2001:1391 - ROUGH ISLAND, Down

Rough Island, Ringcreevy, is located on the north shore of Strangford Lough in County Down, midway between Newtownards and Comber. The island, formed from a small drumlin, is joined to the mainland during low tide by a narrow artificial causeway. It was first identified as being of archaeological significance in 1936, when Professor Hallam Movius undertook an excavation near the summit on the island’s northern side and uncovered a late Mesolithic shell midden along with an associated hearth (Movius 1940). A second programme of excavation on the island was carried out by Tom McErlean in 1997 as part of the Environment and Heritage Service’s intertidal study of the lough (Excavations 1997, 76). The island’s south and south-west shoreline is under threat from coastal erosion and McErlean investigated a layer of eroding dark sand, rich in medieval pottery.

Although limited in scale and duration, McErlean’s work had suggested that some form of medieval occupation may have occurred at the island’s southern end. In light of this, the Environment and Heritage Service, DOE NI, wished to determine the risk that coastal erosion might pose to any archaeological resource in this area. The School of Archaeology and Palaeoecology, Queen’s University Belfast, were therefore approached to undertake a third programme of archaeological investigation on the island, during a six-week period between 6 August and 14 September 2001. This was also used as an opportunity to reinvestigate the area of prehistoric activity identified by Movius in 1936. The procurement of samples for radiocarbon dating from the shell midden and any associated hearths, for example, could enable a more precise chronology for the prehistoric occupation of the island to be constructed, while a reinvestigation of the shell midden would allow a better assessment of its food potential for the inhabitants.

Nine trenches were laid out in Area 1 to investigate whether the area in the vicinity of McErlean’s 1997 trench was the location of a medieval settlement. A similar stratigraphy was encountered in all nine trenches, with no evidence of anthropogenic activity other than in the ploughzone, where some medieval pottery sherds and a mix of modern and prehistoric artefacts were encountered. Trench 42, which lay adjacent to McErlean’s trench, was later extended westward by 0.7m to connect with the 1997 cutting. Excavation to c. 0.87m revealed the layer of dark sand containing medieval pottery but it was found to extend only c. 0.7m beyond the area investigated by McErlean’s trench. While three more clusters of medieval pottery sherds were noted within this matrix, no further evidence was found to suggest that this layer represented the remains of a medieval occupation and no evidence of medieval settlement was discovered in Area 1.

Twelve small test-trenches (Trenches 1 to 12) were opened in Area 2, in the area between Movius’s main trenches, to investigate the prehistoric shell midden. Excavation revealed that the midden measured at least some 8.3m by 8.1m and was 0.21m in depth, lying immediately below a dark brown modern humus layer. The midden was quite unconsolidated and loose in nature, and there were no layers visible within its matrix. Its slight depth would tend to suggest that the material accumulated over a short yet continuous period of time. Sherds of Western Neolithic pottery were recovered from both within the midden and in the charcoal-stained yellow sandy soil that lay beneath. A pit containing a combination of large stones and shells was discovered below the shell midden. It also yielded sherds of Western Neolithic pottery. While samples have been submitted for radiocarbon dating, the presence and location of this pottery would suggest a possible date of c. 4000 BC for the deposition of the midden.

A number of stake-holes were also discovered under and beside the midden. Unfortunately, it was not possible to identify whether these had been cut before or after the shell midden had been deposited. Two rows of post-holes—running parallel to each other for a distance of c. 4.5m—were also identified, along with a semicircle of stake-holes, c. 2.3m in diameter, interrupted by the edge of the trench. These do not seem to reflect the remains of substantial architecture, but they could represent evidence for tent structures, windbreaks or fish-drying stakes.

A further two trenches (Trench 20 and Trench 21) within Area 2 were opened in order to investigate whether evidence of prehistoric activity survived on the island’s summit. With the exception of the discovery of some early Mesolithic artefacts in disturbed contexts, nothing of archaeological value was discovered in either trench.

In advance of excavation, a geophysical survey had been undertaken on the western side of the island by Geoquest Associates. A magnetometer survey identified an anomalous area, possibly caused by burning. Trench 40 was opened, therefore, to elucidate the nature of this reading. Nothing of archaeological significance was encountered, however, and it is likely that the reading was caused by modern ferrous debris in the area.

A final trench (Trench 30) was opened on the north-west side of the island to investigate a layer of shells, some 0.18–0.33m in depth, eroding into the sea. Excavation proved this to be a natural feature.

Reference
Movius, H.L. 1940 Report on a Stone Age excavation at Rough Island, Strangford Lough, Co. Down. Journal of the Royal Society of Antiquaries of Ireland 70, 110–42.

School of Archaeology and Palaeoecology, Queen’s University, Belfast BT7 1NN