Pigs, people, and proximity: a 6000-year isotopic record of pig management in Ireland

Abstract

The ways that pigs interact with humans are more flexible than other livestock. This plasticity means that pig behaviour can evidence a tremendous range of cultural phenomena, some of which may not otherwise show up in the archaeological record. We explore how people and pigs interacted in Ireland over 6000 years (4000 BC–AD 1900) from the perspective of isotopic zooarchaeology, using a large sample of pigs from 40 sites. Results demonstrate continuity and dramatic change. While pig diets show an emphasis on pannage throughout much of the period, husbandry was fundamentally reconstructed in the early medieval period. Through prehistory, pigs were herded in areas distant from human settlements, whereas later they were relocated to live near people. We explore potential implications of these patterns at a range of scales, from economics, to perspectives on zoonoses, and animal agency. While syntheses of a similar scope are needed for other areas of Europe, these findings may reflect a uniquely Irish trajectory of human–animal relationships.

Full text

Pigs, people, and proximity:
a 6000-year isotopic
record of pig management
in Ireland
Eric Guiry1,2, Fiona Beglane3, Finbar McCormick4, Eric
Tourigny5 and Michael P. Richards6
1Department of Anthropology, Trent University, 1600 West Bank Drive, Peterborough, Ontario,
Canada K9L 0G2
2School of Archaeology and Ancient History, University of Leicester, Mayor’s Walk, Leicester LE1
7RH, UK
3Centre for Environmental Research Innovation and Sustainability, School of Science, Atlantic
Technological University, Ash Lane, Sligo F91 YW50, Ireland
4School of Natural and Built Environment, University Road, Queen’s University Belfast, Belfast
BT7 1NN, UK
5School of History, Classics and Archaeology, Newcastle University, Newcastle upon Tyne NE1
7RU, UK
6Department of Archaeology, Simon Fraser University, Education Building 9635, 8888
University Drive, Burnaby, British Columbia, Canada V5A 1S6
EG,0000-0002-1467-1521; FB,0000-0002-6619-0586
The ways that pigs interact with humans are more flexible
than other livestock. This plasticity means that pig behaviour
can evidence a tremendous range of cultural phenomena,
some of which may not otherwise show up in the
archaeological record. We explore how people and pigs
interacted in Ireland over 6000 years (4000 BC–AD 1900)
from the perspective of isotopic zooarchaeology, using a large
sample of pigs from 40 sites. Results demonstrate continuity
and dramatic change. While pig diets show an emphasis
on pannage throughout much of the period, husbandry was
fundamentally reconstructed in the early medieval period.
Through prehistory, pigs were herded in areas distant from
human settlements, whereas later they were relocated to
live near people. We explore potential implications of these
patterns at a range of scales, from economics, to perspectives
on zoonoses, and animal agency. While syntheses of a similar
scope are needed for other areas of Europe, these findings
may reflect a uniquely Irish trajectory of human–animal
relationships.
© 2025 The Author(s). Published by the Royal Society under the terms of the Creative
Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits
unrestricted use, provided the original author and source are credited.
Research
Cite this article: Guiry E, Beglane F, McCormick F,
Tourigny E, Richards MP. 2025 Pigs, people, and
proximity: a 6000-year isotopic record of pig
management in Ireland. R. Soc. Open Sci. 12:
241300.
https://doi.org/10.1098/rsos.241300
Received: 1 August 2024
Accepted: 18 November 2024
Subject Category:
Earth and environmental science
Subject Areas:
biogeochemistry, palaeontology, ecology
Keywords:
archaeology, historical ecology, stable isotopes,
animal proximity, woodland, pannage
Author for correspondence:
Eric Guiry
e-mail: eric.guiry@trentu.ca
Electronic supplementary material is available
online at https://doi.org/10.6084/
m9.figshare.c.7577661.

1. Introduction
Pigs hold a special place in many cultures and regions of the world. We recognize their intelligence and
social nature as standing apart from other livestock and this helped to create diverse ways of husband-
ing and otherwise relating to pigs [1]. Research on the ways in which pigs have been incorporated into
human societies and landscapes through time can open valuable windows onto human activity in the
past. Such work can, for instance, reveal clues about what resources were available for human societies
as well as the decisions people made about both how to practice agriculture and how to situate other
intelligent, agentive beings in their worlds [1,2]. In the context of archaeological interpretations, the
complexity and broad scope of pig behaviour, from the perspective of both intelligence and dietary
range, represent challenges and opportunities. On the one hand, their sheer flexibility (in terms of the
diverse possible ways of raising pigs) means that evidence for specific behaviours and approaches
to pig husbandry might be equivocal. On the other hand, these same qualities make pig husbandry
something that could be moulded into unique sets of practices in different places and times, something
which would be well positioned to signify cultural differences and meanings.
In this paper, we explore the ways in which pigs have been husbanded at a macro-scale across
Ireland over 6000 years (ca 4000 BC to AD 1900, including 40 sites; figure 1) within a dietary framework
structured by isotopic compositions of archaeological pig and cattle bones. Documenting what pigs
ate at different times offers perspective on continuity and change in Irish human–animal relationships
that could not easily be achieved using other lines of archaeological evidence or through smaller-scale,
site-level approaches. These data establish an interpretive context in which, at a coarse scale, we can
observe patterns in the degree to which pigs lived near people. We explore both of these narratives
(i.e. diet continuity and proximity to people) as evidence for what may have been a distinctly Irish
trajectory of human–animal relationships. We also highlight some of the implications these perspec-
tives have for other research areas, including for understanding trends in potential for zoonoses,
the structuring of connections between styles of animal husbandry and settlement patterns in late
prehistoric Europe, and approaches to comprehending animals as persons.
2. Context
2.1. Pig ecology
The domestic pig (Sus scrofa domesticus) is closely related to the wild pig (Sus scrofa), which is often
termed ‘wild boar’, the phrase used here, although a boar is technically the male of either taxon. The
wild and domestic forms can freely interbreed. Zooarchaeologically the two are differentiated mainly
by the size of bones and teeth, with wild boar being substantially larger than prehistoric and historic
domesticates [3,4]. Wild boars maintain an ecological niche that overlaps with that of humans. They are
predominantly herbivorous, consuming a range of plant materials including nuts, seeds, mast (acorns),
roots, herbs, and fruit, which are high in carbohydrates. As omnivores, they will also consume carrion,
young birds and mammals, and eggs of ground-nesting species [5], which would have the effect of
increasing the protein and fat contents of the diet. Extensively herded, domestic pigs can be allowed
to forage in woodland and may also, for example, glean grain in fields after harvesting. Alternatively,
pigs can be kept intensively in sties and fed with a variety of foodstuffs including grain, as well as
waste food from human settlements including vegetable trimmings, meat scraps, and dairy byproducts
such as whey. Pigs can also be fed with human excrement, which is a source of nutrients including
animal protein, although this has significant risks of parasite transfer [6]. Sty feeding can offer more
animal-protein-rich diets than woodland feeding, and therefore typically promotes faster weight gain.
Extensive and sty-based husbandry methods can also be combined on a seasonal basis [7]. Given these
factors, extensively herded domestic pigs are likely to have similar isotopic signals (see below) to wild
pigs, with differences from this likely to be caused by seasonal or year-round sty-based feeding.
Skeletally, pigs are fully grown at ca 42 months, based on fusion of the long bones [8]. Given
adequate food, piglets from modern breeds initially grow quickly, before the growth rate tails off and
a maximum weight is achieved at ca 1000 days (2.75 years) [9]. In modern pig husbandry, following a
carefully controlled high-protein diet, the age at slaughter is typically ca 6 months. However, this was
not the case archaeologically, with ages of 1.5–2.5 years more typical in Ireland [10]. In his ethnographic
study of woodland-oriented pig rearing in 1960s Spain, Parsons [11] noted that for these extensively
managed herds, food supplies varied through the year and growth was slow. In their first autumn,
2
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

the juvenile pigs could not fatten on the mast crop, which was high in carbohydrate, but lacked the
protein needed for skeletal development. Instead, the pigs were kept for another year, and based on
a January farrowing the main period of slaughter was at 20–24 months, with these older pigs having
been fattened on mast (acorns and other tree nuts) immediately prior to slaughter.
2.2. Pigs in Ireland’s past
While archaeological and historical evidence for human–pig relationships in Ireland is reviewed where
relevant in §4, it is worth outlining some of the broader patterns evident in the literature. The richest
body of evidence for early Irish pig husbandry has been assembled by Kelly [12] from law-texts,
wisdom-texts, annals and tales recorded in the early medieval period (ca AD 600–800). Compared
with the archaeological evidence from prehistory, these sources provide a highly detailed interpretive
framework (see §4).
McCormick has synthesized the zooarchaeological evidence for prehistoric- and historic-period
pigs in faunal assemblages from across Ireland (for reviews, see [13–20]). While these sources note
considerable gaps in the available faunal record during the prehistoric period, generally cattle remains
are more frequent than those of pigs [13]. Notably pig-rich assemblages occur at Late Neolithic
Newgrange, Bronze Age Lough Gur, and Early Iron Age Navan Fort [13]. Interpreting evidence for
human–pig interactions from these assemblages is challenging, however, because it is conceivable that
they could reflect human activities, such as ceremonial feasting, considered special or unusual by their
participants. In the context of this limited body of evidence, knowledge about the roles pigs played
during the prehistoric period in Ireland remains limited [13,14].
By contrast, archaeological evidence for pigs becomes more abundant in the historical period,
with evidence pointing to greater variation in the importance of pigs relative to other livestock
across different site types and environments [14,17,20]. Combining documentary and archaeological
evidence, we can see a more complete picture of typical management practices [12,17]. Pigs were
Figure 1. Map showing site locations. Symbols: yellow are sites with pigs only; green are sites with pigs and baseline data from cattle;
purple are sites with baseline data from cattle only. See electronic supplementary material, figure S1, for version listing sites and
electronic supplementary material S2 for version showing sites by time period.
3
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

single-farrowing (spring), with up to nine piglets per litter. Husbandry practices typically incorporated
feed from both dairy products and pannage as well as grain-based resources (see §4). The peak age
for pig slaughter was between the ages of 1.5 and 2.5 years at most sites, which likely optimized the
quantity of meat produced relative to the feed input, though they tended to be killed at an earlier age
in some urban contexts [14]. Male pigs were typically dispatched younger than females. While some
pork was consumed fresh, much was cured by salting for later use [17].
2.3. Isotopic ecology
Stable carbon (δ13C) and nitrogen (δ15N) isotope composition of collagen extracted from archaeological
bone can tell us what animals ate. Moreover, to the extent that the abundance of specific foods may
differ across space and time, animals’ bone collagen δ13C and δ15N compositions can offer clues about
where they lived. The core concept is that the ways in which carbon and nitrogen are sourced and
cycled across environments create distinctive isotopic signatures in select foods, which are then passed
on to their consumers. In turn, patterns in these signatures across animal bone assemblages can serve
as indicators for environmental and cultural processes at a wide range of scales [21,22].
While we review key dimensions of these isotopic indicators as they become relevant throughout
§4, it is worth outlining some of the main axes of isotopic variation in Ireland. As omnivores and
opportunists, interpretation of pig isotopic compositions can be highly complex, and assessment
of individual pig isotopic compositions will almost always come with caveats about potential for
equivocality in interpretations. By contrast, when data from pigs are aggregated across larger site-,
community- and regional-scale assemblages, robust patterns can be discerned.
Ireland is a C3-dominated environment, and hence δ13C variation is expected to occur across a
relatively narrow spectrum [23], in contrast to the kinds of larger-scale isotopic differences expected
from mixed consumption of both C3 and C4 plants [24]. As outlined in table 1, variables include the
canopy effect (i.e. from feeding in denser woodlands [26]), consumption of photosynthetic (e.g. leaves)
versus non-photosynthetic (e.g. mast, grain) plant tissues [29], trophic level [34] and use of aquatic
resources including lacustrine, riverine and wetland habitats [37]. Given that trophic enrichment is
small for δ13C (see §6) and that the vast majority of our sample comes from places where marine foods
would not be abundant, we expect δ13C variation in our dataset to be governed by the canopy effect
and consumption of non-photosynthetic plant foods (e.g. mast, grain).
Previous work by Hamilton et al. [39] has highlighted the potential relevance of fungus consump-
tion for explaining variation observed in pig δ13C values, but we do not believe this offers a more
realistic explanation compared to mast/grain feeding. Hamilton et al.’s [39] work proposing a potential
importance of fungi for pig δ13C values was aimed at comprehending an apparent paradox in isotopic
patterns observed across the British Neolithic and Iron Age. Namely, why pigs that, based on context,
should have been consuming more mast in woodlands, habitats which are generally associated with
low δ13C values (table 1), seem to have high δ13C values. What has since become clear is that mast
consumption will be associated with higher, not lower, δ13C as it is a non-photosynthetic, and thus
13C-enriched, plant tissue (for a review, see [27]). In that context, the need for a woodland-associated
source of 13C-enriched foods (i.e. fungi) to explain these paradoxical isotopic patterns is met by the very
thing (mast) that was expected to be driving pig husbandry into woodlands.
Pig δ15N compositions are governed primarily by trophic position and baseline variation [40]. In
contrast to δ13C, δ15N has a large trophic enrichment factor (TEF) meaning that it undergoes a substan-
tial shift (ca +3.6‰) at each trophic step [34]. For this reason, pigs eating some animal protein (e.g.
meat, dairy, or even human faeces) will have higher δ15N on average than their herbivorous counter-
parts. It also means that particularly young pigs that are still consuming, or have recently been weaned
from sow’s milk will have elevated δ15N relative to adults in the same population [41]. As aquatic
environments have additional trophic levels (i.e. carnivores that eat other carnivores), consumption of
certain fish can also result in very high δ15N values [42]. In addition to these TEF-related variables,
there is considerable potential for variation in baseline δ15N. Specifically, nitrogen cycling and sourcing
processes at the plant–soil level, which are impacted by diverse human (e.g. manuring, stocking rates)
and non-human phenomena (e.g. mycorrhizal changes), change the baseline starting point from which
δ15N trophic enrichment occurs in both terrestrial and aquatic environments (for reviews, see [35,40]).
Upward baseline shifts have been observed in Irish prehistory and are thought to reflect increasing
human land management impacts [43]. It is therefore essential to consider the baseline when interpret-
ing pig δ15N.
4
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

Lastly, it is worth highlighting that, due to the complexity of carbon and nitrogen sources and
cycling in wetlands, animal husbandry practices that make systematic and sustained use of bogs, fens
and marshes could contribute to increased isotopic variation for pigs and cattle in our dataset (for
reviews, see [35,44]). This would particularly be the case for use of food webs that were based on
submerged aquatic vegetation (e.g. phytoplankton, macroalgae etc.), which can produce foods with
isotopic compositions conceivably spanning a range of >25‰ for both δ13C and δ15N across short
distances within a single aquatic system [35]. This heightened potential for variation occurs because
submerged aquatic vegetation must rely on carbon and nitrogen sourced from the water column,
which is in turn governed by a varied range of (often spatially heterogeneous) cycling processes.
Emergent aquatic vegetation, by contrast, sources carbon mainly from the atmosphere and will have
much less δ13C variation (i.e. similar to terrestrial plants). Recent work has suggested that flooding
may affect isotopic compositions of willow trees, potentially driving their δ13C upwards or downwards
under conditions of shorter- and longer-term flooding (with extreme flooding altering foliar δ13C by
as much as −1.2 ± 0.5‰ on average [36]). More work is needed, however, to establish whether this
also occurs in plants that have evolved adaptations requiring freshwater wetland conditions. It is also
worth pointing out that the anoxic conditions established by permanent waterlogging can influence
Table 1. Some key axes of carbon isotope variation for Irish omnivores.
axis δ13C detailed description of axis in context of Ireland
denser canopy
versus open
air
↕canopy cover can lead to lower δ13C in animals feeding near ground level in closed woodland areas
relative to those feeding in open pasture lands [25]. Causes include: (1) slower air movement/
exchange under canopies, which allows woodland foliage to incorporate more 13C-depleted CO2
(released from decomposing vegetation in soils) and (2) influence of lower light levels on
understory foliage, which impacts ability to discriminate against 13C [26]. Based on a large sample
of archaeological Irish cattle, δ13C variation associated with this open–closed canopy feeding
spectrum spans ca <3‰ [23]
leaves versus
nuts/seeds
↕consumption of non-photosynthetic plant tissues such as mast (tree nuts) can lead to higher δ13C in
animals feeding on pannage compared to those feeding on vegetation [27, 28]. This is caused
by an isotopic fractionation that occurs as nutrients are remobilized in the process of forming
different portions of plants’ tissues, such that non-photosynthetic structures (e.g. nuts, seeds)
become 13C-enriched relative to photosynthetic structures (e.g. leaves, stems) [29]. While the
isotopic compositions of mast are expected to vary based on a wide range of environmental
factors, including the canopy effect [30], based on archaeological and contemporary observations,
a spectrum of no-mast to intensive-mast feeding could result in a δ13C range of ca 4‰ [31, 32].
Cereals and grains should also result in similarly elevated δ13C
animal protein ↑consumption of animal protein can lead to higher δ13C in omnivorous and carnivorous animals [33].
This is caused by processes occurring as amino acids are synthesized, processed, and assimilated
from diet to consumer tissues, which, at the whole-protein level (i.e. collagen) selectively retain
13C-enriched building blocks. Compared to trophic enrichment for δ15N, this effect is small, and is on
average thought to be ca 0.5‰ per trophic level for δ13C (though variation is known; see [34])
aquatic foods ↕consumption of aquatic foods, including those from marine, lacustrine, riverine, and wetland habitats,
can drive animal δ13C in both directions (for a review, see [35]). Plants growing in flooded areas
draw CO2 from an isotopically homogeneous source (the atmosphere), but may show greater
isotopic variation based on water availability, including lower δ13C due to waterlogging [36]. More
variation can be expected for submerged primary production, which draws CO2 from the water
column. These effects stem from differences in the way carbon is cycled in aquatic, relative to
terrestrial, systems. Marine foods of animal or plant origin are generally thought to be 13C-enriched
because the primary CO2 source in marine environment has more 13C than the source used by
terrestrial plants [37]. By contrast, freshwater carbon sources can incorporate CO2 from a wide
range of isotopically distinctive materials. This can include CO2 respired from breakdown of
terrestrial vegetation, resulting in extremely low δ13C for terrestrial consumers of freshwater fish
in some regions [38]. By contrast, carbon cycling in some areas can serve to increase baseline
freshwater δ13C and to reduce primary producer discrimination against 13C, leading to high δ13C
overlapping with most of the marine range [35]. Overall, a potential δ13C range for marine and
freshwater foods in Ireland could be as wide as 30‰
5
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

the nitrogen cycling, causing emergent aquatic vegetation to have higher δ15N values [45]. While use
of wetland resources from food webs based on either submerged or emergent aquatic vegetation will
remain a potential source of variation, we do not expect that this will be a primary driver of variation
in pig or cattle isotopic compositions in this study. This is because the broad scale, aggregating
approach we take should obviate potential influences from occasional and more spatially localized
instances of intensive wetland use.
2.4. Isotopic work on Irish pigs
Previous interpretations of archaeological pigs in Ireland have begun to sketch out a narrative for
broader trends in husbandry. Most direct interpretations have focused on the medieval period and are
limited in nature because the studies from which they were drawn were focused on, and used pigs as a
baseline for interpreting human diet and mobility. Knudson et al. [46] noted that isotopic compositions
of Hiberno-Norse pigs (n = 12) in County Dublin do not indicate marine protein consumption. Ryan
et al. [47] noted that isotopic compositions from early medieval-period pigs (n = 7) in County Meath
could reflect variation in life stage, location and animal-management practices. McKenzie et al. [48]
noted that a later medieval pig from County Donegal had an isotopic composition that could reflect an
omnivorous diet. Work by Madgwick et al. [49] and Guiry et al. [23] offer interpretations that are more
animal-oriented and are therefore an exception to this human-focused trend. Madgwick et al. [49],
examining 19 pigs from Iron Age Navan Fort in County Armagh, found that while pigs were mainly
herbivorous, some may have been omnivorous. Exploring isotopic variation among 178 pigs and wild
boar from sites across Ireland, spanning the later Holocene (including many from the samples used
in this study), Guiry et al. [43] showed that pig δ15N increased through time across the island. While
this trend was thought to reflect increasing baseline δ15N (i.e. at the plant–soil level), higher δ15N in
pigs from later periods was recognized as evidence for consumption of animal protein. Building on this
existing work, in this paper we interpret data from 238 pigs from 40 sites. While some of these data
have been sourced from the literature, 88% of δ13C (n = 209) and 40% of δ15N (n = 93) values are new
(see §3). We use the large body of archaeological herbivore data available from Ireland [23,43,48,49]
to correct for baseline variation. Specifically, we use mean values from 344 cattle from 49 widely
distributed sites (figure 1) aggregated by time period (and site, where sample sizes allow; see §6) to
establish what herbivorous δ15N should look like. We then compare pigs to these baseline data to
evaluate the extent to which they were herbivores or omnivores (see §6).
2.5. Hypotheses and research questions
Based on previous isotopic and zooarchaeological work, as well as information from the historical
record, we can structure the following research questions and hypotheses.
(1) To what extent did pannage and animal products feature in medieval pig diets?
Hypothesis. If historical narratives indicating the dual importance of animal protein1 and mast
for pig husbandry [12] are correct, we will see relatively high δ15N and δ13C values on average
in pigs from that timeframe. This could include grain consumption, which, like mast, will result
in elevated δ13C values relative to most other foods (i.e. foods derived from photosynthetic
plant tissues or products of animals eating these foods). Existing isotopic evidence shows some
support for the importance of animal protein in pig diets evidenced in the early documentary
sources (especially dairy byproducts, such as whey) [43], though baseline variation has not been
corrected for in past work, nor has the relevance of pannage and grain been considered.
(2) Looking further back in time, to what extent does the dual-diet approach to pig husbandry (i.e.
animal protein and pannage) that is featured in the early medieval period literature, characterize
pig husbandry practices in earlier periods?
Hypothesis. If pig husbandry in Ireland has always involved feeding pigs waste generated from
human settlements and other activities and driving pigs to pannage, then we will see a similar
pattern to that expected for the medieval period, involving higher δ15N and δ13C values relative
to contemporaneous herbivorous livestock. While existing isotopic evidence suggests that at
1Human-generated animal protein may include scraps and waste from human meals and or byproducts from hunting, agricultural
activities or industrial processes or human or domestic animal excrement. From the perspective of impacts on pig bone collagen
δ15N, all of these potential foods falling under the umbrella of animal protein are interchangeable, serving to drive up δ15N.
6
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

least some pigs were more herbivorous in the prehistoric timeframe, interpretation of relative
omnivory has been limited by baseline variation. Moreover, the pool of published δ13C data
available for prehistoric pigs has been too small to consider trends in pannage.
3. Results
We analysed 302 pig specimens from 38 sites, of which 209 samples from 34 sites produced isotopic
compositions associated with viable quality control metrics (electronic supplementary material, table
S6). All 209 δ13C values are new whereas 93 of these δ15N values are new and 116 are previously
published [43]. Of the samples failing quality control, most (95%) were ultrafiltered rather than NaOH
treated (see §6). We exclude a further four samples from interpretations based on issues of low
chronological resolution or osteological indicators suggesting samples are from younger pigs, which
could have δ15N values influenced by milk suckling [41]. To this we can add previously published
isotopic compositions from 33 pigs from seven sites [46,48,49] giving a total of 238 pigs from 40 sites
(all data listed in electronic supplementary material, table S6). Data are visualized in figure 2 and
means and s.d. for all periods are shown in table 2.
The Neolithic–Early Bronze Age sample includes isotopic compositions from two specimens
identified by the original zooarchaeologists as wild boar [43,50]. We note that these samples fall within
the main clustering of data (electronic supplementary material, figure S2), and their exclusion does not
change Neolithic–Early Bronze Age sample means significantly (table 2). Due to the small size of our
sample from this early period, we include data from these specimens in the Neolithic–Early Bronze
Age group in the following comparisons.
During sampling, bones were identified and recorded by an experienced zooarchaeologist. Due to
specimen fragmentation, which prevented direct observation of age-related morphological indicators,
it was not always possible to establish age based on fusion; however, care was taken to exclude
samples from small and therefore young juvenile pigs. Given the relatively short window of milk
suckling for piglets, which naturally wean by 17 weeks and may be artificially weaned before this
[51], and the efforts we have made to exclude samples from this age category, we do not expect that
patterns in pig isotopic variation are driven by this factor. Two pig specimens, both from historical
contexts at the coastal site of Greencastle, County Down [52], show isotopic compositions consistent
Figure 2. Isotopic compositions for pigs grouped by time period shown in bivariate space (a), with violin and box plots showing kernel
density for δ15N for raw (b) and baseline-corrected (c) δ15N. For (b,c): Neo EBA = Neolithic and Early Bronze Age; LBA = Middle Late
Bronze Age; IA = Iron Age; EM = early medieval; LM = later medieval; PM = post-medieval. For (c), baselines for all later periods have
been corrected to the earliest (Neolithic) period (see §3 for description of methods).
7
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

Table 2. Means and s.d. for pigs as well as statistical and means comparisons of pigs with cattle from respective time periods and sites (where n > 5 for both cattle and pigs). For raw data and sources, see electronic
supplementary material, table S6. Statistically significant differences are shown in bold font. For full statistical results, including normality and variances tests as well as means comparisons, see electronic supplementary
material, table S7.
grouping pig cattle δ13C comp. δ15N comp.
n = sites δ13C (‰) δ15N (‰) n = sites δ13C (‰) δ15N (‰) pig–cattle p = pig–cattle p =
timeframe comparisons
Neolithic–Early Bronze Age 16 7 −21.7 ± 0.7 5.2 ± 0.5 36 13 −22.4 ± 0.6 5.1 ± 0.7 0.62 0.001 0.08 0.709
Neolithic–Early Bronze Age without boar 14 5 −21.6 ± 0.8 5.2 ± 0.5 —
Middle–Late Bronze Age 28 5 −22.2 ± 0.5 6.5 ± 1.1 70 12 −22.8 ± 0.5 6.3 ± 0.6 0.60 <0.001 0.21 0.001
Iron Age 91 6 −22.3 ± 0.4 6.4 ± 0.9 44 6 −22.4 ± 0.4 6.4 ± 0.9 0.19 0.009 −0.02 0.894
Middle–Late Bronze and Iron Age 119 11 −22.2 ± 0.4 6.4 ± 0.9 114 18 −22.7 ± 0.5 6.3 ± 0.8 0.44 <0.001 0.07 0.961
early medieval 42 6 −21.9 ± 0.4 8.5 ± 1.2 71 7 −22.2 ± 0.4 7.1 ± 1.0 0.29 0.001 1.42 <0.001
later medieval 43 10 −21.4 ± 0.4 7.6 ± 1.1 60 7 −22.4 ± 0.5 6.3 ± 1.3 1.01 <0.001 1.32 <0.001
post-medieval 18 8 −21.9 ± 0.5 8.3 ± 1.7 63 15 −22.4 ± 0.5 6.4 ± 1.1 0.49 <0.001 1.84 <0.001
medieval and post-medieval 103 20 −21.7 ± 0.5 8.1 ± 1.3 194 22 −22.3 ± 0.5 6.6 ± 1.2 0.63 <0.001 1.47 <0.001
intra-site comparisons
Ross Island (Early Bronze Age) 5 — −21.4 ± 1.0 5.0 ± 0.6 8 — −22.3 ± 0.5 5.3 ± 0.5 0.95 0.040 −0.37 0.273
Haughey's Fort (Middle–Late Bronze Age) 14 — −22.3 ± 0.5 5.9 ± 0.7 51 — −22.9 ± 0.5 6.3 ± 0.7 0.59 <0.001 −0.40 0.062
Dún Ailinne (Iron Age) 15 — −22.2 ± 0.4 6.3 ± 1.0 15 — −22.2 ± 0.4 6.8 ± 0.9 0.02 0.897 −0.55 0.135
Navan Fort (Iron Age) 51 — −22.2 ± 0.3 6.3 ± 0.7 20 — −22.6 ± 0.4 6.1 ± 0.9 0.33 0.001 0.16 0.428
Mountgorry (early medieval) 9 — −22.0 ± 0.2 8.0 ± 1.2 16 — −22.2 ± 0.3 6.5 ± 0.9 0.27 0.022 1.50 0.002
Navan Road (early medieval) 6 — −21.8 ± 0.3 9.1 ± 1.0 18 — −22.1 ± 0.4 7.4 ± 0.7 0.29 0.101 1.70 <0.001
Ratoath (early medieval) 12 — −21.9 ± 0.4 9.3 ± 1.2 15 — −22.3 ± 0.4 7.9 ± 1.0 0.39 0.025 1.34 0.003
Stalleen (early medieval) 12 — −21.8 ± 0.3 7.9 ± 0.6 15 — −21.9 ± 0.2 6.4 ± 0.4 0.04 0.660 1.53 <0.001
Greencastle (later medieval) 5 — −21.5 ± 0.3 7.8 ± 1.1 23 — −22.5 ± 0.4 5.6 ± 0.8 0.99 <0.001 2.18 <0.001
Nobber (later medieval) 17 — −21.5 ± 0.2 7.7 ± 0.8 9 — −22.6 ± 0.6 6.1 ± 1.3 1.10 0.001 1.55 0.001
Eyre Square (post-medieval) 6 — −21.8 ± 0.5 8.0 ± 1.4 11 — −22.2 ± 0.4 6.2 ± 1.2 0.40 0.024 1.84 0.010
8
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

with some consumption of marine foods. While these samples do not pass the conservative quality
control criteria we use in this study [53], they would pass liberal quality control (see §6) and we offer
brief interpretations in electronic supplementary material, text S1 and figure S3.
Collectively pigs show considerable δ13C and δ15N variation (figure 2). To establish how much of
this variation is driven by baseline versus dietary change, we compared mean pig with cattle isotopic
compositions grouped by time period (table 2). During the Neolithic–Early Bronze Age (figure 3b,c),
Middle–Late Bronze Age (figure 4b,c) and Iron Age (figure 5b,c), mean pig δ15N values are close to
those of cattle from their respective time periods (means differ by 0.0–0.2‰), with no significant
differences (table 2). This suggests that during these prehistoric periods, pigs were largely herbivorous.
By contrast, during the early (figure 6b,c), later (figure 7b,c) and post-medieval (figure 8b,c) periods,
mean pig δ15N was substantially higher than that of cattle (means differ by 1.3–1.8‰) and consistently
shows significant differences (table 2). This suggests that during the historic period, pigs were largely
omnivorous (see §4). With respect to δ13C, across all temporal comparisons (figures 3–8, panels (a)
and (b)), mean pig δ13C values were consistently elevated (by 0.2–0.9‰) above, and were in all cases
significantly different from, cattle from their respective time periods (table 2). These interspecific δ13C
differences indicate that, relative to cattle, pig diets were on average focused on more 13C-rich foods
(for consideration of interspecific differences in digestive physiology, see §6).
To further verify these interspecific time-period-level patterns, we also perform the same compari-
sons between means for isotopic compositions of cattle and pigs at the same sites, using data from
all sites with at least five cattle and five pigs from one time period (table 2). Such comparisons could
be made at one Neolithic–Early Bronze Age, one Middle–Late Bronze Age, two Iron Age, four early
medieval, two later medieval and one post-medieval site (figures 3–8 panels (b) and (c)). Of these
22 comparisons, nearly all (11 of 11 for δ15N and 8 of 11 for δ13C) conform to the patterns observed
at the time-period level in terms of both order of means (e.g. where pigs had higher means than
cattle) and statistical significance (table 2). With respect to the three exceptions (Dún Ailinne, Ratoath,
Navan Inner Relief Road 1–3), in all instances pigs still produced mean δ13C values that were higher
than their respective cattle group, but these differences were not statistically significant (table 2).
These interspecific site-level comparisons, therefore, broadly confirm and support the picture we see
Figure 3. Isotopic compositions of Neolithic and Early Bronze Age pigs and cattle shown in bivariate space (b) and as violin and
box plots showing kernel density for δ13C (a) and δ15N (c). Data are shown by site for sites with isotopic compositions for at least
five individuals from each taxon. For (b), symbols are grey for sites with fewer than five individuals from each taxon. For statistical
significances, see table 2.
9
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

at the time-period level. Moreover, the fact that these sites, located across several counties in Ireland
(electronic supplementary material, figure S1), show isotopic patterns that correspond with those of
their broader-scale temporal period, support our expectation, based on previous δ15N baseline work
[43], that isotopic patterns are driven more by temporal than geographical factors. We can also note
that these sites include a range of socio-economic situations, including sites that are more rural and
more urban, and sites of higher and lower status (see electronic supplementary material, table S1),
suggesting that patterns transcend these site type categories (see §4).
For the Iron Age, there is a particularly large sample of pigs, which is >50% composed of specimens
from a single site, Navan Fort. To establish that temporal patterns in pig–cattle comparisons for this
period were not driven by relatively localized phenomena influencing pig diets at Navan Fort, we
also compare pigs from Navan Fort (n = 51, mean δ13C = −22.2 ± 0.3‰, Shapiro–Wilk W = 0.945, p =
0.020; mean δ15n = +6.3 ± 0.7‰, Shapiro–Wilk W = 0.984, p = 0.707) to the rest of the sample of pigs
from the Iron Age period, which are sufficiently similar to be pooled (n = 40 from five sites, mean
δ13C = −22.3 ± 0.4‰, Shapiro–Wilk W = 0.983, p = 0.798; mean δ15n = +6.5 ± 1.1‰, Shapiro–Wilk W =
0.947, p = 0.060). No significant differences were observed for either δ13C (Mann–Whitney U = 954, p
= 0.600; Vargha–Delaney A = 0.532) or δ15N (Mann–Whitney U = 957, p = 0.885; Vargha–Delaney A =
0.482). This comparison suggests that the relatively large number of samples from Navan Fort has not
disproportionately influenced Iron Age isotopic patterns.
To evaluate the significance of chronological change in levels of pig omnivory across time we
also compared baseline-corrected mean δ15N between time periods (compare figure 2b,c). This was
accomplished by sequentially comparing mean pig δ15N between adjacent time periods but, for each
comparison, we adjusted the latter temporal groups’ δ15N values according to the baseline offset
observed between groups of cattle from those respective time periods (electronic supplementary
material, tables S8 and S9; see §6 for more detail). Differences in baseline-corrected mean pig δ15N
between time periods were all small (within 0.5‰) except for during the Iron Age to early medieval
transition, which showed an upward δ15N shift of 1.4‰ after baseline adjustment (figure 2c; electronic
supplementary material, table S9). Statistical significances for these comparisons follow the same
Figure 4. Isotopic compositions of Middle–Late Bronze Age pigs and cattle shown in bivariate space (b) and as violin and box plots
showing kernel density for δ13C (a) and δ15N (c). Data are shown by site for sites with isotopic compositions for at least five individuals
from each taxon. For (b), symbols are grey for sites with fewer than five individuals from each taxon. For statistical significances, see
table 2.
10
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

pattern, with no significant differences (i.e. p ≤ 0.05) observed except for the comparison between the
Iron Age and the early medieval groups (where p ≤ 0.001; electronic supplementary material, table
S9). Together these comparisons support the trends observed in cattle–pig comparisons between time
periods. Specifically, pig δ15N consistently shows a herbivorous diet from the Neolithic until the end
of the Iron Age. Pig δ15N then underwent an upward shift reflecting a substantial increase in intake
of animal protein in the early medieval period, after which time this meaningful level of omnivory
remained relatively consistent until the final, post-medieval timeframe.
Lastly, given the complexities of discerning the relative impact of the two major sources of
δ13C variation (i.e. the canopy effect versus mast/grain consumption; see table 1 and §4), we
compare δ13C means from pigs grouped by time period to cattle from the early medieval period
[23]. The high δ13C observed in cattle during the early medieval is thought to reflect the zenith
for open-land pasturing (for consideration of climatic factors on δ13C, see §6), and provides a
convenient benchmark beyond which mast consumption is more likely. In nearly all cases, pig δ13C
means were higher than those of early medieval cattle. The exception being Iron Age pigs, with a
mean δ13C that is 0.1‰ lower than that of cattle (see electronic supplementary material, table S10,
for results). Differences were significant for comparisons involving Neolithic–Early Bronze Age,
later medieval, and post-medieval pigs, but not for the Middle–Late Bronze Age and Iron Age
comparisons (electronic supplementary material, table S10).
Considering the higher-trophic-level nature of pigs from the early, later and post-medieval periods,
we compared the δ13C means of these groups, adjusted for trophic 13C enrichment, to cattle from
the early medieval period (electronic supplementary material, table S10; see §6). Pig trophic level
(TLpig) for the early, later, and post-medieval periods ranged from 0.4 to 0.5 (electronic supplementary
material, table S10; see §6). In all cases, mean pig δ13C was higher than that of early medieval cattle
(by 0.2–0.3‰), but the difference was only significant for the later medieval comparison (electronic
supplementary material, table S10).
Figure 5. Isotopic compositions of Iron Age pigs and cattle shown in bivariate space (b) and as violin and box plots showing kernel
density for δ13C (a)and δ15N (c). Data are shown by site for sites with isotopic compositions for at least five individuals from each taxon.
For (b), symbols are grey for sites with fewer than five individuals from each taxon. For statistical significances, see table 2.
11
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

4. Discussion
Results from 238 specimens show considerable variation in pig isotopic compositions across the
later Irish Holocene (ca 4000 BC to ca AD 1900; see figures 2c and 6–8a for baseline-corrected data).
Considering pig isotopic compositions relative to published cattle data, which serve as a faunal
baseline (see §6), two patterns emerge. On the one hand, a striking and comparatively rapid change
in pig trophic behaviour occurs at the transition from the Iron Age to the early medieval period. On
the other hand, there appears to be 6000 years of continuity in use of mast from woodlands, the same
resource that we know from historical records [12] had become synonymous with pig husbandry by
the early medieval period. We will consider the isotopic evidence for each of these processes, and their
attendant implications, in turn.
4.1. Pigs and proximity: nitrogen isotope compositions show change
One starting point for considering change in pig diets is the diet of wild boar. As noted above,
these are naturally woodland dwelling and generally herbivorous, subsisting on mast in the autumn/
winter and leaves, seeds, fruit and roots throughout the year. Limited animal protein in their diet
may include carrion, young mammals and birds, and eggs [5]. This ecological niche is evidenced by
two zooarchaeologically identified wild boar samples (electronic supplementary material, figure S2)
showing the expected predominantly herbivorous diets. These results are isotopically similar to those
of zooarchaeologically identified domestic pigs at that time. This similarity suggests that Neolithic and
Early Bronze Age pig management may have allowed pigs to mimic their wild counterparts.
Figure 6. Isotopic compositions of early medieval pigs and cattle shown in bivariate space (b) and as violin and box plots showing
kernel density for δ13C (a) and δ15N (c). Data are shown by site for sites with isotopic compositions for at least five individuals from
each taxon. For (b), symbols are grey for sites with fewer than five individuals from each taxon. For (a), a second violin plot for early
medieval pigs grouped by time period (‘TL corrected’) shows trophic-level adjusted δ13C (per electronic supplementary material, table
S10). For statistical significances, see table 2.
12
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

This approach to pig management continued for millennia. Binned by time period and corrected
for baseline changes, mean δ15N values (figure 2c) clearly show that during the Neolithic–Early Bronze
Age, Middle–Late Bronze Age and Iron Age, pigs were largely herbivorous. By contrast, pigs raised
during the early, later and post-medieval periods were omnivores with considerable animal-protein
intake (potentially including both meat and dairy; ranging from TL 0.4 to 0.5, where 0 is fully
herbivorous, and 1 is fully carnivorous; electronic supplementary material, table S10).
Before exploring broader δ15N trends as evidence for changing Irish pig husbandry practices, it
is worth pointing out that, at the individual specimen level, there are likely some exceptions to the
overall pattern of prehistoric pig herbivory, particularly during the Middle–Late Bronze and Iron Ages.
Previous work has documented δ15N baseline shifts, which began to move dramatically sometime in
the Middle–Late Bronze Age and onwards [43]. In aggregate, at the island-wide scale, these baseline
changes appear to occur in unison. However, we expect that if sufficient data were to become available
to examine how these changes unfolded over smaller segments of space and time we would see
a heterogeneous patchwork of isotopic compositions with changes happening at different rates in
different places centring on the later Bronze Age. In that context, moving from the aggregate time-
period level to assess diet at the individual level, it becomes difficult to evaluate the extent to which
isotopic compositions from specific pigs represent higher trophic positions (i.e. more omnivorous
diets) rather than fluctuations in 15N content of local plants and other food sources, such as those which
can arise from changing cultivation and other land management processes [40]. However, considering
the particularly high δ15N values for some specimens from the Middle–Late Bronze Age (e.g. samples
with δ15N > ca 8.0‰, n = 3 of 28, electronic supplementary material, table S6, figure 4b, some of which
were definitely adults) it seems likely that at least some pigs were consuming diets rich in animal
protein at this time. Nonetheless, at scale, and compared with the early medieval, the evidence we
have generated for later prehistoric pigs still suggests that this was a comparatively rare husbandry
pathway at that time. In that context, it is worth noting that a very small (+0.2‰) but significant
elevation was observed in mean δ15N of pigs relative to cattle during the Middle–Late Bronze Age
Figure 7. Isotopic compositions of later medieval pigs and cattle shown in bivariate space (b) and as violin and box plots showing
kernel density for δ13C (a) and δ15N (c). Data are shown by site for sites with isotopic compositions for at least five individuals from
each taxon. For (b), symbols are grey for sites with fewer than five individuals from each taxon. For (a), a second violin plot for later
medieval pigs grouped by time period (‘TL corrected’) shows trophic-level adjusted δ13C (per electronic supplementary material, table
S10). For statistical significances, see table 2.
13
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

(table 2) suggesting that pigs in this time period could have consumed slightly more animal protein
on average than would be expected for herbivores. However, this 15N enrichment is small enough
(representing only ca 0.05 TL, or about 5% of a trophic level step from herbivore to carnivore) that it
remains inconsequential for our interpretations. We further note that during the Iron Age, pigs and
cattle produced virtually identical mean δ15N values (table 2). On the whole, it is therefore clear that,
like their Neolithic and Early Bronze Age counterparts, later prehistoric pig keepers did not take pains
to involve substantial amounts of animal protein in pig diets.
To explore the temporal shift observed in mean pig δ15N (figure 2c, table 2), we consider the
implications from two perspectives. First, we reflect on what these patterns mean for our understand-
ing of animal husbandry before the early medieval. Second, we ask what structural changes must have
occurred at the early medieval juncture and subsequent periods to generate these patterns.
In order to unpack what these isotopic patterns mean for our understanding of prehistoric pig diets,
it is best to start in the middle of the story, and ask why it was that pigs would have begun consuming
large quantities of animal protein and what processes drove that change. It is widely held that pigs will
eat virtually any food, but among this almost limitless potential menu, sources of animal protein are
particularly valuable from the perspective of both the pigs and their associated keepers. We know, for
instance, that, granted access, pigs will consume meat and other animal products preferentially [54,55].
Because pigs will readily consume animal products, and feeding animal-protein-rich food to pigs helps
increase weight gain [55,56], it is likely that knowledge of the value of adding animal protein from
waste products of human activity to pigs’ diets is of nearly the same antiquity as pig husbandry itself.
In that context, if we accept as our baseline premise that feeding edible animal-protein-rich waste
from human activities to pigs has always represented a potentially lucrative pathway to generating
more pork, it is reasonable to ask why pigs husbanded in Ireland prior to the early medieval evidently
consumed so little animal protein. Taking a hypothetical approach, one could hypothesize that there
was less human-generated animal protein to go around and, therefore, that on average very little
animal protein could be diverted from human, or indeed other domesticate (i.e. dog), use towards
raising pigs. In turn, this would suggest that sources of human-generated animal protein (i.e. livestock
agriculture and hunting and fishing activities) were less productive during prehistory. Against the
Figure 8. Isotopic compositions of post-medieval pigs and cattle shown in bivariate space (b) and as violin and box plots showing
kernel density for δ13C (a) and δ15N (c). Data are shown by site for sites with isotopic compositions for at least five individuals from
each taxon. For (b), symbols are grey for sites with fewer than five individuals from each taxon. For (a), a second violin plot for
post-medieval pigs grouped by time period (‘TL corrected’) shows trophic-level adjusted δ13C (per electronic supplementary material,
table S10). For statistical significances, see table 2.
14
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

wider corpus of evidence from prehistoric agricultural practice [57], however, it seems implausible
to suggest that, on average, areas of the landscape with denser prehistoric human populations (e.g.
settlements, farmsteads) would not have generated meaningful amounts of edible animal-protein-rich
waste. In that context, an alternative approach is to ask why prehistoric people would apparently
ignore this special, and arguably one of the most economically valuable, dimensions of pig husban-
dry—their capacity to convert human-generated sources of waste, especially those that are rich in
animal protein, into more food for humans [56].
We note that cultural factors could offer one possible explanation. For instance, it is conceivable that
the consumption of animals may have been seen as culturally inappropriate for pigs, perhaps because
this behaviour would make pigs too close to humans. These possibilities, however, remain speculative
and, while the limited nature of existing evidence for prehistoric human–pig relationships in Ireland
[13] means that this potential issue cannot be adequately assessed, we assume that it does not offer
a more parsimonious explanation than scenarios in which pigs living near humans could access a
representative selection of human-generated food scraps.
With the backdrop of these questions and the apparent interpretive paradox to which they give
rise (i.e. why prehistoric people did not direct available edible animal-protein waste to raising
pigs), we can offer a more parsimonious explanation—that in the context of prevailing prehistoric
settlement patterns, environmental resources and pig ecology, it made less sense to prehistoric
farmers to husband pigs in ways that involved comparatively animal-protein-rich foods. More
specifically, our data may reflect a scenario in which prehistoric pig husbandry systems in
Ireland were more often centred (for most of the year anyhow) away from areas with denser
human populations. For our purposes, areas of denser human population would be settlement
locations associated with a sustained availability of human-generated animal protein (i.e. food
scraps, agricultural byproducts, human excrement), such as small concentrations of farms or even
individual farmsteads. Reasons for this may have been manifold, involving rationales embedded
in seamless human–animal–environmental worldviews that, today, we might recognize as being
not only economic, but social, cultural and religious in nature. While we have little concrete
evidence with which to explore the latter, we can consider some of these processes in terms of the
opportunities they offered farmers to generate larger numbers of pigs and thus more pork.
In that context, we can hypothesize both ‘pull’ and ‘push’ factors that could have served to guide
pig husbandry locations away from where more people lived. We recognize that describing ‘pull’
factors in this way presupposes that pig husbandry has a default setting in which pigs will be
husbanded close to people. Indeed, as our data suggest, this may not have been the case in the distant
past and, instead, the ‘default’ approach to early pig husbandry could be management of pigs in
locations away from denser human areas. After all, these woodlands were the same habitats preferred
by wild boars, making them ecologically well suited for domestic pigs as well. Acknowledging this
issue of presentism, we proceed using these terms to frame our interpretation as the ‘push–pull’
spectrum serves to neatly illustrate potential trends.
With respect to factors that could have pulled pig husbandry away from sources of human-gen-
erated waste, it may have been that woodlands and other ecosystems further afield were consid-
ered more productive habitats for raising pigs than locations such as settlements or other areas
with relatively dense human populations that were richer in human-generated animal protein. In a
scenario where highly productive ecosystems for pigs were abundant and widely accessible, commu-
nal husbandry (via swine herders) could have offered a better approach to pork production than
more blended systems that were traditional by the early medieval period (e.g. involving rearing of
pigs at individual farmsteads, ringforts, for a large portion of the year, followed by communal swine
herding during the masting season; see below). An alternative, though not mutually exclusive factor
could be that prevailing settlement patterns were less conducive to keeping pigs on farms or in
settlements during the prehistoric period. We know comparatively little about settlement patterns in
Ireland in later prehistory (for a review, see [58]), but it is conceivable that they were organized in
such a way that keeping pigs closer to human habitation would have represented a liability (relative
to later periods) and that this served as a ‘push’ factor. Such liabilities could stem from differences in
defensibility, mobility or care and management needs of pigs relative to other more important foci of
activities in areas with denser human groups (i.e. the places we would expect humans to generate more
animal-protein-rich waste for potential pig feed).
While the prehistoric pig isotopic data alone cannot serve as a basis for weighing the relative
influence of the potential ‘pull’ (of other habitats) or ‘push’ (of potential incongruencies of farm-
based pig husbandry and the other responsibilities faced by prehistoric farmers) factors, additional
15
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

perspective can be gained from taking a longer view. Zooming out to consider the dataset as a whole,
what is abundantly clear is that whichever factor drove the animal-protein-light diets of prehistoric
pig husbandry, the ‘barriers’ to connecting pigs with human-generated animal-protein-rich food waste
had disappeared or been overcome by the early medieval period. Or, put in less presentist terms,
farmers, for one reason or another, chose to husband pigs in a different way that included more animal
protein. The emergence of these new animal husbandry practices with comparatively animal-protein-
rich diets during the early medieval provides a historically referenced anchor point from which to
retrospectively reconsider these potential ‘push’ and ‘pull’ factors in prehistory.
Textual evidence for early medieval Irish agriculture provides a robust framework for fleshing out the
kinds of husbandry practices responsible for generating what, in the context of our dataset, represented
a blending of new and traditional (see below) forms of animal husbandry. As in the prehistoric period,
during the medieval and post-medieval periods, the richest sources of animal protein for pigs would have
been areas of concentrated human activity. Specifically, pigs were raised in connection with a broad range
of ready sources of human-generated animal-protein-based foods, including anything along a spectrum
from household-level waste to byproducts, such as whey, from farmstead-level to quasi-industrial-scale
monastic operations [12,20]. What is most salient for considering patterns in prehistoric husbandry are
sources showing that, on average, medieval pig husbandry also involved a substantial focus on pannage in
woodlands [12] (see also [59,60]). In turn, this suggests that woodland resources, situated at some remove
from human settlements, were still being used for pig husbandry during part of the year. In that context it
is fair to ask, if typical approaches to early medieval pig husbandry still involved the travel and time spent
away from human population centres for pannage (as evidenced in both historic and prehistoric pig δ13C
values; see below), what other changes account for the contrast of pigs spending the remainder of their
time in closer proximity to more densely populated areas (as evidenced by greater access to animal protein
during the historic period)?
The thirteenth-century agriculturist and widely read English scholar Walter of Henley noted that it
was much more economical to rear pigs ‘in the forest, or in woods, or waste, or in marshes’ rather than
at the farm [61]. He continues that:
for whoever will keep swine for a year from the cost of the grange [farm] alone, and count the cost and the
allowance for the swine and swineherd, together with the damage they do yearly to the corn, he shall lose twice
as much as he shall gain, and this will soon be seen by whoever keeps account.
It may therefore be suggested that the change in strategy to farm-based rearing of pigs in early
medieval Ireland was occasioned by necessity rather than choice. Several broad factors could explain
this.
Plunkett [62] notes that there was extensive woodland clearance throughout much of Ireland during
the early medieval period, which, in turn, could have restricted the availability of mast. The documen-
tary evidence also suggests that woodland was not as extensive and tended to be confined to poorer,
marginal land [12], the implication being that woodland on better lands tended to be cleared for
agricultural use. However, there is no evidence for sudden widespread woodland clearance with the
onset of the early medieval. It was a gradual process occurring at different times in different parts of
the country [63]. This decline in woodland alone cannot account for the sudden change in pig diet at
the beginning of the period. Moreover, at least some woodland survived, providing opportunities for
pannage in at least some regions in later periods.
The second possible factor is changes in land ownership and settlement. The documentary evidence
indicates that much of the surviving woodland was privately owned [12], thus potentially restricting
pannage rights and mast availability for many, if not most, farmers. The early laws indicate ‘that
early Irish society attached great importance to the principle of the private ownership of property’
[64]. Indeed Mytum [65] argued that the change from kin-group land ownership, which would have
facilitated more general pannage rights, to private land ownership coincided with the advent of
Christianity in Ireland. A sudden change in settlement also occurs at this time. The settlement type in
Iron Age Ireland was of dispersed rural unenclosed round houses, which was essentially unchanged
since the Bronze Age [66]. The roundhouse tradition continued but in the early medieval period the
house became surrounded by a formidable ditch. The enclosed, defended, privately owned farmsteads
(i.e. ringforts), which number in their tens of thousands, now became the type-site. The law tracts
imply that pig sties and pens for other domesticates were located within the ringforts [12], and the
presence of pigs is demonstrated by an abundance of pig lice (specifically Haematopinus apri) recovered
from a waterlogged ringfort level at Deer Park Farms, County Antrim [67]. Perhaps a change to private
16
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

ownership of land and livestock somehow necessitated the change in strategy of rearing pigs from a
free-range regime of woodland and wasteland to one where, for the majority of the year, they were
reared at the farm, with short periods of pannage in woodland, if available. Indeed, it has been argued
that the ringfort was devised primarily to protect livestock rather than humans [68].
A third factor to be taken into consideration is the increasing importance of cattle in the early
medieval period [23,68]. This would have led to increased availability of milk and associated byprod-
ucts such as whey and buttermilk. These byproducts of the cheese- and butter-making processes
represent potentially valuable sources of animal protein for pig husbandry. While some would have
been used for human consumption, they have been important elements in pig feed up to modern
times. Their availability, therefore, could represent an important ‘pull’ factor, drawing pig husbandry
closer to areas of human habitation.
Having considered processes that may have contributed to husbandry of pigs in closer proximity
to humans in the medieval and post-medieval periods, we can now return to our consideration of
‘push’ and ‘pull’ factors responsible for more distant pig husbandry practices in the prehistoric period.
While there remain questions of the relative scale of different potential pools of pig food resources, the
overall pattern seems to suggest that something more fundamental or structural shifted approaches to
settlement between the prehistoric and early medieval periods, which made husbandry of pigs in or
near human habitations more preferable. In other words, the ‘push’ factors had diminished and the
pull factors (e.g. availability of dairy byproducts; ringforts for livestock defensibility) had increased. In
that context, it could be that changing social, political, and economic forces during the transition from
late prehistory to the early medieval period altered the landscape of risks and benefits associated with
husbandry of pigs in areas further from farmsteads and other centres of agricultural activity. Whatever
the cause, when considered in the light of the early medieval texts, which demonstrate that husbandry
in areas away from settlements continued on a seasonal basis [12], the trends we see in the prehistoric
period suggest that shifting economics and settlement patterns could have been drivers of the changing
approaches to husbandry evidenced in pig δ15N.
4.2. Of mast and meat: carbon isotope compositions show consistency
In contrast to the δ15N evidence, a dietary trend that is consistent across the dataset is that pig
δ13C values are broadly in line with at least some pannage feeding throughout all time periods. We
recognize that opportunities for pannage husbandry would have been diminished as woodlands were
converted to open fields in the medieval and post-medieval periods, but, as outlined in early medie-
val texts, pannage continued where surviving woodlands allowed [12,69,70]. Pigs show significantly
higher mean δ13C relative to cattle in all respective time periods (figures 3–8; table 2). Pig δ13C
means are also higher than those for respective cattle in all intra-site comparisons, most of which
were statistically significant (figures 3–8; table 2). The main levers governing carbon isotope variation
between cattle and pigs differ in ways that make their δ13C values challenging to compare directly (see
above). Moreover, small systematic interspecific differences in how digestive physiology influences
bone collagen isotopic compositions cannot be ruled out, though it is possible that these could serve
to dampen, rather than artificially inflate, the interspecific pattern observed here (see §6) [71]. A result
of these challenges is that patterns definitively identifying pannage will be rare when pig isotopic
compositions do not sit near the extreme positive end of the δ13C range [72]. Cattle diets are, from an
isotopic perspective, constrained relative to pigs. Because cattle are herbivores and cannot consume
large quantities of mast [73,74], at this aggregated and island-wide scale variation in their isotopic
compositions is primarily driven by one factor—the canopy effect—making interpretation of variation
in cattle δ13C relatively straightforward [23]. By contrast, in addition to the canopy effect, δ13C variation
in pigs will be influenced by two other 13C-enriching factors—animal protein (e.g. meat and dairy)
and non-photosynthetic plant tissues (namely mast, but also grains). Each factor requires nuanced
consideration before the question of pannage can be addressed.
First, to assess the impact of mast consumption on δ13C of pigs relative to cattle it is important
to consider the carbon isotopic ecology of different food resources within woodlands and between
woodlands and more open areas (table 1). This is because mast is produced in woodlands, a habitat
in which typical foods will be 13C depleted relative to the same foods from more open areas [25]. For
this reason, while mast is 13C enriched relative to woodland foliage [27,29], that 13C enrichment begins
from a lower (i.e. more 13C-depleted) starting point. Mast consumption is thought to impart up to a 4‰
δ13C increase to the bone collagen isotopic compositions of mast consumers relative to foliage feeders
17
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

in the same habitats [27,29]. If we consider the δ13C range across the entire published assemblage of
archaeological cattle from all sites and time periods in Ireland as a proxy for extremes in woodland
versus open-land pasturing, we see a spread of 2.9‰ (−24.2 to −21.3, n = 344, conservative C : N criteria
applied) [23]. In that context, starting from the lower end of this observed δ13C range, even moderate
mast consumption by pigs could result in δ13C values that fall within the range of cattle. It is also worth
noting that this scenario represents a conservative interpretive framework because mast-generating
woodlands do not always have dense canopies, and consumption of mast from more open woodlands
would generate even higher (more distinctive) pig δ13C values [30].
Taking an even more conservative approach, rather than considering pig δ13C relative to cattle
binned by respective time period, or even at individual sites, we can further compare pig δ13C
means from all time periods relative to cattle from the early medieval period (electronic supplemen-
tary material, table S10). While anachronistic, this approach is more conservative because the early
medieval is a period for which both archaeological and historical evidence suggest that open-land
grazing (which should promote some of the highest possible terrestrial herbivore δ13C values outside
of C4 consumption) of cattle reached its zenith [23,75]. Even in that more extreme comparative context
we find that pigs still consistently have significantly higher mean δ13C values across time (electronic
supplementary material, table S10). This comparison serves to underscore the extent to which pig δ13C
is, by and large, higher than even the most 13C-enriched temporal grouping of non-mast feeders.
It is worth bearing in mind that, in addition to mast, husbandry could have provided pigs with
diets rich in a wide range of other non-photosynthetic plant tissues. For instance, adding grains and
cereals to pig feed should result in a 13C enrichment similar to mast consumption. In that context,
particularly in the medieval and post-medieval periods, as brewing and distilling operations become
more centralized and pigs more sty-based, grain-based waste could have joined mast as a factor
elevating pig δ13C relative to cattle. While cattle would consume grain, they also consume these plants’
comparatively 13C-depleted photosynthetic tissues such as straw.
Second is meat and dairy consumption, which can result in a small TEF for δ13C (ca +0.5‰ per
trophic level, though considerable variation has been noted [34]). As outlined above, consumption of
animal protein has been established based on corresponding δ15N values, with pigs from the early
medieval onward showing clear trophic differences from herbivorous cattle. It is therefore likely that
medieval and post-medieval pig δ13C values are slightly elevated (by perhaps 0.2–0.5‰; corresponding
to trophic levels of 0.4–0.5 based on mean cattle–pig δ15N offsets; electronic supplementary material,
table S10) relative to herbivorous animals. This means that, in addition to pannage feeding, the larger
differences we see between mean pig and cattle δ13C in early, later and post-medieval Ireland (table 2)
may be amplified by a partial trophic offset. However, even factoring in these differences, mean pig
δ13C from all three historical periods remain higher than the peak mean δ13C for early medieval cattle
(though this difference is statistically significant only for the later medieval comparison; electronic
supplementary material, table S10; figures 6–8). In other words, the δ13C differences we have observed
between cattle and pigs are not attributable to trophic enrichment of 13C in pigs with diets richer in
animal protein.
Having considered some of the nuances of the carbon isotope ecology of pig diets, the time series
of δ13C we have presented are consistent with the expectation that mast would have been an important
component of Irish pig diets. While variation in pig δ13C is observed through time, which may reflect
an ebb and flow of the importance of mast and other non-photosynthetic plant tissues (e.g. grain),
the complexities of carbon sources and cycling, compounded by variation from a 13C TEF, mean
higher resolution interpretations remain out of reach without further analyses (see below). We also
note that further work comparing our medieval and post-medieval results based on sites’ relative
access to woodland and grain byproducts could allow for more detailed interpretations. For instance,
pig δ13C variation between urban sites—which may have more access to grain byproducts and less
access to mast—and castle and ecclesiastical sites—which would have better access to mast from
private woodlands—could help unpack the cultural processes behind these patterns at a finer scale.
Nonetheless, given that the centrality of pannage feeding in Irish pig husbandry is well documented in
the earliest writing [12], and that woodlands are a natural habitat for wild boars, it is no surprise that
our results suggest that this practice has been part of Ireland’s agricultural heritage since the beginning
of farming on the island. This is supported by the similarity between the zooarchaeologically identified
wild boars and contemporary domestic pig results presented in this study. In that context, it is worth
pointing out that recent evidence has shown a degree of pannage-focused husbandry, including mean
pig δ13C beyond what we have observed in this wider dataset, at Newgrange during the Late Neolithic
(for discussion, see [72]).
18
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

4.3. Broader implications and consequences of changing pig proximity
We can also ask what these data mean for human–animal relationships in the medieval and post-medi-
eval periods, eras that served as the precursors for present-day ontologies for understanding our own
relationships to animals [76,77]. What implications, for instance, would husbandry of pigs closer to
settlements and increases in the sharing of food resources have for humans? What other traces should
these processes have left in the archaeological record? And what implications might the trajectory of
changing approaches to pig husbandry in Ireland have for our understanding of pig husbandry in
other regions of Europe?
Early medieval texts offer detailed information on the ways in which pigs were husbanded in
Ireland, shedding considerable light on the roles that pigs played in early Irish societies. For instance,
a rich tapestry of sources outline a wide range of norms about what pigs should and should not
be fed, how many pigs should be kept and where they should reside relative to humans’ dwellings,
and importantly when and for what duration pigs should be sent to woodlands [12]. These sources
are invaluable for contextualizing our early medieval data. On the one hand, the sources explain,
and reinforce many times over, the importance of animal protein, especially in the form of dairy
products, in pig diets. On the other hand, the connection between pig husbandry and pannage in
oak woodlands in the autumn is also a well-developed theme. We therefore have an interpretive
framework that blends what had obviously been a long-standing tradition of woodland husbandry
in the autumn/winter with new, more animal-protein-oriented feeding regimes during other times of
the year. Together this set of practices seems tailormade to explain the isotopic compositions we have
observed in early medieval pigs and highlights a new set of Irish pig–human relationships in which,
compared with prehistory, the distance between where pigs were husbanded and where most humans
lived was, on average, shorter.
Compared with the early medieval period, later and post-medieval pigs show similar dietary
patterns, no doubt supported by a continuation of more closely quartered pigs in or near densely
human-populated areas. However, in contrast to the early medieval period, when most settlement was
of a dispersed, rural nature, these latter periods were a time of increasing urbanization, beginning
with the arrival of the Vikings and then the Anglo-Normans [16,17,20]. The more frequent presence
of pigs in what were becoming dense urban settlements during the later and post-medieval periods
is well attested in the historical sources, which include a wide range of narratives following themes
of urban pigs as public nuisances [78–80]. These stories about the trials of humans and pigs living in
close proximity serve to highlight not only the challenges pigs posed to public order, but also to health.
Urbanization and the husbandry of animals in urban settings is thought to have been a major factor in
the pace and nature of development of zoonoses [81–83]. In that context, it is worth noting that, to the
extent that our data show that pigs in Ireland began living in closer quarters to humans by the early
medieval, the risk they presented with respect to the development and communication of zoonotic
diseases may have antecedents reaching at least that far back. At the same time, we can also recognize
the reverse implication, that of increased risk for zooanthroponosis that humans would pose to pigs
when living in closer quarters [84].
While beyond the scope of the present study, it would also be worth considering these data in
comparative contexts with Britain, much of which, unlike Ireland, came under Roman governance
[85]. The arrival of Roman influence in Britain, occurring ca 350 years before the end of the Irish Iron
Age, brought forms of settlement centralization [86] that, within our interpretive framework, could
have led to more animal-protein-focused pig diets and could therefore be marked in the collective
isotopic compositions of British post-Roman-occupation fauna. We recommend caution, however,
when transposing Irish patterns to Britain and Europe. It is possible that pig husbandry in Ireland, and
the implications this has for wider understanding of Irish settlement and farming habitats, unfolded
in ways that were unique to the island. For this reason, factors influencing the ways that pigs were
husbanded in, say, England could cumulatively result in pigs being raised closer to human settlements
earlier in time than in Ireland. While numerous studies have examined archaeological pig isotopic
compositions in England, Scotland and Wales [87–93], none have done so on the scale and from the
baseline-corrected perspective that we have here. This means that the broadscale patterns we have
observed cannot be readily compared.
19
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

4.4. Perceiving pigs: paragons or pariahs?
Our data suggest that the typical ways in which people interacted with pigs would have changed over
time. During prehistory, when pigs lived at some remove from areas of dense human population, the
day-to-day practical knowledge of pig management (i.e. their habits, ecology and biology), including
the direct experience required to know pigs as individuals (i.e. their group dynamics and personalities
[94]; see also [1]), may have been concentrated in the minds of relatively few specialized swine herders
and others supporting that role. By contrast, while medieval and later swine herders would have
maintained this role, during the pannage season at least [12], and the close connection with pigs
that it entails, the shift to pigs living in closer association with people would have provided more
exposure, including opportunities for day-to-day interaction and observation between pigs and people
from a larger cross section of society. In other words, more people would have been forming more of
their understanding of pigs—both abstractly as social and biological creatures and, in cases of closer
sustained interaction, as individuals—based on firsthand experience.
While it is well beyond the scope of this study to unpack these themes in detail, the macroscale
trends in the amount of day-to-day interaction between pigs and people revealed by our data offer
novel opportunities to think about trends in the occurrence and abundance of meaningful two-way
human–pig relationships over time. Here, isotopic data can contribute to interpretations of the past
that do not centre on the human experience and, rather, allow us to consider how the agency of
pigs would have influenced both animal and human decisions. Theoretical approaches that attempt to
comprehend animals as persons, or as agents with a capacity to influence (rather than unidirectionally
be influenced by) people [94] are firmly rooted in the ‘Animal Turn’ within the humanities [95,96].
While these approaches hold considerable promise for advancing the archaeology of human–animal
relations, they require knowledge about direct, sustained interactions between humans and animals (a
prerequisite for a genuine two-way exchange between persons [97,98]). This makes these approaches
challenging to contextualize because nuanced evidence for sustained relationships between people and
animals (even domestic livestock) in the archaeological past can be rare (e.g. much zooarchaeological
data, for instance, reflect animal deaths rather than animal lives (although, see [99,100])). They are also
typically, perhaps necessarily, focused on interactions located at the scale of the individual or small
group. In this context, the approach we have used here, involving large-scale aggregation of δ15N data
(contextualized with supplementary δ13C data) as proxy for proximity to humans, may offer a useful
strategy for framing such narratives at a different, larger scale (for work at a similar scale, see [101]).
5. Conclusion
This is the first study to offer a detailed later-Holocene-scale picture of pig husbandry. While com-
mendable work in numerous other studies has offered diverse insights into human–pig interactions,
including studies that are both spatially and temporally extensive [102–111], our study has benefited
from a synergy among key contextual elements—a robust herbivore baseline for correcting δ15N shifts,
large sample sizes, and a relatively simple interpretive context from an isotope ecology perspective.
These assets allow our interpretations to home in on specific changes in human–animal relation-
ships, changes that imply significant turning points in areas of wider archaeological interest beyond
husbandry, including settlement patterns, landscape management and economic structures. Our data
show that for more than 4000 years, pigs were husbanded in ways that limited the kinds of foods
available to them and mimicked the natural diet of their mainly herbivorous wild ancestors. During the
early medieval that pattern changed and pigs, on average, began eating considerably greater amounts
of animal protein. With respect to the prehistoric period, we suggest that these patterns could reflect
a confluence of two factors that offer viable explanations: settlement patterns and the richness of
pig habitats at locations removed from denser human populations. We suggest that change in land
ownership and the consequent change in farming organization may have been a stronger driver.
Furthermore, with respect to the medieval and post-medieval periods, new opportunities, including
human-generated, animal-protein-rich food waste and the increasing availability of byproducts from
dairy and other forms of agricultural and industrial processing, became a cornerstone of pig husban-
dry. We would like to emphasize that the simplified categories we have used here—i.e. pigs being
more-or-less herbivorous versus strongly omnivorous—are almost certainly masking a great deal of
complexity in the lived experience of people and pigs through time. In that context, we acknowledge
20
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

that, both before and after the transition we have observed, there were likely manifold distinct ways
people raised and related to pigs.
It is also worth pointing out some of the many potential avenues of future work that could help
to unpack these patterns. At a basic level, we believe that further work exploring δ13C and δ15N in
both pigs and baseline fauna would help to refine spatiotemporal dimensions of the transition from
herbivory to omnivory. In that context, exceptions to the broad trends that we have observed will likely
be identified. For instance, larger-scale husbandry operations that fed pigs using brewing and baking
byproducts, as attested to in the historical record, would systematically generate pig remains with less
animal-protein-influenced isotopic compositions. Such a scenario would produce livestock with lower
δ15N and higher δ13C values, and could, for instance, account for a small number of pigs from Viking
Age settlements, in both Ireland [46] and England [87], that show more herbivorous diets. Likewise,
more refined approaches to estimating trophic position, such as δ15N analyses of single amino acids
[112], could also offer more detail about variation in animal-protein consumption among prehistoric
pigs. While more costly and therefore often applied on smaller scales, such techniques can generate
data for establishing both baseline and trophic level from a single pig sample and can therefore offer
interpretations at the individual level, compared to the population level to which our interpretations
are necessarily limited. We would also suggest further work exploring trends in animal health and
pathology through time. For instance, changes in pig husbandry, involving living in closer proximity
to humans, could have left osteological markers in the form of trends in the abundance and kinds of
pathologies preserved in time series of animal remains [81,113,114]. Changes in diet could also have
been recorded in the calculus and metabolites preserved on teeth and in bones [115,116]. Against all
these potential future avenues, however, this study highlights the importance of scale and the value of
integrating larger quantities of data from broad spatial and temporal cross sections of archaeology.
Lastly, it is worth reflecting on our findings in the context of discussion on connections between
Ireland’s earliest historical narratives and late prehistory (for a review, see [117]). There has been much
debate about the value of transposing information from ethnographic and historical processes to earlier
time periods for the purpose of interpreting archaeological patterns. In that context, it is interesting to
consider that the two patterns for pig diets noted in the early medieval period literature—husbandry
based in part on human-provisioned animal protein, on the one hand, and on woodland pannage, on
the other—are not uniformly represented in the Neolithic–Early Bronze Age, Middle–Late Bronze Age
or even most comparatively recent, Iron Age, past. While our data show a consistent importance of
pannage extending from early prehistory through to the historic past, the emphasis on animal protein
in medieval pig diets, and its absence in earlier periods, clearly illustrated the value of exercising
caution when transposing historical processes to earlier times.
6. Methods
6.1. Sample
Our samples were collected for a project examining baseline δ15N variation throughout the Irish
Holocene and a subset of the data presented here were published in a paper exploring that theme [43].
We have since been able to analyse a larger number of pigs from many sites and, by applying more
stringent quality control metrics and considering both δ15N and δ13C from these new and previously
published data [43,46,48,49,118], we are now able to offer interpretations focusing specifically on pigs.
Samples come from 38 sites (see electronic supplementary material, table S1) dating from the
Neolithic to the post-medieval periods and spanning much of the island (figure 1). Our chronological
framework for grouping samples is Neolithic–Early Bronze Age, ca 4000 BC to ca 1500 BC (n = 27);
Middle–Late Bronze Age, ca 1500 BC to ca 500 BC (n = 40); Iron Age, ca 500 BC to ca AD 400 (n = 86);
early medieval period, ca AD 400 to ca AD 1100 (n = 46); later medieval period, ca AD 1100 to ca AD
1550 (n = 58); and post-medieval/early modern period, ca AD 1550 to ca AD 1900 (n = 45). Samples
were identified and recorded using standard zooarchaeological criteria [119], and young juveniles were
excluded on the basis of either fusion, tooth eruption and wear, or size, as appropriate.
6.2. Collagen extraction and isotopic analyses
Collagen extractions followed two protocols, both modified from Longin [120]. Samples with a ‘SUBC’
prefix (n = 203) were demineralized in 0.5 M hydrochloric acid (HCl), then neutralized in Type 1
21
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

water (resistivity = 18 MΩ cm) and refluxed in a 10−3 HCl (pH3) solution in an oven at ca 65−70°C
for 36–48 h. Gelatinized samples were then processed through 45−90 μm Ezee filters and 30 kDa
molecular weight cut-off filters (ultrafilters) and then frozen and lyophilized. Samples with ‘IUBC’
prefix (n = 99) followed the same protocol with two changes: (i) Ezee and ultrafilters were not used;
and (ii) a sodium hydroxide (NaOH) pre-treatment step (i.e. samples were soaked in a 0.1 M NaOH
solution in an ultrasonic bath for 15 min cycles until solution remained clear) was performed after
demineralization but before refluxing. We recognize that these two methods may not be equally
effective at removing humic acids, the main potential contaminants for bone collagen, and therefore
systematic δ13C differences could occur between samples treated with each method (for a review, see
[53]). However, we apply the strictest possible quality control criteria, including conservative C : N [53]
as well as standard carbon (>13.8%) and nitrogen (>4.0%) concentration metrics [121]. With respect
to C : N, we have applied the conservative criteria, which were modelled to ensure measured bone
collagen δ13C values deemed viable are within −0.5‰ of their biogenic δ13C value. Virtually all samples
fell in the ‘<−20.00‰’ δ13C category, which allows for associated C : N values of no greater than 3.50
under the conservative criteria. By contrast, liberal criteria are less stringent and tolerate humic acid
contamination-induced δ13C alterations of up to −1.0‰. As noted in §3, these criteria could technically
be applied to two pig samples that showed evidence of marine diets (see electronic supplementary
material, figure S3). These two samples produced δ13C values falling in the ‘<−20.00‰’ and ‘−19.99
to −18.00’ δ13C categories, which allow for associated C : N values of no greater than 3.70 and 3.55,
respectively. In the context of these quality control metrics, we note that regardless of whether the two
methods used for collagen extraction are equally effective at purifying collagen extracts (i.e. mainly
removing humic acids), application of the conservative C : N criteria will ensure that any data used
(regardless of what extraction protocol was applied) will be comparable, with minimal influence from
contamination.
Isotopic compositions were measured on 0.5 mg subsamples of collagen using an elemental
analyser coupled to an isotope ratio mass spectrometer in the Archaeology Chemistry Laboratory
at the University of British Columbia. Measurements for 36% of samples were replicated. Isotopic
compositions were calibrated using a two-point calibration curve anchored to international standards
and accuracy and precision were monitored using a variety of check standards. Known and long-
term observed isotopic compositions and s.d. for calibration and check standards are presented in
electronic supplementary material, table S2. Standard deviations for calibration standards from all
analytical sessions are presented in electronic supplementary material, table S3. Means and s.d. for
check standards and sample replicates are shown in electronic supplementary material, tables S4
and S5, respectively. For δ13C and δ15N: systematic errors [μ (bias)] were ±0.11‰ and ± .15‰, respec-
tively; random errors [μR(w)] were ±0.13‰ and ±0.20‰, respectively; and standard uncertainties were
±0.16‰ and ±0.25‰, respectively [122].
6.3. Baseline adjustments
Previous research has shown that Irish herbivore δ15N values have increased over the Holocene,
evidencing significant shifts in δ15N baseline variation occurring at the plant–soil level [43]. This
phenomenon complicates the interpretation of potential patterns in higher trophic level animals,
including omnivores such as pigs, because δ15N shifts in non-herbivorous animals could be linked to
variation both in δ15N baseline and animal protein consumption. To obviate this issue while interpret-
ing temporal patterns in pig δ15N values, we take two approaches.
First, the significance of changing levels of pig omnivory through time was evaluated by sequen-
tially comparing baseline-corrected mean δ15N between adjacent time periods. For each comparison,
the latter temporal groups’ δ15N values were adjusted according to the baseline offset observed
between groups of cattle from those respective time periods. For instance, to establish the extent to
which the change in pig δ15N between, say, the early medieval and the later medieval periods could be
driven by omnivory rather than by fluctuations in baseline δ15N, we adjusted the isotopic composition
of the later medieval pigs by the difference in mean δ15N (−0.7‰) that we had observed between cattle
from the early medieval and later medieval periods. For a list of groups compared, see electronic
supplementary material, table S9. We selected cattle for this purpose because they offer, by far, the
largest published dataset of δ15N and δ13C values (also used in other comparison, see below), including
n = 344 samples with associated quality control metrics meeting conservative C : N criteria [53].
Second, we aggregated isotopic data from across large numbers of sites for each time period. This
offers a ‘zoomed out’ view of trends that is less likely to be impacted by site-specific processes, which
22
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

may be more prone to reflect localized and otherwise unusual husbandry practices. To help further
confirm that site-level patterns are in line with overall trends, where sample sizes permitted (i.e. for
sites with more than five samples per taxon, see below), we performed site-level statistical comparisons
to verify that these matched patterns observed at the more spatially encompassing time-period level.
The importance of mast in pig diets, which is 13C enriched relative to most other woodland foods in
Ireland, was established in part through comparisons between cattle (which cannot eat large amounts
of mast) and pigs (which will preferentially eat mast). For more detail, see §§2–4. Some of these
comparisons require adjustments to account for the influence of a minor trophic enrichment factor
(TEF) for δ13C (ca 0.5‰ per trophic level (TL) as compared with ca 3.6‰ for δ15N [34]). Specifically,
historical pig temporal group δ13C means were adjusted according to TL (electronic supplementary
material, table S10; figures 6–8 panels (a)). TL for pigs (TLpig) was established by subtracting the mean
δ15N for cattle from the mean δ15N for pigs in each time period and then dividing this number by
the expected δ15N TEF (i.e. 3.6‰). The amount of adjustment for pig δ13C was then calculated by
multiplying TLpig by the TEF for δ13C (i.e. 0.5‰). This adjustment was then subtracted from pig δ13C
values for respective time-period groups. Comparisons were then performed using recalculated mean
pig δ13C for the early, later and post-medieval periods versus cattle from the early medieval period.
6.4. Other interpretive considerations
While our dataset is large by archaeological standards, spatial coverage per time period does not
provide an opportunity explore geographical variation in detail. Nonetheless, we use the opportunity
created by our comparison of site-level and time period-level trends (see above) to consider whether
regional patterns emerge. We also note that previous work on Irish archaeological pig and cattle
isotopic compositions, aggregated at the island-wide scale, indicates that broadscale patterns reflect
temporal changes rather than geographical factors [23,43]. Moreover, we note that these previous
analyses [23,43] showed no correlation between trends in archaeological faunal isotopic compositions
(either δ13C or δ15N) and known climate oscillations across the Irish Holocene [123–125], suggesting
that, at this island-wide, aggregated scale, changing environmental conditions may be less important,
relative to human impacts, in driving isotopic variation.
We are aware that interspecific differences in isotopic diet-to-collagen spacing can occur and,
as such, there could be systematic variation in the way that cattle and pigs eating the same
diet express δ13C and δ15N values. With respect to δ15N, while a wider range of inter-specific
variation has been observed in diet-to-collagen spacing (also known as trophic enrichment factors),
evaluation of the extent to which these could be applicable to our interpretations is complicated.
This is due to the diverse range of variables (e.g. nutritional status, diet composition, growth
rate) which can, at both the inter-specific/intra-digestive-physiology and the intra-specific levels,
dramatically impact diet-to-collagen spacing size [126–128]. These factors cannot be known for
animals from the archaeological past. Despite the wide range and complexity of this variation,
broadscale syntheses of observations across a taxonomically diverse range of species suggest
among mammals diet-to-collagen spacing for δ15N is on average similar across taxa [34]. We
note that across intra-site interspecific comparisons in this study (table 2), pigs provide examples
of cases where average δ15N values are higher or lower than those of cattle, suggesting any
systematic difference, if present at all, would likely be very small.
With respect to δ13C, systematic differences have been documented between diet and collagen
relative to other tissues from the same animals [129,130], which are thought to be linked with differen-
ces in respective species’ digestive physiologies [131]. Studies examining cattle and pigs independently,
as well as other taxa, have shown that, in addition to digestive physiology, diverse variables, again
including factors such as nutritional status and diet quality, can dramatically impact diet-to-collagen
spacing for δ13C between animals for the same species [126,132,133]. Other recent work exploring
spacing between enamel (bioapatite) and diet among animals with differing digestive physiologies has
suggested that animals with ruminant-fermenter digestive systems, such as cattle, have δ13C values
that are ca 1‰ higher than those with non-coprophagous hindgut-fermenter digestive systems, such as
pigs [71]. However, bone collagen and bioapatite are constructed and maintained using materials from
distinctive biomolecular pools within the body, making the enamel–collagen tissue spacing notoriously
difficult to characterize [131]. To concretely resolve the question of whether, and under what condi-
tions, there may be a consistent and meaningful interspecific diet-to-collagen offset in δ13C between
cattle and pigs, in-depth controlled feeding experiments are needed to quantify the impacts of different
23
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

diets, nutritional statuses and environmental conditions on bone collagen δ13C in cattle and pigs raised
under the same conditions. In summary, as with δ15N, because intraspecific differences in diet–collagen
δ13C are contingent on basic factors like diet quality, which are unknowable for archaeological animals,
it is not possible to correct for the potential effects of digestive physiology on δ13C. In that context,
we note that we are not aware of research suggesting that pigs and cattle should have a pronounced
difference (e.g. >1‰) in δ13C diet–collagen offset.
Considering the above, while we recognize that it is possible that under certain conditions, cattle
and pig bone collagen isotopic compositions may reflect diet in slightly different ways, we do not
attempt to apply a correction when comparing cattle and pig δ15N and δ13C.
6.5. Statistics
Statistical comparisons were conducted in PAST version 4.13 [134]. We tested for normality of sample
group distributions using Shapiro–Wilk tests. Where groups were not normally distributed, a Mann–
Whitney U test was used to compare group means. In cases where compared groups were normally
distributed, a Levene’s test was used to evaluate homogeneity of variance. Where variance was equal
between groups, a Student’s t-test was used to compare means. Where variance was not equal between
groups, a Welch’s t-test was used to compare means. Effect size was determined by Cohen’s d for
parametric tests (i.e. Student’s t and Welch’s t) and Vargha–Delaney A was used for non-parametric
tests (i.e. Mann–Whitney U).
Ethics. This work did not require ethical approval from a human subject or animal welfare committee.
Data accessibility. All data are included in the paper and the electronic supplementary material [135].
Declaration of AI use. We have not used AI-assisted technologies in creating this article.
Authors’ contributions. E.G.: conceptualization, data curation, funding acquisition, investigation, methodology,
project administration, visualization, writing—original draft, writing—review and editing; F.B.: conceptualization,
investigation, methodology, resources, writing—review and editing; F.M.: resources, writing—review and editing;
E.T.: writing—review and editing; M.P.R.: resources, writing—review and editing.
All authors gave final approval for publication and agreed to be held accountable for the work performed
therein.
Conflict of interest declaration. F.B. undertook the original faunal analysis of many of these assemblages on a paid
consultancy basis.
Funding. This project was supported by the Social Science and Humanities Research Council of Canada (Banting
postdoctoral fellowship, E.G.), the Wenner-Gren Foundation (fieldwork dissertation grant, E.G.), and the Ireland
Canada University Foundation (Craig Dobbin Fellowship Program, E.G.).
Acknowledgements. For permission to sample and for logistical help, we thank: N. O’Connor and E. Ashe (National
Museum of Ireland); K. Neil (Ulster Museum); J. Lyttleton (then University College Cork); R. MacDonald and J.
Hepburn (then University of British Columbia); T. Kahlert (then IT Sligo); G. Stout (then National Monuments
Service) and M Stout (then Dublin City University); R. Crumlish, C. Jones (National University of Ireland Galway);
D. Moore (Moore Group); R. Ó Baoill (Queens University Belfast, Centre for Archaeological Fieldwork); S. Ní
hAodha, S. Scully and C. McConway (then Archaeological Development Services Ltd); F. O’Carroll and S. Mandal
(then Cultural Resource Development Services Ltd); S. Johnston (then Arch Tech Ltd.); and R. Gillespie (Mayo
County Council).
References
1. Kreiner J. 2020 Legions of pigs in the early medieval west. New Haven, CT: Yale University Press. (doi:10.12987/yale/9780300246292.001.0001)
2. Albarella U, Dobney K, Ervynck A, Rowley-Conwy P. 2007 Pigs and humans: 10,000 years of interaction. Oxford, UK: Oxford University Press.
3. Albarella U, Payne S. 2005 Neolithic pigs from Durrington Walls, Wiltshire, England: a biometrical database. J. Archaeol. Sci. 32, 589–599. (doi:
10.1016/j.jas.2004.11.008)
4. Payne S, Bull G. 1988 Components of variation in measurements to distinguish wild from domestic pig remain. Archaeozoologica II 1–2, 27–66.
5. Putman RJ. 2008 Pigs. In Mammals of the British Isles: handbook (eds S Harris, D Yalden), pp. 563–564. Southampton, UK: Mammal Society.
6. Copado F, de Aluja AS, Mayagoitia L, Galindo F. 2004 The behaviour of free ranging pigs in the Mexican tropics and its relationships with human
faeces consumption. Appl. Anim. Behav. Sci. 88, 243–252. (doi:10.1016/j.applanim.2004.03.013)
7. Albarella U, Manconi F, Albarella U, Dobney K, Ervynck A, Rowley-Conwy P. 2007 Ethnoarchaeology of pig husbandr y in Sardinia and Corsica. In
Pigs and humans: 10,000 years of interaction. Oxford, UK: Oxford University Press. (doi:10.1093/oso/9780199207046.003.0027)
8. Silver IA. 1963 The aging of domestic animals. In Science in archaeology: a survey of progress and research (eds D Brothwell, E Higgs), pp. 250–
268. London, UK: Basic Books.
24
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

9. Danfaer A, Strathe A. 2012 Quantitative and physiological aspects of pig growth. In Nutrition and physiology of the pig (eds KEB Knudsen, NJ
Kjeldsen, HD Poulsen, BB Jensen). Copenhagen, Denmark: SEGES Danish Pig Research Centre.
10. Beglane F. 2016 The faunal remains. In The Bective Abbey project: archaeological excavations 2009–12 (eds G Stout, M Stout), pp. 106–153.
Dublin, Ireland: Wordwell.
11. Parsons JJ. 1962 The acorn-hog economy of the oak woodlands of southwestern Spain. Geogr. Rev. 52, 211. (doi:10.2307/212957)
12. Kelly F. 1997 Early Irish farming. Dublin, Ireland: School of Celtic Studies.
13. McCormick F. 2007 Mammal bone studies from prehistoric Irish sites. In Environmental archaeology in Ireland (eds EM Murphy, NJ Whitehouse),
pp. 77–101. Oxford, UK: Oxbow.
14. McCormick F, Murray E. 2007 Excavations at Knowth, vol. 3: Knowth and the zooarchaeology of early Christian Ireland. In Royal Irish Academy
monographs in archaeology. Dublin, Ireland: Royal Irish Academy.
15. Pigière F, McCormick F, Olet L, Moore D, O’Carroll F, Smyth J. 2022 More than meat? Examining cattle slaughter, feasting and deposition in later
4th millennium BC Atlantic Europe: a case study from Kilshane, Ireland. J. Archaeol. Sci. 41, 103312. (doi:10.1016/j.jasrep.2021.103312)
16. McCormick F. 1991 The effect of the Anglo-Norman settlement on Ireland’s wild and domesticated fauna. In Animal use and culture change (eds
PJ Crabtree, K Ryan), pp. 41–52. Philadelphia, PA: University Museum of Archaeology and Anthropology, University of Pennsylvania.
17. McCormick F, Murray E. 2017 The zooarchaeology of medieval Ireland. In The Oxford handbook of zooarchaeology, pp. 195–213. Oxford, UK:
Oxford University Press. (doi:10.1093/oxfordhb/9780199686476.013.15)
18. Beglane F. 2009 Meat and craft in medieval and post-medieval Trim. In Uncovering medieval Trim: archaeological excavations in and around Trim,
Co. Meath (eds M Potterton, M Seaver), pp. 346–370. Dublin, Ireland: Four Courts Press.
19. Beglane F. 2009 Long pig’s feet from Iron Age Trim. In Uncovering medieval Trim: archaeological excavations in and around Trim, Co. Meath (eds M
Potterton, M Seaver), pp. 57–71. Dublin, Ireland: Four Courts Press.
20. Beglane F. 2023 The fat of the land. In Food provisioning in complex societies: zooarchaeological perspectives (eds L Atici, BS Arbuckle), pp. 71–
91. Denver, CO: University Press of Colorado. (doi:10.5876/9781646422562.c004)
21. Szpak P. 2022 Why zooarchaeology needs stable isotope analysis. In Isotope research in zooarchaeology, pp. 248–270. Boca Raton, FL:
University Press of Florida. (doi:10.2307/j.ctv2zx9q60.15)
22. Guir y EJ, Gaulton BC. 2016 Inferring human behaviors from isotopic analyses of rat diet: a critical review and historical application. J. Archaeol.
Method Theory 23, 399–426. (doi:10.1007/s10816-015-9248-9)
23. Guir y E, Beglane F, Szpak P, McCormick F, Teeter MA, Cheung C, Richards MP. 2023 Changing human-cattle relationships in Ireland: a 6000-year
isotopic perspective. Antiquity 97, 1436–1452. (doi:10.15184/aqy.2023.163)
24. Tieszen LL. 1991 Natural variations in the carbon isotope values of plants: implications for archaeology, ecology, and paleoecology. J. Archaeol.
Sci. 18, 227–248. (doi:10.1016/0305-4403(91)90063-u)
25. Bonafini M, Pellegrini M, Ditchfield P, Pollard AM. 2013 Investigation of the ‘canopy effect’ in the isotope ecology of temperate woodlands. J.
Archaeol. Sci. 40, 3926–3935. (doi:10.1016/j.jas.2013.03.028)
26. van der Merwe NJ, Medina E. 1991 The canopy effect, carbon isotope ratios and foodwebs in Amazonia. J. Archaeol. Sci. 18, 249–259. (doi:10.
1016/0305-4403(91)90064-v)
27. Guiry E, Orchard TJ, Needs-Howarth S, Szpak P. 2021 Isotopic evidence for garden hunting and resource depression in the late woodland of
northeastern North America. Am. Antiq. 86, 90–110. (doi:10.1017/aaq.2020.86)
28. Guir y EJ, Orchard TJ, Royle TCA, Cheung C, Yang DY. 2020 Dietary plasticity and the extinction of the passenger pigeon (Ectopistes migratorius).
Quat. Sci. Rev. 233, 106225. (doi:10.1016/j.quascirev.2020.106225)
29. Cernusak LA etal. 2009 Why are non-photosynthetic tissues generally ¹³C enriched compared with leaves in C₃ plants? Review and synthesis of
current hypotheses. Funct. Plant Biol. 36, 199–213. (doi:10.1071/FP08216)
30. Styring AK , Jirdén E, Lagerås P, Larsson M, Sjöström A , Ljung K. 2024 Carbon isotope values of hazelnut shells: a new proxy for canopy density.
Front. Environ. Archaeol. 3, 1351411. (doi:10.3389/fearc.2024.1351411)
31. Sarà M, Sarà G. 2007 Trophic habits of Muscardinus avellanarius (Mammalia Gliridae) as revealed by multiple stable isotope analysis. Ethol. Ecol.
Evol. 19, 215–223. (doi:10.1080/08927014.2007.9522563)
32. Selva N, Hobson KA, Cortés-Avizanda A, Zalewski A, Donázar JA. 2012 Mast pulses shape trophic interactions between fluctuating rodent
populations in a primeval forest. PLoS ONE 7, e51267. (doi:10.1371/journal.pone.0051267)
33. DeNiro MJ, Epstein S. 1978 Influence of diet on the distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42, 495–506. (doi:10.
1016/0016-7037(78)90199-0)
34. Caut S, Angulo E, Courchamp F. 2009 Variation in discrimination factors (Δ15N and Δ13C): the effect of diet isotopic values and applications for
diet reconstruction. J. Appl. Ecol. 46, 443–453. (doi:10.1111/j.1365-2664.2009.01620.x)
35. Guir y E. 2019 Complexities of stable carbon and nitrogen isotope biogeochemistry in ancient freshwater ecosystems: implications for the study
of past subsistence and environmental change. Front. Ecol. Evol. 7, 313. (doi:10.3389/fevo.2019.00313)
36. Fan R, Morozumi T, Maximov TC, Sugimoto A. 2018 Effect of floods on the δ13C values in plant leaves: a study of willows in northeastern Siberia .
PeerJ 6, e5374. (doi:10.7717/peerj.5374)
37. Chisholm BS, Nelson DE, Schwarcz HP. 1982 Stable-carbon isotope ratios as a measure of marine versus terrestrial protein in ancient diets.
Science 216, 1131–1132. (doi:10.1126/science.216.4550.1131)
38. Losey RJ, Guiry E, Nomokonova T, Gusev AV, Szpak P. 2020 Storing fish? A dog’s isotopic biography provides insight into Iron Age food
preservation strategies in the Russian Arctic. Archaeol. Anthropol. Sci. 12. (doi:10.1007/s12520-020-01166-3)
25
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

39. Hamilton J, Hedges REM, Robinson M. 2009 Rooting for pigfruit: pig feeding in Neolithic and Iron Age Britain compared. Antiquity 83, 998–
1011. (doi:10.1017/s0003598x00099300)
40. Szpak P. 2014 Complexities of nitrogen isotope biogeochemistry in plant-soil systems: implications for the study of ancient agricultural and
animal management practices. Front. Plant Sci. 5, 288. (doi:10.3389/fpls.2014.00288)
41. Katzenberg MA, Pfeiffer S. 1995 Nitrogen isotope evidence for weaning age in a nineteenth century Canadian skeletal sample. In Bodies of
evidence: reconstructing history through skeletal analysis (ed. A Grauer), pp. 221–235. New York, NY: Wiley-Liss.
42. Minagawa M, Wada E. 1984 Stepwise enrichment of ¹⁵N along food chains: further evidence and the relation between δ¹⁵N and animal age.
Geochim. Cosmochim. Acta 48, 1135–1140. (doi:10.1016/0016-7037(84)90204-7)
43. Guir y E, Beglane F, Szpak P, Schulting R, McCormick F, Richards MP. 2018 Anthropogenic changes to the Holocene nitrogen cycle in Ireland. Sci.
Adv. 4, s9383. (doi:10.1126/sciadv.aas9383)
44. Guiry EJ, Orchard TJ, Needs-Howarth S, Szpak P. 2022 Freshwater wetland-driven variation in sulfur isotope compositions: implications for
human paleodiet and ecological research. Front. Ecol. Evol. 10, 953042. (doi:10.3389/fevo.2022.953042)
45. Guiry E, Noël S, Fowler J. 2021 Archaeological herbivore δ¹³C and δ³⁴S provide a marker for saltmarsh use and new insights into the process of
¹⁵N-enrichment in coastal plants. J. Archaeol. Sci. 125, 105295. (doi:10.1016/j.jas.2020.105295)
46. Knudson KJ, O’Donnabhain B, Carver C, Cleland R, Price TD. 2012 Migration and Viking Dublin: paleomobility and paleodiet through isotopic
analyses. J. Archaeol. Sci. 39, 308–320. (doi:10.1016/j.jas.2011.09.014)
47. Ryan SE, Reynard LM, Crowley QG, Snoeck C, Tuross N. 2018 Early medieval reliance on the land and the local: an integrated multi-isotope study
(⁸⁷Sr/⁸⁶Sr, δ¹⁸O, δ¹³C, δ¹⁵N) of diet and migration in Co. Meath, Ireland. J. Archaeol. Sci. 98, 59–71. (doi:10.1016/j.jas.2018.08.002)
48. McKenzie CJ , Murphy EM, Guiry E, Donnelly CJ, Beglane F. 2020 Diet in medieval Gaelic Ireland: a multiproxy study of the human remains from
Ballyhanna, Co. Donegal. J. Archaeol. Sci. 121, 105203. (doi:10.1016/j.jas.2020.105203)
49. Madgwick R , Grimes V, Lamb AL, Nederbragt AJ, Evans JA, McCormick F. 2019 Feasting and mobility in Iron Age Ireland: multi-isotope analysis
reveals the vast catchment of Navan Fort, Ulster. Sci. Rep. 9, 19792. (doi:10.1038/s41598-019-55671-0)
50. Woodman P, Mccarthy M, Monaghan N. 1997 The Irish Quaternary fauna project. Quat. Sci. Rev. 16, 129–159. (doi:10.1016/s0277-
3791(96)00037-6)
51. Jensen P, Recén B. 1989 When to wean—observations from free-ranging domestic pigs. Appl. Anim. Behav. Sci. 23, 49–60. (doi:10.1016/
0168-1591(89)90006-3)
52. Ó Baoill R. 2001 2001:309—Greencastle, Down. Excavations.ie: database of Irish excavation reports.
53. Guiry EJ, Szpak P. 2021 Improved quality control criteria for stable carbon and nitrogen isotope measurements of ancient bone collagen. J.
Archaeol. Sci. 132, 105416. (doi:10.1016/j.jas.2021.105416)
54. Kyriazakis I, Emmans GC, Whittemore CT. 1990 Diet selection in pigs: choices made by growing pigs given foods of different protein
concentrations. Anim. Sci. 51, 189–199. (doi:10.1017/s0003356100005298)
55. Rose SP, Kyriazakis I. 1991 Diet selection of pigs and poultry. Proc. Nutr. Soc. 50, 87–98. (doi:10.1079/pns19910014)
56. Westendorf ML. 2000 Food waste as animal feed: an introduction. In Food waste to animal feed, pp. 3–16. Ames, IA: Iowa State University Press.
(doi:10.1002/9780470290217)
57. Whitehouse NJ etal. 2014 Neolithic agriculture on the European western frontier: the boom and bust of early farming in Ireland. J. Archaeol. Sci.
51, 181–205. (doi:10.1016/j.jas.2013.08.009)
58. Cooney G. 2012 Reading a landscape manuscript: a review of progress in prehistoric settlement studies in Ireland. In A history of settlement in
Ireland, pp. 17–65. London, UK: Routledge. (doi:10.4324/9780203025192-8)
59. Szabó P. 2013 Rethinking pannage: historical interactions between oak and swine. In Trees, forested landscapes and grazing animals (ed. ID
Rotherham), pp. 51–61. London, UK: Routledge.
60. Wealleans AL. 2013 Such as pigs eat: the rise and fall of the pannage pig in the UK. J. Sci. Food Agric. 93, 2076–2083. (doi:10.1002/jsfa.6145)
61. Lamond E . 1890 Walter of Henley’s husbandry: together with an anonymous husbandry, seneschaucie and Robert Grosseteste’s rules. London, UK:
Longmans, Green and Co.
62. Plunkett G. 2007 Pollen analysis and archaeology in Ireland. In Environmental archaeology in Ireland (ed. NJ Whitehouse), pp. 221–240. Oxford,
UK: Oxbow Books.
63. Plunkett G. 2020 A palynological perspective on ‘An archaeology of Ireland for the information age’. Emania 25, 39–43.
64. Kelly F. 1988 A guide to early Irish law. Dublin, Ireland: Dublin Institute for Advanced Studies.
65. Mytum HC. 1992 The origins of early Christian Ireland. London, UK: Routledge.
66. Becker K, Armit I, Swindles G, Stanley M. 2017 New perspec tives on the Irish Iron Age: the impact of NRA development on our understanding of
later prehistory. In Stories of Ireland’s past (eds M Stanley, R Swan, A O’Sullivan), pp. 85–100. Dublin, Ireland: Transport Infrastructure Ireland.
67. Kenward H, Hall A, Allison E, Carrott J. 2011 Environment, activity and living conditions at Deer Park Farms: evidence from plant and
invertebrate remains. In Deer Park Farms: the excavation of a raised rath in the Glenarm valley, Co. Antrim. Northern Ireland archaeological
monographs (eds C Lynn, JA McDowell), pp. 498–547. Belfast, UK: Stationery Office.
68. McCormick F. 1995 Cows, ringforts and the origins of early Christian Ireland. Emania 13, 33–37.
69. Beglane F. 2018 Forests and chases in medieval Ireland, 1169–c .1399. J. Hist. Geogr. 59, 90–99. (doi:10.1016/j.jhg.2017.11.002)
70. Beglane F. 2015 Anglo-Norman parks in medieval Ireland. Dublin, Ireland: Four Courts Press.
71. Cerling TE, Bernasconi SM, Hofstetter LS, Jaggi M, Wyss F, Rudolf von Rohr C, Clauss M. 2021 CH₄/CO₂ ratios and carbon isotope enrichment
between diet and breath in herbivorous mammals. Front. Ecol. Evol. 9, 638568. (doi:10.3389/fevo.2021.638568)
26
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

72. Guir y E, Beglane F, Carlin N, Orton D, Teeter M, Szpak P. Submitted. Pigs, pannage, and the solstice: isotopic insights from prehistoric feasting at
Newgrange.
73. González-Barrio R, Truchado P, García-Villalba R, Hervás G, Frutos P, Espín JC, Tomás-Barberán FA. 2012 Metabolism of oak leaf ellagitannins
and urolithin production in beef cattle. J. Agric. Food Chem. 60, 3068–3077. (doi:10.1021/jf300718k)
74. Cedervall A, Johansson HE, Jönsson L. 1973 Acorn poisoning in cattle. Nord. Vet. Med. 25, 639–644.
75. McCormick F. 2008 The decline of the cow: agricultural and settlement change in early medieval Ireland. Peritia 20, 209–224. (doi:10.1484/j.
peri.3.632)
76. Thomas R. 2009 Bones of contention: why later post-medieval faunal assemblages in Britain matter. In Crossing paths or sharing tracks?, pp.
133–148. Woodbridge, UK: Boydell and Brewer. (doi:10.1515/9781846157103-017)
77. Thomas R, Fothergill BT. 2014 Foreword. Anthropozoologica 49, 11–18. (doi:10.5252/az2014n1a01)
78. Cantwell I. 2001 Anthropozoological relationships in late medieval Dublin. Dublin Hist. Rec. 54, 73–80.
79. Adelman J. 2020 Civilised by beasts. Manchester, UK: Manchester University Press.
80. Nugent J. 2009 The human snout: pigs, priests, and peasants in the parlor. Senses Soc. 4, 283–301. (doi:10.2752/
174589209x12464528171851)
81. Bartosiewicz L. 2022 Ancient zoonoses. In Zoonoses: infections affecting humans and animals, pp. 1–23. Dordrecht, The Netherlands: Springer.
(doi:10.1007/978-3-030-85877-3_54-1)
82. Tanga C, Remigio M, Viciano J. 2022 Transmission of zoonotic diseases in the daily life of ancient Pompeii and Herculaneum (79 CE, Italy): a
review of animal–human–environment interactions through biological, historical and archaeological sources. Animals 12, 213. (doi:10.3390/
ani12020213)
83. Uhl EW, Thomas R. 2022 Uncovering tales of transmission: an integrated palaeopathological perspective on the evolution of shared human and
animal pathogens. In Palaeopathology and evolutionary medicine, pp. 317–349. Oxford, UK: Oxford University Press. (doi:10.1093/oso/
9780198849711.003.0017)
84. Messenger AM, Barnes AN, Gray GC. 2014 Reverse zoonotic disease transmission (zooanthroponosis): a systematic review of seldom-
documented human biological threats to animals. PLoS ONE 9, e89055. (doi:10.1371/journal.pone.0089055)
85. Johnston E. 2017 Ireland in late antiquity. Stud. Late Antiq. 1, 107–123. (doi:10.1525/sla.2017.1.2.107)
86. Redfern RC. 2020 Changing people, changing settlements? A perspective on urbanism from Roman Britain. In Bioarchaeology and social theory:
the bioarchaeology of urbanization, pp. 25–47. Dordrecht, The Netherlands: Springer. (doi:10.1007/978-3-030-53417-2_2)
87. Hammond C, O’Connor T. 2013 Pig diet in medieval York: carbon and nitrogen stable isotopes. Archaeol. Anthropol. Sci. 5, 123–127. (doi:10.
1007/s12520-013-0123-x)
88. Jones JR, Mulville J. 2016 Isotopic and zooarchaeological approaches towards understanding aquatic resource use in human economies and
animal management in the prehistoric Scottish North Atlantic Islands. J. Archaeol. Sci. 6, 665–677. (doi:10.1016/j.jasrep.2015.08.019)
89. Madgwick R, Mulville J, Stevens RE. 2012 Diversity in foddering strategy and herd management in late Bronze Age Britain: an isotopic
investigation of pigs and other fauna from two midden sites. Environ. Archaeol. 17, 126–140. (doi:10.1179/1461410312z.00000000011)
90. Schulting RJ etal. 2019 The ups & downs of Iron Age animal management on the Oxfordshire Ridgeway, south-central England: a multi-isotope
approach. J. Archaeol. Sci. 101, 199–212. (doi:10.1016/j.jas.2018.09.006)
91. Hamilton J, Thomas R. 2012 Pannage, pulses and pigs: isotopic and zooarchaeological evidence for changing pig management practices in later
medieval England. Mediev. Archaeol. 56, 234–259. (doi:10.1179/0076609712z.0000000008)
92. Millard AR, Jimenez‐Cano NG, Lebrasseur O, Sakai Y. 2013 Isotopic investigation of animal husbandry in the Welsh and English periods at
Dryslwyn Castle, Carmarthenshire, Wales. Int. J. Osteoarchaeol. 23, 640–650. (doi:10.1002/oa.1292)
93. Stevens RE, Lightfoot E, Hamilton J, Cunliffe BW, Hedges REM. 2013 One for the master and one for the dame: stable isotope investigations of
Iron Age animal husbandry in the Danebury environs. Archaeol. Anthropol. Sci. 5, 95–109. (doi:10.1007/s12520-012-0114-3)
94. Knight J. 2005 Animals in person: cultural perspectives on human-animal intimacies. Oxford, UK: Berg.
95. Rit vo H. 2007 On the animal turn. Daedalus 136, 118–122. (doi:10.1162/daed.2007.136.4.118)
96. Over ton NJ, Hamilakis Y. 2013 A manifesto for a social zooarchaeology. Swans and other beings in the Mesolithic. Archaeol. Dialogues 20, 111–
136. (doi:10.1017/s1380203813000159)
97. Dwyer PD, Minnegal M. 2020 Person, place or pig: animal attachments and human transactions in New Guinea. In Animals in person (ed. J
Knight), pp. 37–60. London, UK: Routledge. (doi:10.4324/9781003135883-3)
98. Theodossopoulos D. 2020 Care, order and usefulness: the context of the human–animal relationship in a Greek island community. In Animals in
person (ed. J Knight), pp. 15–35. London, UK: Routledge. (doi:10.4324/9781003135883-2)
99. Argent G. 2010 Do the clothes make the horse? Relationality, roles and statuses in Iron Age Inner Asia. World Archaeol. 42, 157–174. (doi:10.
1080/00438241003672633)
100. Szpak P, Millaire JF, White CD, Longstaffe FJ. 2014 Small scale camelid husbandry on the north coast of Peru (Virú Valley): insight from stable
isotope analysis. J. Anth. Arch. 36, 110–129.
101. Orton D. 2010 Both subject and object: herding, inalienability and sentient property in prehistory. World Archaeol. 42, 188–200. (doi:10.1080/
00438241003672773)
102. Frémondeau D, De Cupere B, Evin A, Van Neer W. 2017 Diversity in pig husbandry from the Classical-Hellenistic to the Byzantine periods: an
integrated dental analysis of Düzen Tepe and Sagalassos assemblages (Turkey). J. Archaeol. Sci. 11, 38–52. (doi:10.1016/j.jasrep.2016.11.030)
27
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

103. Aiken M. 2019 Pig husbandry at Kastro Kallithea: an isotopic study of pig husbandry in Hellenistic Thessaly. Master thesis, University of Alberta,
Canada.
104. Kennedy JR, Guiry EJ. 2023 Exploring railroad impacts on meat trade: an isotopic investigation of meat sourcing and animal husbandry at
Chinese Diaspora sites in the American West. Int. J. Hist. Archaeol. 27, 393–423. (doi:10.1007/s10761-022-00663-6)
105. Madgwick R, Lamb AL, Sloane H, Nederbragt AJ, Albarella U, Pearson MP, Evans JA. 2019 Multi-isotope analysis reveals that feasts in the
Stonehenge environs and across Wessex drew people and animals from throughout Britain. Sci. Adv. 5, u6078. (doi:10.1126/sciadv.aau6078)
106. Halley DJ, Rosvold J. 2014 Stable isotope analysis and variation in medieval domestic pig husbandry practices in northwest Europe: absence of
evidence for a purely herbivorous diet. J. Archaeol. Sci. 49, 1–5. (doi:10.1016/j.jas.2014.04.006)
107. Guiry E, Jones BM , deFrance S, Bruseth JE, Durst J, Richards MP. 2018 Animal husbandry and colonial adaptive behavior: isotopic insights from
the La Belle shipwreck fauna. Hist. Archaeol. 52, 684–699. (doi:10.1007/s41636-018-0142-7)
108. Balasse M, Cucchi T, Evin A, Bălăşescu A, Frémondeau D, Horard-Herbin MP. 2018 Wild game or farm animal? Tracking human-pig relationships
in ancient times through stable isotope analysis. In Hybrid communities, pp. 81–96. London, UK: Routledge. (doi:10.4324/9781315179988-5)
109. Cucchi T, Dai L, Balasse M, Zhao C, Gao J, Hu Y , Yuan J, Vigne JD. 2016 Social complexification and pig (Sus scrofa) husbandry in ancient China: a
combined geometric morphometric and isotopic approach. PLoS ONE 11, e0158523. (doi:10.1371/journal.pone.0158523)
110. Vaiglova P, Reid REB, Lightfoot E, Pilaar Birch SE, Wang H, Chen G , Li S, Jones M, Liu X . 2021 Localized management of non-indigenous animal
domesticates in Northwestern China during the Bronze Age. Sci. Rep. 11, 15764. (doi:10.1038/s41598-021-95233-x)
111. Guiry EJ, Staniforth M, Nehlich O, Grimes V, Smith C, Harpley B, Noël S, Richards MP. 2015 Tracing historical animal husbandry, meat trade, and
food provisioning: a multi-isotopic approach to the analysis of shipwreck faunal remains from the William Salthouse, Port Phillip, Australia. J.
Archaeol. Sci. 1, 21–28. (doi:10.1016/j.jasrep.2014.10.001)
112. McMahon KW, Newsome SD. 2019 Amino acid isotope analysis: a new frontier in studies of animal migration and foraging ecology. In Tracking
animal migration with stable isotopes (eds KA Hobson, LI Wassenaar), pp. 173–190, 2nd edn. Waltham, MA: Academic Press. (doi:10.1016/
B978-0-12-814723-8.00007-6)
113. Thomas R. 2019 Nonhuman animal paleopathology—are we so different? In Ortner’s identification of pathological conditions in human skeletal
remains, pp. 809–822. San Diego, CA: Elsevier. (doi:10.1016/b978-0-12-809738-0.00023-5)
114. Bendrey R, Martin D. 2022 Zoonotic diseases: new directions in human-animal pathology. Int. J. Osteoarchaeol. 32, 548–552. (doi:10.1002/oa.
2975)
115. Alldritt I et al. 2019 Metabolomics reveals diet-derived plant polyphenols accumulate in physiological bone. Sci. Rep. 9, 8047. (doi:10.1038/
s41598-019-44390-1)
116. Weber S, Price MD. 2016 What the pig ate: a microbotanical study of pig dental calculus from 10th–3rd millennium BC northern Mesopotamia.
J. Archaeol. Sci. 6, 819–827. (doi:10.1016/j.jasrep.2015.11.016)
117. Gleeson P. 2020 Archaeology and myth in early medieval Europe: making the gods of early Ireland. Mediev. Archaeol. 64, 65–93. (doi:10.1080/
00766097.2020.1754646)
118. Schulting R. 2014 The dating of Poulnabrone. In Poulnabrone, Co. Clare: excavation of an early Neolithic portal tomb (ed. A Lynch), pp. 93–113.
Dublin, Ireland: Stationery Office.
119. Davis SJ. 1992 A rapid method for recording information about mammal bones from archaeological sites. Ancient Monuments Laboratory Report
19/92.
120. Longin R. 1971 New method of collagen extraction for radiocarbon dating. Nature 230, 241–242. (doi:10.1038/230241a0)
121. Ambrose SH. 1990 Preparation and characterization of bone and tooth collagen for isotopic analysis. J. Archaeol. Sci. 17, 431–451. (doi:10.
1016/0305-4403(90)90007-r)
122. Szpak P, Metcalfe JZ, Macdonald RA. 2017 Best practices for calibrating and reporting stable isotope measurements in archaeology. J. Archaeol.
Sci. 13, 609–616. (doi:10.1016/j.jasrep.2017.05.007)
123. McClung LC, Plunkett G. 2020 Cultural change and the climate record in final prehistoric and early medieval Ireland. Proc. R. Ir. Acad. 120C,
129–158. (doi:10.1353/ria.2020.0014)
124. Kerr TR, Swindles GT, Plunkett G. 2009 Making hay while the sun shines? Socio-economic change, cereal production and climatic deterioration
in Early Medieval Ireland. J. Archaeol. Sci. 36, 2868–2874. (doi:10.1016/j.jas.2009.09.015)
125. Swindles GT etal. 2013 Centennial-scale climate change in Ireland during the Holocene. Earth Sci. Rev. 126, 300–320. (doi:10.1016/j.earscirev.
2013.08.012)
126. Warinner C, Tuross N. 2009 Alkaline cooking and stable isotope tissue-diet spacing in swine: archaeological implications. J. Archaeol. Sci. 36,
1690–1697. (doi:10.1016/j.jas.2009.03.034)
127. Sponheimer M et al. 2003 Nitrogen isotopes in mammalian herbivores: hair δ15N values from a controlled feeding study. Int. J. Osteoarchaeol.
13, 80–87. (doi:10.1002/oa.655)
128. Warinner C, Tuross N. 2010 Brief communication. Tissue isotopic enrichment associated with growth depression in a pig: implications for
archaeology and ecology. Am. J. Phys. Anthropol. 141, 486–493. (doi:10.1002/ajpa.21222)
129. van der Merwe NJ, Vogel JC. 1978 13C Content of human collagen as a measure of prehistoric diet in woodland North America. Nature 276,
815–816. (doi:10.1038/276815a0)
130. Ambrose SH, Norr L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of
bone collagen and carbonate. In Prehistoric human bone, pp. 1–37. Cham, Switzerland: Springer. (doi:10.1007/978-3-662-02894-0_1)
131. Hedges REM. 2003 On bone collagen–apatite‐carbonate isotopic relationships. Int. J. Osteoarchaeol. 13, 66–79. (doi:10.1002/oa.660)
28
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300

132. Sponheimer M etal. 2003 An experimental study of carbon-isotope fractionation between diet, hair, and feces of mammalian herbivores. Can.
J. Zool. 81, 871–876. (doi:10.1139/z03-066)
133. Froehle AW, Kellner CM, Schoeninger MJ. 2010 FOCUS: effect of diet and protein source on carbon stable isotope ratios in collagen: follow up to
Warinner and Tuross (2009). J. Archaeol. Sci. 37, 2662–2670. (doi:10.1016/j.jas.2010.06.003)
134. Hammer Ø, Harper DA, Ryan PD. 2001 PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4,
9.
135. Guiry E, Beglane F, Tourigny E, McCormick F, Richards MP. 2024 Supplementary material from: Pigs, people, and proximity: a 6000-year isotopic
record of pig management in Ireland. Figshare. (doi:10.6084/m9.figshare.c.7577661)
29
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 12: 241300