1
Vegetation changes and woodland management associated with a prehistoric to 1
medieval burnt mound (fulachta fiadh) complex at Ballygawley, Northern Ireland 2
3
4
ABSTRACT 5
This paper examines the impact on woodlands associated with burnt mound use 6
from floodplain sediments and peats, using a combination of pollen, non-pollen 7
palynomorphs, micro- and micro-charcoal and worked wood for the first time. We 8
present this data from a multi-period burnt mound complex, dating from the Late 9
Neolithic to the Medieval period, at Ballygawley, Co. Tyrone, Northern Ireland, to 10
reconstruct vegetation changes from the Neolithic onwards to establish the 11
significance of these changes, in particular on woodlands, whilst the burnt mounds 12
were in use. The findings from the macroscopic charcoal suggests the most abundant 13
trees were commonly, but not exclusively, exploited. Local woodland was seemingly 14
unaffected by use of burnt mounds during the Neolithic and early Bronze Age based 15
on pollen evidence. A sustained increase in microscopic charcoal coincides with a 16
permanent decrease in alder-carr woodland during a period of near continuous 17
burnt mound use between 1725 and 530 BC and a second phase of high microscopic 18
charcoal values, c. AD 880, corresponds to the end of the penultimate phase of burnt 19
mound use. Evidence from the worked wood indicates that some form of woodland 20
management may have used for hazel from the Neolithic onwards. 21
22
Keywords: prehistory, medieval, burnt mounds, pollen, non-pollen palynomorphs, 23
charcoal, woodland management, Northern Ireland. 24
25
INTRODUCTION 26
Burnt mounds or ‘fulacht fiadh’ are a common feature in the Irish and British 27
archaeological record (Brindley et al. 1989; Buckley 1990; Feehan 1991), dating from 28
the Neolithic to the medieval period (Anthony et al. 2001; Ó Néill 2009), but their 29
function has been long debated (see Hawkes 2013) and as such they remain an 30
2
archaeological enigma. Until, recently, few palaeoenvironmental studies have 31
focussed specifically on understanding the function and wider environmental 32
context of burnt mounds (e.g. Innes 1998; Gonzalez et al. 2000). Brown et al. (2016) 33
summarises some of the recent palaeoenvironmental work done in Ireland and 34
suggests that the most likely function of burnt mounds studied was textile 35
production or activities related to hide cleaning and tanning. Wheeler et al. (2016) 36
also used a palaeoecological approach to place a number of Late Neolithic and Early 37
Bronze Age burnt mounds into an environmental context at two locations in County 38
Tyrone, Northern Ireland. Palynological results from their study showed that activity 39
horizons at each site shared similar characteristics: high microscopic charcoal values, 40
repetitive fluctuations in tree and shrub taxa, increased Sphagnum, and the presence 41
of non-pollen palynomorphs (NPPs), all of which could be diagnostic indicators of 42
burnt mounds in palynological records. While the data did not allow Wheeler et al. 43
(2016) to ascribe a specific function for the burnt mounds, the ‘seesaw’ pattern of 44
tree and shrub pollen, combined with the macroscopic charcoal data, indicate 45
possible species selection and management of the local woodland species for 46
fuelwood. 47
48
This study focusses upon a burnt mound complex at Ballygawley, where activity 49
dates from the Late Neolithic to the medieval period. We present microfossil data 50
from a new sampling site within the burnt mound complex, together with new 51
charcoal (spanning the mid Bronze Age to the medieval period) and wood 52
technology data (from burnt mound troughs). The aims of the paper are to: (i) 53
reconstruct the vegetation history of a burnt mound complex using pollen, NPPs, 54
3
microscopic charcoal data and anthracological data; (ii) examine the wood types and 55
technology used in trough construction, specifically in relation to woodland 56
management (iii) identify vegetational changes, particularly on woodlands, 57
associated with the use of the burnt mounds and (iv) compare the results found in 58
this new investigation to those presented in the previous investigation of the same 59
burnt mound complex by Wheeler et al. (2016). 60
61
The term ‘burnt mounds’ is used throughout the paper ‘to refer to a site which 62
contains one or more mounds or stone-spreads containing burnt or heated stones 63
with, or without, an associated trough or pit but with no connotation as to function’ 64
following Brown et al. (2016:3). 65
66
ARCHAEOLOGY AND SITE DETAILS 67
Palaeoenvironmental sampling was carried out as part of the archaeological 68
evaluation and excavation strategy associated with the A4/5 road improvement 69
scheme between Dungallen and Ballygawley, Co Tyrone, Northern Ireland, 70
undertaken by Headland Archaeology Ltd (Figure 1). 71
72
The site is located in low lying pasture land several hundred metres east of the 73
Ballygawley Water, on the edge of the floodplain at the foot of higher ground 74
formed by drumlin topography where layers of peat were discovered, overlain by 75
alluvial silts and clays. Prior to excavation the site was well-drained pasture with a 76
diverted stream, which archaeological excavation and the analysis of historical maps 77
4
suggesting that it is a modern diversion of a natural meandering stream present on 78
site for over c. 5000 years. A number of palaeochannels and alluvial islands that 79
formed a migrating channel system were recorded. Radiocarbon dating has shown 80
that the channels represent the various stages of a migrating channel system, which 81
has generally moved southwards over time, and that the site was used from the Late 82
Neolithic to the medieval period. Within the channel system, 23 burnt mound groups 83
were discovered, together with ten timber and wattle-lined troughs with associated 84
pits and hearths (Bailey 2010; Bamforth et al. 2010). A radiocarbon chronology of 85
these features indicates that they were used from c. 3350 BC to AD 1270 (Wheeler et 86
al. 2016). The earliest radiocarbon date from a burnt mound is dated to 2897-2671 87
BC (GU-17361) and the youngest has been dated to AD 1041-1220 (GU-17375) 88
(Figure 2; Table SI1). Of those burnt mounds radiocarbon-dated, 10 were in use 89
during the Neolithic and Copper Age, 12 during the Bronze Age, 4 in the Iron Age and 90
5 in the medieval period. Some burnt mounds were in use during more than one 91
archaeological period. A hiatus of activity of approximately 900 years occurred 92
between the late Iron Age and early medieval period at the site, yet the overall 93
longevity of activity indicates that people returned in order to use hot stone 94
technology. A monolith of 168 cm depth, named BG-M1, was taken from a section to 95
the east of the majority of excavated burnt mounds (Figures 1 and 3). It is located 96
close to several burnt mounds approximately 20 metres to the north-east and south 97
west: BM9003 is dated to the Late Neolithic and BM9009 has been dated to the 98
Early/Middle Bronze Age (Table SI1). No burnt mound deposits were found at BG-M1 99
so the stratigraphy comprises deposits of peat, alluvial silts and clays. The 100
5
stratigraphical relationship between BG-M1 and burnt mounds 9003 and 9009 can 101
be seen in Figure 3. 102
103
METHODS AND MATERIALS 104
105
Microfossils 106
Sub-samples from BG-M1 were taken over 5 cm intervals. Each 1 cm
3
sub-sample 107
was prepared for microfossil analyses following Barber (1976) including the 108
additional step of density separation (Nakagawa et al. 1998). A sum of 500 total land 109
pollen (TLP) was achieved for all sub-samples. Data are expressed as a percentage of 110
the TLP, with spores and aquatic taxa excluded from the TLP sum. NPPs were also 111
counted during routine pollen analysis (cf. van Geel 1978; van Geel et al. 1982/1983, 112
2003) and they are expressed as a percentage of TLP plus total NPPs. Rare types are 113
indicated by a cross (+), where one cross is equal to one pollen grain or NPP. 114
Microscopic charcoal was counted in three fractions (<21μm, 21-50μm, and >50μm). 115
Identification, including cereal-type pollen, was aided by reference keys in Fægri et 116
al. (1989), Moore et al. (1991), Beug (2004) and Reille (1999), and supported by a 117
modern type-slide reference collection housed at the University of Aberdeen. As the 118
separation of Myrica gale from Corylus avellana-type can be difficult these pollen 119
grain types are classified as Corylus avellana-type (Edwards 1981). Plant 120
nomenclature follows Stace (2010). Basic land use designations interpreted from the 121
pollen records follow Brown et al. (2007). Loss on Ignition percentages (LOI) were 122
also determined (Schulte and Hopkins 1996). 123
124
