Kilda Archipelago Research Articles

The Last of the Great Auks: Oral History and Ritual Killings at St Kilda

The story of the killing of the ‘last’ great auk (Pinguinus impennis) in Britain, apparently put to death as a witch at Stac an Armin in the St Kilda archipelago c. 1840, is well known. However, other accounts claim that an auk was killed on the main island, Hirta, having been condemned to death by the celebrated men’s ‘parliament’. The historical veracity of three differing stories, which recount discreditable deeds in a deeply Christian community, is evaluated; it seems that fewest difficulties are raised if two great auks were killed, one on Hirta and the other on Stac an Armin. It is argued that this kind of avicide was a ‘ritual’ killing, to be understood in its historical context. The auk-killing probably took place in the mid to late 1840s, after the St Kilda minister had departed in the wake of the Disruption of 1843 - a particularly unsettling time within this small island community. A possible sighting of a pair of great auks on Soay (St Kilda) in 1890 is also briefly discussed.

Population surveys of burrow-nesting seabirds on the St Kilda archipelago: results and insights from the 2019 Seabirds Count census

Population surveys of burrow-nesting seabirds on the St Kilda archipelago: results and insights from the 2019 Seabirds Count census

Soay Sheep: The Back-story

The primitive race of Soay sheep from the St Kilda archipelago in northwest Scotland has played an important role in narratives of the history of domestic sheep. The Soays, apparently a ‘Bronze Age’ race of sheep, were probably confined to the precipitous isle of Soay as soon as ‘Iron Age’ sheep were introduced to Hirta, St Kilda's main island, owing to the competitive edge of the ferocious Soay rams over the new arrivals. In the 1880s, Pitt-Rivers, following his archaeozoological interests, was the first to keep Soays in his park, their epic journey from the edge of the Atlantic to southern England enabled by his acquaintance with their owner. In the early twentieth century, Soays featured in animal bone reports for archaeological sites, were kept in parks and involved in breeding experiments, particularly around Horsham, Sussex (where their owner lived), and in Edinburgh. The transfer of 107 Soays to Hirta in 1934 and 1935, after humans had evacuated St Kilda, was a remarkable feat, enabling the important long-running Soay Sheep Project. The historical exploitation of ‘feral’ Soay sheep by the islanders of St Kilda has significant cultural ramifications.

Decline of Leach’s Storm Petrels Hydrobates leucorhous at the largest colonies in the northeast Atlantic

Leach’s Storm Petrel Hydrobates leucorhous has undergone substantial population declines at North Atlantic colonies over recent decades, but censusing the species is challenging because it nests in burrows and is only active at colonies at night. Acoustic playback surveys allow birds present in nest sites to be detected when they respond to recordings of vocalisations. However, not all birds respond to playback on every occasion, response rate is likely to decline with increasing distance between the bird and the playback location, and the observer may not detect all responses. As a result, various analysis methods have been developed to measure and correct for these imperfect response and detection probabilities. We applied two classes of methods (calibration plot and hierarchical distance sampling) to acoustic survey data from the two largest colonies of breeding Leach’s Storm Petrels in the northeast Atlantic: the St Kilda archipelago off the coast of northwest Scotland, and the island of Elliðaey in the Vestmannaeyjar archipelago off the southwest of Iceland. Our results indicate an overall decline of 68% for the St Kilda archipelago between 2000 and 2019, with a current best estimate of ~8,900 (95% CI: 7,800–10,100) pairs. The population on Elliðaey appears to have declined by 40 –49% between 1991 and 2018, with a current best estimate of ~5,400 (95% CI: 4,300–6,700) pairs. We also discuss the relative efficiency and precision of the two survey methods.

Unpacking multi-trophic herbivore-grass-endophyte interactions: feedbacks across different scales in vegetation responses to Soay sheep herbivory

Grazing can induce changes in both plant productivity and nutritional quality, which may subsequently influence herbivore carrying capacity. While research on Soay sheep (Ovis aries L.) dynamics on Hirta Island in the St. Kilda archipelago has elucidated the complexity of population drivers, including parasites, the role of herbivore-generated feedbacks as an intrinsic regulating factor remains unclear. The sheep lack large predators and every 3–9 years undergo population crashes (overcompensatory mortality). We investigated the effects of grazing on (1) sward productivity and (2) quality (toxicity) of the primary forage species, red fescue (Festuca rubra L.), which is highly infected by an alkaloid-synthesizing fungal endophyte. Grazing had a negative impact on both forage quantity and quality. At higher sheep densities, impacts on sward growth were magnified, resulting in a nonlinear relationship with plant productivity. Simultaneously, endophyte hyphal load (and by inference, toxicity) peaked close to the time of a crash. A greenhouse experiment showed that alkaloid concentration in F. rubra increased in response to artificial defoliation. We conclude that at high sheep densities, grazing-mediated reductions in productivity, together with sustained alkaloid production, are likely to influence sheep dynamics. Future research should consider the interactive effects of forage toxicity, quantity, and nutritional content.

Influences of Domestication and Island Evolution on Dental Growth in Sheep

Domestication and island evolution can lead to changes of life history along the slow-fast gradient. Shifts of life history patterns, in turn, are potentially related to alterations of patterns and timing of tooth eruption. Schultz’s rule predicts an earlier eruption of molars relative to premolars as fecundity increases during the domestication process. On the other hand, evolution on a predator-free, resource limited island might lead to a generally slow life history and delayed tooth eruption, as in the Plio-Pleistocene Balearic caprine Myotragus. In this study, we investigate tooth eruption and its relation to life history in a unique sheep population that is an example of both domestication and island evolution: the ancient and feral Soay sheep (Ovis aries) of the St. Kilda archipelago, Scotland. Tooth eruption timing and sequence is investigated in a comparative framework featuring new data on other domestic sheep (O. aries), including European mouflon (O. a. musimon), as well as wild sheep (O. vignei, O. cycloceros, O. arkal, O. orientalis, O. ammon). These data indicate that the order of eruption is similar in wild and domestic sheep, despite the fundamental life history changes that came about with domestication. However, in contrast to other domestic sheep breeds, Soay sheep erupt their teeth at an absolute older age and also tend to grow more slowly, which resembles the evolutionary trend in island-adapted Myotragus. Despite these similarities, Soay sheep do not share the slow life history pattern inferred for Myotragus, highlighting the distinctive nature of tooth eruption in Soay sheep.

Enabling Capabilities: Innovation and Development in the Outer Hebrides

The existing challenges to non-profit heritage organisations are increasing in these volatile and austere climates. Particularly, existing organisational and strategic mission developments are needed to enable more resilient organisations. This paper explores the relationship between Ionad Hiort and the Glasgow School of Art’s Institute of Design Innovation (INDI) as a six-month case study for understanding how design innovation can build creative capabilities within this unique partnership. In this case, the “building” of a new type of heritage centre on the west coast of the Isle of Lewis in Scotland. It focuses on the concept of innovation in approaching complex situations as an alternative to existing means of problem solving. Through a series of workshops and engaging informal sessions a new strategy was built, which focused the group’s potential. The outcome being a new way to build confidence within communities and build new relationships, enabling new creative capabilities within a rural island community.

Late Quaternary glaciation in the Hebrides sector of the continental shelf: cosmogenic nuclide dating of glacial events on the St Kilda archipelago

The St Kilda archipelago lies ~65 km west of the Outer Hebrides and ~60 km east of the Atlantic shelf break, and represents a key site for testing the assertion that during the Last Local Glacial Maximum (LLGM; c. 27 ka) the British–Irish Ice Sheet (BIIS) extended to near the shelf edge in all sectors. Two consistent cosmogenic 36Cl exposure ages averaging (≥) 81.6±7.8 ka for perched boulders at 290 m altitude demonstrate that the last ice sheet failed to over‐run high ground on the largest island, Hirta. 36Cl and 10Be exposure ages for glacially emplaced boulders on low ground indicate deposition by small, locally nourished glaciers that last occupied a north‐facing valley (Gleann Mòr) at c. 30.9±3.2 ka, prior to extension of the last ice sheet to the outer shelf, and a south‐facing valley (Village Bay) at c. 19.2±2.3 ka, several millennia after the LLGM. Our dating evidence is consistent with previous interpretations of lithostratigraphical, seismostratigraphical and geomorphological evidence and confirms that the last ice sheet failed to encroach on St Kilda. A simple ice‐flow model demonstrates that even if thin, low‐gradient ice lobes encircled the archipelago during the LLGM, the ice margin can only have reached the outermost moraine banks, ~40 km west of St Kilda, under extremely low (<2 kPa) driving stresses, implying either surge‐like transient streaming behaviour at the ice‐sheet margin or that the moraine banks relate to an earlier, more extensive ice sheet. The final glaciation of the Village Bay area at c. 19.2±2.3 ka was out of phase with the behaviour of the BIIS, which was undergoing net retreat during this period.

Exposure to viral and bacterial pathogens among Soay sheep (Ovis aries) of the St Kilda archipelago.

We assessed evidence of exposure to viruses and bacteria in an unmanaged and long-isolated population of Soay sheep (Ovis aries) inhabiting Hirta, in the St Kilda archipelago, 65 km west of Benbecula in the Outer Hebrides of Scotland. The sheep harbour many metazoan and protozoan parasites but their exposure to viral and bacterial pathogens is unknown. We tested for herpes viral DNA in leucocytes and found that 21 of 42 tested sheep were infected with ovine herpesvirus 2 (OHV-2). We also tested 750 plasma samples collected between 1997 and 2010 for evidence of exposure to seven other viral and bacterial agents common in domestic Scottish sheep. We found evidence of exposure to Leptospira spp., with overall seroprevalence of 6·5%. However, serological evidence indicated that the population had not been exposed to border disease, parainfluenza, maedi-visna, or orf viruses, nor to Chlamydia abortus. Some sheep tested positive for antibodies against Mycobacterium avium subsp. paratuberculosis (MAP) but, in the absence of retrospective faecal samples, the presence of this infection could not be confirmed. The roles of importation, the pathogen-host interaction, nematode co-infection and local transmission warrant future investigation, to elucidate the transmission ecology and fitness effects of the few viral and bacterial pathogens on Hirta.

Bad coat, ripped genes: cryptic selection on coat colour varies with ontogeny in Soay sheep

A fundamental goal for evolutionary biologists is to connect the processes of natural selection and genetic drift with the maintenance of genetic variation in the wild. There are now scores of examples in which mapping phenotypes to genotypes has identified the molecular basis of traits in natural populations, but documenting the fitness consequences of these genes has proven more elusive. A long-term study of several thousand unmanaged sheep on the remote Scottish island of Hirta in the St. Kilda archipelago provides a unique opportunity to make these connections between genotype, phenotype and fitness. Since 1985, there has been a steady decline in the frequency of sheep with a rare uniform ('self'-) pigmentation pattern as opposed to the more common wild-type pattern of dark upper body and pale belly. In this issue of Molecular Ecology, Gratten et al. (2012) link variation at the agouti signalling protein (ASIP) gene with 25 years of pedigree information and individual fitness data to investigate the contributions of natural selection and genetic drift to this decline. Consistent with the decrease in frequency of self-type sheep, Gratten et al. demonstrate that recessive genetic variants at ASIP are associated with reduced lifetime fitness. But surprisingly, these variants show the opposite trend to the self-type phenotype by increasing throughout the study period. This paradoxical result occurs because heterozygotes that harbour the majority of self-type alleles have increased in frequency. This pattern would not have been detectable if only phenotype had been monitored because the pigmentation pattern of heterozygotes is indistinguishable from that of wild-type homozygotes. The study provides an excellent example of the importance of measuring selection at both the phenotypic and genetic level, and demonstrates how long-term pedigrees can be used to link alleles to phenotypes and ultimately to fitness.

Heritability not missing—genetic basis of sexual weaponry uncovered

The chase to uncover the genetic underpinnings of quantitative traits of ecological and evolutionary importance has been on for a good while. However, the potential power of genome‐wide association studies (GWAS) as an approach to identify genes of interest in wild animal populations has remained untapped. Setting technical and economic explanations aside, the sobering lack of success in human GWAS might have fed this restraint. Namely, while GWAS have successfully identified genetic variants associated with hundreds of complex traits (e.g. Ku et al. 2010), these variants have generally captured only a low percentage of variance in traits known to be highly heritable—an observation came to be known as the ‘missing heritability’ (Maher 2008; Aulchenko et al. 2009). Hence, if the vastly resourced human studies have been unsuccessful (but see: Yang et al. 2010), why should we expect that less resourced studies of wild animal populations would be able do better? In this issue of Molecular Ecology, Johnston et al. (2011) prove this line of thinking wrong. In an impressive and what may well be the most advanced gene mapping study ever performed in a wild population, they identify a single locus (RXFP2) responsible for explaining horn phenotype in feral domestic sheep from St Kilda (Fig. 1). This same locus is also shown to account for up to 76% of additive genetic variance in horn size in male sheep: this contrasts sharply with most human GWAS where mapped loci explain only a modest proportion of genetic variation in a given trait. The Soay sheep of the St Kilda archipelago are a primitive feral breed of domestic sheep. Pictured are a male with vestigial horns (=‘scurred’; left) and two normal‐horned males (centre and right). Photograph courtesy of Peter Korsten.image

Survival of the ttest: a comparison of medicine and health on Lord Howe Island and St Kilda

Lord Howe Island and the St Kilda archipelago have many similarities, yet their communities had totally disparate outcomes. The characteristics of the two islands are compared and contrasted, and it is hypothesised that the differences in health and diseases largely explain the success of one society and the failure of the other.

Estimating Population Size and Hidden Demographic Parameters with State‐Space Modeling

Recent research has shown how process variability and measurement error in ecological time series can be separated using state-space modeling techniques to combine individual-based data with population counts. We extend the current maximum likelihood approaches to allow the incorporation of sex- and age-dependent counts and provide an application to data from a population of Soay sheep living on the St. Kilda archipelago. We then empirically evaluate the performance and potential of the method by sequentially omitting portions of the data available. We show that the use of multivariate time series extends the power of the state-space modeling approach. The variance of measurement error was found to be smaller for males and the sex ratio of lambs to be skewed toward females and constant over time. Our results indicated that demographic parameters estimated using state-space modeling without relevant individual-based data were in close agreement with those obtained from mark-recapture-recovery analyses alone. Similarly, estimates of population size obtained when the corresponding count observations were unavailable were close to those from the entire data set. We conclude that the approach illustrated here has great potential for estimating hidden demographic parameters, planning studies on population monitoring, and estimating both historical and future population size.

Land use history of Village Bay, Hirta, St Kilda World Heritage Site: A palynological investigation of plaggen soils

This paper presents findings based on a palynological investigation of artificially accreting (plaggen) soils from the settlement of Village Bay, Hirta, in the St Kilda archipelago, which was perhaps the most distant and inhospitable outpost of sustained human habitation in the British Isles. The soils were developed principally through the addition of turf ash and seabird waste, although some ash may have been derived from upland peats. It is assumed that the woodland pollen signal (much lower in the soils than in an upland peat site nearby) represents off-island sources. Corylus avellana-type pollen (frequent in upland sites), along with Potentilla-type, may provide markers in the Village Bay profiles for the addition of ashed hillside turf, and possibly peat, to the plaggen soils. Cereal-type pollen is well represented through the profiles and is often strongly associated with the record for Chrysanthemum segetum (corn marigold), a frequent indicator of arable land. The Brassicaceae signal may partly reflect the cultivation of cabbages; Chelidonium majus (greater celandine) may have been grown for medicinal use. Soil mixing has rendered radiocarbon dating meaningless at this site, but the establishment of a change in cultivation regime before AD 1830 may have been identified from the patterns of pollen concentration and preservation in the profiles.

A New Twist on an Animal Model Reveals Environmental Constraints on Selection

No one knows for sure when or how they got there, but Soay sheep (Ovis aries) arrived on the isolated, wind-swept Scottish island of Soay sometime during the Bronze Age, thousands of years ago. In the 1930s, a flock was relocated to Hirta, an adjacent island in the St. Kilda archipelago, after villagers evacuated to the mainland. Since then, the sheep have been left mostly to their own devices, save for ongoing encounters with scientists keen to test a range of evolutionary hypotheses on this unique population. Mark-and-recapture studies have revealed important clues to the ecology of the population, which periodically crashes, typically when a severe winter follows a robust reproductive season. These studies have also generated a wealth of information on birth and death rates, reproduction, pedigree (maternal and paternal lineages), and phenotype (such as birth weight). Alastair Wilson, Loeske Kruuk, and colleagues took advantage of a 20-year study of birth weight in Soay sheep—which covered nine generations and significant climatic fluctuations—to examine the impact of environmental quality on the strength of selection and the amount of genetic variation associated with a fitness-related trait. Evolution should occur when selection acts on a trait, such as birth weight, that has a genetic basis. Birth weight is closely linked to juvenile mortality, with larger lambs having higher survival and higher fitness. The first six months of life are critical; lamb deaths during this crucial window account for nearly a quarter of all deaths. Even though birth weight has all the raw materials for evolution—a positive correlation between birth weight and survival, and a heritable basis for variation, attributed to maternal genetic effects—previous studies found no evidence for evolution of the trait. Since annual birth weight had not increased over the 20 years, the researchers hypothesized that a variable environment could limit its micro-evolution by limiting heritability, selection, or both. Wilson et al. found evidence that that this is indeed the case. To study the micro-evolutionary dynamics of a wild population, they adapted a theoretical model long used by animal breeders. Quantitative genetic models measure the genetic contribution (which typically involves multiple genes) to variation in a given trait, such as milk production in dairy cows. With these models, plant and animal breeders maintain or develop desired traits by predicting an evolutionary path based on the strength of selection and the amount of genetic variance underlying the trait. Some of these models assume a constant environment, which is fine for breeders, who can artificially control their subjects' environment, but proves problematic for evolutionary studies of wild populations. To overcome these limitations, the researchers incorporated a statistical tool called random regression analysis. With this approach, they could not only model the effects of environmental variation on selection and heritability but also maximize the statistical power of their dataset. This is a particularly effective technique for natural populations, for which the data are inevitably limited. It also allowed the researchers to model the genetic variance underlying lamb birth weight as a function of environmental quality (poor if lamb survival was low, good if survival was high), creating a more realistic evolutionary picture. (Lamb birth weight is primarily influenced by “maternal genetic effects,” which derive from gene products of the mother's genome, rather than those of the lambs themselves.) The most powerful modeling result showed that the extent to which birth weight is heritable—that is, the variance caused by maternal genetic effects—increased along with environmental quality. But when Wilson et al. modeled the relationship between birth weight and fitness (defined as lamb survival) across environments, they found that the strength of selection followed a different pattern. Fitness increased with birth weight, and with environmental quality, but the positive relationship between fitness and birth weight—indicating the strength of selection—became weaker in better environments. Environmental quality shapes the trajectories for both genetic variance and the strength of selection. Lambs are not growing bigger and bigger because there's a lack of heritable variation for selection to act on in harsh environments and there's a lack of selection to act on higher genetic variation during favorable conditions. These results emphasize the importance of using biologically realistic models to predict the evolution of traits in wild populations—an important component of genetic restoration for endangered species, for example. They also confirm the value of long-term studies on wild populations to better detect patterns of selection and genetic variation in the natural world.

St Kilda: the pre-Improvement clachan

Archaeological studies of Hirta, the main island of the St Kilda archipelago to the west of the Western Isles of Scotland, have debated the location and character of the settlement demolished during the Improvement phase of the 1830s. This paper argues that archaeological evidence for the clachan is recoverable, and that it is in very good agreement with the drawings made by Acland and Atkinson in the early 19th century.

Modelling non-additive and nonlinear signals from climatic noise in ecological time series: Soay sheep as an example.

Understanding how climate can interact with other factors in determining patterns of species abundance is a persistent challenge in ecology. Recent research has suggested that the dynamics exhibited by some populations may be a non-additive function of climate, with climate affecting population growth more strongly at high density than at low density. However, we lack methodologies to adequately explain patterns in population growth generated as a result of interactions between intrinsic factors and extrinsic climatic variation in non-linear systems. We present a novel method (the Functional Coefficient Threshold Auto-Regressive (FCTAR) method) that can identify interacting influences of climate and density on population dynamics from time-series data. We demonstrate its use on count data on the size of the Soay sheep population, which is known to exhibit dynamics generated by nonlinear and non-additive interactions between density and climate, living on Hirta in the St Kilda archipelago. The FCTAR method suggests that climate fluctuations can drive the Soay sheep population between different dynamical regimes--from stable population size through limit cycles and non-periodic fluctuations.

Conditional Methodology for Individual Case History Data

Summary Key ecological studies involve the regular censusing of populations of wild animals, resulting in individual case history data which record when marked individuals are seen alive and/or found dead. We show how current conditional methods of analysing case history data may be biased. We then show how a correction can be applied, making use of results from a mark–recovery–recapture analysis. This allows a simple investigation of the effect of time-varying individual covariates such as weight that often contain missing values. The work is motivated and illustrated by the study of Soay sheep in the St Kilda archipelago.

Sex-ratio variation in Soay sheep

We studied the effects of ecological variables on the birth sex ratio of Soay sheep (Ovis aries) lambs on the island of Hirta, in the St Kilda archipelago, Scotland. Both individual- and population-level models were constructed. In the individual-based model, only population size was significantly associated with the sex of a lamb, with the probability of giving birth to a male lamb being positively associated with population size. However, this model explained a very small proportion of the variance in birth sex ratio. A multiple regression analysis of the annual population birth sex ratio also showed a slight increase in the proportion of males born in years following high autumn population density, but this result was not statistically significant. Population growth rate, Julian birthday, litter size, mother's age and weight, and the weather conditions during the gestation and neonatal period did not explain significant variation in the birth sex ratio.

Age, sex, density, winter weather, and population crashes in Soay sheep.

Quantifying the impact of density, extrinsic climatic fluctuations, and demography on population fluctuations is a persistent challenge in ecology. We analyzed the effect of these processes on the irregular pattern of population crashes of Soay sheep on the St. Kilda archipelago, United Kingdom. Because the age and sex structure of the population fluctuates independently of population size, and because animals of different age and sex respond in different ways to density and weather, identical weather conditions can result in different dynamics in populations of equal size. In addition, the strength of density-dependent processes is a function of the distribution of weather events. Incorporating demographic heterogeneities into population models can influence dynamics and their response to climate change.