Human disturbance has been observed to have effects on wildlife. These effects can be either positive or negative, depending on the study species; however, most of the research done to date has demonstrated negative effects that are reflective in the behaviour and reproductive biologyof the organisms. I focused on Cassin’s auklet, a burrow-nester species to study the effects of increasing tourism on the islands along the peninsula of Baja California, Mexico. I also studied the effects of adverse climatological conditions on breeding success of this species. First, using some biometrics of the individuals captured, I tested a method for aging Cassin’s auklet based on the iris colour. My results corroborate what was proposed before, juveniles have a dark-brown iris that shades into a complete white as they become adults. I also used morphometrics to determine the sex of the individual. When making intra-pair comparisons, it is possible to estimate the sex of the individual based on bill measurements. In 100% of the cases, individuals were correctly sexed using bill depth and width, as corroborated with DNA analysis. Second, using distance as a measure of disturbance, I looked onto the effects of people walking around the island. In general, breeding success was lower in those sites closer to the path or the village; older and more experienced individuals represented the vast majority of the burrows further away from the disturbance source. After that, using an experimental approach, I manipulated the amount of disturbance received by the chicks and recorded growth rate, as well as fledging weight, that may influence the possibility of survival for the chicks. I found that at the early stages of development, chicks grew at the same rate; however, chicks in the experimental groups reached a lower peak weight, compared to chick in the control group, and once fully feathered, chicks in the experimental group had a higher rate of weight loss, fledging lighter and earlier than control chicks. I also had the opportunity to explore the effects of adverse climatic conditions on breeding success. My results show that under a warming of the upper layer of the ocean, as was registered in 2005, Cassin’s auklets struggle to maintain a good body condition and, although attempting to breed, abandon the nest later on, to guarantee survival and another attempt to breed the following season should the condition improve. Breeding success decreased from nearly 70% in 2004 to less than 10% in 2005. Body condition was also lower in 2005, with a slight improvement the following year.

Changes in environmental conditions in early life can cause changes in the tempo and pattern of growth and development in animals. Natural selection favours processes that enable animals to make decisions that maximise Darwinian fitness. These decisions are influenced by trade-offs between current and future benefits. An episode of poor conditions (i.e. reduced nutrition, low temperature and changes in photoperiod) is generally linked to a slowing of growth. If adequate conditions are restored after this episode, growth rate is accelerated and normal adult size can be reached; in other words, ‘compensatory’ growth occurs. Compensatory growth has benefits in enabling a return to the typical size-at-age growth trajectory. Although this ability to alter growth rate provides a degree of adaptability, there is now increasing evidence that resource allocation to rapid growth carries various long-term costs. While there is experimental evidence that poor environmental conditions in early life can induce subsequent compensatory growth, little is known about the long-term effects of compensatory growth on locomotor and reproductive performance, and on lifespan. In this thesis, I investigated how different growth trajectories affected subsequent performance (i.e. locomotory capability, reproduction and lifespan), and how any such effects were influenced by the perceived time until the key life history event of reproduction. Using juvenile three-spined sticklebacks (Gasterosteus aculeatus), I showed that temperature manipulations early in life in three temperature treatments (low, intermediate and high, independent of food supply) or food restriction (with a constant temperature) affected skeletal growth trajectory not only during the manipulation itself, but also during a subsequent compensatory phase. To investigate the effects of time of year, all experimental groups of temperature and food manipulations were replicated at different seasonal periods (= Winter or Spring); to manipulate apparent time of year while holding initial size and maturity constant, a photoperiod manipulation was also undertaken at both seasonal times (ambient or delayed photoperiod). While there was compensatory growth (i.e. accelerated growth) in the food manipulation, temperature manipulations induced both positive compensatory growth (i.e. growth acceleration following exposure to low temperature) and also ‘negative’ compensatory growth (decelerated growth following exposure to high temperature). The outcome of these changes was that fish in all treatment groups reached the same average size by sexual maturity, despite having different growth patterns. However, early growth trajectories influenced both pre-breeding swimming endurance and its decline over the course of the breeding season, such that swimming ability was negatively correlated with compensatory growth whereas ‘negative’ compensatory growth reduced swimming ability less (Chapter 2). Reproductive investment (males: sexual ornaments and ability to build nests; females: first clutch size and mean egg size) was negatively affected by compensatory growth; positive effects of ‘negative’ compensatory growth on reproduction were found (Chapter 3). Interestingly, the effects of growth rate on subsequent swimming and reproductive performance were greater when the perceived, or actual, time until the breeding season was shorter (Chapter 2 and 3). These results implied that increased metabolic rates and cellular damage (e.g. oxidative stress) induced by compensatory growth negatively affected subsequent performance, while decelerated growth reduced the damage levels and so later performance was less affected. Under food manipulation, there were similar patterns: compensatory growth (i.e. accelerated growth) negatively affected locomotor and reproductive performance and the time until the breeding season altered the effects on performance (Chapter 4). To further examine trade-offs between growth rate and fitness parameters such as future reproductive investment and rates of senescence, I developed four theoretical models of increasing complexity with different growth-damage scenarios, ranging from assuming that the animal maximises growth regardless of any costs, through assuming a relationship between growth rate and mortality risk, to assuming growth leads to damage accumulation and that the animal is able to apportion resources between somatic growth, gonadal growth and investment in repair of damage. The models predicted that growth trajectories strongly influenced future reproductive investment irrespective of body size at the time of breeding, presumably due to the effects of damage accumulation in the run up to the breeding season; the predictions of the most complex model were closest to the experimental data on egg production (Chapter 5). Lifespan was different among treatment groups and also influenced by early growth trajectories. Compensatory growth negatively affected lifespan whereas ‘negative’ compensatory growth extended lifespan. Lifespan in female sticklebacks was positively related to egg production. Male sticklebacks lived for a shorter time when they showed less growth between their first and second breeding seasons, and a greater change in the duration of having a red throat between the first and second breeding season (an indicator of reproductive senescence). The costs of compensation were strongest when the perceived time until breeding was shortest (Chapter 6). Consequently, this thesis shows that environment conditions in early life have substantial effects on subsequent performances and lifespan. Moreover, results in this thesis strongly support the time-stress hypothesis, that is the time available until the onset of a key life history event, in this case reproduction, influences outcomes.

Diseases caused by parasites are responsible for immense human and animal suffering, declines in biodiversity, and substantial economic losses across the globe. It is therefore important to understand how parasites spread through and persist within natural populations, so that control interventions that aim to reduce individual infection risk can be designed and implemented appropriately. Crucially, most parasites exist within multi-host communities, and often appear to infect multiple sympatric species, all of which potentially play a role in parasite persistence. However, certain species may contribute disproportionately to transmission and be nearly completely responsible for the persistence of a parasite within a community. Identifying such “key hosts” therefore offers a means to appropriately target control interventions to maximise success. However, assessing host species contributions to parasite transmission within multi-host communities is a challenging task, and much insight can be gained from studies of model host-parasite systems. In this thesis, the transmission dynamics responsible for the persistence of several endemic Bartonella parasites (bacterial flea-borne haemoparasites) are investigated within wild sympatric populations of wood mice (Apodemus sylvaticus) and bank voles (Myodes glareolus) in northwest England. Bartonella infections were first identified to species-level according to an existing method based on length polymorphism of a fragment of the 16S-23S internal transcribed spacer region (ITS). Broad patterns of prevalence suggested that some species were host generalists while others were host-exclusive, indicating that different transmission dynamics underlie the persistence of each Bartonella species within the rodent community. Attempts to identify key transmission hosts for each Bartonella species based on the effect of past host population densities on infection risk proved inconclusive. However, finer-scale characterisation of Bartonella infections, using DNA sequencing, found that Bartonella species that appear to be generalists actually comprise a complex of genetic variants, the majority of which are host-specific, suggesting that transmission between host species is uncommon and limited to a relatively few host-shared variants. Furthermore, detailed characterisation of the flea community infecting wood mice and bank voles found that these Bartonella vectors were host-generalists, and that at least two flea species were able to transfer between individuals of different host species. This suggests that a lack of between-species transmission is likely to arise through different compatibility between host species and Bartonella variants, rather than as a result of current ecological encounter barriers (e.g. through differential Bartonella-flea or flea-rodent specificity). The results of an experimental manipulation of between-species transmission within these wild communities support the notion that between-species transmission of Bartonella parasites is uncommon. Across three woodland sites, bank voles were treated with a veterinary insecticide to remove their fleas and therefore reduce the rate of transmission of Bartonella from treated bank voles to the rest of the rodent community. Following treatment, risk of bank vole infection with bank vole-exclusive Bartonella variants was reduced, but there was no affect on the risk of bank vole infection with host-shared variants, nor risk of infection in wood mice with either wood mouse-exclusive or host-shared variants. Importantly, these treatment effects were best identified by grouping the parasites on a ‘functional’ (i.e. host-exclusive versus host-shared variants) rather than a taxonomic (i.e. Bartonella species) basis. Together, these findings highlight the importance of characterising parasite infections to as fine a scale as possible, and the value of using a combination of observation, genetic and experimental approaches to understand parasite transmission within complex natural multi-host systems.

I conducted a series of studies which looked at influences on avian clutch size. Firstly I examined the traditional view that the demands of rearing chicks create a bottleneck at which clutch size is shaped by natural selection. I considered whether instead other stages such as incubation might also be important. I proposed that reproductive demands at each stage of the breeding season may be interdependent, and by developing a mathematical model, I formalised the argument and showed that data on the relationship between the number of offspring and the expenditure of resources at many stages of the season could reveal the importance of natural selection on clutch size at each stage. I then reviewed the literature on the importance of incubation for clutch size determination. Results indicated that metabolic demands of incubation were appreciable and that the incubation of enlarged clutches imposed penalties on the adults. In a field study of kittiwakes I found that breeding success was depressed during incubation and chick rearing by enlargement of clutches and broods respectively. I measured metabolic rates of kittiwakes during incubation and found them to be comparable with those during chick-rearing. Secondly, I examined whether individual adults within populations differed in their reproductive capacities (i.e. whether there was a range of 'adult quality') and whether this could then affect clutch size. In a study of kittiwakes I found clusters of birds with similar breeding performance, but found that these clusters did not persist between years. In a study of swifts, I found that some individuals were consistently good breeders but that this had negligible effects on the distribution of lifetime reproductive success between individuals. I then examined whether the low clutch sizes and high survival of swifts might reflect a bet-hedging strategy in a fluctuating environment, but found little evidence of this. I looked at whether differences in the amount of space available at the nest site could account for differences in clutch sizes of kittiwakes, but could find no such evidence. Lastly I developed a theoretical model to look at how clutch size might be affected by changes in reproductive effort with age. I examined whether the predictions of optimality models were borne out by the more appropriate population genetics approach and found that in birds the optimality models are robust.

Increased interactions between humans and wild animals in and around protected areas have raised the risks for sharing diseases and parasites among them. Regular surveillance with intervention for these infections in such areas is therefore necessary for improving animal health and for controlling any spill-over of animal diseases into nearby human populations. Although both humans and non-human primates in the Gombe area in western Tanzania are infected with schistosomiasis, it is not known whether strains of their schistosomes are epidemiologically and genetically distinct. The distribution and transmission risk factors for the disease in these areas are also not well known. This study investigated the infection patterns of schistosomiasis in humans and non-human primates in Gombe National Park and surrounding villages of Mwamgongo, Bugamba, Kiziba and Mtanga and related the infection prevalence and intensity to locality and other demographic characteristics such as age and sex in humans. It also examined the dosage and number of praziquantel tablets administered to school children based on their weight and compared this to the dosage they would receive based on their height. The parasite fauna of baboons and vervet monkeys in Gombe was also examined to determine the parasite species assess whether their infection levels have changed over time. Snails were also sampled so as to gain a clear understanding of the species present in an area, their local distribution and infection status. The results showed a significant variation of S. mansoni prevalence between age groups in humans, which also depended on site. The parasite egg counts (intensity) also varied significantly between age groups and across study sites. The dosage range of praziquantel in mg per kg of bodyweight predicted by height was 23-43 (average: 35.2) while the dosage range given to children based on their weight was 29-78 (average: 45.2) and this variation was statistically significant (p > 0.0001). Overall, six children (5.3%) received a praziquantel dosage below the recommended range (30-60 mg/kg) based on their weight while two children (2.6%) would have received the drugs above the optimum range based on their height. The parasites identified in baboons and vervets included Trichuris spp., Physaloptera spp., hookworms and unidentified nematodes, while Paragonimus spp., Streptopharagus spp. and Schistosoma mansoni were exclusively detected in baboons. Molecular analysis of baboon schistosome eggs confirmed them to be S. mansoni. A GLM analysis indicated that the interaction between season and baboon troop was a significant predictor of parasite prevalence and intensity in baboons. Snails obtained from all streams except at Mtanga were identified as Biomphalaria pfeifferi based on morphology and DNA analysis. These findings indicate a high infection of intestinal schistosomiasis in these areas of western Tanzania, suggesting that the distribution of the disease in the country could be more widespread than previously thought. The results also confirm that the infection of intestinal schistosomiasis in the area is focal, with marked variations between adjacent villages. The study shows that while both weight and height estimate the amount of praziquantel dosage that is within an acceptable range, weight tends to underestimate the drug while height slightly overestimates it. It is therefore essential to conduct further field studies to test the usefulness of the dose pole in praziquantel distribution and evaluate the extent to which the method could be wasting the drug by giving more than the necessary dosage. As most parasites diagnosed in baboons and vervets are capable of infecting humans, these animals can potentially serve as reservoirs of human helminths given the regular human-wildlife interactions in the area. The implication of these observations to wildlife conservation and public health issues in the area has been explored.

Mammalian monogamy is puzzling from and evolutionary perspective because it is unclear why males, which have the potential to father a great many offspring, should choose to associate with only one female. This project investigated the behaviour of a socially monogamous (pair-living) population of bat-eared foxes in Laikipia, Northern Kenya, and had two principal aims. The first aim was to identify the selective forces that operate to maintain social monogamy in the study population. The second aim was to determine whether bat-eared foxes mate exclusively with their social partners (i. e. if they are genetically as well as socially monogamous). Chapter I summarizes by background to the research: Broadly speaking, theories advanced to explain the evolution of monogamy fall into two categories; those that proposing that monogamy occurs when male assistance is required for successfW reproduction, and those that proposing that aspects of female spatial and/or temporal distribution make it impossible for even the most competitive males to gain more than one mate. Chapter 2 describes the study site and general methods employed. Chapter 3 examines whether a requirement for paternal care maintains social monogamy by investigating the parental roles of males and females: I found that females invest very heavily in reproduction, feeding at close to maximum rate throughout lactation and suffering increased mortality rates during this period. Consistent with previous studies of the species, I found that males are heavily involved in the rearing of young, spending significantly more time than females close to breeding dens, and contributing to all aspects of cub care. The importance of male care was revealed by the fact that, after statistically controlling for the confounding effects of territory quality, the male den attendance was significantly associated with cub survival. Chapter 4 investigates factors other than the requirement for male care that may prevent males from achieving polygynous status: Social monogamy was not enforced because males were incapable of defending sufficient resources to support more than one female, as some male territories contained sufficient food to support two or more females. I found, however, that because females occupied largely exclusive ranges and had synchronized fertile periods, it was probably impossible for even the most competitive males to successfully defend more than one fertile female. Chapter 5 investigates the mating tactics of bat-eared foxes by comparing their behaviour during and outside the mating season: Neither male nor female foxes increased their home range sizes during the mating season, demonstrating that they do not roam widely in search of extra-pair mates. Time-budget data suggest that this may be because bat-eared foxes have little time available to engage in activities other than foraging. The behaviour of mated partners wass highly coordinated, particularly during the mating season, and the close proximity of mated partners did not reduce their feeding rate. Chapter 6 uses DNA microsatellite analyses to establish the paternity of bat-eared fox cubs: We found that for the vast majority of cubs (42 of 44) social fathers were most likely to be their true fathers. These data demonstrate a high level of genetic monogamy in the study population. Chapter 7 summarizes data from the thesis: I conclude that, although male care enhances offspring survival, there are circumstances under which males may gain from polygyny. Males are probably unable to attain polygynous status, however, because the spatial and temporal distribution of females, combined with intense competition for mates makes it impossible for them to defend more than one mate. Consistent with observations of occasional polygynous breeding from other bat-eared fox populations, I conclude that polygyny could only a viable male strategy if compliant females were willing to co-ordinated their behaviour. I argue that the high levels of genetic monogamy observed are probably consequence of the species insectivorous diet, which leaves individuals with little time to engage in activities other than foraging, and makes it easy for males to guard their own partners.

The oocyte of Xenopus laevis has been established as a system with which to study the export of many secretory proteins. When oocytes are micro-injected with hen oviduct mRNA, ovalbumin and lysozyme accumulate at different rates in the surrounding incubation medium. This thesis concerns an investigation of the molecular and cellular basis for this non-parallel secretion. Kinetic studies confirm that the intrinsic rate of lysozyme secretion from oocytes is approximately twelve times that of ovalbumin and show that not all of the lysozyme is available for export. This slower rate of ovalbumin export is maintained following injection of a range of concentrations of oviduct mRNA or of purified ovalbumin messenger, the latter having been obtained by hybridisation to cloned ovalbumin complementary DNA. These results suggest that the differential rates of secretion of these two proteins observed in oocytes are not the consequence of competition for amphibian or avian factors and show that oviduct-specific proteins are not required for ovalbumin secretion. To analyse the cellular basis for this non-parallel secretion, oocyte-fractionation protocols have been employed: if organelles of the secretory pathway could be separated, then it might be possible to distinguish between two possibilities. Firstly, that this non-parallel secretion reflects the relative retardation of ovalbumin during stages preceding the formation of the final exocytotic vesicles (containing both ovalbumin and lysozyme). Secondly, the two proteins might be segregated from one another and secreted by different tracks along the same pathway. Microsomal fractions containing oviduct proteins have been isolated from oocytes using sucrose gradients, but a demonstrable separation of Golgi- and ER-derived microsomal subfractions has not been obtained. Analysis of pulse-labelIed and pulse-chased oocytes on sucrose gradients show that non-secretory lysozyme slowly enters a novel high density oocyte compartment. Movement of lysozyme into this compartment may be prevented by the use of monensin.

Atg8 proteins play a major role in autophagy. Atg8 is involved in the formation of autophagosome and also, serves to recruit selective autophagy receptors, autophagy regulators and autophagy substrates. These receptors, regulators and substrates are characterized by the presence of the LIR motif that mediates Atg8 interaction. Despite the discovery of several Atg8 interactors in human and yeast, knowledge of these interactors in Drosophila is limited. The known Atg8a interactors (Drosophila homolog of Atg8) in Drosophila include Ref(2)P, functions in selective autophagy of ubiquitinated substrates whereas Atg1, functions in autophagy initiation. We searched for novel Atg8a interactors in Drosophila and their role in selective autophagy of ubiquitinated protein aggregates and autophagy regulation. Using a bioinformatics approach, we identified novel putative LIR-containing proteins in Drosophila proteome and characterized three of the promising candidates in vivo using various biochemical and molecular biology techniques. One of the candidates was UbcD4, a ubiquitin ligase containing putative LIR and UBA (involved in ubiquitin binding) domains. Despite this, we could not detect UbcD4 interaction with Atg8a, suggesting that the putative LIR is not functional. Localization and western blot analysis showed that UbcD4 accumulates as aggregates under Atg8a-deficienct conditions and colocalizes with various markers of protein aggregation. Knockdown of UbcD4 indicated that UbcD4 mediates aggregate formation in old flies but not in young flies. Thus, we identified a novel component of ubiquitinated protein aggregates that mediate aggregate formation in Drosophila brains under autophagy-deficient condition. The second candidate was PAR1, a serine/tyrosine kinase with putative LIR and UBA domains. Interaction studies demonstrated that PAR1 interacts with Atg8a. Further, PAR1 is not a major component of protein aggregates formed in response to Atg8-deficiency, demonstrating that PAR1 is not found with ubiquitinated protein aggregates and thus does not participate in the removal of ubiquitinated proteins through ubiquitin-dependent selective autophagy. Further, we showed that overexpression of kinase-dead PAR1 induces mCherry-Atg8a puncta (a marker of autophagy) under fed condition, suggesting a novel role of PAR1 in autophagy regulation. The third candidate was Sequoia, a putative LIR, and ZnF-C2H2 domains (involved in DNA-binding) protein. Sequoia interacts with Atg8a, and it does not accumulate under Atg8a-deficient condition. Knockdown of Sequoia and overexpression of LIR-mutated Sequoia induces autophagy under fed conditions in the larval fat body. Additionally, expression analysis indicated that Sequoia acts as a repressor of Atg7. Thus, we identified a novel Atg8-ineracting protein that negatively regulates autophagy under fed condition. To further the understanding of selective autophagy in Drosophila, we characterized the role of Ref(2)P in mediating the cytotoxicity associated with the expression of mutant huntingtin (Htt), a toxic protein that accumulates in Huntington’s disease (HD) brains. We examined whether overexpression of Ref(2)P ameliorates HD-associated phenotypes in a Drosophila HD model. This HD model faithfully recapitulates HD-associated phenotypes such as Htt aggregation, motor dysfunction, and short lifespan. It was observed that overexpression of Ref(2)P does not rescue the above-mentioned HD-associated phenotypes. These analyses suggest that selective autophagy receptor Ref(2)P does not mediate Htt-induced toxicity in Drosophila. Finally, a label-free shotgun proteomics was employed to identify proteins with increased accumulation in Atg8a-deficient versus wild-type conditions to further the knowledge of autophagy mechanisms. An increase in the accumulation of 69 and 57 proteins was identified in Triton-soluble and Triton-insoluble fraction respectively. These proteins may rely on autophagy for their degradation. These proteins were further searched computationally for the occurrence of the putative LIR motif. The results presented here open the pathway for the discovery of novel Atg8a interactors and autophagy substrates and thus provide insights into novel mechanisms of autophagy in Drosophila.

Communication can be crucial to the profitability of reproduction by allowing individuals to attract and select an appropriate mate. Across mammals, successful reproduction can depend on the ability of individuals to gain information such as relatedness, health parameters and breeding status from potential mates. Although visual and auditory signals are utilised, scent is a crucial and ancient form of communication yet, with the exception of certain model systems, we understand little of how it functions in wild mammals. This thesis will focus on the mechanistic role of odour signals: what information they contain and how they may facilitate reproductive decision-making in the banded mongoose (Mungos mungo). I use a wild but habituated population to conduct experimental odour presentations showing these mammals are capable of discriminating scents based on sex, familiarity, relatedness and female reproductive state. The ability of odours to encode such multi-modal information suggests they may facilitate key behavioural processes such as kin recognition, mate-choice and competitive interactions. However, the discrimination of pregnancy specifically implies scent cues function within reproductive decision-making, attracting males to receptive mates. The gastro-intestinal parasite community of this banded mongoose population was also screened, allowing the ability of odour cues to advertise parasitic infection to be tested. Observations show highly parasitised individuals scent-mark less frequently, suggesting marking behaviour indicates quality in terms of parasite burdens. Furthermore, experimental odour presentations show that banded mongooses exhibit behavioural aversions toward odours of heavily infected individuals. Scent cues, in the banded mongoose system, thus appear to encode a multitude of information relevant to reproduction.

Artificial Light at Night (ALAN) is among the fastest growing anthropogenic influences on the natural environment. ALAN has been suggested to affect the behaviour and physiology of nearly all vertebrates and invertebrates by reducing the distinction between day and night, and by altering the cues that activate nocturnal behaviours. Information is particularly scarce for freshwater ecosystems, many of which are close to sources of ALAN. This thesis examines the behavioural and physiological impact of broad spectrum ALAN on Atlantic salmon (Salmo salar L.) and their invertebrate prey. After reviewing available literature of the effects of ALAN on freshwaters (Chapter 1), a series of empirical field and laboratory experiments examined the impact of ALAN on i) invertebrate drift in an experimentally artificially lit stream (Chapter 2) to determine the influence of ALAN on the primary food source of Atlantic salmon; ii) the dispersal behaviour (Chapter 3) and cortisol stress response in dispersing Atlantic salmon fry (Chapter 4); and iii) the diel pattern of foraging and refuging in Atlantic salmon parr (Chapter 5). ALAN impacted the drifting behaviour of invertebrates from contrasting taxa with a divergent effect of ALAN between taxa and functional feeding groups (FFGs), with some increasing and others decreasing under part-lighting. In dispersing Atlantic salmon fry, ALAN disrupted the timing and periodicity of nocturnal dispersal behaviour, at all experimental light intensities (1 – 8 lux). However, this behavioural change was not the result of a cortisol stress response. Finally, ALAN affected activity levels of Atlantic salmon parr through disrupting the amount and timing of refuging behaviour, with fish housed under high intensity ALAN found to refuge 28 % more than those in the control treatment. These results highlight the complex nature of the response of both Atlantic salmon and their invertebrate prey to ALAN, whereby the influence of ALAN can be difficult to generalise between taxa and species’ life stages. Moreover, this thesis provides evidence to inform proposed mitigation strategies and advocates an increase in natural unlit areas.