Chronology

Radiocarbon dates from seven pig (this study) and three cattle (Guiry et al. Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a) bones provide a clear answer to RQ1, that the fauna excavated by O’Kelly was not contemporaneous with the use of Beaker pottery in Ireland (c. 2450–2050 BC). When all 10 dates are considered, we see activity over a two-century period sometime between 2760 and 2480 cal BC (Figure 6 and 7: Model 1). These samples represent material excavated in both the eastern and western units, indicating that this temporal window captures activity across the site. When the pig dates are treated as a group, these indicate a short phase of depositional activity during the late 27th century to early 25th century cal BC (Figure 7: Model 2). Contra O’Kelly et al.’s (Reference O’Kelly, Cleary, Lehane and O’Kelly1983) hypothesis suggesting these faunal materials reflect ‘Beaker-period occupation’, this shows that these bones pre-date the adoption of this ceramic type and were contemporary with the use of Grooved Ware at Newgrange (Carlin Reference Carlin2017; Reference Carlin2018). We acknowledge that our sample size remains small, and that future dating evidence could show greater variation.

In answer to RQ2, the cohesiveness of these tightly clustered radiocarbon dates does not provide further evidence for stratigraphic mixing. Instead, they represent a unified chronological horizon pre-dating 2450 BC which may seem at odds with the multi-period nature of other deposits outside the tomb, including artefacts such as Beaker pottery that post-date 2450 BC. However, it is our understanding that the specimens from under the YCB were found in their primary context, having been protected from later disturbance by this bank, while those from over the bank seem to have been subjected to redeposition. This is based on Van Wijngaarden-Bakker’s (Reference Van Wijngaarden-Bakker1986) observation of concentrations of anatomically adjacent elements, especially vertebrae and scapulae, suggesting the deposition of intact portions of pig skeletons under the YCB. She noted that the pigs above the bank showed broadly similar but less pronounced patterns of concentrations of these elements. In contrast, those from the multi-period deposits where the bank was absent were far more disconnected. While the excavation records do not enable us to add further information to this, our Model 2 results suggest that pigs excavated from deposits both above and below the YCB were slaughtered over a very short, potentially contemporaneous, window of time and that those above were subject to later disturbance. In this scenario, the estimated construction date range for the YCB based on pig dates in Model 3 (Figure S3), which assumes that deposits from below and above the YCB were both in their primary contexts, may dramatically over-constrain the timeframe for construction of the YCB. Larger-scale dating of the Newgrange faunal assemblage has potential to add nuance to this narrative.

By contrast, our radiocarbon results are well suited for defining a terminus post quem for the YCB’s construction. Model 4, which allows for the possibility that pig samples from above the YCB are in secondary contexts — while still assuming that their initial deposition was part of the same episode of activity as the other pig samples — indicates that the YCB’s construction began after 2560–2450 cal BC (95 % probability). This terminus post quem is within the period when Grooved Ware was still in use and is before the estimated date for the adoption of Beaker pottery. This is at odds with the presence under the YCB of so-called ‘rusticated ware’, ‘Neolithic-derived bowl forms’, and ‘Beaker-associated coarse ware’, which were all originally considered part of a contemporaneous Beaker assemblage (Cleary Reference Cleary1983). However, these may well represent earlier styles of pottery, given that other Newgrange ceramics which were also treated as Beakers have subsequently been recognised as Grooved Ware (Roche Reference Roche1995; Brindley Reference Brindley1999). That accords with the absence of radiocarbon-dated pigs that overlap with the use of Beakers, either below or above the YCB. However, this can only be resolved through future work, including further reassessment of the pottery. Below, we return to considering the implications of our results for understanding the construction and function of this bank.

Interesting patterns emerge when dates from pigs are considered separately from cattle. The very short event duration modelled for pigs compared to cattle has implications for species-specific activities, which addresses RQ3 regarding taxonomic or spatial patterns in human–animal interactions (Figures 6 and 7: Model 2). The pigs date to 2600–2440 cal BC with a median duration estimate of 25 years under the simplest model (Model 2) and, at an extreme, the determinations could plausibly reflect a single event (ie less than one year). That these pig remains come from a range of locations suggests that the event or series of events that created the pig-rich deposits at Newgrange was highly constrained in time (Figure 7: Model 2). With respect to documenting human–pig interactions at Newgrange, a key aim of this paper, we prefer Model 2 over the other models because it focuses specifically on the pigs and relies on no other assumptions.

It is not clear why cattle (n = 3), with a modelled median duration estimate of 678 years, should show more temporal spread than pigs (n = 7, median duration estimate of 25 years). While the larger sample size for pigs might have been expected to produce somewhat tighter duration estimates, even artificially reducing sample size to only three specimens to match the cattle (by sampling all possible combinations of three pig dates) still resulted in a maximum median duration estimate of only 117 years (Figure S2). We believe the difference between cattle and pigs reflects a real pattern in the temporality of human–animal interactions at the site (see Appendix S8) but note that additional dates from cattle are needed to explore this interspecific pattern further.

In contrast to the cattle, the floruit of pig dates suggests an intense phase of pig-related activities representing feasting. If we assume that the seven pig dates are representative of the pig-centric nature of the assemblage, given the MNI evidence that at least 200 pigs (>150 from the western unit alone) were slaughtered and deposited in the ‘Beaker layers’, the shortest extreme of the modelled durations (ie an event duration of one year) seems less plausible. According to Van Wijngaarden-Bakker’s (Reference Van Wijngaarden-Bakker1986) calculations, based on MNI, these pigs represent a minimum of about 12,000 kg of dressed meat and fat, and such an outlay of livestock being expended for a single event seems unlikely, although not impossible, particularly if not all the meat was eaten, and some was instead deposited, perhaps as offerings. Assuming feasting consumption of a generous 2 kg of meat per person per day this would equate to 6000 portions, or enough to feed over 850 people for a week (Appendix S9). It is nevertheless easier to imagine a scenario in which these pigs reflect episodic feasting events, involving the consumption of very considerable quantities of meat, which took place over a period of a few years to a few decades. Notably, the shorter duration (be it one, five, or 50 years) we have observed in these modelled dates complements other patterns in the distribution and isotope compositions of pig bones at the site. Dense concentrations of pig remains, including numerous associations of specific elements consistent with the deposition of articulated portions, as well as a specialised and yet broadly consistent set of husbandry practices (see below), suggest that at least some of these deposits derive from a relatively short time span.

The temporally constrained nature of the pig dates suggests that the feasting events or processes that led to the pig-rich deposits may have been connected to the subsequent construction of the YCB, and potentially both may have served related goals. They were preceded by the construction of the large timber enclosure, but they all form part of a dynamic series of intense episodes of Grooved Ware-associated activity at the front of the large passage tomb in the two centuries c. 2700–2450 cal BC, immediately following the closure of its entrance (Carlin Reference Carlin2017; Smyth et al. Reference Smyth, Geber, Carlin, O’Sullivan, Griffiths, Hofmann, Cummings, Bjørnevad-Ahlqvist and Iversen2025). Evidence for intense midwinter pig feasting events like those at Newgrange has been identified at various timber or earthen monuments, including Durrington Walls, Mount Pleasant and West Kennet palisaded enclosures in southern Britain (eg Craig et al. Reference Craig and Parker Pearson2015; Bayliss et al. Reference Bayliss, Cartwright, Cook, Griffiths, Madgwick, Marshall, Reimer, Bickle, Cummings and Hofmann2017; Greaney et al. Reference Greaney, Hazell, Barclay, Ramsey, Dunbar, Hajdas, Reimer, Pollard, Sharples and Marshall2020). These large-scale seasonal gatherings of people and animals also occurred for short durations during the 26th and 25th centuries and were directly associated with monument building activities and the use of Grooved Ware. The Newgrange hearths may well have played a role in the cooking of pigs during Grooved Ware-associated feasting activities, similar to examples within small buildings at Durrington Walls, where large-scale feasting remains were enclosed by a henge in the 25th century BC (Craig et al. Reference Craig and Parker Pearson2015).

The resemblance of the Newgrange YCB to the banks of nearby embanked enclosures or henges (their counterparts in Britain) suggests that the YCB was built to further encircle the exterior of the passage tomb; other passage tombs at Brú na Bóinne are enclosed in the same way (Cooney Reference Cooney2006; Carlin Reference Carlin2017; Davis & Rassmann Reference Davis and Rassmann2021). While the original extent of the YCB is hard to reconstruct, by circumscribing the tomb’s southern and western perimeter, this bank directly continues a longer sequence of incomplete enclosing actions around the monument. However, it remains unclear if the absence of the YCB from the excavated area east of the tomb’s entrance is because it was subsequently removed or simply never constructed. O’Kelly et al. (Reference O’Kelly, Cleary, Lehane and O’Kelly1983) suggested it may have been created specially to cover earlier features, such as the underlying stone hearths. More recently, these hearths have been regarded as evidence for Grooved Ware-associated activities because of their resemblance to those found within Orcadian dwellings (Richards Reference Richards2005; Smyth Reference Smyth2011). Both interpretations of the YCB can be combined, based on comparison with the contemporaneous Grooved Ware-associated practices within henges in Britain, which were built through communal acts of labour that enclosed and memorialised the location of recent past activities involving feasting and social gatherings (eg Brophy & Noble Reference Brophy and Noble2012; Pollard Reference Pollard2012).

Husbandry

Isotopic data reveal the extent to which Newgrange pig husbandry was focused on woodland resources (RQ4) and, in turn, offer insights on seasonality of deposition at the site (RQ5) and more broadly, animal husbandry across Ireland (RQ6). Because we found no significant differences between contexts for both isotopic compositions, all pig data are treated as one interpretive unit. We consider Newgrange pig isotopic compositions in two ways. First, we compare pig data with published cattle isotopic compositions from Newgrange (Guiry et al. Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a). Cattle are herbivorous but are likely to have fed more on grass and foliage than tree mast, which is poisonous to them (Cedervall et al. Reference Cedervall, Johansson and Jonsson1973; Dixon et al. Reference Dixon, McPherson, Rowland and Maclennan1979; González-Barrio et al. Reference González-Barrio, Truchado, García-Villalba, Hervás, Frutos, Espín and Tomás-Barberán2012). This means that cattle can provide a baseline against which to compare more ecologically and dietarily flexible pigs (see also Appendix S3). Second, we consider the Newgrange pig data within the wider corpus of isotope evidence for pig husbandry spanning the Neolithic to post-medieval periods in Ireland (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025).

The relationship between Newgrange pig and cattle δ ¹⁵N, with pigs producing a mean δ ¹⁵N value that is 0.7 ‰ lower than that of cattle, may seem surprising. This is because pigs and cattle are omnivores and herbivores, respectively, and δ ¹⁵N has traditionally been seen as an indicator of trophic position, such that animals incorporating meat into their diets are expected to show higher δ ¹⁵N than animals that do not (DeNiro & Epstein Reference DeNiro and Epstein1981). This serves to underscore that animal protein, including from aquatic sources, was likely unimportant in the diets of pigs slaughtered at Newgrange. The question remains, however, of why pigs, even those that were largely herbivorous, should have lower average δ ¹⁵N values than cattle at the same site. Acknowledging the many complexities of nitrogen sources and cycling (Szpak Reference Szpak2014), we consider three possible explanations in Appendix S10. We find that while it is not possible to establish definitively whether these animals were raised in the same area(s), and therefore the extent to which their δ ¹⁵N values should be comparable, it is parsimonious to proceed on basis that they were.

It is worth reviewing key temporal and biological dimensions of our interpretive context. Given the timing of mast abundance in Ireland’s woodland habitats (masting occurring in autumn, September–October; Kelly Reference Kelly1997) and the refined slaughter profile for pigs at Newgrange described above (potentially starting as early as August, but with a peak likely focused later in the year and completed by the end of the calendar year) we would expect isotope compositions of mast-fed pigs to show ¹³C enrichment especially in those slaughtered shortly after a period of mast-driven weight gain. There are two reasons for this. First, because mast as a non-photosynthetic tissue is ¹³C enriched, relative to foliage from the same plants (Cernusak et al. Reference Cernusak, Tcherkez, Keitel, Cornwell, Santiago, Knohl, Barbour, Williams, Reich and Ellsworth2009), there should be a clear signal in the bone collagen of mast consumers (Guiry et al. Reference Guiry, Orchard, Royle, Cheung and Yang2020; Reference Guiry, Orchard, Needs-Howarth and Szpak2021), with a difference of +2 to +4 ‰ being common (eg Sarà & Sarà Reference Sarà and Sarà2007; Guiry et al. Reference Guiry, Orchard, Needs-Howarth and Szpak2021).

Second, while the isotopic compositions of bone collagen are generally thought to reflect long-term (multiyear, or multidecadal) averages of consumers’ diets, these signals will be weighted toward periods of faster growth (Hobson & Clark Reference Hobson and Clark1992; Hyland et al. Reference Hyland, Scott, Routledge and Szpak2021). This will particularly be the case for animals slaughtered before or soon after growth plateaus (Warinner & Tuross Reference Warinner and Tuross2010). Pigs at Newgrange were largely dispatched between 27 and 30 months and, compared to older skeletally mature pigs, will have isotopic compositions that are more influenced by recently consumed foods. Food surfeits during the mast season are known to drive periods of accelerated weight gain (potentially more than doubling pig weights in a two- to three-month pannage period, if modern traditional breed Iberian figures can be transposed; eg Parsons Reference Parsons1962). This would be associated both with continued skeletal growth and considerable skeletal remodelling as the musculoskeletal system responds to changes in load bearing. Because the majority of pigs at Newgrange appear to have been slaughtered during the autumn–winter, we would expect isotopic signals associated with mast feeding, which are being recorded during a period of accelerated skeletal growth and adjustment, to have been amplified relative to periods of dietary intake earlier in life when growth was slower. It is important to note that pigs that were dispatched later, in the following year, would be expected to have lower δ ¹³C values, reflecting proportionately more materials from subsequent (non-mast) dietary intake.

Cattle from all time periods at Newgrange (n = 38) have δ ¹³C values consistently falling towards the lower end of the wider pan-Irish cattle isotope dataset (n = 388; Figure 4; Guiry et al. Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a). At the other end sit cattle from early medieval sites (n = 71), when cattle husbandry had become a key driver of land management practices (Kelly Reference Kelly1997; Guiry et al. Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a). Because cattle grazing and foddering focuses on foliage, and does not preferentially incorporate mast (Cedervall et al. Reference Cedervall, Johansson and Jonsson1973; Dixon et al. Reference Dixon, McPherson, Rowland and Maclennan1979; González-Barrio et al. Reference González-Barrio, Truchado, García-Villalba, Hervás, Frutos, Espín and Tomás-Barberán2012), the primary driver of δ ¹³C variation in Ireland is likely to be the canopy cover (Bonafini et al. Reference Bonafini, Pellegrini, Ditchfield and Pollard2013; for review see Guiry et al. Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a). In that context, lower cattle δ ¹³C at Newgrange suggests that many of these individuals were husbanded in (or with resources from) environments with denser canopy cover, likely woodland habitats, though natural landscape variation must also be expected. By contrast, cattle from the early medieval period, which represent animals more often grazed on open lands, provide a reference point for the highest δ ¹³C values that would be expected from non-mast feeding herbivores.

These patterns offer two important anchor points for interpreting pig data. First these data confirm that woodland habitat was likely being used for the husbandry of animals that were slaughtered at Newgrange. Second, cattle provide a baseline for the range of δ ¹³C variation expected in domestic herbivorous diets that do not incorporate tree mast. Using this interpretive framework, we can now address RQ4, which asks whether pig husbandry at Newgrange was, as hypothesised by McCormick (Reference McCormick, Murphy and Whitehouse2007), strongly woodland oriented. Our finding that pigs have significantly higher δ ¹³C relative to cattle at Newgrange, and even compared to cattle from the early medieval period (Figure 4) supports this, pointing to a husbandry regime in which pigs consumed large amounts of tree mast before slaughter (Appendix S11).

Fungus consumption has also been considered in previous research (Hamilton et al. Reference Hamilton, Hedges and Robinson2009; see Appendix S3 for further discussion). Pigs foraging in woodland habitats often have access to fungi, and these could be ¹³C enriched relative other C₃ foods. However, fungi are also typically ¹⁵N enriched relative to local plants with which they have symbiotic relationships (Hobbie & Högberg Reference Hobbie and Högberg2012; Szpak Reference Szpak2014; O’Regan et al. Reference O’Regan, Lamb and Wilkinson2016). Therefore, given that our pig δ ¹⁵N values are on average lower than those of herbivorous cattle (Appendix S10), this explanation is unlikely to be a primary driver of the higher pig δ ¹³C values at Newgrange.

We can also examine how Newgrange pig isotopic compositions compare with a wider dataset across the Irish Holocene (Figure 5B; Knudson et al. Reference Knudson, O’Donnabhain, Carver, Cleland and Price2012; Guiry et al. Reference Guiry, Beglane, Szpak, Schulting, McCormick and Richards2018; Reference Guiry, Beglane, Szpak, McCormick, Teeter, Cheung and Richards2023a; Reference Guiry, Beglane, McCormick, Tourigny and Richards2025; Madgwick et al. Reference Madgwick, Grimes, Lamb, Nederbragt, Evans and McCormick2019a; McKenzie et al. Reference McKenzie, Murphy, Guiry, Donnelly and Beglane2020). Isotopic compositions from Neolithic and Early Bronze Age pigs are extremely rare, but the small pool of existing data shows wide ranging δ ¹³C and very low δ ¹⁵N, suggesting diverse husbandry practices, including both woodland and non-woodland oriented husbandry, with no evidence for diets with high quantities of animal protein (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025). Within this context, pigs from Newgrange show less variable diets (Figure 5C) with a similar degree of herbivory (Figure 5D) and a greater emphasis (ie slightly higher mean δ ¹³C) on mast feeding (Figure 5A). The mast-focused diets of Newgrange pigs are particularly distinctive when compared with pigs from the Middle–Late Bronze and Iron Ages (n = 119; Figure 5A), which show more variable diets (Figure 5C), with a trend to lower δ ¹³C and a general uptick in δ ¹⁵N. This increase in δ ¹⁵N has previously been observed in published Irish pig isotopic compositions and is thought to be at least partly explained by broadscale human impacts on the cycling of nitrogen in soil brought on by changing land management practices in this period (Guiry et al. Reference Guiry, Beglane, Szpak, Schulting, McCormick and Richards2018; for an English example see Madgwick et al. Reference Madgwick, Mulville and Stevens2012). In that context, as with Newgrange and other earlier prehistoric sites, pigs from Middle–Late Bronze and Iron Age sites were largely herbivorous (Figure 5D). Pigs from the medieval and post-medieval periods show more extreme variation (Figure 5C) centered on moderate δ ¹³C but higher δ ¹⁵N values (Figures 5A and 5D). The extent to which δ ¹⁵N is elevated for this latter group reflects prevailing increases in animal protein consumption (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025).

Against this backdrop, we can now address RQ6, which queries whether there are parallels between the zooarchaeological evidence for an unusual abundance of pigs at Newgrange and isotopic evidence for specialisation in how those pigs were raised. We acknowledge that the extent to which we can address this is limited by generally smaller sample sizes at comparison sites, although broad insights can still be gained by comparing the Newgrange results with pig populations at wider spatial and temporal scales. For contemporaneous and later sites, isotopic compositions of Newgrange pigs show on average higher (statistically significant for all but the Neolithic/Early Bronze Age group) δ ¹³C values (Figure 5A). It is therefore clear that at Newgrange a unique pattern in pig husbandry was practised, involving a stronger focus on woodland resources than was common at other times or in most areas across Ireland for which we have data. Moreover, the supplemental animal protein sources (eg food scraps, agricultural byproducts) which appear increasingly important to Irish pig diets through the historical period (Figure 5D) are notably absent at Newgrange, as has been observed more widely for Neolithic/Early Bronze Age pigs across Ireland (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025).

Pannage and woodlands

These comparisons between pig diets at Newgrange and broader spatiotemporal trends serve to further underscore the significance of woodland resources. In his review of early medieval texts on Irish farming, Kelly (Reference Kelly1997) found that pigs maintained a high profile across major textual sources. These highlight the antiquity of associations between pigs and oak woodlands in Irish agriculture (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025). Kelly’s (Reference Kelly1997) analyses of law texts, reflecting ecological knowledge on pig husbandry, indicate that, while pigs could be fed a variety of foodstuffs, mast was widely thought to be the most important. An excellent example is seen in laws that establish a legal distinction between acorn-fed pigs (muca for mesruth) and those raised on any other food (muca dema) (Kelly Reference Kelly1997). By the early medieval period, pig husbandry practices were so deeply entwined with oak woodland in Irish thinking that they had become proverbial. A notable example from the tale Togail Bruidne Da Derga (the destruction of Da Derga’s hostel), highlighted by Kelly (Reference Kelly1997), is the phrase ‘a pig which dies before the acorn crop’ (‘mucc remi-thuit mess’) which is used as a metaphor for a lost opportunity.

This historical context helps to further constrain isotopic interpretations. Given the entrenched connections between pigs and oak woodland in the early medieval period and given that the natural habitat of wild boar is woodland, it is possible that these concepts represent older traditional ecological knowledge. Combining documentary and isotopic data from the Iron Age and medieval periods (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025), it is clear that pig δ ¹³C values from these periods represent animals that, on average, had diets partly, or largely, focused on mast.

That Newgrange pig δ ¹³C values are on average higher than those from these later time periods is telling (Figure 5A; Appendix S12; Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025) and suggests that pig husbandry at Newgrange was even more closely tied to woodland resources than in the early medieval period, when the pig–pannage connection was first documented in Ireland. However, given the consistency in the revised fusion and tooth age-at-death estimates, higher δ ¹³C values could simply reflect highly regularised seasonal slaughter shortly after pannage-driven weight gain. Even so, it is tempting to suggest that the traditional ecological knowledge codified in the early medieval period is rooted in practices from a much deeper past. Indeed, based on the isotopic patterns observed in Newgrange pigs, it seems likely that the meaning of the early medieval proverb highlighted by Kelly would be easily recognised by those feasting at Newgrange 3500 years earlier.

Given that both the cattle and pigs slaughtered at Newgrange in the earlier 3rd millennium BC were husbanded in very specific ways that were intimately connected with woodlands, it is relevant to consider where these woodland habitats were located. Related to this, significant oak resources would have been required for the ongoing construction of timber monuments within Brú na Bóinne between c. 3000–2450 BC (Davis & Rassman Reference Davis and Rassmann2021). One might presume that pigs at Newgrange were being fed in dense oak woodland nearby. While current paleoenvironmental data lack sufficient detail to conclusively address this (Appendix S2), there is no indication that suitable mast-rich woodlands were present in the immediate area. Instead, Newgrange was seemingly in a largely open landscape with a pastoral character, generally regarded as prime land for mixed farming (Cooney Reference Cooney2000; Davis et al. Reference Davis, Lewis, Turner, Foster, Ostericher, O’Donnell, Collins, McClatchie, Allen and McCormick2017).

This suggests that the woodland-oriented husbandry practices we have observed reflect at least some degree of animal movement, a possibility that was also suggested by Van Wijngaarden-Bakker (Reference Van Wijngaarden-Bakker1986) for cattle. As detailed above, both the cattle and pigs at Newgrange display very specific age structures. While this may be taphonomic, it could also indicate that the animals consumed at Newgrange were neither born nor raised there. If they did come from a wider catchment, then the cohesive approach to their husbandry is even more striking. Such a possibility is supported by the isotopic evidence from the pigs slaughtered during midwinter feasting at Durrington Walls, which show high δ ¹³C values like those observed at Newgrange, but which also have other isotopic compositions that suggest diverse geological origins (Madgwick et al. Reference Madgwick, Lamb, Sloane, Nederbragt, Albarella, Parker Pearson and Evans2019b). There, the interpretation was that pigs were brought to Durrington Walls from up to 100 km away by various communities who assembled for midwinter feasting and monument building activities. It was also suggested that the number of pigs being raised specifically for midwinter feasting may have been so large that they could not be supported by locally available resources and were instead driven from settlements to woodlands specifically to be fattened on mast, before being brought to Durrington Walls.

There is considerable overlap in the evidence for animal husbandry and mobility practices, as well as short-lived feasting activity and associated monument building observed at Durrington Walls and Newgrange during the 26th and 25th centuries. However, this forms part of a much longer-term pattern of assembly of people, animals, and materials at Newgrange and Brú na Bóinne, at certain times of the year. At least five centuries previously, the Brú na Bóinne passage tombs incorporated multiple types of local and non-local stone and other materials brought from various places up to 40 km away by those who seasonally journeyed to participate in ongoing construction and ceremonial activities (Cooney Reference Cooney2000; Brady 2007; Reference Brady2018; Eogan & Cleary Reference Eogan and Cleary2017; Carlin et al. Reference Carlin, Smyth, Frieman, Hofmann, Bickle, Cleary, Greaney and Pope2025). As discussed below, Newgrange and Knowth continued to be a focus for large-scale gatherings and ceremonial activity until at least c. 2450 BC. During this time, these monuments and their surroundings saw an ongoing series of changes: tomb construction ceased by 2900 BC, whereas earthen or timber monument building peaked after 2700 BC. Yet, many aspects of the passage tomb-associated ceremonial activity, including communal gatherings, persisted until at least c. 2450 BC, (Carlin Reference Carlin2017; Eogan & Cleary Reference Eogan and Cleary2017). It would not be surprising if pigs, cattle, and oak timbers were also brought there in significantly increased numbers during the first half of the 3rd millennium BC. This is an issue requiring further work that can be readily addressed through other, widely available isotopic techniques and additional radiocarbon dating (see below).

Winter solstice at Newgrange after 2700BC

The temporality of human–pig–environment relationships has implications for our understanding of broader cultural and temporal trends at Newgrange. We have not generated comparable isotopic data from fauna contemporaneous with the construction and use of the large tomb at Newgrange (c. 3305–2700 BC; Lynch et al. Reference Lynch, McCormick, Shee Twohig, McClatchie, Molloy, Schulting, OCarroll and Sternke2014) so it is not possible to evaluate whether woodland-focused pig husbandry represented a new development during the 25th or 26th centuries BC or, more likely, was a continuation of even older traditions (Guiry et al. Reference Guiry, Beglane, McCormick, Tourigny and Richards2025). However, by considering the cycle of pig husbandry, we can address RQ5 (seasonality of site use) to explore a core theme — the winter solstice — around which activity at the site had been organised.

Considering the ways in which these isotopic analyses build on and refine existing zooarchaeological narratives is key to informing this exploration. The importance of mast supports Van Wijngaarden-Bakker’s (Reference Van Wijngaarden-Bakker1986) assessment that most pigs were dispatched within a six-month window in the autumn–winter seasons. While her methods could not be more specific about when this occurred, our isotopic analyses and refinement of the toothwear and bone fusion data suggests that slaughter was shortly after mast feeding had propelled pigs to their peak annual weight. Taking the medieval pannage season (September–October; Kelly Reference Kelly1997) as a benchmark, this suggests that pigs would have been slaughtered around the turn from autumn to winter, likely in the November to December window. We acknowledge slaughtering pigs at this time of year may reflect secular and/or economic factors, such as cooler weather facilitating easier meat preservation and the need to cull animals before they become harder to feed in winter, but as noted above, the scale, spatial patterning, and tight temporality of the pig assemblage at Newgrange all strongly suggest this pattern reflects a broader cultural phenomenon.

If we accept, as most now do (Mount Reference Mount1994; McCormick Reference McCormick, Murphy and Whitehouse2007; Carlin Reference Carlin2018), that these pigs reflect ritual, ceremonial, or at least social, feasting, rather than settlement, then it is difficult to ignore the timing of these activities. Together these lines of evidence suggest that people were physically present for feasting events leading up to and incorporating the winter solstice between the 26th and 25th centuries BC. The extent to which this directly echoes aspects of Newgrange’s construction and use (Hensey Reference Hensey2015), especially the gathering of people for the winter solstice, seems more than a coincidence. It is also difficult to ignore the location of these feasting remains outside the entrance. This deposition seems to represent a continuation of a long-standing tradition of enhancing the front of developed passage tombs through activities that emphasised their entrances (eg Carlin Reference Carlin2017). It is, therefore, reasonable to suggest that in the context of Newgrange, the greater concentration of faunal remains near the entrance relates directly to the celebration of the winter solstice, a phenomenon that had been a defining feature of this monument’s design, construction, and use since 3300 BC. Even though the sealing of the tomb c. 2700 BC prevented the beams of the winter solstice sunrise from being witnessed as they illuminated the passage, the solstice seems to have remained important. In that context, and addressing RQ5, we offer an alternative hypothesis for the origins of this faunal assemblage: that these pig remains reflect winter solstice-oriented ‘ceremonial’ activities outside the Newgrange tomb for at least two centuries (2700–2450 BC) later than previously thought.

This new evidence for the persistence of midwinter gatherings at Newgrange into the 3rd millennium BC confirms the extent to which emerging cultural influences from across the Irish Sea, such as Grooved Ware-associated feasting were incorporated within the Irish passage tomb tradition (see Carlin Reference Carlin2017; Carlin & Cooney Reference Carlin, Cooney, Card, Edmonds and Mitchell2020). Our results demonstrate the evolving nature of activity at Newgrange throughout this millennium, indicating continuity of both place and practice; Newgrange certainly did not go into decline either with the closure of the tomb, or the adoption of Grooved Ware in Ireland. Instead, activities continued at Newgrange during much of the 3rd millennium BC in much the same ways as beforehand, often directly referencing the pre-existing passage tomb monuments (Carlin & Cooney Reference Carlin, Cooney, Card, Edmonds and Mitchell2020). This includes concentrations of ceramics found outside the entrance (O’Kelly et al. Reference O’Kelly, Cleary, Lehane and O’Kelly1983) and in the timber structures that enclosed another nearby tomb (Site Z), resulting from Grooved Ware-associated feasts during communal gatherings and ceremonial undertakings between 2700–2450 BC (for reviews, see Carlin 2017; Reference Carlin2018). There is a growing body of evidence, albeit much of it from geophysical and aerial survey, and therefore undated, suggesting that activity around Newgrange, including monument construction using oak timbers and/or earth rather than stone, escalated considerably between c. 3000–2450 BC. Furthermore, some of these structures and enclosures also display a concern with solstice alignments (Carlin & Cooney Reference Carlin, Cooney, Card, Edmonds and Mitchell2020; Davis & Rassmann Reference Davis and Rassmann2021). It is therefore reasonable to suggest that those who continued to occupy this landscape were aware of the significance of the winter solstice at Newgrange and maintained a concern for solar events.

If this is true, another question emerges — when did this knowledge disappear? This has been explored by many researchers (Carey Reference Carey1990; Stout & Stout Reference Stout and Stout2008; Waddell Reference Waddell2014; Hensey Reference Hensey, Papadopoulos and Moyes2017). O’Kelly (Reference O’Kelly1982) considered sun-oriented themes in medieval narratives such as De Gabáin int Síde (the taking of the otherworld mound) and Tochmarc Étaíne (the wooing of Étaín) as potential evidence for the preservation of knowledge of Newgrange’s solar orientation. In that context, he asked if it could be that the tomb builders also ‘planted the seeds of Irish oral literature’, even pondering whether such medieval narratives could now be seen ‘not as a window on the Iron Age but as one on the Late Neolithic?’ More recent scholarship has highlighted potential pitfalls in transposing knowledge from these medieval narratives, which may at least partly represent coeval inventions, into the prehistoric past (Gleeson Reference Gleeson2020). Acknowledging this issue, our findings still show that concentrated activities, including feasting, occurred around the time of the winter solstice, centuries after the monument was sealed off. It is striking to reflect that such activities continued some generations after the last people experienced the solstice phenomenon from within the monument, suggesting a considerable degree of collective memory.