MSO9350 - Late Roman or early medieval iron smelting site in Blacklake Wood (Monument)

A late Roman or early medieval iron smelting site is visible as a mound of slag adjacent to Windball Hill Road. It has been heavily disturbed by forestry work. Gradiometer survey of the site took place in advance of excavation.

Centred at SS 9040 2868 is a mound of slag and earth, identified by the Exmoor National Park Archaeologist. It lies in coniferous woodland, called Blacklake Wood, to the west of Kennel Farm, adjacent to and on the south side of Windball Hill road. The mound is subrectangular in shape, measuring about 32 metres by 11 metres, and 1.2 metres high. At its west (upslope) end it fades into the natural slope, whilst on the north it is abutted by a stone bank marking a drop of 1.0 metres to the road. The mound contains large amounts of slag, and is either therefore associated with iron smelting or represents a dump of material, possibly connected with the construction of the road. Analysis of the slag is being carried out by the Exmoor National Park Authority as part of its Early Ironworking Project. The mound has been heavily disturbed by forestry work: it appears to have been used as a loading platform, and its top, rutted and uneven, is approached by a former track from the west. The stumps of felled coniferous trees are visible on the mound. Surveyed at 1:500 scale, 22 October 1996. [1-3,24] A charcoal sample was submitted for assay in August 1998.The sample was taken from depth of up to 0.5 metres below present ground level at highest point of the slag heap. The sampled deposit contained a high level of smelting slag and frequent charcoal inclusions, mostly roundwood indicative of coppicing. It also contain fragments of smelting slag and furnace lining, possibly indicative of a collapsed or dismantled furnace. [5] The charcoal sample returned a date range of AD 415-650 (calibrated), placing the smelting activity in the early medieval period. [6] Analysis of the charcoal sample supports an interpretation of a very casual use of fuel with little or no recycling of fuel debris, perhaps indicating unrestricted access to fuel resources. The charcoal was produced mainly from coppiced wood, mostly oak and birch with some hazel and alder. [10] A four week season of excavation took place in 2004 at Blacklake wood as part of the Exmoor Iron project. Two trenches examined smelting waste from two distinct and physically separate areas. Excavation did not identify any stratigraphic relationship between the two trenches but analysis identified several phases of activity separated by breaks or periods of reorganisation. The ceramic assemblage was very small, amounting to 5 sherds in total. Two were identified as South Devon ware, a grey ware typical of the pottery of east Devon and south Somerset in the Roman period. One sherd was an imitation flanged bowl dating to the late 3rd and early 4th Century. Samples of technological debris, including bulk samples, were also collected from both trenches. Similarities between these samples and those from Sherracombe Ford suggest potential in a programme of comparative analysis. A noteworthy find arising from the on site processing of the technological debris was the contents of a small shaft furnace, frozen in mid smelt. Analysis of this find could provide valuable information about the smelting regime at this site. A programme of radiocarbon dating took place following excavation, returning 10 new dates that suggest the site dates to the 3rd and 4th centuries AD, perhaps continuing into the 5th. These date appear to conflict with the dates achieved during the pilot study, described above [6] which placed the site into early 5th to mid 7th Century. [11-16] Finds of slag reported at above National Grid Reference. [18] Gradiometer survey of the site took place in February 2004 as part of the excavation planning process of the Exmoor Iron project. Five anomalies indicating in situ areas of burning potentially associated with iron working were identified along with several uncharacterised archaeological features and possible structural evidence. [19] A sample of possible iron bloom from the Blacklake Wood excavation was submitted for microstructural analysis. Despite being extensively corroded, analysis confirmed the provisional identification of the object as bloom. [21] Reconsideration of the earlier radiocarbon measurements that returned early medieval dates [6] suggests the variance from the more recent samples [16, 23] that returned secure Roman dates may have arisen from the methodology of dating unidentified charcoal bulk samples, which may contain charcoal of different actual ages. [22] Damaged by forestry work and the construction of the road. The extent of the damage has not been ascertained, it is possible that further research may enhance the understanding of this site. [24] This record was enhanced as part of the National Record of the Historic Environment to Exmoor National Park Historic Environment Record data transfer project. [25]