thomas, 7th Earl of Ormond ie c1476–1512. One of these loops at Granagh pierces the corner of a tower house integrated with a residential block like at Newcastle. A possible sequence is that the Earl of Ormond brought back the idea of such loops from England c1485–90 and his rival Desmond then copied them c1495–1500. That at Kilkea, probably of the time of the 8th Earl of Kildare, c1500–13, adds to the impression that such loops were a badge of status, ie the prerogative of earls and Irish chiefs of similar rank. The Welsh castle of Raglan has both large plain circular gunports and crossloops with bottom oillets probably of the 1460s but there are no references to firearms being used in Ireland in that decade. Bourchiers, a Desmond castle already referred to several times, which may be as late as the 1540s, has a pair of gunport roundels set near, but not at the base of, vertical loops with chamfered edges.

Chapter XIII

The Missing Topmost Room of the Tower

In western Irish towers of the period c1460–1530, the largest and most splendid room was usually set at the top, above an upper vault. In that location the walls could be made thinner, resulting in a wider space, and the windows larger. This room was the only heated space in such towers as originally built, and its fire was in a hearth set in the middle of the floor, hence the importance of a vault over the main room below to provide a fireproof floor. A second vault lower down wasn’t essential to this scheme and was dispensed with in some towers. The topmost room was open to a hipped roof covered in thatch (or possibly stone tiles), through which the smoke escaped by means of some sort of louvre. A hearth set in the middle of the floor of a room seems to have been particularly important to the culture of families adhering to Irish customs. Towers further east in Ireland had mural fireplaces when first built and this enabled the principal room to be set lower down in the building with the main sleeping spaces provided above it, as was the normal arrangement in the tower houses of Scotland and northern England.

Up until about twenty years ago the topmost rooms of western Irish tower houses were described by writers (including the present author) as halls, and these spaces are commonly presented as such in towers open to the public. However several recent post-graduate theses have now established (at least amongst professionals in the field of castle studies) that tower houses in Ireland were essentially private spaces. The principal room in these western tower houses was a great chamber in which a select few would dine and be entertained in an evening. The unheated rooms further down in the tower at Castle Matrix were sleeping spaces, whilst the room at ground level would be used for storage.

The importance of this topmost room is such that it is unthinkable that it never existed at Castle Matrix. It would be like building a lighthouse without ever providing it with a beacon light. Even if a problem such as a temporary shortage of labour, materials or finance caused a hiatus in the process of building a tower, the topmost room would have been added eventually, unless the manorial complex of which it formed part had ceased to function, for which there is no evidence here. A likely scenario which might account for the loss of this room is that when the Southwells first arrived the tower may have been lying derelict for over thirty years since the conflicts of 1579–81, and the thin walling above the upper vault had either been deliberately destroyed or become quite ruinous. The Southwells would have been used to having their main living spaces no higher than a single storey above ground level and heated by mural fireplaces rather than hearths in the centre of floors. Thus it would make more sense to the Southwells to add other rooms with mural fireplaces set much lower down beside the tower, and remove a topmost room which was so damaged that it needed considerable rebuilding. To them such a room would have seemed both old fashioned and to have associations with a culture and way of life alien to that which they were used to.

Chapter XIV

The Ways of the Irish

Suggesting that none of the sleeping spaces in the tower had any means of heating them as originally built needs some further comment. Interestingly, a letter written in 1595 by the elder Sir John Dowdall (father of the Sir John who owned Castle Matrix for a few years) sheds some light on this. Speaking of the Irish, he says “they are a nation bred idly and in looseness of life, men full of agility and strength. They can endure hardness of diet. They care neither for good lodging, cold nor wet...” As already mentioned the sharing of beds was normal and would have helped people to keep warm at night, so the builders of towers like Castle Matrix saw no need for providing fireplaces in their bedrooms. Fireplaces needed wood to burn, and servants to collect and carry the wood. This amounted to quite a lot of work in a building with multiple fires burning night and day for much of the year. A similar attitude prevailed in the remoter parts of the Scottish Highlands in the 18th century when local clansmen continued to wade across rivers even though bridges had begun to be provided as part of the new road system laid out under General Wade. Highland clansmen thought of bridges as being for those less agile and hardy than themselves, ie town-based merchants, women and English soldiers.

The central hearth in the topmost room seems to have been largely ceremonial and for lighting and cooking as much as for providing heat. Trying to provide warm and properly cooked food and a supply of fuel in a dining chamber only accessible from ground level by a narrow spiral staircase of up to sixty steps would have been regarded as inconvenient and old-fashioned in the rest of Europe. Rooms which are clearly identifiable as being kitchens are almost unknown in medieval Irish castles of any date or form. Where kitchens do very occasionally exist in Irish tower houses, their fireplaces and ovens are either clearly later insertions to an older structure, or the whole building dates from close to 1600 (ie is post-medieval by English standards). By that time English culture was gradually starting to prevail amongst the landed classes of the west of Ireland.

In 1591 the German Ludolf von Munchhausen said that Irish towers were only used as refuges and that the owners or tenants lived in poorly-built adjacent houses with central hearths. Luke Gernon c1620 describes dining with an Irish chief in his great chamber high up in a tower, and says that although a good and varied supply of alcoholic drinks were on offer upon arrival, throughout the stay, and again on departure, cuisine was primitive, meat for instance being “ill-cooked and without sauce“.

Furnishings of all kinds also seem to have been sparse in Irish towers and their associated outbuildings until quite late in the 16th century. In medieval towers in Ireland the windows rarely had glass panes and were closed with internally-opening shutters rather than curtains. Clothes were usually kept in chests instead of wardrobes. Tables and seats of any kind were uncommon, and carpets and wall-hangings such as tapestries and pictures in frames only became the norm in the 17th century.

Chapter XV

Garderobe Provision at Castle Matrix

Towards the west end of the north side of the third storey main room of the tower is a set of steps leading down into a third storey room over the south end of the extension, which is otherwise of two storeys only. This room has a fireplace at the east end and closeby a section of panelling is hinged to allow a discreet means of access to a small and oddly shaped room set towards the eastern end of the northern haunch (slope) of the lower vault. The small room is lighted by a slit angled north-eastwards to a loop in the tower north wall. Although it currently has no sign of a chute or drain, it is possible that this room was a garderobe. If this chamber was a garderobe then this adds extra support to the suggestion that the entrance doorway has been moved, since the outlet at the bottom of the chute would have been right alongside the west side of the existing doorway.

The chamber is not shown on a 20th century set of plans supplied by the current owner to the author. The chamber must surely have once been reached from within the tower since a feature of this kind only accessible from an adjoining building is an unlikely arrangement in Ireland, although the English county of Cumbria does have several towers (most of them 14th century) which are integrated with contemporary adjoining hall blocks. It is unclear how the chamber was originally accessed, but there may have been steps down to it from the east side of a north-facing window embrasure of the third storey main room. This window embrasure was subsequently replaced by the fireplace now in that position, such an alteration being common in Irish towers of this particular type, as at Gleninagh in Co Clare and Clara in Co Kilkenny.

Accessing a garderobe chamber by steps would have been difficult in the middle of the night because of the dark so such rooms are not normally positioned in the haunch of a vault. Another possibility is that originally there was a passage lying between the upper part of the second storey main room and a subsidiary room at that level. This would have allowed direct access to the presumed latrine from the spiral staircase, thus making it accessible to everyone in the tower (even if negotiating steps was needed). Most late 15th century towers in Co Clare with a tier of subsidiary chambers over the entrance have a garderobe accessed in this manner.

Although the author now considers that the angle-loop in the tower south-west corner opened directly onto the third storey main room by means of a long embrasure, there remains the possibility that his original assumption of this angle-loop having lighted a latrine chamber was correct. If so, such a chamber would have probably been L-shaped and reached by a doorway opening either off an adjacent side of the western or southern window embrasure, or between one of those embrasures and the intervening corner. The space behind the angle-loop here was probably filled in to help support the weight of extra material added to the top of the upper vault to make it a water-tight roof (probably in the 17th century), this being a function the vault wasn’t originally intended to serve.

Chapter XVI

The Parapet and Corbelling on the Tower

Although the embattled parapet on the tower with its double-stepped merlons looks authentic to an untrained eye from below, the low sills of the crenels set hardly above roof level betray it as being most likely of the 1830s. However the south-west and north-east corners project out on corbels that look authentic. The five corbels at the north-east corner are the more common type of blocks each extending further out and having the exposed outer ledge edge rounded, although there is some variation in their individual form. The corbels at the south-west corner are much taller and have an inverted pyramidal form with the edges chamfered, this form being distinctively Irish. There is nothing like them in England or Scotland. Corbels of the latter type appear on over fifty Irish buildings dating between c1460 and the 1640s and there is no evidence of the feature ever being revived later. Such corbels probably first appeared in Co Tipperary, where (and in Co Kilkenny) they mostly support box machicolations, although two of the earliest ones respectively support a garderobe overhanging a wall (Fethard), and arcading under a parapet on an abbey church (Holy Cross).

Other examples of pyramidal corbels in Co Offaly, and further west in counties Galway, Clare, Limerick, Cork and Kerry, mostly support corner bartizans on buildings of c1550–1630, or older ones heavily modified in that period. Also, the few instances of where the two types of corbel are found together in the same structure again all appear to be later than the 1550s.

Many of the corner bartizans supported by pyramidal corbels are set lower down than the level of the battlements, a feature specific to Irish buildings, and usually found in towers and fortified houses with other features regarded as quite late, such as mural fireplaces, roofs with vertical gables, gunloops and Elizabethan-style windows. There is a slight possibility that the pyramidal corbels at Castle Matrix are an original feature which once supported a roofed corner bartizan containing a chamber opening off the now-lost topmost room, as at Ballyallinan not far away to the south, which appears to be mid 16th century. However it does not appear that bartizans of that type appeared earlier than c1540, and Castle Matrix is almost certainly at least forty years older than that.

A possible explanation of the use of two types of corbels is that c1620 the Southwells, after removing the original top storey, had a new parapet at a lower level put on the tower at Castle Matrix, and that this parapet had the more common type of open bartizans set at diagonally-opposed corners. The combination of two different corbel types and the irregularities between the bracket corbels on the north-east corner might be explained by the use of left-over materials found in a builders’ yard in the city of Limerick, where there still remains a ruined merchant’s house adorned with a row of several corbels of the inverted pyramidal type. If this theory is correct, this second parapet at Castle Matrix, with the exception of the corbels, may only have lasted twenty or thirty years until being knocked off sometime in the 1640s or 1650s, and then replaced in the late 1830s.

Chapter XVII

Who Built Castle Matrix?

As already explained there must have been a manor house and outbuildings somewhere on the estate since the early 13th century, so the tower may have been added to a pre-existing group of buildings. Lime mortar takes a long time to set, so the construction of a tower like that at Castle Matrix would have taken a minimum of three building seasons, most likely four or even five, since each of the two vaults would have probably needed to set over winter before the shuttering was removed and work proceeded upwards. Several of the tower’s features suggest that it is no earlier than the 1460s. Tradition accords Thomas, 7th Earl of Desmond, the honour of having built the tower, although there is in fact no actual contemporary written record of its construction. Thomas succeeded his father in early 1463 and ruled for five years until in February 1468, whilst attending a meeting of the Irish Parliament at Drogheda, he was arrested by John Tiptoft, an over-zealous Lord Deputy recently sent over from England. His execution following a trial and conviction on a charge of coign and livery shocked Irish society, amongst whom the 7th earl was highly regarded. Coign and livery was a form of maintenance for armed retainers common amongst the Gaelic-speaking parts of Ireland at that time.

However, the gunport high on the south side, which appears to be contemporary with the walling containing it, suggests a construction date later in the 15th century, or even the first decade of the 16th century. Unfortunately none of the few other gunports in Ireland with which it can be compared can be closely dated. Small cannon were in use elsewhere in Europe by the mid 14th century, and may have been deployed at Carlow in that period but there are no further references in an Irish context until the late 15th century. In other words we wouldn’t really expect buildings in Ireland to be fitted with gunports or gunloops before the last fifteen or twenty years of the 15th century. This would tend to suggest that the builder of the tower at Castle Matrix was Maurice FitzThomas, who ruled as 9th Earl from 1487 to 1520. Maurice was on the attacking rebel side when the city of Waterford was unsuccessfully besieged in 1495 and would have seen the effectiveness of the defenders’ cannon. Since the gunport at Castle Matrix is quite high up it is just possible that the lower parts of the tower, ie up to and including the lower vault, were actually built by his elder brother, James FitzThomas, 8th Earl just prior to his murder in 1487, and that Maurice simply completed the building. Either of the 8th or 9th earls may have decided that although the nearby towers at Lisnacullia and Cappagh then newly built or under construction were not a military threat because they were held by allies, yet the earls’ own manor of Rathkeale should also have the dignity and security of possessing a stone tower. The murder of his elder brother and the implications of rebellion against the earls’ authority in the locality of Rathkeale might have acted as an additional motive for Maurice having built the tower c1490.

Chapter XVIII

How Defensible Was Castle Matrix?

Irish tower houses like Castle Matrix have long been regarded as castles, with the military functions that implied. However castles are now regarded as having multiple and complex functions and were not always as military in purpose as once thought. The native Irish didn’t tend to conduct the sort of sieges where attackers camped up outside and spent weeks undermining or battering fortress walls. Towers without bawns were never designed to withstand that sort of assault. Hardly any of these towers had an internal water supply. Most tower houses only needed to protect those sleeping within from a night raid, usually with stealing animals as its main aim. Throughout the 16th and 17th centuries strong outer defences were necessary for any Irish castle to resist attack for more than a few days.

Experiments have demonstrated that oak doors secured with bars could be shattered with a rudimentary ram in less than an hour, although the participants in such experiments were not being distracted or harmed by missiles from defenders. In any case many Irish tower entrance doorways were fitted with a hinged grille called a yett. These opened outwards which could be interpreted as a sign of weakness and might suggest towers were designed to be secured from outside, but the author quite recently saw a demonstration by a tower house restorer of how a yett could actually be closed and secured from within the building by means of chains and levers. Interestingly, no contemporary account of the capture of an Irish tower by English forces in the late 16th century mentions a yett.

Machicolations would also have helped in the defence of a tower although they are not as common a feature as might be expected. In contrast to the situation in England and Scotland, machicolations seem to have become increasingly popular in Ireland as the 16th century wore on and were sometimes added when towers were remodelled in the period c1580–1620. It has already been suggested that bartizans were added at Castle Matrix towards the end of that period. Machicolations were still occasionally being provided on new houses in Ireland up until the conflicts of the 1640s, which tends to confirm Ireland’s permanently unsettled state.

Irish towers like Castle Matrix often have the staircase terminating at the level of the topmost room, from which there is a separate staircase up to the battlements. Because the topmost room was an area to which only a select few were admitted, this implies that up until the late 16th century (when towers began to be built with a single staircase direct from the doorway to the roof) only high status members of the household and their guests normally had access to the battlements. This suggests that sentries on tower house wall-walks were only posted during times of anticipated unrest and not as a matter of routine every night. In general most tower houses could be defended overnight but were not fortresses intended to house permanent garrisons. Like the machicolations, the gunport at Castle Matrix should be regarded mainly as a deterrent and not evidence of military functions normally taking precedence over domestic ones.

Further Reading

An Examination of the Architectural Features of the Tower Houses in the Gowran Barony, Samantha Morris, The Old Kilkenny Review, 2018.

Castles Restored From Ruins, Mike Salter, 2018, particularly gazetteer entries for Ballybur, Ballyportry, Ballytarsna, Caherkinmonwee, Claregalway, Clonbrock, Moyode, Newtown and Tullaun.

Cross-cultural occurrences of mutations in Tower House Architecture, Rory Sherlock, Journal of Irish Archaeology, Vol XV, 2006, pp73–91.

Detecting Privacy and Private Space in the Irish Tower House, article by Gillian Eady, in Chateau Gaillard number 24, pages 69–75, 2010.

Early Guns in Ireland, S. de hOir, article in Irish Sword, 1983.

Irish Castles and Castellated Houses, Harold Leask, 1941.

Irish Tower houses, a Tipperary case study, C.T. Cairns, 1987.

Medieval Castles of Ireland, Mike Salter, 2024, particularly pages 50–125.

Proceedings of the Royal Irish Academy: articles by Thomas Westropp on the castles of Co Clare (1900) and Co Limerick (1907).

Space in Tower Houses in SE Kilkenny, thesis, John K. McCarthy, 2007.

The Castles of Connacht, Mike Salter, 2003.

The Castles of Leinster, Mike Salter, 2004.

The Castles of North Munster, Mike Salter, 2004.

The Castles of South Munster, Mike Salter, 2003.

The Castles of Ulster, Mike Salter, 2003.

The Medieval Castles of Ireland, David Sweetman, 1999.

The Other Clare: numerous articles by Richard Cronin and Martin Breen about tower houses and their owners in Co Clare.

The Tower House in Ireland, origins, dating, function, Dave Swift, 2005.

The Tower Houses of Co Limerick, PhD thesis, C. Donnelly, 1995. See also the same author’s Typological Study of Tower Houses in Co Limerick, 1999.

The Tower Houses of Co Wexford, Alfred Jordan, 1990.

The Tower Houses of West Cork, PhD thesis, Mark Ratcliffe Salmon, 1998.

The theses mentioned above can be viewed online.

See also the Historic Environment Viewer website for monuments in Ireland.

Acknowledgements

Firstly a very big thanks to Bob Tiernan, current owner of Castle Matrix, for whom this report has been prepared. Thanks are also due to former owner Elizabeth O’Driscoll for granting access to the castle in 2004, to Tim Robinson, for help with transport and accommodation on my 2026 visit to the castle and making useful comments on a draft of this account, to John and Margaret Carignan for also helping provide accommodation during the same trip, to Val Webb for proof-reading this account, and to Sarah McCutcheon, Limerick County Archaeologist, for enabling an internal inspection of Bouchiers Castle at Lough Gur, which helped to shed considerable light on some of the features of Castle Matrix because of similarities between these towers.

Mike Salter, Malvern, Worcestershire, July 2026

· · ·