Eachtra Journal

Issue 3 [ISSN 2009-2237]

Archaeological Excavation Report E2771 - Monganstown 1, Co. Westmeath

Metalworking site Final Archaeological Excavation Report, Monganstown 1 N6 Kinnegad to Kilbeggan Co. Westmeath

Metalworking site

August 2009

Client: Westmeath County Council Culleen Beg Mullingar Co. Westmeath

E Number: E2771 Ministerial Order No.: A001/01 Licensee: John Lehane

Contact details: The Forge, Innishannon, Co. Cork. Written by: John Lehane and Penny Johnston Tel.: 021 470 16 16 Fax: 021 470 16 28 E-mail: [email protected] Web Site: www.eachtra.ie E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Table of Contents

Acknowledgements...... iv 1 Non-Technical Summary...... 1 2 Scope of the Project...... 1 3 Receiving Environment...... 2 3.1 Geology...... 2 3.2 Soils and their uses...... 2 3.3 Topography...... 2 4 Archaeological and Historical Background...... 3 4.1 Bronze Age c. 2500-500 BC...... 3 4.2 Iron Age c. 500 BC-500 AD...... 3 4.3 Early Medieval c. 500-1100 AD...... 3 4.4 Later Medieval c. 1100-1650 AD...... 4 4.5 Post-medieval c. 1650-20th century...... 5 4.6 Placenames and Townlands...... 5 5 Site location and topography...... 6 6 Results of the Excavation...... 6 6.1 Area 1...... 6 7 Artefactual Remains...... 10 8 Environmental Remains...... 11 9 Industrial Residues...... 11 10 Discussion...... 12 11 Conclusion...... 13 12 Bibliography...... 14 13 Figures...... 15 14 Plates...... 21 15 Appendices...... 27 15.1 Appendix 1: Context Register...... 27 15.2 Appendix 2: Stratigraphic Matrix...... 33 15.3 Appendix 3: Radiocarbon dates from Monganstown 1...... 34 15.4 Appendix 4: Assessment of Industrial Residues from Excavations at Monganstown 1 ...... 35 15.5 Appendix 5: Charred plant remains and charcoal from Monganstown 1...... 54

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ iii E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Table of contents cont.

15.6 Appendix 6: Analysis of charcoal assemblages from Monganstown 1...... 59 15.7 Appendix 7: Chemical analysis of industrial residues...... 67 15.8 Appendix 8: Lithics Finds Report for A001/01 – Monganstown 1, Co. West- meath ...... 73

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ iv E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Acknowledgements

Senior Archaeologist: John Tierney Licensee: John Lehane Field staff: Ray Riordan, Caroline Healy, Helen Butler, Julian Stroud, Deidre Gleeson, Christina Murphy Additional Post-Excavation Work: Antonia Doolan, Sara Camplese, Illustrations: Enda O’Mahony, Deidre Gleeson Text: John Lehane, Áine Richardson, Antonia Doolan, Penny Johnston Specialists: Neil Fairburn, Marcos Martinón-Torres, Abigail Brewer, Penny Johnston, Mary Dillon, Farina Sternke, Queen’s University Belfast 14Chrono Centre

Works were carried out on behalf of Westmeath County Council and were funded by the National Roads Authority under the National Development Plan 2000 - 2006.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ v E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

List of Figures Figure 1: General situation of site along the route of the new road N6 Kinnegad to Kilbeggan Discovery Series Map...... 15 Figure 2: Monganstown 1 in relation to trenches of archaeological centreline testing that was carried out along the route of the new road...... 16 Figure 3: RMP sites near the route of the new road N6 Kinnegad to Kilbeggan...... 17 Figure 4: The excavated site showing three areas of excavation (Areas 1, 2 and 3) at Monganstown 1, Co. Westmeath (A001/01)...... 18 Figure 5: Area of excavation in Area 1 at Monganstown 1...... 19 Figure 6: Sections through furnaces C.57, C.60 and C.63...... 20

List of Plates Plate 1: General site photograph...... 21 Plate 2: General site photograph...... 21 Plate 3: C.24 post-excavation...... 22 Plate 4: C.63 mid-excavation...... 22 Plate 5: C.40 post-excavation...... 23 Plate 6: C.15 post-excavation...... 23 Plate 7: C.2 post-excavation...... 24 Plate 8: C.66 mid-excavation...... 24 Plate 9: C.50 mid-excavation...... 25 Plate 10: Flint A001/01:13:1...... 25 Plate 11: An example of clay wall remains within furnace C.63...... 26

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ vi E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

1 Non-Technical Summary This report details the results of archaeological excavations carried out at Monganstown townland, along the line of the N6 realignment between Kinnegad and Kilbeggan (Figure 1). Situated approximately 2 km west of Kinnegad, the site is one of fourteen new archaeological sites identified during test excavations carried out in 2004. Following the identification of archaeological remains in situ during the testing phase, it was recommended that preservation by record would be required. The main concentration of archaeological features was confined to an area of approximately 15 m2 and was excavated under ministerial order A001/001. Evidence was recovered showing the existence of early medieval charcoal production pits, and Iron Age ironworking features, as well as other possible smelting and smithing furnaces close by. Several other pits were excavated, and had been used for the disposal of slag, although they may have originally had a different function.

2 Scope of the Project This archaeological services project was carried out on behalf of Westmeath County Council, County Buildings, Mullingar, Co. Westmeath. The project was funded by the National Roads Authority under the National Development Plan 2000-2006. The purpose of the project was to conduct archaeological site investigations within the lands made available for the scheme and to assess the nature and extent of any new or potential archaeological sites uncovered. There were two contracts; Contract 1 (Kinnegad to Tyrellspass) undertaken by Eachtra Archaeological Projects and Contract 2 (Tyrellspass to Kilbeg- gan) carried out by Valerie J. Keeley Ltd. and Cultural Resource Development Services Ltd. This report covers results from Contract 1, Kinnegad to Tyrellspass. Phase 1 of the project (archaeological centreline testing of the route) was carried out in June and July 2004 under licence (04E0908) issued by the Department of the Environment, Heritage and Local Government (DoEHLG). The principal aim of this phase of the project was to investigate known and possible sites of archaeological interest along the route of the proposed road scheme and to investigate the remainder of the route. This was done by a programme of centreline and offset testing (Figure 2). In addition Phase 2 included the resolution of identified sites which were excavated in the townlands of Monganstown, Farthingstown, Kiltotan Collinstown and Rattin. This phase of the project was carried out between January and March 2005 and excavations were carried out by two licensed directors under the direction of a senior archaeologist. In total fourteen sites were excavated during this phase of works and were carried out under Ministerial Order. The sites were situated near the western end of the scheme, in County West- meath, and were found in the townlands of Farthingstown and Kiltotan Collinstown (in the Barony of Fartullagh) and Monganstown and Rattin (in the Barony of Farbill).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 1 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

3 Receiving Environment 3.1 Geology The bedrock geology is mostly comprised of Lower Carboniferous rocks, mainly limestone, which overlies Devonian Old Red Sandstone (Holland 1981; Riada Consult 2003, 58). Some sills of Carbon- iferous volcanic rocks also pass through the bedrock sequences. The dominant topographical feature of Croghan Hill, 7 km southeast of Tyrrellspass, is comprised of shallow intrusive basaltic and dolamitic rocks formed by volcanic activity (Riada Consult 2003, 59). Superficial drift deposits overly the bedrock, varying from impermeable clay to permeable gravel (Ria- da Consult 2003). Glacial features such as eskers and kames dominate an otherwise flat landscape; the eskers are punctuated by sand and gravel quarries that provide good quality building materials (Casey 2002).

3.2 Soils and their uses The soil type encountered in the area (Grey-Brown Podzolic) covers 3.43 % of Ireland, on the southern limit of the north to west Drumlin belt across the northern half of the country (Gardiner and Radford 1980, 91). The lighter Grey Brown Podzolics are ‘good all-purpose soils’ and the heavier Grey Brown Podzolics are better for pasture production (Ibid., 27). Although the soil is technically a fertile type, it has a high clay content which results in poor drainage and peat accumulation in the area is widespread. This is particularly the case along the western portion of the road route, (where the sites from this project were found), which has been covered by the growth of fens and raised bog. These peat lands have generally been worked and, while residual peats are often present, they do not tend to exceed 1 m thickness (Riada Consult 2003, 61).

3.3 Topography The landscape followed by the route of the new road from Kinnegad to Athlone is generally lowlying, ranging from the low undulating drift cover east of Athlone to the flat plains of the central boglands and moraine near Kinnegad. Only a 4 km stretch of the corridor east of Tyrellspass rises above 100 m in height, most of the land undulating gently along the northern extremities of the Bog of Allen. Outside the area of bogland the landscape is typified by regular enclosed fields, bordered by densely overgrown banks with mature hedgerows of ash, elder and hawthorn. This uniform landscape is broken up by streams, eskers and rivers; the River Brosna and its tributaries drain the western part of the study area, while the land east of Rochfortbridge is drained by the Yellow River and other smaller tributaries of the River Boyne (Casey 2002). The moist climate combined with the low-lying condition of much of the area ensures seasonal flood- ing, limiting the land-use capability to livestock grazing punctuated by infrequent tillage. In areas of marginal land close to the edges of the raised bogs the pasture is criss-crossed by drainage ditches without the usual accompanying enclosing bank (Casey 2002).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 2 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

4 Archaeological and Historical Background (based on an Archaeological and Historical Background by Orlaith Egan) The sites within this study area are located in a rich multi-period archaeological landscape (Figure 3) and several monuments have already been assessed in the original EIS report (Riada Consult 2003). Recent excavations along the routes of new roads have added significantly to the list of known sites and the newly discovered sites from this part of the N6 road will add further knowledge to the overall understanding of the area. The area is associated with ancient routeways of unknown date; a large togher discovered by R.A. S Macalister in the nearby townland of Baltigeer in the 1930s may possibly have linked up with the Slighe Dala or Slí Asail, two ancient routeways which led to Tara and Connacht. One of the five great ancient roads of Ireland, the Slí Mór, is also thought to have passed through the area. The earliest indication of archaeological activity within the area was the recovery of a stray Mesolithic Bann flake (IAWU 2002) and a stone axe (SA 1989:17), potentially of Neolithic date (c. 4000-2500 BC), found in the townland of Rattin (IAWU 2001). However, all the archaeological sites excavated in the area have been identified as of Bronze Age date or later.

4.1 Bronze Age c. 2500-500 BC The earliest known evidence of settlement is represented by an Early Bronze Age (c. 2500-1500 BC) togher discovered by the Irish Archaeological Wetland Unit in the townland of Rattin (IAWU 2001). Finds from the area include a socketed bronze axe head from the Late Bronze Age (1936:1873 NMI) which was recovered near Kinnegad townland (exact location unknown). Bronze Age burnt mounds are also reasonably common; one definite example (CHS 20) and several potential sites were discovered during fieldwalking of the proposed route (Riada Consult 2003, 247). A burnt mound was excavated at Kiltotan Collinstown 12 and at Kiltotan Collinstown 13 an anomalous pit produced a Middle-Late Bronze Age radiocarbon date.

4.2 Iron Age c. 500 BC-500 AD There is a general dearth of evidence from this period in the Irish archaeological record. However, two Iron Age radiocarbon dates from iron smelting pits were produced at this site at Monganstown 1.

4.3 Early Medieval c. 500-1100 AD The record of Early Historic activity in the study area and the surrounding countryside is rich. An ancient monastic site founded at Clonfad (WM27: 067, WM27: 066) to north of the townland of Rattin, consisted of a possible church site, a rectangular enclosure, a graveyard, crosses, and Bishop’s grave. St. Etchen was bishop of Clonfad or Cluain-fota-Baethain in the sixth century and the annals of Ulster record his death between 578-84 AD. He is reputed to have ordained St. Columcille and St. Colmáin mac Lúacháin of Lynn and many others. The monastery survived into the eight century AD as Blahmac, an abbott of Clonfad was killed in 799 AD (Gwynn & Hadcock 1970). Sites with evidence for craft/industry from this period include charcoal/metalworking pits at Monganstown 1, which produced two Early Historic radiocarbon dates, and material from a small metalworking pit at

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 3 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Kiltotan Collinstown 14, which also returned an Early Historic date. The most common settlement monuments of this period are ringforts (also known as rath or lios). These are interpreted as enclosed farmsteads and they generally consist of a circular ditch outside an earthen bank (constructed with the upcast from the excavation of the ditch). Larger examples may have more than one ditch and a bank forming the enclosure. At Kiltotan Collinstown 14 there are four ringforts located within 1 km of the site. The nearest, WM033-061, is situated just 350 m to the north. East of the site are WM033-065 and WM033-066, at distances of 700 m and 400 m respectively. The fourth ringfort, WM033-062, is about 840 m north-northeast of the site. Also within 1 km of the site is a fifth recorded monument, WM033-068, listed as an earthwork. At Monganstown 1 a further two ringforts (WM 027:069 and WM 027:070) are located in the area between the site and the town of Kinnegad, with one, just 500 m from the site. The evidence suggests that the area was quite intensely settled during the Early Historic period.

4.4 Later Medieval c. 1100-1650 AD The villages of Tyrrellspass and Rochfortbridge both date from the medieval period and are located near the sites examined in this project. Rochfortbridge is located 11 km southwest of Kinnegad within the Barony of Fartullagh. The village is named after the Rochforts, a French family who settled in Ireland in 1243. Before the Rochforts established themselves in the area the Tyrrells, a powerful An- glo-Norman family, held the Barony of Fartullagh. This included the lands around Rochfortbridge and the parish of Castlelost. In the 13th century (c. 1411) the Tyrrells built a castle that consisted of a motte and bailey (a stone castle came later) in Tyrrellspass, to the northwest of the town. It guarded the western entrance to the Barony of Fartullagh and it remained the centre of power for the Tyrrells until the Cromwellian Invasions (1650). They also built a semi-fortified manorial church on the castle lands which contained an effigy of armoured Knight John Tyrell dating to 1607. The site of another castle (WM 027: 071) and a bridge (WM 028: 003), reflecting further settlement in the later medieval period, are located in the townland of Kinnegad. A defensive castle was also constructed in Rattin (WM34: 008), built to defend extensive Anglo-Nor- man territories in the midlands. The lands were owned by Hugh de Lacy but passed into the possession of Sir John Darcy and his descendents when he became Chamberlain and Steward of the household of King Edward ΙΙΙ, Chief justice of England and Peer of the realm. In the Insurrection of 1641 Nicholas Darcy forfeited Rattin and the greater part of his estates (Bardon 1913). The remains of a sixteenth century towerhouse (called Rattin Castle) are found in the townland today. Test trenches c. 1500 m to the northwest of the castle failed to produce any archaeological remains (Conway 1999, 298). How- ever, phase 1 test excavations north of Rattin towerhouse found the remains of a probable timber and brushwood trackway. There is also a recorded earthwork site (WM 34:007) located to the west of the castle but its function and date are unknown. The only excavated site from this project that dated to the Medieval period was a small furnace at Kil- totan Collinstown 14 which produced a fifteenth century radiocarbon date.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 4 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

4.5 Post-medieval c. 1650-20th century The post-medieval archaeology examined during the project included the remains of field systems and vernacular architecture. Post-medieval field boundaries and ditches were excavated at Farthingstown 009, Farthingstown 011, Monganstown 2, Kiltotan Collinstown 12, Kiltotan Collinstown 13 and Kil- totan Collinstown 14. The field systems at Kiltotan Collinstown are probably related to a nineteenth century farmstead pictured on the 1st Edition OS map and located immediately adjacent to the site at Kiltotan Collinstown 12. Vernacular architecture was also examined at the site of Rattin 4, where a nineteenth century farmhouse was tested and recorded prior to road construction. South of Rattin 4 the post-medieval Clonfad mill still stands. This mill has been subjected to a standing building survey and has been reported separately. In addition there are two demesnes located within the study area (Farthingstown House and Side- brook House), both found in the townland of Farthingstown. The term ‘demesne’ originates from Norman French and indicates the portion of an estate retained by a feudal lord for his own use. The term is also used to indicate the extent of a wealthy landowners post-medieval landholding, immediately adjacent to their home. Most Irish examples typically consist of a big house with associated buildings, ornamental and recreational grounds, and perhaps the remains of an elaborate boundary wall (Riada Consult 2003, 249).

4.6 Placenames and Townlands The sites were excavated in the townlands of Farthingstown, Kiltotan Collinstown, Rattin and Mon- ganstown. The townland of Farthingstown lies in the parish of Castlelost in the barony of Fartullagh. It covers a substantial area containing c. 1802 acres. It is known in Irish as Baile na Feóirlinge meaning ‘town of the farthing’ (Walsh 1957). It was known as ‘Ballyneforlin alias Fardingston’ in the inquisi- tions of the seventeenth century (Inq. Car. Ι no. 129). The townland of Kiltotan and Collinstown is also located within the parish of Castlelost and the Barony of Fartullagh. It lies south and southeast of the old mail coach road from Tyrrellspass to Dublin and borders part of the County of Offaly. In 1837 it consisted of c. 320 acres, which was mainly of arable and pastureland but included a narrow stretch of bog, which bordered the parish of Newtown. Kiltotan is known in Irish as Cill Toiteáin meaning ‘the church of the conflagration’. Collinstown is known as Baile Choileáin translated as ‘the town of Collins’. (OS Namebooks). The townland of Rattin is located in the west of the parish of Killucan within the barony of Farbill. It is known in Irish as Rath Aitinne meaning ‘Rath of the furze’. The lands of Rattin were formerly part of the lands of Clonfad, situated to the west. The name Monganstown is derived from ‘the town of the Mongans’ and the townland covers an area of 483 acres.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 5 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

5 Site location and topography The site is situated approximately 2 km west of Kinnegad town, near the eastern end of the new route (Figures 1-3). It is 500 m south of the current N6, at Chainage 57240 (field 123/National Grid Co-or- dinates 258000/244650), in the townland of Monganstown, in the parish of Killucan and the barony of Farbill. The excavation was located 150 m north of, and in the flood plain of, the Kinnegad River, which forms the county boundary between Westmeath and Meath. It is on a slightly raised ridge in a level boggy field, which has been in use as pasture for cattle and sheep. It appears that the drier ground of the raised ridge would have attracted occupation in the past. The Boyne Valley Drainage Scheme in the 1950’s is responsible for the presence of deep field drains up to 3 m in depth close to the site.

6 Results of the Excavation Three separate areas (Area 1 to 3) were stripped of topsoil during the original test excavations of the new route (Figure 4). Area 1 was the largest, being 53 m long (east to west) and 20 m wide, though of an irregular shape. The features identified included two charcoal production pits, six furnaces and several other pits and postholes (Figure 5, Plates 1 and 2). Area 2 was located 19 m west of Area 1 and just 4.4 m west of Area 3. It was 20 m long (northeast to southwest) and 10 m wide. The area was opened in order to uncover a possible linear feature identified during the testing phase but the only features visible were the remnants of a lazy-bed system. Area 3 was then opened between Areas 1 and 2, 4.4 m to the east of Area 2. It measured 20 m by 4 m with a long-axis running northeast to southwest. No archaeological evidence was uncovered in this area. A full record of the excavated contexts is presented in the context register in Appendix 1 while the stratigraphic matrix is presented in Appendix 2.

6.1 Area 1 6.1.1 Charcoal pits Two large charcoal production pits (C.21 and C.24) were identified during excavations. The pits had a similar morphology, being sub-rectangular in plan with almost vertical sides and flat bases. Both pits were almost 3 m long, between 1.2 m and 1.3 m wide and 0.2 m deep, with C.21 the larger of the two. A large amount of charcoal remained in situ in both pits. The larger of these pits (C.21) measured 2.85 m north-south x 1.3 m in width x 0.26 in depth. It was situated in a corner of the site, approximately 20 m northwest of the main concentration of features and 14 m from the other charcoal pit (C.24). It contained an undisturbed charcoal rich fill (C.14) which was interpreted as the original fill of the pit following charcoal manufacture. Analysis of this material has shown that oak was the main raw material used for the manufacture of this charcoal and it has produced an Early Historic radiocarbon date of AD 898-920. There was no indication why this material was not utilised following its production. The second pit (C.24) measured 2.8 m northwest-southeast by 1.2 m in width and a maximum of 0.2 m in depth (Plate 3). This pit was identified during the test excavation and the upper fill (C.25) was material that was re-deposited as a result of archaeological test trenching. This overlay the original

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 6 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

charcoal fill (C.13) which contained a piece of burnt struck flint (A001/001:13:1). The charcoal was identified as oak and it yielded an Early Historic radiocarbon date of AD 897-1024. The orientation of the individual wood pieces could be discerned lying along a northwest-southeast alignment, the same orientation as the pit itself.

6.1.2 Furnaces Six furnaces were excavated on the site; C.40, C.47, C.57, C.60, C.63 and C.74. Four of these (furnaces C.47, C.57, C.60 and C.63) were situated roughly in the centre of the site (Figure 5). The most southerly furnace (C.47) measured 0.54 m in length and 0.5 m in width. The lower fill (C.46) was composed of burnt clays that were interpreted as being part of the superstructure of the furnace. A total of 8 kg of slag were also retrieved from this deposit and chemical analysis indicated that this slag was from a bloomery furnace (Appendix 7). The second fill (C.48) was topsoil-like and probably accumulated in the furnace after the original furnace contents were removed. The smallest furnace on the site (C.57, Figure 6) measured 0.4 m by 0.3 m in plan. It contained two fills (C.58 and C.59); the lower fill (C.59) consisted of various slag and oak charcoal elements. The upper fill (C.58) was a layer of slumped topsoil that filled the furnace following removal of the original contents. Two furnaces (C.60 and C.63) were of similar size and morphology; both were sub-circular with aver- age dimensions of 0.78 m by 0.61 m by 0.42 m. The furnace C.60 contained two fills (C.61 and C.62). The basal fill (C.62) was a charcoal-rich deposit found at the base of the cut (Figure 6) and interpreted as the remnants of the original charcoal (identified as ash) used as fuel in the furnace to attain the high temperature necessary for smelting. A total of 9 kg of slag were retrieved from this deposit. The upper fill (C.61) was similar to the upper fills of other furnaces, and contained topsoil-like sediments as well as 12 kg of slag. The furnace C.63 (Plate 4) differed from all the other furnaces in that it contained three fills (C.64, C.65 and C.75). The basal fill (C.75) was a layer of nearly pure charcoal that was recorded at the bottom of the cut (Figure 6) and it contained 14 kg of slag. It was similar to C.62, the basal fill in furnace C.60. This basal layer was overlain by a deposit of burnt clay (C.65) with possible structural elements (burnt clay and charcoal) that may imply a domed superstructure that collapsed into the furnace when it was broken to extract smelted ores. A total of 15 kg of slag remained in the deposit. The uppermost fill of this furnace (C.64) contained small amounts of slag and some oak charcoal and was probably disturbed furnace fill that accumulated in the furnace after its use. The largest furnace was C.74/C.40. This was an elongated, sub-rectangular pit/furnace (C.74) with a small circular furnace (C.40) at its southern end (Plate 5). This was found on the western edge of the main grouping of features in the centre of the site. C.40 was c. 0.55 m in diameter and 0.32 m in depth and it contained two fills (C.39 and C.38). The primary fill (C.39) was a dark slag rich mix, with 15 kg of slag retrieved along with charcoal from a diffuse-porous wood type. This context returned an Iron Age radiocarbon date of BC 361-113 (UB 6940, see Appendix 3). The upper fill of the furnace (C.38) was interpreted as a mix of topsoil and collapsed materials (including 3 kg of slag and ash, hazel and yew charcoal) that filled the furnace subsequent to the use of the feature. Recovered from this fill were two large pieces of vitrified clay lining that were originally part of the furnace; they may have formed

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 7 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

part of a shaft (see Appendix 4) or they may have been from the clay walls of a bloomery furnace. One of these pieces has a small round hole towards its base for the insertion of bellows. The edges of C.74 extended northwards from this circular bowl, approximately 1.3 m in length by 0.5 m in width. These dimensions made it roughly 1 m longer than the other furnaces. This part of the furnace was filled by two deposits (C.72 and C.73), the basal fill (C.73) being very disturbed and composed of burnt clays (some possible fragments of superstructure), slag pieces (13 kg), charcoal and topsoil-like material. The presence of topsoil-like material in these fills may indicate some disturbance by ploughing in the general area, which resulted in the formation of homogenous brown fills mixed with debris from the furnace which had been dismantled to remove the processed ores. The upper fill (C.72) was clayey sand topsoil that probably slumped into the furnace void following a period of disuse. The irregular morphology of this feature implies that it was severely disturbed after smelting/smithing.

Furnace cut Length Width Depth Fuel/Charcoal type Clay walls Vitrified clay lin- present ing present C.47 0.54 0.5 Oak and yew Yes (in C.46) No C.57 0.4 0.3 Oak No No C.60 0.8 0.6 0.35 Ash No Yes (in C.61) C.63 0.76 0.62 0.5 Oak Yes (in C.65) Yes (in C.65 and C.75) C.74/C.40 1.3 0.55 0.32 Ash, hazel and yew Yes (in C.73 and Yes (in C.38 and (diameter of C.38) C.39) C.40 0.55 m)

6.1.3 Stakeholes One stakehole (C.6) was situated in the southeast corner of the site. It had only one fill (C.5) indicating that it was formed by a driven post/large stake, which was subsequently removed. A similar stakehole (C.15, filled by C.16) was situated in the northwest corner of the site (Plate 6). The charcoal from this stakehole was identified as oak (Appendix 6) and it probably represents the remains of a burnt post (Appendix 6). Whether these two features formed part of larger structures is unknown, but there is no evidence for associated stakes or postholes. However, it is quite likely that some form of shelter was erected around the metal working site as this would have been necessary to protect against the elements.

6.1.4 Clay Extraction Pits Six pits (C.2, C.23, C.45, C.52, C.66 and C.79) were excavated on the site. They ranged in size from 4 m to 0.80 m in length, with depths of between 0.67 m and 0.10 m. The first pit (C.2) was an irregular cut and measured 3.84 m northwest-southeast x 0.80 m in width x 0.58 m in depth. It was situated in the northeast corner of the site (Plate 7). The primary fill (C.4) contained oak and hazel/alder charcoal, fragments of clay and 5 kg of slag waste. The other fills (C.1 and C.3) both contained charcoal (including oak, ash and elm) and slag, with most slag being retrieved from C.1, where 2 kg of slag was found. The oak charcoal identified from C.1 returned an Iron Age radiocarbon date of BC 338-46 (UB 6940, see Appendix 3). This pit was probably originally used as a quarry pit, for extracting clay that was used to build the superstructure of the nearby furnaces. It

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 8 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

was evidently later backfilled with waste material from the iron-working process taking place in the surrounding area. A similar pit (C.69) was situated approximately 1 m southwest of C.2. It measured 3.3 m southeast- northwest x 0.8 m x 0.4 m in depth and it had two fills (C.70 and C.71). The basal fill of the pit (C.71) contained small to moderate slag and oak charcoal inclusions (see Appendix 6 for charcoal identifications) and it represents deliberate deposition of waste material from furnace activities. One fragment of a hazelnut shell was recovered from this context, indicating that waste from domestic activities may have been used to set the furnaces alight (Appendix 5). The upper fill (C.70) consisted of a mix of redeposited material that included topsoil. This pit was interpreted as a quarry pit and, like the other quarry pit C.2 and the disused furnaces, it was backfilled with waste from metalworking that was carried out at the site. Another linear quarry pit (C.45) was found 2 m south of C.2 and C.69, and it was located beside a deposit of orangey gravely clay. It measured 3.8 m east-west x 0.8 m x 0.4 m in depth and it had three fills (C.44, C.56 and C.55). The basal fill (C.44) contained frequent charcoal and slag inclusions as well as some plant remains. It was overlain by C.56, a deposit of red burnt clay, probably the lining of a furnace. This deposit also produced plant remains. The plant remains from both contexts were in a poor state of preservation but may represent domestic material that was used as tinder for the furnace fires; C.44 produced one charred unidentified seed, one stem or straw fragment, and one indeterminate fruit stone and C.56 produced one hazelnut shell fragment during plant remains analysis (Ap- pendix 5). Both of these deposits evidently represented redeposited furnace waste. The uppermost fill (C.55) contained only small amounts of charcoal, probably re-deposited from the surrounding area, and no industrial residues and may have been a result of natural silting.

Another pit (C.23) was situated at the southeast edge of the main body of features and to the south of the pit C.45. There was no evidence of burningin situ or any other indicator of use. It had four fills (C.31, C.30, C.22 and C.32). The primary fill of the cut (C.32) seemed to represent a deliberate back- filling episode. Other small deposits such as C.22 and C.30 contained small amounts of slag (there was 1 kg in C.22) and represented rubbish deposited in the pit. The uppermost fill (C.31) may have been the result of natural silting. The remaining pits were centrally located and they produced the most complex stratigraphy on the site. The earliest pit was C.79 which was situated next to one of the charcoal pits (C.24) at the northwest corner of the main concentration of the archaeology. The pit (C.79) measured 2.9 m northwest-southeast x 1.78 m x 0.67 m in depth. It had two fills (C.68 and C.29). The lowest was C.68, a deposit of silty sand with 10 kg of slag, interpreted as a layer of disposed slag. This underlay a spread of dark sediment with frequent charcoal inclusions (C.29) and it filled the pit and spread outside its confines, covering an area of approximately 3.2 m by 1.3 m. This spread (C.29) was also cut by another pit (C.66) which measured 2.5 m north-south x 0.85 m x 0.23 m in depth and was filled by C.67, a silty sand with occasional slag and charcoal (Plate 8). This pit (C.66) was in turn cut by a crescent-shaped pit (C.52) which was situated in the centre of the site. It measured 3.5 m east-west x 1.1 m x 0.3 m in depth and it followed and respected the edge of a natural deposit of orange gravely clay. It was filled by

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 9 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

layers C.53 and C.54, both of which had inclusions of slag and charcoal pieces. The charcoal-rich spread (C.29) that partially filled the earliest pit (C.79), was overlain by a thin layer of re-deposited industrial waste material (C.33) where charred seeds and hazel nut shell fragments were found (Appendix 5) along with fragments of oak charcoal (Appendix 6). This spread had dimensions of 1.0 m north-south x 0.4 m x 0.1 m and it was cut by another pit (C.36) which measured 1.5 m east- west x 1.45 m x 0.16 m and was filled by C.34, which contained frequent slag and moderate charcoal inclusions. It was cut by another pit (C.37) which measured 0.8 m north-south x 0.5 m x 0.1 m, and was filled with a layer of slag-rich industrial waste (C.35). It was postulated during excavation that some of these pits were dug for the purpose of extracting deposits of ore from the subsoil, as many of the pits were situated in proximity to natural deposits of orange gravely clay. Small lumps or nodules of possible iron ore and manganese deposits were identified along the sides of natural deposits during the excavation and it was surmised that higher concentrations were mined out at the time the site was in use. However, it is extremely unlikely that these pits could have produced anything near the amount of ore required for smelting purposes (Neil Fairburn, pers. comm.). The pits were found adjacent to the furnaces and their positions appear deliberate. It is therefore possible that some of the pits were excavated in order to extract clay for the manufacture of the furnace superstructures, for example clay domes, and were later utilised for the disposal of the industrial waste.

6.1.5 Pits of unknown function A shallow depression or possible pit (C.10) was located in the southeast corner of the site and may have been associated with a nearby stakehole (C.6). It was irregular in plan and measured 0.5 m east-west x 0.45 m x 0.07 m in depth and had one charcoal-flecked fill (C.9). Another shallow pit (C.50) was located to the west of the main concentration of archaeological features (Plate 9), it measured 0.54 m east-west x 0.5 m x 0.15 m and it had only one fill (C.51). A third pit (C.78) situated nearby, measur- ing 2.05 m southwest-northeast x 1.1 m x 0.21 m and containing two fills (C.76 and C.77). The lower fill (C.76) consisted of a thin layer of charcoal while the upper fill (C.77) represented deliberate back- filling. The absence of evidence for burning in situ suggested that the charcoal production occurred elsewhere.

7 Artefactual Remains Only one find was recovered from this site, a burnt flint (A001/01:13:1, Plate 10) which was examined by Farina Sternke (Appendix 8). It was a type that was probably made in the Early Mesolithic or the Neolithic, although the find spot is probably typically Mesolithic (an island of raised ground surrounded by marchland). This is an early find that has evidently been disturbed and incorporated into deposits associated with later archaeological activity.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 10 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

8 Environmental Remains Twenty-one bulk soil samples and thirteen charcoal samples were analysed for the retrieval of charred seeds and charcoal. The flots for charred seeds were examined by Penny Johnston and Abigail Brewer (Appendix 5) and the charcoal from the site was examined by Mary Dillon (Appendix 6). Charred seeds were only found in six samples, these were mostly fragments from hazel nut shells. Most of the charcoal from the site was identified as oak, but ash and yew were also significant elements of the assemblage.

9 Industrial Residues Visual inspection of the industrial residues from the site was carried out by Neil Fairburn (Appendix 4) and chemical analysis of three samples was carried out by Marcos Martinón-Torres at the Archaeo- logical Science Laboratories at the Institute of Archaeology in London (Appendix 7). Several of the samples resembled tap slag which suggested that the smelting occurred in a shaft furnace. If this was the case it is one of the earliest examples known from Ireland, where iron smelting technology is generally considered deeply conservative (see Appendix 4). Chemical analysis suggested that the material, although visually resembling tap slag, was actually material that accumulated at the bottom or around the walls of a bloomery furnace (Appendix 7). The composition may be because this slag was a by- product, rather than a product, of the smelt (Appendix 4). Chemical analysis also indicated that the ore source was high in manganese and that the smelting process was quite efficient. High manganese (Mn), strontium (Sr) and barium (Br) are the indicators that could help to identify the potential ore source in the vicinity.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 11 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

10 Discussion All of the features excavated at this site were associated with iron working and included charcoal pits, furnaces and various extraction and/or disposal pits filled with charcoal and slag. In total, twenty-three cut features were excavated at the site and forty-eight deposits (fills and spreads) were investigated. The stratigraphic matrix from Monganstown 1 was fairly simple, with no difficult sequence of inter-cutting features, implying a relatively short-lived phase of metalworking activity. Metalworking residues excavated across the river at Kinnegad 2 (Murphy 2003 and Murphy 2004, 510) indicate that the general area around Monganstown was a favoured location for metalworking. The location of these sites is dependant on two primary factors; access to fuel and access to ore. Access to bog ore was presumably relatively easy in the midlands as much of this part of the country is covered by raised mires. The bog may also have provided a source of fuel as peat has been identified as a fuel used to fire furnaces in some parts of Scotland in prehistory (Pleiner 2000, 129). However, the primary source of fuel at this site was identified as oak charcoal (Appendix 5 and 6). The fact that charcoal production pits were excavated at the site indicates that the metalworking was carried out near the location of the fuel source; Pleiner (2000, 118) suggests that it is more economical to locate the site near the fuel source rather than the ore, especially when fuel is in short supply, as ore is smaller in volume and lighter in weight than the charcoal required for smelting and it is therefore less problematic to transport it. The method of making charcoal was probably relatively simple, involving the excavation of a pit, filling it with timber, covering in vegetation to limit the oxygen supply, then lighting the fire and allowing it to smoulder, slowly turning the wood to charcoal. This production process was similar to that identified at Kinnegad 2 (Murphy 2003) It is clear that substantial iron working was carried out at Monganstown 1. In total 61 kg of industrial residues were recovered from a number of pits and furnaces during the archaeological excavation. The industrial residues consisted of material associated with iron working including furnace and amorphous slag from most contexts. An iron smithing PCB (Plano Convex Bottom), was found in C.62 (from furnace C.60) and pieces of vitrified furnace lining were found in several of the deposits within furnaces (C.60, C.63 and C.40) and in slag-rich waste deposits that backfilled three clay extraction pits (C.2, C.23 and C.45) and in another small pit (C.50), presumably cleared after use of (Appendix 4). Fragments of vitrified clay-lining from Monganstown were very large and they indicate a feature with an internal diameter of c. 0.3 m. In some cases this indicates a feature considerably smaller than the width of the excavated furnaces (generally c. 0.5 – 0.6 m wide) and it may suggest that these furnaces were the bases of shaft furnaces (see Appendix 4). There was also a considerable quantity of material that resembled fluid tapped slag (27.5 kg), the type of slag that is produced in shaft furnaces. However, chemical analysis on a limited number of samples suggested that it was actually slag from a bloomery furnace (see Appendix 7). In this case the vitrified clay linings found among the slag and the fragments of clay walls found within the furnaces may actually be the surviving remains of domes or clay walls temporarily built over small bowl furnaces. Partial remains of the superstructure of several furnaces survived in the form of collapsed clay walls (in C.47, C.63, C.74 and C.40, see an example in Plate 11).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 12 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

In general, simple bowl furnaces are thought to have been utilised at the beginning of the Iron Age and into the Early Historic and medieval period in Ireland (Scott 1991). The identification of shaft furnaces has important technological implications as the procedures used in shaft furnaces were more efficient. There is a possibility that there was a shaft furnace at Monganstown 1, indicating a relatively sophisticated smelting process and suggesting a very early date for use of this technology; the earliest known examples are from a small quantity of tap slag retrieved at Farranstack, Co. Kerry (03E0171) where the furnace was identified as a shaft furnace producing radiocarbon dates of between 1020-1270 AD (Dowd and Fairburn 2005). However, chemical analysis has yet to confirm this issue for certain. What is clear is that the smelting process was relatively efficient, perhaps due to the qualities of the ore used, rather than because of the technology (Appendix 7). The Monganstown 1 radiocarbon results demonstrate a wide period of use at the site, with deposits from the area associated with metalworking producing Iron Age dates and deposits from charcoal production pits indicating Early Medieval activity. This suggests a long period of intermittent activity at the site. Such repeated usage is perhaps down to the ideal location, on a slightly raised island that probably attracted occupation in the past, as the area has been extensively drained since the 1950s. The only artefact from the site, a flint that was disturbed from its original depositional context, indicates that occupation in the general area extends even further back than the radiocarbon dates suggest.

11 Conclusion Charcoal pits and furnaces associated with iron working were excavated at Monganstown 1, with material from the site producing both Iron Age and Early Historic dates. The industrial residues were primarily identified as the remains from iron smelting with a small amount of iron smithing residues. The results from this site will feed into a growing collection of discoveries that is changing our knowledge about metalworking technology available in Ireland in the past.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 13 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

12 Bibliography Bardan, P. 1913. ‘Fairbill Topography’, Typed manuscript, Westmeath County Council Library

Casey, M. 2002. ‘N6 Kinnegad to Athlone Aerial Survey’, section of the N6 Environmental Impact Statement, 2004.

Conway, M. 1999. ‘Rattin’, in Bennett, I. (ed.) Excavations 1999: Summary account of archaeological excavations in Ireland. Bray, Wordwell.

Dowd, M. and Fairburn, N. 2005 ‘Excavations at Farranstack, Co. Kerry: evidence for the use of shaft furnaces in medieval iron production’ Journal of Irish Archaeology XIV, 115-121.

Gardiner, M.J. and Radford, T. 1980. Soil Associations of Ireland and Their Land Use Potential. Soil Survey Bulletin No.36. An Foras Talúntais, Dublin.

Gwynn, A. and Hadcock, R.N. 1970. Medieval Religious Houses. Ireland. London.

Holland, CH (ed.) 1981. A Geology of Ireland. Scottish Academic Press, Edinburgh.

Irish Archaeological Wetland Unit, UCD, 2001. ‘Cavemount, Esker & Derryhinch Bogs, Cos Meath, Offaly & Westmeath’, in Peatland Survey Report 2001.

Irish Archaeological Wetland Unit, 2002. ‘Fieldwork 2000, Counties Westmeath and Offaly’ in I. Bennett (ed.) Excavations 2000. Bray, Wordwell.

Murphy, D. 2003. ‘Archaeological Excavation Report for site K-E-K M4 Motorway, Contract 1, Kinnegad 2, Co. Westmeath, Licence 02E0926’, Unpublished report by Archaeological Consultancy Services Ltd.

Murphy, D. 2004. ‘Kinnegad 2, Multi-period smelting site 25838 24541 02E0926 Co. Westmeath’ in I. Bennett (ed.) Excavations 2002. Bray Wordwell.

Ordnance Survey field name Books of the County of Westmeath, 1837.

Pleiner, R. 2000. Iron in Archaelogy: The European Bloomery Smelters. Oxford, Oxbow Books.

Riada Consult, 2003. ‘N6 Kinnegad to Athlone Dual Carriageway Environmental Impact Statement’, Unpublished report for Westmeath County Council.

Scott, B. 1991. Early Irish Ironworking. Belfast, Ulster Museum.

Walsh, Rev. Paul, 1957. The Placenames of Westmeath. Dublin

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 14 Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ E2771 | A001/01 13

Figures Monganstown 1,Co.Westmeath I

Monganstown Monganstown Clonfad Mill 2 1 Rattin 3 Rattin 5

Rattin 4

Farthingstown 6 SSUE

Farthingstown Farthingstown 10 11 Farthingstown Kiltotan and 8 3: EachtraJournal-ISSN2009-2237 Kiltotan and Collinstown Collinstown Site 12 Site 14 Farthingstown 9

Kiltotan and Collinstown Site 13

0m 2000m

Figure 1: General situation of site along the route of the new road N6 Kinnegad to Kilbeggan Discovery Series Map 15 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 1 Km Legend: Archaeological sites outline C.P.O. Trenchs Test 0 m

00 57500.0 - 1 e t 0 i 0 s

n g w n o k i r s t o n w a r n o o r M I

57000.000

5650 0.000

d l - e i 2 0 F l 0 a v w n e i t o d s e m n M a

t e g s e r t y o a s L M

56000.000

55500.000

00 55000.0

54500.000 y a - c k w 3 a r 0 t

0 n n i e t t d a o R o W

54000.000

00 53500.0 g n i d l i 000 u - 00. 4 B 30 5 l 0 - 0 5

v a 0 n i e i t 0 h t t d n i r a e t t a R m a e t R H s o P N

0 00 0. 50 52 Figure Monganstown 2: archaeological trenches of to 1 in relation centreline testing that was carried out the of new along road the route

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 16 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Monganstown 1 Monganstown 2 Rattin 3 Rattin 4 Clonfad MillClonfad Rattin 5 N 2km 0 km Motte Earthwork Site Ecclesiastical Castle Site Potential Ringfort Rectangular enclosure Barrow Farthingstown Farthingstown 6 Farthingstown Farthingstown 8 Farthingstown Farthingstown 7 Farthingstown Farthingstown 9 Farthingstown Farthingstown 10 Farthingstown Farthingstown 11 Kiltotan and Collinstown 13 Site Kiltotan and Collinstown 12 Site Kiltotan and Collinstown 14 Site Figure RMP 3: Kinnegad the near of newN6 road sites the route Kilbeggan to

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 17 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 5 m N C.2 200 m C.45 C.23 C.69 C.6 C.8 0 C.10 1 m C.78 C47 C.57 C.52 C.63 C.66 C.27 C.20 C.79 C.74 C.60 C.50 C.24 C.40 C.12 0

000 57500. C.21 C.15 C.18 /01 2 2 1 3 0 6 6 4 4 e A0 0 t 4 4 i n 2 2 s

w g N N o

t n i 3 3 s 2 2 n r k 9 9 o 7 7 w 5 5 ng a n 2 2 o

o r M E E I

57000.000

Kinnegad River Figure 4: The excavated site showing three areas of excavation (Areas 1, 2 and 3) at Monganstown 1, Co. WestmeathFigureshowing The (A001/01) 4: excavatedsite three 1, Monganstown Co. at 3) areasof (Areas 1, 2 excavation and

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 18 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 5 m N

C.2

C.45 C.23 45 E 0 N C.69

C.6

C.8 0 C.10 1 m F C.78 C47 C.57 E

T1 C.52 40 E 0 N C.63 C.66 S1 C.27 C.20 C.79 C.74 C.60 L1 C.50 K1 C.24 C.40 35 E 0 N C.12 30 E 0 N Extent of Excavation Indicating areas of different soil type Indicating areas of different C.21 Possible Clay Extraction Pits Non-Archaeological C.15 C.18 Furnace Charcoal Pit 20 E 0 N Monganstown A001-001 Post-Excavation Plan Monganstown Figure Area 5: excavation in of Area Monganstown 1 at 1

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 19 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

42.20 E 42.66 E 0 N E F 0 N

59 57

10 cm 0 50 cm

63.55 E Baked clay 36.42 E 1.00 N 2.28 N K L

60 62

10 cm 0 50 cm

W E S1 T1

10 cm 0 50 cm

Figure 6: Sections through furnaces C.57, C.60 and C.63

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 20 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

14 Plates

Plate 1: General site photograph

Plate 2: General site photograph

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 21 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Plate 3: C.24 post-excavation

Plate 4: C.63 mid-excavation

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 22 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Plate 5: C.40 post-excavation

Plate 6: C.15 post-excavation

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 23 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Plate 7: C.2 post-excavation

Plate 8: C.66 mid-excavation

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 24 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Plate 9: C.50 mid-excavation

Plate 10: Flint A001/01:13:1

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 25 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Plate 11: An example of clay wall remains within furnace C.63

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 26 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples None Soil 3 None None None CharcoalSoil 17 3 Charcoal 2 None Soil 4 Slag 2 Charcoal 4 None None Soil 1 Slag 5 Charcoal 1 Soil 5 None Finds None None None None None 1 None None None None None None None None None - - - Description Friable darkFriable brown silty sand with moderate charcoal flecks. fine Frequent sub-angular pebbles. Only fill shallow of depression C10. Irregular shape in plan. Break imperceptible, slope at top of sides smooth & gentle. Break slope at baseof gradual imperceptible. to Base flat. Filled C9. by Natural like feature. material. Topsoil Irregular shape interpreted as tree root Firm black charcoal 98% fill with very occasional small occasional stones. Very slag. In C24. Black charcoal 100% fill. C21. over Soft. C81, Under Sub-circular stakehole. Break sharp. slope at top of Sides steep and smooth. Break slope of at base gradual. & filled Over Under C82. C16. by Irregular break slope gradual of shape. Top at N & E sides, Sides vari sharp at S & W. ously steep, irregular, undercut and vertical. Soft greyish fine black siltyFrequent pebbles, moderate clay. charcoal flecks. Only fill of C8. Mid brownish black charcoal, soft siltyFrequent clay. flecks large to pieces. Moderate peb bles and stones. Over under C3, C81. Irregular/linear in plan. Rounded corners. Break sharp top: slope of S, gradual imper N, Sides: steep N & S, E & W mod. undercut. S side ceptible Breaks E & W. base: slope of gradual. C4, Fills: C82. over 2 & 4. Under Cs 1, Soft mid reddish Occasional brown sandy clay. moderate, to fine coarse, to sub angular pebbles. Occasional small & medium sub rounded stones. Occasional charcoal and slag. C4. over C1, Under Firm dark brownish black. Occasional medium sized pebbles small to stones, sub rounded angular.to Moderate amount charcoal of flecks pieces. to Occasional burnt clay & slag. C2. over C3, Under Soft mid-brown silty clay with occasional charcoal. sub-angular Frequent fine pebbles, occasional sub-angular C6. over small C81, stones. Under Sub-oval in plan. N break slope gradual, of others sharp. Sides steep vertical. to Base break slope sharp,of base flat. Slight NE-SW orientation. Filled C5. by L x W D Dimensions (m) 0.5 E-W x 0.45 x 0.07 x 0.45 E-W 0.5 x 0.07 x 0.45 E-W 0.5 x 0.2 max x 1.2 2.8 E-W x 0.26 2.85 N-S x 1.3 N-S x 0.27 x 0.14 0.33 0.5 E-W x 0.19 x 0.16 x 0.19 E-W 0.5 0.5 E-W x 0.19 x 0.16 x 0.19 E-W 0.5 3.84 NW-SE x 0.80 x 0.08 3.84 NW-SE x 0.80 x 0.58 x 0.11 0.87 NW-SE x 0.52 NW-SE x 0.64 x 0.42 3.75 0.24 NE-SW x 0.11 x 0.21 0.24 NE-SW x 0.11 x 0.21 - - Type Fill C10 of Shallow depres sion Fill C12 of Natural feature C24Fill pit of C21 Fill pit of Stakehole Natural feature. ral feature Fill of pit C2Fill pit of pit of Cut C2Fill pit of C2Fill pit of Fill stakehole of C6 Stakehole Fill C8, of natu Appendix 1: Context RegisterAppendix 1: Appendices Context 9 10 11 12 13 14 15 8 1 2 3 4 5 6 7 15 15.1

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 27 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples Soil 5 Soil 28 None None None None Slag 1 None None Soil 14 Soil 13 None None Finds None None None None None None None None None None None None None - - Description Dark greyish brown silty sand with charcoal Moderate fine 10%. sub-angular pebbles, oc casional medium and coarse subangular pebbles. under Over C81. and fill C15, of like material.Topsoil Filled natural C18. over feature. C81, Under Irregular C82. cut. over Interpreted as natural feature. C17, Under like material.Topsoil Filled natural C20. over feature. C81, Under Interpreted as natural feature. C82. over and filled Under C19, by rectangularSub break sharp, slope of in plan. sides near Top vertical, bottom break of slope gradual, base flat. Orientation and N-S. C82. filled over Under C14, by Dark brown clayey silt with fine slag. c. Frequent 10% sub-angular pebbles, occasional medium sub-angular C32. C30, over pebbles. Under C82. C30, over 32, C22 & C32. Under & slagPit Filled dump. C31, by Sub-rectangular charcoal break in plan,slope production of pit rounded corners. Top sharp, sides near vertical, bottom break slope gradual. of Orientation NW-SE. and Under filled C82. over C13, by Mid brownish black firm and charcoal sandy silt. 5%) Moderate slag flecks. (approx Ap pears be to a mixture topsoil of and dumped waste material slag i.e. and charcoal. Under C13. over C81, Mid yellowish brown friable silty sand. occasional Very charcoal flecks. over 81, Under and fill C27. of Natural feature formed vegetation by action. root irregular Very in shape and profile, no obvious C26, cut. C82. Under over Mid yellowish brown soft siltyOccasional clay. varied stones, very occasional charcoal. Interpreted as non-archaeological. and over fill C81, C49. Under of L x W D Dimensions (m) 0.33 E-W x 0.27 x 0.14 E-W 0.33 x 0.26 2.85 N-S x 1.3 x 0.45 x 0.85 E-W 0.9 x 0.55 x 1.35 E-W 1.57 2.8 x 0.2 max NW-SE x 1.2 max x 0.14 x 1.2 2.8 E-W 0.68 NE-SW x 0.14 x 0.32 0.68 NE-SW x 0.14 x 0.32 N-S x 0.26 x 0.06 0.31 - - - Type Fill of C15 stake Fill C15 of hole Fill natural of feature C18 Natural feature Fill natural of feature C20 Natural feature Charcoal pro duction pit C23Fill pit of Pit Charcoal pro duction pit. C24Fill pit of Fill natural of feature C27 Natural feature Fill natural of feature C49 Context 16 17 18 19 20 21 22 23 24 25 26 27 28

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 28 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples None Soil 7 Soil 6 Soil 8 Soil 15 Soil 16 Slag 3 31 Soil 18, Slag 4, 11 Soil 19 Charcoal 8 Soil 11 Finds None None None None None None None None None None None None None - - Description Dark greyish brown sandy silt with circa Under 20% slag. frequent charcoal Very (10%). & C66.C33 Over C68. Light brownish grey firm moderate sandy fine clay, pebbles. Moderate small pieces slag. of C22,Under C23. over Light greyish brown soft Deliberate sandy clay. backfill C30. over C81, Under pit. of Light grey occasional sandy clay, fine pebbles. Deliberate over backfill 31, Under pit. of C22. Primary fill C23. of Dark yellowish black sandy silt soft charcoal-rich C36, spread. C29. over Under Dark yellowish brown firm sandy silt with frequent slag, moderate charcoal. C37, by Cut only fill C36. of Mid greyish brown soft gravelly silt. Occasional slag. Only fill C37. of Sub-circular Sides midden concave. pit. Break slope base: of gradual. Filled C34. by disposal for pit Possible slag, of oval break irregular. but slope gradual, of Top except at NE imperceptible. Sides convex and smooth. Orientated N-S. Filled C35 by Mid blueish brown friable sandy silt. Moderate angular, sub-rounded fine pebbles, oc casional sub-rounded medium & coarse pebbles, occasional sub-rounded, rounded small stones, moderate flecks & small C39. charcoal. over C81, Under Brownish black charcoal-rich firm clayey silt. Moderate rounded fine pebbles, occasional sub-rounded medium pebbles, occasional sub-angular, sub-rounded coarse pebbles, oc casional flecks charcoal, moderate small charcoal, occasional flecks, medium, large, burnt moderate small C38, C40. over clay, Under burnt clay. Sub-rectangular break slope sharp of rounded corners. except Top at S & NE gradual. Sides N & S convex, W & E concave. Base break slope sharp of except at NE gradual. Filled C38, by C39. Dark brownish black soft sandy silt. Moderate sub-angular fine pebbles, moderate small & large slag, frequent medium slag, frequent flecks charcoal, moderate small charcoal. C45. over C55, Under L x W D Dimensions (m) 3.2 N-S x 1.3 x 0.1 3.2 N-S x 1.3 x 0.15 x 0.9 E-W 1.1 x 0.08 x 1.3 E-W 1.5 x 0.1 x 0.7 E-W 0.9 x 0.1 N-S x 0.4 1.0 x 0.17 x 1.1 E-W 1.5 x 0.1 0.8 N-S x 0.5 x 0.16 x 1.45 E-W 1.5 x 0.1 0.8 N-S x 0.5 x 0.32 N-S x 0.54 1.74 x 0.15 N-S x 0.49 1.6 x 0.32 N-S x 0.54 1.74 x 0.3 x 0.7 3.8 E-W - Type Layer C23Fill pit of C23Fill pit of C23Fill pit of Spread mid of den material C36Fill pit of Fill midden of cut C37 midden of Cut dump C36 disposal of Cut C37 pit Fill C40 of (C38) Fill C40 of (C39) Furnace (shaft smelting) Fill C45 of Context 29 30 31 32 33 34 35 36 37 38 39 40 44

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 29 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples Soil 20 Charcoal 5 Slag 8 Soil 21 Charcoal 7 Soil 10 Soil 12 Soil 24 Soil 25 Finds None None None None None None None None None None None None None None None - - Description erate medium pebbles, moderate angular, sub-rounded coarse pebbles, frequent sub-angu Linear pit rounded corners. Top breakLinear slope N sharp-gradual, of rounded corners. pit Top S gradual, N W gradual, sides N E irregular, S concave, W smooth, base break N S sharp, slope of E gradual, W Filled C56. C44 by imperceptible. C55 Orientated E-W. Black charcoal-rich C48 C47. over Under bottom layer. Furnace. C46, Under C82. over Mid yellowish brown friable clayey silt. Occasional rounded fine, medium & coarse pebbles, rare sub-rounded small stones, occasional small slag, occasional flecks charcoal. C46 topsoil, over Under irregularVery in plan and profile. Breaks slope and of sides vary from imperceptible to vertical. Interpreted as caused vegetation. by Filled C28. by Filled breakC51 by Circular slope sharp. of Orientated E-W. Top pit. Mid orange brown firm sandy sub-rounded silt.sub-angular Frequent fine pebbles, mod sub-roundedlar, small stones. Over C50. & slagPit Curving dump. linear rounded corners break top slope N S gradual, of E W imperceptible, sides N convex, S smooth, W irregular, E concave, base break slope N S of sharp, E W gradual. C54. Filled C53, by Mid greyish black firm sandy silt. Occasional sub-angular small stones, moderate small slag, frequent flecks charcoal. C52. over C54, Under Mid greyish brown firm sandy silt. Occasional sub-angular small stones, medium, large slag. Over C53 Mid greyish brown compact silty sand. sub-angular Frequent fine pebbles, moderate sub- angular medium pebbles, occasional flecks charcoal. C45. topsoil, over Under Red burnt clay lining possible bowl furnace. Mid over reddish orange C55 Under soft clay. C44. furnacePossible circular. Filled C59 C58, by Mid reddish brown firm silty sand. Occasional sub-roounded fine pebbles, occasional flecks charcoal, moderate small charcoal, occasional flecks slag, moderate small slag. C59 topsoil, over Under Dark brown/black C57. over sandy C58, silt. Under L x W D Dimensions (m) 3.8 E-W x 0.8 x 0.4 3.8 E-W x 0.5 0.54 N-S x 0.26 x 0.06 0.31 x 0.15 x 0.5 E-W 0.54 x 0.15 x 0.5 E-W 0.54 x 0.3 x 1.1 E-W 3.5 SE-NW x 0.2 x 0.7 3.12 2.08 SE-NW x 0.17 x 0.56 x 0.14 x 0.7 2.8 E-W x 0.2 x 0.1 0.27 SW-NE Type Cut C45 Cut Fill C47 of Furnace Fill C47 of C49 Cut Pit C50 Fill pit of Pit Fill C52 of Fill C52 of Fill C45 of Fill C45 of C57 Cut Fill C57 of Fill C57 of Context 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 30 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples Slag 15 Slag 10 Soil 22 Soil 23 Slag 13 Soil 30 Charcoal 9 Soil 26 Slag 12 Soil 9 Charcoal 6 Slag 9 Finds None None None None None None None None None None None None None None None - - - - - Description . Sub-circular possible furnace. break slope sharp, of sides N S Top Possible extractionPossible & slag pit Irregular dump. with pit rounded corners. Base break of slope gradual. Filled C67. by Dark orange sandy clay with mod. Slag?? Dark brownish black compact silty sand. Under C66 over C81, Mid-dark grey weakly cemented silty sand. Sub-rounded fine pebbles, sub-angular me dium & coarse pebbles, sub-rounded small & medium stones, moderate medium Slag. C36. over C29, Under & slagPit Linear dump. breal with slope sharp, of rounded corners. sides N & S ir Top regular, NW concave, base break slope N imperceptible of S sharp. Filled C71. by Mid brown soft sandy silt. Occasional aub-angular small stones, moderate flecks slag. C71. topsoil, over Under Black clayey silt with freq slag & charcoal? Dark brownish grey soft sandy silt. Occasional sub-angular small stones, moderate flecks slag & charcoal. under OverC70. C69, Collapsed material. Light pinkish orange compact Moderate sandy sub-rounded & clay. sub-angular fine pebbles, occasional sub-rounded & rounded medium pebbles. Over C73, under topsoil. Orange burnt clay furnace walls. Light reddish grey silty sand. Less than burnt clay 5% & charcoal. C72, C74. Under over Sub-circular construction cut. Filled C72, by C 73. Mid reddish brown firm clayey silt. Occasional rounded fine & coarse pebbles, occasional rounded & sub-rounded medium pebbles, occasional sub-rounded small stones, moder orangeYellowish burnt clay? Mid pinkish brown compact/firm siltyOccasional clay. rounded fine, medium & coarse pebbles, frequent flecks charcoal, moderate small char coal, occasional medium charcoal, occasional flecks & medium burnt clay & charcoal, moderate small burnt clay & charcoal. C64, C63. Under C75, over cave, base break slope gradual of imperceptible. to Filled C64, by and C75. C65 ate flecks occasional burnt clay, small moderate & medium flecks, burnt clay, small & medium slag & charcoal. C65. topsoil, over Under Some burningSome situ in stepped, E irregular, base break slope gradual. of C62. Orientated N-s. Filled C61, by Mid brownish orange firm & compact Moderate sandy sub-rounded fine clay. pebbles, occasional sub-rounded small C62. topsoil, over stones. Under Dark black soft sandy silt. Charcoal C60. over C61, rich. Under Sub-circular furnace Possible with rounded corners. Break sharp, slope top of sides con L x W D Dimensions (m) 2.5 N-S x 0.85 x 0.23 N-S x 0.85 2.5 x 0.23 N-S x 0.85 2.5 x 0.05 N-S x 0.45 1.4 SE-NW3.3 x 0.8 x 0.4 SE-NW2.9 x 0.15 x 0.48 x 0.4 SE-NW3.3 x 0.45 0.52 E-W x 0.4 x 0.22 x 0.4 E-W 0.52 x 0.15 x 0.4 E-W 0.55 0.8 N-S x 0.35 x 0.6 x 0.2 0.8 N-S x 0.56 N-S x 0.18 x 0.6 0.78 x 0.5 x 0.62 E-W 0.76 Type Cut Fill C66 of Fill C66 of Pit Fill C69 of Fill C69 of Fill C74 of Fill C74 of Furnace Fill C63 of Fill C63 of Furnace Fill C60 of Fill C60 of Cut Context 66 67 68 69 70 71 72 73 74 64 65 60 61 62 63

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 31 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 Samples Soil Slag 27 14 Soil 29 Finds None None None None None None N/A - Description Black charcoal Dark layer. bluish black friable silt.charcoal. 100% C63. over C65, Under Black charcoal dark layer, black, frequent flecks, small & medium charcoal. C77, Under C78 over Collapsed material. Light brownish grey weakly cemented silty sand/pebbles. Frequent sub-angular & sub-rounded fine & medium pebbles, occasional sub-rounded small stones, moderate flecks C76. topsoil, over silt. Under Sub-rectangular break in NE plan,slope of imperceptible, rounded corners. grad SE Top ual sharp, to SW & NW gradual, sides NE smooth, irregular, SE SW & NW concave, base break slope NE of & SW imperceptible, sharp, SE NW gradual. Orientated NE-SW. Filled C77. C76, by Oval/irregular. Break SW W SE slope top gradual, of NE NW sharp, sides NW SW irregular, NE concave, SE break slope base of NE SW gradual, imperceptible, SE NW sharp. Orientated NW-SE. C68 Filled C29, by Mid brownish green clayey sand found close worked stone C14 to Possible L x W D Dimensions (m) x 0.03 x 0.38 E-W 0.42 N-Sx 0.24 x 1.26 1.34 x 0.21 x 1.1 2.05 SW-NE x 0.67 NW-SE x 1.78 2.9 Extensive Extensive Type Fill C63 of Fill C78 of Fill C78 of Pit Pit Fill Topsoil Subsoil Find Context 75 76 77 78 79 80 81 82 Surface

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 32 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 44 57 60 69 78 45 74 29 68 79 Topsoil C.81 Natural C.82 32 36 67 72 23 33 66 73 24 22 34 52 40 47 Appendix Stratigraphic 2: Matrix 2 13 54 6 9 10 14 21 16 15 25 13 31 30 35 37 54 53 38 39 48 46 51 50 55 56 58 59 61 62 70 71 77 76 15.2

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 33 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.3 Appendix 3: Radiocarbon dates from Monganstown 1

Analysis by 14Chrono Centre, Queen’s University Belfast

Context Sample Charcoal Lab no. δ13C 14 C Date 2 Sigma calibration Period 14 2 Oak UB6937 -26.0 1050+/-29 cal AD Early Medieval 898-920 AND 946-1026 13 3 Oak UB6938 -27.0 1056+/-30 cal AD Early Medieval 897-921 AND 942-1024 1 4 Oak UB6939 -24.0 2110+/-31 cal BC Iron Age 338-330 AND 203-46 39 8 Diffuse- UB6940 -26.0 2168+/-31 cal BC Iron Age porous 361-270 AND 264-151 AND 138-113

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 34 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.4 Appendix 4: Assessment of Industrial Residues from Excavations at Monganstown 1

By Neil Fairburn October 2006

Summary The excavations at Monganstown, Co. Westmeath have produced a small, but exciting, collection of industrial residues. The slag that was recovered from the excavation has come from shaft furnaces and includes tap slags and fluid furnace slags. Radiocarbon dates that have been obtained from iron working features have provided an Early Iron Age date and demonstrates at long last that the technology of producing iron in a shaft furnace was in Ireland from the beginning of the Iron Age. As slag finds in Ireland have not been reported very often from a shaft furnace and are at the moment considered to be very rare, their presence at Monganstown in an archaeological context is extremely noteworthy. They should now be recognised as coming from an iron smelting site that was using a shaft furnace and not a bowl furnace. Slag that was recovered from pit feature C.2 and chemically analysed is likely to be from the remains of a slag pit shaft furnace. It is possible that the pit is in fact the slag pit base of the shaft furnace. Simi- larly, the slag recovered from the later furnace feature C.47, which has also been chemically analysed, is also likely to be from the remains of a slag pit shaft furnace. The presence of the slag pit furnaces may explain why furnace C.40 has cut into the earlier furnace C.74. Monganstown can now be added to the limited but growing evidence for the use of the shaft furnace in the iron smelting process in Ireland, and also now demonstrates at long last that the technology of producing iron in a shaft furnace was in Ireland. Monganstown is also the first reported use of the slag pit shaft furnace in Ireland. A type that was in regular use on the European continent, but of which there is a lack reported evidence of use in Ireland and Britain. In total 61kg of industrial residues were recovered from a number of pits and furnace bases during the archaeological excavation. The industrial residues consisted of material associated with iron working: Bloomery/Furnace Slag, 27.5kg of a fluid slag, an iron smithing PCB, amorphous slag and pieces of vitrified furnace lining. The furnace lining fragments were of size that it was possible to determine the internal diameter of one shaft furnace located on the site, which was 0.3m. It is suggested the recovered material is primarily the residues of iron smelting with a small amount of associated iron smithing. Chemical analysis undertaken by Dr Marcos Martinón-Torres from University College London confirmed that the slag from C.4 and C.47 had slowly cooled, indicated by the shape of the fayalite crystals and small quantities of wűstite dendrites, which are consistent with slag that solidified relatively slowly within or below the furnace. Thus suggesting that this slag is the by-product of a slag pit shaft furnace. The analysis of the slag has shown that the slag is largely fayalitic, with high manganese oxide concentrations and the near absence of free iron oxide or metal, altogether point to an efficient iron smelting technology, which is contrary to the use of the bowl furnace that has been demonstrated to be highly inefficient and thought to be in continuous use in the Westmeath area. The presence of the slags from shaft furnaces and a slag pit shaft furnace along with slags from other

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 35 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

contexts associated with furnace structures will now challenge Scott’s 1991 (p.159) statement of no shaft furnaces in Early Ireland into question and this will now need to be revised.

Introduction Excavations on the N6 at Monganstown was undertaken by John Lehane for Eachtra Archaeological Services Ltd and recovered a quantity material that was associated with metalworking and this was submitted for assessment and analysis. The majority of the material was recovered from features that have been identified as possible shaft furnaces as well as a possible base of a slag pit shaft furnace. Other material was recovered from a number of other contexts from pits within an area around these furnaces. It is not known if any structures were present surrounding any of the possible furnaces, but stakeholes recorded during the excavation may indicate that this was the case.

Iron Working Background When an archaeologist excavates a site that has the remains of iron production, the assemblage of finds will mainly consist of burnt ore, charcoal, slag and fired clay. The charcoal, slag and clay form integral and inseparable parts of the metallurgical process. This inevitably means that to understand the site; first it is necessary to identify and interpret the slag and burnt clay remains; and second one has to understand the basic technology of iron production that has produced the assemblage. The manufacture of an iron artefact from iron ore can be separated into three distinct processes. The smelting of the ore in a furnace, which will produce a bloom of iron as well as fayaltic slag residues; the primary smithing consolidation of the iron bloom into a billet; and finally secondary smithing - the shaping of the billet into an object. Each of these processes will produce a range of residues. Work by Brian Scott (1991) has indicated the range of material that might be found on early ironworking sites in Ireland, and recent work by Peter Crew and Thilo Rehren on material from the excavations at Ráith Na Ríg, Tara, Co. Meath has highlighted the types of industrial residues that are diagnostic of both iron smithing and non-ferrous metalworking (Crew and Rehren 2002). However, aside from this recently published work, very little analytical work has been carried out on early Irish ironworking sites. Edwards (1996) notes that in past excavations in Ireland the slag was often not recorded system- atically or rarely analysed to provide additional information about the activity that produced it. Experimental and analytical work on iron production and iron working residues in Britain, particularly work by Peter Crew, Snowdonia National Park, at Bryn y Castell, Crawcwellt and Llwyn Du, Wales, (Crew 1986, 1989, 1990, 1991, 1998, Crew and Crew 1995) and Gerry McDonnell, Bradford University (1988), along with work in Europe by Radomír Pleiner (Pleiner 2000), has clearly shown the nature of the archaeological evidence for iron production and for secondary smithing, and archaeologists can now identify the range of metalworking activity on sites in Ireland and Britain more confidently.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 36 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

The Ironworking Process The base material for making iron is ore. Any iron ore needs to have a sufficient concentration of iron minerals for conversion to metallic iron, but the reducibility and availability of ore will influence the willingness to use lower-grade ores. Therefore it is reasonable to accept that mineral composition, both of the iron mineral, the location and availability of ores had the most influence over the production of iron. The commonest iron ores in a geological sense are hematites, limonates and carbonates. The largest concentration of these iron ores in Ireland is in County Wicklow (Scott 1991, 153-154). However, one of the major sources of iron ore and the most likely to have been used for iron smelting in Ireland, was bog ore. A relatively easy ore to extract once it had been located. In theory there are plenty of sources of bog ore around Ireland, but this aspect has not been studied but is often cited. An analysis of the iron working slags recovered from the rath at Mullaghbane, Co Tyrone (Harper 1972) and the ringfort at Cush, Co Limerick (O’Riordain 1940) showed that bog ore was used on these sites for the manufacture of iron. Before it was smelted, it was necessary for the mineral ore to be broken into smaller fragments and roasted to drive off excess water and also to change the ore to an iron oxide. On most sites, the roast- ing of the ore would have taken place in a simple bonfire on to which was heaped the iron ore and a bonfire erected around it. This may have taken place close to the furnace, but equally may have taken place near to the ore source. Ore and fuel, usually charcoal, are fired in a furnace in a reducing atmosphere to remove unwanted material (the slag) and produce raw iron. The furnaces were either a shaft furnace or a bowl furnace. The air-supply for the burning fuel is provided by forcing the air through blowing holes in the sides of the furnace with use of bellows, or by natural draught as has been proposed for the Medieval furnaces at Stanley Grange, Derbyshire (Challis 2002). Shaft furnaces could have been up to 1.5m tall and some 0.30m in diameter and were built with clay. Considerable quantities of clay would have been used in the construction of furnaces and it could have been an important aspect in the location of a site. It is the large quantity of the vitrified remains of the furnace superstructure that helps to identify the remains of a shaft furnace. At the base of the majority of the shaft furnaces was a small arch that enabled the slag to be removed. The efficiency of the shaft furnace improved in Britain with the arrival of the Romans, who went on to industrialise iron producing areas like the Weald (Cleere and Crossley 1986), using efficient large furnaces similar to the one found at Laxton (Jackson and Tylecote 1988 and Crew 1998a). For example, the known Roman sites in the Weald produced tens of thousands cubic metres of slag and tonnes of iron (Cleere and Crossley 1986). The Post Roman period in Britain has produced no evidence to suggest great centres of smelting comparable with the Roman occupation (Cleere and Crossley 1986, 87). The evidence that can be gathered suggests that the Roman metallurgical techniques for smelting iron were not passed onto the native population of Britain (Tylecote 1986, 179). Instead the evidence suggests that the native population returned to a pre-Roman Iron Age tradition of producing iron probably with smaller shaft furnaces. The two possible explanations Tylecote (1986) give for this are that the old techniques were reintro-

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 37 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

duced by the migration of peoples from north east Europe and more likely- the economic conditions did not warrant the same large scale production as the Roman period did. The main notable development, however, in the Medieval period in terms of iron production, comes later with the introduction of waterpower in the fourteenth century (Cleere and Crossley 1986,106). The shaft furnace was loaded from the top with alternate layers of iron ore and charcoal. Molten slag was drawn off through an arched opening at the bottom and ran away in channels in which it solidified as a tapped slag. The arched opening allowed air to be drawn into the shaft but bellows could also be used. Shaft furnaces achieved a higher temperature, so increasing the carbon content of the iron and also could produce larger amounts of iron. In Europe, particularly in Poland and Sweden, the shaft furnace was often not tapped and a pit was dug beneath the shaft to collect the fluid slag as it became molten and dropped down the shaft. These furnaces are known as slag pit furnaces. The pit was usually lined with wood and brush to support the charge of the furnace, as the slag became molten, it flowed down and burnt away the material and dropped into the pit. As the slag flowed and dropped it created flow like slags, similar to the flowed tapped slags. When the pit beneath shaft had filled up the, the furnace was moved to a new location and a new pit dug and the process started again. During smelting the iron minerals are reduced and broken up by reaction with the burning charcoal. The unwanted minerals and elements, of which silicon is the most important, react with part of the iron oxides from the ore and with the clay and fuel ash form a liquid slag that falls to the bottom of the furnace. The clay used in construction of the furnace and possible air-pipes is subjected to the high furnace temperatures as well. If sufficiently heated it may melt, resulting in vitrified clay, or is even dissolved into the slag. The final product of smelting, besides the residues of ash and slag, is called a bloom. This is a rough, often spongy mass, containing metallic iron flakes and nodules that have sintered together, mixed with bits of slag, partially reduced ore, charcoal and parts of the furnace clay. The bloom gives the early production technology its name of “the bloomery process”. The bloom has then to be refined. This is done by hammering the bloom into a smaller piece, consoli- dating the iron particles and this is part of the primary smithing process. Primary smithing requires that the iron bloom is heated again until red-hot and soft, and subsequently hammered to squeeze out remaining slag and consolidate the fragment into a workable shape. This piece of iron, called a billet, can then be worked on again and made into an artefact. The slag that is squeezed out during the primary smithing of the bloom will end up in the bottom of the hearth, and it differs from smelting slag in that it is more or less magnetic and less fluid. These slags forming just above the bottom of the hearth are very characteristic and are often described as smithing hearth cakes or more frequently as Plano Convex Bottoms (PCB’s). These slags are sub-circular convex-convex shaped and usually magnetic. The smithing process hardly changed from the Prehistoric period through to the Medieval period, leaving similar residues. The smithing of the bloom can be done anywhere. Quite often this primary smithing was carried out on the smelting site. The bloom is heated in a hearth or forge. The hearth doesn’t need a purpose built structure but would require a shelter from the elements for the smith and also so as to provide low

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 38 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

light for the smith to be able to judge the temperature of iron. Early ironworking hearths were situated at ground level, while some Roman, and later, Medieval, hearths were positioned at waist height. The anvils, positioned close to the hearth, to strike the red hot bloom quite often utilised a large flat-topped stone or a large wooden block. The hammering of the metal bloom produces further waste products; slag spheres, solid balls or vesicular balls of slag that can fly for a considerable distance; hammer scale, small flat and thin pieces of magnetic metal. Hammer scale is usually a prime indicator of smithing and can be used to locate where the process was taking place. However hammer scale can also be produced during the consolidation of the bloom, If a large amount of smithing has taken place, the residues from this process can become trampled in to the floor around smithing area and form a cemented smithing pan. The pan is a conglomerate of highly magnetic material, dust, hammer scale, slag spheres and some other non-related material. Secondary smithing is the process that turns a refined billet into an artefact or implement and is carried out in the same way as the primary smithing and leaves the same sort of residues described above. This is the sort of small-scale smithing work is the sort of work that would be expected to be undertaken within the small nucleated farmsteads and enclosures. It has been suggested that many people were using basic iron technology to make and repair simple artefacts (secondary smithing), but the actual production of iron and manufacture of complex iron artefacts were still being produced by specialised smiths (Mytum 1992). Documentary evidence from the Irish Annals suggests that the blacksmith was held in high esteem and that the forge was a central part of the community (Scott 1987, Edwards 1996, 86).

Assessment All of the material was washed and cleaned by the author and visually examined by eye in 2005 (Fair- burn 2005d). The slags were weighed and were assessed on a typological appearance, which has been outlined by the work of Bachmann (1982), Crew (2002) and English Heritage (2001). Where surfaces had been fractured and a clean break could be observed, these were examined by using an x30 hand lense and then additionally examined under a microscope with x70 and x400 objective lense. The slags could not be cut open for examination, as there was not a licence to alter issued by the National Museum of Ireland. The material was considered to be very interesting, as it appeared to have come from a shaft furnace with an Early Iron Age date. Further analysis was needed to obtain an understanding of the composition and construction of the slags as very little analysis of these early slags has been undertaken within Ireland. It was also needed to provide a useful comparison with other material that has been excavated in the area and analysed. It is understood, that some analysis of excavated material recovered along the route of the nearby N4 at Kinnegad, has been undertaken, with results as yet un-published. The results of this analysis from Kinnegad suggested that the material was all the result of smelting in a bowl furnace (Photos-Jones 2003).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 39 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Results Quantification of the industrial residues from the excavations at Monganstown.

Context Sample Type Qty Total No Weight F47 21 Furnace slag 7 1960 F39 21 Furnace lining 5 705 F39 11 Tap slag 33 1808 F39 11 Slag unreduced 3 204 F39 11 Amorphous 17 1334 F39 4 Furnace slag 3 660 F39 4 Amorphous 9 424 F51 1 Furnace lining 1 108 F51 1 Tap slag 5 106 F51 1 Amorphous 13 112 F13 17 Furnace slag 1 505 F13 17 Amorphous 2 24 F13 17 Tap slag 1 20 F39 19 Furnace slag 2 130 F39 19 Amorphous 22 116 C71 23 Amorphous 25 176 F68 10 Furnace slag 3 565 F68 10 Amorphous 7 250 F68 10 Tap slag 3 140 C56 12 Amorphous 3 46 F38 18 Tap slag 16 394 C55 10 Tap slag 4 320 C55 10 Amorphous 16 70 F58 24 Amorphous 3 30 F42 7 Tap slag 13 154 F42 7 Amorphous 2 22 F1 1 Fluid slag 6 104 F1 1 Amorphous 21 302 F46 8 Furnace slag 3 1210 F46 8 Amorphous 12 174 F46 8 Fluid slag 8 418 F59 25 Amorphous 32 118 F59 25 Tap slag 4 46 C53 Furnace slag 2 60 C53 Amorphous 6 58 C53 Tap slag 2 148 F38 3 Furnace Lining 1 68 F38 3 Tap slag 13 1880 F61 12 Amorphous 4 106 F61 12 Tap slag 9 2402 F61 12 Furnace Lining 7 3136 F38 31 Furnace Lining 4 3355 F41 6 Furnace Lining 1 3500 F41 6 Furnace Lining 6 560 F41 6 Tap slag 10 414 Loose Tap slag 24 4705 Loose Ore 1 30 C44 11 Tap slag 20 1030 C44 11 Furnace Lining 1 100 C44 11 Furnace slag 1 1885

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 40 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Context Sample Type Qty Total No Weight F4 5 Amorphous 9 1175 F62 9 Tap slag 44 2352 F62 9 Slag cake pcb 1 464 Loose Tap slag 55 2950 Loose Furnace Lining 1 50 Loose Amorphous 2 72 Loose Amorphous 1 28 F65 15 Tap slag 2 116 F65 15 Furnace Lining 33 3290 F4 Furnace Lining 4 696 F4 Amorphous 3 154 A002 17 Furnace Lining 1 274 F75 14 Tap slag 58 1742 F75 14 Furnace Lining 1 102 F75 14 Amorphous 3 236 F22 1 Tap slag 18 2865 F22 1 Furnace Lining 4 118 F22 1 Amorphous 33 1330 F1 2 Fluid slag 35 3345 F1 2 Furnace Lining 2 585 F1 2 Amorphous 18 545 F1 2 Furnace slag 4 2650 Total 416 61331

Amorphous slag As with most assemblages there is a quantity of material that is difficult to classify and is termed ‘amorphous’. Amorphous slags do not have any distinguishing characteristics and are amorphous in shape and are often small. They could be from either the smelting or the smithing process. In this case all of the material was recovered from a furnace and therefore can be considered to be part of the smelting process. A small number of these slags had calcite in their matrix suggesting that a limestone flux or gypsum flux had either been used when the iron was smelted. This is similar to the material that was recovered from the excavations at Castledermot by Eachtra - Licence 04E0750 (Fairburn 2005c). The flux will make the slags more fluid and allow the separation of the metal from the waste to be more efficient, hence the low ratio of iron oxide (FeO) in the slags and a raised silica content (SiO2).

Fluid slag This slag was recovered from a furnace feature C.47 and pit C.2 and probably represents the last use of a slag pit shaft furnace. This type of slag is typical of a slag that would be found in the base of furnace that is using a pit beneath it to drain of this waste material. A number of the pieces of the Monganstown slag had wood impressions, suggesting that the slag had formed around pieces of wood. This would be similar to examples from Sweden and Iron Age examples from the Weald in Britain. It is thought that the wood was criss crossed at the base of the furnace, over a pit, and stopped the slag from forming a large lump in the furnace. The wood, while it was intact, provided a riddling affect and allowed the slag to run through. Some larger pieces of the slag show how the slag has run down the inside of the furnace and down

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 41 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

into the slag pit and then it has flowed flat, creating an L shaped pieces of slag, which are sometimes referred to as “elephants feet”. This clearly demonstrates that the slag was very viscous and flowed easily, perhaps facilitated by the criss crossed wood.

Tap slag Characteristic finds from iron smelting sites dating from late prehistoric times up until the 15th Cen- tury are pieces of slag with flat rounded or distorted bottoms and contorted upper surfaces with flow patterns. These are known as tap slags. The molten slag was drawn off through an arched opening at the bottom of a shaft furnace and ran away in channels in which it solidified. Tap slag has a characteristic ropey shape resembling a flow of lava. This type of slag represented the largest quantity of material recovered from the Monganstown excavations. This slag was recovered from the furnace features possibly represents the last use of a furnace as the material was not removed from it. The slag is normally removed and discarded and the furnace repaired ready for the next smelt.

Furnace Slag or Bloomery Slag The bloomery slag is usually represented by the largest amount of material recovered by weight, which is not surprising as these slags are usually large and heavy. The quantity of these slags produced from the furnace is dependant on the quality of the ore (Crew 1991). The bloomery slag from the furnace was typical of material described by Tylecote (1986) containing partially reduced ore and charcoal. Some of the slag exhibited could be considered as viscous with very small runs, suggesting that the slag may have been raked out during the tapping process.

Smithing Hearth PCB Smithing PCB’s or Plano Convex Bottoms (PCB’s) are diagnostic of the smithing process. The smithing process produces as well as hammer scale and small slag spheres, residues that consolidate in the bottom of the hearth as PCB’s as fayaltic slag lumps. These fayalitic slags are similar in composition to furnace slags but are distinguishable by their shape, a curved base and a level contorted surface. Their production is still poorly understood but the process that produces them is well documented. The Monganstown PCB is a small thin plate of slag and represents material that has accumulated in a very very short time period within the smithing hearth.

Vitrified Clay Lining This material consists of clay that has been vitrified on one side in the high temperature area of the furnace or the smithing hearth. Vitrified lining is produced by a high temperature reaction between the clay lining and the alkaline fuel ashes or slag. It can be difficult to identify if pieces of vitrified clay come from a furnace or a hearth structure. Smelt- ing sites usually produce significantly larger quantities than smithing sites, because of the difference in the size of the structures.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 42 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

The lining has probably been made from the local clay and has oxidised to an orangey red colour. Where one face of this lining has been exposed to high temperatures, it has started to vitrify to a slightly vesicular vitreous material with a black colour. Vitrified surfaces on furnace lining can vary in colour from black through to olive green, which in turn reflects the temperature conditions and the make up of the silica content in the clay. A number of the Monganstown furnace lining pieces were very large weighing over 3kg and are part of a shaft furnace superstructure. The curve of the shaft furnace can be seen in three of the pieces and the internal diameter of the shaft can be ascertained – 0.3m. One piece of furnace lining has clearing come from the base as it has a large amount slag adhering to it.

Hammer Scale Magnetic bluish-grey flakes of iron oxides formed on the surface of iron when heated for smithing. Primary and secondary smithing produces hammer scale when a hot iron object is struck. It is usually found in the area where the smithing was carried out. Hammer scale was noticed amongst the debris and dust accompanying the collection of slags.

Spheroidal Hammer Scale Spheroidal hammer scales consist of round droplets, which can be hollow to varying degrees and are mainly formed during the primary smithing of an iron bloom. The small droplets, some of which, but not all, are magnetic are expelled from the bloom when the hot bloom is struck during the consolidation process. They can also be formed by the welding process of two iron objects. The Spheroids recovered from the excavation vary in size from 2mm to 9mm and there are also some fragments of expunged slag, which have not formed spheroids. The location of the material would be indicative of the area in which the smithing of the bloom took place. Spheroiadal hammer scale was noticed amongst the debris and dust accompanying the collection of slags.

Analysis Three samples of the slag were sent to Dr Marcos Martinón-Torres from University College London for bulk chemical analysis. Two samples were taken from Context C.4 and one from C.46 The aim of this analysis was to obtain a chemical breakdown of the slag to understand the process by which it had been formed. It was also hoped to use the analysis to compare with the work undertaken on material recovered from Kinnegad in 2003. The detailed report of the analysis is presented as Appendix 7 (excavation report appendix). The analysis of the slag showed that the slags were fayalitic, had cooled slowly and had a high Iron Oxide (FeO) content with moderate amounts of manganese and alumina. This would suggest that slag was from a fairly efficient furnace, as the iron content was not high, and that a non-slag tapping shaft furnace must have been used to account for the slow cooling of these slags. The analysis also showed that the slags had high Strontium (Sr) and Bromine (Br) and this provides a useful geological indicator of the ore used, which was a source of bog ore. The manganese content has probably come from the ore, (bog ore) and has assisted the ore separation process.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 43 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Discussion The only certain indications for an iron smelting site are primarily the presence of ores and tap slags and also the presence of pieces of the furnace superstructure. Without either of the two reliable indicators, a site should be considered to be an iron smithing site. Slag is not datable in itself, but it is an important indicator of the site activities. At Monganstown only a relatively small amount of slag was recovered from the features on the site as well as the truncated remains of furnaces. The majority of the slag was a mixture of tap slag furnace slag and fluid slag and can only have come from slag tapping shaft furnaces and also from at least one non slag tapping shaft furnace with slag pit beneath it and not from bowl furnaces as is so often erro- neously described. A small number of significant parts of the furnace superstructure and one piece of ore were recovered at Monganstown, and indicates that iron smelting was taking place and that shaft furnaces were being used for this. We do know that at least a small amount of iron smithing has taken place as a small PCB was recovered from close to the furnace. The smithing was probably connected with the primary smithing of the bloom from the furnace. The small quantity of material found at Monganstown would not be indicative of large scale production associated with trading, as there are insufficient quantities of slag. However it is possible that further deposits of slag and iron working features lie in the area beyond the excavation. The slag was recovered from the now truncated bases of the furnaces and this may suggest that the furnaces were not used again, which is typical of a slag pit furnace, but not of slag tapping shaft furnace. The truncated base is usually all that remains of the shaft furnace and is in effect just a bowl or hole in the ground. The remains of this feature demonstrate how the erroneous bowl furnace myth still persists, as the feature with its distinctive slag potentially is not recognised for what it is or assessed. It is therefore perceived as a bowl in the ground with slag and therefore must be a bowl furnace. However, in this case, a large quantity of the furnace superstructure was also recovered along with a reasonable quantity of fluid tap slag from the features and this clearly shows that shaft furnaces were being used. Experimental work by Crew and others has shown how difficult it actually is, even today, to produce iron using the same techniques utilising our own modern knowledge and technology. In fact, so far, no experiment in iron production has actually reproduced material from the slags to a bloom that is comparable to any material recovered from a prehistoric iron smelting site. Peter Crew took 27 experimental smelts before he produced something that could be considered to be the bloom of iron. To get to that stage a lot of manpower, a lot of charcoal and a lot of ore was used. Therefore it can be deduced that if modern technology and knowledge cannot easily produce the iron it must have been very hard in the prehistoric period. Analysis of iron working slags recovered from excavations of a number of sites by ACS Ltd on the N4 at Kinnegad, showed that the ore was not being smelted properly and that these smelts would have probably produced poor quality, if any, iron at all. The excavations and the subsequent analysis of the slags from Kinnegad II reported that the iron smelting on these sites had been undertaken in bowl fur-

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 44 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

naces Photos-Jones 2003). These so-called bowl furnaces are now known from experimental work to be highly inefficient and produce very limited quantities of workable iron. Radiocarbon dates from the Kinnegad II site gave dates of around 810-420 Cal BC. However, two later dates were also obtained of 420-360 and 370-50 Cal BC (Photos-Jones 2003 and Murphy 2004), which are broadly similar in date to the Monganstown site. It should also be noted that Photos-Jones reported that slags from later Medieval sites from excavations on the N4 were also considered to have been smelted in bowl furnaces, and on this basis it was suggested that bowl furnace had not been replaced by shaft furnace by iron workers in Ireland, therefore it was still in use the Medieval period. This view is discussed further later in the report. The material from the Kinnegad II excavations supported by the radio carbon dates suggests that iron working, possibly in bowl furnaces, was being undertaken in Ireland during the transitional period between the Bronze Age and the Iron Age - the early beginnings of the Iron Age. This was also thought to be the case in Britain, but there are now no documented sites for the bowl furnace iron smelting work, as the two examples Kes Tor and Chelmes Combe, which are often quoted as early bowl furnace sites, have been recently re-examined and are not iron smelting sites nor are they Iron Age (Pers Comm Peter Crew). Scott quotes numerous examples in Ireland (1991, 159), but these should perhaps all now be reassessed in a similar way to the British examples. Scott in 1991 suggested that only the sites at Dressogagh, Co Amargh (Collins 1966) Ballyvourney and Garryduff, Co.Cork (O’Kelly 1952 and 1962), excavated prior to 1970 could be considered to be iron smelting sites. However, none of these early excavated site or sites excavated afterwards between 1970 and 1988 have been dated to the Iron Age and all were considered to be using the bowl furnace, hence Scott’s quote “All of the archaeological evidence to date indicates that during the period under review, iron was smelted in Ireland only in what has been called the bowl furnace.” It is not clear how the iron smelting technology evolved. As the technology of iron smelting was ad- vancing in the Bronze Age/Iron Age transitional period, it is possible that the use of the bowl hearth/ bowl furnace was employed, perhaps on the Kinnegad sites, but was found very early on to be inefficient (as we now know) and was gradually replaced with the shaft furnace. This advance can now, at long last, be clearly seen in Ireland within the excavation of the site at Monganstown and also at Lisnagar Demesne 1 (Fairburn 2006). Lisnagar Demesne 1, Co. Cork (licence 03E01510) was excavated by Deirdre Murphy of ACS Ltd during the construction of the Fermoy Bypass. Excavation uncovered a metalworking area with evidence for a slag tapping shaft furnace. A radiocarbon date of 390-50 Cal BC was obtained from suitable material that was mixed in with the slag, which had been recovered from the furnace feature, which also provides an Early Iron Age date for the use of the shaft furnace in Ireland (Ibid). To summarise at this point of the report, the material recovered from the furnace at Monganstown, indicates that iron smelting was taking place, using slag tapping shaft furnaces and non-slag tapping slag pit furnaces. In addition a small iron smithing PCB was also recovered and this would suggest that the iron bloom from the smelting process was probably consolidated on site.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 45 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Kinnegad II and Monganstown The N4 Kinnegad II excavations and the nearby N6 Monganstown excavations suggest a clear line of technological change. The reason for this is not clear and this is not reflected in the radiocarbon dates, as the activities at both sites are broadly similar in time. However the chemical analysis from both sites shows indicates two different activities. Monganstown is efficient with a low iron content within the slags and Kinnegad II with an inefficient high iron content within the slags. It has been suggested that Kinnegad II was using a bowl furnace, a known and highly inefficient process to produce iron while the excavation evidence and analysis shows that the efficient shaft furnace was being used a short distance away at Monganstown. The excavations on the N4 and now on the N6 also suggest that a large amount of iron working was being undertaken in the Irish Midlands of West Meath. This may be due to the presence of suitable bog ores in the area, perhaps from the Bog of Cullen. Analysis from Kinnegad II and now from Mon- ganstown supports the use of bog ore as the ore source for the smelting of the iron. The ore is rich in manganese and should therefore be relatively easy to smelt. The source of charcoal for the furnaces may additionally indicate that there was abundance of wood, shown to be oak and this may have been managed and coppiced. Oak has been shown from numerous excavations in Ireland and Britain to be the usual main choice of charcoal for smelting.

Monganstown and Ironworking in Ireland The ironworking debris material recovered from Monganstown was not large and it is impossible to say on how many occasions iron smelting has taken place. Experimental work has shown that the iron smelting process, to produce a bloom, followed by its smithing, would be expected to produce a lot more slag waste, with at least 7kg of slag waste per episode from a small shaft furnace (Crew 1991). As only 3.5kg of material was recovered from the excavation it probably is only representative of a small number of smelts. The low quantity of material from Monganstown could suggest that the other material is still wait- ing to be found outside of the excavation area, possibly in the area outside of the road take. However, the low quantity of slags and lack of other furnaces and hearths, might suggest that there is no more material to be found. The size of the small PCB recovered from a nearby pit is the only indication of primary smithing and is probably only from one episode of smithing work. This may suggest that the blooms were smithed elsewhere and support the view that other material is outside of the excavation area. However, early shaft furnaces and even later Roman furnaces from Britain are now known to have been inherently inefficient, as also was the material produced in bowl furnaces in Ireland. The slag from these early iron working sites still remained rich in iron and this was often plundered for re-smelting during the 18th and 19th centuries. It was also used in road building and sometimes in agriculture even for manuring to aid soil drainage. The amount of slag on a site can provide information about the metal working processes that took place. The 7th –8th century ringfort at Lisleagh, Co. Cork produced over 800 kg of slag (Scott 1991, 158). Later Roman and Medieval sites in Britain can have upwards from one tonne to hundreds of

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 46 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

tonnes of slag. Many of these later slag heaps were huge and may once have been larger still. Therefore as stated above, 61kg from Monganstown is not a large quantity of slag Smelting sites are usually located close to the resources that are required to produce the metal. Any ironworking site should be close to the ore source, wood for the charcoal and clay for the furnace lining, as transporting the raw materials from one place to another would have been a waste of time and effort. Peter Crew has made an approximate calculation that to produce 1kg of iron with a shaft furnace; ancient smelters would have needed some 15kg of ore, 100kg of charcoal and this would have taken 25 man days work (Crew 1991). In comparison, Tylecote’s experiments with a bowl furnace, undertaken in laboratory conditions, which bear no relation to actual field conditions, estimated that it would require 5-7kg or ore and 60kg of charcoal, but did not quantify the required man-days work (Tylecote 1986, 133). These experiments are discussed further later in the next section of the report.

Bowl Furnace or Shaft Furnace A large number of smelting furnaces have been identified in Ireland and Britain, mostly in Britain, and are of the shaft type with provision for slag tapping. In Ireland the only smelting furnace identified has until recently, according to excavation reports and the Excavation Bulletins, been the bowl furnace. As Scott wrote in 1991, “the principle difficulty in interpreting ironworking structures on Irish sites is the probability that hemispherical holes in the ground showing evidence of intense heat is as likely to have resulted from smithing as from smelting or even a combination of both”. Unfortunately this is still the case twelve years on, with very few ironworking sites in Ireland studied in any depth and disseminated through publication. Therefore, it is difficult to find comparable sites from published sources on work in Ireland with the exception of the influential work by Scott (1991). Ironworking residues seem to be dismissed by field archaeologists as assuming Scott’s conclusion ‘its iron working slag and must be from a bowl furnace’ as definitive. A bowl furnace has been described by Scott and Tylecote (Scott 1991 and Tylecote 1986) as a furnace consisting of an open or possibly covered bowl-shaped depression in the ground, which may or may not be lined with a clay ceramic or a layer of refractory stones. Bowl furnaces are currently, through the work of Scott, thought to be the only type of furnace that was used in Ireland from the Iron Age through to the Medieval period (Scott 1991, Pleiner 2000). There is some debate as to whether or not the bowl furnace could function properly to produce a large quantity of iron for a sizable iron bloom. The iron blooms, which have been found so far in Ireland, such as at Carrigmurrish Cave and Broth- ers’ Cave in Co. Waterford (Scott 1991, 162), are too big to have come from a bowl furnace, probably indicating that larger furnaces of the shaft type were used. Tylecote’s experiments have shown that bowl furnaces had similar charges of ore and charcoal like the shaft furnace, but these were placed vertically on the side of the furnace, opposite the tuyère-blowing hole. The slag then ran downwards and consolidated into a disc like lump of slag (furnace bottom) or even formed on the side of the hearth below the tuyère. The iron bloom then formed on top of the slag. The conglomerate of slag and iron was removed by lifting it from the top. Tylecote showed that it would then have been necessary to break this conglomerate up to separate the iron from the slag and

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 47 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

then it was reheated to form a small bloom. Characteristic finds from shaft-furnace sites are large pieces of furnace lining and pieces of slag that have solidified in the channels with rounded bottoms and a contorted upper surface with flow patterns. These are known as tap slags. It is the absence of these tap slags and large quantities of clay furnace structural evidence from sites in Ireland, which has led to Scott (1991) and Pleiner (2000) to suggest that the shaft furnace was not used in Ireland during the Irish Iron Age and the Medieval period. Scott suggests that the lack of innovative technology was blocked by socio-political developments and subsequently poor communication between craftsmen in Ireland and with Britain (1991, 213-214). Excavations on Irish sites have turned up a number of the so-called curved furnace bottoms, which are believed to have come from the bases of bowl furnaces. It is more likely that these furnace bottoms are in fact smithing hearth plano-convex bottoms (PCB’s) that the sites were actually iron smithing rather than smelting sites and the ‘bowl furnaces’ were in fact smithing hearths. This would at least explain the lack of tap slag and furnace lining material. However, the material from these sites needs to be re- examined with the benefit of the recent work by Crew and McDonnell before this can be confirmed. The bowl furnaces from Early and later Irish sites where smelting has been confirmed, are probably but not exclusively the remains of low shaft furnaces. This can now be demonstrated from the Mon- ganstown and Lisnagar Demesne 1 excavations. Tylecote’s experiments with bowl furnaces found that they were much more successful if the bowl was actually enclosed with a dome or a low shaft. This type of furnace would have allowed the bowl to have a large diameter and the shaft or dome would have produced the necessary reducing conditions to smelt a reasonable quantity of iron. The slag from this furnace is unlikely to have been tapped, and the bloom would have had to be brought out through the top of the structure or the structure was broken open to obtain the bloom. If the structure had been broken it might explain the lack of lining structure around the feature, as this material would have been more widely dispersed rather than concentrated in one area after the furnace had been aban- doned. However, the bowl furnace would not produce a tapped slag or a significant fluid slag. Bowl furnaces associated with iron smelting were identified by O’Kelly at the ringfort of Garryduff 1 and at Ballyvourney Co, Cork, The furnaces on these two sites were reported as small pits, lined with clay into which the charcoal and iron ore would be placed. The slag it was also reported was not tapped but had formed in the bottom of the furnace, while the iron had been left to form a bloom (O’Kelly 1952 and 1962). Evidence of iron smelting during the Early Medieval period is known from other sites besides Bally- vourney and Garryduff 1. These include Garranes, Co. Cork (O’Riordain and Hartnett 1943), Laith- more, Co.Tipperary (Tylecote 1986, 188), Lagore Crannog, (Hencken 1950) Clogher (Scott 1983), the ecclesiastical centre on Church Island, Co Kerry (O’Kelly 1958). None of the material from any of the sites above has been re-examined in the light of the recent work and new discoveries in the past 15 years. Examination of iron working material from recent excavations along the route of the M4 motorway between Kilcock, Co Meath and Kinnegad, Co. West Meath by ACS Ltd has suggested that the majority of the recovered material was from the smelting of iron in bowl furnaces, with the excavated sites and material dating between the Iron Age and the Medieval period (Photos Jones 2003).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 48 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

More recently, excavation work in County Kerry by Eachtra has uncovered definite evidence of the use of a shaft furnace for the production of iron in Ireland at Ballydowney and Farrananstack (Fairburn 2003a and Fairburn 2003b). Work by ACS Ltd at Cappakeel during work on the Monasterevin Bypass and at Newtown 1 on the Mullingar Bypass also has undated evidence for slag tapping shaft furnaces (Fairburn 2004 and 2005). Excavation work by ACS ltd at Corrin 1 and Lisnagar Demesne 1 on the N8 Fermoy Bypass (Fairburn 2005a and 2005b) has also recovered slag that can only have come from a slag tapping shaft furnace. The evidence from these sites and also now from Monganstown combined with the large blooms from Carrigmurrish Cave and Brothers’ Cave in Co. Waterford appears to con- tradict Scott’s 1991 statement of that there were no shaft furnaces in Early Ireland into question and therefore this statement will now need to be revised. It does seem bizarre that the innovative technology of the shaft furnace was being employed a short distance away across the Irish Sea in West Wales, and yet with clear evidence of trade and raiding between Ireland and Britain, particularly Wales, during the Iron Age through to the Early Medieval period, that this efficient method of iron production was not brought to the Irish Blacksmith and utilised. The later arrival of the Hiberno Norse with their metallurgical traditions should also have brought this technological advance on the bowl furnace to the Irish shores, likewise with the Anglo – Normans. Yet the myth persists that only the bowl furnace was employed in Ireland and this is seen in nearly all of the reports that appear in the Irish Excavation Bulletins. Radio carbon dates from excavations on the N4 Kinnegad and analysis of the recovered material has appeared to support the view of Scott (Photos Jones 2003). However, the evidence of the use of the shaft furnace from the analysis of material from other excavated sites and also now from Mongan- stown, combined with the obvious trade and technology links with Britain and the Hiberno Norse, shows that Scott’s statement can no longer be fully supported and needs to be revised. As more sites are excavated and more material analysed, this view should eventually change to one where it is accepted that the shaft furnace was used for the production of iron from the Iron Age through to the Medieval period. Monganstown joins the small but growing list of sites in Ireland where it can be shown that the shaft furnace was in use in Early Iron Age Ireland. Additional sites will eventually be recognised as more excavated industrial residues are studied and excavators become more familiar with the products of iron working technology.

Conclusion The excavations at Monganstown have produced a small amount of industrial residues from the iron working smelting and iron smithing process. The largest amount material recovered from the excavation was tap slag. This type of fluid slag can only have been produced in a shaft furnace and not a bowl furnace. This material has provided further evidence, to substantiate the use of the shaft furnace in the production of iron in Ireland in the Early Iron Age. The material from the excavated furnace has shown that early iron working in Ireland made use of the same technology to produce iron, namely the shaft furnace, to that used in Britain and Europe and that the country was not solely reliant on the inefficient bowl furnace as previously thought.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 49 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

It is clear that this site will eventually make a major contribution to our understanding of the cultural sequence of the metalworking in the region and in the country. The well stratified nature of the deposits and, indeed, the presence of directly datable residues with the ironworking slag will allow this assemblage to be used by researchers in the future in a much more meaningful way than has previously been possible.

Recommendations Further follow up work at Monganstown would be useful in attempting to pinpoint the extent of the metalworking site, if there is one, and to look for the characteristic dump of waste slag. This ideally should be done with a combination of non-invasive geophysics augmented with trial trenching. Work by Peter Crew et al has shown the benefits of using geophysics to pinpoint metalworking sites and also to obtain archaeomagnetic dates from them (Crew 2002 and Crew, Smekalova and Bevan 2002).

Acknowledgements Thanks to Peter Crew for discussions on Irish iron working and metalworking in general. Thanks to Dr Marcos Martinon-Torres from UCL for undertaking the chemical analysis. Thanks to all the staff at Eachtra Ltd for their help in dealing with my numerous questions, particularly Jacinta Kiely. A special thanks is also due to Jacinta Kiely and John Tierney at Eachtra Ltd for their patience and continued support for the assessment and analysis of metalworking residues. Slag really is interesting and it can tell you so much about a site.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 50 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Bibliography Bachmann, H.G 1982 The Identification of Slags from Archaeological Sites, UCL Institute of Archaeology Occasional Publication No.6, London.

Challis, K. 2002, A Medieval iron smelting site at Stanley Grange, Derbyshire, Historical Metallurgy Vol 36, pp.24-32.

Cleere, H. and Crossley, D. 1986, The Iron Industry of the Weald, Leicester University Press.

Collins, A.E.P. 1966, Excavations at Dressogagh Rath, County Amargh, Ulster Journal or Archaeology 29, pp.117-129.

Crew, P. 1986, Bryn y Castell Hillfort - a late prehistoric iron working settlement in North-West Wales. In: Scott, B.G. and Cleere, H. (Eds) The Crafts of the Ancient Blacksmith, Proceedings of the 1984 Symposium of the UISPP Comite pour la Siderurgie Ancienne, Belfast.

Crew, P. 1989, Excavations at Crawcwellt West, Merioneth, 1986-1989, Archaeology in Wales 29, 11- 16.

Crew, P. 1990, Excavations at Crawcwellt West, Merioneth, 1990, Archaeology in Wales 30, 46-47.

Crew, P. 1991, the experimental production of prehistoric bar iron, Historical Metallurgy 25.1, 21-36.

Crew, P. 1998a, Laxton Revisited: a first report on the 1998 excavations,Historical Metallurgy Vol 32.2, pp.49-53

Crew, P. 1998, Excavations at Crawcwellt West, Merioneth, 1990-1998: A late prehistoric upland ironworking settlement, Archaeology in Wales 38, 22-35.

Crew, P. 2002, Magnetic mapping and dating of prehistoric and Medieval iron-working sites in North- West Wales, Archaeological Prospection 9, pp.163-182.

Crew, P. and Crew, S. 1995, Medieval bloomeries in North-West Wales. In: Magnuson, G. (ed), The importance of ironmaking: technical innovation and social change, 43-50, Jernkontorets Berghistorika Utskott.

Crew, P. and Rehren, T. 2002, Appendix1: High Temperature Workshop Residues from Tara: Iron, Bronze and Glass, in Roche, H. (ed) Excavations at Raith Na Rig, Tara, Co. Meath 1997, Discovery Programme Reports: 6, Royal Irish Academy/Discovery Programme, Dublin

Crew, P., Smekalova, T. and Bevan, B. 2002, High resolution magnetic surveys of prehistoric and Medieval iron smelting furnaces in North-West Wales. In: Norbach, L. and Voss, O. (eds) Prehistoric and Medieval direct smelting in Scandanavia and Europe: aspects of technology and

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 51 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

society, 209-222, Aarhus University Press.7

Edwards, N. 1996, The Archaeology of Early Medieval Ireland, London.

English Heritage, 2001, Guidelines for Archaeo-Mettalurgy.

Fairburn, N. 2003a, Assessment of the Metalurgical Residues from the archaeological investigations at Ballydowney, County Kerry, Unpublished report for Eachtra Archaeological Services.

Fairburn, N. 2003b, The Evidence is Out There: Metallurgical Residues from Farrananstack, Co. Kerry, Unpublished report for Eachtra Archaeological Services.

Fairburn, N. 2004, Assessment of Industrial Residues from Excavations at Cappakeel 3, Unpublished report for Archaeological Consultancy Services Ltd.

Fairburn, N. 2005 Assessment of Industrial Residues from Excavations at Newtown 1, Unpublished report for Archaeological Consultancy Services Ltd.

Fairburn, N. 2005a Assessment of Industrial Residues from Excavations at Corrin 1, Unpublished report for Archaeological Consultancy Services Ltd.

Fairburn, N. 2005b Assessment of Industrial Residues from Excavations at Lisnagar Demesne 1, Unpublished report for Archaeological Consultancy Services Ltd.

Fairburn, N. 2005c Assessment of Industrial Residues from Excavations at Castle Dermot, Unpublished report for Eachtra Archaeological Services Ltd.

Fairburn, N. 2005d Assessment of Industrial Residues from Excavations at Monganstwon, Unpublished report for Eachtra Archaeological Services Ltd.

Fairburn, N. 2006, Analysis of Industrial Residue from Excavations at Lisnagar Demesne 1 Final Report, Unpublished report for Archaeological Consultancy Services Ltd.

Harper, A.E.T., 1972, The excavation of a Rath in Mullaghbane Townland, Co. Tyrone,Ulster Journal of Archaeology 35, pp.37-45.

Hencken, H.O’N. 1950, Lagore crannog: an Irish royal residence of the seventh to tenth centuries AD, Proceedings of the Royal Irish Academy 53C, pp.1-247.

Jackson, D.J. and Tylecote, R.F. 1988, two new Romano-British ironworking sites in Northamptonshire – A new type of furnace, Britannia 19, pp.275-298.

McDonnald, J.G. 1988, Ore to artefact – a study of early ironworking technology in Slater. In: E.A. and Tate, J.O. (eds) Science and Archaeology, 283-93. British Archaeological Reports, British

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 52 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Series 196.

Murphy, D. 2003 ‘Archaeological Excavation Report for site K-E-K M4 Motorway, Contract 1, Kinnegad 2, Co. Westmeath, Licence 02E0926’. Unpublished report by Archaeological Consultancy Services Ltd.

Mytum, H. 1992, The Origins of Early Christian Ireland, Routledge, London.

O’Kelly, M.J. 1952, St Gobnet’s house, Ballyvourney, Co. Cork, Journal of the Cork Historical and Archaeological Society 57, 18-40.

O’Kelly, M.J. 1958, Church Island, near Valencia, County Kerry, Proceedings of the Royal Irish Academy 59C, pp.57-136.

O’Kelly, M.J. 1962, Two ring forts at Garryduff, Co. Cork,Proceedings of the Royal Irish Academy 63C, 17-25.

O’Riordain, S.P. 1940, Excavations at Cush, Co. Limerick, Proceedings of the Royal Irish Academy 45C, pp.83-181.

O’Riordain, S.P. and Hartnett, P.J., 1943, the excavation of Ballycatteen Fort, Co. Cork, Proceedings of the Royal Irish Academy 49C, pp. 1-43.

Photos-Jones, E. 2003, “If it works, do not fix it”: Early (and later) Iron working in Ireland along the KEK-M4 Motorway, unpublished report for ACS Ltd, Drogeheda.

Pleiner, R. 2000, Iron in Archaeology: The European Bloomery Smelters, Prague.

Roche, H. 2002, Excavations at Raith Na Rig, Tara, Co. Meath 1997, Discovery Programme Reports: 6, Royal Irish Academy/Discovery Programme, Dublin.

Scott, B.G. 1987, The Status of the Blacksmith in Early Ireland, in Scott, B.G. and Cleere, H. (Eds) The Crafts of the Ancient Blacksmith, Proceedings of the 1984 Symposium of the UISPP Comite pour la Siderurgie Ancienne, Belfast.

Scott, B.G. 1991, Early Irish ironworking, Belfast.

Tylecote, R.F. 1986, The Prehistory of Metallurgy in the British Isles, The Institute of Metals, London.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 53 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.5 Appendix 5: Charred plant remains and charcoal from Monganstown 1

By Penny Johnston and Abigail Brewer

Introduction This report describes the plant remains retrieved from a site excavated at Monganstown (A001/001). The samples were taken from features such as charcoal production pits and metalworking pits associated with smelting and/or smithing. While charcoal was frequent in all the sampled deposits, other plant remains such as charred seeds were not common and only a small assemblage of remains was retrieved.

Methodology The samples were collected on site as bulk soil and were processed using a simple flotation method. Each sample was saturated in water to allow carbonised plant material to float; this “flot” (the floating material) was then poured into a stack of geological sieves and trapped in the sieve meshes (the smallest measured 250µm). When all the carbonised material was collected the flot was air-dried prior to storage in airtight plastic bags. The samples were scanned under low-powered magnification and the results are presented in Tables 1 at the end of this report. Sorting and identification of the flots was carried out using a low-powered binocular microscope (magnification x10 to x40) and identified seeds were separated and stored in sealed glass phials. Nomenclature and taxonomic order follow Stace (1997), although in order to facilitate easy reading of this text the scientific names are included only in the table of identified seeds presented at the end of this report (Table 2). Separate samples were taken for charcoal analysis; these were air-dried and each charcoal piece was snapped by hand to reveal the anatomical features necessary for identification, which was carried out using a binocular microscope (magnification range up to x50). The low-level of magnification available meant that while it was possible to separate different wood types from each other, it was not possible to identify non-oak species. The magnification range was, however, sufficient for the identification of oak, which made up the bulk of this assemblage. Identification was carried out with reference to the online wood identification key “Wood Anatomy”.

Results: Charred seeds In total, twenty-one samples from the site were scanned for charred plant material and environmental remains (results are in Table 1 at the end of this report). Only six samples contained the remains of charred seeds; C.1 (S.1), C.7 (S.4), C.33 (S.16), C.44 (S.11), C.56 (S.12) and C.71 (S.23). In each of these the plant remains were present in only very small amounts; four samples contained hazel nut shell fragments, one had the remains of an indeterminate fruit stone and an indeterminate seed was also found. One fragment of hazelnut shell, a fragment of straw/stem and two indeterminate, broken seed fragments were recorded in C.56 and C.44, fills of a charcoal making pit. Samples from C.33 a spread of

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 54 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

midden material, C.7 the fill of a natural feature and C.71 the fill of a pit also contained fragments of hazel nut shell. Waste hazel nut shells would occasionally have been burnt with fuel as a disposal mechanism (Monk 2000, 75) and their presence at this metalworking site is probably a result of this. Similar conclusions may perhaps be drawn to explain the presence of other small pieces of charred plant material such as the fruit stone fragment. Two indeterminate seeds were recovered from C.33 and C.44. The seeds could not be identified due to damage caused by heating. A fragment of straw or stem was found in a sample from C.44. Fragments of straw and other plant stems are often found together with other by-products of crop processing such as chaff and weed seeds in archaeobotanical assemblages. Metalworking sites excavated by Eachtra Archaeological Projects at Ballydowney, Co. Kerry (02E0055), Farranstack, Co. Kerry (03E0171) and Tracy’s Hill, Co. Waterford (00E0059) have also produced few macro-plant remains, despite extensive sampling. One reason for this is probably the high temperatures in the furnaces which would have burnt to ash, rather than charred, any plant remains that may have been present. The infrequent plant remains that occurred at these sites include cereal grains, chaff, weed seeds and hazelnut shells. These are likely to be the result of crop processing waste and hazelnut shells being used as tinder or fuel.

Results: Charcoal identification The charcoal from eight specially selected charcoal samples were examined, taken from charcoal production pits, furnaces and slag dumps. Two samples were taken from charcoal production pits; C.4 (Charcoal S#2) and C.13 (Charcoal S#3). The only fuel type identified from these pits was oak. Charcoal forms when it is burned in an inad- equate supply of oxygen, and oak produces good charcoal as it tends to burn more slowly than woods that are less dense such as, for example, willow. These factors ensure that oak charcoal is an ideal fuel for activities such as metalworking, where temperature control is necessary. Four samples of charcoal were taken from the lower fills of four furnace bowls; C.59 (Charcoal S#5), C.62 (Charcoal S#6), C.39 (Charcoal S#8) and C.64 (Charcoal S#9). The fuel types in three of these were also identified as oak, but in the fourth (from C.39) a diffuse porous type of wood was found. This could not be identified to species with low-powered magnification. The recovery of mostly oak charcoal fuel from the smelting/smithing pits complements the recovery of prepared oak charcoal found in the charcoal production pits nearby. Two samples were taken from slag dumps; C.4 (Charcoal S#1) and C.1 (Charcoal S#4). All of the charcoal from these samples was also identified as oak.

Conclusions The plant remains assemblages from this site were all small in quantity and poorly preserved. Hazel nut shells and fruit stone fragments from furnaces probably represent waste used as fuel or tinder. The results of charcoal identification suggest that oak was the main raw material for fuel, in the form of charcoal, at the site. Analysis of charcoal residues in slag from Johnstown 1, Co. Meath also identified oak as the predominant fuel type in use (Miller and Ramsay 2003). Some hazel was also found

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 55 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

at Johnstown 1 and it is possible that the diffuse–porous wood type identified at Monganstown was also hazel. Although this is an unusual type of wood to use as fuel in a smelting furnace it may have bee used as kindling or as wattle superstructure in the furnace; this coincides with the evidence from analysis of the Monganstown plant remains, which suggests that some domestic or agricultural residues (such as waste plant remains) may have been used as fuel, including hazel nut shell and fruit stone fragments as well as some straw. This material was probably used as tinder to get the fire in the furnace going initially before the larger charcoal fragments caught fire.

References Miller, J. and Ramsay, S. 2003 ‘Fuel analysis from metalworking slag from Johnstown 1’ in Photo-Jones, E. ‘Analysis of metallurgical waste from the M4 Kinnegad-Enfiled-Kilcock Motorway Scheme, the Johnstown 1 site’ Report by Scottish Services for Art and Archaeology for Archaeological Consultancy Services Ltd.

Monk, M. 2000 ‘Seeds and Soils of Discontent: an environmental archaeolgical contribution to the nature of the early Neolithic’, in: A. Desmond, G. Johnson, M. McCarthy, J. Sheehan and E. Shee Twohig (eds) New Agendas in Irish Prehistory, 67-87. Wordwell. Wicklow.

Stace, C. A. 1997 New Flora of the British Isles (2nd edition) Cambridge, Cambridge University Press.

“Wood Anatomy” accessed 25th January 2005 at http//:www.woodanatomy.ch

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 56 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Table 1: Scanning record from Monganstown, Co. Westmeath (A001/001)

Context no. Sample no. Seeds Charcoal Invertebrates Modern 1 1 Spore * **** Roots * 5 2 **** Roots *

7 4 Corylus avellana * **** * (modern) Roots * Spores* 9 31 *** Roots ** 13 17 **** Roots * 16 5 **** 17 28 Spores * **** Roots * 25 4 Roots **** 30 19 **** 33 16 Spores* and indet. seed **** 38 18 Corylus avellana * ** Roots ** 44 11 Spore* **** Fruit stone fragment* Stem/straw * 46 20 **** 53 ? Spore* **** Roots ** 55 10 ** Roots ** 56 12 Corylus avellana * **** Roots * 59 25 **** 62 9 **** Stem * 71 23 Corylus avellana * **** Roots * 75 27 **** 76 28 ****

Table 2: Identified remains

Cut 8 45 45 69 Context 1 7 33 44 56 71 Sample 1 4 16 11 12 23 Hazel nut shell frag- 2 1 1 1 ments (Corylus avellana) Indeterminate fruit 1 stone fragment (cf Prunus spp.) Stem/Straw fragment 1

Indeterminate seed 1 1

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 57 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Table 3: Identified Charcoal from Monganstown 1, Co. Westmeath (A001/001)

Context No Charcoal Feature Type Identifications Comment Sample No. 4 1 Slag dump in Oak (Quercus spp.) 2 fragments Very large pieces with growth extraction pit rings close together 14 2 Charcoal produc- Oak (Quercus spp.) 14 fragments tion pit 13 3 Charcoal produc- Oak (Quercus spp.) 6 fragments tion pit 1 4 Slag dump in Oak (Quercus spp.) 4 fragments extraction pit 59 5 Lower fill in fur- Oak (Quercus spp.) 10 fragments nace bowl 62 6 Lower fill in fur- Oak (Quercus spp.) 6 fragments nace bowl 39 8 Lower fill in fur- Diffuse porous wood charcoal; 11 nace bowl fragments examined 64 9 Charcoal layer at Oak (Quercus spp.) 6 fragments Bark frequent base of furnace Possible oak bowl (cf Quercus spp.) 2 fragments

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 58 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.6 Appendix 6: Analysis of charcoal assemblages from Monganstown 1

By Mary Dillon

Introduction This report details the results of charcoal analysis from archaeological excavations carried out at Mon- ganstown 1, approximately 2 km west of Kinnegad in Westmeath in the townland of Monganstown. The excavation was located 150 m north of, and in the flood plain of, the Kinnegad River, which forms the county boundary between Westmeath and Meath. It is on a slightly raised ridge in a level boggy field at the eastern end of the new section of road. Six furnaces and two charcoal production pits indicate that the site had an industrial/metallurgical role. Several other pits were excavated, and had been used for the disposal of slag, although they may have originally had a different function. Two of the furnaces were dated to the early to mid Iron Age, while the two charcoal production pits dated to between the late 9th to 11th century

Methodology Bulk soil samples were collected on site and were processed post-excavation using a simple flotation method. Each sample was saturated in water to allow the carbonised plant material to float, which was then poured off into a series of sieves (1 mm and 250 µm), trapping the ‘flot’ (floating material). This was air-dried and stored in air-tight plastic bags. The flots were sorted and scanned for plant material and charcoal using a low-powered binocular microscope (magnification x 10 to x 40). All charcoal fragments of 2 mm or greater were identified. Each fragment was prepared for microscopic examination by fracturing it by hand and thereby exposing a clean surface along transverse, radial and tangen- tial planes. All three planes were examined at a range of magnifications (x 5 to x 100) under a Nikon stereo microscope. For reference literature the website “wood anatomy” was consulted. The number and weight of fragments were recorded for each charcoal type. The following details were also noted: Ring curvature. The analysis of ring curvature can be regarded as broadly indicative of age, and this was noted during charcoal analysis. On the basis of the degree of ring curvature the charcoal fragments were classified according to age categories as follows: strong curvature, <5 y (years); medium curvature, 6-10 y; weak curvature, 11-15 y; curvature negligible, 15+ y, i.e. mature wood. Where the pith was present, the age of the fragment was noted (equal to the number of rings present). Insect and fungi infestation. Holes, round and larger than vessels, are usually regarded as caused by burrowing insects; their frequency was noted.

Results In all, 437 charcoal fragments were identified in the thirteen samples. In Figs. 1 and 2 percentage frequencies of the various charcoal types based on fragment count and dry weight, respectively, are shown. Bar charts (Figs. 3-5) illustrate the frequencies of the various wood

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 59 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

types in different feature types.

The most frequent charcoal type overall is oak at 88.6%, or 85.3.5% by weight (Fig. 2). This is followed in descending frequency by ash (8%), willow/popular (1.4%), yew (0.9%), elm (0.5%), hazel (0.5%) and hazel/alder (0.2%). When considered on a weight basis, the overall picture remains broadly similar apart from yew, which makes up 7% of the percentage weight (the fragments were very large).

As regards the individual features, there is some variation between feature and feature type (see bar charts Figs. 3-5). C1 from Pit 2, which contained slag waste, contained a majority of ash. Similarly C62 from Furnace 60 contained only ash. This contrasts with the other features which were dominated by oak. C46 from Furnace 47 is distinctive in that, in addition to oak, it contained large branch pieces of yew wood (see below).

There are two time periods represented on the site - the early to mid Iron Age (two of the furnaces) and the later Early Historic Period (the two charcoal making pits). There are no notable differences in the charcoal assemblages from these periods. Only one structural feature -a posthole- was found. Like the other features it contained a majority of oak wood. Given that this sample was purely composed of oak, it may be the remains of a burnt-out post. This feature is not represented in bar charts 3-5.

Willow /Popular Hazel/Alder 1.4% 0.2%

Hazel 0.5% Yew Elm 0.9% 0.5%

Ash 8.0%

Oak 88.6%

Fig. 1 Pie chart showing percentage frequencies of wood types

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 60 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Hazel/Alder 0.004% Willow/Popular 0.1% Yew 7.0% Hazel 0.02%

Elm 0.2%

Ash 7.3%

Oak 85.3%

Fig. 2 Pie chart showing percentage weight of wood types

120

100

0 Oak Ash Willow /Popular

Fig. 3 Bar chart showing wood type counts from charcoal pits

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 61 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

140

120

100

0 Oak Ash Elm Hazel / Alder

Fig. 4 Bar chart showing wood type counts from waste pits

120

100

0 Oak Ash Hazel Yew

Fig. 5 Bar chart showing wood type counts from furnaces

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 62 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Ages and forms of wood present While ring curvature does not give a strict qualitative indicator of age, the results nevertheless are useful in that they give insights into the age of the wood that was burned. Mature wood (c. >15 y) is presumed to be derived mainly from large branches and trunks, while young wood (<15 y) is presumed to derive from young branches and twigs. The vast majority of the fragments analysed derive from mature wood, i.e. the rings have low curvature and furthermore they are, by and large, narrow (<2 mm) which is typical of mature wood. The main exceptions are as follows: (i) The ash fragments from C14 (Charcoal pit 2) were whole branch pieces, c. 10 y old. (ii) The yew fragments from C46 (Furnace 47) were also whole branch pieces c. 15 y old. Most of the wood contained no insect burrowing holes, a feature that is common in dead wood. As the wood would have to have been seasoned prior to charcoal making, it must have been stored care- fully to avoid burrowing insects. The exception to this is the yew wood from C46 (Furnace 47). This indicates that this wood was possibly gathered off the forest floor, where it was subject to insect attack. It may also indicate that it was structural wood, (perhaps a pole given its size) and was burrowed by insects at this stage.

Discussion Comments on charcoal taxa Quercus (oak) The charcoal data as a whole show that oak was the main tree burnt on the site (Figs. 1 - 5). It accounts for 88.6% of all charcoal fragments identified. The charcoal from Monganstown 1 likely represents the remains of charcoal burning rather than wood burning (given that the samples originate from charcoal making pits, furnaces for smelting and pits used to dump furnace waste). High temperatures of between 1100-1250˚ are needed to smelt iron (Waddell 2000, 286). Oak is slow burning and gives out substantial heat as it burns which would have made it a natural choice. Historically, oak was one of the preferred woods for making charcoal along with alder and strawberry tree (Arbutus unedo) (Feehan 2003, 299-355). There are two native species of oak in Ireland, namely the Sesile oak Q.( petraea) and the pedunculate oak (Q. robur). The former tolerates poor, acidic edaphic conditions (e.g. it is the main canopy tree of Atlantic oak woodlands, e.g. Killarney National Park; Kelly 1981) and the latter is characteristic of the more fertile parts of Ireland such as the site area. Unfortunately, it is difficult to distinguish these species on the basis of wood anatomy (Grosser 1977). On the basis of present-day distribution pat- tern and ecological preferences, it is probable that the charcoal derives mainly, if not exclusively, the pedunculate oak The ring patterns of the charcoal indicated that oak was from mature wood i.e. trunks and large branches were utilized. The narrow ring widths of the oak wood suggest the trees grew in dense woodland. In the Iron Age the main canopy trees were oak, ash and elm (see Molloys pollen diagram in Dillon, forthcoming). Oak would have been deliberately selected to fuel the iron smelting process. As the Early Historic period progressed, oak would have become less widespread but still would have been common. It is likely that the centres for smelting were situated in areas where mature oak wood was abundant. In medieval England we know that oak was coppiced to provide a self-renewing supply of

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 63 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

timber, without the need to clearfell woodlands. There is no evidence of coppiced oak being used at Monganstown, possibly because there was a plentiful supply of oak woodland.

Fraxinus (ash; F. excelsior). Ash makes up 8% of the assemblage. It makes great fuel, burned green or dead, and this may have influenced its selection. Ash is only prominent in two samples, and one sample consisted of branch wood c. 8cm in diameter. This may have been structural rather than fuel wood.

The hazel Corylus( Avellana), willow/popular (Salix/Populus) and hazel/alder (Corylus /Alnus) may have been used as kindling wood, as it is much less dense than oak and (apart from alder) does not make good charcoal. It may also have been use to provide a wattle frame for the furnaces.

Land-use history based on pollen analytical data As part of an integrated study into the history of land use in the Blackwater Valley, Co. Meath, a core was taken 7 km to the north-east of Kells, about 20 miles north-east of Monganstown 1 (Dillon et. al. forthcoming). This pollen core shows that trees decline and farming indicators rise in the period 300 BC-AD 20. Tall canopy trees, especially elm and ash, were cleared but most striking is the clearance of hazel which emphasises the intensity of farming during this period. In the period AD 20-500 oak, ash, elm and hazel increase alongside the decline in pastoral and arable indicators. Oak is well represented in the Iron Age pollen profile but would have shared the canopy with elm and ash trees. Hazel woods would have been plentiful. The unbroken curve forTaxus suggests that yew was present but was not frequent (Taxus is generally regarded as over-represented in pollen records). There is an intensification of farming activity, at the expense of woodland cover, during the period from the 9th to 11th centuries.

Comparative studies The remains of iron smelting (slag and charcoal) were found in a kiln the Early Historic period site at Randalstown, near Navan in Co. Meath. The charcoal analysis results from Randalstown were very similar to those from Monganstown 1. The charcoal from the Randalstown contexts was exclusively oak, and was very clean, suggesting that it was the remains of charcoal used to fuel the smelting process (Dillon 2006). At Johnstown 1, a nearby site in Co. Meath where evidence of metalworking was found, oak and hazel charcoal were recovered, suggesting that these two wood types were predominantly used to fuel the metallurgical activities. It is concluded that the hazel wood was probably not used in the smelting process itself, but in some activity relating to it (Miller and Ramsey, 2003). The two comparative studies above had similar findings to those from Monganstown 1 i.e. oak was predominately used to make charcoal for smelting iron.

Summary The charcoal assemblages from Monganstown 1 were dominated by oak wood, which was collected as charcoal-making fuel for iron smelting. Oak is well documented as a reliable wood for charcoal mak-

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 64 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

ing and was traditionally used as such. It would have been deliberately selected from woodlands, which we know from a semi-local pollen profile, were not dominated by oak, but by a combination of oak, ash, elm and hazel. The curvature of the rings and the rings widths of the oak charcoal indicate that mature oak was sourced from closed woodlands to fuel the iron works at Monganstown.

Acknowledgments Pollen analysis at Emlagh Bog was carried out by Dr. Karen Molloy of the Paleo-environmental Unit, NUI, Galway, published in Dillon et. al. forthcoming

References Dillon M. 2006. People and the Environment. Towards an Anthropology of Woodlands in Prehistoric and Early Historic Ireland. Unpublished thesis, NUI, Galway.

Dillon M., Newman C., Molloy K, M. and O’Connell M. (forthcoming).

Environment and ritual in a late Iron Age context: an example from Raffin Fort, Co. Meath, Ireland in BAR I.S. Charcoals From the past. Cultural and Palaeoenvironmental implications.

Feehan J. 2003. Farming in Ireland. History, Heritage and Environment. UCD, Dublin.

Grosser, D., 1977. Die Holzer Mitteleuropas. Berlin: Springer-Verlag.

Kelly, D.L., 1981. The native forest vegetation of Killarney, south-west Ireland: an ecological account. Journal of Ecology, 69, 437-472.

Miller J. Ramsay S. 2003 Fuel Analysis from Metalworking Slag from Johnstown I. In: Photos-Jones E. Analysis Metallurgical Waste from the M4. Kinnegad-Enfield-Kilcock. Motorway Scheme, the Johnstown I site. Report produced by Scottish Analytical Services for Art & Archaeology for ACS, Drogheda

Waddell J. 1998. The Prehistoric Archaeology of Ireland. Galway University Press, Galway

“Wood Anatomy” at http//:www.woodanatomy.ch

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 65 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

61.5 S.5 C.69 C.69

55.5 S.9 C.63 C.64

17.3

? C.60 C.62

73.5 S.5 C.57 C.59 3

37.7

27 S.20 C.47 C.46 1

0.02

0.06

S.18 C.40 C.38

1.9 S.16

C.33

27.2 S.5 C.15 C.15 C.16

2 50

15.5 141 S.2 C.21 C.14

10.3 S.17 C.24 C.13

0.8

30 S.3 C.24 C.13

0.1

25.8 S.1 C.2 C.4

2 12 10

1.2 6.8 7 S.4 C.2 C.1 - - ) ) Popu / Alnus Alnus ) ) ) ) Salix/Popu Salix ) ) ) a) ) / Alder ( ) / Alder ( spp.) Corylus Corylus avellena Corylus Corylus avellena Quercus spp.) Quercus Ulmus glabr Ulmus Taxus baccata Taxus Ulmus glabra Ulmus Taxus baccata Taxus Fraxinus excelsior Fraxinus excelsior ) ) Yew ( Yew Willow/Popular Willow/Popular ( lus Hazel ( Hazel ( Elm ( Ash ( Oak ( Fragment count Yew ( Yew lus Willow/Popular Willow/Popular ( Hazel ( Hazel ( Elm ( Ash ( Weight (g)Weight Oak ( Sample Cut Deposit Table of raw data from charcoal data raw from of identification Table

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 66 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.7 Appendix 7: Chemical analysis of industrial residues

Analytical report on iron slag samples from Monganstown, Ireland Marcos Martinón-Torres Institute of Archaeology, UCL. July 2006

Analytical method

Bulk chemical compositions were obtained by polarising energy dispersive X-ray fluorescence analysis (ED-XRF) of pressed powder pellets. The analysis was carried out with a Spectro X-Lab Pro 2000 instrument, using an evaluation method with standards optimised for iron-rich materials. In order to assess precision and accuracy, each specimen was analysed three times, and alongside standard reference materials. All the resulting data are given in Table 1.

Pellets were prepared by milling 10 g from the core of the slag in a tungsten carbide swing mill with counterweight, down to a particle size of approximately 50 μm. This powder was dried and thoroughly mixed with a small fraction of industrial wax, and subsequently subjected to 15 T of pressure in a 32 mm diameter mould.

A cross-section of approximately 1 cm3 section of the same slag lump was mounted in epoxy resin and polished down to a grain size of 0.25 μm. These sections were examined under plane polarised (PPL) and cross-polarised (XPL) reflected light in a Leica DM LM metallographic microscope.

Analysis and interpretation

Three slag samples were received, sized approximately 8 to 10 cm in the longest axis. Macroscopically, they show a dark orange to brown colour, indicative of some external rust.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 67 1 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Sample F4/F2 A has the typical bulbous appearance of bloomery iron smelting furnace slag. On the other hand, both F4/F2 B and F46/F47 have plano- to concavo-convex profiles, which at first could be suggestive of smithing slags. The analytical study, however, confirms all three as derived from bloomery iron smelting, without any indication of tapping or smithing.

The chemical composition of the three samples falls within a consistent compositional range, although with some variation in the silica to iron oxide ratios. Most remarkable are the high manganese oxide concentrations, around 8 wt%, which are accounted for by the generally low iron oxide levels (43-60 wt%). The use of manganese-rich iron ores can facilitate the production of slag – and, consequently, of iron – as manganese replaces iron oxide in the slag, resulting in slag that is poorer in iron, i.e. a more efficient smelting process. Other noteworthy aspects in the chemical data are the relatively high levels of strontium and barium. Considering the low contribution of ceramic materials to the slag formation – as reflected in the low alumina levels – it can be assumed that Sr and Br constitute a geological signature of the ore used. This, in addition to the high Mn, could be used as indicators to try to identify potential sources of iron ores in the vicinity.

In cross section, all three slag samples show a layered structure and significant accretions of clay and small stones on one side at least. This could indicate that all of them solidified whilst in direct contact with either the surrounding soil – if the slag was tapped –, a smithing hearth bottom, or the bottom/walls of a bloomery furnace. On the basis of the microstructural examination, however, the first two hypotheses can be rejected.

Sample F4/F2 A shows a very homogeneous structure. This consists of relatively large, mostly euhedral, fayalitic blocks, some with skeletal development but none showing the elongated shapes of fast-cooled systems. The interstitial glassy matrix shows some acicular crystals, possibly enriched in calcium oxide (bulk CaO is >3 wt%), whose development is also indicative of a slow cooling rate. Small wüstite dendrites are only found in very localised areas. This is consistent with slag that solidified slowly inside a furnace, and where little free iron oxide is present. In fact, this is the sample showing the lowest bulk iron oxide concentration. Only towards the area in direct contact with clay or soil can some heterogeneity be noticed, namely the presence of some leucitic phases and small crystals of magnetite. Both of these may be explained as a result of the localised chemical environment, which would be richer in oxygen and clay oxides such as alumina and potash.

Sample F4/F2 B is richer in iron oxide (although still within relatively low ranges), and this is reflected in the higher abundance of dendritic formations of wüstite noticed under the microscope. The specimen mounted for cross section shows the interface between two slag

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 68 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

layers. Whilst both of them show developed fayalitic structures, the main difference between the two is the different abundance of wüstite dendrites. This is probably a reflection of two subsequent episodes of slag deposition within the furnace during the same smelt, one of them being slightly richer in iron oxide. The absence of external flow textures or an oxidation skin of magnetite between the two layers – wich would form upon solidification in air –, and the fact that the interface between the two is not clear-cut – some crystals clearly develop across the two layers – allow to rule out the possibility of this constituting tap slag. Furthermore, the internal homogeneity and the lack of sand or charcoal inclusions within the slag are against the interpretation of this as smithing debris, in spite of the plano-convex profile. In fact, the clay- and stone-rich layer is addhering to the flat side of the slag lump, which probably constitutes the bottom of a non-tapping bloomery furnace.

Finally, sample F46/F47 is also comparable to the above samples, with a developed fayalitic structure and little free iron oxide present. Again, in spite of the layering, this slag clearly cooled slowly and inside a smelting furnace. The only noteworthy feature is the presence of a few small sulphide inclusions within a devitrified matrix. As in F4/F2 A, an area richer in magnetite can be noticed near the furnace wall area.

In sum, the samples analysed are indicative of bloomery iron smelting, most likely in a non- tapping furnace. The ores smelted where rich in manganese and, to a lesser extent, strontium and barium, which is reflected in the composition of the slag. The slag samples contain relatively low iron oxide concentrations, therefore it can be inferred that the technology was efficient for a bloomery furnace.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 69 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Figs. 1 (left) and 2 (right). Monganstown F4/F2 A. Left, microstructure of the slag, showing quite developed, euhedral fayalite blocks (light grey) in a glassy matrix (PPL, x50, width 2 mm). Right, view of the same area under cross-polarised light, where the internal reflections reveal the bright acicular crystals within the glassy matrix.

Figs. 3 (left) and 4 (right). Monganstown F4/F2 A. Left, area of the slag showing the development of small wüstite dendrites in the interstitial glass within a fayalite-dominated system (PPL, x50, width 2 mm). Right, detail of the same area under higher magnification (PPL, x200, width 0.5 mm).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 70 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

Figs. 5 (left) and 6 (right). Monganstown F4/F2 B. Left, interface between two layers of slag deposited within the furnace. Note that the bottom layer is richer in free iron oxide, with more abundant wüstite (white), but also that some fayalite crystals (light grey) cut across the two layers (PPL, x50, width 2 mm). Right, detail of the bottom layer, showing eutectic intergrowths of fayalite and wüstite (PPL, x100, width 1 mm).

Figs. 7 (left) and 8 (right). Monganstown F46/F47. Left, microstructure of the slag, showing developed euhedral fayalite crystals resulting from slow cooling (light grey), and bright magnetite in some areas of the glassy matrix (PPL, x100, width 1 mm). Right, detail of the microstructure showing some bright sulphide inclusions within a devitrified matrix (PPL, x200, width 0.5 mm).

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 71 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237 % Sum 67.14 76.20 68.90 80.40 98.87 99.92 69.39 80.52 67.68 80.48 100.82 100.82 100.07 100.07 100.81 100.81 100.29 100.29 100.28 100.28 100.03 100.62 100.62 53 58 49 Ba 136 145 142 6 ppm 3271 3271 7457 7457 3328 3328 3299 3299 6087 6087 7450 7450 6020 6020 7410 6108 6108 95 84 96 86 77 77 61 83 108 287 299 109 299 109 111 ppm ZrO2 Y 56 18 25 56 18 16 55 19 22 18 25 20 108 105 105 ppm cted elements cted aterials, shown 298 235 703 301 237 238 292 379 708 379 705 387 SrO ppm 1191 1191 1193 1193 1183 1183 20 42 19 21 39 34 32 42 43 Se bdl bdl bdl bdl bdl bdl ppm 49 71 47 48 39 65 36 66 41 408 391 177 397 177 177 ZnO ppm

9 7 6 9 9 7 10 12 15 16 14 10 10 12 bdl ppm CuO 93 71 48 50 70 94 13 77 64 41 19 36 73 28 26 204 NiO ppm sults for standard reference m % FeO 27.55 38.44 43.05 43.00 43.32 50.02 32.38 43.09 59.35 32.57 44.15 43.83 59.01 50.00 31.84 59.15 50.06 % 1.25 0.28 0.28 8.00 8.19 1.14 0.28 0.28 8.30 1.13 8.12 8.14 8.31 8.20 1.12 8.18 8.17 MnO 3

% 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.06 0.07 0.07 0.07 Cr2O 01 17 63 65 98 % 3.92 3.74 3.75 2.85 3.76 3.74 3.68 3. 2.83 3.99 3. 2.83 3.99 CaO 22.60 21. 21. 20. n as obtained by ED-XRF in three different runs, with all dete % ified values and the analytical re 0.32 0.59 0.55 0.55 0.19 0.34 0.58 0.55 0.21 0.20 0.34 0.58 0.34 0.57 0.33 0.34 0.32 K2O % 1.00 0.30 0.31 0.32 0.29 0.28 0.31 0.33 0.34 0.29 0.28 0.29 0.29 0.26 0.56 0.58 0.57 SO3 % 0.80 0.30 1.01 1.09 0.69 0.44 0.29 1.09 0.99 0.71 2.02 1.11 1.03 0.69 2.06 2.07 2.05 P2O5 40 % 6.97 7.02 6.85 7. SiO2 40.06 31.66 25.65 17.81 40.79 40.59 25.44 31.59 25.64 31.83 16.26 16.39 16.36 29 36 84 22 72 % 1.86 1.93 1.87 1.84 1.92 2. 1. 2.69 2.68 4. 4. 4. 4.26 10.62 12.64 12.61 12.54 Al2O3 48 % 1.28 1.26 1.22 0.18 0.25 0.40 0.42 0.17 0.22 1.50 0.43 0.17 0.20 1.49 1.43 1. 1.73 MgO % bdl bdl bdl 0.47 0.72 0.69 0.18 0.27 0.42 0.39 0.41 0.25 0.23 0.37 0.25 0.09 0.07 Na2O certified F4F2B certified F4F2B analysed analysed F4F2 F4F2 F4F2 F4F2 F4F2 F4F2 F4F2B analysed analysed analysed analysed analysed analysed F46F47 F46F47 analysed analysed F46F47 F46F47 analysed F46F47 F46F47 here to illustrate the instrument’s accuracy. Table 1. Chemical composition of the slag samples from Monganstow from the slag. The bottom rows present a comparison of the cert ECRM681-1 ECRM681-1 BCS301-1 BCS301-1 BCS301-1 BCS301-1 Monganstown Monganstown Monganstown BCS301-1 ECRM681-1 ECRM681-1 ECRM681-1 Monganstown Monganstown Monganstown Monganstown Monganstown Monganstown Monganstown Monganstown Monganstown BCS301-1 ECRM681-1 ECRM681-1 Monganstown Monganstown Monganstown

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 72 E2771 | A001/01 Monganstown 1, Co. Westmeath ISSUE 3: Eachtra Journal - ISSN 2009-2237

15.8 Appendix 8: Lithics Finds Report for A001/01 – Monganstown 1, Co. Westmeath

Farina Sternke Department of Archaeology, University College Cork

One lithic find (A001/01:13:1) from Monganstown 1, Co. Westmeath was presented for analysis.

Condition:

The lithic which is a flint survives in burnt and incomplete condition.

Context:

This find was recovered from context 13.

Technology/Morphology:

The find A001/01:13:1 is an unretouched cortical flake produced on a beach pebble with a soft stone hammer.

Dating:

The find is morphologically undiagnostic. Based on its technological characteristics, it is possible that it belongs to a residual Early Mesolithic phase on the site. Alternatively, it could represent evidence for Neolithic activity in the area. An association with the Iron Age and Early Medieval occupation of the site cannot be established.

Interpretation:

The single lithic find from the excavation at Monganstown 1, Co. Westmeath is simple cortical flake which is either associated with an Early Mesolithic or Neolithic occupation. Given the site location on slightly raised ground on an island surrounded by marshland, the possibility of it being evidence for residual Early Mesolithic activity in the area certainly has to be considered. However, it is difficult to support this argument with a single isolated find.

Permalink: http://eachtra.ie/index.php/journal/e2771-monganstown-1-co-westmeath/ 73