Ireland’s fallow deer: their historical, archaeological and biomolecular records

The Anglo-Normans first introduced fallow deer ( Dama dama ) to Ireland in the thirteenth century, however no biomolecular research has previously been undertaken to examine the timing, circumstances and impact of the arrival of this species. This study combines historical, zooarchaeological, genetic and iso-topic data from both medieval and post-medieval samples to address this lack of research. The paper identifies a peak in the presence of fallow deer in Ireland between the thirteenth and fourteenth centuries, with a corresponding peak in documentary evidence for their presence in the thirteenth century. The deer are predominantly male, and from castle sites, supporting the historical evidence for their link with elite hunting. The English origin of the source populations shows correspondence between the documentary evidence, suggesting a western bias—and genetic evidence—with a similarity to southern and western England. Furthermore a stable isotope study identifies two possible first-generation imports, one dating from the medieval period and one from the post-medieval period. As is the case with many islands, Ireland’s modern mammalian fauna contains a high proportion of introduced animals, Lever (2009) listing fourteen natura-lised species. These have gradually replaced Ireland’s ‘native’ species, several of which, including grey wolf ( Canis lupus ) and brown bear ( Ursus arctos ), have been brought to extinction (Foster and Chesney 1998, 242). Aside from extinc-tions, Ireland’s list of presumed native species is also diminishing as a result of biomolecular research; recent studies have revealed that some animals that Abstract

The Anglo-Normans first introduced fallow deer (Dama dama) to Ireland in the thirteenth century, however no biomolecular research has previously been undertaken to examine the timing, circumstances and impact of the arrival of this species. This study combines historical, zooarchaeological, genetic and isotopic data from both medieval and post-medieval samples to address this lack of research. The paper identifies a peak in the presence of fallow deer in Ireland between the thirteenth and fourteenth centuries, with a corresponding peak in documentary evidence for their presence in the thirteenth century. The deer are predominantly male, and from castle sites, supporting the historical evidence for their link with elite hunting. The English origin of the source populations shows correspondence between the documentary evidence, suggesting a western bias-and genetic evidence-with a similarity to southern and western England. Furthermore a stable isotope study identifies two possible first-generation imports, one dating from the medieval period and one from the post-medieval period.
As is the case with many islands, Ireland's modern mammalian fauna contains a high proportion of introduced animals, Lever (2009) listing fourteen naturalised species. These have gradually replaced Ireland's 'native' species, several of which, including grey wolf (Canis lupus) and brown bear (Ursus arctos), have been brought to extinction (Foster and Chesney 1998, 242). Aside from extinctions, Ireland's list of presumed native species is also diminishing as a result of biomolecular research; recent studies have revealed that some animals that were traditionally assumed to be autochthonous, for example red deer (Cervus elaphus) and wild boar (Sus scrofa), are, in fact, Neolithic and Mesolithic imports, respectively (Carden 2012;Carden et al. 2012). Whilst the non-native status of these wild animals is somewhat surprising, the introductions of other alien species are well documented, having occurred during the historic era. For instance, documentary and zooarchaeological evidence have established that the fallow deer (Dama dama) was imported to Ireland during the medieval period (McCormick 1991;Beglane 2015a, 23-37;2015b). Prior to their importation to Ireland, fallow deer were restricted to England and Wales and to the Mediterranean (Fig. 1). Given the restricted global distribution of Dama dama at this time, the source population was most likely England, where fallow deer had been maintained as an exotic species within the parks since the eleventh century . It is thought that the Anglo-Norman elite introduced both the fallow deer and the concept of parks to Ireland as a package (McCormick 1991;Beglane 2015a, 23-37;2015b). This elegant species is now a feature of the Irish landscape, with some populations continuing to be kept in parks, for example, the iconic herd of Phoenix Park, Dublin (Pl. 1), whilst others are raised on deer farms or have escaped to live ferally across the island .
Since the history of Irish fallow deer is considered to be 'known', there has been no biomolecular research akin to that for red deer or wild boar to explore the precise timing, circumstances and impact of the arrival of this species. taBle 1-Sites with fallow deer remains. (Beglane 2005;2007;Butler 1989;n.d.-a;n.d.-b;Crabtree and Ryan 2009;Denham and Murray n.d.;Duffy 2010;Holmes 2009;MacManus 1995;McCormick 1997, 837;2004;n.d.;McCormick and Murray 2011;Murphy 1999;Murray n.d.;Whelan 1979). Note that sites with only known misidentifications have not been included. Yet detailed cultural-scientific research concerning the arrival of fallow deer to other islands such as Britain Sykes et al. 2016), Mallorca (Valenzuela et al. 2016), Crete (Harris 2014), Rhodes (Masseti 2002) and Barbuda (Perdikaris et al. 2017) has highlighted the rich, and often unexpected, information that such studies can provide. With this in mind, and to expand existing knowledge of this culturally important species, this paper presents the results of a multidisciplinary investigation of Ireland's fallow deer, combining four independent lines of evidence: historical, zooarchaeological, stable isotope and genetic analysis. Beglane (2010a;2010b;2015a;2015b) has previously reviewed the historical and zooarchaeological evidence for the introduction of fallow deer, which is summarised below. This research formed the basis of our analysis. Site locations and datasets are summarised in Table 1

Zooarchaeological methods
All known zooarchaeological records of fallow deer were synthesised and their percentage representation was calculated for each assemblage (Table 1). Material from six sites was available for further analysis: Newtownstewart Castle,

Materials and methods
Co. Tyrone; Greencastle, Co. Down; Trim Castle, Co. Meath; Cornmarket Street, Dublin City; Cratloemoyle, Co. Clare and Kilkenny Courthouse. These elements were examined for evidence of pathology and taphonomy (e.g. burning, butchery, gnawing) with their ageing data and osteometrics also being recorded (Table 2). In recognition that the osteometrics could be made more meaningful if viewed against a baseline derived from modern animals of known age and sex, measurements from 227 fallow deer (119 male and 108 female) from Phoenix Park, These modern data not only facilitate comparison with raw archaeological measurements but also offer an important opportunity to make the disparate archaeological dataset comparable through the calculation of log ratios. The methods and benefits of log ratios are discussed in detail elsewhere (e.g. Sykes et al. 2013). In this study, average measurements from skeletally mature adult individuals from the Phoenix Park population (male average plus female average  divided by two) were used as the standard to calculate the log ratios for the bone widths of the archaeological fallow deer. Following zooarchaeological recording of the specimens, a subset of samples was used for isotope and genetic analysis. This included samples from Greencastle, Trim Castle, Cratloemoyle, Kilkenny Courthouse, Newtownstewart and Cornmarket St, Dublin (Table 1). The data from the Irish specimens were compared to medieval fallow deer from England, the dataset for which is published in Sykes et al. (2016). That article provides full details of the analytical methods that were employed for this study but they can be summarised as follows.

Genetic analysis
All DNA analyses were carried out at the Department of Biosciences, Durham University. Ancient DNA was extracted and amplified by Polymerase Chain Reaction (PCR). The process of PCR produces a large number of DNA copies from a single DNA template to produce a readable sequence of DNA base pairs (A, C, G and T). In this analysis, a small subsection of the mtDNA control region, totalling 532 base pairs in length, was examined. The 532 base pairs were compared across 17 medieval and post-medieval Irish samples. Alignment of each sample sequence allows relationships to be assessed. These samples were aligned to 78 medieval English samples previously described in Sykes et al. (2016). The relationship among haplotypes (the total number of unique sequences across the samples) was examined by constructing a median joining network. The resulting data provided an assessment of matrilineal diversity within and between sample sites, together with an indication of the demographic history of sampled populations.

Isotope analysis
In order to consider issues of fallow deer diet and, potentially, management, carbon and nitrogen isotope analysis was conducted for the largest available samples of Dama remains: those from Trim Castle and Greencastle, as well as from Cratloemoyle. Whilst δ 13 C and δ 15 N data are generally not used in provenancing studies, fallow deer stable isotope ratios, particularly carbon, do vary significantly across wide geographical ranges  and have the potential to highlight the presence of first-generation imports to Ireland. To explore this issue in more detail, δ 13 C and δ 15 N data for modern fallow deer from Ireland and Britain were analysed to ascertain how environmental variations between the two islands might be expressed isotopically. All analyses were undertaken at the NERC Isotope Geosciences Facilities at the British Geological Survey, Keyworth.
New zooarchaeological information, genetic data and isotope results were obtained for a total of 31 specimens from the sites under consideration.

Results and discussion
These results are shown, by site and specimen, in Table 2. Before these data are examined in detail, it is important to first review the findings of the historical and zooarchaeological synthesis.

History and zooarchaeology
The Greeks and Romans maintained parks filled with a range of animals such as deer, wild pigs, wild goats and wild sheep. These could be extravagant features, such as the c. 34-acre park owned by Quintus Hortensius at Laurentum, south of Rome. This contained a dining room from which diners could view animals that were summoned by blowing a horn (Jennison 2005, 133-6;Toynbee 1973, 16). These Roman parks and the fallow deer within them spread across the Empire, with fallow deer present, for example, in southern England (Sykes et al. 2006). Although they went into decline in the post-Roman period, a number of new parks were created and existing parks did continue to be used. Charlemagne, for example, had one park containing 'multitudes of antlered stags' (Allsen 2006, 40), a deliberate imitation of a Roman status symbol used as a means of legitimising this new imperial power (Andrén 1997, 470). During the Norman period, probably in the late eleventh or early twelfth century, the zooarchaeological evidence suggests that fallow deer were reintroduced to England directly from the Mediterranean, most likely from the Norman island of Sicily (Sykes and Carden 2011) (Fig. 1).
Documentary evidence shows that fallow deer, and the parks in which they were kept, were introduced to Ireland by the Anglo-Normans. The earliest record of fallow deer definitely bound for Ireland is in 1225, when William, Earl Marshal, received a gift of 20 females 'to convey them to Ireland' (CDI, i, no. 1323). However, some years earlier, in 1213, the Archbishop of Dublin, Henry de Loundres, was granted 30 fallow deer from Brewood Park, near Coventry in England (CDI,i,no. 477). Henry had recently been appointed to the See of Dublin and it is tempting to surmise that these animals were imported to Ireland along with his household (Beglane 2015a, 47-8).
In England, where park creation reached its apogee, fallow deer were the main species kept within medieval parks (Rackham 1986, 125). In Ireland, evidence for both parks and fallow deer is much more limited. There is evidence that hunting and the consumption of venison were less important to elites in Ireland, and it is likely that many medieval parks did not contain fallow deer. This species did not filter down the social scale below the first tier of the Anglo-Norman elite, and instead many of the parks that were established became preserves for timber and cattle (Beglane 2015a, 191-2;Beglane 2015b). Nevertheless, where fallow deer were introduced, they were well suited to the park environment and parks functioned as live larders as well as venues for hunting (Fletcher 2011, 97-103;Cummings 1988, 48-9, 53-5). For elite hunters, deer could be hunted par force, which meant that a single animal was tracked using a scenting dog, called a lymer, and then chased using greyhounds. Alternatively, deer could be hunted using 'drive', 'ambush' or 'bow and stable' methods. In all of these cases beaters drove a group of deer towards waiting hunters, who then used spears or bows and arrows to bring down their quarry. This latter method was suitable for use in larger parks, or, if deer were deliberately released, then a longer drive or a par force chase was possible. Professional parkers also used nets and fences to trap deer for live transport or for the provision of venison (Gilbert 1979, 54;Cummings 1988, 32-72;Beglane 2015a, 43-4).
The zooarchaeological representation of fallow deer through time (Fig. 3) shows that, with the exception of Maynooth Castle-where an early fifteenth-century sample has the largest proportion of fallow deer bones anywhere in Ireland-the highest frequency of the species is between the thirteenth and fourteenth centuries. Most are found at castles, but even in such contexts, fallow deer represent only a small proportion of the assemblages. The same overarching pattern is demonstrated by the frequency of documentary references to fallow deer and parks in Ireland (Fig. 4). These also appear in the thirteenth century, reach a peak in the 1250s but then diminish, and references are rare in fourteenth-century documents.
A number of the documentary references relate to royal movements of deer or gifts to Irish magnates and as such these often contain information concerning the source of Irish populations (Beglane 2015a, 47-8 It is noticeable that the majority of the recorded English source populations are located to the west of the country, which is perhaps unsurprising given the destination of the deer: sourcing them as close to Ireland as possible would have reduced transport costs. The origins of Ireland's founding populations can be examined further through genetic analysis, especially since some of the locations

Demographics
Demographics were analysed on samples that gave a positive genetic result (Table 2). Due to the limited sample size, there are insufficient data to gain a detailed understanding of how Irish fallow deer herds were managed. Dental ageing was not possible since no mandibles were recovered; however, examination of the state of epiphyseal fusion of the post-cranial bones shows that the majority of these derived from skeletally mature individuals (Table 2). This age profile is to be expected given that fallow deer were recent arrivals to Ireland at this point and must have been deemed rare exotica, more important to be maintained and displayed alive than consumed. This is also indicated by the log-ratio distribution (Fig. 5), which provides an indirect measure of sex, being skewed towards the positive end of the scale, suggesting an over-representation of larger individuals, which are likely to be males. It has been argued elsewhere Sykes et al. 2016) that a predominance of males within parks suggests an emphasis on the display of large adult bucks, with their antlers rendering them a decorous addition to the landscape. However, elite medieval hunting cultures also tended to focus upon the capture and consumption of prime-aged male animals, as they were perceived to be a more noble quarry than female deer  Fig. 17). (Cummins 1988). An elite preference for hunting male fallow deer would, therefore, explain their over-representation in high-status assemblages.
Genetics DNA analysis determined that one of the purported fallow deer specimens (434 from Trim Castle: McCormick and Murray 2011) was in fact red deer (Cervus elaphus). Male fallow deer are similar in size and form to female red deer and it is not always possible to separate these based solely on visual examination, especially in the case of fragmentary specimens. A number of other samples from Trim Castle (433, 438, 441 and 528) failed, as did specimen 431 from Cratloemoyle, probably due to poor DNA preservation, a common problem when working with ancient DNA. Excluding these specimens, DNA was extracted and successfully amplified for 532 base pairs from 17 samples ( Table 2). The median joining network (Fig. 6) depicts the relationship between the haplotypes detected in the medieval populations from England and all samples from Ireland.
Most of the Irish specimens share haplotypes that are common in medieval England (seq_1 and seq_2), while two other Irish haplotypes (seq_3 and seq_4) are within one mutational step from the most common English haplotype (seq_1). Sequences 1 and 2 include some of the earliest-dated Dama specimens from England, namely those from the eleventh-to twelfth-century assemblages from Carisbrooke Castle (Isle of Wight), Faccombe Netherton (Hampshire), Goltho (Lincolnshire) and Dudley Castle (West Midlands). None of the Irish samples share haplotypes that are found exclusively in northern or eastern England (seq_5, seq_6, seq_10, seq_13, seq_15), which may support the suggestion  from the historical evidence that most deer were sourced from western and southern locations.
Only haplotypes 1-4 are found in Ireland, out of 17 haplotypes identified among all English and Irish regions sampled. We can use a binomial test to assess the probability that this would happen by chance if the same Irish sample size (17) was drawn from the north and south of England independently, given their respective haplotype proportions. Among the 30 northern English samples (from Goltho, Hungate York, Binchester and Belton House), 56% are haplotypes 1-4 (found in Ireland), and 44% are not. That means the probability of sampling a 'non-Irish' haplotype is 0.44. For the south (Sparsholt Villa, Cheddar Palaces, Faccombe Netherton, Carisbrooke, Hyde Abbey, Lewes Castle and Eynsham Abbey) this probability is 0.11. From the binomial test, the probability of drawing only haplotypes 1-4 from the north would therefore be p=0.00005. The same analysis for the south gives p=0.14. So, we can accept the null hypothesis that the Irish sample could have been drawn from the south, but it is unlikely to have been drawn from the north. Dudley Castle is geographically in between these regions, and-based on historical data-a possible origin; however, the haplotype distribution there is most similar to the north (53% 'non-Irish' haplotypes). It should be noted, however, that this test is based on a small sample size from the English distributions, and if one or more small founder events established the Irish population, this could have distorted haplotype frequencies.

Isotope analysis
Issues of translocation can also be considered through stable isotope analysis, especially since Miller et al. (2016) established that English populations of fallow deer through time are significantly different in δ 13 C to those from Ireland (Mann-Whitney post hoc pairwise comparisons: U = 209, P < 0·05 two-tailed). This is likely to be due to broad differences in temperature, rainfall and vegetation regimes across the two islands, resulting in local environmental conditions that influence the diet, and thus dietary isotope values, of the fallow deer.
To further support the idea that it is possible to broadly separate fallow deer bone collagen values according to wider regional locale by analysing their δ 13 C values, we compare isotope values of modern fallow deer from Phoenix Park, Dublin, and those published from Wytham Woods, Oxfordshire (Bonafini et al. 2013) (Fig. 7). These results show that the modern Phoenix Park fallow deer tend to have lower δ 13 C values than those from Wytham Woods. In terms of δ 15 N there is less clear separation, as observed in Miller et al. (2016), but the trend is for higher values in the Irish samples when compared to the English fallow deer values.
It is with these trends in mind that we consider the δ 13 C and δ 15 N bone collagen isotope values of the fallow deer and other cervids from Trim Castle, Greencastle, Cratloemoyle (n = 13, Table 2), and those from medieval English sites   (Fig. 8). The archaeological specimens from Ireland and  Bonafini et al. (2013). These values have not been corrected for the modern fossil fuel effect (Friedli et al. 1986). England (Fig. 8) show more overlap in distribution than is exhibited by the modern populations (Fig. 7). Despite this, the trend shown by the modern samples, with Irish specimens returning lower values in δ 13 C than those from England, is exhibited by the ancient material. The difference between these results, from two adjacent countries separated only by a narrow sea, is significant . This offers the possibility of identifying first generation imports amongst the Irish specimens, with the caveat that, at present, there are still few Dama results from Irish sites and more data are needed regarding the isotope ecology of Irish archaeological fauna as a whole.
Specimen TC433 stands out as anomalous, with an exceptionally high N value, but it is noteworthy that this specimen was pathological (see Table 2), which is known to inflate δ 15 N values (Katzenberg and Lovell 1999). Beyond this, all the Greencastle specimens and the majority of the those from Trim Castle have the lowest δ 13 C values in the distribution, which is consistent with the individuals either having been raised in Ireland, or at least having lived there for several years prior to death. There are two exceptions: specimen TC445 from Trim Castle, with a δ 13 C value of -22.2‰, and the specimen from Cratloemoyle are well within the range expected for English fallow deer. The rate of turnover for δ 13 C and δ 15 N values in fallow deer bone collagen is unknown, but it is estimated at between 10-30 years in humans (Libby et al. 1964, Stenhouse andBaxter 1976). Given this time-lag, and knowing from historical evidence that fallow deer were frequently brought to Ireland from England, the isotope data are consistent with the possibility that these individuals are first generation imports.
Specimen TC445 from Trim Castle is dated to the late thirteenth to early fourteenth centuries, and it is tempting to suggest that it represents the introduction of new bloodstock to the park at Trim when the castle changed hands in 1307. At that time Geoffrey de Geneville retired from his position as lord of Trim, leasing his lands to Roger Mortimer (de Mortuo Mari) (Hartland 2001, 472). Dating to the sixteenth to eighteenth centuries, the Cratloemoyle specimen is much later and is coincident with a later wave of deer introductions that occurred during the post-medieval and early modern periods. This took place as a result of the various plantations of English and Scottish settlers during the sixteenth and seventeenth centuries and the subsequent fashion for landscaped amenity parks in the eighteenth and nineteenth centuries (Beglane 2015a, 184-7, 193;Costello 2015, 164-71).
Biomolecular analyses of all available archaeological fallow deer from Ireland have revealed that Irish populations were most probably drawn from those established in southern and western, rather than northern, England, a pattern consistent with the documentary evidence. Amongst the material analysed, two specimens, one from Trim, the other from Cratloemoyle, appear to be first generation imports as their carbon and nitrogen isotope values are more consistent with English deer. Specimen TC445 from Trim Castle is one of the earlier Conclusion examples, dated to the thirteenth or fourteenth centuries, the period in which both zooarchaeological and historical evidence suggest the greatest levels of deer importation and emparkment (Figs 3 and 4).
When fallow deer were first introduced to Ireland, they would have been valuable and exotic creatures. It would therefore have been important to prevent them from being poached and to ensure that they had sufficient fodder during the winter months. As a result, there is a strong association between the recorded ownership of fallow deer and of medieval parks (Beglane 2015a, 23-37;2015b). While parks and fallow deer became ubiquitous in England, the same is not true in Ireland, where relatively few medieval parks were created and where there are few documentary records or zooarchaeological samples of fallow deer (Beglane 2010a;2015a). It has been argued that the relative lack of evidence for fallow deer and for parks in Ireland, and the fall off in their numbers in the late medieval period reflect economic, political and cultural differences between Ireland and England (Beglane 2015a, 45, fig.16, fig.17, 192).
Records for fallow deer, both historical and zooarchaeological, become sparse after the fourteenth century but there are sporadic occurrences of the species in the post-medieval and early modern period. The various post-medieval plantations of English and Scottish settlers into Ireland, and the subsequent development of landscaped amenity parks to surround great houses, saw renewed interest in the development of parks (Beglane 2015a, 184-7, 193;Costello 2015, 164-71). As a result, fallow deer had a renaissance in post-medieval and early modern times. The sixteenth-to eighteenth-century specimen from Cratloemoyle may represent this later wave of fallow deer introduction.
Although separated by several hundred years, Ireland's waves of fallow deer introduction are united by their symbolism. In both cases the introduction of fallow deer, and the parks in which they were housed, formed part of a 'vocabulary of power' utilised by a colonial elite to legitimise their rule over the local population. It is interesting to note that this same vocabulary has been employed by colonial powers for millennia. Starting with the Romans, fallow deer and deer parks have been used as symbols of dominance by the Byzantines, Carolingians, Normans, Anglo-Normans, and latterly by the British (Perdikaris et al. 2017). This study has demonstrated the rise and subsequent fall in the medieval fallow deer population in Ireland. Through genetics and isotopes, it has highlighted that they failed to become established on the island, necessitating their reintroduction in the post-medieval period. This mirrors the historical evidence for the progress of English control over Ireland, showing how zooarchaeological evidence for a single species can shed light on broad social and political issues.
This work was supported by the British Arts and Humanities Research Council [grant numbers AH/I026456/1]. The authors acknowledge the contribution of the zooarchaeologists and excavators who provided samples and access to unpublished data for this project: Vincent Butler, Pam Crabtree, Sean Denham,