Indicated — Carrickmines Great

Watermill associated with Carrickmines Castle (DU026-005001-) and bawn (DU026-005002-) which is located in the townland of Carrickmines Great (Carraig Mhaighin Mhór), in the parish of Tully in the barony of Rathdown. A millrace runs from the site identified on the OSi 1st edition map as 'site of an ancient mill' to the N of Carrickmines castle. The Draft Conservation Plan stated that; 'In addition to the defences, Carrickmines Castle [DU026-005----] functioned as a working manor throughout the medieval period. Excavations uncovered abundant evidence in the form of structures and artefacts reflecting this activity. Evidence for a series of four medieval structures (probably houses, with one probable industrial building containing furnaces), a thirteenth-century horizontal water-mill and a pair of corn-drying kilns were also uncovered during archaeological excavations' (Bolton 2015, 32). Breen (2012, 44-5, 76-87) recorded the following details about the excavated mill at Carrickmines; 'The earliest medieval features excavated at Carrickmines were located in the area immediately south-west of the unexcavated ‘island’. These features have been divided into two phases of undefended pre-castle structures, comprising two possible building structures with related kiln (medieval phase 1) and a horizontal mill with associated features (medieval phase 2). The features which compose the horizontal mill, with its accompanying head- and tail-races, provide a clear stratigraphic watershed. The mill phase truncated two specific areas of early activity. One was represented by the kiln (features F868, F867 and F675) and the other was a large linear boundary ditch (F82). Whereas the kiln has been interpreted as belonging to the earliest phase of activity, associated with Structure A to the Southwest (phase 1), linear feature F82 appeared as an isolated feature which truncated Structure A and was cut by the headrace of the mill and therefore separated phases 1 and 2. The second medieval phase was represented by a horizontal mill, the remains of two buildings overlying each other on the same site, known as Structure C, and another medieval house, Structure D. The mill Although the superstructure of the mill had been removed or destroyed after it went out of use, three below-ground elements survived, although truncated by later features. These consisted of the head-race, which directed the water towards the wheel, the wheel-pit in which the wheel was situated, and the tail race, which channelled the water back into the river. The Headrace The headrace of the mill diverted water from the Glenamuck river to the wheel-pit of the horizontal mill. Two distinct phases were identified; both phases had the same overall structure and the second phase almost completely overlay the first. This first phase of the headrace consisted of three separate features: a feeder ditch (F74), an earth-cut sluice funnel F1436, and the race proper F783. Most evidence of the physical relationships between these three features had been destroyed by the second phase re-cuts. The headrace extended from the top of the South field in the west, through Trenches T12, T7, T1, T5, finally entering the wheel-pit in T10. The second phase of the headrace incorporated a number of re-cuts of the original components, such as the addition to the funnel of a stone-built sluice F1435, two re-cuts of the feeder ditch (F223 and F229) and a re-cut (F94) of the original headrace F783. The total length of the headrace (including all the later re-cuts and features) was approximately 77.00m. The width of F74 was approximately 4.00m, but F94 and F783 were approximately 1.50m wide. After this point the width of the headrace remained narrow, with occasional wider sections which appeared to be more random than intentional. The alignment of the headrace curved from WNW-ESE in the west to ENE-WSW further to the east and was respected by the large moat F75 (located between 1.10m and 3.00m north and northwest of the millrace). This suggests that even if the latter feature is later in date, the course of the headrace would at least have been noticeable at the time when the moat was created. The moat turned sharply towards the NE after 48.50m (measured from the western baulk in T7) echoing exactly the turn of the headrace. Up until this point the millrace and the moat ran parallel to each other, but whereas the moat turned sharply, the bend of the millrace had a gentler angle and the alignments of the two features diverged. The Feeder Ditch F74 The Glenamuck River flowed to the north and west of the site. West of the site, it was originally closer: part of the old course was found in Trenches W10–W11. It was 114m from the site of the mill itself. The point where the water was drawn off the Glenamuck river was in an unexcavated area, south of W12. The north-south ditch F4353 in Trench T13 may represent part of it, but the first continuous part was a shallow channel in T12, called F4415. Two features were associated with F4415. On its base was a small platform of stones which seem to have been arranged with flat sides facing up in a roughly rectangular pattern. An irregular and quite patchy deposit of ash with charcoal and mortar inclusions (C4715) covered a large area of the platform. A possible drain, F4420, ran into F4415 from the north. The eastern end of F4415 was defined by an outcrop of bedrock which would have formed a natural weir. From here, the water flowed into the headrace proper, F74. At this point there were three features associated with F74. A small ditch or large drain, F4413, ran at an acute angle to F74 divided from it by a ridge of natural and bedrock before merging with it. It was difficult to say which was earlier. Another possible drain, F4417, was cut into its southern edge. It appears to have been contemporary with F74 as the fills of the two were continuous. A stone drain, F4392, was cut into the fill of F74. It had a base composed of flat stones, and was walled with side stones, but there was no evidence that it was covered. F74 measured 50m in length from the westernmost part exposed in Trench T12 to where it reached F1436 in Trench 5. Because of the later re-cut features F223 and F229, the full extent of F74 in T5 could only be estimated to where this met the funnel F1435. The average width of F74 was 4.00m, and the average depth was about 0.70m, reaching a maximum depth of 0.80m. It was uniform in section and maintained a U-shaped profile along its entire length. The feature ran parallel to boundary ditch F635 and its subsequent re-cut F649, mentioned above in connection with Structure A, suggesting either that traces of this ditch may still have been in existence during Phase 2, or the boundary which it represented may still have been active. The water sluice funnel F1436 The two successive water sluice funnels which were incorporated into the headrace were, as mentioned above, F1436 and re-cut F1435. The re-cut truncated the original almost completely and both of them were, in turn, heavily truncated by F281, a post-medieval stone platform. What remained of F1436 was a small, semicircular arc-shaped cut with fill C3319 which was truncated on the Northern side by millrace F94 and on the western side by the sluice re-cut F1435. This primary earth-cut funnel was for the most part cut into the natural soil (C3). F1436, as far as we can ascertain from the scant evidence, was semi-circular plan but, owing to truncation, only a small portion remained visible to the south-east of its successor F1435. Because it has been truncated, it is difficult to determine what, if any, components it contained, but F1435 definitely incorporated a semicircular arrangement of stones. Four of these survived, but the distribution and layout suggested that there may originally have been more. If F1436 was simply a narrower section of the feeder ditch, without a barrier, it would have allowed water to funnel through it with a more powerful flow. A similar narrowing would also have been what differentiated the feeder ditch F74 from the mill-race proper, F783. The re-cut, F1435 was cut slightly to the south of F783, but they still mostly overlay each other. It is likely that F1436 originally extended the full width of F74, and could have measured up to 4.00m in length. The headrace F783 F783 represented the part of the original headrace that existed east of F1436. This feature measured approximately 27m in length and was re-cut by F94, which followed the exact same course of the original cut. This length is an approximate value, measured from the eastern edge of F229 and F223 up to the point where F631 truncated the feature in the east. This channel was shallow and wide and this may have led to the need for a re-cut water funnel (F1435). This feature was shallower in the west (0.25m–0.50m) than its later re-cut F94, but deeper (0.62m) towards the east. It is not known how wide it may have stretched in the east in Trench 5 as it was truncated along its entire southern length by F94 and in the extreme west by F281 but the maximum remaining width was measured to 1.40m. F783 had a regular U-shaped cut with a non-perceptible break of slope at the base, changing from gradual to sharp at the top. The sides were exclusively concave. A narrow strip of the base in parts of the centre of the feature was flat but otherwise it was rounded. It is unclear whether F783 and F74 once formed a single feature as the re-cut features F229, F223 and F94 had destroyed any remains of their relationship. It is likely however, that the original headrace was a simple undifferentiated ditch construction which became more complex through later remodelling such as the sluice funnel (F1435) and the possible weir feature (C3312). The original mill-race must have been found inefficient, so that extensive re-modelling was necessary to produce a more controlled supply of water with sufficient volume and velocity. F783 had been truncated by the tertiary fosse (F631). There is no indication that there was a mill-pond to collect and hold the water between the headrace (F783) and the wheel-pit (F816). The remains of the original millrace F783 had already been severely truncated by F94 and some of the fills were disturbed by this. The break of slope at the top was sharp along both edges and at the base gradual to imperceptible. The southern side was slightly concave and the north side was slanted at an estimated 40º inclination. The base was flat, falling SW-NE along the course of the feature and the natural slope of the ground, and cut through the natural soil, which contained inclusions of natural stone that protruded through the surface. Where the feature truncated the kiln bowl (F868), parts of the remains of the upstanding masonry (F1428/C2077) formed the southern side of the cut, making it vertical in cross-section. As mentioned above, it was likely that this feature was the same as F74, however the later re-cuts would have destroyed all evidence of this. F783 contained altogether seven fills of silty/clay composition, which would have been deposited during the time when the feature was in use. Most of these were smallish deposits that appeared to have been disturbed by the later re-cut. Along the southern edge of the cut, several fragments of collapsed masonry (C1641) from the wall of the kiln bowl (F1428/C2077) had been deposited, most likely when the millrace was constructed. Three deposits of silts and clay, C1697, C1696 and C3339, were deposited along the base of the feature. C3339 occurred between the stones in the bowl wall and was initially thought to be the same as C1524 in T5, but it subsequently became clear that the two were separated by a substantial layer of silty sand and gravel (C1511) along the northern edge of the cut. This context contained large amounts of flecks of oxidised clay and charcoal and has been interpreted as collapse of material from the edge of the ditch where it truncated the kiln bowl during the life of the original mill-race. A small, circular deposit of yellow sand (C1736) was deposited at the base of the feature (on top of C1696). It is possibly the same as context C1511. A large dressed granite stone (0.62m x 0.53m was set into the northern edge of C1511 along the stretch of the mill-race. This architectural stone was re-deposited from elsewhere. It resembles a large baffle stone which would have been used in a kiln or similar feature. The nearby kiln, however, did not include any stones of the right dimensions for a baffle stone of this size. Along each side of the feature, below the bottom fill of F94 and above C1511, two almost identical contexts of mid brown sandy silt/sandy silty clay were excavated (C1867, C2008). They probably represented a single context divided when the millrace was re-cut by F94. Finds from the original millrace were few. Only two of the bottom contexts contained finds. One sherd of Leinster Cooking Ware was recorded in C1697, and in C3339 a pivot stone, possibly related to the mill, was recovered. The mixed context C1511 contained a number of medieval sherds from both T5 and T10, but the other fills were empty of finds. Unlike F94, F783 contained finds only of medieval date but, because of the high level of disturbance, it is likely because that some of the finds were intrusive. For example, the sherd of Leinster Cooking Ware discovered in the bottom fill C1697 could, because the context was mostly thinner that 150mm in depth, have sunk down into the base of the feature. The re-modelled sluice feature F1435 The re-cut of the original water sluice funnel (F1436) had a much more solid construction made up of boulders which formed a semi-circular sluice (Figures 21–22). This construction allowed the water to pass through a gap between the two central boulders. The boulders were firmly bonded allowing no water to pass between them. F1435 did not survive intact due to truncation by post-medieval features. It is possible that there was a sluice gate in the small gap between the two middle stones, allowing water to build up west of it in F74 and allowing variations in the rate of flow. Because the water was allowed to settle, there had been extensive silting in the area west of the sluice. It is not known how many stones may have originally formed F1435, or what shape the overall cut made in plan. There were no clear indications of the boulders of the sluice ever having formed a circular (or other) shaped feature or even that the cut for these curved around to form a dam on the west side of the sluice. The build-up of silt, however, suggests that this feature caused water to accumulate in F74, whether or not there was an actual dam. What remained of F1435 measured 2.50m in length (N-S), 1.10m in width (E-N) 0.25m in depth (the depth of C3318). The latter fill (a very compact dark brown-grey silty sand with small stone inclusions and ample medieval pottery finds) had been truncated, but appeared to have originally been level with the tops of the boulders. As in the case of the original sluice feature (F1436), the length of F1435 must have been equal to the width of F74 (4.00m) in order for the whole construction to work. The cut of F1435 had a flat base where the boulders of C3317 lay flush, and the cut itself was steep-sided as seen in the section of the excavated southern edge of F94. This flat base could also be seen immediately to the north of the boulders C3317, which indicated the continuation of F1435 northwards. Any additional boulders would have been removed from this section by truncation. The remaining stones all had vertical facings and formed a regular arc in plan, with a N-S alignment and assuming that the area to the north took on a similar shape in construction one would be looking at a semi-circular structure in plan. The northernmost boulder measured 0.40m x 0.40m x 0.56m, the middle boulder measured 0.70m x 0.60m x 0.52m and the most southernmost boulder measured 0.90m x 0.60m x 0.52m. Context C3318 was deposited in the space between the two most southerly boulders and the cut of F1435. This context also filled the gap between the southernmost boulder and the middle boulder of F3317. C3318 is contemporary with C3317 and would have acted as a compact sealant barrier, between the boulders, if they were placed as close as possible together to accomplish this task effectively. Therefore it is likely that they were originally as high as the tops of the boulders. F1435 and F94 are contemporary features. The most northern boulder lies on the base of F94, whereas the other two boulders lie on the flat base of F1435, and all three boulders are undoubtedly part of the sluice feature, conclusively linking F1435 with F94. F1435 was located at the most westerly point of F94, where F94 abruptly terminated there was a difference in height between the base of F74 and the base of F94 of the order of 0.37m which would have allowed a better flow of water through the gap. The opening in the sluice measured 0.12m in width and 0.64m in height and was situated between the northernmost boulder and the middle boulder. This gap between the stones remained open during the lifetime of the mill-races but there could have been a temporary closure mechanism like a sluice gate. This would control the volume of water entering F94. This would have improved directional accuracy, leading flowing water straight into the middle of F94. The water would have dropped 0.37m from the base of F74 down to the base of F94, where F1435 meets F94 in the opening between the stones. The re-cut of the headrace F94 This re-cut of the headrace followed the exact course of the original headrace (F783) but was cut wider and deeper in its section in T5 and narrower and shallower in T10. The western part of the original race (inside T5) was nearly completely truncated by the re-cut (F94) and the only part of it that was still intact was the northern most edge. At this end the re-cut could have been wider than the original millrace because the southern side and the base were no longer evident. Although it became narrower along its course, this re-cut of the channel continued east towards Trench 10 but approximately 1.40 west of the section (at the point where the remains of the bowl wall of the kiln was discovered along the south edge), F94 had narrowed down to fit within the perimeter of F783 and in this section, the feature measured 1.40m width (in comparison to the 2.25m width of F783). The total length of the feature was equal to that of F783 and measured 27m. The feature was 0.66m in width in the west and 1.40m in the east and the depth ranged from 0.74 in the west to 0.48m in the east. F94 had a regular U-shaped profile throughout its length, and although some undulations in the bedrock did occur, they would not have obstructed the waterway. The break of slope at the top appeared sharp along the northern edge and gradual along the southern. The north side had a gentle slope of approximately 45º (changing to slightly concave towards the south) with a sharp to imperceptible break of slope at the base. The south side was steeper with only a slightly concave profile, which changed to a gradual break of slope at the base. The base was generally flat although in places slightly uneven. F94 contained four silt/clay fills (C1512, C1714, C1793 and C1523) and one backfill (C1509), which all but two of which contained medieval finds only. The bottom silty clay (C1523) contained two sherds of post-medieval pottery and three iron fragments. The last phase of the silting-up process, context C1512, contained one sherd of post-medieval Blackware and an iron fragment, possibly from a key. These finds could not be anything but intrusive and the result of post- medieval field integration as the millrace, being contemporary with the mill, should not be later than 13th century in date. C1523 and C1524 appeared throughout the length of F94, from the weir feature down to the mill. These silts undoubtedly built up at times when water was allowed to settle in F94, but for the most part water apparently flowed efficiently through this feature. The possible weir, C3312 was located within F94 and consisted of a rough stone wall which spanned the width of F94, situated approximately 1.50m east of F1435, allowing a pool to form between these two features. This would have been particularly effective if there was a sluice gate on F1435. The horizontal mill The wheel-pit and tailrace were given a single feature number, F816, because no clear distinction between the pit and the tailrace could be established. Essentially, the wheel-pit narrowed down from a large, deep sub-oval cut into a linear feature, which decreased in depth and width as it continued northeast, leading the water away from the mill. This transformation from wheel-pit to tailrace was gradual but seemed to start in the corner of T10 and TK and as the feature entered the south end of TN, it took on the form of a linear feature. The entire stretch of the tailrace in TN measured 14.20m. This included the section which had been obliterated by the revetted fosse and was measured from the southern baulk in TN and to the point where F498 cuts F816. The feature became both narrower and shallower the further north it extended. The tailrace disappeared in the east part of Trench N as it was truncated by the medieval moat F498. The revetted fosse also truncated the tailrace after approximately 6.00m into TN and had cut away close to 4m of its length. This established that the mill pre-dated the revetted fosse enclosure and perhaps also the Northern enclosure, which F498 represents. On the other hand, the courses of F265/F75 and the headrace F94 (/F74), were both curvilinear and parallel to each other to some extent, which suggests that the tailrace was still in existence at the time of the creation of the South-western Enclosure, of which F75/F265 formed the major part. The mill does not appear to have been in use at this time, but the mill-race may have been seen as an additional obstacle to attackers. The house structure in Trench N may be contemporary with this phase of activity in the South field. The wheel-pit F816 The large sub-oval shaped cut for the wheel pit, containing what remained of the mill structure, was approximately 10.50m long, but as there was no clear distinction between the end of the pit and the beginning of the tail-race, this is not exact (Figures 19–20). Its orientation was SW-NE. The width at the SW end of the feature was c. 4.50mm and the widest part, at the middle, 5.40m. The narrowest point of the feature was at the north-eastern end, where it measured 3.60m. The wheel pit was positioned in a field which had a marked natural slope and although the base of the feature respected this slope slightly, it levelled out to a much more gentle slope. This resulted in the depth being much greater at the south-western end at 1.547m and steadily decreasing along the feature towards the northeast to a maximum depth of 0.969m. The wheel pit was cut into a yellow sandy natural soil (C3) in the north and south and protruding bedrock in the SW. The break of slope at the top was sharp all around the edge of the feature except at the rounded SW terminal. The south-western end was cut into the bedrock: this created a drop into the pit. The break of slope at the base was gradual all round the feature. All the sides were concave except where the cut bedrock created a stepped edge. The base was flat with an inclination following the natural slope of the ground towards the NE. The cut bedrock could have supported structures such as a sluice and possibly a trough and a dam feature to the west of the wheel pit. However, the tertiary fosse (F631) cut across the westernmost perimeter of the wheel pit, destroying any evidence which may have been present. There were no traces of a mill building evident around the edges of the wheel pit, although it is possible that some of the stone rubble (C2038) could have formed part of such a structure. It is likely that when the mill went out of use, any above-ground structures were dismantled and the materials re-used elsewhere. The only remains of the structure were those buried beneath the silt layers which had accumulated on the base of the pit. The wheel pit had been truncated by both the secondary and tertiary fosses (F637 and F638). After the mill had gone out of use, the wheel-pit was partly backfilled with large stones. These were mostly large natural boulders, but a number of them were worked. The worked stones included four millstones (Plate XI), a quernstone and a staddle stone, as well as two further possible examples of the latter two. In addition to some pieces of dressed building stone, a possible pivot stone and an elongated voussoir from an arch or doorway were found. All these stones were probably dumped into the wheel-pit at the same time, but they sank down to different levels in the soft fills below. The wooden structure (C3232/F1397) consisted of a platform made up of 12 large pieces of wood (Plate XII). Wooden planking, beams and other worked wood pieces were recovered from the feature. Neither trough nor wheel were found: these had probably been removed before or after the mill was destroyed. In addition to this several smaller pieces of wood were found, both worked and un-worked. The largest timbers were three long oak beams (00E0525:3232:1–3). Underneath these timbers a flat surface was situated, consisting of 6 individual planks, most of which were ash (see sections 9.10 and 9.11 below). Also, a number of miscellaneous pieces of worked wood were recovered from in or around the mill structure. The wooden structure was covered in a silty, organic layer (C2173) which was had accumulated at least partly after the destruction of the mill structure. From this layer was retrieved several pivot stones, pieces of quern-stones, medieval pottery and the stick-pin. This silt was also found under the planks in C3232/F1397 but because of the high water content it would have flowed between and under the planks. The dating evidence The three long beams were dated by dendrochronology (see section 9.20 below). Two of them produced usable dates, having been felled in or after the late 11th and early 12th centuries respectively. These dates seem too early, in view of the finds of 13th-century pottery in Structure A and the kiln, which are stratigraphically earlier than the mill race. They are, however, termini post quem, so the actual felling date may be much later. A stick-pin found in the bottom of the wheel-pit, and Dublin Type Coarse Ware pottery in the lower levels, are consistent with a 12th-century date, but must be redeposited. The kiln, in particular, rules out a pre-13th century date for the mill, as it underlies the headrace. The mill-race appears to have been still open when the South-western Enclosure, F265/F75, was being constructed. It cannot have been in use, as the tail-race had already been truncated by the moat, F498, of the North-western enclosure. The fills of the wheel pit Six of the top fills of the wheel pit F816 were excavated and recorded as backfill contexts. C1348, C1336, C1335, C1892 were smaller contexts of mottled coloured soils, mainly sandy silts and silty clays, which contained occasional inclusions of stones and pebbles and occasional inclusions of charcoal. Finds from these context included only occasional sherds of Leinster Cooking Ware (C1348 and C1335) and one worked piece of flint (C1348). The most extensive backfill layer, C1696, contained mixed grey/brown/orange sandy silt and contained occasional inclusions of clay, stones, mica fragments and charcoal. This context was truncated by both the secondary and tertiary fosses. The finds from this context were all of medieval date and included occasional sherds of Leinster Cooking Ware, one piece of ferrous metal, animal bone and one flint flake. The fill below C1696 (C2383) probably originated during the creation of the bank F793 as it was slippage of soil from the bank into the wheel pit. This context was deposited along the southern edge of the feature and along the northern edge a similar deposit (C2252) of collapsed re-deposited material was excavated. C2252 appeared to be an earlier deposit than C2383, which covered the natural soil (C3) to the north and had been truncated by the medieval moat F265/F75, and has been interpreted as collapse of spoil which had accumulated on the edge of the feature when it was being created. No traces of a bank on top of this was seen and it seems likely that F793 only reached as far north as the edge of this feature. No finds were recovered from either C2383 or C2252. The last layer of silt which accumulated within this feature (C2408) was deposited a considerable time after the mill structure had fallen into disuse. This context included brown/grey silty clay, which contained occasional flecks of charcoal and reddish clay as well as 17 sherds of medieval pottery, mainly Leinster Cooking Ware but also 4 sherds of Dublin Type Ware, 1 sherd of Dublin Type Fine Ware and 1 sherd of Dublin Type Coarse Ware. Below this, a third layer of collapse bank material (C2256) had been deposited along the southern edge of the feature. This context proved to be the same as one of the main sterile, core/bottom contexts of the bank material (C1895), which due to its coarse nature would have subsided into the shallow depression that remained in evidence from the wheel pit at the time of the construction of the bank. A further two contexts, identified as post-dating the mill, were excavated underneath C2256 and these represent the first silting up of the wheel pit subsequent to its being abandoned. C2410 comprised brown/grey silty sand with large inclusions of clay and occasional inclusions of flecks of charcoal and red clay. Two sherds of Leinster Cooking Ware and one piece of worked flint were retrieved from this context. C3077 was a thin layer of sand and gravel under C2410, resting on top of and around the mill timbers. At the base of the pit the remains of the mill timbers (C3232/F1397) were situated. The silts would have only partially covered the timber and, because of this, some of the protruding wood had dried out in antiquity (this was especially the case with timber 00E0525:3232:1). Most of the stone artefacts found in the wheel pit were associated with milling or related activities, except for the voussoir fragment. This fragment is a piece of architectural stone, and given its fine finish and the fact that it was probably not local in origin, it is likely to have come from a building more prestigious than the mill. It is also much later, dating from the 16th or 17th century. Tailrace of mill F816 The wheel pit narrowed down from a deep sub-oval pit into a linear feature, which decreased in depth and width as it continued northeast. It therefore merged gradually into the tailrace. The tailrace was broken up by two later constructions. The revetted fosse truncated the tailrace approximately 6m into TN (12m from the estimated start of the feature in T10), cutting away close to 4m of its length. The furthest surviving part of the tailrace was in the east part of Trench N where it was truncated by the 13th Century moat F498. The mill therefore pre-dated the revetted fosse enclosure and the North-western Enclosure structure, as F498 represents the latter structure while the medieval moat F265/F75 defines the South-western Enclosure. Examining the courses of both F265/F75 and the headrace F94/F783/F74, they were both approximately curvilinear and ran parallel to each other and this could indicate that they, at least at one point, were contemporaneous features. The entire stretch of the tailrace in TN measured 14.20m (20.5m from the start in Trench 10 to the point where F498 truncated it in the north-east), including the section removed by the revetted fosse and measured from the southern baulk in TN and to the point where F498 cuts F816. The feature became both narrower and shallower the further north it extended. Close to the revetted fosse the feature displayed a complex stratigraphy. 15 fills were recorded in this section alone. The sections through the wheel pit in T10 and the tailrace in TN were compared, but because of the distance between them and possibly because of the difference in depth along the feature, the contexts could not be matched up. The stratigraphy became less complex east of the revetted fosse and it was noted that some of the contexts continued from the section on the west side. Six fills, all of which contained finds of medieval date, were recorded in section at the corner of Trench K. Again, these could not be matched up with the contexts in Trench 10. East of the revetted fosse the cut of the tailrace was recorded as linear with sharp break of slope at the top, gradual at the base with slanted sides and with a slightly convex to flat base. West of the revetted fosse, the break of slope at the top was gradual and at base gradual to imperceptible. The sides changed from very slightly convex along the north side to concave along the south side and the base was flat. The difference in width and depth between these two sections of the tailrace was substantial. The depth recorded in the west was 0.90m, whereas that in the east measured 0.40m. The width only differed by 0.10m between the western section (3.70m) and the eastern section (3.60m) although the difference between the widest part of the race west of the fosse to the narrowest part of the feature east of the fosse was approximately 2m (4.70m-2.70m). The end of the mill One or two of the timbers showed signs of being charred on the surface, but the charring seemed superficial and none of the timbers were burnt all the way through to the core. Whether this is evidence for a destructive fire is difficult to assess. No burnt soil was found in or around the wheel pit and because nothing of the superstructure survived, it was impossible to be certain. Alternatively, the mill could have gone out of use because of poor quality construction. It has been shown above that the headrace had needed to be re-cut several times. The timber in the surviving part of the structure was of poor quality, with young trees of mixed species (possibly from the local; environment) such as oak, ash and alder'. Bolton, J. 2015 Draft Carrickmines Castle Conservation Plan 2015-2025. DLR Co. Co., County Hall, Marine Road, Dún Laoghaire, County Dublin. Breen, T. 2012 Report on the Findings of the Archaeological Excavations at Carrickmines, Co. Dublin - Carried out under licences 00E0525, 02E1532 and Ministerial Directions. M50 South-eastern Motorway Scheme 4 volumes. Prepared on behalf of V.J. Keely Ltd. for Dun Laoghaire-Rathdown County Council. See attached map and plan showing location of the 13th century watermill associated with Carrickmines Castle, courtesy of Jason Bolton 2015, 32, 33. See attached site plan showing location of wheel pit, headrace and tailrace of watermill at Carrickmines Castle, taken from excavation report under licence No. 00E0525, 02E1532, Breen 2012, Vol. 4. Compiled by: Geraldine Stout and Padraig Clancy Revised by: Caimin O'Brien Date of revised upload: 30 December 2022