1 Introduction During the monitoring of topsoil removal as part of the Lough Mask Regional Water Supply Scheme Stage III under licence 01E0314, a deposit of burnt sandstone was identified at the edge of a zone of peat in the townland of Carrowcor, Co. Mayo (Guinan 2002 and 2003a, 2003b). This site was designated Carrowcor II (Figs. 1&2). No surface evidence to indicate the presence of this site was visible prior to the commencement of development works in the area. After discovery, the site was fenced off and the zone of archaeological interest protected. Cleaning and an initial investigation were conducted to determine the nature and extent of the archaeological deposits. As the site had been uncovered during works, and lay within the pipeline wayleave, a decision was made to excavate it under a licence granted by Duchas (excavation licence 01E1081) in the autumn of 2001 (Appendix 1). Geographical location The site was situated c.76m OD, on the side of a third class road c. 2.3km from Bekan village and c. 10km from Ballyhaunis town in south Co. Mayo (Figs. 1 and 2). GPS co-ordinates were obtained from Mayo County Council’s GPS unit. The site is located at 141508.67 (easting), 278490.73 (northing). The surrounding landscape is one of undulating pasture-land interspersed with low-lying zones of deep peat. The environment around the site today is mainly pastoral, but contains plants associated with drier and more disturbed habitats, probably resulting from modern drainage activities (McClatchie 2002). 2 Archaeological Background The townland of Carrowcor (124 acres) is one of 45 townlands in the civil parish of Knock. No fulachta fiadh were recorded from Carrowcor or the immediate neighbouring townlands prior to this work. However, during monitoring on the Lough Mask Regional Water Supply Scheme Stage III (Guinan 2002 and 2003a), a fulacht fiadh had been discovered 30m west of Carrowcor II. This site was designated Carrowcor I and was excavated under separate licence (01E0680) (2003c and 2004). This monitoring also resulted in the discovery of six previously unrecorded fulachta fiadh in adjoining townlands relatively close to the two sites in Carrowcor. The first of these fulachta fiadh was discovered and excavated in Bekan townland 01E0679 (Guinan 2003d and e), c.2.2km east of Carrowcor. A second fulacht fiadh was recorded during monitoring, in a road cut pipe trench, also in the townland of Bekan c.3.2km east of Carrowcor (Guinan 2002). Two more sites were discovered and excavated in the townland of Cloonbulban (01E0681) 2.8km to the south-east (Guinan 2003f and g). In the nearby townland of Cloontumper, 3.4km to the south-east, two additional fulachta fiadh were recorded. One of these sites (Cloontumper I) was discovered as a spread of burnt sandstone and charcoal in ground disturbed during the reinstating of pipeline wayleave. In the same field just outside the pipeline wayleave, c. 20m to the west, a second possible fulacht fiadh (Cloontumper II), with visible surface remains, was discovered. This extant monument is a roughly semi-circular shaped mound c. 0.5m high. The only recorded monument in the townland of Carrowcor is an enclosure (MA102:013) located 200m north of the excavation site. In the adjoining townland of Lissaniska there are six recorded enclosures (MA092:066; 067; 068; 069; MA102:014 and 015). A number of other prehistoric sites are known from the wider area (Comer and O’ Muraile 1986). One of the most notable monuments is a Bell Barrow located c 3km to the south-east in Belesker townland (MA102:051). A wedge tomb in Greenwood Td (MA092:074) is located 3.3km to the north-east (De Valera & O’ Nuallain 1964, Sheehan 1987-88, Guinan 2001a, 2003c). However, there is a complete absence of any recorded monuments within a 3km radius, south and south west of the excavation at Carrowcor. This in part may be explained by the boggy nature of the land in the area. It is possible that this area was occupied by zones of open water or a lake in prehistory as extensive deposits of marl were encountered during monitoring. 3 A number of finds of Bronze-Age date have been recorded from the environs. These include a hollow wooden tube, bound with bronze ribbon, from a bog in Bekan townland and a dugout canoe has been recorded from Bekan Lake. South-east of Carrowcor a bronze socketed spearhead (NMI find no. 1983:66) was found in Bracklagh townland in 1983 (Kelly, 1986, 34). The Excavation The site consisted of an irregular spread of burnt sandstone within a rich charcoal matrix. It was excavated from a base line established along a 0.5m baulk set out across the centre of the burnt spread, parallel to the Bekan/Barneycarroll road. An area 10.5m north/south by 13m east/west was excavated up to the limit of the acquired pipeline wayleave (Fig 3). Excavation showed that the site extended beyond the wayleave acquisition line to the south where a substantial area of the site remains unexcavated (see note on reinstatement p.25). The site was initially de-sodded and cleaned up to the limits of the wayleave acquisition line (Plates 1-4). After initial site work it became apparent that the site had been subjected to modern agricultural disturbance caused by land reclamation and intense furrow activity. The northern edge, nearest the Barneycaroll/Bekan road, was bounded by a deep field drain and disturbed by the local group water scheme, whose pipe was laid in the area. The Stratigraphy of Carrowcor II, (Figs. 3-9, Plates 1-14) Sod (C1) consisted of a dark brown silty clay, 0.05m in depth which overlay a mid-brown clay topsoil (C2) with occasional fine pebble inclusions ranging from 0.15m-0.4m in depth. This topsoil sealed all archaeological strata (Plate 1). Topsoil was removed to reveal a redeposited layer of a compact, mixed, brown and yellow clay with moderate stone inclusions (C3). This deposit ranged in depth from 0.02-0.4m. This episode of redeposition was the result of land reclamation which the land owner indicated took place in the early 1980’s (Figs 7-9 Plate 1). Beneath this zone of disturbance the original pre-reclamation sod line was discovered (C4). This consisted of a very compact mid-brown layer of fibrous roots in clay c. 0.03m in depth. This was underlain by the pre-reclamation topsoil (C5) which was a compact greyish brown 4 clay with occasional soil inclusions. This was 0.1-0.15m in depth. Context 6 which underlay this original topsoil was a dark greyish brown clay (Figs. 3, 7 and 8, Plates 8-11) The layer was 0.12m deep. It had moderate stone inclusions 50% of which consisted of burnt sandstone. The layer was mixed - a result of ploughing disturbance - and represented the interface between the original topsoil (C5) and the burnt mound proper (C7). A low mound of burnt sandstone within a black peat and charcoal matrix (C7) was exposed immediately below (C6) (Figs. 3 and 7-9 Plates 5-6 and 9-13). This compact deposit was largely composed of burnt sandstone cobbles (50-200mm in diameter) but also contained unburnt pebbles. The surviving mound which ranged from 0.15-0.27m in depth was heavily disturbed by furrow activity. It measured 6.2m north/south by 6m. A substantial area of the mound which extends to the south beyond the excavation limits, remains unexcavated. The burnt mound was heavily disturbed by a series of seven cultivation furrows (C14) (Figs. 3-7, Plate 9). These generally ran north/south or east/west indicative of cross ploughing. The furrow fill varied depending on the material they cut through. In general the fill (C14A) comprised a mix of burnt sandstone, charcoal, lumps of natural boulder clay, marl, topsoil and peat. The furrows averaged 0.3 – 0.1m in width while they ranged in length from 1.3m-3m. The depths ranged from 0.10m to 0.15m. They were generally U-shaped in profile and cut through the burnt mound and the underlying marl into the boulder clay (C10) (Plate 9). Excavation revealed the mound of burnt sandstone (C7) was underlain by an extensive deposit of marl (C8), a natural lacustrine deposit (Figs. 4, and 7-9, Plates 7-13). The cream coloured marl contrasted with the black charcoal filled furrows which cut into it. The marl covered the entire area of the excavation and was noted extensively in the vicinity during topsoil stripping and subsequent pipelaying. The deposit was extremely wet underfoot and whiter in colour towards the south-western side of the site. Frequent snail shell inclusions were visible throughout this deposit. It measured an average of 0.1m in depth but varied greatly across the site. Marl (C8) was underlain by a deposit of mid-brown peat (C11) (Figs. 5 and 7-9, Plates 5-6 and 9 and 12-13). This natural peat growth varied in depth from 0.06 – 0.1m in depth and covered an area of 7.5m by 4m. It overlay compact light grey boulder clay (C10) a natural glacial deposit which underlay the entire area and contained occasional cobble inclusions (Fig. 6, Plates 14 and 15). 5 Beyond this basic stratigraphy a number of other contexts were identified which illustrate the extensive nature of post depositional disturbance to which the site had been subjected. At the south-western fringe of the site an isolated deposit of gravel with some grey clay (C9) covering an area of 1.7m north-northwest/south-south-east by 1.2m was identified. This deposit was 0.06m–0.1m deep and represented material associated with land reclamation. Another disturbed context associated with land reclamation was (C13), a compact grey silty sand with frequent stone inclusions (Fig. 7). This occurred in two separate deposits at the south-western and south-eastern ends of the site and was 0.07m in depth.. Deposit 1 located in the south-western part of the site measured 3.3m by 2.5m in area. Deposit 2 located in the south-eastern part of the site measured 2.5m by 2.5m in area. Underlying Deposit 1 was a disturbed mid-brown very fine peaty clay (C12), which lay in the south-western part of the site on the natural boulder clay (C10) (Figs. 3 and 7 Plates 8 and 9). It covered an area 3.2m northsouth by 2.08m and was 0.15-0.2m deep. Context 15 an irregular featured was half-sectioned (D-D1) (Fig 4), revealing a shallow natural depression 1.5m north/south by 2.2m with a maximum depth of 0.2m. This depression was filled with a dark brown peat with coarse pebble inclusions (C15A) (Fig. 4). 6 Context Type Description Munsell Interpretation 1 Deposit 10yr-3/3 Brown Natural deposit. Overlay and sealed entire site 2 Deposit 10yr-3/3 Brown Natural deposit. Overlay and sealed entire site 3 Deposit 10yr-5/4 Yellowish brown Evidence of reclamation circa 1985 10yr-3/3 Brown Pre- reclamation sod line 10yr-3/3 Brown Pre- reclamation topsoil 10yr-4/2 Dark greyish brown Interface layer between topsoil and burnt mound. Mixed layer resulting from ploughing disturbance 7.5yr-2.5/1 Black Heat shattered sandstone in a black charcoal and peat matrix fulacht fiadh material 10yr-7/3 Very Pale Brown Natural lacustrine deposit which underlay the burnt mound. 10yr-5/1 Grey Redeposited gravel possibly associated with land reclamation 4 Deposit 5 Deposit 6 Deposit 7 Deposit 8 Deposit 9 Deposit Sod Dark brown silty clay Litter layer Depth: 0.05m Topsoil Mid-brown clay Depth: 0.15-0.4m Infrequent stone inclusions Redeposited, mixed, brown & yellow clay Depth: 0.02-0.4m Moderate stone inclusions Compact Original sod line Layer of fibrous root and clay Mid- brown Depth: 0.03m Very compacted Original topsoil Greyish brown clay Depth: 0.1-0.15m Compact Occasional stone inclusions Dark greyish brown clay Area: 10m east/west by 5m Depth: 0.12m Moderate stone inclusions 50% burnt sandstone Disturbed Burnt stone in black peat and charcoal matrix Area 6.2m north/south by 6m Burnt sandstone cobbles 50-200mm in diameter Depth: 0.15-0.27m Extends beyond wayleave Includes unburnt pebbles Disturbed by furrow activity Marl Cream Whiter at south-western edge of site Frequent snail shell inclusions Very wet deposit Average depth: 0.1m but varies greatly across site Grey clay Area 1.7m northnorthwest/south- southeast by 1.2m Depth: 0.06-0.1m Numerous (70%+) well rounded, well sorted pebbles Located in the south-western part of the site 7 10 Deposit 11 Deposit 12 Deposit 13 Deposit 14 Cut 14A Deposit 15 Natural hollow 15A Deposit Grey compact boulder clay Natural subsoil 15% very angular cobble inclusions Larger rocks also present Peat Mid-brown Depth: 0.06-0.1m Area 7.5m east/west by 4m Sterile Largely confined to southern side of site Very fine peaty clay Mid-brown Area 3.2m north/south by 2.08m Depth: 0.15-0.2m Confined to the south-western corner of the site Redeposited grey silty sand Compact with frequent stone inclusions. Resting above (C12) Depth approx: 0.07m 2 deposits one in the southwest (3.3m by 2.25m) and a second in the southeast (2.5m by 2.5m) Associated with land reclamation Furrows Cuts through burnt stone Orientated north/south or east/west indicative of cross ploughing 7 identified Cuts through burnt stone (C7), marl (C8) and into boulder clay (C10) U- shaped profiles Length: 1.3-3m Width: 0.3-1m (varies) Depth: 0.1-0.15m Furrow fill Comprised a mix of burnt sandstone, charcoal, lumps of natural boulder clay, marl, topsoil and peat. Irregular Shallow Area: 1.5m by 2.2m Depth: 0.2m Fill of natural depression (C15) Dark brown peat Area 1.5 north/south by 2.2m Depth: 0.1–0.2m 25% poorly sorted coarse pebble inclusions 2.5yr-7/1 Light Grey Natural glacial deposit underlying the entire area 10yr-4/3 Brown Natural peat growth generally underlying burnt mound and marl deposit 10yr-4/3 Brown Natural deposit overlying boulder clay 10yr-4/2 Dark greyish brown Redeposited material resulting from land reclamation in the 1980’s N/A Modern cultivation furrows Mottled Mixed furrow fill N/A Natural depression 10yr-3/3 Brown Fill of natural depression Table 1 Context List from Carrowcor II 8 Archaeobotanical Analysis During the course of the excavation a number of samples were taken from various contexts for archaeobotanical analysis. Four samples from the fulacht fiadh were subsequently submitted for analysis. This work was undertaken by Meriel McClatchie, Archaeological Services Unit, Department of Archaeology, University College Cork. Methodology (extract from McClatchie 2002). The archaeobotanical material was extracted from each soil sample using a combination of flotation, wash-over and wet-sieving techniques. A combination of these techniques is the most suitable method for the recovery of material as the processes cause the least amount of damage to fragile plant remains and are also effective in recovering small weed seeds. The soil sample is placed into a bucket and soaked in water in order to disaggregate the deposit. The length of time allowed for the soaking of each sample is dependent upon the cohesiveness of the deposit. The water, including the flot, (Flot is the material that floats and is suspended in the water, such as seeds and charcoal), is then poured through a graded bank of sieves (the washover technique) containing mesh measuring 1mm, 500µmm and 250µmm, leaving the residual material that does not float at the bottom of the bucket. This process is repeated until no more material floats to the top of the water. The residue is then collected by wet-sieving which involves it being placed into a 2mm sieve and washed thoroughly in a concentrated flow of water using a hose. This method recovers denser organic material such as nuts and bone. The scanning, sorting and subsequent identification of the archaeobotanical material in all samples is carried out using a Wild 8 zoom stereo-microscope, with magnifications ranging from x6 to x50. Each sample is scanned in order to extract the archaeobotanical material. The seeds are then sorted into general groupings on the basis of visual comparison of their morphological features. The seeds were identified by comparison to reference material in the U.C.C. Archaeology Departmental collection of modern diaspores and the drawings from various seed keys (Anderberg 1994; Beijerinck 1947; Berggren 1969; Berggren 1981; Katz et al. 1965). Some of the seeds were distorted or fragmented and identified to genus level only. The identified taxa are listed by context in the order that they appear in Flora Europaea (Tutin et al. 1964-83). 9 Plant remains present Four soil samples from a fulacht fiadh at Carrowcor II were submitted for analysis. Non-wood plant remains were not recovered from any of the examined deposits. The presence of substantial quantities of land snails in a number of deposits (Table 2) may represent organic or rotting material that would have attracted the land snails. No botanical evidence was recorded to reveal activities at the fulacht fiadh. Conclusions A low level and rather narrow range of flora was recovered from deposits. Evidence for open, damp environments was regularly identified. There is likely to have been woodland available locally, as evidenced by the recovery of flora associated with woodland margins. The absence of charred plant remains besides wood is another indicator that wood was readily available as a fuel source during the various periods of activity, as the remains that could be expected from burning turves or other vegetation were not identified. The Samples The following is a summary of the context and content of the four samples which were submitted for archaeobotanical analysis. Sample no. 3 was taken from context 8, marl located under topsoil at the edge of the burnt mound and contained charred wood and land snails. Sample no. 4 was taken from the main mound. It consisted of burnt sandstone within a charcoal matrix (C7). It contained charred wood. Sample no. 8 was taken from the marl layer (C8) underlying the burnt mound (C7). It contained charred wood and land snails. Sample no. 10 was taken from a peat deposit (C11) which underlay the burnt mound (C7). It contained charred wood, anaerobically preserved wood and land snails. 10 Site Townland Carrowcor 01E1081 Sample Charred Charred number seeds wood Anaerobically Anaerobically preserved preserved seeds wood Anaerobically preserved vegetative material Anaerobically Anaerobically preserved preserved mosses beetles Land snails 3 . ◊ . . . . . ◊ 4 . ◊ . . . . . . 8 . ◊ . . . . . ◊ 10 . ◊ . ◊ . . . ◊ II Carrowcor II Carrowcor II Carrowcor II ◊ = present Table 2 Organic material recorded in deposits from 01E1081 Species identification of charcoal samples (extract from O’ Carroll, 2003) Introduction [One charcoal sample was] …. sent for species identification prior to 14C dating and also to give an indication of the range of tree species, which grew in the area, as well as the utilisation of these species for various functions. Wood used for fuel at prehistoric sites would generally have been selected at locations close to the site. Therefore charcoal identifications may, but do not necessarily, reflect the composition of the local woodlands. Larger pieces of charcoal, when identified, can provide information regarding the use of a species. Methodology The process for identifying wood, whether it is charred, dried or waterlogged is carried out by comparing the anatomical structure of wood samples with known comparative material or keys (Schweingruber 1990). The identification of charcoal material involves breaking the charcoal piece so as a clean section of the wood can be obtained. This charcoal is then identified to species under an Olympus SZ3060 x 80-zoom stereomicroscope. By close examination of the microanatomical features of the samples the species were determined. The diagnostic features used for the identification of charcoal are micro-structural characteristics such as the vessels and their arrangement, the size and arrangement of rays, vessel pit arrangement and also the type of perforation plates. 11 Each species identified from each sample was weighed in grammes. A minimum of 5 grammes of charcoal is needed for a conventional 14 C date but 25 grammes is the preferred amount. Sample no. Context no/ Locational info. SPECIES Weight Prunus Hazel 7, burnt mound material 4 Prunus, hazel & alder (5g), (1g) Alder (1g) Table 3 Species identification results from 01E0680 There were three species identified from Carrowcor II. These were prunus, alder and hazel. Discussion There are three species types present in the charcoal remains. These were hazel (Corylus avellana) and alder (alnus glutinosa) and Prunus (blackthorn). They are all native to Ireland. Alder is a widespread native tree and occurs in wet habitats along streams and riverbanks. Alder also grows regularly on fen peat. It is an easily worked and split timber and does not tear when worked. Alder is commonly identified from wood remains associated with wet/boggy areas. Hazel was very common up to the end of the 17th century and would have been used for the manufacture of many wooden structures such as wattle walls, posts, trackways and baskets. McCracken (1971, 19) points out that “it was once widespread to a degree that is hard to imagine today”. With the introduction of brick, steel and slate the crafts associated with hazel became obsolete, and today the woods that supplied hazel have diminished rapidly. Hazel is normally only about 3-5m in height and is often found as an understory tree in deciduous woods dominated by oak. It also occurs as pure copses on shallow soils over limestone as in The Burren in Co. Clare and survives for 30 to 50 years. Its main advantage is seen in the production of long flexible straight rods through the process known as coppicing. & 12 Prunus was identified from burnt mound material at Carrowcor II. It is a very durable wood and is as strong as oak. Blackthorn (Prunus spinosa) is a thorny shrub found in woods and scrub on all soil types. In a woodland situation it is more likely to occur in clearings and at the woodland edges. It is more likely that the charcoal identified from this assemblage was blackthorn rather than cherry. CONCLUSIONS A total of three species were identified from the site. Alder indicates local wet condition along riverbanks or peat bogs. Hazel would have grown in drier conditions preferring free-draining soils and nutrient rich clays. The Prunus spp. identified is indicative of those species, which may have grown locally in hedgerows or as scrub nearby to the sites. Finds Excavation uncovered 13 lithic artefacts, a polished stone axe and 58 fragments of animal bone (Plates 16-19). In addition a clay pipe stem and a range of modern material including glass, iron and plastic, came from the topsoil or contexts of redeposition resulting from land reclamation. The faunal material was sent to Catherine Bonner for osteoarchaeological analysis, the results of which are detailed below. The Faunal Material (extract from Bonner 2003). Introduction A small quantity of faunal material was submitted for osteological analysis following archaeological excavations of a fulacht fiadh which were carried out as part of the Lough Mask Regional Water Supply Scheme Stage III. The bones were retrieved entirely by hand. Methodology Where possible, bone fragments were identified to species and skeletal element using the Queen’s University of Belfast comparative skeletal collection. Differentiation between sheep and goat could not be achieved and the remains of these species were categorized as sheep/goat. Epiphyseal fusion data is based on Silver (1969). Due to the material’s fragmented state, metrical analysis could not be carried out. 13 Analysis A summary of the results of analysis is presented in Table 4. Overall, the material was found to be quite poorly-preserved. Of the 58 fragments examined only thrirteen could be identified to species and skeletal element. Species identified include cattle (Bos taurus), horse (Equus caballus), sheep/goat (Ovis Capra) and pig (Sus domesticus). The remains of bird and fish are absent from the samples. Carrowcor II (01E1081) Most of the animal bone which was recorded for Carrowcor II comprises tooth fragments (over 90%). Because of their protective enamel, teeth have a better survival rate than of other bones and therefore, their high proportion at this site is probably indicative of unfavourable soil conditions for bone preservation. Context 6 (interface between topsoil and burnt mound) produced the largest sample of animal bone. All of the identifiable fragments from this context are teeth. Species identified include cattle (N=2), horse (N=2), pig (N=1), and sheep/goat (N=1). The pig tooth was identified as a right male lower canine. Its size and morphology indicate that the animal represented was of the domestic variety. Two teeth (1 cattle molar, 1 horse incisor) were recovered from Context 5 (buried topsoil). The horse incisor showed considerable wear. Conclusion The faunal assemblage which was presented for analysis was extremely small, fragmented, and poorly-preserved and as a result, facilitated little interpretation. As mentioned above, four main domesticates (cattle, horse, sheep/goat and pig) are the only species represented. The remains of wild animals are absent from the assemblage. The high proportion of teeth retrieved from Carrowcor II suggests that the paucity of bone is due to poor preservational properties of the soil. Poor bone preservation is argued to be a feature of fulachta fiadh, as approximately only 10% of sites excavated have produced animal bone (Dr. John O Neill pers. comm.). Soils with low pH values have an extremely unfavourable effect on bone preservation (Gill-King 1997) and therefore, it has been claimed that possible acidic soil conditions may account for these poor bone survival rates 14 (Waddell 1998). However, anthropogenic factors, such as the transferral of meat after cooking or the disposal of food waste to deter vermin when the sites were in use, should also be considered (Dr. John O Neill pers. comm.). Site Carrowcor Find No. 2 Context No. Species Anatomy Details No. frag 14 5 5 Cattle Tooth frag. Molar frag. 8 5 1 Horse Tooth Incisor 29 10 29 Unid. - - 25 8 2 Unid. - - 23 6 1 LM LB S/root action visible 14 Unid. - - (01E1081) 1 6 1 Pig Tooth R/Male/Lower canine 15 6 1 Cattle Tooth P4 6 6 1 S/G Tooth Molar 5 6 1 Cattle Tooth Molar 2 6 1 Horse Tooth Mandibular molar 24 6/7 1 Horse Tooth Mandibular molar Key: LM=large mammal; MM=medium mammal; S/G=Sheep or goat; Unid.=unidentified; LB=longbone; L=left; R=right; P=proximal; D=distal; S=shaft; FU=fused; Un=unfused; B=burnt bone. Table 4 Results of analysis of faunal remains from Carrowcor II 01E1081 The Lithic Material The thirteen lithic artefacts uncovered during the course of excavation (Plates 16-19), Appendix 8) were was sent to Dr. Sarah Milliken Department of Archaeology, University College Cork, for analysis, the results of which are detailed below. 15 Lithics Report (Milliken, 2003) Raw material provenance The artefacts are predominantly made of chert, with one made of quartz. Some of the chert may be derived from the Castlebar River Limestone Formation, where bands of chert nodules are common, or it may be derived from the Quaternary deposits which cover much of the area and which were deposited during periods of glaciation, either directly by the huge ice sheets that spread across the country, or by meltwater flowing from this ice as it melted. The ice broke off pieces of the rock over which it passed, and this sediment was deposited as glacial till. The quartz would also be derived from this glacial till. Technology There are 13 artefacts (Appendix 6). In the majority of cases, the pebbles were broken using the bipolar technique, whereby the pebble was placed on a stone anvil and struck with another stone. This technique is used when the pebbles are too small to hold them in the hand, and it results in the simultaneous production of flakes, chunks and split pebbles. In other cases the pebbles were broken using direct percussion, whereby a pebble core was held in the hand and a series of flakes was struck from it using a hammerstone. Typology There is one end-scraper and one hammerstone. The artefacts are not chronologically diagnostic beyond a generic attribution to the later prehistoric period (Neolithic/Early Bronze Age). Polished stone Axe (plate 19) Introduction The polished stone axe was recovered c. 1m north of the burnt mound resting on natural boulder clay and covered with peat mottled marl (C8). The stratigraphical position of the axe suggests that it may have been deposited prior to the use of the burnt mound, possibly into the lake which predated the site. The occurrence of stone axes from both lake and lake shore contexts is documented in Ireland (Cooney and Mandal 1998, 35-6). There is a possibility that the axe may relate to activity at the fulacht fiadh as the marl on which it rests is very wet and 16 soft and the axe may have simply sank into the marl having been deposited, discarded or lost during the lifetime of the site. Recording the Axe The stone axe below has been recorded using the criteria defined by the Irish Stone Axe Project (Cooney and Mandal 1998). Completed stone axe measuring 205mm long, 86mm wide and 52mm thick. Face shape - oblique butt (FS05). Cross section - oval (CS05). Edge shape - curved asymmetrical (ES04). Profile - symmetrical medium (PO2). Blade profile - symmetrical junction 1/2 (BP09). Butt shape - oblique (BU07). Facets - none between sides and faces. Primary treatment - unknown (PT04). Secondary treatment - blade ground and polished. - sides ground and polished. - faces ground and polished. - butt ground not polished. Hafting - Circa two thirds of the length of the sides from the butt towards the blade end were pecked after the axe was ground and polished. Description - The sides were pecked after the axe was ground and polished. Pecking leads from the butt towards the blade end. Pecking designed to facilitate hafting. Junction between the sides and the edge is rounded and clearly defined. The junctions between the sides and butt are marked and angular producing an oblique butt. The edge is curved and slightly asymmetrical. The blade area is smooth and merges with the faces of the axe. The faces are smooth and well polished. The oblique butt is slightly rough and unpolished. Some minor areas of edge damage visible. Striae are visible on the blade area of the axe. Petrology The axe was examined macroscopically by Dr. Stephen Mandal who classified the axe as a greywacke sandstone. 17 Discussion The form of words fulacht fiadh or fulachta fiadh (plural) which is used in Ireland to describe mounds of burnt stone has only been in use since the 19th century. The term is based on early Irish literary references from the 9th century AD where the word fulacht appears on its own with a variety of spellings and meanings. These are listed by O’ Drisceoil who has found fiftynine passages in the early Irish literature where the term fulacht appears ‘fulucht, fulocht, folucht. Inadh fulachta (cooking place) occurs, as do fulucht, fianachta, fulacht fiansae, folach fiann, fulacht fian and fulachta na bhFian’ (1990, 158) The term can have a variety of meanings such as ‘cooking place of deer’, ‘cooking place of game’ or ‘cooking-place of the wild’ (O’ Kelly 1954). In Cormac’s Glossery one of the earliest of the literacy sources the name takes the form fulacht fiansae or ‘cooking place of the roving huntsman'. In later sources such as Geoffrey Keating’s history Foras Feasa ar Eirinn written between 1633 and 1637 fulacht fian is the usual form. Fian here is a reference to the Fianna warriors of the mythical Finn mac Cumhaill. "From Bealtaine until Samhain, the Fian were obliged to depend solely on the products of their hunting and of the chase as maintenance and wages from the kings of Ireland, thus, they were to have the flesh for food, and the skins of the wilde animals as pay. But they only took one meal in the day and night and that was in the afternoon. And it was their custom to send their attendants about noon with whatever they had killed in the morning's hunt to an appointed hill, having wood and moorland in the neighbourhood, and to kindle raging fires thereon and put into them a large number of emery stones, and to dig two pits in the yellow clay of the moorland, and put some of the meat on spits to roast before the fire, and to bind another portion of it with suagans in dry bundles, and set it to boil in the larger of the two pits and keep plying them with the stones that were in the fire, making them seethe often until they were cooked. And these fires were so large that their sites are today in Ireland burnt to blackness, and these are now called Fulacht Fian by the peasantry." (P.S Dineen ed 1908 326-9). 18 The Function of Fulachta Fiadh. Fulachta fiadh have been the subject of interest and academic study in Ireland since the mid19th century (Trench 1885-86), (Quinlan 1885-86), (Forsayeth 1911). However it was not until the 1950’s that this monument type was studied at a scientific level. M.J O’Kelly seminal work on the subject (1954, 105-155) was holistic in its approach, combining the results of his excavations with a review of the evidence from the early Irish literature and a program of experimentation. O’Kelly excavated two burnt mound at Ballyvourney in Co. Cork. His experimentation demonstrated how these monuments could have been used as cooking places. Using a reconstructed trough filled with water he showed that stones heated in a fire and placed in the trough would bring the water to the boil in 30 to 35 minutes. This water could be maintained simmering at boiling point by adding the occasional hot stone to the trough. In this experiment a 4.5kg leg of mutton wrapped in straw was ‘cooked through to the bone and free of all contaminants’ in three hours and forty minutes. ‘Thus we satisfied ourselves that such a trough made in the ground could be used effectively for the cooking of meat in a manner described the early Irish literature’. (O’ Kelly 1954, 122). In more recent years similar experiments have been undertaken by Christy Lawless in Co. Mayo (1990). Despite the fact that it has been demonstrated that fulachta fiadh could be used for cooking there is no agreement in the literature that they were always used in this way. Although Keating’s description of burnt mounds in his account of the Fianna (quoted above) appears to suggests they were cooking sites, Gillespie (1991, 69-70) has urged caution in how this account is interpreted. Citing other literary references from the late 16th and early 17th century it is suggests that fulacht fiadh should not be considered simply as cooking sites but temporary camps where a range of activities, of which cooking is only one element, were carried out. General problems associated the value of Irish literary sources have been discussed by O’ Drisceoil (1990 157-164). References to fulachta fiadh are found in an assortment of texts from early Irish law tracts, glossaries and hagiographies to histories, poetry and annals and span a very large time frame from 9th to the 18th century. This presents a huge time gap between their use in the Bronze Age and references to their use in medieval texts. Written in old, middle, early modern and modern Irish the language of these texts is very difficult to interpret and translate. In addition many accounts are fanciful and would have had a long life in oral tradition before being written down, affecting their reliability as sources for interpreting the function of fulachta fiadh. As O Drisceoil has noted these sources have limited 19 archaeological value ‘It would be wrong to uncritically accept the early Irish literary evidence in an interpretation of the cultural, social and economic contexts of fulachta fiadh, in dating them and in ascribing a function to them’ (1990, 157) An alternative interpretation to cooking is that burnt mounds may have been used as saunas/sweat houses or for bathing (Barfield and Hodder 1987, O’ Drisceoil 1988, Barfield 1991). The proximity of these sites to a source of water as well as the many ethnographic parallels for sweat bathing are cited in support of this hypothesis. If these sites were used as saunas the trough which is a feature of many excavated sites could have been used to collect water for steam production or used as a source of cold water for pouring over the body. The general absence of animal bone from excavated fulachta fiadh is also used to argue against the interpretation of fulachta fiadh as cooking places. However the presence of acidic soils, which would greatly affect preservation of bone, offers one possible explanation for the lack of this type of material. It is also possible that there is a behavioural explanation for the lack of bone on excavated fulachta fiadh. Quite simply cooked food may have been removed from these sites for consumption elsewhere, perhaps in nearby settlements. While bone is often absent from fulacht fiadh this is by no means always the case. Excavation of a fulacht fiadh from Fahee South in Co. Clare, produced a large number of animal bones perhaps reflecting the low acidity of the soil in the area (O’ Drisceoil 1988). Beyond cooking and bathing the possibility that fulachta fiadh may have functioned as laundries and centres of textile production where clothes were prepared, fulled, washed and or dyed has suggested (Jeffery 1991). Fulling is the process of cleaning, shrinking and thickening clothes. It has been argued that the debate about the function of fulachta fiadh should take into account the general range of uses to which hot stones can be put. Barfield lists the following suggested uses for hot stone technology - storage heating, beer making, canoe manufacture, drying meat or fish, fumigation to eliminate flies, butter production, pottery firing, leather preparation, metal working, fulling, steam for birthing, the extraction of grease from animal bones, salt production and the use of steam to bend long timbers for use in building (Barfield 1991, 6264). 20 Fulachta fiadh should perhaps not be thought of simply as cooking sites but rather multifunctional features of temporary settlements used for both cooking and washing. While this discussion has focused on the possible domestic uses of fulachta fiadh the possibility that some of these sites may have had a role within the ritual/spiritual world of prehistoric communities should also be considered. While the position of fulachta fiadh in the wider settlement patterns of the Bronze Age is not very well understood, associations between these sites and other contemporary monument types has been noted across a number of regions. In the south Limerick region, survey has shown that fulachta fiadh are one component of an extensive range of Middle Bronze Age sites, including burial and domestic sites forming an integrated settlement pattern (Cooney and Grogan 1994, 124). Similarly Waddell (1998, 117) points out that a number of fulachta fiadh, near Carron, Co. Clare, seem to be part of a complex of sites including field system, cairns and wedge tombs (Hayden 1994). In the Monavullagh Mountains, Co. Waterford a group of fulachta fiadh have been found in close association with house sites, stone circles and multiple cist cairns (Buckley 1991, 7). In Mooghaun, Co. Clare an association between the distribution of fulachta fiadh and standing stones has been identified (Grogan 1996). Perhaps a greater understanding of these ubiquitous monuments will emerge from more intensive regional landscape studies which place more emphasis on integrating these sites into the broader settlement patterns of Bronze Age Ireland. With so many previously unknown fulachta fiadh being discovered every year across the country this presents a challenge for Irish archaeology. Fulachta Fiadh in Co. Mayo As recently as 1986, there were only two fulachta fiadh known from County Mayo. One from Dooros Td near Balla (Buckley & Lawless 1986) and another near Ballina. The fulacht fiadh at Dooros is of particular note as an undecorated gold dress fastener, is recorded from this area, (NMI reg. no. 1934:5600) (Cherry 1990,53). Intensive fieldwalking by Christy Lawless between 1986 and 1987 within a 16 sq. km area in the centre of the parish of Turlough lead to the discovery of 130 new sites (Buckley and Lawless 1986), (Lawless,1990). The number of recorded sites in Turlough parish continues to increase with some 152 fulachta fiadh now known (Lawless 2001, 84). Additional sites have been recorded in Killasser, near Foxford (O’ Hara 1991) and the Clare Island survey has discovered several new fulachta fiadh (Gosling 21 1994). For Co. Mayo as a whole, the Sites and Monuments Record classifies 258 sites as fulacht fiadh or fulachta fiadh (Gibbons et al 1991, 17). Over the last five years archaeological monitoring associated with major infrastructure projects in Co Mayo has lead to the discovery and excavation of a large number of new sites, adding greatly to the corpus of fulachta fiadh known from the county. While many of these sites have yet to be fully published, a quick survey of the Excavations Bulletins suggests that in excess of 50 fulachta fiadh have been excavated in the county over the last number of years. New sites have been identified as a result of monitoring works during the construction of the Knock/Claremorris by-pass (Nolan 2000a-c, 2003, King 2003, Zajac 2000, 2002, 2003). Additional concentration of fulachta fiadh have also come to light during works associated with Claremorris sewage and Westport main drainage schemes (Gillespie 2000, 2001 and 2003a-d, 2004 a-c). The discovery of fulachta fiadh around Ballinrobe have been reported upon in by Gerry Walsh (1995a and b and 2000). Drainage works have exposed fulachta fiadh at Lough More, Bofeenaun (Moloney and Keane 1992). Additional sites have been excavated near Ballina (King 2000), Knock, (King 2004), Crossmolina (Zajac 2004), Claremorris (Nolan 2003) and a site at Coolavally, near Ballinrobe (Ryan 1996). Work by the writer in recent years in south Co. Mayo has resulted in the discovery and excavation of many additional new sites (Guinan 2002 and 2003a). A re-appraisal of our understanding of burnt mounds was undertaken in 1988 and a series of papers published in Burnt Offerings: International Contributions to Burnt Mound Archaeology (Buckley 1990) This was followed by a second international burnt mound conference in Sandwell in the UK (Barfield and Hodder 1991). Since then an explosion of archaeological activity across the country in the 1990’s has greatly increased our data on the subject. On a national level the numbers of fulachta fiadh known in Ireland continues to grow. Waddell (1998,174) put the number of known fulachta fiadh across the country at 4500. With the current rate of discovery the actual number of sites is probably far in excess of this. At present it can be said that the collection of archaeological data on burnt mounds is currently outpacing our ability to synthesise the available evidence. 22 The Fulacht Fiadh at Carrowcor II The site at Carrowcor II was the disturbed remains of what was originally a much larger mound. Evidence of significant modern post depositional disturbance affecting the site was discovered. The sod, topsoil and layer of compact redeposited, mixed, brown & yellow clay which covered the site was the result of recent agricultural land reclamation. The mound itself was subject to further agricultural disturbance in the form of cross ploughed cultivation furrows which cut through the archaeological strata, resulting in a degree of mixing. While there is no obvious surface source of water at the site at present, extensive land reclamation in the area may have altered pre-existing watercourses. Extensive deposits of marl which underlay the site and surrounding environs may suggest the existence of open water in the past. In any event the water table lies close to the surface in this area and even a relatively shallow trough would gather sufficient water. The land owner indicated that there are a series of springs located along the base of the glacial ridge north of the site. An area 10.5m north/south by 13m was excavated up to the limit of the acquired pipeline wayleave. Excavation showed that the site extended beyond the wayleave acquisition line to the south. No boiling pit or trough was found in the excavated portion of the site but it is possible that one or more may be located in the unexcavated part of the site. Archaeobotanical analysis of environmental samples revealed no non-wood remains. Wood appears to have been the exclusive source of fuel at the sites as there was a general absence of any other charred plant remains. All the environmental evidence suggests that the materials used at the site were available within the immediate locality. Three species of wood - hazel, prunus and alder - were identified from charcoal samples recovered from the site investigated at Carrowcor II. Alder suggests local wet condition along river banks or peat bogs. Prunus is a thorny shrub found in woods and scrub on all soil types. The hazel would have grown in drier conditions preferring free-draining soils and nutrient rich clays. The site was located at the interface between a glacial ridge and an area of low- lying deep peat so it is probable that these materials were readily available within the immediate area. 23 Although a quantity of faunal remains (pig, horse, sheep/goat and cattle) were recovered from the site (Bonner, 2003) this material largely came from contexts disturbed by modern agricultural activity. In addition to a range of lithic debitage, a chert end scraper and a quartz hammerstone, a polished stone axe was found at the edge of the site. A number of stone axes have been found close to or associated with fulachta fiadh. In the 19th century Quinlan (1885-6, 392) reported the discovery of three axeheads ‘within a few feet of the burnt stones’ of a fulacht fiadh at Clonkerdon, Co. Waterford. However there is some confusion about these artefacts as it is not certain whether they were made of stone or bronze (Cherry 1990, 53). At Kiltrassy in Co. Kilkenny a stone axe made of schist was found within ‘a destroyed fulacht fiadh exposed by ploughing’ (Cherry 1990, 52). At Sheephouse in Co. Meath a burnt stone found within the burnt mound of a fulacht fiadh was ‘probably the mid-section of a polished stone axe’ (Campbell 1994, 73). A polished mudstone axe was found close to a Middle Bronze Age fulacht fiadh dated 1606-1400BC at Leedaun, Co. Mayo (Walsh 2000, 311-2). Polished stone axes was recovered during excavations at Gorteen in Co Limerick (Deevy 2000, 166), and in Clare Co. Mayo (Nolan 200b, 228). Recently a polished stone axe was found within a fulacht fiadh (site 23 at George’s land in Co Tipperary during archaeological excavation in advance of N8 Cashel By-pass (www.nra.ie, 2004, 3). Dating the Fulacht Fiadh at Carrowcor II Export and alteration licences were obtained to send charcoal extracted from sample no. 4 for radiocarbon dating, to the Scottish Universities Research and Reactor Centre. (Appendices 2 and 3). Sample no. 4 was taken from the southern side of the main body of the burnt mound (C7). 7 grammes of charcoal, was extracted from this sample. This comprised of Prunus (5g), Hazel (1g) & Alder (1g). This charcoal was dated by the University of Arizona AMS facility to 2890+/- 45 BP (Cal BC 1220 –920) (AA-54325 GU-10830), (Appendices 4 and 5). The date from the site suggests that much of the activity at this fulacht fiadh took place in the early stages of the Late Bronze Age (1200-600 BC) date. The 13 lithic artefacts recovered from the site were mainly debitage but also included one end scraper and one hammerstone. In 24 relation to the chronology of these artefacts Milliken (2003) remarks that “The artefacts are not chronologically diagnostic beyond a generic attribution to the later prehistoric period.” The polished stone axe recovered from the site is of a type whose main period of production is the Neolithic (4000-2500 BC). However “the use of stone axes begins in the early Mesolithic and continues well into the Bronze Age” (Cooney and Mandal 1998, 3). Excavations undertaken by the author in recent years on a number of fulachta fiadh within the general area of the site at Carrowcor II have produced a number C14 dates ranging from the Final Neolithic to the Late Bronze Age. At the neighbouring fulacht fiadh, Carrowcor I, located only 30m to the west, two radiocarbon samples from the burnt mound dated the main activity at the site to Cal BC 1749 and Cal BC 1896-1685, placing the site at the end of the Early Bronze Age (2300-1700 BC) (Guinan 2004). Despite its proximity and similarity, the Late Bronze Age date (Cal BC 1220 –920) for the site at Carrowcor II indicates that fulacht fiadh activity in this area occurred as two distinct episodes circa five hundred years apart. The excavation of a fulacht fiadh in Bekan townland, c. 2.2km east of Carrowcor, produced dates indicating that activity involving burnt stone technology took place there in the Middle Bronze Age (1700-1200) (Guinan 2003d and e). Also in Bekan townland a charcoal sample extracted from a fulacht fiadh found in a road cut pipe trench 3.2km east of Carrowcor has produced an Early Bronze Age date (Guinan forthcoming). Two fulachta fiadh, excavated in Cloonbulban townland, c. 2.8km to the south-east of Carrowcor, produced the earliest dates with activity here occurring in the Final Neolithic and Early Bronze Age (Guinan 2003f and 2003g). 25 Summary and Conclusion A large spread of burnt sandstone (Carrowcor II) was identified in Carrowcor townland, near Claremorris, Co Mayo during topsoil stripping in advance of pipelaying associated with the Lough Mask Regional Water Supply Scheme Stage III. The site consisted of a low mound of heat shattered sandstone within a compact, dark charcoal rich matrix. The burnt mound was heavily disturbed by a range of post depositional processes, the most significant of which was the damage caused by a number of cultivation furrows that cut through the site. No boiling pit or trough associated with the burnt mound was found. Such features may well be present in the portion of the site which lies outside the pipeline wayleave and so remain unexcavated. Thirteen lithic artefacts, made from chert and quartz were recovered from Carrowcor II. In the majority of cases, the chert pebbles were broken using the bipolar technique. Amongst the assemblage were one chert end scraper and a quartz hammerstone. A polished stone axe was also recovered from the site. Archaeobotanical analysis of soil samples revealed no botanical evidence for activities at the fulacht fiadh. Three species of wood (alder, prunus and hazel), were identified from charcoal taken from the site. One radiocarbon date was obtained from charcoal samples taken from the site. This indicate that activity at the site took place in the Late Bronze Age. Note on site reinstatement Excavations took place solely within the Mayo County Council acquired wayleave. After consultation with Duchas the following actions were undertaken to protect the surviving section of the site along the wayleave acquisition line. 1. Heavy-duty teram was used to cover the exposed site section created by the wayleave acquisition line (Plate 15). 2. A layer of gravel was placed on the teram to keep it is place against the section. 3. Normal topsoil reinstating was carried out under strict archaeological supervision. 4. The precise survey location was supplied to Duchas so that the site could added to the RMP.