Land use and mobility during the Neolithic in Wales explored using isotope analysis of tooth enamel

Abstract Objectives The nature of land use and mobility during the transition to agriculture has often been debated. Here, we use isotope analysis of tooth enamel from human populations buried in two different Neolithic burial monuments, Penywyrlod and Ty Isaf, in south‐east Wales, to examine patterns of land use and to evaluate where individuals obtained their childhood diet. Materials and Methods We employ strontium (87Sr/86Sr) and oxygen (δ18O) and carbon (δ13C) isotope analysis of enamel from adjacent molars. We compare strontium isotope values measured in enamel to locally bioavailable 87Sr/86Sr values. We combine discussion of these results with evaluation of new radiocarbon dates obtained from both sites. Results The majority of enamel samples from Penywyrlod have strontium isotope ratios above 0.7140. In contrast, the majority of those from Ty Isaf have 87Sr/86Sr values below 0.7140. At Penywyrlod oxygen isotope ratios range between 25.9 and 28.2 ‰ (mean 26.7 ± 0.6 ‰, 1σ, n = 15) and enamel δ13Ccarbonate values range between −18.0 and −15.0 ‰ (mean −16.0 ± 0.8 ‰, 1σ, n = 15). At Ty Isaf oxygen isotope ratios exhibited by Neolithic individuals range between 25.4 and 27.7 ‰ (mean 26.7 ± 0.6 ‰, 1σ, n = 15) and enamel δ13Ccarbonate values range between −16.9 and −14.9 ‰ (mean −16.0 ± 0.6 ‰, 1σ, n = 15). Discussion The strontium isotope results suggest that the majority of individuals buried at Penywyrlod did not source their childhood diet locally. One individual in this group has strontium isotope ratios that exceed all current known biosphere values within England and Wales. This individual is radiocarbon dated to the first few centuries of the 4th millennium BC, consistent with the period in which agriculture was initiated in Wales: the results therefore provide evidence for migration during the transition to farming in Wales. In contrast, all individuals sampled from Ty Isaf post‐date the period in which agriculture is considered to have been initiated and could have sourced their childhood diet from the local region in which they were buried.

The nature of land use and residence patterns during the Neolithic in Britain has also been the subject of intense debate. Some authors suggest the first farmers in Britain were fully sedentary and argue that communities obtained the majority of their dietary resources, keeping livestock and cultivating crops, close to permanently occupied settlements (Bogaard et al., 2013, pp. 12589;Rowley-Conwy, 2003. These authors place emphasis on the discovery of substantial timber buildings and argue that mobility of early farming communities was limited by the demands of cereal cultivation (Jones, 2000(Jones, , 2005Jones & Rowley-Conwy, 2007;Rowley-Conwy, 2000Rowley-Conwy & Legge, 2015).
Others, however, suggest that there was greater diversity in subsistence and settlement patterns. These authors highlight evidence for exploitation of a varied range of resources during the early Neolithic in Britain, including use of both wild plants and cereals (e.g., Bishop, Church, & Rowley-Conwy, 2009, pp. 86, 90;Stevens, 2007, pp. 381-382) and suggest that evidence for cattle herding could indicate that some members of the community were residentially mobile (Thomas, 2013, pp. 411;e.g., Schulting, 2013, pp. 321;Serjeantson, 2011;Viner, 2011). A varied range of evidence for occupation may also be indicative of diversity in settlement systems during the early Neolithic in Britain (e.g., Anderson-Whymark & Thomas, 2012;Brophy, 2015;Sheridan, 2013). In addition to the remains of substantial timber buildings, the presence of ephemeral structures, pits, lithic scatters and middens could indicate communities moved episodically between occupation sites that were located in different geographical areas to obtain their dietary resources (Garrow, Beadsmoore, & Knight, 2005, pp. 155;Whittle, 2003, pp. 43).
In view of these debates we applied strontium ( 87 Sr/ 86 Sr) and oxygen (d 18 O) isotope analysis of tooth enamel, which can be used to provide evidence for the location from which individuals may have obtained their childhood diet (e.g., Bentley, 2006;Montgomery, 2002Montgomery, , 2010Slovak & Paytan, 2011, pp. 743-744), to explore patterns of land use and mobility during the Neolithic. Here we present the results of analysis of individuals from two burial monuments, the long cairns of Penywyrlod (Talgarth) and Ty Isaf, located less than 5 miles (8 km) apart amongst a cluster of monuments in the Black Mountains, Powys, southeastern Wales (Figure 1). Currently archaeological evidence for the nature of settlement in this region is very limited, few structural remains of Neolithic date have so far been discovered and surface lithic scatters dominate the archaeological record (Makepeace, 2006;Olding, 2000).
The long cairns of Penywyrlod (Talgarth) and Ty Isaf are part of the Cotswold-Severn group, a concentration of burial monuments located in the regions around the Severn Estuary: south-east Wales, the Cotswolds, Somerset and Wiltshire in southern Britain (Darvill, 2004, pp. 71-72). Although sharing a common geographical distribution, long cairns within the Cotswold-Severn group are diverse in form. Penywyrlod (Talgarth) is a substantial long cairn over 60 metres in length. Partial excavation of the north-eastern side of the monument revealed three lateral chambers which contained co-mingled and disarticulated human remains ( Figure 2, Luff, Brothwell, & O'Connor, 1984;Savory, 1973Savory, , pp. 187, 1984. Following excavation a sample of human bone recovered from one of the excavated chambers (chamber NEII) was radiocarbon dated to between 3960 and 3640 cal BC (95% confidence, OxCal v. 4.2; Britnell & Savory, 1984, pp. 29). In contrast, Ty Isaf is approximately 30 metres in length and is a composite monument consisting of two distinct elements: a long cairn with two opposing lateral chambers (chambers 1 and 2; Figure 3 over page) and a circular rotunda containing a transepted passage grave (chamber 3; Figure 3; Grimes, 1939, pp. 123-124). Human remains within the transepted passage grave in the rotunda date from the mid 4 th to earlier 3 rd millennium BC (Bayliss et al., 2011b, pp. 537, 546-547, see Radiocarbon Dating below).
Strontium isotope analysis is undertaken on enamel as it is highly resistant to diagenesis (e.g., Budd, Montgomery, Barreiro, & Thomas, 2000;Trickett, Budd, Montgomery, & Evans, 2003 Montgomery et al. (2006) and . Locally bioavailable strontium isotope ratios on Devonian sandstones derived from measured values in plants collected within a radius of 10 miles, approximately 16 km, of Penywyrlod and Ty Isaf (Table 3) NEIL ET AL. | 373 levels (Graustein, 1989, pp. 492;e.g., Blum, Taliaferro, Weisse, & Holmes, 2000) and as enamel does not remodel once mineralized, strontium isotope ratios directly reflect sources to which an individual was exposed during tooth formation (Bentley, 2006;Montgomery, 2002). Strontium isotope ratios in modern plants and waters have been shown to vary with the age and composition of the underlying lithology (e.g., Evans, Montgomery, Wildman, & Boulton, 2010;Warham, 2011;Willmes et al., 2013). Comparison of 87 Sr/ 86 Sr values to mapped bioavailable values can therefore be used to evaluate whether an individual obtained dietary resources from the area in which they were later buried or whether they sourced their diet from a region further afield.  Evans, Eckardt, & Lewis, 2010, pp. 155;Evans et al., 2010;Montgomery, Evans, & Wildman, 2006, pp. 1628 Evans et al., 2010;Montgomery et al., 2006Montgomery et al., , pp. 1628Shand, Darbyshire, Gooddy, & Haria, 2007, pp. 254, 256), highlighting the potential for heterogeneity in bioavailable strontium isotope ratios on lithology of Lower Palaeozoic age. Samples of deep groundwater taken on lithology of Lower Palaeozoic age in central Wales can give values up to 0.7152 (Shand et al., 2007, pp. 256 et al., 2010, pp. 155). However, as rocks of this age crop out in a very limited geographical area, values such as this have so far remained difficult to reproduce (Lucie Johnson, pers. comm.), with water sampled close to the Malvern Hills having given a value of 0.7132 (Montgomery et al., 2006(Montgomery et al., , pp. 1628) and plants recording a mean 87 Sr/ 86 Sr value of 0.7128 6 0.0040 (2r, n 5 13, Chenery et al., 2010, pp. 155-156).

| Oxygen and carbon isotope analysis: Principles
The oxygen isotope composition of water also varies geographically with factors such as temperature, latitude, altitude and distance from the coast (e.g., Gat, 2010;Mook, 2005, pp. 89-98). Britain receives most of its rainfall from a westerly direction and contemporary groundwaters in western Britain therefore record higher d 18 O values than those in eastern Britain (Darling, Bath, & Talbot, 2003, pp. 189-190).
The use of oxygen isotope analysis for geographic provenancing of human individuals is based on the premise that, although there is some contribution from respiratory oxygen and chemically-bound oxygen in food, a significant component of the d 18 O values in mammalian bioapatite derives from ingested fluids which can therefore reflect values in drinking water (e.g., Daux et al., 2008Daux et al., , pp. 1146Levinson, Luz, & Kolodny, 1987;Longinelli, 1984;Luz, Kolodny, & Horowitz, 1984;Luz & Kolodny, 1985;Kirsanow & Tuross, 2011;Podlesak et al., 2008). Evans, Chenery, and Montgomery (2012, pp. 759)  3 ) fraction of tooth enamel of Holocene age is considered to be resistant to diagenesis (e.g., Koch, Tuross, & Fogel, 1997;Zazzo, 2014) and was analysed in the present study. tion of results must, however, give consideration to the potential influence of culinary practice (e.g., stewing foods and brewing : Brettell, Montgomery, & Evans, 2012b;Daux et al., 2008Daux et al., , pp. 1144, or the consumption of fluids that have undergone fractionation through biological processes (e.g., cow's milk, Camin, Perini, Colombari, Bontempo, & Versini, 2008, pp. 1695Kornexl, Werner, Rossmann, & Schmidt, 1997, pp. 22;Lin, Rau, Chen, Chou, & Fu, 2003, pp. 2193), on the d 18 O values of ingested fluids (Lightfoot & O'Connell, 2016). Breast milk can have a higher d 18 O value relative to fluids consumed by the mother and teeth which form while an infant is being breast fed (e.g., deciduous molars) may therefore have higher values than teeth that form later in childhood (Britton, Fuller, T€ utken, Mays, & Richards, 2015, pp. 8;Roberts, Coward, & Ewing, 1988, pp. 625;Wright & Schwarcz, 1998, pp. 14). Cerling, 2014). Current understanding of dietary composition in the European Neolithic is primarily based on analysis of d 13 C and d 15 N values in bone collagen, which predominantly reflect the protein component of the diet and support routine exploitation of C 3 terrestrial sources of protein during the early Neolithic in Britain (e.g., Richards & Hedges, 1999, pp. 893;Richards et al., 2003). In contrast, d 13 C carbonate values in bioapatite reflect the isotope composition of the diet as a whole, including lipids and carbohydrates (Ambrose & Norr, 1993, pp. 2;Jim, Ambrose, & Evershed, 2004). Individuals who obtain the majority of their diet from C 3 terrestrial sources may be predicted to have

| Radiocarbon dating
The program of radiocarbon dating conducted by Bayliss et al. (2011b, pp. 537, 546-547) found that the majority of sampled individuals buried in the transepted passage grave within the rotunda at Ty Isaf dated from the mid to late 4 th millennium BC. However, individuals buried in the adjoining long cairn ( Figure 3) have not been radiocarbon dated.
The excavator was unable to determine which of the two monuments was constructed first: it is possible that both of these monuments were in use at the same time; alternatively, the rotunda could have been inserted into the long cairn at a later date (Grimes, 1939, pp. 137-138).
In addition, radiocarbon dating of bone from the rotunda also revealed the presence of several individuals dated to the earlier third millennium BC (OxA-14248, 2900-2670 cal BC and OxA-14250, 2860-2490 cal BC, Bayliss et al., 2011b, pp. 537;OxCal v4.2, IntCal13), which is argued to indicate that Ty Isaf was the focus of secondary burial activity during the later Neolithic (ibid.). Fragments of Bronze Age pottery found during excavation could also suggest the site was a focus for activity during later periods (Grimes, 1939, pp. 125, 130 and 135-136).
All burials sampled by the present study from Ty Isaf were therefore radiocarbon dated to assist in evaluating whether individuals buried in the long cairn are contemporary with those in the rotunda and whether the site remained of importance to communities beyond the 4 th millennium BC.

| Sample selection
Burial assemblages from Penywyrlod and Ty Isaf long cairns consist of highly fragmentary disarticulated and co-mingled human remains. Care therefore had to be taken to avoid the potential for duplication of isotope results through inadvertent sampling of cross matching fragments of dentition (e.g., mandibular and maxillary) that could belong to the same individual. Only teeth from the left mandibular dentition were selected for sampling. Nine different human individuals from Ty Isaf were sampled (Table 1). Left mandibular dentition from nine different individuals from Penywyrlod was sampled (Table 2). In addition, two pre-existing chips of core enamel taken in 2003-2004 from right mandibular third molars from Penywyrlod by a project unrelated to the present study were also analysed to obtain isotope ratios: 74.23H/ 9.5.19/P27 and 74.23H/9.16/P27 (Table 2).
Information on the contexts of sampled specimens (where available) is given on the labelling and documentation associated with the collections now stored in the National Museum Wales and is provided in Tables 1 and 2. No dentition attributed to the rotunda monument at Ty Isaf could be located that met the criteria for sampling specified above. Four of the specimens from Ty Isaf lacked documentation detailing their excavation context (those listed in Table 1 as being of 'undocumented' context). These specimens could either have been excavated from the rotunda or the lateral chambers of the long cairn at Ty Isaf.
The human burial assemblage from Penywyrlod was recovered during a partial rescue excavation that was instigated following damage to the monument (Savory, 1973(Savory, , 1984. It therefore represents a sample of what may be present at the site. The current study undertook analysis of tooth enamel from individuals buried in two of the three excavated chambers (Chambers NE II and NE III) and in a cist in the forecourt, within the lower revetment wall of the south-east horn of the monument ( Figure 2). Few remains were recovered during excavation of Chamber NE I (Savory, 1984, pp. 18;Britnell & Savory, 1984, pp. 6) and no dentition from this chamber was available that met the criteria for sampling specified above. During assessment of the collections it was found that several specimens from Penywyrlod possessed the same generic museum accession number. To differentiate between specimens, each was assigned an additional unique code, corresponding to the page number on which the specimen is listed in the sampling application made to the National Museum Wales (e.g., P21 refers to page 21 of the sampling application).
The sex of the majority of sampled individuals cannot be determined with confidence owing to the fragmentary nature of the assemblages and disarticulation of cranial remains from other skeletal elements used for sex attribution. Facial reconstruction commissioned by the National Museum Wales for the purpose of gallery display suggests individual 74.23H/9.23/P22 is male. Wherever possible, the approximate age of individuals at death (Tables 1 and 2) was determined from stage of dental eruption and tooth root development, following AlQahtani, Hector, and Liversidge (2010). Individuals who have fully erupted permanent dentition with fully formed third molar tooth roots are denoted as 'adult' in Tables 1 and 2. Where possible, consecutively mineralizing molars were sampled to compare isotope ratios at different stages of childhood: formation of the second molar crown occurs between approximately 2.5 6 0.5 years and 8.5 6 0.5 years of age (AlQahtani et al., 2010;Hillson, 2014, pp. 31, 55-56). Timing of third molar formation is more variable (Liversidge, 2008, pp. 313), with initial cusp formation from approximately 8.5 6 0.5 years of age and crown completion by approximately 14.5 6 0.5 years (AlQahtani et al., 2010). Strontium and oxygen isotope analysis was undertaken on samples of bulk enamel, and isotope ratios therefore represent the weighted average of all dietary sources exploited during the period in which the enamel was forming (Montgomery, 2010, pp. 333). An enamel chip of approximately 20-30mg in weight from each tooth was utilized for strontium isotope analysis and of approximately 10mg in weight for oxygen isotope analysis.
A limited number of fragmentary animal remains were also recovered during the excavation of Ty Isaf (Cowley, 1939, pp. 141-142), including a loose cattle third permanent premolar tooth (Table 1). This was sampled to compare strontium isotope ratios with those of the human burial group. Although formation of the tooth crown proceeds from cusp to cervix (Hillson 2005, pp. 156), complexity in the process of enamel maturation and averaging of strontium in the body pool prior to incorporation into enamel may limit the study of 87 Sr/ 86 Sr values at high chronological resolution in human populations (e.g., Montgomery, Evans, & Horstwood, 2010). In bovine high crowned (hypsodont) teeth it may, however, be possible to detect variation in strontium isotope ratios by taking spatially separated samples along the axis of tooth formation (e.g., Viner, Evans, Albarella, & Parker Pearson, 2010). One sample of enamel, of a similar size to those sampled from each human tooth, was therefore excised from the top (cusp) and one at the bottom (cervix) of the tooth crown. As the earliest forming enamel at the cusp can be eliminated through dental attrition the location from which these samples were excised was recorded in millimetres with respect to the cervical margin (Table 1). The sampled cattle tooth was a loose permanent third molar. In modern cattle the third permanent premolar crown begins to form at approximately 11-12 months of age and formation of the crown of this tooth is complete by approximately 24-30 months (Brown, Christofferson, Massler, & Weiss, 1960).

| Sample preparation and laboratory analysis
Initial preparation of enamel samples for isotope analysis was undertaken in the laboratories at Durham University following procedures developed by Montgomery (2002, pp. 131-138). The enamel surface was mechanically cleaned using tungsten carbide dental burrs and a flexible diamond edged rotary saw was then used to excise a chip of core enamel. To remove any adhering dentine, exposed surfaces of the enamel chip were again thoroughly abraded using dental burrs. Resulting chips of core enamel were transferred to clean sealed containers. Dental saws and burrs were cleaned ultrasonically for five minutes and rinsed three times in high purity de-ionized water between preparation of samples.  and 2). 14 C results are reported as conventional radiocarbon ages, quoted in conventional years BP (before 1950 AD) (Millard, 2014;Stuiver & Polach 1977). Calibrated ages (Tables 1 and 2)

| R E S U L T S
Eight of the nine human individuals from Ty Isaf are assigned by AMS radiocarbon dating to the mid 4 th to early 3 rd millennium BC (Table 1). ppm, 1r, n 5 15). In contrast to Ty Isaf, the majority of the teeth (12 of 15) sampled from Penywyrlod have strontium isotope ratios that are higher than 0.7140, including both individuals from excavated chambers NE II and III (see Figure 2) and the child whose remains were recovered from a cist in the lower part of the outer revetment wall close to the forecourt of the monument (Savory, 1984, pp. 22-23 (Snoeck et al., 2016). However, with the exception of Ferriter's Cove in the far south-west of Ireland, where cattle bone in Mesolithic contexts has been interpreted as a failed episode of colonization during the mid 5 th millennium BC (Sheridan, 2010a;Woodman, Andersen, & Finlay, 1999) and Magheraboy causewayed enclosure, Co.
Although areas in closer geographic proximity within England  Figure 1). Only two individuals in the burial group from Ty Isaf have strontium isotope ratios higher than 0.7140.
Both adult 39.190/310 and sub-adult 39.190/58 were exposed to more radiogenic sources of strontium than are known to be bioavailable locally during formation of their second permanent molar crown, between approximately 2.5 6 0.5 years to 8.5 6 0.5 years of age (AlQahtani et al., 2010;Hillson, 2014, pp. 31, 55-56). These two individuals could have derived their childhood diet from similar source areas to those discussed above for the group sampled from Penywyrlod.
Penywyrlod and Ty Isaf are part of a cluster of early Neolithic long cairns in the Black Mountains. While their construction appears to reflect a common theme, it is argued that differences in the morphology of these monuments may indicate they were built and used by different social groups (Wysocki & Whittle, 2000, pp. 600). The strontium isotope results could support this suggestion. While they are situated less than 5 miles (8 km) apart, these monuments differ in their morphology and the 87 Sr/ 86 Sr results are consistent with the two burial groups having obtained their diet from different geographical locations.
The majority of those buried at Penywyrlod appear to have obtained NEIL ET AL.
| 383 their childhood diet from outside the immediate area in which they were buried, whereas results from Ty Isaf could support the hypothesis that individuals sourced their diet locally. It is possible that this contrast may be the result of a difference in the chronology of the two burial groups as, unlike the individual from Penywyrlod who was radiocarbon dated by this study, all individuals from Ty Isaf post-date 3650 cal BC.
Further radiocarbon dates are being sought from Penywyrlod to examine this possibility.
The rules defining who was selected for burial in such monuments have also often been debated (e.g., Whittle et al., 2007, pp. 134). The results of this study suggest that, rather than acting as a focus for burial of individuals who sourced their diet from a diverse range of geographical areas, the individuals who were chosen for burial within each monument could have obtained their diet from a similar geographical location.
Where adjacent consecutively mineralizing second and third molar teeth were available for sampling from Ty Isaf, the majority of tooth  6 and 7). At this site, several individuals exhibited a change in strontium isotope ratios from a value lower than 0.7085 to higher than 0.7105 between adjacent teeth. In southern Britain, with the possible exception of the Lizard Peninsula in Cornwall, lithologies that routinely give measured biosphere 87 Sr/ 86 Sr values below 0.7085 are geographically separated from those that give values higher than 0.7105 and the shift between these two values can therefore be interpreted to indicate that individuals changed the location from which they obtained their diet during childhood Neil, Evans, Montgomery, & Scarre, 2016a Brettell et al., 2012b), or by consumption of fluids that have undergone fractionation through biological processes (e.g., cow's milk; Camin et al., 2008Camin et al., , pp. 1695Kornexl et al., 1997, pp. 22;Lin, Rau, Chen, Chou, & Fu, 2003, pp. 2193 (Darling et al., 2003, pp. 189, 191;Evans et al., 2010). However, before reaching this conclusion the possibility that there is greater spatial and temporal variability in the oxygen isotope composition of groundwaters of closer regions that can record biosphere 87 Sr/ 86 Sr values higher than 0.7140 (e.g., central Wales) might be considered. It is possible that altitude effects, for example, may induce greater local variability in the isotopic composition of groundwaters within Wales (Darling et al., 2003, pp. 189), resulting in lower d 18 O values in areas of higher elevation (Mook, 2005, pp. 94). In addition, although it has been argued that climatic variation during the Holocene is unlikely to have been sufficiently significant to have influenced d 18 O values exhibited by human populations (Evans et al., 2012, pp. 758), the precise effect of local climatic variability on d 18 O values in groundwaters over time (e.g., at a seasonal, inter-annual or decadal scale) in north-western Europe remains poorly understood and new proxies are currently being sought to study localized variation in past precipitation levels and temperature (e.g., Young et al., 2012Young et al., , 2015.  Darling et al., 2003;IAEA/ WMO, 2016;L ecolle, 1985;Millot, Petelet-Giraud, Guerrot, & N egrel, 2010) and archaeological populations excavated in Lower Normandy (e.g., Brettell et al., 2012a, pp. 127 & 132) have also recorded values comparable to those exhibited by individuals buried at Penywyrlod.
The deciduous molars of two children sampled by this study (from Penywyrlod) have not been included in the above comparison of oxygen isotope ranges, as they begin formation in utero (AlQahtani et al., 2010) and, following birth, values within these teeth may be influenced by consumption of breast milk, which has a higher d 18 O value relative to fluids consumed by the mother (Britton, Fuller, T€ utken, Mays, & Richards, 2015;Roberts et al., 1988;Wright & Schwarcz, 1998). The deciduous second molar of child 74.23H/9.18/P20 from Penywyrlod Multiple factors may be influential in determining strontium concentrations in mammalian tissues (see reviews by Burton & Wright, 1995 kidneys, Kobayashi &Suzuki, 1990) occurs at successive trophic levels within a food chain (Blum et al., 2000;Burton, Price, & Middleton, 1999). As a consequence, herbivores can record higher strontium concentrations than human populations. Strontium concentrations in cattle enamel of prehistoric date excavated in Britain frequently exceed 150 ppm (e.g., Minniti, Valenzuela-Lamas, Evans, & Albarella, 2014, pp. 310;Towers, Montgomery, Evans, Jay, & Parker Pearson, 2010, pp. 511;Viner et al., 2010; also see Neil et al. 2016 (Viner et al., 2010(Viner et al., , pp. 2185. One factor that has been invoked to explain the particularly high strontium concentrations observed within human archaeological populations in an Atlantic facing island such as Britain is the potential for marine-derived strontium to contribute to diet, owing to the high concentration of strontium in seawater (Odum, 1957) and aerial deposition through seaspray (e.g., Whipkey, Capo, Chadwick, & Stewart, 2000).
High strontium concentrations are also recorded in association with other types of environment (e.g., hot arid climates, e.g., Buzon, Simo- range. This suggests that this individual was exposed to non-local sources of strontium and obtained their diet from further afield, with the closest areas to record comparable values being the Malvern Hills and central Wales, as discussed above. These isotope ratios were recorded in enamel of deciduous first and second molars which begin formation in utero. As such, it is possible that they could reflect dietary resources that were exploited by the mother (Montgomery, 2002, pp. 44 Evans et al. (2012, pp. 759) argue to represent occupation of western Britain. However, as formation of deciduous molars continues in the months following birth, it is possible that oxygen isotope ratios in enamel from these teeth could also be influenced by consumption of breast milk, which may confer higher d 18 O values than drinking water (Britton et al., 2015;Roberts et al., 1988;Wright & Schwarcz, 1998). The remains of this individual, dated to 1670-1500 cal BC (95% confidence; SUERC-57789), were found outside Chamber 2 of the long cairn (Table 1). As prominent features in the landscape, Neolithic long cairns were sometimes re-used for burial in later periods (Darvill 2004, pp. 214-232). The AMS radiocarbon results support the suggestion of the excavator that Ty Isaf was reused for burial during the Bronze Age (Grimes, 1939, pp. 135-136).

| C O NC LU S I O N S
The majority of individuals buried at Penywyrlod have strontium isotope ratios that exceed the local biosphere range, suggesting they  (Snoeck et al., 2016). However, the attribution of this individual to this area may also be problematic on archaeological grounds as at present there is limited evidence to suggest that Neolithic material culture and practices were established in Ireland any earlier than the 38 th century BC (Bayliss et al., 2011a, pp. 805-808;Cooney et al., 2011, pp. 663-668;Schulting et al., 2012, pp. 31;Whitehouse et al., 2014).

AUTHOR CONTRIBUTION
S.N. designed the research; S.N., J.E. and G.C. undertook the analysis; J.E. provided strontium isotope results from locally collected modern plants (Table 3)