Scottish Animal Welfare Commission: welfare of isolated populations of feral animals - report

2. Background and definition of area of analysis

Concerns over the welfare of feral small ruminants

Complaints, petitions and concerns have been raised about the welfare of both types of feral small ruminant living in Scotland but for different reasons. In our investigations we considered both aspects of concern, as a number of the welfare experiences of these animals are common to both species, but here we summarise the main issues discussed with us separately.

Soay sheep on St Kilda

For a number of years there have been concerns raised by the petitioners (most recently PE2021¹, which was discussed at the Citizens Participation and Petitions Committee, 19^th March 2025) that large losses of Soay sheep on Hirta occurred annually due to starvation and parasitism. The petitioners called for the islands’ owners, the National Trust for Scotland, to intervene to prevent suffering; they also held that, as feral animals, the sheep were subject to the protection of the Animal Health and Welfare (Scotland) Act 2006³. The Scottish Government took a different view of the legal status of the St Kilda sheep. The legal issues are not subject to this report but as part of the consideration of this petition the charity, OneKind, stated that,

‘given there was no option for the sheep population to disperse, there is a moral obligation to address the high levels of winter starvation but cautions that any proposals to reduce levels of winter starvation should be subject to animal welfare impact assessments.’

It is this area that we address in this report.

In our discussions with the petitioners, they reiterated their view that the Soay sheep are domestic sheep and have been subject to artificial selection by humans except for the last 90 years. They cited the annual mortality of up to 600 sheep on Hirta (derived from the Soay Sheep Project’s annual reports⁷) as evidence that these sheep are unsuited to life unmanaged. A central theme of their concerns is that death from starvation is unpleasant and prolonged, that animals were left to die on Hirta even though they could have been euthanised (sharing video evidence of their own observations from April 2026), and that this would not be permitted for domestic sheep under the Animal Health and Welfare (Scotland) Act 2006³. In support of their views we discussed the impact of starvation in sheep with an expert, who provided an overview of the effects on the biological functioning of sheep and the clinical progression of fatal starvation. In his view, ‘because of the duration of suffering experienced during fatal starvation, it is considered one of the most severe types of animal welfare issues encountered in veterinary forensic practice’.

The counter position was expressed by the National Trust for Scotland (NTS), who manage the archipelago, and the Soay Sheep Project (SSP, at the University of Edinburgh) that these animals are wild and there is no clear biological evidence that the Soay sheep are meaningfully different to other wild mammal populations. They also note that, however unpleasant, starvation and exposure to harsh weather in the winter is a common occurrence for wild animals (e.g., rabbits, birds, foxes, etc.) and, in their view, this is similar to what occurs on Hirta for Soay sheep. This reflects the Scottish Government opinion that the Scottish Government considers ‘this unique historical flock [is] protected by the Wild Mammals (Protection) Act 1996, in the same way as any unowned and unmanaged population of wild deer’. We note that in the past (1960s and 1970s) there have been some experimental interventions for some of the sheep on Hirta (castration of some animals, treatment with anthelminthics). However, these were conducted for research purposes and not with the intention of managing the population. These sorts of interventions would be unlikely to be considered acceptable in modern research ethics.

The key area of concern in this species is the welfare of the sheep and whether there is a need for a change in the management of the animals to improve their welfare. The situation for feral goats also raises welfare concerns, but this is markedly different, concerning the impact of management (for environmental and conservation reasons) on the welfare of the animals.

Feral goats on Langholm/Newcastleton

The main argument within Petition PE2151² (discussed at the Citizens Participation and Petitions Committee, 11 February 2026) was to grant protected status to the populations of feral, primitive goats in the Scottish Borders. This was prompted by a decision by an investment company to cull around 80% of the feral goat population at Langholm Moor during the breeding season. Feral goats have no protection under wildlife management legislation in Scotland, and the main approach to their management has been to reduce numbers to prevent damage to forestry and habitats. This petition received support from more than 20,000 respondents, reflecting local concerns about the management of the goats, and it was noted in discussions that this is one of the largest responses ever received by the committee.

In our discussions with the petitioners they outlined concerns that goats were treated as a separate category of animal (invasive, non-native) and were not afforded the protections of the Wildlife and Countryside Act 1981, which provides protection to species which are native to Great Britain and endangered. The goats are described as ‘feral’ and not native to Great Britain and so not protected. The petitioners were seeking the same protections during management as are given to red deer, as there is no closed season for culling of goats. This raised concerns about pregnant and lactating goats being killed and leaving dependent young to starve to death. As part of our discussions, we were told that goats managed south of the border in the Cheviots have been identified as a separate breed of ‘British Primitive Goat’ and have been added to the Rare Breed Survival Trust watchlist⁸. Options for management of these goat populations, recognising that goats can be damaging to other land uses, were described. From the perspective of this report, we do not address the legal status of these animals, but consider the issues from a welfare perspective.

History of domestication and arrival of small ruminants in Scotland

Wild sheep (Asian mouflon), Ovis gmelini, and wild goats, Capra aegagrus, were some of the earliest species to be domesticated, dating from around 11,000 years ago in the Fertile Crescent of the Middle East. Subsequently, small ruminants travelled with Neolithic farmers through the northern Mediterranean to mainland Europe and the British Isles, and through the southern Mediterranean to North Africa⁹. The domestic sheep, Ovis aries, is primarily descended from the Asian mouflon, O. gmelini, through multiple domestication events, although other wild sheep species have contributed to local breeds through hybridisation^10-13. The European mouflon from Corsica, Sardinia and Cyprus has recently been suggested to represent ancient introductions of Asian mouflon to the Mediterranean, but the most recent genetic data have been interpreted to suggest that these are early breeds of domestic sheep, closely related to the Soay sheep¹⁴. The domestic goat, Capra hircus, is primarily descended from the wild goat, C. aegagrus, involving also multiple domestication events, but hybridisation with other wild goat species has contributed too^9,15,16. There is useful information on the behaviour and adaptations of these wild and early domestic species, which will be used in this report where it aids understanding of the situation for feral animals.

The first stage of domestication is represented by the formation of loose ties, for example by the sharing of watering places, but there is evidence of confinement and breeding control by Neolithic farmers, and the presence of ‘breeds’ by the Bronze Age¹⁷. Domestic sheep arrived in the Iberian peninsula around 5000-7000 years ago and small ruminants were farmed in Britain around 6000 years ago. Archaeological records suggest that these sheep were small, brown animals with horns that resemble the current Soay sheep. In addition, these sheep had primitive cycles of growth and shedding of their fleeces, which pre-date human selection for fleeces that can be harvested for wool¹⁸.

The exact time of arrival of sheep onto the remote islands of St Kilda is unclear, but there are arguments that Soay sheep represent a Bronze Age breed that would have been transported in open boats from the Western Isles of Scotland to the archipelago around 2000 BCE¹⁹. Archaeological assemblages of pottery and food remains indicate habitation by Iron Age and Bronze Age farmers, which is the likely time of arrival of sheep on St Kilda, and when regular human inhabitation may have begun. Sheep were kept on four islands of the Archipelago (Hirta, Boreray, Soay and Dùn), although there is only evidence of human habitation on Hirta and Boreray²⁰. When newer, improved (in terms of providing human-selected traits), breeds appeared on the island of Hirta in the Iron Age, there is evidence that the Soay sheep may have been restricted to Soay at that point. This is supported by historical records that document the presence of feral sheep on Soay in the late Middle Ages and that the island is named ‘sheep’ (Soay in the Norse language) well before the 1400s. The Iron Age sheep, representing the early developments of breeds such as Hebridean, Manx Loghtan and Blackface, were preferred over Soay sheep for their growth, meat and wool characteristics. However, the genetic and behavioural characteristics of Soay sheep, especially the aggression of Soay rams which out-competed the newer breeds, meant that the St Kildans did not tolerate the presence of Soay sheep on Hirta. Therefore, Soay sheep were removed to the island that became known as Soay, and more modern breeds (resembling current Hebridean sheep) replaced them on Hirta and Boreray. In 1697 a population of 400 sheep (believed to be direct descendants of the Scottish Dunface sheep) were present on Boreray²¹. In 1870, Blackface and Cheviot rams were introduced to Hirta and may also have been used on Boreray. Populations of sheep deriving from these different breeds remain on Boreray to the present day (known as Boreray Blackface).

The St Kildans appear to have utilised the sheep on Soay and Boreray for meat, skins and fleeces (collected by plucking the wool from the animals), harvesting animals to supplement their own farmed sheep on Hirta, but no real evidence of ‘farming’ of these animals appears to exist. Seton²², for example, writing in 1877 and drawing on other sources, suggests there was a stable population of around 1500 animals across the four islands and reports that ‘the native breed, which run wild, are only caught at plucking time, chiefly on the smaller islands’ adding that ‘the sheep received very little attention’. In the early part of the twentieth century, a number of Soay ewes and rams were imported to the mainland from Soay and were established on estates and parks as a novelty¹⁹. These animals, supplemented by 24 sheep that were taken from Hirta in 1963, form the origins of the Soay sheep found on the mainland. In 1930, following health challenges, the impact of tourism and the First World War, the remaining St Kildans petitioned to be removed from the islands, taking their cattle and sheep with them. In 1932, the then owner of the islands, the 5^th Marquis of Bute, moved 107 Soay ewes and rams to Hirta, possibly as a means of managing the land or as part of plans to make the island a tourist attraction. Within 20 years, when a census of the sheep was conducted, the population had risen to around 1000 animals. Except for the sheep removed in 1963, the Soay sheep on Hirta have not been removed or added to for nearly 100 years; those on Soay and Boreray have been almost entirely unmanaged for far longer. There is some limited evidence that some of the male lambs on Soay may have been castrated – based on diary notes from Alice MacLachlan in 1906²³, and a suggestion in Clutton Brock et al. (1996)²⁴ that some of the sheep moved from Soay to Hirta may have been castrated. However, the inaccessibility of Soay and Boreray, the lack of any signs of human habitation on Soay, and the writings from visitors to St Kilda, do not support the idea of active management of these sheep. Genetic data taken from the sheep on Hirta suggest that some admixture of genes from modern breeds (expressed as polymorphisms for coat colour and pattern) is consistent with some addition of modern breed genetics 150 years ago²⁵. Feulner et al (2013)²⁵ explain these findings as evidence of genetic selection for the light coat colour by natural selection. Additionally, Gannon and Geddes (2016)²⁰ report that the sheep on Soay were reserved to the owner and islanders were forbidden to introduce other rams to the island, so that these additional genes may have come from planned introductions by the island owner. It seems most likely that the sheep on these islands were viewed as part of the harvestable resources provided by the archipelago, along with fowling expeditions for seabirds, to supplement the sheep, cattle and crops grown on Hirta.

Goats are not a native species in Britain, but they have been kept continuously since their introduction around 6000 years ago by Neolithic settlers, probably alongside the arrival of sheep into the British Isles²⁶. However, the archaeological record for the presence of goats in Scotland is patchy for the Neolithic period, although they are considered to be present from the Iron Age onwards²⁷, and in substantial numbers from the 12^th to 15^th centuries. Goat farming was common in Scotland for many hundreds of years with goat herds observed across Scotland alongside farming of sheep and cattle. It is apparent that the more lucrative farming of cattle and sheep led to far greater numbers of these species being kept by the late 1700s²⁷, and a decline in the farming of goats.

Although primitive goat breeds have been studied to understand their adaptations to the environment and their genetic diversity, little is known about the history of feral goat populations, unlike the Soay sheep records. However, genetic analyses of primitive goats suggest shared ancestry with Irish breeds and support a relationship with arrival of these animals before the Bronze Age with Neolithic farmers, at least into England and the Northumberland border with Scotland²⁶. The period of time when these animals may have become naturalised is also unclear and may differ markedly depending on the population. For example, the populations of primitive goats on Rum were left behind when the human population emigrated to Canada as part of the forced evictions in the clearances between 1750 and 1860, and a similar history may also apply to other populations. There is anecdotal evidence that the naturalised Cheviot goat populations in Northumberland were abandoned much earlier, by the monks of Lindisfarne when they fled inland to avoid a Viking invasion in 793 CE, although there is no historical evidence for this claim²⁶. For some island populations globally, there is evidence of animals deliberately being relocated by European sailors to serve as a food source on long journeys²⁸, which may also reflect Viking practices. Whitehead²⁹ produced a gazetteer of wild goat populations in the British Isles, listing 141 historic and then contemporary populations in Scotland. He provided further information on 127 of these, of which 53 were extinct or functionally extinct (single-sex populations or single individuals) and 74 were extant at that time. There has been no more recent survey in Scotland. In September 2025 the Mammal Society carried out a population survey of feral goats in the Langholm and Newcastleton Hills, Dumfries and Galloway, but no results have been published so far.

Definition of feral status

The term ‘feral’ is widely considered to describe animals that were previously domesticated and under the management of humans and are now free-living animals with no human control. Categorisation of animals as feral is related both to their genetic adaptation to living with humans (domesticated) and to their current state of dependence on humans for care (managed or not)⁶. Rudge (1984)³⁰ suggested that feral small ruminants can be divided into three categories: 1) those that derive from ancient populations (e.g., Soay sheep, Corsican and Sardinian mouflon and agrimi), and notes that until recently some of these populations have been considered to be truly wild (e.g., Corsican and Sardinian mouflon). These animals are considered to be ‘functionally wild’ or ‘naturalised’, even if the animals are in their current environments through historical human actions⁶; 2) Recent feral small ruminants, which were abandoned in 17^th and 18^th centuries by sailors or as a result of the land-use changes that occurred in the Clearances, and may also be naturalised to their current environment (which seems to be the most likely rationale for the presence of primitive goats on islands and elsewhere); and 3) Strays from transhumance or mustering, which tend to be only temporarily living without human management and are largely contiguous with farmed sheep or goats. To this list we add a more recent type of feral animal, which derive from re-wilding and conservation projects, such as Exmoor ponies used in coastal grazing management and projects such as Oostvaarderplassen. However, in these latter cases domestic animals are deliberately being used and changed within a specific conservation context, which is different to our focus.

Gambourg et al⁶ further consider that feral animals, although of domestic origin, display wild-type behaviours towards humans (typically fearfulness and avoidance) and thus cannot be considered to be tame, and also demonstrate other wild-type responses, such as free-roaming. They also describe a process of ‘feralization’, which can be regarded as the opposite of domestication, where genetic and morphological changes can occur over time as the feral animal shows adaptation to the current environment and the influence of natural selection. Of relevance here is also the seminal work of David Wood-Gush and Alex Stolba in the 1980s that demonstrated that wild-type behaviours are preserved in domestic animals, even if those animals have been prevented from expressing those behaviours (such as nest-building in domestic sows) for generations³¹. This demonstrated that domestic pigs were capable of living successful lives without continuous human interventions within a single generation, providing that the environment did not contain predators or competition for resources. Similar responses are also seen in ruminants. Therefore, domestication has not engendered animals unable to cope and survive without human intervention, at least for those animals that are the subject of this report.

Adaptation of feral small ruminants

Sheep and goats are remarkably robust and adaptable as species, occurring in some of the harshest and most severe environments on earth. They are found in hill and mountainous regions across the world, existing in the Himalayas and Rocky Mountains, where the temperature extremes between summer and winter can exceed 40°C. Wild sheep are found both in hot and arid desert environments, and live inside the Arctic Circle in Canada, Alaska and Siberia. They are able to cope with thermal extremes, poor quality forage and limited availability of water. In addition, wild, feral and domestic small ruminants can exploit a wide range of food sources, including fruit, leaves, lichens, heathers and even seaweed (especially the North Ronaldsay sheep where their diet is almost exclusively seaweed)³².

Sheep and goats show seasonality in their reproductive cycles, voluntary food intake, fat metabolism, and pelage and horn growth. Seasonality is an adaptation that allows animals to cope with fluctuating food supplies and climatic variability. In temperate regions, reproductive effort is timed to coincide with the growing season of plants. Ewes of temperate breeds begin to show oestrous cycles in late summer and autumn, ensuring that lambs are born in spring as new grass begins to grow. Various studies have shown that the breeding season is very short for some northern latitude breeds (e.g., Scottish Blackface), but longer for Galway, Suffolk and Finnish Landrace breeds and their crosses³³. Rams also show seasonality in testis size and libido, which varies with breed^34,35. Seasonal changes are most pronounced in some rams (mouflon) and occur later than in feral and domestic breeds (Soay, Shetland, Herdwick, Scottish Blackface, Norfolk and Wiltshire Horn). Defined seasonal cycles of horn growth also occur in spring, which coincides with changes in plasma prolactin and a resurgence of pelage growth, in mouflon, Soay, Wiltshire Horn, Herdwick and Shetland sheep³⁴, but is not seen in breeds selected for greater wool growth. In addition to the seasonal changes in reproduction, temperate sheep show seasonal declines in appetite, voluntary food intake, metabolic rate and bodyweight gain in winter. Soay, Scottish Blackface and Shetland sheep have a greater seasonal variation in voluntary food intake than Dorset Horns, suggesting an adaptation, which can be affected by degree of human selection for genetic traits, such as a longer oestrus cycle, or impacts of living at higher latitudes.

Breed differences and adaptation are also seen in many behavioural traits in small ruminants, including fearfulness, flight distances, flocking and tolerance of social contact, home-range and ‘hefting’, and adaptations to support lamb survival^32,36. Research studies of fearfulness, together with responses to predators, suggest that the less-selected hill and upland breeds and more primitive and feral breeds have a greater reactivity to the same stressor and take longer to recover than do more selected and specialised lowland breeds. Overall, analysis of different sheep breeds, including feral and primitive breeds, suggests a continuum of adaptation with less managed and selected sheep (typically wild species, primitive, feral and hill breeds of small ruminants) displaying greater environmental adaptation and survival-related traits than those of more selected and specialised sheep and goat breeds.

Focus of this report

The goal of this report is to review the welfare of feral ruminant species in Scotland and consider the ethical responsibilities we may have to these species. It is apparent from the historical review that these animals have found themselves in their current environments due to their past association with humans, even if this was hundreds or even thousands of years ago. Small ruminants did not evolve in the British Isles, but were brought here by Neolithic farmers. However, some populations have been living successfully without human control or management for many hundreds of years. They could be considered ‘functionally wild’, or ‘naturalised’ even if not considered as a truly wild animal⁶. Without human intervention these animals will have adapted and changed genetically in response to the environment, in some cases for more than 1000 years. In situations without predation, animals that can adapt are able to exploit these new locations and the survival of these animals for many generations, as well as the increase in population numbers, without human control or support, suggests they have become well adapted to their locations. Nevertheless, it is noteworthy that welfare is not the same as adaptation, and the costs and benefits of adapting to the environment bear further inspection.

In conducting this review, we have used the Five Domains model of animal welfare³⁷, which has underpinned our welfare considerations in other SAWC reports. This considers the welfare of animals from the perspective of four physical domains (three survival critical: health, physical environment and nutrition; and one situational: behavioural interactions, including interactions with the environment, other members of the same species and with humans), which integrate to impact on the fifth domain, mental state, which is the animal’s welfare state. This framework situates welfare with the experiences, feelings or affective state of the animal (as does SAWC’s definition of welfare) and recognises that welfare can be both positive (e.g., animals experiencing pleasure, comfort, anticipation, etc.) and negative (e.g., animals experiencing pain, fear and discomfort). We firstly consider the welfare state of feral animals based on the current situation and then consider the probable welfare impact of a range of potential human interventions. Finally, we provide some recommendations for future approaches to the welfare management of feral animals.