This is an archaeology website dedicated primarily to the medieval castle of Dunamase in County Laois, Ireland, but I will add information about other archaeological sites in which I have been involved. It contains the complete text of a report on work at Nenagh Castle. If you go to the very end and scroll backwards you will see that I put short summaries of all the work I am involved in and update it regularly.

I would be grateful if anyone who knows of any references to Dunamase that were written in the 17th century, or describe the rebuild in the 18th century, could email them to me at bjhodkinson@eircom.net . Many of the later histories describe in detail events in the 1640's and 50's but I can find no contemporary references to these events.


NB. for regular visitors. During my last update there seems to have been an error in transmission and some material has got axed, I dont know why.

1) an essay on recent archaeological work at Dunamase. Now published in Kenyon & O'Conor (eds) 2003, "The Medieval Castle in Ireland and Wales". ISBN 1-85182-726-9

2) a history of Dunamase, with references. (not there yet, but watch for the new Co. Laois History/ Archaeology Journal due out in February 2003)

3) Summaries of excavation works 2002 (Didn't do too well on keeping this bit up to date)



At the very end of this there is a copy of the first draft of the final report on excavations at Cormacs Chapel, Cashel. It's warts and all and subject to some changes. It is text only and does not include specialist reports. But it may be of use to somebody out there.


1) A Summary of Recent Work at the Rock of Dunamase, Co. Laois, Ireland

1. Introduction.

The last few years have seen a spate of articles about Dunamase, most of which have come into print between 1993 and 1997 whilst the present writer was conducting excavations on the site on behalf of what is now Duchas; The Heritage Service. Two of these articles, by McNeill (1993) and O'Conor (1996), use pieces of fieldwork as the starting point for a discussion of aspects of the castle and its history, while Delany's booklet (1996) uses the occasion of a reprint of O'Leary's (1909) article as an excuse to present a more updated view of the castle. The Urban Survey of Co. Laois (Bradley, no date) contains a description and history while the present writer has published an article outlining the first two seasons' work at the castle (Hodkinson, 1995).

The excavations have, of course, brought forward new evidence about Dunamase which means that all of the above works are outdated to various degrees. The purpose of this present paper is, therefore, to provide an updated overview of the interpretation and phasing of the castle, utilising both the excavated evidence and standing fabric and, also, to try and set Dunamase in its historical context. It is not intended as the definitive final word on the subject but as a platform for further debate. The historic references are so few that there are several possible interpretations for the dating of the early phases of the castle and no "correct" solution will ever be forthcoming. A best solution on the balance of probabilities is the most that can be offered.

As it presents itself today, Dunamase consists of four basic elements, upper and lower masonry wards plus inner masonry and outer earthwork barbicans. These are layered from the summit of the Rock down the main approach route from the south-east, to present defence in depth along the vulnerable approach route. The summit required little more in the way of defence than a simple wall following the edge of the Rock, with a postern gate on the south side.

The excavations have concentrated on three areas, the gate of the inner barbican, the gatehouse leading from the inner barbican to the lower ward together with sections of the curtain wall to either side, and the building at the top of the hill which has usually been referred to as the keep but is in fact a hall.

2. Description and phasing of the castle.

2.1. The Early Christian Period.

Dunamase originated not as an Anglo-Norman castle but as Dun Masc, an Early Christian dun. Its first and only mention in the annals is for the year 843/4 when it was plundered by the Vikings and the abbot of Terryglass killed in the raid (AI, sub anno 844, AFM, sub anno 843). It should be noted here that there have been no finds to date to suggest that the site was occupied in Prehistoric times.

The walls of this early dun survive on the south-east approach to either side of the surviving gatehouse of the lower ward. The best preserved section lies to the north where there are two dry-stone walls; one, more or less on line with the castle curtain, resting on the back of an earlier one which is 2-3 m east of the curtain. The later gatehouse cuts across both wall lines and has removed all trace, but south of the gatehouse the remains of the outer wall can be seen as a series of larger stones at the base of the curtain and the inner one was probably represented by a row of stones on the inside of the curtain. North of the gatehouse, the outer wall stands to a height of at least 1m (base level of the wall was not uncovered) and is composed of some massive limestone blocks rising the full height with smaller stones interspersed between. It is unclear if this outer wall formed a revetting of the slope or whether it was a free standing wall, with inner edge.

The later wall stands to a height of c 50cms and is made of uncoursed limestone rubble. In the northern section the curtain wall of the castle stands on top of this later wall and it is impossible to gauge whether it functioned as a revetting wall or was free standing. To the south, if the large stones on the inside are correctly interpreted as the wall line, then it would seem that it was a revetting wall as there is only a single line of stones with bank behind.

The entrance to the dun lies c 7m north of the gatehouse, where there is a c 2.5m wide opening through both walls. In the passageway the subsoil on the southern side was the bedrock while the northern part was clay. A post hole in the centre of the passage cut into the clay next to the bedrock, and a cobble surface on the northern half which appeared to run under the end of the earlier of the two drystone walls can be put together with an edge noted in the area inside the wall to suggest the possibility of an earlier pre-wall phase with a slightly different approach line.

Immediately inside of the walls there were a number of post and stake-holes, associated with both phases, but none of these made any clear pattern. There was, however, a bank of earth separating an upper and lower phase, and the top of the bank was roughly at a level with the top of the inner wall, mirroring the situation south of the gatehouse. No direct stratigraphic link could be made from the bank to the wall because of the superimposed castle walls. The lower phase of features is assumed to go with the earlier wall (and possible earlier phase) and the upper with the later wall. A feature cut into the top of the bank and largely hidden by the curtain wall, could be interpreted as the start of a palisade trench running northwards. One side of a possible palisade trench survives between the southern end of the north section of the walls and the later gatehouse, as a step down immediately in front of the later curtain. South of the gatehouse, however there is no such feature inside the supposed line of the upper wall.

Excavation of other areas on the inside of the wall line revealed probable Early Christian features but there was no clear structure to them. At the top of the hill some pre-hall layers probably belong to this phase though there were no features or finds of note to date them.

Two stratified pieces of decorated metalwork from near the entrance to the dun have been dated by Raghnall O'Floinn (personal comment) to the 9th century. One, a small stud, was found in the bank against the later wall and the other, a horse trapping?, came from a slightly later pre-castle deposit. An unstratified silver penny of Ecgberht of Wessex (802-839) also fits in with this group. It was found on the surface within two metres of the other two pieces at the start the 1995 season. The 1994 season had stopped at what was (correctly) believed to be the Early Christian/ Norman interface and it is, therefore, virtually certain that the coin eroded out of the surface during the winter break.

2.2. The early phases of the castle.

The arrival of the Normans in the late twelfth century saw a refortification of the Rock. The two surviving elements which belong to this early phase are the hall at the top of the hill and an early gate-tower into the lower ward. The latter is dealt with first because it is this writer's feeling, and for which there is only slight proof, that if one of these elements was built before the other then it is was the gate-tower.

The gate-tower.

The gateway consisted of a rectangular tower c 10m by 8m with its long axis running parallel with the earlier walls. It was placed directly over the entrance of the dun with its outer wall more or less flush with the later of the two dun walls. The building today survives to a height of between 1 and 3m, but it is clear, from a straight join with the later curtain wall, that it rose at least to the later curtain wall-walk level. The defence of the building seems to have been solely from the wall head. This building henceforth is referred to as the gate-tower to distinguish it from the later gatehouse.

The tower had simple openings through the centre of each long wall; the outer one, with visible jambs, at c.2.2m being slightly wider than the inner at c2m. There are no indications on the interior of a first floor level, nor are there any surviving windows. Access to the first floor is believed to have been external timber stair. The evidence for this is threefold; firstly the square upright void of a rotted or removed timber at the north-west corner, preserved in a batter added against the tower; secondly the presence of a possible corbel stone on the outside of the tower between the void and the doorway, (unfortunately this stone, at surviving wall top level, slipped or was kicked out one winter); and thirdly the fact that the top level of the added batter on the west wall seems to have been dictated by the corbel, whereas the top level is higher once it turns the corner to the north wall. One final feature of the tower to mention is an external render, which survived to a height of c. 1m between the inner door and the end of the later batter, and which still survives in part sandwiched between tower wall and batter.

The existing curtain butts up against the gate-tower on either side and is clearly of later build. There are no contemporary masonry remains associated with the gate-tower, so either it was a free standing tower astride the old entrance to the dun utilising the low dun walls as the sole perimeter defence, or there was an associated timber palisade. The physical evidence for a palisade is none existent because the later masonry curtain sits on what would have been its line but from knowledge of Norman military engineering it is clear that there had to have been an associated palisade which stretched at least across the length of the south-east approach and probably surrounded the summit.

The hall.

At this early phase of the castle there is no evidence for a division into an upper and lower ward, so the hall building was free standing within a single enclosure. The building appears to be unique within an Irish context as an example of a ground floor hall with a two storey solar block at the northern end set transversely to the hall. The full external dimensions of the building are 35.5m north-south by 20m, east-west, with walls 2.8m thick; internally the hall measures 21m by 14.4m.

There are original opposed entrances in the east and west walls, giving a screens passage along the inside of the hall/solar division. The main formal door is believed to be that in the west wall because it is slightly wider, while the east door is interpreted as more of a service entrance. Around the west door it is possible to identify two hinge positions on the northern side and the drawbar hole on the south, but the details of the east door are obscured by a later blocking.

Prior to excavation it was possible to state that the hall end of the building was lit by a light in an embrasure in the west wall of the building towards the southern end. Two further embrasures have been revealed by excavation, one in the middle of the south wall and one in the east wall, more or less opposite that in the west wall. The eastern embrasure was modified in post medieval times by extending it downwards and knocking it outwards to become a door.

The hall had a suspended wooden floor sitting on joists which rested on the inside of a plinth which can be traced around the inside of the whole building. The starting point for this plinth appears to be the highest surviving point on the natural bedrock, near the outside of the west door, with the remainder of the outline of the hall built up to this level . The main wall rises from this with the joists resting on a narrow offset.

The western embrasure had a stone arrangement built up from ground level in front of it. This had a narrow chute at the northern end with, it is thought, a similar chute at the destroyed southern end. This arrangement was clearly designed to transfer something from above to below floor level. It is therefore interpreted as the waste chute for a water cistern within the embrasure. There was a similar though much less well preserved example inside the embrasure in the southern wall.

Analysis of the standing structure reveals something of the roofing arrangement. A gable at the east end of the solar shows its ridge ran east-west, while details in the surviving section of the hall's west wall show its ridge to have run north south.

The hall roof is an interesting arrangement, paralleled this writer has been told at Castle Rising (pers. comm. by several members of the Castles Study Group on visit to Dunamase in 1998). The wall has been raised to a level approximately 1m below the wall-walk and then tie beams have been laid across the hall. These beams run through the thickness of the wall and so could possibly have held a hoarding, at least, over the approach to the main door. A wall plate was then laid across the tie beams and the rafters set into the wall plate; the whole wall was then raised up to wall-walk level through the skeleton of the roof, before the roof was finished. There is no clear evidence as to the material used for the roof except to state that no ceramic or stone tiles have been found. There was a small lead smelting furnace, inside the hall at the building level but this is believed to be associated with the water cisterns rather than producing the roofing material. A wooden shingle roof seems the most likely option.

The inside of the hall has only been partially excavated and so some of the detail of the internal arrangement is unclear. The span of the hall is so great that there may have been some form of aisling but none has yet been found. The location of the fireplace also remains to be discovered.

One final feature of the hall is the stairs up to the wall-walk which start directly above the west door. It is somewhat affected by later works but is believed to have been accessed from a wooden gallery over the screens passage which would also have given access to the solar. At wall-walk level the stairs rise to a door of which the jamb and drawbar hole of one side remain.

The solar end of the building is much affected by later works. The whole of the north-west corner, with its late medieval style doorway and intra-mural stair is a post-medieval rebuild, while much of the ground floor has been refaced both internally and externally. It would seem that two openings in the north wall at ground floor level may be original lights which have been broken out and enlarged. Otherwise there is one blocked embrasure at first floor level and another in the east wall, both of which are original. The present north-south wall dividing the ground floor of is a later build as is much of the present cross wall between hall and solar. The original ground floor plan of this end of the building is, therefore, unclear.

The added batters.

Both the gate- tower and the hall have had stone batters added against the base of the walls, presumably to add to their defensive capabilities.

Outside the gate-tower the batter starts to the north of the inner door, runs around the north-west corner where, as noted above, it was built around a timber upright, and along the north wall to disappear in under the later curtain. On the east side of the tower north of the entrance, the early wall which is slightly east of the tower, has been cut back slightly and stones added to the top to give a batter effect, while south of the opening there is a proper batter which turns the south-east corner to disappear in under the curtain wall. On the inside of the curtain there is no batter along the south wall nor along the west as far as the door. It is felt that the batter has been removed from these walls at the time of the subsequent rebuild to improve access around the new arrangement.

On the hall building the batter starts flush with the north side of the west door and runs around the northern end of the building to stop just short of the east door. It then picks up on the other side of the door to run to a point just short of the south-west corner. There is no batter from here to the main door. This south west corner is in very poor condition and it is felt likely that the batter did originally turn the corner and run some way along the west wall. There is, however, evidence to suggest that it would have stopped some way short of the door. Both doorways in the building have later fore-buildings or porches added to them and the walls of these rise up the batter except for at the south side of the west door, where the forebuilding directly abuts the main hall wall. It could be argued that the batter in this area had been removed when the forebuilding was added, but there is one other consideration to take into account. The main formal door is sited on a slight outcrop of rock which makes access from the south the only option and had the batter continued into the south side of the door then it would have greatly narrowed the access space.

There is a clear time interval between the building of the gate-tower and the addition of its batter so the two do not appear to planned as one unit. One indication of this is the layer of render between the batter and the tower wall, and another is the rather ad hoc arrangement along the east side. With the hall building, however, there is some indication that the batter may have been planned from the outset. No render has been noted between the batter and the wall. Also, a slight quarrying back of the hillside which seems to have taken place at the south-eastern corner of the tower, stops roughly in line with the outer edge of the batter whereas one might expect the quarrying to extend into the hall wall line if the batter not been planned at that stage.

If one assumes that the batters are contemporary with each other, then the possibility arises that the gate-tower predates the hall building. There is also a reasonable lapse of time between the addition of the batter to the gate-tower and the early 13th century remodelling of the castle. On the north side of the tower there was a build-up of over 30cm of deposits against the batter, separating the build level of the batter and the construction cut for the early 13th century re-modelling. While it is difficult to assess the rate of accumulation, because the main access route to the hall swung northwards from the gate-tower causing erosion downhill, we probably have to think in terms of at least a decade or two rather than in single-figures of years for this accumulation.

2.3. The early 13th century rebuild.

In the early part of the 13th century the defences were remodelled entirely in stone using a more aggressive form of defence. The old gate tower was blocked up and a new gatehouse built to the south projecting forward of the new curtain line to provide depth of defences along the gate passage as well as permitting flanking fire along the base of the curtain. This phase was built mostly in uncut limestone, but the one feature which does distinguish it from the previous and subsequent phases is the use of cut sandstone around loops, portcullis slot, doorways etc.

The emphasis of this new phase is on archery. Besides the wall-walk there is a lower tier of embrasures, entered at ground level and equipped with plunging loops, of which only one survives in a more or less intact state, just south of the new gatehouse. In all there are twelve loops across the south-east approach, with a new south-east corner tower to give flanking fire. To maintain the regular spacing of these loops one was incorporated into the blocking of the gate-tower and another seems to have been knocked through the tower wall to the north. In order to disguise the fact that there was a blocked gate, a new section of batter was added in front of the blocking, connecting the ends of the earlier one. Square holes in the section of the curtain between gatehouse and gate-tower, suggest that there may have been an external hoarding along the northern stretch of wall, but there is no evidence for such on the south side of the gatehouse.

Because the front of the gatehouse was largely demolished, there had been some speculation, prior to excavation, as to the form of the gatehouse. It can now be seen to be square fronted with two narrow vaulted chambers flanking a central passage. Each of the chambers has two loops in its north and south wall and one firing straight out. The chambers are so narrow that the front three loops must have been serviced by a single archer.

At first floor level there was a single large room, entered from the south-west corner. The absence of vaulting over the passage, indicates that there was a plank floor between front and back arch, which presumably could have been lifted in a crisis to provide a fighting platform above and either side of the central passage. The stub ends of the north and south walls preserve the remains of embrasures, and there is enough space to the missing front of the building for two more. The room is also equipped with the only visible fireplace in the whole castle. This is in the north wall and sits awkwardly in relation to the embrasure, so it is believed to be a later insertion.

Within the gate passage there is depth of defence. The first obstacle is the portcullis, the front of which is protected by inward facing loops from the side chambers and probably by a murder hole above. The second barrier is the main gate set just over half way up the passage and marked by an offset in the walls. It is approached along what is effectively a chasm with fighting platforms above on either side. There is also an the inward facing loops in each of the flanking chambers which are offset from one another to prevent accidentally shooting someone on one's own side. The entrances to the chambers lie behind the gate and that on the south side is wider than that on the north, presumably to give a shelter for the guard or porter. Behind the main gate there are back projections inside the curtain which prolong the passage and presumably the fighting platform on either side, while there may have been a second gate closing off the end of the passage. There is no surviving evidence for this second gate as neither of the back projections survive to the likely hinge level.

As noted the old gate tower had its eastern opening blocked and at the same time the inside opening was narrowed down by the insertion of stone piers against either side. Within the building the ground was levelled upwards with a thick deposit of sandstone chips, which suggests that it may have been used as the mason's workshop during the rebuild. Subsequently it appears to have been a smithy as the sandstone is covered with iron scale and slaggy material. There is, however, no furnace surviving but this may have been in the south-west corner where there is a large hole cut from the level at which the smithy fell out of use. An alternative interpretation of the sandstone is that it was deliberately used to level up with to give a solid fire-proof floor to the smithy.

Elsewhere the entire circuit of the hill top was enclosed with the stone curtain. The postern gate, on the basis of the use of sandstone for jambs and loops, is contemporary and a garderobe tower projecting over the cliff edge on the west is also believed to be of this date. It was probably at this time that the large inner ward was subdivided into upper and lower wards by the addition of cross-walls running from the hall building to the north and south curtain. Both these walls can be seen to butt up against the hall batter and thereby post-date it, but there are no architectural features currently visible which by the use or non-use of sandstone could indicate phasing. The southern wall is continuous to the postern gate, while the northern section has gaps, showing that the entrance into the upper ward must have been on this northern side.

As noted above the hall building was poorly defended. The choice of a ground-floor hall with two virtually undefended doors was either an arrogant statement of supremacy, or plain stupidity in what was effectively a Marcher lordship. This weakness was recognised and addressed during the rebuild with the provision of forebuildings, or porches against both doors. That around the west door is equipped with two sandstone loops in the west wall and one in the north, while the door is in the south wall. A pivot stone survives at the base of the west side of the door, but it is not clear if there was a drawbar behind. There is no indication of the original height of the building, but it need not have had a second storey. A drain hole through the west wall suggests that it was not roofed.

Very little survives of the eastern forebuilding. The doorway here was on the northern side and the only section surviving to any height was the west jamb and wall running into the batter. There is however a drawbar hole running into a hole cut into the batter.

2.4 The barbicans.

The inner barbican

The inner barbican is a triangular enclosure with an entrance at the apex, through a D-shaped tower. In style it continues the theme of defence by archery noted in the rebuild phase, with plunging loops in embrasures under the wall walk. The wall is continuous from under the south-east angle tower to the gate and from the gate north-westwards there is a short section with three ground level loops, beyond which point the wall is missing.. The circuit is then completed by a much narrower wall running at right angle to the slope and abutting the inner curtain. This last section is effectively little more than a screen wall but is so contrived to join the curtain just beside a loop to allow flanking fire along its length..

The gatehouse lacks its rear wall and part of its western wall and has undergone a number of modifications which have disguised its true nature. The front of the gateway has the remains of a two slot counter-weight drawbridge, which is not now visible. The current Office of Public Works wooden stile, which stands below the inside edge of the front arch, is conveniently placed for descriptive purposes. The back of the drawbridge pit coincides more or less with the stile, while the top of the stile is roughly at the original road level. The drawbridge must have opened onto a pier or trestle bridge to cross the dry moat outside. When the woodwork had been removed from the gate, access to the interior would have been exceedingly difficult, so at some undetermined time someone has effectively bulldozed down through the gate pushing material into the moat to create the present day pathway. Subsequent refacing within the gateway has partially obscured the scarring this caused. The only other feature of the entrance is an alcove on the west side, presumably for a guard or porter.

At first floor level there is a single chamber with three loops, one of which has access to a murder hole in the arch over the gate. The stub of the west wall preserves one side of a door giving onto the wall walk running off south-westwards

It has been argued that the barbican and curtain form part of one overall plan and are more or less contemporary (O'Conor 1996, 105) but, in this writer's opinion, this is not so. While the overall design concept, i.e. defence by archery at two levels, is the same, there are two hard pieces of evidence to suggest a chronological difference. The first of these is that the screen wall section of the barbican butts up against the curtain and is not tied into it, while the second is that there is no use of sandstone in the barbican. The form of the two gateways is also completely different, one rounded the other rectangular. The gatehouse is a step along the way to the classic twin towered gateway, with chambers flanking a central passage while the barbican is simply a tower with central passage. The choice of a round fronted tower is also somewhat strange because equipping it with a drawbridge creates difficulties which would not arise with a square tower. To quote McNeill on the barbican gate (1993, 238), "Dunamase reflects not the latest design for castle gates, but the device of an inexperienced mason", which is in marked contrast to the quality of the gatehouse.

Unfortunately there were no finds from the excavation of the barbican which can be used to date its period of construction. In this writer's earlier article (Hodkinson 1995, 20), a dating to the middle or later 13th century was suggested, based partly on his understanding of the development of the drawbridge. On the basis of his work at Nenagh Castle, where a drawbridge pit of the first quarter of the 13th century was uncovered (Hodkinson 1999), the writer is now prepared to accept that an earlier dating is possible, but still feels that there has to be a significant time lag between the building of the curtain and the barbican.

The outer barbican.

It was O'Conor's survey of the outer barbican which formed the starting point of his paper on Dunamase and his detailed plan is reproduced in that work. It shows a more or less triangular enclosure, with massive ditches cut into the bedrock on two sides and the natural slope on the third. O'Conor has noted flattish areas along the inside of the ditch which he interprets as the base for timber defences. This writer accepts O'Conor's interpretation and refers the reader to his paper for the details. There is, however, one small detail which seems to be absent from the plan. On the rock outcrop below the south-east angle tower of the curtain and outside the inner barbican, there is a fragment of wall with another section to the east. This appears to be a screen wall across the ditch between the inner and outer barbicans, and as such it would have prevented a sudden rush for the gate along the base of the ditch by a group approaching around the side of the rock where there is something of a blind spot. The dating of the outer barbican is unclear.

2.5 The later medieval and post medieval periods.

The excavated evidence suggests that Dunamase was abandoned by the middle of the 14th century and this ties in with the historical sources which stop in the 1330's. Some histories suggest that it was taken over by the O'Mores, which may be true for the land but they did not use the buildings. There are no archaeological features or finds which can be used to suggest that the castle was inhabited in the later medieval period. If anything there is evidence of slow decay, best exampled by the inside of the gate tower, where there are two superimposed features cut into rubble. Both features are circular with a clay floor and because the lower one burnt it is possible to say that there was a wicker superstructure. A spread of charred grain suggest that the structures were used for storage, and given that they were in the corner of the deserted building they probably represent a cache hidden by a local farmer from raiding parties. There is no close dating for the structures and they could even be early Modern in date. They do however separate an upper and lower phase of rubble, showing that the destruction of the tower was not a single event. Indeed even within the upper phase of rubble it was possible to trace phases of collapse rather than a single event. Reference to later medieval occupation in various histories is incorrect.

The histories also relate a re-occupation in the middle of the 17th century, and a siege and destruction of the castle by the Cromwellians (e.g.O'Byrne, 1856); Cromwell's Lines, a ring-fort marked on the OS maps, is the site of the batteries in local tradition, and a nearby hill is called Hewson Hill, presumably after the Cromwellian General of that name. Again the excavated evidence does not support this; there are no cannon or musket balls to support a siege, no traceable living surfaces, no building phases, and while there is some 17th century pottery the amount is small and in the case of one small group comes from the upper part of the rubble. Furthermore the writer is conducting an ongoing trawl through contemporary sources and has yet to find a reference to Dunamase in the mid 17th century.

One cannot, however, dispute that Dunamase is a ruin and that much of the destruction has been done by gunpowder. The questions which cannot be answered with certainty are by whom and when? It is very likely that it was slighted in the Cromwellian period, given the wide-scale destruction of castles ordered in England in the same period. There is no evidence, however, to contradict a suggestion that it could be a hundred years earlier, or any time in the intervening 100years. The establishment of Fort Protector in what is now Portlaoise, may have given the impetus for destruction. A large castle in reasonable repair within four miles of the new fort could have been viewed as dangerous.

There are some architectural details of the hall which do suggest a late medieval/ early modern date, but close scrutiny reveals that these are all later insertions. Towards the end of the 18th century, Sir John Parnell attempted to restore the castle for use as a banquetting hall and brought in architectural details from other ruins, incorporating some into the building and leaving others lying around unused. There is a fine late medieval style door into the ground floor of the solar through the west wall, which on closer examination has different moulding on either side and is in fact part of two different doors. There is a window in the east wall of the solar area where part of the hood mould has been left out, and a corner angle light in the north-west corner way above head height. From the excavation of the hall a number of cut pieces of moulded limestone have been recovered which relate to nothing on the building.

Works which can be attributed to this Parnell phase are, the complete rebuild of the north-west corner and insertion of the aforementioned later medieval door; the blocking of the light inside the embrasure in the north wall of the solar at first floor level; the refacing of much of the interior of the ground floor of the solar as well as parts of the exterior; the insertion of the window with hood moulding in the east wall; the insertion of rectangular windows in the ground floor of the north wall which may have been smaller original lights knocked through, the subdivision of the ground floor by a wall containing fireplaces side by side facing into opposite rooms; a rebuild of the cross wall between hall and solar incorporating a stair at the west end partially blocking the original west door; and the blocking of both of the original doorways. The insertion of the windows in the north wall and the door in the west wall necessitated destruction of part of the original batter. In the case of the door it was necessary both to remove part of the batter and build up the ground outside of it. It is believed that the curving wall around the top of the cistern can be attributed to this phase, giving a broad sweep around the northern end of the hall and up to the new doorway.

The only feature which does not tie in neatly with this is a new door in the east wall formed by breaking downwards through the base of the window embrasure and outwards through the wall and batter. This door leads directly to a massive block of collapsed masonry leaning against the inside of the wall and blocking access. The explanation for this feature is probably a change of plan, because the doorway does not seem to have been finished. The southern side of the cut through the wall has been faced but there is no matching facing of the opposite side.

The Parnell phase also seems to have included the breaking up of some sections of collapsed masonry. While it is possible to take the existing sections and work out where they originally stood, there is a section of the west wall unaccounted for. It had been cleared away before another large section of masonry had landed on the base of the wall. This second piece probably came from the south wall which had slipped outward at the base and fallen inwards to leave an overhang, which was later broken off. It is also probable that the refacing inside the barbican gate-house and parts of the southern end of the curtain belong to this phase. A small lime kiln at the back of the gatehouse dates to the post-medieval phase.

3. The Finds.

The number of registered finds from the site stands currently at just over 5,500. This number, however, includes bags of animal bones which have been given a collective registration number so the actual number of finds is much greater.The site, being a dry limestone outcrop has produced no wooden, leather or plant materials other than charred remains. Apart from that there is a large range of finds which appear to reflect a largely masculine household on the Rock. The number of finds which one associates with traditional female activities e.g. spinning and weaving, are few.

There is a range of medieval pottery, mostly local wares, with imports from England, and more specifically the South-West dominating the imported wares. There is a small amount of French wares. A handful of potsherds have been converted into circular gaming pieces/ counters. Later medieval wares are absent and post-medieval wares are few, reflecting the disuse of the castle after c. 1350. There is a large amount of metalwork, most of which is in good condition. Among the ironwork it is possible to list weapons and armour, in the form of a range of arrowheads (studied by A. Halpin for his PhD thesis on medieval archery in Ireland), fragments of mail and an assortment of knives (some of which may have been tools rather than weapons); door furniture, in the form of fittings, locks and keys; and horse equipment e.g. shoes and nails, rowel spurs and various buckles. Copper alloy objects include two fragments of Early Christian metalwork, an assortment of stick and ring pins. There is one minute fragment of gold with filigree decoration which appears to have been a panel on an Early Christian object which was broken up in the 13th century. There are ten coins of medieval date, one of which is a 9th century penny of Ecgberth of Wessex, two are from Irish mints, (one of John as Lord of Ireland and one as king), five are English long and short cross pennies and, one is a Scottish halfpenny. The 10th is a jeton from S.E. France dating to the late 14th century which was found in the early destruction layers of the hall and this is the only object from the site which can be positively dated to the late 14th or early 15th centuries. Among the other materials there are dice as well as evidence for dice-making in the form of waste. These fit neatly with a stone nine-man morris board and the previously mentioned pottery counters/ gaming-pieces. Fragments of 10 skulls were found in the early destruction levels of the gate tower, and these are assumed to have once adorned the roof of this structure. Examination of the skulls by Lauren Buckley has shown that they were all beheadings.

4. A Brief Medieval History

The first contemporary reference to a castle at Dunamase appears to be two mentions in the Pipe Roll of 1211-12, where William Marshal is recorded as rendering 36 for the land on which the castle stands and Geoffrey Luttrell, who had custody of the castle at the time, is recorded as rendering 353-6s-8d for the Irish of Dunamase. The Histoire ( lines 14128-9), written at a slightly late date, records the handing over of Dunamase to William Marshal by Meyler FitzHenry in 1208. The castle is clearly older than any of these dates but its early history remains obscure. There are four people for whom cases have been, or can be, made as founder of the first castle, Strongbow, William Marshal, Meyler FitzHenry and Geoffrey de Costentin.

The argument for Strongbow as founder seems to rest solely on Orpen's belief that the castle was built between 1171 and 1176 ( Orpen 1911-20, i , 373-375; O'Conor 1996, 110-11). Strongbow was active in Laois (Song 1176-1255) and it is tempting to see the choice of the massive hall block as a harking back to the Great Tower at Chepstow Castle, heart of Strongbow's honour of Striguil in South Wales, but there is no solid evidence to put forward. William Marshal's case is also weak. In 1189 the lordship of Leinster passed to Marshal by virtue of his marriage to Isabella, the product of Strongbow's marriage to Aoife, daughter of Dermot McMurrough. It appears that Marshal did not receive full seisin of Leinster until c1194 (Flanagan 1989, 134) because John, Lord of Ireland, had arrogated it to his own use, while Crouch (1990, 92-107) has argued that he did not gain effective control over the lordship until 1208 because of the hostility of his sub-tenants, especially Meyler FitzHenry. Despite an earlier visit by Marshal to Ireland in 1200-01 (Crouch 1990, 80) there seems to have been little opportunity for him to have commenced the building of the castle.

In recent times a strong case has been made for Meyler FitzHenry, who was actually in possession of the castle in 1208. He was amongst the early Norman settlers and, in 1181, was deprived of his cantred of Carbury and given the province of Laois in exchange. In the words of Giraldus Cambrensis, "Thus they assigned this remote border area to a borderer and true son of Mars" (Exp., 195). Giraldus also records the building of a castle at Timahoe, c 5 miles from Dunamase (ibid.). O'Conor makes the case that this castle was probably the first to be built in the area and was intended as the caput of FitzHenry's lands. Timahoe's absence from the later record and FitzHenry's appearance at Dunamase is taken by O'Conor to indicate a transfer of the caput to Dunamase and that FitzHenry was thus responsible for its construction (1996, 111) . Flanagan, however, argues that although FitzHenry was granted Laois (1988, 236) he was in all probability only granted custody of Dunamase, its principal castle, and furthermore suggests this may be one background cause for the antipathy between Marshal and FitzHenry which manifested itself early in the next century. Her argument, however, presupposes the existence of a castle at Dunamase in 1181 and thus she is apparently following the school of though which has Strongbow as founder.

The fourth possibility, Geoffrey de Costentin, has been neglected in recent accounts of the castle and is only connected to Dunamase in peripheral literature. Geoffrey was recorded as holding Laois at the Council of Oxford in 1177 though the circumstances behind this are not recorded (Roger of Howden, i ,164). He is also recorded as being enfeoffed by Hugh de Lacy in Meath (Song, 3154-55). Flanagan (1988, 236) citing the grant to FitzHenry of 1181 then dismisses Geoffrey from the Laois equation, while O'Conor has suggested that the grant to Geoffrey was purely speculative (pers. comm.). There is, however, some evidence that Geoffrey was both active and had a longer interest than O'Conor or Flanagan would allow. Clark, in his introduction to the Register of Tristernenagh, a priory founded by Geoffrey, notes the place name Drumcostentin in Leighlin (1941, ix) and uses this to suggest that Geoffrey was active outside Meath and Connaught where he later held most of his lands. Clarke even suggests (ibid) that Geoffrey was appointed constable of Strongbow's new castle of Dunamase. As late as 1200 Geoffrey had an interest in Laois, when perhaps tellingly, he gave up all claim to Laois in favour of FitzHenry in exchange for a cantred in Connaught (CDI ,22 ). It is thus plausible that the 1200 deal saw the transfer of Dunamase from Geoffrey to FitzHenry who then transferred his caput there. Following this hypothesis, it is possible to argue that the remodelling of the castle was undertaken by FitzHenry in the period 1200-1208.

It was only in 1208 that Marshal gained full control of Dunamase and shortly after that in 1210, King John, took it back into his own hands where it remained until 1215 (CDI, No. 644). If the remodelling of the castle is truly Marshal work, as most of the authorities at the head of this paper seem to agree (and Knight, 84), then it has either to have taken place 1208-1210 or post 1215. The period of two years seems a little too short to plan and execute a major rebuilding programme and had the rebuild been on-going when it was taken into John's custody in 1210, we might expect to see some expenditure on it in the surviving Irish Pipe Roll of 1211-1212. In 1213 Marshal left for England, (Crouch , 110) where he subsequently became embroiled in the events surrounding Magna Carta and the subsequent minority of Henry III, so the later date appears less likely for the rebuild. Perhaps the argument for FitzHenry remodelling the castle is the stronger for this. If anything is Marshal work then it is likely to be the inner barbican, though it is possibly to the Younger rather than Elder William that we should look, because of the time lag between the remodelling and the addition of the barbican which was discussed above

After the death of William Marshal the Elder, the lordship of Leinster passed successively through the hands of his sons, William, Richard, Gilbert, Walter and Anselm all of whom died without issue. In 1234 Gilbert had been allowed to inherit from his brother Richard who was killed while in rebellion against the Crown, but the price that was paid included the surrender of Dunamase. It was restored later the same year (CPRH3, 48, 53 & 65) With the death of the youngest brother Anselm in 1245, the Lordship of Leinster was divided among the daughters of William the Elder and Dunamase fell to the youngest, Eve, who was married to William de Braose. William had died well before 1245 and his wife only survived until 1246, after the death of Anselm, but before the division was completed. Dunamase thus passed to Eve's eldest daughter Maud and by her marriage into the hands of the Mortimer family, which then held the castle into the 14th century. The year 1330 saw the downfall of the Mortimers when Roger, Earl of March and lover of Isabella, mother of Edward III, was executed for treason and his lands escheated to the crown. While the lands in question undoubtedly included Dunamase, what is not clear is whether the Mortimers were physically in possession of the castle at the time. Clyn, recording the death of Lysaght O'More under the year 1342, lists among the deceased's achievements the destruction of Roger Mortimer's castle of Dunamase and the usurpation of the lordship. This suggests that the castle was already in Irish hands before Mortimer's downfall in 1330, but in 1334 the crown granted the manor of Dunamase, "an escheat by the forfeiture of Roger de Mortuo Mari, Earl of March" to Fulke de la Freigne for 10 years rent free (CPR, 561). The following year Fulke complained that Dublin had not released it to him and on June 8th the King, commanded its release. (CCR, 401). This letter includes the added complication of a claim on Dunamase, by representatives of the Earl of Kildare, then a minor. Could it be that the grant of 1334 was somewhat speculative in that Fulke was expected to win the castle back before he could enjoy its use? This could possibly explain why it was granted rent free. Whatever the case, after the 1335 letter Dunamase seems to drop out of the royal record and no references of the later medieval period have been noted by the writer.


Printed Sources

"The Annals of Innisfallen" ed. MacAirt, S. Dublin 1988.

"The Annals of the Four Masters" ed, J. O'Donovan, Dublin 1856.

CCR "Calendar of Close Rolls" 1333-1337.

CDI " Calendar of Documents Relating to Ireland" vol 1 ed. H. Sweetman, 1875.

Clyn. "The Annals of Ireland by Friar John Clyn & Thady Dowling", ed. R. Butler. Dublin 1849

CPR. "Calendar of Patent Rolls", 1330-1334

CPRH3 "Calendar of Patent Rolls of Henry III", 1234-124

Exp. "Expugnatio Hibernica", by Giraldus Cambrensis, ed. by A.B. Scott and F..X. Martin, Dublin 1978.

Histoire. "L'Histoire de Guillaume le Marechal", ed. Meyer, P., 3 vols. Paris.1891-1901.

Pipe Roll. "The Irish Pipe Roll of 14 John, 1211-12."Ulster Journal of Archaeology Vol. IV, supplement, ed. O.Davies and D.B. Quin, 1941.

Roger of Howden. "Gesta regis Henrici secundi Benedicti abbatis"

W.Stubbs, Rolls series., London 1867.

Song."Song of Dermot and the Earl," ed.G.H. Orpen, 1892. Oxford.


Secondary Sources

Bradley, J. no date, "The Urban Archaeological Survey; Part VI. Co. Laois" (Limited distribution), Office of Public Works, Dublin.

Clarke, M.V. (ed) 1941 "Register of the Priory of the Blessed Virgin Mary at Tristernagh", Dublin.

Crouch, D, 1990 " William Marshal; Court, Career and Chivalry in the Angevin Empire 1147-1219", London and New York

Delany, D. and O'Leary, E. 1996 "The Rock of Dunamase", Vicarstown

Flanagan, M.T. 1988. "Henry II and the Kingdom of Ui Faelain", in "Settlement and Society in Medieval Ireland,; Studies presented to F.X. Martin, o.s.a. " ed. J. Bradley, Kilkenny.

Flanagan, M.T. 1989. "Irish Society, Anglo Norman Settlers, Angevin Kingship; Interactions in Ireland in the late 12th Century", Oxford.

Hodkinson, B.J. 1995 "The Rock of Dunamase". Archaeology Ireland Vol. 9 No. 2

Hodkinson, B.J. 1999 "Excavations in the Gatehouse of Nenagh Castle". Tipperary Historical Journal 1999, 162-182.

Knight, J.K. 1987. "The Road to Harlech; aspects of some early thirteenth-century Welsh Castles" in Kenyon J.R. and Avent R. (eds) Castles in Wales and the Marches: Essays in honour of D.J. Cathcart-King, Cardiff, 75-88.

McNeill, T.E. 1993 "The Outer Gate House at Dunamase Castle Co. Laois". Med. Arch, 37, 236-8.

O'Byrne, D. 1856 "The History of The Queen's County", Dublin.

O'Conor, K. 1993 "Dunamase Castle, Co. Laois". Journal of Irish Archaeology. 7, 97-115.

O'Leary, E. 1909 "The Rock of Dunamase". Journal of the Co. Kildare Archaeological Society. Vol . 7 No. 2.

Orpen, G.H. 1911-20 "Ireland under the Normans, 1169-1333", 4 vols., Oxford.



02E0035 Kilcolman, Co. Tipperary, Vicinity of Ringfort 18116/17816 RM TN020-043

Monitoring of the topsoil stripping for a new house and garage on a greenfield site revealed no features or deposits of archaeological interest.







Excavations in and around Cormac's Chapel were undertaken as part of the restoration programme and were financed by the OPW. There was a short season in 1992, followed by more extensive work in 1993. The excavation number for the site is 92E202. Short reports on the work have been published in Excavations Bulletin for 1992 and 1993 and a more detailed preliminary statement was printed in the Tipperary Historical Journal 1994.

Three areas were investigated and numbered in the order in which the trenches were opened. The enclosed area to the north of the chancel of Cormac's Chapel was Area 1. Area 2 lay outside the north door of the chapel and Area 3 comprised of the chancel, eastern part of the nave and the north tower of the chapel itself (fig. 1).

It was not possible to excavate the whole of any of the areas simultaneously. Throughout the excavation the exterior of the chapel was covered in scaffolding which could not be undermined, so Area 1 had to be excavated in three stages and Area 2 in two stages. The interior was excavated in two halves to allow one side of the chapel to be underpinned before the second half was excavated. Artificial lighting was necessary at all times in Area 3 and, in overcast weather, in the other two areas.

2. AREA 1.

Area 1 was enclosed on the north by the cathedral chancel and on the west and south by Cormac's chapel. In the eighteenth century the fourth side was closed off and the area converted into a sacristy for the cathedral. This new east wall was taken down to foundation level, by the OPW, in November 1991 and, in the course of this work, the base of a high cross was discovered incorporated into the fabric (Harbison, 1993). The foundations of this wall formed the eastern limit of the excavation. At the eastern end of the site, the cathedral foundations were found to overlie an earlier mortared wall, context 67, which is interpreted as an earlier church predating Cormac's chapel. The earliest burials in the area predate this earlier building and for this reason it is proposed to present the development of the graveyard prior to the description of the structures. The earlier church appears to have been erected at phase 3 of the graveyard, while both Cormac's chapel and the cathedral are associated with phase 5.

2.1. The early deposits and phase 1 of the graveyard (Fig 2).

Uneven fissured limestone bedrock formed the base layer over the whole area, and sloped slightly down from the north-east to the south-west. The earliest features, which appeared upon removal of the lowest layer, were rock-cut but there was no discernible pattern to them and no datable finds were retrieved from them.

Context 502 was a sub-circular feature, ca. 0.3m in diameter and 0.15m deep. It lay close to the wall of Cormac's Chapel at the east end of the area. Its single fill was 503 a sticky brown soil.

Context 504 lay to the west of 502. It was unclear whether this was a feature in its own right or a part of the foundation trench of the chapel.

Context 398 lay to the north of 504 and was sub-rectangular in form, measuring ca. 1.4m north-south by 0.5m east-west and up to 0.4m deep. Its fill, layer 399, was a sticky grey brown clay containing angular stones.

Context 396 was a sub-oval post-hole ca. 0.6m in diameter by a maximum 0.5m deep, lying to the north of 398. Its fill, layer 397, was a light brown soil which may have been part of the post-packing.

Context 500 lay to the west of 398 and was a sub-oval post-hole ca. 0.6m in diameter by ca. 0.3m deep. The exact point from which this feature was cut was unclear because a later pit, contexts 38 and 333, had cut down to the bedrock in this area. The fill of the feature, layer 501, was a sticky brown clay containing several large lumps of mortar, some of which had flat finished faces.

The earliest stratigraphy was very badly disturbed by the graveyard, which later covered the area, and survived only as scattered patches. The most extensive area lay in the north-west corner, in towards the foundations of the later Gothic Cathedral, where the deposits measured no more than ca. 0.35m thick. The upper part of the sequence of deposits described below could be contemporary with phase 2 of the graveyard.

Layer 314, and its equivalent layer 57, (henceforth = is used to denote a layer which was given two numbers) was a dark brown silty soil flecked with charcoal which directly overlay the bedrock. 314 contained a parallelepiped antler die. It was sealed by an uneven patchy cobbled surface, 312=58, which was in turn was sealed by a light brown clay soil, layer 305=390, which contained animal bone. In places the latter layer directly sealed the bedrock, suggesting that there had been either some wear or removal of the underlying layers.

Overlying layer 305=390 in the north-west corner was a very compact layer of white mortar, 40 which was partly sealed by a thin ash deposit 55=392. Further to the south west a spread of white lime and sand 369=99, overlay brown soil 305=390. This white layer partly sealed a grave, 391, cut through layer 390 into the bedrock. Immediately to the north of grave 391 was a second grave, 395, also cut through layer 305=390 but this was not sealed by lime 369=99. A third rock-cut grave, 386, lay to the north-east of grave 395 and was partly sealed by a stony grey-brown clay, 365=98, which also sealed lime 369=99. Clay 365=98 was itself sealed by 53=342 a light-brown clay soil which was the uppermost surviving layer of the early stratigraphy. Layer 342 contained a single offcut from a comb tooth plate which was the only unequivocal evidence from the whole excavation that combs had been made on site.

Phase 1 of the graveyard therefore consists of the three rock-cut graves 391, 395 and 386. They have an east-north-east west-south-west orientation which distinguishes them from the second phase of the graveyard where the orientation is more truly east-west. The skeleton in grave 391 was of an adult male over 50 years of age and was complete but for the feet. It lay with its arms extended and hands by the side of the pelvis. That in grave 395 was of a child, probably no older than three months. It too was more or less complete and lay with its left arm by its side and right hand resting on the pelvis. Its legs were crossed. The third skeleton, in grave 386, was of a female probably between 35 and 40 years of age. It lay with its hands by the pelvis while its lower legs continued in under the later wall, context 67, and were not lifted.

Two other burials, 56 and 62, possibly belong with this phase of the graveyard. In both cases pairs of adult's feet projected into the area while the remainder of the body lay in the area occupied by the north tower of the chapel. In the case of skeleton 62 a short portion of the lower legs was seen and these had a north-east south-west orientation matched only by a single burial found in Area 2 (below). It is, however, quite possible that this strange orientation was caused by displacement of the lower legs when the foundation trench for the north tower was cut. It was not possible to place these two skeletons stratigraphically and they could equally well belong to phase 2 of the graveyard. If the sole criterion of rock-cut graves is used to define phase 1, then they belong to phase 2 as both skeletons rested on the bedrock. Both burials, which are illustrated with the phase 2 burials (fig. 3) were left in-situ.

2.2. Graveyard Phase 2 (Fig. 3).

There are twenty mixed adult and children's burials which definitely belong to this phase, with a further eleven which probably do so. The orientation of the skeletons, in contrast to the phase 1 burials, is more or less east-west and they seem to take their orientation from a structure lying further to the south in Area 3. A line of four and possibly five stake-holes, context 354, running along the north edge of the area from ca. 1.2-1.8m east of the north tower may mark a northern limit to the graveyard at this phase and a single wooden stake, 367, at the west end of the later wall, 67, may serve a similar purpose. The stratigraphic placing of these features is, however, not absolutely certain since they may have been truncated by the later graveyard phases.

The main group of 20 skeletons is concentrated towards the eastern and southern part of the area. They lie within a ca. 2.5m wide band which stops ca. 1.2m short of the north tower and does not extend as far north as the later Cathedral wall. It should be noted that the foundation trench for the chapel was not vertical sided and so in plan some skeletons appear to be under the (later) chapel walls.

325 was a newborn child, lying on its side with legs flexed, directly under the north-east corner of the chapel.

327 was a ca. 15 year old lying under the corner of the chapel. The skeleton lacked the skull and lower limbs which had been removed by the foundation trench for the chapel and by later graves.

324 was a partial adult skeleton cut by the chapel foundations and by the grave cut for the overlying skeleton 327. Only part of the vertebral column and left ribs and arms were retrieved, while parts of the right side were trapped in under the foundation trench which had cut and removed the lower limbs.

93 was a 50 year old or more adult, badly disturbed by the grave cut for the later phase 4 skeleton 92. Only the left side from knee to neck survived. Context 319, the grave cut associated with the burial, accounted for the partial removal of the earlier burial 303 to the

303 was a male over 50 years of age, its pelvis, legs and part of the right arm had been removed by the grave cut for skeleton 93. 316 was the grave cut associated with the skeleton.

383 was a fragmentary adult to the west of 303. Only parts of the left arm and scapula and ribs survived, the rest having been removed by the cut for the later burials 303 and 388.

388 was a male over 50 years of age lying to the west and south of 383. Only the cranium was missing and it lay with arms extended and hands beside the pelvis. The cranium had rested on a pillow stone with upright stones on either side but only the lower mandible remained. The associated grave cut, 389, accounted for the removal of parts of skeletons 383 and 384 and showed that skeleton 388 was a later insertion than 303.

384 was a male in the age range 35-45 years lying to the south of the legs of 388. Only the pelvis, part of the vertebral column and right ribs and arm survived. The right hand rested on the pelvis. The skeleton was cut by 389 the grave cut for skeleton 388, while its own insertion was probably responsible for the disturbance to part of skeleton 385.

385 was a newborn child lying to the south and slightly east of 384. Its left arm, pelvis and leg and lower right leg were absent and its right arm was extended by its side.

372/382 was an approximately 5 year old child which was accidentally double numbered. Only parts of the left arm and leg survived, having been almost totally removed by the cut for a later phase 4 skeleton.

366 was another approximately 5 year old child, lying to the west of skeleton 385. Only the lower right arm and leg survived. The grave cut 364 for the phase 4 skeleton 363 was responsible for most of the disturbance.

311 was an adult which lay ca. 1m to the north of skeleton 325, in a slightly rock-cut grave 322. Only the skull survived , the remainder of the skeleton having been removed during the insertion of the later foundation 71.

306 was a baby no older than 6 months which lay to the west and slightly south of skull 311. It was a complete skeleton with the left hand resting on the pelvis and the right hand by its side and with the legs resting in a very bowed position. 313 was the associated grave cut.

83 was an adult female which partially overlay the baby 306. Context 318, the grave cut for skeleton 317, had removed the upper torso while the later pit 73 accounts for the absence of the lower right arm and femur and later foundation, 71, for both feet and half of both lower legs.

381 was an approximately 50 year old male lying to the west of 83. The skeleton lacked its skull and left arm. The right arm rested on the pelvis.

93 was a newborn? child lying between skeletons 381 and 383. It lacked the skull and the insertion of skeleton 303 had removed the lower limbs.

387 was a 35-45 year old male of which only the lower part of the body, except the feet, survived. The right lower arm lay by its side and some finger bones suggest that the left hand may have rested on the pelvis. The missing feet may be 308 which are discussed below.

302 was a 40-45? year old female to the north of 387. Only the scapulae, some of the vertebrae and right pelvis survived, the grave cut for the later skeleton 95 and feature 71 accounting for the missing portions.

394 was a newborn ? child to the north of 387. Only the skull and part of the vertebral column remained.

376 was an approximately 6 year old child with the pelvis and legs missing. It is something of an outlier to the main group, lying much further to the west than the remainder.

Of the remaining eleven skeletons which probably belong to phase 2, 307, 308 and 309 were pairs of adult feet which protruded from the east facing section during stage 2 of excavation of this area, and it was expected that the remainder of the bodies would appear during the third stage. This did not happen. It is possible that 308 were the feet of 387 in which case there is a slight drawing error. The other two pairs could not be matched up with anything.

The remaining eight skeletons are all very fragmentary and with the exception of 321 and 358 lie outside the band represented by the other twenty.

56 and 62 have been discussed above in phase 1.

506 lay close to the wall of the Cathedral and was not excavated. Only a fragment of right arm and the skull were visible.

368 was an adult scapula and clavicle. It lay to the east of 506 and was badly disturbed by the cut for the later skeleton 361.

378, a sub-adult, lay to the east of 368 and was disturbed by the cuts for skeleton 361 and the cathedral foundations. Only the femur survived and it is thus possible that this should really have been categorised as loose bone.

357 was a female, 50-60 years old which lay to the south of 378. Only the right arm, pelvis and part of the right leg survived. It was partly disturbed by the cathedral foundations and partly by the later burial 336.

321 to the south of 357 was a group of articulated foot bones.

358 lay to the south of 321 between skeletons 303 and 384 and was a group of articulated adult foot bones.

2.3. Graveyard phase 3 (Fig. 4).

At phase 3 there is a distinct change in the alignment of the burials from the more or less east-west of the phase 2 burials to a more east-north-east west-north-west orientation reminiscent of phase 1. This new alignment is more or less parallel with the chancel of Cormac's Chapel but it cannot be that structure which is influencing the alignment since it is of later date. It is therefore believed that they are aligned with wall 67 which is described in greater detail below.

15 skeletons form the phase and it is noteworthy that they are all children or babies. Also of possible significance is the fact that, with one exception, the area occupied by the skeletons was mutually exclusive with the main group of twenty in phase 2. There was a distinct group of eight in the north-west corner of the area, and a second group of four at the east end, to the north of the main phase 2 group. The remaining three are scattered between the two groups.

The group of eight in the north west corner is especially interesting in so much as they appear, in plan, to have been deliberately laid with their heads against the north tower. This, however, is impossible, since the tower is later in date and so it would seem that the wall of the tower is built upon an earlier boundary or demarcation line which the skeletons are respecting. It is further possible to suggest that this putative boundary was laid out immediately prior to the insertion of the burials. The mutual exclusivity of phase 2 and 3 burials at this end of the area suggests that there may have been an even earlier north-south boundary forming a western limit to the phase 2 burials. In other words the phase 3 burials may have been inserted into an area created by a slight westwards extension to the phase 2 graveyard. There is also a very marked stratigraphic division between the group of eight skeletons and the later burials in the area. The terminal level of the 1992 trial excavation was not bedrock but the surface of layer 37=347 which, because it was free of visible human bone, was felt to be below graveyard level and close to the bedrock. When this layer was removed at the start of the 1993 season, the sequence of 8 burials came as something of a surprise. At the surface of 37=347 were a few medium sized flattish stones, initially interpreted as a rudimentary paving, may be remains of grave markers.

48 was a 9-12 month old baby complete but for the foot bones which lay
in the east section of stage 1 but were not retrieved at stage 3. The left arm was bent while the right lay by its side. The associated grave fill was context 50.

47 was a ca. 18 month old baby which lay to the north of 48 and overlay skeleton 54. It lacked the lower legs and lower right arm. Both arms rested by its side and there was a possible pillow stone for the head. The associated grave fill was context 51.

54 was another ca. 18 month old and was overlain by skeletons 47 and 42. It was complete but for the hands and feet.

42 was a complete 9-12 month old baby lying with its arms by its side and slightly bowed legs. The skull which lay in under the tower foundation was left in situ. The associated grave fill was 44.

46 was a 6-9 month old baby lacking its lower limbs which had been truncated by skeleton 41. The associated grave fill was 49.

41 was an 18 month- 2 year old child which may have been placed in the same grave as 46 because they are more or less superimposed. The right arm was flexed over the rib cage and the left lay by the side of the body. The associated grave fill was 43.

45 was an approximately 6 month old baby. The skeleton was more or less complete with both arms at its side and its legs flexed side by side.

39 was approximately 18 months old. It lay on a slightly different alignment to the others being slightly more north-east south-west than the remainder. It was more or less complete with arms by its side. the skull was left in situ.

The eastern group of four skeletons is listed below.

304 was a newborn baby. Only the skull right ribs and arm remained.

91 was a ca. 1 year old baby. It lay to the south of 304 was disturbed by the later feature 71 and by the grave cut for skeleton 88.

301 was a child 6-10 years old, badly disturbed by the insertion of the later burial 95.

95 was a child under 10 years old. The skeleton overlay 301 but was in part under 91. The insertion of the later feature 71 had removed the lower body.

The remaining three skeletons were as follows.

377 was a sub-adult. The skeleton had been disturbed by the insertion of the Cathedral wall and the later cut 61 and only three ribs remained.

352 was approximately 3 years old. The skeleton partly overlay phase 2 skeleton 381. Only the lower arms, vertebrae, pelvis and both femurs were present, the rest having been removed by later features including cut 355.

334 was no older than 5 years. It lay between 352 and the main western group. Only part of the right side survived, the rest was removed by the grave cut for the later skeleton 14.

2.4. Graveyard phase 4 (Fig. 5).

A group of five adult/sub-adult skeletons ran parallel with and were cut by context 371, the foundation trench of Cormac's Chapel. Their alignment is similar to that of phase 3 and distinct from that at phase 2. It is not possible to say with absolute certainty whether these graves are contemporary with phase 3 or represent a distinctly separate but slightly later phase. If they are contemporary with phase 3 then there is a marked distinction within the graveyard with an area for adult burial and one for children. Further out from the wall, where the skeletons were not cut by the foundation trench, it was more problematic to interpret which of the remaining burials belong to the phase. There was a mixture in the orientations in the remaining skeletons still to be dealt with but it was possible to divide out eleven skeletons which may be contemporary with the others at phase 4 using the criteria of a similar orientation to the definite phase 4 skeletons and their low stratigraphic positioning. Interestingly the eleven thus chosen slot neatly between the eastern and western groups at phase 3 and so seem to form a cohesive group. If these eleven can be considered as a group contemporary with the other five, then the phase is definitely later than phase 3.

The five skeletons cut by the foundation trench of Cormac's Chapel are;-

323 a ca. 50 year old male, lying under the north-east corner of the chapel. Only the left hand side of the skeleton , from the neck to the pelvis remained, the right hand side having been removed by the foundation trench and the lower limbs by the later feature 71. The surviving right hand rested on the pelvis.

374 a ca. 17 year old female lying to the west of 323. The skull and most of the left hand side, excluding the lower arm, survived, the remainder was removed by the foundation trench. The grave cut for the skeleton was 379.

373 a ca. 17-18 year old female lying directly over 374. Only the pelvis and lower vertebrae survived.

370 a male of 50 years or more overlying 373. This skeleton lay slightly askew to the two underlying ones, with an alignment part way between phase 2 and 4. This was, however, probably caused by slumping in the ground. The upper body was more or less complete but the lower limbs were missing. The right arm rested on the pelvis. The northern side of the grave cut, 380, was stone lined.

375 a sub-adult. Only the left leg and foot remained the rest having been removed by the foundation trench.

The following 11 skeletons are probably contemporary with the above;-

92 a male of 50 years or more lay to the north of skeletons 323 and 374. The skeleton was more or less complete but for its feet. The left hand rested on the pelvis while the right hand lay by its side. Unusually for this site the legs were crossed at the ankle, left over right. The associated grave cut, 320, had removed much of the underlying phase 2 skeleton 93.

353 an elderly adult of indeterminate sex, lay to the north-west of 92. Only the lower mandible and upper torso with both humeri survived. The lower body was disturbed by the insertion of skeleton 87, and context 317, the cut for skeleton 318, probably caused the removal of parts of the left side.

345, a female probably in the age range 30-40 years, lay to the west of 353. The skull, parts of the left arm and both legs were missing, but both hands rested on the pelvis. The later skeletons 331, 335 and 87 all cut 345.

363, an adult male, lay to the west of 345. The skull and the legs were absent, the left hand rested on the pelvis and the right arm was bent upwards. The skeleton was cut by the later pit 332 and the grave cut for the burial was context 364.

359 a ca. 5 year old female(?) was the only child's grave in this group. It overlay 363 and only the lower vertebrae, pelvis and both femurs survived. The right hand lay under the pelvis. It was cut by pit 333.

317, a ca. 17 year old male, lay to the north of skeletons 363 and 345. The skeleton was more or less complete but for the skull. A disarticulated skull lay at the right shoulder of the skeleton but appeared not to belong to this skeleton. There were possible pillow on which the skull would have rested. The left arm lay by its side while the right hand rested on the pelvis. The associated grave cut was 318.

96, an adult, of which only the lower legs and feet survived, lay to the north-east of 317. Its associated grave cut was context 300.

348, a female probably 50 years old or more, lay to the north of 317. The skeleton was complete and the re-interred bones of at least two other individuals were placed around it. The right arm lay at its side while the left hand rested on the pelvis. The associated grave cut is 350 and two disarticulated skulls registered as from graveyard fill 69 are probably associated with the cut. The skeleton overlay 360.

360 was a male(?) of 50 years or more. The skull, part of the vertebral column, right ribs and humerus, and left scapula were present. The pelvis was missing but the right femur and both lower legs were present. The legs were crossed at the ankles, left over right.

349 was an old adult male. Only the skull, left scapula and humerus survived, the rest having been removed by the cut for skeleton 348.

362, an adult, lay to the north of skeletons 348 and 360. The left scapula and some ribs were all that remained.

2.5. Graveyard phase 5 (Fig. 6).

The remaining 40 skeletons cannot be subdivided further and are a mix of adults and children, male and female. None of them were cut by the foundations of Cormac's Chapel, so all post-date that structure, but several respect the more east-west orientation of the cathedral. Some of the upper graves were truncated by the foundations for the 18th century sacristy floor, so the phase has reasonably firm date brackets, i.e. between 1134 and the 18th century.

87 was a male between 30 and 40 years of age which had sunk into the underlying graves. The skeleton was complete but for the feet which had been removed by the later feature 71, though the right arm and leg were left in situ. Bones from at least one other individual were stacked in the grave. Both its arms lay at its side.

82 was a female of 50 years or more, overlying 87 but with its head slightly further to the east. It too had sunk into the underlying grave, displacing the vertebral column, and had its lower limbs removed by feature 71. The left arm lay across its stomach and its right arm was by its side.

78 was a female, probably over 60 years old, overlying 82. Traces of very badly decayed wood fibre around the body suggests that it may have been a coffin burial. The hands rested on the pelvis. The associated grave fill was context 84.

331 was an adult male which lay to the north-west of 78. Only the lower legs and feet survived, the rest of the body having been removed by the later intrusive feature 333. The left foot rested on the right.

335 was a male of 50 years or more, lying to the north of 331. The skull, shoulders and upper vertebrae were present but the remainder of the skeleton had been removed by feature 331 and the grave cut for skeleton 331. The left hand rested on the pelvis.

332 was an adult male overlying 335. Only part of a humerus and some of the right ribs survived. It was cut by feature 333 and the cut for skeleton 331.

338 was a male between 40 and 50 years of age, lying to the north-west of 332. The skeleton was complete but for the feet which, had been removed by the later pit 73. The skull rests between vertically set stones and other stones line the edge of the grave cut. Both arms lay at is side.

75 was a pregnant 30 to 35 year old, with associated foetus 76. The skeleton lacked the left lower arm, left, and part of the right, legs. It was cut by later features 71 and 79. The right hand rests by its side. In the pelvic and chest regions were a scatter of small bones which have been identified as a foetus.

77 was a male adult, lying to the south of 75. Only the right arm remained, the rest having been removed by 75.

81=339 was a male aged between 45 and 50 years, lying to the north of and partly overlying skeleton 338. It was complete but for the feet which had been removed by feature 71. The hands were crossed on the pelvis. Some wood fibres under the coffin indicate the possible presence of a coffin

341 was a child under 1 year old, lying to the west of 339. The skeleton was very fragmentary and only parts of the vertebral column, right ribs, pelvis and right leg were retrieved. The remainder had been removed by the cut for skeleton 337.

337 was a 60 year old or more male, lying to the north of 339 and 341. The skeleton lacked the lower leg and right foot and both hands rested on the pelvis. The grave cuts for skeletons 339 and 340 account for the missing portions.

344 was a male, aged between 50 and 60 years, lying to the north of 337 and under skeleton 340. The skeleton lacked its right arm, lower left arm and lower legs.

340 was an old adult, of indeterminate sex, overlying 344. Fragments of the vertebral column and ribs only. The remainder had been removed by the general truncation of the overlying graveyard.

85 was an adult female to the east of 344. Only the left arm survived.

86 was an adult to the north of 85, represented only by a pair of tibia.

351 was an adult lying to the north of 86. Only the feet remained, the rest having been removed by the cut for skeleton 336.

361=88 was an adult male (?) lying between skeletons 336 and 334. It is believed that legs 88 excavated in stage 2 match 361 found in stage 3. Complete skeleton with arms by its side. 90, the fill of the associated grave cut 89, contained the bones of at least one other individual. 310 a disarticulated skull was found close to 88.

336 was 25 to 30 year old male, lying to the north of 361 and tucked slightly in under the Cathedral foundations. Complete with hands resting on the pelvis.

32 are two fragments which were wrongly matched on site and are subdivided here.32a was the left arms and ribs of a child 10-15 years old, the rest of which had been removed by feature 333.32b was a 30 to 40 year old female. Only the pelvis and legs survived with part of an associated coffin 356. Context 355 was the associated grave cut.

27 was an elderly male, lying to the north of 32a. The skull and some ribs were all that survived.

31 was a female ca. 21 years old lying to the north of 27. The skeleton was more or less complete with possible pillow stones for the head. The hands were crossed on the pelvis.

18 was a child ca. 2 years old, overlying 31. The skeleton was more or less complete, with both hands resting on the pelvis.

15 was the crushed skull of a child of 5 years or more, lying between 18 and the north tower.

14 was a male adult lacking parts of the left arm and lower limbs.

30 was a ca. 40 to 45 year old male, lying to the north of 14. Only the left arm and leg and part of the pelvis with the right hand resting on it remained. The rest had been removed by the cuts for skeletons 19 and 31. This skeleton had a ring on its right hand.

19 was a ca. 8 year old child, overlying 30. The skeleton was more or less complete with its hands resting on the pelvis. There were possible pillow stones for the head.

29 was an adult male skull, lying by the wall of the north tower.

26 was a female between 30 and 35 years of age lying in the north-west corner of the area. The skeleton was more or less complete with its left arm resting on the pelvis and the right arm extended by its side.

20 was a 15 to 18 year old, overlying 26. Only the shoulder blade and a few articulated vertebrae survived.

13 was a ca. 45 to 50 year old male, overlying 20. The skeleton lacked its skull and the left arms and feet had been removed on the insertion of skeleton 12.

12 was a female 35 to 45 years of age. The skeleton was complete down to the knees; its right arm rested on the pelvis and left by its side.

24 and 25 were the disarticulated remains of what turned out to be three individuals stacked within the grave for either 23 or 26. The bones were of an elderly female, an elderly male and a child.

28 was a 15 to 16 year old male, lying under skeleton 23. The skeleton was complete, with its head resting on pillow stones and both arms by its side.

23 was a male 60 years or more old, overlying 28. The skeleton lacked part of the pelvis and right humerus. The right arm was extended by its side and the left rested on the pelvis.

97 was the disarticulated skull of a ca. 6 year old child lying at the surface of the graveyard.

79 was a pair of adult male legs which somehow seems to have escaped being planned and whose position is uncertain. The remains, however, lay close to skeleton 82.

2.6. The graveyard fill.

With the constant churning of the area by the insertion of over one hundred burials the graveyard fill has become very mixed and it cannot be divided up into a meaningful sequence except early and late.

346 was part of the lower fill. It was dark brown in colour and mixed with some lime and mortar, probably deriving from the underlying layers 40 and 369 discussed at a) above. It lay towards the north and west of the area.

94 was the lower fill to the south and east end and was a light brown clay soil. It contained fragments of antler waste.

72 was similar to 94 but at a higher level. Finds from the layer include a spindle whorl and antler waste.

17 was a mid grey-brown soil associated with the earlier part of phase 5.

11 was a mid grey brown compact soil containing mortar, which post dates 17.

9 was a compact mid brown soil at the upper levels of phase 5.

8=16=69 was the upper dark brown clay soil flecked with charcoal and mortar.

4=68 was the uppermost level within the graveyard and represents the trampled interface between 8=18=69 and the overlying foundation.

315 was a pair of wooden stakes lying at the east end of the site. It is impossible to tell from which level these have been inserted into the graveyard

2.7. The buildings.

There are, as noted above, the remains of three mortared stone buildings within Area 1; the present cathedral, Cormac's chapel and a short section of wall, 67, re-utilised as foundation for the cathedral (Figs. 4 and 5).

When Cormac's Chapel was built the graveyard had already been in use for a considerable period of time with over half the burials in Area 1 pre-dating the building. The foundations 371=326 were cut through to the bedrock in a U-shaped trench which narrowed towards the base. Unlike the inside of the building (see Area 3 below) the outer face of the foundations appear in part at least to have been deliberately laid. The lower part of the foundations were voided rubble but with the semblance of careful placing at the edge of the trench. Upon this lower foundation were coursed uncut stones which had their greatest depth of three courses under the tower and narrowed to one course as one moved eastwards along the wall. These stones averaged 0.3-0.4m in length and 0.1-0.2m high. On top of these was a plinth, consisting of a single course of uncut stone projecting ca. 0.2m out from the wall, upon which the ashlar masonry of the chapel itself sits. There is no comparable plinth on the inside of the chapel where the foundation trench is much wider and did not hug the wall line (see area 3 below).

326=371 was the construction trench for the north wall of the chapel, and context 505 its fill. The chapel's north wall, both ashlar and foundations, was context 33 and the north tower and its foundation was context 34. Context 63 was the construction cut for wall 34 and context 64 was the voided rubble fill at the base of cut 63.

The Cathedral wall rests on two distinctly different foundations. From Area 2 in the west unto roughly the mid-line of the doorway from the sacristy into the chancel, the foundations, 35=343, are clearly of the cut and fill type, i.e. a trench, more or less the width of the overlying wall, backfilled to ground level with stone and mortar upon which the faced wall sits. Just west of the doorway a large flat limestone block is incorporated into the upper foundation and appears to run through the wall to be visible on the inside of the chancel. The significance of this block, if any, is unclear. From the mid line of the doorway eastwards into the section the foundation is much wider and more carefully laid. This section, 67, also appears to be on a slightly different alignment to the overlying wall and to be inserted into the graveyard at phase 3 while the rest of the foundation did not appear until phase 5. There is no absolute proof for the assertion that wall 67 was built at phase 3 but it certainly post-dates the phase 1 graveyard, because it overlies grave 386. The excavated section did not cut any other graves, which suggests that it must be early in the sequence and it quite clearly pre-dates phase 5. It is at phase 3 that there is a marked change in orientation in the graves and it seems as though at this time the influence of an early church building in Area 3 (see below) was waning and that a new building was influencing the orientation of burials. This new building cannot be either the Gothic Cathedral or Cormac's Chapel because neither was built at the time. The only candidate is the wall 67 and so it is therefore interpreted as part of an earlier church. The short length of wall is mortared, rests on the bedrock and survives to a height of 2-3 stones or roughly 0.3m. The full thickness of the wall is masked by the later Cathedral but it projects 0.4-0.5m out from the overlying wall. It is assumed that the wall must have turned in under the doorway into the chancel of the Cathedral because there were sufficient remains of the undisturbed early stratigraphy to show that it had not continued any further west. It is unfortunate that the sacristy door was inserted into the Cathedral wall at just the point where 67 and the Cathedral foundations join, because the work around the doorway obscures exactly what happened to wall 67. Excavation of the inside of the Cathedral might clarify the position.

2.8. The post graveyard features (Fig. 7).

The level of the graveyard was lowered across the whole area at the time that the sacristy was built between the chapel and cathedral. This truncation was numbered 61. The level was then built up again with a ca. 0.3m thick spread of rubble which was presumably the base for the floor of the sacristy. This rubble was variously numbered, 60=330=3=7 and continued eastwards off the site where it ran under 328 the east wall of the sacristy.

Under the rubble layer were a number of features which were probably truncated by the general lowering of the level.

71 lay at the east end of the site and continued eastwards into the section. It ran from the edge of wall 67 and southwards to the chapel wall. The fill of the feature, 70, was large voided limestone rubble. It is assumed that 71 is a drainage feature/ foundation for the sacristy.

73 lay ca. 0.2m to the east of 71 and was a sub-oval pit ca. 0.5m in diameter and 0.4m deep. The fill of the feature, 74, was mostly disarticulated bone, possibly re-interred when the sacristy was built.

65 lay ca. 0.8m to the south-west of 73 and was a sub-rectangular pit measuring 0.66m x 0.32m by 0.32m deep. The fill, 66, like that of 73, was re-interred disarticulated bone.

38=333 In the south west corner was a large sub-rectangular pit measuring ca. 2.2m east-west by 1.2m north-south cutting down to within 0.2m of the bedrock. It is this feature which accounts for the absence of skeletons in this corner from at least phase 2 onwards. The fill of the feature 10=21 and 22=329 was a mixed mid-brown clay containing sandstone fragments, mortar, small stones as well as fragments of green glaze pottery, olive jar and window glass.

Sealing the rubble 60=330=3=7 were modern sand and gravel layers 1, 2 and 59.

Cutting these very modern layers was context 5, the trench excavated by Mr. Con. Manning around the blocked doorway from area 1 into the chancel of the chapel. Its backfill was 6.

36 was the context number given to cleaning over the surface of stage 1 after the winter.

3. AREA 2.

A small trench just outside the north doorway and running for ca. 2m along the Cathedral transept wall was excavated in 1992. In 1993 the area was excavated in two main stages because of the scaffolding. The eastern half of the area was excavated first and then the western half. (Figs. 8-10).

Towards the end of the last century or early this most of the area had been reduced in level and backfilled with rubble presumably to help drainage. This disturbance had removed the stratified deposits almost down to bedrock in some places while other parts stood a little higher. Only along parts of the surrounding walls was anything like the full stratigraphy left intact where there was a narrow ribbon of deposits sitting between the foundation trenches and the central reduced area.

The fissured limestone bedrock 115=117 formed the basal layer for the whole area, with, here and there, some orange clay, 116, between the fissures.

Two features were cut into the bedrock from a very low level. 139 was a linear east-west feature lying just in from of the tomb niche to the east of the north doorway. It ran from under the north tower and westwards for a distance of ca. 1.35m. It widened westwards from 0.4-0.5m and was 0.3-0.4m deep. Its full extent is unknown but it was not recognised as a feature within the north tower and so presumably terminated somewhere within the thickness of the tower wall. Its fill, 138, was small to medium sized stones in a dark orange-brown earth. The purpose of the feature is unclear but despite its east-west orientation it does not appear to have been a grave.

The second feature, 140, was cut into the bedrock by the northern edge of the area and continued in under the Cathedral foundations. The area visible was sub-rectangular in plan ca. 0.5m east-west by at least 0.4m north-south. It measured ca. 0.5m at its deepest though there was a slightly higher ledge in the south-west corner. The fill of the feature, 141 was a medium brown clay containing some bone and small stones.

Both 139 and 140 were sealed by a layer of dark-orange brown clay 134 measuring up to 0.3m in thickness. Towards the south-east the layer was slightly more orange in colour and was given context number 135. To the west the same layer was also numbered 114 and overlain by a greasy brown clay 111.

In the centre of the area the clay layer 134 was sealed by a small area of flattish slabs, 145, which appear to be a deliberately paved surface. This paving had presumably been more extensive but was disturbed by the modern intrusion. The remaining area measured ca. 1.7m east-west by 1m north-south. To the west some stones on the surface of layer 111 may be displaced remnants of the paving. The stone paving was sealed by a thin charcoal spread 110=127. Charred grain was found between the stones of 145, at the interface of 127 and 134 (see appendix 3).

The lower limbs of an adult skeleton, 148, fit into the stratigraphic sequence somewhere here though it is not possible to give a more precise account of the level from which it was inserted because of the modern disturbance. The upper part of the body had been removed by the later circular feature 146=147. The orientation of the burial, north-east south-west, is completely different to all the other burials on site, with the possible exception of 62 in area 1.

146=147 was a circular stone lined feature cut through and into the bedrock, possibly from the same level as the paving 145 or slightly higher. The cathedral transept and chancel walls were built over and had protected the north-west quadrant of the feature. It was cut ca. 0.55m into the bedrock and at the base was ca. 0.6m in diameter widening upwards to ca. 1.1m. At its highest point on the north-west it measured 1.45m from base to topmost surviving stone but to the south-east the stone lining had been removed down to bedrock level. The sides were of randomly coursed mortared limestone. The feature had been cleaned out early this century and the sole fill was the rubble backfill which covered the whole of the area. The purpose of the structure is unclear but it is very definitely not a well because its base is several metres above the water-table in the well at the external north-west corner of the Cathedral crossing. That explanation literally does not hold water. If, as is possible, it functioned at the level of the paving then it may be connected with the charred grain at the interface of 127 and 134, but there was absolutely no evidence for a flue and it is unlikely to have been a corn-drying kiln. However the fact that the upper surviving stone is ca. 0.4m above the paving suggests the possibility that it is of a slightly later date than the paving. Layer 105, was cut by the feature and probably post-dates the paving. Feature 146=147 did not extend into the area excavated in 1992 but its presence was suspected where it accounts for what was then termed a "phantom feature", 106.

At the level above the charcoal 110=127 the upper deposits survive mainly as bands of material around the periphery of the area. In the north-east corner was a deposit of light brown sticky clay, 125, measuring up to 0.3m thick which continued into the Cathedral and Chapel tower foundations. Under the Cathedral a light brown sticky clay 126, possibly the same as 125, overlay the charcoal. At the west end a small deposit of light brown clay 109 sealed the charcoal but 109's relationship to 125 and 126 is unfortunately lost. 109 was sealed by a brown clay 105 which dipped down towards the circular feature 146=147 and appeared to be cut by that feature.

The foundations of the north tower cut through layer 125 which was the upper stratified deposit at the east end. The line of this foundation cut was carried on northwards to the cathedral wall by a north-south edge 123. East of the edge was a fill of mortar and small angular limestone, 124, which was heavily iron panned. No opposing edge was seen in area 2 and it is assumed that the opposing edge was the edge of foundation trench 63 in area 1. The function of feature 123 is assumed, in part at least, to be the foundation trench for the north tower but the fact that it continues in under the Cathedral wall which is over 1m to the north at this point suggests that 123/124 may have an earlier function which has been masked by the tower. A clue to this is provided by two large flat slabs of limestone one of which looked to have been deliberately set along the edge of 123, while the second, to the north of the first, looked to have been displaced from a similar position. At one point the charcoal layer 127 fractionally, by 10mm or so, overlay the stone set along the edge. At the time of excavation this was believed to be a product of the compression of layers which caused slight lateral movement towards the voids among the rubble fill 124 thus giving a false relationship. With hindsight this may have been a genuine relationship which would suggest that the lower part of 123 is a separate earlier feature. What is certain is that at no other point within the excavated areas does the foundation trench extend so far out from the wall because such a wide foundation was totally unnecessary. If this alternative explanation is acceptable then the feature fits into the stratigraphy as roughly contemporary with the paving 145.

At the east end a row of five burials were aligned parallel to the chapel with their feet respecting the north wall line of the tower. The major modern disturbance has removed any clear evidence of which level they were cut from but the indications are that they post-date the construction of the chapel. The most northerly skeleton lies partially under the Cathedral wall and so pre-dates its construction. The burials are therefore believed to date between 1134 and the mid 13th century.

131 adult male skeleton lacking its skull and part of its right arm, the left hand rested on the pelvis. The grave-fill 133 was a medium brown sticky clay containing charcoal and small fist sized stones.

132 was an almost complete male adult (50+). A disarticulated left leg found in the overlying grave 130 is believed to be the missing bit of 132. The arms were crossed on the pelvis. The associated grave fill was 144.

130 was a female? child's (ca. 7yrs) skeleton complete but for its hands, with the left arm by its side and right forearm on its chest. The grave fill was 136.

129 was a complete old adult male with its right arm across its body and left arm by its side. The fill 128 was a grey brown sticky clay containing mortar, charcoal and bone.

149 lay under the cathedral and was left in situ. Only parts of the right side were seen.

A sixth skeleton adult male (30-35?) 112 lay further west immediately outside the porch of the north door. The grave, 107, cut clay 105 but only the sacrum, left pelvis and femur remained. The right hand side had been removed by the insertion of a later rectangular stone feature, 103=142, and modern disturbance accounts for the absence of the lower limbs. Part at least of the upper body could have been removed during the construction of the cathedral but this could also have taken place at the time that 103 was inserted. At the base of the grave cut was a deposit of loose iron panned voided stone, 113, with some grey clay between. Overlying this was the upper fill , 108, a gritty sandy grey layer containing stones up to fist size and mortar. The whole grave seemed to be water-washed by the run off from the chapel.

Within the porch of the north door lay 103=142 a rectangular stone lined feature measuring ca. 0.9m east-west by 0.7m north-south. The east wall of the feature was missing, removed by the massive intrusion into the area at the turn of the century. It appears to have had a slab base but this had been broken through to expose the underlying bedrock when it was emptied. The fill was part of the modern backfill 143. The walls of the feature were made of mortared roughly coursed limestone and the internal face was rendered. At its highest point it rose ca. 0.9m above
the base which brought it roughly level with the lower ashlar coursing at the west side of the doorway. The exact stratigraphic relationship of the feature to the chapel has been destroyed by the modern intrusion but it is believed that it was inserted into the doorway after the chapel was built rather than that the chapel or porch was built around it. The reasons for this assumption are twofold. Firstly the feature cuts an earlier skeleton 112 which itself is quite a late feature in the surviving stratigraphy. This skeleton has the same orientation as the row to the east and would therefore seem to be of a similar date i.e. after 1134. Secondly on the east side of the doorway there seems to have be some modifications made at the base of the ashlar which may be connected with the feature. A long flat slab has been inserted at the base of the ashlar, some of the adjacent sandstone has been cut back and some limestone blocks have been inserted into the doorway. The slab is currently flush with the wall but had it or a second one originally extended across the doorway then it would have served as a lid for the feature. The feature is tentatively interpreted as a reliquary shrine.

At the turn of this century the whole of this enclosed are was dug out.104=122=143 a sticky grey clay with scattered fist sized stones seemed to be a trample associated with this work and it was sealed by a ca. 0.45m thick deposit of voided rubble, 101=119, from which came an 1897 Queen Victoria halfpenny which dates the activity to recent times. It is perhaps worth noting that local memory of the work survives. When the well like feature 146 was exposed it came as no surprise to one of the OPW guides on the rock, who stated that her father had always told her that there was a well there. The family appears to have had close connections to the Rock throughout the century. A small lens of light brown clay, 121, was sandwiched between 101=119 and 122 by the cathedral wall. This is assumed to be part of the layer 125 which collapsed into the hole before it was backfilled with 101=119. In front of the tomb niche, to the east of the door, the edge of the disturbance was numbered 120. The whole of the stratigraphy was topped off with ca. 0.2m of compacted gravel, 100=118.

102 was the context number given to the foundation of the south transept of the cathedral. It was built by the cut and fill method as was the wall of the cathedral forming the northern edge of the area.

4. AREA 3.

A small trench in the north east corner of the chancel was opened in 1992. In 1993 the line of the southern section of the 1992 excavation was extended westwards and the northern half of the nave excavated first, in order to allow the north wall to be underpinned. The southern half of the area was then excavated as stage 2. Figs. 11-13.

The fissured limestone bedrock, context 218, formed the base layer within the chancel and north-east corner of the nave, but elsewhere there was an orange clay subsoil, context 217. The ground sloped slightly from north-east to south- west and the orange clay occupied the lower area as well as being found in the fissures in the bedrock in the higher area.

In the area under the chancel arch there was a large scale disturbance which cut the stratigraphic links between the nave and chancel. Because of this the parts of the interior of the chapel are presented separately.

4.1.The chancel.

The earliest features in this area were two shallow rock-cut post-holes 415 and 417 both of which were irregular in shape because of the nature of the rock into which they are cut. The maximum dimension of both is ca. 0.5m and depth 0.2m. Further to the west on the same line was a similar sized cut into the bedrock, context 418. This lay underneath a later intrusive feature, 277, and so it is not absolutely clear whether 418 is in fact another post-hole or part of the later intrusion. A similar situation pertains to the east where possible post-hole 435 underlay a modern cut 205. Seen in isolation 418 and 435 could easily be disregarded but taken together with 415 and 417 they do seem to be part of a row of four post-holes. This row is even more convincing when compared with a slightly later row of four lying immediately to the south. The orientation of both rows is more or less east-west and the burials at phase 2 of the graveyard in Area 1 appear to take their orientation from these alignments.

The second row lay just to the south of the first and consists of post-holes 276, 275, 413 and 416. Only 416, under the later intrusive feature 277, was in any way doubtful. The dimensions of the holes in this second row were altogether larger than the first. The two eastern post-holes 275 and 276 each had two distinct fills and a second identifiable "cut" noted at a higher level. In both cases the surrounding stratigraphy ran over the lower post-hole cut and up to the edge of the second cut. It is believed that these second "cuts" are not evidence for re-utilisation of the post-holes but are in fact the post-pipes of the posts themselves. In other words there was a build up of deposits against the posts which have subsequently either rotted in situ or been pulled out to leave a void which was then backfilled. In the case of 276, the lower fill 270, was a friable grey clay containing some fist-sized stones. Context 266 was the post-pipe and its fill, 267 was a loose grey-brown clay containing several medium sized stones which may have been part of the original packing. The lower fill of post hole 275 was a friable grey clay, 271, while 262 was the post-pipe and 263 the upper fill. It is possible to suggest that the post which had stood in this post-hole was 0.25m in diameter. In the case of the third definite post-hole it was not possible to identify a post-pipe. The fill 414 was a loose grey-brown clay. An ash layer 244, of which more below, appeared to seal the very edges of the pit and to have been deposited around the post.

The two rows of post holes are interpreted as representing the wall line of a two phase early church.

Sealing post-hole 415 of the northern row and partly sealing post-hole 413 of the southern row was a spread of white ash, 244, with a slight concentration of charcoal along its southern edge. The spread was sub-rectangular in plan measuring 1.6m north-south by 1m east-west and the burning appears to have occurred in situ because the clay in the fissures of the limestone was slightly reddened. Seven stake-holes, 412 (a-g) were set into the ash, all occupying peripheral positions, which suggests, possibly, that the full extent of the layer was visible when they were inserted.

A thick deposit of dark-grey clay with a high charcoal content, 264=261=209, sealed ash 244 and the earlier row of post-holes. It also sealed the second row of post-holes but was the layer which formed the edge to the post-pipes. The layer extended across the chancel but only at one point, between the later intrusions 229 and 231, did its full original thickness, of ca 0.4m, survive. A stake-hole, 268, on the south edge of post-pipe 264 was probably pushed in from the surface of 264. It measured 0.08m in diameter by 0.25m deep.

Overlying 264=261=209 was a ca. 0.15m thick deposit of light brown friable clay, 260, and above that was a series of burials which had been cut on the east by the foundation trench for Cormac's Chapel and on both sides by the later intrusions 229 and 231. All of the skeletons were fragmentary but their orientation does not match that of the underlying rows of post-holes. The line is more east-north-east west-south-west akin to that of graveyard phases 3 and 4 in Area 1. They are therefore interpreted as contemporary with phase 4, on the grounds that the skeletons are a mix of children and adult unlike phase 3 which is exclusively children.

255 was a female? adult, only parts of the vertebral column and right ribs and arms survived.

254 overlay 255 and was an adult male between 35 and 45 years of age. Only the chest and vertebral column remained in situ. The burial was disturbed by a later one, 253, and at the time that the later grave was dug the skull of 254 appears to have been removed and repositioned on the chest.

253 lay further to the west and was a child of ca. 18 months. It was more or less complete but for some parts of the left side. A grave cut, 256, was identified around part of the burial.

A light to medium grey silty clay, 251, sealed burials 253, 254 and 255 and was itself overlain by another skeleton, 246, which is believed to belong to the same phase as the others. It was an adult female skeleton ca. 23 yrs. of age, of which only the lower jaw, vertebral column and ribs and part of the right pelvis survived. The legs were missing but some articulated foot bones, 250, are believed to be belong to the same skeleton. Context 245, a light to medium grey silty clay, surrounded the skeleton.

Further to the west and cut into 209, was another grave, 410. All that remained of the skeleton, 409, after disturbance by the later feature 277 was a pair of articulated adult feet. The grave fill, 411, was a loose grey brown friable clay mixed with charcoal. This burial is believed to belong to the group described above although its position in the stratigraphic sequence was less clear than the remainder. The lack of a coffin distinguishes it from the adult post-medieval burials in this area. Skeleton 408 on the edge of later feature 229 occupies a similar uncertain position in the stratigraphy to 409. Even less of this skeleton remained, only a left adult foot.

To the north of the main group were two other skeletons which probably belong to the same phase. 216 was an adult male. All that remained of the skeleton was the pelvis, upper part of the left femur and some hand bones lying in the pelvic region. Layer 209, upon which the skeleton rested had been badly disturbed and these later intrusions account for the absence of the rest of the skeleton. A short-cross penny of ca. 1200-1210 was found on the disturbed surface of layer 209 ca. 0.2m to the north of the pelvis of 216. This coin cannot have come from within layer 209 because the layer quite clearly predates the chapel. It is just feasible that the coin was associated with the skeleton, though no trace of a grave cut was noted around the body. If that is the case then the burial is separate from those described above, predates the post-medieval burials and is the sole burial in the chancel during medieval times. The more likely explanation for the coin is, however, that it was trampled into the surface of layer 209 when the later disturbance took place.

The second skeleton in the northern part of the chancel was 214 a single articulated foot of a 10-15yr old child, lying in the north-east corner, sealed by a light brown layer 212. This sealing layer appears to have been part of the fill of the foundation trench of the chapel which has spilled south-eastwards when one of the late intrusions into layer 209 occurred.

The building of the chapel itself is the next event in the stratigraphic sequence. The construction trench was traced around the inside of the chancel, though in several places it had been disturbed by later intrusions. Along most of the north wall it was obscured by later graves, but at the north-east corner it turned quite wide of the corner and was clear along most of the east wall. From the south-east corner the inside edge of the trench ran under the wall line, hugging the wall quite closely, but emerged again near the south side of the chancel arch. The trench received several context numbers, 234 along the north wall, 235 around the north side of the chancel arch and 248 in the south east corner, while the fill of the trench, a loose stone rubble with decayed sandy mortar, was variously numbered, 206, 215, 222, and 249. The wall itself rest on a loose voided rubble which is a marked contrast to the regular coursing noted in the foundations on the outer side of the north wall.

The remaining features in the chancel fall into two groups, a series of burials of probably 18th century date and a whole series of intrusions which post-date the burials.

From north to south across the chancel the burials were.

204 an adult female skeleton in coffin 203 in grave cut 202. The bones within the coffin were partially demineralised and very soft. Within the coffin some short lengths of silk ribbon remained tacked to the sides. The coffin was equipped with handles and along the bottom edge were closely set flat and roughly T-shaped pieces of iron. A disarticulated female child's (ca. 15yrs) skull 211 was set at the foot of the coffin within 202.

Immediately to the south was grave cut 207 in which there were the remains of two burials. 227 was the upper part of a coffin containing 228 the partially demineralised skeleton of a 25-35 year old female. The skull was in reasonable condition and it was possible to see from green stains that a shroud had been pinned at the temples and on the lower jaw by the back molars. The lower part of the coffin had been removed by later intrusions. Below 228 was coffin 208 containing adult skeleton 210 which was largely demineralised. The coffin was equipped with flat fittings along the base of the sides and a piece which presumably sat on a gabled lid.

To the south and west of 207 was the foot of a coffin, 287. The rest of the coffin and its contents had been cut and removed by the later feature 277.

East of 287 was grave cut 213 which contained at least two burials. Of the earlier one, only the headboard of the coffin, 289, remained. Slightly further to the east was the female? adult skeleton 242 in coffin 241. This lacked its skull and was cut and removed from the pelvis downwards by feature 231. On either side of the inside of the coffin was a row of bronze tacks, presumably used to hold some sort of lining in place.

To the south and west of 242 were two very fragmentary skeleton neither of which had coffins. 274 was the pelvis and lower arm of a male? child (10-15yrs). 288 was the right lower right leg and foot of a child 10-15 yrs.

The remaining skeleton was a complete baby, 9mth-1yr. 232, lying in coffin 233.

The later intrusions from south to north;

229 an east-west cut 2.5m long by 0.5-0.6 wide and 0.35m deep bottoming out in 209=262=264. Although this feature was rather grave like there was no coffin or articulated bone within it. The fill, 230, was a fairly compact grey clay containing lumps of mortar and plaster, large stones and some loose human bone.

231 lay ca. 0.4-0.5m to the north of 229, with an upstanding strip of early stratigraphy between the two. This feature appeared linear but the edge ran into grave 213 and became somewhat amorphous. The fill was indistinguishable from the overlying general layer 201=220.

Pit 205 cut feature 231 and was ca. 0.9m in diameter. It was excavated in relatively recent times and seems to have been for reburial of loose bone. Its fill contained a foil sweet wrapper with the name MILROY printed on it.

219, another pit, continued under the centre of the east end of the chancel and could possibly be the eastern continuation of 231. Its fill was similar to the general layer 201=220. The stone floor of the apse/alcove at the east end sits on what looks suspiciously like modern mortar and would seem therefore to have been reset in recent times. In such a circumstance it is likely that feature 219 is connected with that work.

Apart from the burials and intrusions listed above it is clear that there had been a number of other late intrusions whose edges were indefinable. For instance in much of the northern half of the chancel layer 209 had been cut away but no definable features could be discerned. The whole of the area was however covered by a very mixed grey sandy layer, 201=220, formed from the continuous churning of deposits. A large amount of disarticulated human bone and loose coffin fittings attest the very disturbed nature of the deposit. It varied in thickness and actually formed the fill of several of the intrusions and its crusted upper surface formed the floor level prior to excavation. That level more or less corresponded with the lowest ashlar course. Only at the very east end was there a thin scatter of yellow sand, 200, overlying 201=220.

4.2.The nave.

The earliest feature in the nave was a scatter of stake-holes, 419, which appeared as shadows against the orange clay subsoil in the southern half of the nave. There were 29 holes, varying from 0.04-0.09m in diameter and from 0.03-0.12mm in depth though there was no discernible pattern to them. A further two holes, 401, were exposed in the base of a later feature and are believed to belong to the same group. It is questionable whether the spread of stake-holes was ever more extensive than that exposed. Further to the north and east the subsoil gave way to bedrock and it could be argued that the stake-holes were simply not recognised there, however, at the west end the orange subsoil did continue across to the north edge of the site and there were no stake-holes in the northern half of the nave.

Post-hole 269, 0.26m in diameter and 0.13m deep was probably contemporary with the stake-holes. Another two post-holes, 400 and 420 may belong with this phase but their positioning within the stratigraphic sequence is unknown. In common with the stake holes 401 these two post-holes appeared in the base of later intrusions and so there is no evidence for the level from which they were inserted.

Sealing the stake-holes and the underlying orange subsoil was a very compact orange-brown clay, 265, containing specks and some small concentrations of charcoal. Its extent, as planned, was co-terminous with the orange subsoil but it is possible that it was originally somewhat more extensive having sunk into the cracks and fissures in the limestone at the north east corner of the area. It was ca. 0.1m thick.

On the surface of layer 265 were three small patches of ash, 403, 404 and 406. 403 was an irregularly shaped spread of mixed white ash and charcoal with maximum dimensions of 0.6 x 0.57m and 0.06m thick. It lay close to the south wall of the chancel. A single stake-hole, 402, 0.05m in diameter and 0.13m deep between 403 and the south wall may be associated with the ash. Further to the north and east 404 was a sub-oval deposit of black charcoal mixed with some white ash, ca. 0.3m in diameter and 0.01-0.02m thick. To the west lay white ash 406 which continued into the west section and had a maximum thickness of 0.03m. At a similar level to 403, 404 and 406 was an irregular patch of ash mixed with lumps of burnt redeposited natural clay, 405, ca. 0.02-0.04m thick.

Feature 405 was cut on its northern edge by a linear feature 407=258 which ran straight from the west section on a north-east line towards the north side of the chancel arch. At more or less the point at which the orange subsoil gave way to the underlying bedrock, the feature bifurcated and both arms faded away before reaching the chancel arch. The feature had a marked U-shaped cross section and depth of ca. 0.15m where it cut into the underlying clay 265 but was more irregular in cross section where it cut the rock. 259 a loose grey-brown clay containing the occasional stone formed the fill of the feature. The feature's function is unclear, having the appearance at the west end, at least, of a single ard furrow. The orientation of the feature is completely different to any other within the chapel and only skeleton 62 in Area 1 and skeleton 148 in Area 2 are comparable. If it was intended to mark a boundary line, then the boundary was very short lived because the layers above it are continuous across the feature.

A compact stony-surfaced grey clay layer, 247, varying in thickness from 0.02-0.08m, sealed clay 265 and linear feature 407=258. Although stony the layer did not have the appearance of a deliberate cobble surface for the stones were too widely set for that. A large amount of animal bone and a few sherds of B-ware pottery were recovered from the interface of the layer with the overlying layer 243, so there was a definite occupation surface at this level. There were no associated structures within the nave, though it is believed that the surface could be contemporary with the church further east in the chancel area.

A small patch of white ash, 252, 0.01-0.04m thick lying by the west section was associated with the layer 247. Overlying 247 was a 0.45-0.5m thick homogenous layer of grey-brown clay, 243, speckled throughout with charcoal. The layer extended over the whole of the nave but had been cut so badly by later intrusions that its upper surface survived only on two small upstanding pedestals of deposits. One of these measured ca. 1.2 x 0.8m was located more or less centrally within the nave, while the other was smaller in extent but continued into the west section. The latter was stratigraphically detached from the former by two later intrusions which cut through to the subsoil and it was therefore separately numbered 296. An animal burrow was identified running within the layer from the centre of the site to the south wall and, at one point close to the wall, it seemed to have widened out into a nest. This nest caused the layer and possibly parts of other contexts to subside into it, and the resulting somewhat mixed deposit was separately numbered as context 295. Given the similarities in thickness of layer 243=296 and 209 in the chancel as well as their relative stratigraphic positions, it was felt that 243=296 was probably the continuation of 209 even though their composition was not exactly the same. A later boundary between the two areas, discussed below, may account for the slight differences, as the two areas subsequently developed in different ways.

The two areas where the full thickness of layer 243=296 survived were both sealed by a thin ash deposits; 281 over 296, and 237 over 243. Ash 281, by the section, was sealed by the later general fill layer 220, but a ca. 0.25m thick deposit of friable light grey clay, 280, sealed ash 237. This latter deposit rose almost to the floor level prior to excavation and was the highest surviving point of the pre-chapel stratigraphy.

The foundation trenches for the chapel, 235 and 279, were cut from a point above layer 243=296, which did not survive. The edge of 235, the trench along the north wall, ran ca. 0.4m out from the wall but had been almost totally removed by later intrusions and only survived to anything near its original height at the west section and in the corner by the chancel arch. At this corner the trench turned south to become the foundation trench for the arch and only here was it possible to get a proper picture of the nature of the trench fill. Here loosely packed stone mixed with grey clay, 236, formed the lower fill and was sealed by the upper fill, 222, an even looser stone and decayed mortar layer. Along the south wall the trench, 279, had been largely obliterated by later intrusions which in places had even undermined the wall. The foundations themselves are of uncoursed rubble.

The remaining features all post-date the building of the chapel and, by their intrusive nature, account for the removal of much of layer 243=296. In common with the chancel area, they can be divided into burials and later pits.

292 was possibly the earliest burial within the nave. Its grave-cut, 290, cut both clay 280 and ash 237 on the upstanding pedestal of early deposits in the centre of the nave, but the later intrusion 272 had removed all but the foot of the grave. Only the articulated foot bones, of an adult, remained. The grave fill, 291, was a mixed grey-brown clay flecked with charcoal and containing no trace of a coffin. What distinguishes this burial from the others within and post-dating the construction of the chapel, was the use of two upright flattish stones deliberately set on edge around the feet. One of the stones had mortar adhering to it, showing that it had been reused from a building. There is no firm evidence for the skeletons date other than the fact that it predates intrusion 272 and probably, but by no means certainly, post-dates the building of the chapel. It is, therefore, possible that this is the only medieval burial to have been interred within the building

The remaining burials appear to be eighteenth century or later.

Grave 225 continued into the west section. It contained a poorly preserved wooden coffin, 226, equipped with iron fittings. Within the coffin were the remains of a 60 year plus male(?). The skull of the skeleton was left in situ in the section, while the lower part of the body had been removed, from the pelvis down, by the later intrusion 272. The remains of at least one other individual had been stacked in the grave.

Burial 223 lay close to the foundation trench for the north chancel arch. All that survived was a pair of feet, belonging to an adult, resting on the badly decayed base of a coffin which had been equipped with iron fittings. The rest of the skeleton and coffin had been removed by the later intrusion 240.

Further to the south was feature 285, measuring ca. 2m east-west by 0.7 wide at its maximum. The feature was truncated by the later cut 238 so that there was only any real depth to it at the very west end. The form of this feature was remarkably grave-like though it contained no skeleton, however, its fill, 286, contained some fragments of wood which may have come from a displaced coffin. The feature is, therefore, interpreted as a robbed out grave.

The remaining features performed unknown functions.

238, which cut grave 285, was a large shallow feature measuring at least 2.3m north-south by 1.3m east-west bottoming out within layer 243. It was cut on the north by feature 240 and to the east by feature 277. The fill of the feature, 239, was a loose gritty grey-brown clay containing small stones and mortar. 239 also contained quantity of human bone, fragments of cloth, the sole of a welted shoe as well as fragments of both window and bottle glass. It may well be that the bone, shoe and cloth were part of the missing fill of grave 285.

To the north of 238 lay 240 which accounts for the disturbance of most of the northern half of the nave. It is quite possible that there were several features here running into one but, if so, it was impossible to differentiate them. The feature bottomed out in layer 243 and cut feature 238 and skeleton 223 as well as the foundation trench of the north wall. The fill of the feature was the general disturbance layer 220.

297 cut and formed the southern edge to the upstanding pedestal of layer 243 and was itself cut by intrusions 239 to the east and 272 to the west. It bottomed out within layer 243. The feature continued up to and under the south wall of the chapel where its southern limit was not explored for fear of collapsing the overlying wall. It is assumed that the original width of the feature was represented by the gap in the foundations which was 1.1m. The northern edge of the feature lay ca. 1.7m out from the wall and its base sloped down towards the wall. The fill of the feature, 298, was a grey-brown clay mixed with a high proportion of orange redeposited natural clay with some stone and redeposited human bone.

293 lay in the south-west corner of the area and continued into the west section. Its northern edge cut ash 281 ca. 1.7m out from the south wall of the chapel and it was cut to the east by feature 272. The fill, 294, was a very mixed grey clay containing a high proportion of stone of varying size and mortar

272 was cut right through to the subsoil. It measured ca. 3.2m north-south and 1.6m east-west. The feature had cut and removed parts of skeletons 224 and 292 as well as features 293 and 297. The fill, 273, was a very loose dry grey-brown sandy clay with spots of mortar and stone throughout and a couple of concentrations of slate. The finds from the feature include a fragment of post medieval pottery, brick and bottle glass.

The general layer sealing all the deposits was the same as that in the chancel area, 220. It was an extremely mixed layer and varied considerably in thickness. Cutting 220 was 221 the trench for the electrical conduit which ran from the south doorway and in under the lower step to the north tower.

4.3. The chancel arch area.

The area under the chancel arch was extremely confused because several features coincided to cut and mask each other. The foundation trench for the chapel is cut at just this point by at least one late intrusion and a burial and the late intrusion may have all but removed a much earlier north-south feature the only remaining traces of which are in front of the door to the south tower.

The foundation trench on the north side of the nave, 235, swings southwards round the north-east corner to a point level with the chancel arch where it disappears. To the east of the arch it can be picked up again ca. 0.25m. beyond the arch where the edge coincides with the north-south edge of a later feature 277. This feature, which measures ca. 2.6m east-west seems to account for the missing portion of the foundation trench but defining the northern limit of the feature was almost impossible because some of the fill of the foundation trench through which it cut had collapsed into the feature to form part of the fill of the feature itself. The western edge to 277 on plan looks to be a southward continuation of the edge of the foundation trench 235 but this cannot be because 277 cuts the intrusive features 238 and 285 which are both clearly later in date than the foundation trench. This western edge continued out to the door to the south tower where, as to the north, the feature accounts for the absence of the foundation trench for the nave wall. It would seem however that the western edge to 277 is masking an earlier feature, and it is possible that the line of the west side of 277 was, in part, dictated by the underlying feature. 277 cuts, as stated above late features 238 and 285, but the edge of the cut actually continues in under the south nave wall at least for a hands length thus appearing to predate the features that it cut! It is quite clear then that there must have been two separate features one pre-Chapel and one, 277, of relatively recent date whose west edges coincide. The early feature was numbered 433 and along the inside of its edge were some large flat stones, 299=434. These stones had been deliberately set along the inside of the edge and ran up to ca. 1.4m from the south wall though was never more than one stone wide or high. The full extent of 433 is unknown because feature 277 has otherwise removed all trace of it. The function of this feature is not entirely clear because its full extent is not known. It is not inconceivable that it originally ran right across the nave because on plan it lines up quite nicely with the edge of the foundation trench of the nave where it swings southwards in the north-east corner. The foundation trench here is quite wide and could be quite deliberately so because it is re-utilising an earlier feature. In other words it is possible that there has been a north-south stone feature across the nave at this point which has been almost totally removed.

Within 277 was a burial of a female? child ca. 6 years old. The skeleton 284 was demineralised and it lay within a very poorly preserved coffin 283. The coffin also contained a field mouse skeleton. Remains of a second individual 284b, a sub-adult male ca. 21 yrs old were found in close association. There was no clear cut for this burial except on the south side where it cut the foundations and was numbered 282. It must be assumed that 277 has removed all trace of the cut on the north side.

The fill of 277 was loose voided stone of varying sizes mixed with a high proportion of yellow sand and mortar, context 278. The fill sloped down from the chancel arch towards the west edge of the cut and it is assumed that most of the fill derived from the chapel foundation trenches which have collapsed into the feature. It was also partly filled and sealed by the general layer 220.

4.4. The north tower.

A late medieval arch and steps have been inserted into the original doorway into the tower and, in the east side of the tower, a late doorway leading through into the sacristy in Area 1 has been blocked up. The starting level within the tower corresponded with the top of the steps and was thus somewhat higher than that in the nave. Upon excavation the base of the ashlar facing within the tower was found to correspond to that in the nave.

The whole of the interior of the tower was filled with a voided stone rubble, 432, which rose from the bedrock almost to the base of the ashlar and it, therefore, appears that the whole of the tower area was excavated out for the foundations. The voided stone 432 was topped off with a ca.0.05m thick layer of yellow sand, 431, which ran more or less to the base of the ashlar. Corresponding with the base of the ashlar was a thin, 0.01-0.02m thick, level of sandy mortar, 430. This mortar continued in under the later steps towards the nave and quite possibly once continued into the nave itself. The layer is interpreted as the bedding layer for the original floor, which is presumed, but not proven, to have consisted of stone flags.

The floor level associated with 430 seems to have functioned up until the insertion of the later steps, when it was raised to that of the top step. At presumably the same time, a small window in the east side of the tower and which was part of the original chapel was converted into a doorway through to the sacristy in Area 1. This had been achieved by lowering the base of the window by removing two courses of ashlar and knocking through. A limestone threshold, 424, was then inserted across the inside of the new opening. This new door was itself crudely blocked at a later, unknown, date. The original base of the window opening would have been between 0.9m and 1m above the original floor level.

Sealing the bedding layer 430 was a ca. 0.2m thick deposit of clean loose yellow mortar and stone, 429, which presumably derived from the conversion of the window. Cut into this layer, in the north-west corner of the tower was a rectangular post-hole, 428, measuring 0.3 by 0.3 by 0.3m deep, containing traces of the wood. The function of the post-hole is unclear. A mixed layer of mortar, grey clay and stone, 427, sealed layer 428 and post-hole 429, and was itself sealed by a grey-brown clay, 426, which contained some mortar and small stone. The surface of 426 was more or less level with the upper step and this was interpreted as the upper level of the build up associated with the steps. The layer above 426, 423, was very compacted and laminated and seemed to have been an earth floor which had built up over a period of time. Layer 423 and some of the lower layers were cut by context 422 which was the continuation of cable trench 221 into the tower. Both cut 422 and layer 423 were sealed by a thin but exceedingly mixed deposit 421 which is presumed to have derived from the work carried out by the OPW in the overcroft of the chapel, when debris was thrown down the inside of the tower. A compacted grey earth, 425, lay between the blocking of the doorway in the east side of the tower and the threshold 424.


From the descriptions of each area as outlined above it is clear that we are dealing effectively with four separate excavations without any direct stratigraphic links between them, i.e. Areas 1, Area 2 and Area 3 nave and chancel. It is therefore necessary to make a few, hopefully informed, assumptions in order to put the whole site into context and tell a meaningful story. The purpose of this short section is therefore to outline those assumptions, so that if anyone is unconvinced with the conclusions below then they have a starting point for deconstructing the story.

1) The burials in the Area 1 phase 2 graveyard are parallel with the proposed early church in Area 3 and are therefore taken to be broadly contemporary with each other.
2) The lower deposits in Area 3, either side of the chancel arch, are deemed to have a common origin. There is admittedly a difference in that the chancel deposit contain a higher level of charcoal, but this is believed to result from different usage of two areas where material was accumulating in one manner.
3) Similarities between the lower stratigraphies of Areas 1 & 2 despite no direct link.

The standing remains on the Rock represent only two of at least four churches known or suspected from the historical sources. Excavation has added a further two to that number and so it was initially thought that the four buildings on site might be the four in the historical sources but the correlation does not work. From historical sources the four churches are, in chronological order, 1) the church which became the cathedral church at the Synod of Rathbreasail in 1111, and which was, possibly, built shortly after the hand-over of the Rock to the church in 1101, 2) Cormac's Chapel built between 1127 and 1134 according to the annals, 3) the Romanesque cathedral of 1169 and 4) the Gothic cathedral. The excavated churches in chronological order are 1) the wooden church under the nave of Cormac's chapel, 2) the stone church under the wall of the Gothic cathedral, 3) Cormac's Chapel and 4) the Gothic Cathedral. The implication of this is quite simply that what we call Cormac's Chapel is in fact the 1169 Romanesque cathedral. This hypothesis was tested with several architectural historians but the consensus seemed to be that the style of the church fitted better at the traditional earlier date. That date for the chapel is, therefore, accepted but, if the 1101 date marks the beginning of church activity on site, then there is a great deal happening between 1101 and 1127. There are four separate phases of graveyard and at least two churches that have to fit into that period. The fact that the graveyard associated with the first stone church extends southwards across the site of the wooden church suggests that the latter had possibly disappeared from view when the churchyard was laid out over it. The stratigraphy would seem to support the idea of some sort of hiatus between the two churches. The associated occupation surface to the west is covered by a thick homogenous layer relatively devoid of finds with a similar build up against the posts of the church, both of which seem to be a product of a lack of activity on the site. If these observations are correct then 26 years is too short a span to accommodate all the events. The conclusion therefore is that the constraint of 1101 has to be disregarded and that the early church predates 1100. Unfortunately the pottery associated with the early occupation surface in the nave has yet to be positively identified, despite showing it to several authorities on pottery, and so an exact date for the church cannot yet be put forwards on that basis, however the presence of B-ware from the later graveyard deposits does hint of activity on the Rock reaching as far back as the 6th century.

A hypothesis for the date of the church can, however, be put forwards on the historical evidence. The 9th and 10th centuries saw a number of king-bishops of the Eoghanacht dynasty based at Cashel and it is not unreasonable to suppose that they had a church on the site. With this in mind it may be that the collapse of the dynasty and its replacement by the Dal Cais accounts for the apparent period of disuse. If we are seeing political events reflected in the archaeological record then the early church is pre-10th century. There is then something of a gap in the use of the site until it is handed over to the church in 1101 when the second church (wall 67 in Area 1) was built.

Taking the site as a whole the following phasing is therefore suggested. The dates of the early phases are tentative at the time of writing.

Phase A. An initial possibly secular use of the site. This would include the stake holes in area 3 and parts of the lower stratigraphy in areas 2 and 3. There appears to be limited interment towards the end of the phase in area 1, phase 1 of the graveyard, and the single early burial in area 2. Suggested date 6th-9th century.

Phase B. The wooden church was established in area 3 and the phase 2 burials in area 1 are orientated upon it. Further to the west the area is open and free of constructions, though in area 3 it was possible to identify a contemporary occupation surface. All the burials are restricted to the east end of the site. There is a substantial build up of deposits over the occupation surface and against the posts of the church. This suggests quite a long period of time for the use of the church. Suggested date pre-1100, possibly 9th-10th century

Phase C. The wooden church has fallen out of use and is replaced by a stone church lying under the present cathedral. Phases 3 and 4 of the graveyard are orientated upon this new building and the graveyard is allowed to extend further south into the area once occupied by wooden church. The western end of the site again appears to have been an open area, with possibly a paved area in area 2. The circular stone feature in area 2 possibly belongs to this phase. There is the possibility of a more formal division between the church/graveyard area and the open area being introduced. Suggested date ca. 1100-1127.

Phase D. The construction of Cormac's Chapel and its early use. The area to the west of the tower is taken into use for burials and phase 5 of the graveyard in area 1 begins. 1127-mid 13th century.

Phase E. The building of the cathedral and continuation of phase 5 of the graveyard in Area 1. The area to the west of the north tower now becomes a cul-de-sac and is not used for further burials. Mid 12th century to ca. 1650. The shrine in the north doorway of the chapel was probably inserted at this time.

Phase F. Modifications to the chapel, insertion of new door and steps into north tower, adaptation of the window to a door and the building of the sacristy in area 1. Phase 5 of the graveyard terminates with the building work. ca. 1650.

Phase G. Use of the interior of the chapel for burials. Ca 1650-1800.

Phase H. Modern disturbances 1800+.

The identification of the early building as a wooden church of possibly two phases rests on the fact that the phase 2 burials in area 1 are aligned upon it. The size and form of the church is debatable however for a number of reasons. Firstly it is not clear if the row of post-holes which constitute the building continued further east under the chapel wall and secondly it is not absolutely definite where the west end lay because the western post-hole could in fact be the product of the later intrusion 277. If one accepts that the four post-holes constitute the complete row, then the building is ca. 4.2m east-west, whereas if the western post-hole is excluded it is 2.6m. Both dimensions fall within the size range known for early wooden churches (cf. plans in Mytum,1992 p.87). What is also not clear is whether the row represents the north or the south wall of the building because there are no return lines at all, however, it is assumed to be the south wall. If it had been the north wall it is likely that the associated graveyard, phase 2 of area 1, would have continued some way into the nave therefore the fact that it does not suggests that it is the south wall. The presumed parallel row has been obscured by the later foundations of the chapel and there is no trace of the return in area 2. The missing north-south dimension cannot have been less than 2.5m or else the eastern post-hole of the north wall would have fallen within the nave, nor can it be greater than 4m because of the southernmost skeleton in phase 2 of area 1. The fact that the building is not less than 2.5m north-south suggests that the east-west dimension is more likely to be the 4.2m than the 2.6m given that the ratio of breadth to length is usually roughly 1:1.5 rather than the 1:1 of the 2.6m.

One feature of the site deserving examination in some detail is the marked division between the eastern and western half of the site from the establishment of the first church up to the building of Cormac's Chapel. This division runs roughly north-south just to the west of the chancel arch and continues out under the cathedral to the north. Of all the burials that pre-date Cormac's Chapel only one lay to the west of this line, the somewhat anomalous 148. There were no recognisable features to suggest that there was an actual physical boundary in the early stages but later there is some indication that a more formal boundary was established. The cut, 123, carrying the line of the west side of the north tower out to and under the Cathedral wall may be one part of this. The stone fill of the feature certainly continued too far north to have been a foundation so there must be some alternative explanation for its existence which may be that an earlier feature was re-utilised for the foundation of the tower. The skeletons of phase 3 at the west end of Area 1 seem to respect a line which must have been in existence prior to the Chapel. A third element was feature 433 on the south side of the nave which appears to have also been a north-south feature which was masked by later activities. If these features can be interpreted together as evidence for a formal division, then a development of the site something akin to that noted by the late Prof. Fanning at Reask can be envisaged. There the division between sacred and profane was introduced in the later stages of development of the site. One can wonder if the actual location of Cormac's Chapel itself is a symbolic harking back to the earlier division, with the chancel arch being deliberately placed on the dividing line to separate the laity in the nave from the clergy in the chancel. The site of the early wooden church may have been obscured by 1127 but that does not mean that the formal division had disappeared too.

Given the royal connections of the Rock it is perhaps surprising that the finds are neither numerous nor of particularly high quality. There was no evidence for any high status crafts such as metalworking while the low status combmaking was the only craft to leave any definite trace, and even that is restricted to a single offcut from a tooth plate and a few pieces of cut antler. The only finds to give any hint of the importance of the site are the few sherds of early pottery, some of which are B-ware and others remain to be positively identified. The area investigated is, however, only a tiny fraction of the total area enclosed and so it is quite possible that any handicraft or living quarters were located sufficiently far away that the material was not spread into the site.

6. THE FINDS (FIGS. 14-21).

The finds are listed by registration number with the context number (c.) being given in the description. All finds have the site prefix 92E202: which is omitted from the following lists. Each area was given blocks of context numbers, so the context number defines the area the find is from. These blocks are as follows:-
Area 1 contexts; 1-99, 300-399 and 500-506.
Area 2 contexts; 100-149.
Area 3 contexts; 200-299 and 400-435.



Early Wares.

The early pottery has been examined by Dr. Richard Warner of the Ulster Museum and much of it identified as B-Ware (Thomas, 1959). Four sherds are of a type not paralleled by Thomas but similar to Warner's Clogher type B3. A single sherd of grey ware, suspected of a Romano-British origin was sent to England for identification by Dr. Jerry Evans. Dr.Warner's and Dr. Evans's comments form the basis for the following descriptions.

10. single bodysherd from c.8; 74 single bodysherd from c.21; 143 two bodysherds from c.22; and seven bodysherds from c.68. Probably Bii amphora. All part of the same vessel, some from the fill of pit 333 the remainder from the upper graveyard level of Area 1 associated with the level from which pit 333 was cut.

146. Body sherd, with grooving. Identical to Warner's Clogher type B3 but not paralleled by the Thomas types. Probably a globular amphora. Date 6th century. From the upper graveyard fill, c.69, in Area 1. Fig. 14.

148. Body sherd, ribbed. Thomas type Bii amphora, mid 6th century. From c.138, the fill of the cut into bedrock in Area 2. Fig. 14.

150. Body sherd, grooved. Thomas type Bi amphora, mid 6th century. From the general fill c.220, in Area 3. Fig. 14.

164. Bodysherd. Probably Clogher type B3. From c.243, the thick deposit overlying the early occupation surface in Area 3.

165. Two conjoining bodysherds. Similar fabric to 164. From c.247, the early occupation surface in Area 3.

166. Two conjoining body sherds from the shoulder of a closed form; exterior decorated with two horizontal grooves and wheel burnished. Hard mid-grey fabric with occasional sand and some fine rounded ironstone? inclusions. Dr. Evans writes,"the most probable origin for the sherds is Romano-British, although they might possible be Gallo-Roman. A Merovingian or Carolingian origin could be possible, but the fabric is much finer and better manufactured than those of that period seen by the writer". From c.247, the early occupation layer in Area 3, the same context as 165. Fig. 14.

172. Handle. Probable Bii amphora. From a possible pre-graveyard layer, c.342, in Area 1. Fig 14.

173. Body sherd. Probable Clogher type B3. From the early clay layer c.390 in Area 1.

174. Body sherd. Bii amphora. From the same layer as 173.

French medieval.

22. Body sherd, off white fabric with external yellow-green glaze. French. From c.102, the cathedral foundation trench in Area 2.

English medieval.

149. Everted rim, early Ham Green cooking ware? From the general disturbance layer c.220, in Area 3. Fig. 14.

154. Two conjoining body sherds, pale grey fabric with limestone inclusions, with buffish exterior surface where unglazed, external glaze clear and green. Ham Green. From same deposit as 149.

157. Small body sherd, off-white to creamy-buff sandy fabric with one red inclusion, external green-brown glaze run. From same deposit as 149.

175. Body sherd. Fine dark grey fabric with external green glaze. From the backfill c.427 within the tower in Area 3.

Local medieval wares.

Type 1. Highly fired reduced fabric with sparse quartz and occasional limestone? inclusions and external green glaze.

151. Six conjoining body sherds of medieval ware giving a profile from base to just above the base of the strap handle. The strap handle has four incised vertical lines; the base is thumbed and there is some rilling especially above the handle. Fine mid grey sandy fabric with some occasional fine grained white inclusions external green glaze becoming sparse below the handle. Fig. 15. From the general disturbance layer, c.220, in Area 3.

152. Bodysherd with rilling. Dark grey sandy fabric with occasional sparse fine white inclusions and lighter grey at the surfaces. External dark green fabric. From same context as 151.

153. Two conjoining body sherds of medieval ware. Dark grey sandy fabric with occasional small white inclusions. The inner surface is a fawn colour while the outer surface is a dull red colour where not covered by green glaze. From same context as 151.

161. Two conjoining sherds of dark grey fabric with white inclusions and lighter grey surfaces, external dark green glaze. From same context as 151.

162. Bodysherd, dark grey fabric with slight fine white inclusions, external green glaze with reddish surfaces where glaze is absent. From same context as 151.

163. Bodysherd, dark grey fabric with white inclusions and external dark green glaze. From same context as 151.

Type 2. Reduced quartz free fabric, less highly fired than type 1, and with some oxidation of the inner margins.

73. Bodysherd, dark grey core with orange internal surface and external green glaze. From context c.21, one of the fills of pit 333, in Area 1

142. Three conjoining body sherds, grey sandy fabric with external green glaze. From c. 22, one of the fills of pit 333 in Area 1.

144. Three body sherds and part of a strap handle, a) and b) have a dark grey fine grained fabric with external dark green glaze, c) is similar except that the internal surface has a buffish colour, d) is a part of a strap handle in similar fine grained fabric though the surfaces have not been reduced to the same extent as a)-c). The handle is decorated with a vertical incised line with short vertically stab marks to either side. Handle conjoins with 170. From the same context as 142.

155. Body sherd, fine dark grey fabric with external green glaze. From the general disturbance c.220 in Area 3.

156. Two body sherds with very dark grey fine grained core and lighter buff orange surfaces and external green glaze. From the same context as 155.

170. Rim/ handle sherd and a body sherd. The strap handle meets the rim just below the lip and has vertical stabbed decoration. Sandy grey core with orange-buff surfaces patchy green glaze. Handle conjoins with 144. From c.329, one of the fills of pit 333. Fig. 14.

171. Base/body sherd, with thumbing around the base. Grey fine grained fabric with occasional white inclusions, orange buff external surface with occasional glaze spots. From the same context as 170.

Type 3. Oxidised sandy red-orange fabric with clear glaze.

21. Base of a strap handle, light grey sandy core, with orange-buff internal surface and external green glaze. From rubble deposit c.3, in Area 1.

158. Two body and one base/body sherds fine orange fabric with spots of clear glaze. From the general disturbance c.220 in Area 3.


147. Base/body sherd of Gravel Tempered Ware From the later disturbed layer c.122 in Area 2.

159. Bodysherd Creamware, from the general disturbance level in Area 3.

160. Two body sherds of Brownware, from the same context as 159.

168 Flat everted rim, bright orange-red fabric with internal green glaze. From c.294, the fill of the late intrusive feature 272 in Area 3.

169 Body sherd. Bright orange sandy, gritty and slightly micaceous fabric with grey slip on the external surface. From c.273 the fill of the late intrusive feature c.272 in Area 3.

271. Body sherd, sandy white fabric with internal and external white glaze and some blue paint externally. From context 60, the rubble deposit sealing Area 1. N.B. this sherd appears to be relatively modern and must be intrusive.


167. Fragment of crucible? From the pre-chapel layer c.260 in Area 3.

Clay Pipe.

11. Stem fragment from the modern layer c.2 in Area 1.
15. Stem fragment from the graveyard layer c.201 in Area 3.
16. Stem fragment from the upper graveyard deposit c.9, in Area 1.
212. Three fragments of stem, one with a diamond pattern, from the general disturbance in Area 3.
213. Fragment of a plain bowl from the same context as 212.
267. Two fragments of stem, from the same context as 212.


269. Fragment of red brick from c.122 in Area 2.
270. Fragment of red brick from c.273, the fill of the late intrusion 272 in Area 3.
461. Half of a yellow brick; width 92mm by 54mm thick. From the rubble layer c.60 sealing the graveyard, in Area 1.



1. Silver short-cross penny. King John, Canterbury mint, Coldwine the moneyer. Ca 1200-1210 AD. Found on the surface of c.209 in Area 3.
84. Bronze halfpenny of Queen Victoria 1897. Found in the rubble deposit c.119 in Area 2.

Copper alloy.

Stick pins.

4. Watchwinder head with slight collar at neck. Length 59mm. From the upper deposit c.201 in Area 3. Fig. 16.
85. Watchwinder head with stabbed decoration below the head. The shank thickens towards the centre. Length 100mm. From c.121, in Area 2. Fig. 16.
86. Kidney headed pin, with interlace decoration, length 82mm. Found in clay layer c.105, in Area 2. Fig. 16.
87. Shank. From c.128, the grave fill associated with skeleton 129 in Area 1. Fig. 16.
88. Shank, from the general disturbance layer c.220 in Area 3.
89. Shank, from the stony surfaced layer c.247 in Area 3.

Shroud pins.

All the pins are ball-headed and lengths are given for all complete specimens. The pins are associated with the post-medieval burials. A representative sample is illustrated in fig. 16.

2. 16mm in length From c.203 associated with skeleton 204.
76. 25mm. From c.125 close to the tomb niche.
77. 4 pins, a)25mm, b)23mm, c) in three pieces d) head only. From c.203 associated with skeleton 204.
78. 4 pins, a) 22mm with traces of silk, b)-d) fragments. From c.208 associated with skeleton 210.
79. 5 pins, a)34mm, b)24mm, c)26mm, d)25mm and e)24mm. From c. 220, the general disturbance layer.
80. 2 pins, a)32mm, silvered? and b) 31mm but in two pieces. From c.221. the cable trench. Fig. 16.
81. 7 fragments of at least 3 pins. From c.227, associated with skeleton 228. From the green staining on the skull it is possible to show that the shroud was pinned at the temples and on the lower jaw by the molars.
82. 3 complete and 5 fragmentary pins all with traces of silvering a)24mm, b)24mm and c)22mm. Associated with skeleton 233.
83. 3 complete and 4 fragmentary pins a)25mm, b)25mm and c)22mm. Associated with skeleton 284 in coffin 283.

Other bronze objects.

3. Ring. Plain band from the right hand of skeleton 30. Fig. 16.
5. Part of a large pin-type object. The D-shaped cross section has a half loop near the end away from the point. The flat surface has irregularly spaced diagonal saw? cuts. The context in which this was found, c.201, was quite recent and it need have no great antiquity. Fig 16.
90. Tacks. 32 tacks set into the inside face of coffin 203; some with textile still adhering. Heads ca. 5mm. in diameter; length 12mm. Fig. 16.
91. 9 tacks. 5mm in diameter and length 12mm. From coffin 241.
92. 13 tacks of same type as previous, found as a group together in the south side of coffin 241.
93. 11 tacks of the same type found in the north side of the coffin 241.
94. Tack. Single tack holding a fragment of silk to the wood of the coffin 208.
95. Buckle, from c.101.
96. Fitting? 43mm diameter hemispherical object, with opposed flanges, the one perforated (for attachment?) the other broken. Found in the general layer, c.220, on the south side of the nave. Fig. 17.
97. Spacer? Thin circular cross sectioned and slightly curved rod, ca. 55mm long, with flattened terminals each with a notch in the end. The central portion is slightly flattened and perforated. From c.220. Fig.17check.....................a
98 fitting? D-shaped cross section, hooked at one end and slightly upturned at the other. From c.20. Fig 16.
99 tack head. 12mm in diameter. From c.222
100 ring. From c.119.
101 wire. 42mm long scrap of rectangular cross sectioned wire, from c.220.
102 horse equipment/fitting? Fragment of a larger object. Quadrant of what was presumably a full circle, with projecting spur and flange at the quarters. From c.220 Fig. 17.
103 Handle. From c.427 (check if this is really the one in the drawing???????????????????
268 stud. Dome headed stud 17mm in diameter, from c.220.

The remaining pieces are offcuts and amorphous lumps.

135 (c.203), 136 (c.220), 137 (c.225), 138 (c.239), 139 (c.243), 140 (c.261) and 141 (c.265).


Coffin fittings.

All coffin fittings come from the late coffins in Area 3. There are three distinct elements apart from coffin nails. The types are illustrated in fig 18.
1) handles, some of which still have their plates for affixing to the side of the coffin.
2) flat base pieces which are set regularly around the side of the coffin at the base angle. The pieces are trefoil shaped with one arm slightly longer than the others, set pointing up. There is usually a triangular perforation at the centre.
3) Ridge pieces consisting of two conjoined trapezoidal plates. When found in a coffin they are in a position which suggests that they were set on the central ridge of a pitched lid.

8. Base piece from next to skull of 211, probably from coffin 203.
26. Four coffin handles from c.201.
27 Two coffin base pieces from c.201. One is illustrated in Fig. 18.
29. Top piece from coffin 208 for skeleton 210. Fig. 18.
30. Five base pieces from coffin 208.
32. Two handles, one with attachment plate, and several base pieces, part of coffin 204.
249. Handle with attachment plate, from coffin 203.
250. Twenty base pieces and thirteen nails from coffin 203.
251. Handle, top piece and seven base pieces together with eight nails, from coffin 208.
252. Six handles, one with attachment plate. From the general disturbance layer c.220.
253. Nine base pieces and two top pieces, from c.220.
257. Two handles with attachment plates, from coffin associated with skeleton 223. One is illustrated on fig. 18.
258. Thirty base pieces, from coffin associated with skeleton 223.
259. Handle, four base pieces and two nails from coffin 226 for skeleton 224.
263. Two top pieces, from c.220.
264. Top piece from grave cut c.225, for skeleton 224.
265. Top bit and two nails from coffin 289.
266 handle, five base pieces and two nails from coffin 289.

6. Shank from c.212.
13. Modern nail with round head and shank, length 76mm, from c.2.
28. Two nail shanks, 76 and 43mm, from c.201.
31. 3 coffin nails, a) 60mm, b) 26mm and c) 20mm, from c.208.
72. 2 nails, lengths 60 and 43mm, from c.21.
236. 4 heavily encrusted shanks, from c.20.
241. 3 nails with small sub-oval heads (a-c), and one shank (d). Lengths a) 82mm, b) 72mm, c) 46mm and d) 80mm, from c.60.
246. 4 nails with small sub-oval heads (a-d). Lengths a) 56mm, b) 37mm, c) 37mm and d) 42mm, from c.122.
248. 5 shanks lengths 78, 68, 47, 51 and 45mm, from c.122..
254. 30 assorted nails, from general graveyard fill c.220.
255. 2 nail shanks from c.221.
260. 5 small coffin nails, from c.233.
261. Shank?, length 101mm, from c.243.
288. Shank, length 49mm, from c.37.

Other iron objects.

235. Key; shank with broken D-shaped end. Length 136mm. From c.68 the interface between graveyard truncation and overlying rubble. Fig. 18.
237. Rectangular flat plate 38x38mm. From c.20
239. Triangular flat plate, height 79mm. From c.57
240. Ferrule?, curved cross section, length 40mm diameter 20mm. From c.57.
242. Encrusted flat plate, length 47mm. From c.69.
243. Rectangular cross-sectioned rod, slightly curved, length 42mm. From c.119.
244. Triangular cross-sectioned rod, slightly curved, length 47mm. From c.119
245. Rectangular cross-sectioned rod, length 63mm. From c.72
247. Curved rectangular cross-sectioned rod, length 100mm. From c.125.
256. Bracket? flat object upturned at one end and coming to something of a point at the other. From c. 221.
262. Part of a door-hinge? flat bar with two perforations. From c.427.


All are fragments of window lead.
214. One fragment., from c.20.
215. Four fragments, from c.122.
216. Three fragments, from c.220.
217. One fragment, from c.230.
219. One fragment. from c 239.



104. Fragment of a single sided composite comb. Fragment of connecting plate with one tooth-plate lacking the teeth and two iron rivets. Decorated. From the upper graveyard fill, c.69. Fig. 19.
105. Fragment of a single sided composite comb. Fragment of undecorated connecting plate with two tooth-plates held by side rivets. From the trample, c.122 at the base of the major disturbance. Fig. 19.
106. Fragment of a double sided composite comb. Single tooth plate with one iron side rivet. From c.295, the layer of disturbance caused by animals. Fig. 19.
107. End tooth plate of a single sided composite comb of a flared end type of comb. It has a raised solid zone, length 33mm. From c.365 one of the layers partly sealing a phase 1 grave. Fig. 19.
108. Crutch headed pin. From c.259, the early linear feature in the nave. Fig. 19.
109. Spindle whorl made from the head of a femur. Diameter 40mm, diameter of perforation 10mm. From c.72, the graveyard fill. Fig. 19.
110. Lathe-turned bone ferrule or terminal, length 23mm and maximum diameter 20mm from c.222, the top of the fill of the foundation trench for the chapel. Possibly from a slightly disturbed part of the layer. Fig. 18.
111. Disc. Plain bone disc with a central perforation. 12mm in diameter; 2mm diameter of the perforation. From c.125, under the tomb niche. Fig 19.
112. Die. Parallelepiped antler die, with ring and dot values. Each end is a 5 while the others are unclear but possibly 6, 2 or 3, 5? and 4? From c.314, one of the lower pre-graveyard layers. Fig 20.
113. Bone toggle. 41mm long section of long bone with cut rounded and polished ends. It appears to have been deliberately dyed black. From c.122 the trample at the base of the large intrusion into the area.

Antler and bone waste.
114. Section of tine lacking the tip; beam end trimmed. Combmaker's waste.
115. Tine tip with chamfer, combmaker's waste?
116. Offcut tine with chamfered tip and notch on the inner surface. Combmaker's waste?
117. Tine tip with chamfer and notch, combmaker's waste.
118. Offcut tine tip with saw cut towards the beam end. Width of cut 1.5mm.
119. Tine tip with knife cut at one end.
120. 3 tine tips, from c. 69.
121. Offcut tine tip missing the very tip, from c.72.
122. Tine tip, from c.94.
123. Tine tip, from c.121.
124. Tine tip, from c. 122.
125. 3 tine tips, from c.220.
126.Offcut tine tip, c.291.
128. 2 tines, from c. 220.
129. 3 fragments of waste, one of which is just the cancellous tissue, from c.72
130. Shed burr, from c.72.
131. Shed burr, from c.94.
132. Shed burr, from c. 342.
133. Offcut from a tooth plate after the comb was assembled, from 342.
134. 4 fragments of tine waste, from 342.
462. Fragment of cut down long bone, from c.114.
463. Fragment of antler split longitudinally, from c.342.


All the fragments of leather are of post-medieval date.
221. 8 fragments of welted shoe sole, from the general disturbance layer, from c.220.
222. 4 fragments of welted shoe upper, from c.220.
223. 1 fragment of hobnailed sole, from c.220.
240. 1 fragment of welted shoe sole from c. 239 the fill of pit feature 238.


The textile fragments are all fragments of post-medieval coffin linings.
9. 2 fragments of wood both with fragments of silk held in position with copper alloy tacks. Part of coffin 203.
220. Several large fragments of silk from c.239, the fill of pit feature 238.


25. Small fragment with reddish brown and dark brown paint, from c.201.
225. Approx. 30 fragments with red-brown paint, from c.220.
226. Large piece with yellow white and black decoration partly covered by a lime wash, which is probably part of the fresco currently being restored. From c.220.
227. 4 small fragments with red-brown paint, from c.222.
228. 1 fragment with red-brown paint, from c.239.
229. 1 fragment with red-brown paint, from c.278.
272. 2 fragments one with burnt and vitrified surface, from c.22



18. 3 fragments with tool marks, from c.3.
19. Fragment of a limestone pillar base, from c.3.
20. Fragment of sandstone pillar base, from c.101.
306. Fragment of sandstone moulding, from c.220.
307, 308 & 309. 3 fragments of sandstone mouldings, from c. 119.
464. Fragment of a limestone pillar, from c.119.
465. Fragment of a limestone pillar with mortar adhering which shows that it was reused before ending up in the backfill, from c.119.
466. Fragment of a limestone gulley, from c.119..

Roof tile.

17. 2 fragments with hour-glass perforation, from c.201.
23. 2 fragments one with perforation, from c.101.
230. Fragment with hour-glass perforation, from c.119.
231. Fragment with hour glass perforation, from c.122.
302. 3 fragments with hour glass perforation, from c.220.
303. 2 fragments of roofing slate, from c.220.
304. Fragment with hour glass perforation, from c.278. Fig. 21.


7. 2 fragments of flint, from c.7.
232 Fragment of flint, from c.122.
233. Spindle whorl 37mm in diameter with 8mm diameter hole. From c.243, the base of the thick deposit overlying the early occupation surface. Fig 21.
234. Whetstone? Fragment of fine grained stone apparently worn on one side, from c.365.
310. Fragment of a piece to receive a door jamb? Part of a rectangular limestone block with notched corners and a central square perforation. From c.143. Fig. 21.
311. Whetstone? Fine grained black stone 80x52x30mm, from c.247.
312. Fragment of a sandstone quern, from c.221. Fig 21.



218. Plain white bead, length 5mm diameter 7mm. From, c.221 the modern cable conduit trench.

Window glass.

There are two types of window glass. Type 1 was, on average, ca. 2mm thick clear glass which has become pitted and encrusted over time. The pieces show evidence of cutting and chipping to shape, and some is painted in a reddish brown. Although found exclusively in the disturbed post-medieval contexts, the type is medieval in date, and the decorated pieces are illustrated in Fig. 22. Type 2 was thinner and of a higher quality pale blue or green glass with no pitting or encrustation. Type 2 is undecorated and believed to be of post-medieval date and is not illustrated.

Type 1.
14. 3 plain fragments, from c.201.
24. 2 plain fragments, from c.201.
176. 2 plain fragments, from c.203.
178. Single fragment of cobalt blue glass with dark painted lattice design, from c. 220. Fig. 22.
179. 8 decorated fragments, from c.220. Fig 22.
180. 29 plain fragments, from c.220.
183. 2 plain fragments, from c.222.
184. Single plain fragment, from c.225.
189. Single plain fragment, from c.241.
191. Single decorated fragment, from c.278. Fig. 22.
192. 2 decorated fragments, from c.278. Fig. 22.
193. 5 plain fragments, from c.278.
195. Single plain fragment, from c.294.

Type 2
70. Single fragment, from c.16.
71. Single fragment, from c.21.
177. Single fragment, from c.213.
181. Apprrox.. 85 fragments, from c.220.
182. Single fragment, from c.221.
185. Single fragment, from c.225.
186. 11 fragments, from c.230.
187. 6 fragments, from c.230.
188. 2 fragments, from c.239.
190. Single fragment, from c.265.
194. Single fragment, from c.273.
196. 10 fragments, from c.294.
197. 4 fragments, from c.22.
198. 3 fragments, from c.60.
199. 5 fragments, from c.119.
200. 48 fragments, from c.122.
201. 12 fragments, from c.125.
202. 5 fragments, from c.128.

Bottles and vessels

12. Bodysherd thick green glass, from c.2.
203. Fragment of green/brown glass, from c.36.
204. Small curved fragment of pale green glass, from c.119.
205. 3 fragments of green bottle including part of a base, from c.122.
206. 3 clear glass bottle fragments, from c.220.
207. 3 green glass bottle fragments, from c.220.
208. Fragment of a modern drinking glass, from c.220.
209. Small fragment of green glass bottle, from c.222..
210. Fragment of clear glass, wrongly attributed to c. 239, probably from c.329.
211. Fragment of green glass bottle base, from c.273.



273 Fragment from a stone in c.291 which was used to line grave c.290.
278 Fragment from c. 40, part of the pre-Cormac's Chapel greveyard fill.
279. Several lumps from the early post-hole c.501.
280. From the early mortar layer c.369.
281. Taken from the foundation under the north chancel wall on the interior.
282. From c. 429 inside the tower.
283. Fragment from c.40, part of the pre-Cormac's Chapel graveyard fill.
293. Taken from the junction of the ashlar and foundations in the nave, approx. 0.7m west of the door into the north tower.

274 (c.252), 275 (c.390), 276 (c.392), 277 (c.403), 284 (c.125), 290 (interface c.127/134), 291 (c.55), 292 (c.342), 294 (c.265), 295 (c.244), 296 (c.264), 297 (c.405), 298 (c.406), 299 (c.404), 300 (c.237) & 301 (c.261).


289 oak stake, c.315B.
287 oak stake c.354.
285 fragments of coffin edge, c.356.
288 oak stake, c.367.


I would like to record my thanks to Christina Forrest who acted as assistant in 1992, Jo Moran who performed the same function in 1993, the excavation team of Cara Murray, Frank Ryan, Delice Osborne, Seamus Taaffe, Miriam McAlinney, Clodagh O'Neill and Emma Quinn, and to Hugh and Michelle Hyland who completed the bone washing in Castleconnell. My thanks also to all the staff of the OPW at Cashel without whose full help and co-operation life would have been much harder. Rhoda Cronin drew all the finds except for the glass, which was done by Patricia Johnson. Finally a thank you to Dr. Richard Warner and Dr. Jerry Evans for commenting on the early pottery.