1999:240 - PROPOSED DUNDRUM TOWN CENTRE, Dundrum, Dublin

Testing was carried out before the development of the proposed Dundrum Town Centre. Three trenches were opened underneath the overflow dam of a millpond (Area A), at a point along a mill-race (Area B) and at a point along the River Slang (Area C), in order to test for the possible survival of archaeological deposits relating to milling in the area. One trench was mechanically opened with a concrete-breaker and a 1.5m back bucket, and a second with a 600mm bucket on a mini-digger; a third trench was excavated by hand. All three trenches were reduced to natural subsoil and/or bedrock.

The site consists of 22 acres contained by the Ballinteer Road to the north, the Sandyford Road to the east, the proposed Wyckham Bypass to the south and the proposed Dundrum Main Street Bypass to the west. The River Slang forms the north-western boundary of the site for approximately one-third of the site's length, before transecting the site and dividing it in half. The central section of the river flows through a deep gully before turning sharply north and entering a wider valley. This section of the site is overlooked by Dundrum Castle on the ridge to the west and has itself an SMR designation, owing to the possibility of the archaeological survival of medieval milling activity along the riverbank.

The area to the south of the river was once occupied by the grounds of Rockmount House, an early 19th-century residence that was demolished in the 1980s. Its walled garden and a portion of its farmyard wall still survive. The southernmost area is now occupied by the grounds of Crazy Prices, while the remainder is waste ground. The north-western sector of the site is occupied by the grounds of the mill-house (which has a List 2 designation on the Dun Laoghaire-Rathdown Development Plan). This consists of the mill-house itself; an 18th-century residence with Victorian alterations, its garden and orchard; a mill-race leading from the River Slang to a large millpond to the rear of mill-house; and an area of upcast to the south of the pond, which is grassed over. The water in the pond was held by an overflow dam consisting of a large granite wall. The level of water in the mill-race and pond was controlled at this point and additionally by a sluice system along the mill-race, which appears to have been blocked relatively recently, draining the pond in the process.

Area A is that area along the eastern bank of the River Slang occupied by the Pye buildings and the access yard. Trench 1 was opened in the yard to the east of the Pye buildings, at a point directly below the overflow dam of the millpond, adjacent to a cut feature that may have originally housed the wheel pit. The earliest cartographic evidence for this specific location dates from 1760, in John Rocque's 'Map of the County of Dublin', which would appear to depict a building in the approximate position of the mill-house, linked by a drafter's line to a building on the riverbank. This building's position in relation to Dundrum Castle on the ridge on the opposite side of the river would place it quite close to that building in the Pye complex immediately to the west of Trench 1.

A trench was mechanically excavated, 14.6m long and 1.5m wide, c. 10m west of the overflow dam and parallel to it. This area is directly to the west of a small building on a raised platform that was demolished recently. It was opened at this point in order to assess the extent and antiquity of the tail-race and to see whether any earlier evidence of milling had survived.

At either end of the trench, decayed granite bedrock lay just underneath the concrete surface. Firmer, undisturbed bedrock lay 0.3m underneath the decayed material. At a point 1.6m north of the southern end, the bedrock was cut to accommodate two 16-inch pipes in a trench extending to the south-west. The cut was filled with loose rubble. The pipes would appear to connect with the channel running from the area that possibly housed the wheel pit and are typical of drainage pipes used from the early part of the 20th century onwards. The pipes rested on decayed granite at c. 1.2m below the surface of the concrete.

From c. 3.2m north of the southern end of the trench, a surface of cinder bricks extended from the edge of the pipe cut. This surface extended from a straight edge north of the pipe cut, and it is likely that the pipes were accessible from it. It extended 6.2m to the north, terminating at a straight edge. The bricks were wire-cut and were laid on a 0.04m bed of fine lime mortar that in turn overlay a layer of black silt. The silt layer was between 0.07m and 0.15m thick, overlying undisturbed bedrock and truncated to the south by the pipe-trench. It extended for c. 6.2m to the north, where it was truncated by the cut for a brick culvert. There were no inclusions in the silt.

A brick culvert of 2ft 8in diameter occupied the remaining 5.4m of the trench. The culvert ran towards the north-west, and its apex was just below the surface of the concrete. It was constructed from a trench cutting through the bedrock, the brickwork rising from granite sides, and filled with brown silt and cinders. The base of the culvert was 1.55m below the surface of the concrete. It appeared to run from a cast-iron pipe extending out from the overflow dam and possibly connects with the tail-race as depicted on the 2nd edition of the Ordnance Survey.

The trench appears to have exposed no features earlier than the mid-19th century, although the respective functions of the culvert and the twin pipes are difficult to extrapolate without more evidence relating to the workings of the ironworks, the laundry and the millpond, which are all evident on various editions of the Ordnance Survey. While the ironworks presumably used the reserves of water held by the overflow dam both as a coolant and to drive the machinery, it is unlikely that the laundry used the silty water of the millpond for washing clothes, although it may have been used to drive primitive tumblers. It would seem probable therefore that the brick culvert relates to functions carried out by the ironworks and that the double pipes may have utilised a previously cut channel through the bedrock to supply the laundry. On the basis of the above evidence, it is likely that if any archaeological evidence of milling in the area survives, it will be found along the riverbank underneath the Pye buildings.

Area B is the area along either side of the mill-race, which flows to the north from its junction with the River Slang, before curving to the west and forming the millpond to the rear of mill-house. Although the water supply has been cut off, the mill-race has not completely drained off into the pond and remains full but stagnant for most of its length. The area west of the race and south of the pond is mostly clear of trees. Its contours indicate that it is composed of the upcast from the millpond, and the plot is referred to as 'waste from millpond' in the records of the Valuations Office. A steep slope to the south of this area runs down to the river, while a wooded slope runs down on the western side to a concrete retaining wall in the Pye precinct. Trench 2 was opened across the mill-race to test for evidence that this feature pre-dates the first cartographic evidence.

The mill-race and pond first appear on Duncan's map of 1821, albeit in a different form to that depicted in subsequent editions of the Ordnance Survey. Although the race as represented here comes off the River Slang at approximately the correct position, as has been stated above, it continues to flow north of a cigar-shaped pond, continuing under the Ballinteer Road before finally rejoining the Slang at a point north of the Catholic church. Despite the eventual destination of Duncan's mill-race, an examination of the frequent meanders of the surrounding watercourses on his map would suggest that he took a fanciful notion of hydrography, which cannot be relied upon for exactness. There is, however, enough there to suggest that an arrangement of mill-race and pond was in existence by 1821.

Trench 2 was opened at either side of the mill-race 2m to the south of the sluice at the head of the millpond. It was excavated to a width of 2m with a 600mm mini-digger bucket in three sections: Trench 2a to the east of the mill-race; Trench 2b through the channel itself; and Trench 2c on the west side of the mill-race. The individual trenches were co-linear and extended for 12.2m.

Trench 2a extended for 4.4m eastwards of the mill-race. The stratigraphy encountered was a slight bank of upcast extending for 2.2m, overlying an old soil horizon and 1.1m of mid-brown, sandy loam. This overlay a very hard, compact, grey silt with frequent granite stones through the matrix, first evident in the western end of the trench at c. 1.6m below the surface. This layer sloped upwards gradually to the east, levelling out at the end of the trench at c. 0.6m below the topsoil. The silting became more ephemeral before disappearing at c. 2m east of the baulk. This overlay compact boulder clay. Undisturbed bedrock was not encountered in a 3.5m test-pit excavated at the eastern end of the trench.

Trench 2b attempted to determine the nature of the bottom of the mill-race by dredging across the channel and holding the water back with the spoil. The mill-race at this point is 3.75m wide. The banks of the mill-race are lined for most of its length with squared-off granite blocks, which have been inserted into the upcast. The bottom of the mill-race does not seem to be so lined, although several granite blocks were dredged up by the bucket. The mill-race appears to have a depth of 1.8m in the centre (where the silting gave way to more solid material), the sides rising steeply forming a wide V-shape. The silt and mud taken from the channel contained 20th-century refuse.

Trench 2c extended for 4m westwards of the mill-race. A 0.3m baulk was maintained at the eastern end, where the stones lining the mill-race were particularly well set, forming a higher bank than that on the opposite side. Solid bedrock was encountered at 2.1m west of the channel, 0.4m below the sod surface, which extended back the length of the trench.

The area between the bedrock and the baulk was excavated by hand to a depth of 1.2m below the sod surface. It was found that the bedrock had been deliberately cut at an almost vertical slope, presumably to create a channel for the mill-race. An examination of the sections revealed that a deposit of soft, grey silt overlay the cut in the bedrock but was itself sealed by the topsoil and the upcast bank running alongside the mill-race.

The silting at either side of the mill-race, which is sealed by the upcast on both sides, would indicate that the present channel is a secondary one. The distance between the cut in the bedrock and the equivalent level of the silting in Trench 2a is c. 7m, which may indicate that either Duncan's map was accurate in depicting a stream at this point or an earlier mill-race had at some point burst its banks, which were later reinforced with the granite blocks.

Area C is that area at either side of the River Slang between its junction with the mill-race and the sharp turn to the north, 115m downstream. The area is heavily overgrown with trees and bushes, and access is consequently difficult. A steep slope of over 3m, which is vertical in places, delineates the southern extent of this area, forming a gorge. The slope across the river to the north is less marked, although the ground rises rather sharply in the area of the millpond upcast. The river, flowing from east to west, drops several metres in its passage through the area, falling over four small weirs. They are not shown on the 1907 Ordnance Survey revision and would appear to be relatively modern. Concrete work is visible in places, although it is unclear whether this is primary or repair.

The banks of the river have been deliberately constructed with large granite boulders that survive well on the southern side. The retaining walls appear to be of drystone construction, but test excavation has indicated the presence of a lime-mortar bonding agent. A formalised pathway also seems to have existed along the southern bank of the stream and is possibly that shown on the 2nd edition of the Ordnance Survey (1869). This also depicts a relative lack of vegetation in the area, suggesting that the trees here today were planted in the 1860s or afterwards to create an afforested riverside walk, presumably for the enjoyment of the occupants of Rock Mount House.

A trench was opened by hand on the southern riverbank, primarily in order to test for earlier milling activity in the area but also to investigate the nature of the riverbank itself. It was 1m wide and 6m long, extending north-south from the riverbank to a near-vertical rock face.

The wall along the southern bank of the river stood to a height of 1.2m from the silt at the bottom, and its upper course protruded from the topsoil. It was constructed of large granite blocks and was roughly coursed.

On removal of a thin layer of topsoil, a layer of rough cobbling was exposed in the centre of the trench. This was composed of roughly laid sub-angular granite stones, with an average diameter of less than 0.2m. A spread of lime mortar ran along the length of the wall and extended outwards to the south for 0.1m. The cobbling lay on a layer of mid-light brown, sandy loam, which contained more granite chippings at the southern end of the trench. Here, at its greatest depth, the layer was 0.2m deep. At the northern end the layer was disturbed by the riverbank wall. The granite bedrock underlies the loam and chippings at c. 0.3m below the upper level of the topsoil.

The pathway and retaining riverbank wall would both appear to be Victorian garden features, and no evidence of earlier features was recorded.

2 Killiney View, Albert Road Lower, Glenageary, Co. Dublin