A main challenge in evolutionary ecology is to elaborate the main ecological factors that vary in a study system and analyse how those factors shape the phenotype of organisms in their natural environment. These ecological variables can have complex effects since most life history traits are correlated. Here I examined the effects of ecological factors on life history traits, and the relationship among life history traits of North Uist three-spined stickleback, Gasterosteus aculeatus (hereafter stickleback). Both abiotic and biotic factors that are commonly measured in aquatic systems, and are known to vary greatly between North Uist lochs, were examined. The traits analysed were body size, age and size at maturation, proportion of fish over one year old, and the fecundity rate. My results showed that there was strong variation in water chemistry and inter-specific competition among the lochs on North Uist. Lochs on the west side of the island were characterised by higher sodium and calcium ion concentrations, higher pH and conductivity, and increased inter-specific competition. As a result, western populations had a faster absolute growth rate, larger body size, and a higher proportion of older fish, indicative of a longer life span. Although western populations had a larger body size, they tended to mature later with a smaller gonadosomatic index. They also favoured a larger number of eggs, but with a smaller egg size. Variations within population showed that larger females had greater fecundity but small egg size. Both plasticity and evolution appear to have contributed to these life history variations. Overall, variation in life history of North Uist stickleback was largely related to measures of resources availability and interspecific competition, except for egg size which was mostly associated with variation in predator density.

Global warming and temperature variation are likely to have profound impacts on fish as ectotherms that are heavily reliant on environmental temperature for growth, development, metabolism and immunity. This study addresses the impact of climate change on the development of infection and immunity in three-spined sticklebacks (Gasterosteus aculeatus) and their common parasites. In addition to thermal consequences on host parasite interactions, the study also addressed the effects of co-infection on parasite intensities and pathology on host swimming ability. Experiments were designed to mimic global warming, temperature variability and stochasticity (Chapters 3-5). Temperature during exposure to Saprolegnia parasitica was the major determinant of high infection prevalence and intensity with historical temperature exposure having little impact (Chapters 3-5). A further contributor to infection risk was higher host body condition (Chapter 3 and 5), attributed to a trade-off between host immunity and condition, higher condition individuals investing less in immunity supported by a decline in β-def expression in high condition fish (Chapter 5). Peak infection intensities in Gyrodactylus gasterostei were dependent on temperature variability and the host’s immune response. In variable conditions, an established G. gasterostei was better able to adapt to a changing environment than the host’s response causing higher peak infection intensities (Chapter 3 and 4). Temperature, and not photoperiod, was the major cause of circannual rhythm in host immunity (Chapter 6). Co-infection between G. gasterostei and A. foliaceus, revealed higher gyrodactylid infection peaks compared to hosts infected with G. gasterostei alone suggesting that immunomodulation by A. foliaceus. Lastly, pathology, rather than drag, reduced burst and long-term swimming performance of sticklebacks infected with A. foliaceus. Many of the factors highlighted have implications for aquaculture. High aquaculture feeding regimens, resulting in higher body condition, co-infection and temperature, all could severely increase morbidity and mortality of fish in a parasite species dependant manner.

Biological invasions can have dramatic detrimental impacts on ecosystems, however they also represent rich opportunities to study the evolutionary processes associated with them. Hybridisation and subsequent introgression are two such processes and are common among native and non-native species. The crucian carp, Carassius carassius (L.), is a European freshwater fish that is threatened throughout much of its native range by several factors including hybridisation and introgression with three non-native species, the goldfish, Carassius auratus (L.), the gibel carp, Carassius gibelio (Bloch), and the common carp, Cyprinus carpio (L.). The conservation of C. carassius is hampered by a lack of phylogeographic knowledge for the species and no knowledge of the extent or impact of hybridisation and introgression. Contemporary genomic approaches such as Restriction Site Associated DNA sequencing (RADseq) can offer unprecedented insights into such research areas, however RADseq comes with several sources of potential bias. Exploratory analyses in Chapter 2 show that two sources of bias in particular, null alleles and over merged ohnolog loci, are highly important in this dataset, but can be filtered using population genetics statistics. The filtered dataset is used in phylogeographic analyses in Chapter 3, along with microsatellite and mitochondrial DNA and show that C. carassius exists as two major lineages in Europe, which diverged approximately 2.26 million years ago, and should be treated as separate units for conservation. These lineages result from the C. carassius postglacial recolonisation routes thtough Europe, which are highly distinct from the general patterns seen in other freshwater fish species. These phylogeographic results showed high similarity between C. carassius in England and those in continental Europe, calling into question the presently assumed native status of C. carassius in England, which has been contentious in the past. Empirical tests of this status using microsatellites showed that, in fact, C. carassius is most likely introduced in England around the 15th century, raising interesting discussions pertaining to their conservation in the England. Lastly, in Chapter 5, microsatellite and RADseq approaches show that hybridisation between C. carassius and non-native species is prevalent where they are sympatric, however backcrosses are rare, and there is no evidence of further introgression between the species studied. Taken together, these results suggest that postzygotic mechanisms of isolation limit interspecific gene flow, and conservationists should focus further research on the direct impacts of non-native species and F1 hybrids.

Salmon became locally extinct from the River Mersey, northwest England, during the 1950s – 1970s due to deterioration in water quality and man-made barriers. Stray salmon began entering the River Mersey in the 1990s but a self sustaining population has yet to become established. The aim of the study was to review and investigate the recent history of the Mersey catchment, the current status of and factors effecting the salmon population and the potential for a natural recolonisation of River Mersey. The requirements of adult and juvenile salmon and homing and staying in salmon were reviewed. The physiochemical requirements of salmon are highly specific with connectivity of fundamental importance to upstream migration. In reviewing the status of the Mersey catchment flow manipulation, obstructions to migration, poor water quality and river modifications were common in all rivers. Adult and juvenile salmon have been captured in the Mersey catchment since 2000 but have consistently been caught in low numbers and smolts have not been captured. Genetic analysis was used to assign salmon entering the Mersey to their region of origin. The Mersey is dependent on stray salmon with the majority from rivers in the Solway and Northwest England areas. A tracking study was used to determine salmon behaviour and route choice in the Mersey catchment and salmon were found to be prevented from moving freely within or upstream of the lower Mersey catchment. Habitat surveys undertaken throughout the catchment revealed a general trend of key habitats existing upstream of barriers and inaccessible to adult salmon. Salmon are unable to recolonise the Mersey catchment in its current state. There are a range of management and restoration options available to restore salmon to the Mersey catchment but a coordinated and concerted effort is required to be successful.

Intraspecific genetic, morphological and life history structuring is evident in many taxa. Where such intraspecific structuring exists, study of the nature of the patterns displayed can reveal much about the evolutionary processes that operate during the early stages of divergence. Intraspecific structuring is particularly prevalent amongst fishes that occupy recently glaciated freshwater systems. One such species, the brown trout, Salmo trutta, was the subject of the work presented in this thesis. Genetic and morphological intraspecific structuring of brown trout was examined across a single but large dendritic catchment, the River Foyle, Ireland. Structuring was examined at three spatial scales (large-scale, compared between major sub-catchments; medium-scale, compared between tributaries within sub-catchments; small-scale, compared between streams within tributaries). The two general aims of the study were to look for any structuring in either phenotype or genotype in brown trout across the catchment and, if this was found, to look for landscape or environmental gradients that might be driving such structuring. Using a suite of 21 microsatellite markers that were chose for their ability to resolve population differences in this species elsewhere, this study identified clear and distinct genetic structuring. Brown trout collected from 28 sampling sites, resolved into 21 genetically distinct and discrete populations using a hierarchical approach implemented in STRUCTURE. The structuring was evidence across all three spatial scales. There was strong evidence of isolation by distance and isolation by environment playing a role in shaping the magnitude of the genetic differences between populations. Landscape variables which are shaped by anthropogenic impacts (urbanised area (measured as the number of houses in the catchment), proximity to farmland (measured as the distance to the nearest farm) and concentration of phosphorus in the water) showed the greatest effects in shaping the genetic population structuring (chapter 2). In a parallel study, the morphological structuring of brown trout from across the Foyle catchment was investigated at three spatial scales. Morphology was measured as the shape of brown trout determined by Geometric Morphometric Analysis of fixed position landmarks identified on photographs of trout taken from 22 sampling sites across the catchment. Very clear, statistically significant differences in morphology (fish shape) were evident for all the 21 sampling sites (one sampling site was removed from the analysis due to small sample size) with Canonical Variate Analysis resolving 21 discrete phenotypic groups. Morphological structuring was evident across all spatial scales (large, medium and small). Analysis showed that genetic distance and geographic distance between morphological groups was significantly correlated with morphology of populations, with morphological groups that were most divergent from each other also being most genetically distinct and geographically more distant. The effect of landscape and environmental variables driving morphology of populations was tested. In-stream substrate composition, water pH, stream order, site elevation, river gradient and the number of houses per km2 (representative of urban area) were all found to have a significant effect on morphology of populations. However, once the effect on morphology on these environmental variables were accounted for the residual effect of genetic distance was not significant (chapter 3). To attempt to discriminate between three alternative population genetic hypotheses for the origin of two alternative life history strategies in brown trout; freshwater residency and anadromy, the genetic structuring of brown trout was examined between life history strategy (anadromy or resident), between three sites and across two years (2013/2014) for brown trout collected from the Foyle catchment. There was no evidence of population structuring being attributed to life history strategy (that is no genetic differences between anadromous or resident trout). There was however strong and clear evidence of five genetic populations based on geographical site. Two sympatric populations were identified at each of two locations. However, both populations in each river were composed of both freshwater resident and anadromous brown trout, although the frequency of each life history strategy significantly differed between these rivers. The results of this study support the concept that partial migration in brown trout is most likely driven by a quantitative threshold trait, where the threshold trait value varies both between populations and between individuals within populations (chapter 4). It is critical, for effective management of the relatively high economic value anadromous component of brown trout populations in a catchment, to be able to identify which tributaries are contributing most to their production. A Genetic Stock Identification (GSI) analytical framework was used to determine the tributary of origin for anadromous brown trout captured from a mixed stock within the River Faughan sub-catchment, River Foyle and to look for any evidence of straying. The results showed that three genetic populations from specific parts of the sub-catchment contributed disproportionately to the production of anadromous brown trout. There was also evidence of straying of anadromous trout, particularly to one tributary elsewhere in the catchment (chapter 5). Taken together this body of work has demonstrated strong genetic and morphological structuring amongst brown trout in this large dendritic catchment. Genetic structuring seems to be at its most extreme when driven by factors which could be regarded as anthropogenic. This raises questions about human effects on the process of genetic divergence. Morphological structuring was, if anything even stronger than genetic structuring. Although there was evidence of genetic divergence between populations of differing morphologies, this neutral genetic differentiation was not a significant driver of morphological variation once landscape and environmental variables, such as substrate composition, driving morphological differences were taken into account. This suggests that the environmental drivers of structuring are greater in magnitude than neutral genetic divergence. Examining genetic structuring between two common morphologies of brown trout (anadromous and freshwater resident) in more detail, revealed no genetic differentiation between life history strategies but there was evidence of differences in frequency of life history between populations. Using the genetic structuring of brown trout as a genetic baseline it was possible to determine which tributaries within the River Faughan sub-catchment produce anadromous brown trout. If some discrete populations in a catchment are contributing disproportionately to the anadromous trout population (as they are in the Foyle) there is a strong risk of over exploitation and a need for enhanced attention in the nursery areas for those populations. These results have significant implications for the management of all trout in the Foyle catchment and elsewhere.

The development of fin damage in Atlantic salmon parr was investigated using social network analysis of behavioural interactions occurring under different feeding and stocking conditions. Four separate experiments were carried out in which groups of fish were subjected to a long food restriction period (30 days) described in Chapter 2, high (30 kg/m3) or low (8 kg/m3) fish stocking densities (Chapter 3), predictable or unpredictable food delivery regime (Chapter 4) or a short food restriction period (10 days) described in Chapter 5. Dorsal fin damage (erosion, splits and fin index) was significantly higher in groups of fish subjected to food restriction periods (short and long), held at high stocking density (30 kg/m3) or with an unpredictable food delivery regime. No other fins were found to be affected by fin damage irrespective of the treatment. The social networks based on aggressive interactions showed higher centrality, clustering coefficients, in-degree centrality, out-degree centrality and less dense networks in groups subjected to food restriction (short and long), low stocking densities 8 kg/m3), and unpredictable food delivery. The high centralities and clustering coefficients indicated separation of fish within the groups into initiators of aggression and receivers of aggression. This separation of roles was seen only in the food restricted group, high density groups and unpredictable food delivery groups. Initiators had higher out-degree centrality while receivers showed high in-degree centrality. Also, initiators of aggressive interactions had less fin erosion, higher final weights and higher body lengths than receivers of aggression. The severity of fin damage was significantly higher when this role differentiation occurred, and it was highly correlated with fin biting events. Additionally, overall aggression was higher in food restricted groups, low density groups and unpredictable groups. The dynamic analyses of networks over time (Chapter 6) showed that fish classified as initiators of aggression did not change this behaviour after normal/control environmental conditions were restored (i.e. ad libitum food delivery, low stocking density or predictable food delivery). The latter result indicates that individual fish maintained their behaviour irrespective of their social context, which is consistent with the definition of behavioural syndromes, personalities or temperament defined for other non-human animals. Overall these studies demonstrated the importance of using social network analysis to clearly identify and quantify roles that individual fish assume within their network group and through time based on their behavioural interactions leading to dorsal fin damage and differential physical characteristics. The results are potentially applicable within the commercial aquaculture industry as a valuable technique to evaluate and improve the welfare of farmed fish.

Placing our understanding of the function of the immune system into a more natural setting remains a fundamental challenge in biology, particularly how natural variation shapes the immune response and what the evolutionary consequences of such variation are. In this thesis, I use the three-spined stickleback as a model system for wild immunological studies. First, I developed a set of markers to measure the expression levels of key immune system genes using quantitative real-time PCR, representing the innate and adaptive immune response, and then used them to address a number of questions. I demonstrated that there are underlying differences in innate and adaptive gene expression levels between populations, as well as in innate immune response potential, which may reflect the contrasting challenges faced in these populations. By sampling individuals from multiple wild populations, I was able to demonstrate how a range of factors contribute to shaping immune system function, including sex, reproductive status, and infection with the common parasites Schistocephalus solidus and Gyrodactylus arcuatus. Next, I exposed laboratory raised fish to natural conditions and examined their response. Again, a range of factors where identified which appear to shape immune expression levels, particularly reproductive investment and infection with G. arcuatus. I also used this approach to demonstrate that immune system variation can be linked to Eda genotype, the gene which controls lateral plate phenotypic divergence during adaptive radiations. Finally, I performed a controlled infection experiment in the laboratory to show that both the innate and adaptive systems respond to Gyrodactylus infection. This thesis provides the basis for further immunological studies in stickleback, and adds to our growing understanding of the relevance of natural variation in shaping the immune response.

The diets of sympatric stream dwelling juvenile coho salmon (Oncorhynchus kisutch) and Dolly Varden (Salvelinus malma) were analyzed in five streams of different age and watershed complexity since glacial retreat in Glacier Bay, southeast Alaska. Interferometric synthetic aperture radar (IFSAR) data were analyzed with geographic information systems (GRASS GIS) to understand watershed change since deglaciation. Spatiotemporal variations of watershed characteristics from GIS were then used to develop a four-dimensional physical habitat template (PHT) and a directional model of stream development. Juvenile coho salmon and Dolly Varden coexisted in all streams with a strong overlap in diet while using different foraging zones. Diet niche breadth for both species increased with the amount of terrestrial insects in the diet as a result of resource depression, as mean fork length (FL) of juvenile fish increased when aquatic insects were a greater part of the diet, but decreased with greater terrestrial input. Using the stable isotopes of δ15N and δ13C, the contribution of marine derived nutrients (MDN) to stream biota was found to be determined by salmon spawner abundance and watershed retention, and an increase in δ15N and δ13C dispersion (SD) was found with stream age, likely causing a difference in the fractioning of δ15N and δ13C. The most significant change with stream age was denudation of watersheds and the development of stream networks. Dietary overlap was linked to fish resource depression, and controlled by stream geomorphology. Juvenile salmonids were found to consume sea lice (Lepeophtheirus salmonis) in the older streams.

River flow regime is fundamental in determining lotic fish communities and populations, and especially of salmonid fishes. Quantifying the effects of human induced flow alteration on salmonids is a key question for conservation and water resources management. While qualitative responses to flow alteration are well characterised, a more intractable problem is quantifying responses in a way that is practical for environmental management. Using data drawn from the Environment Agency national database, I fitted generalised linear mixed models (GLMMs) using Bayesian inference to quantify the response of salmonid populations to the effects of impounding rivers, flow loss from rivers due to water abstraction, and the mitigating effects of flow restoration. I showed that in upland rivers downstream of impounded lakes, the magnitude of antecedent summer low flows had an important effect on the late summer abundance of 0+ salmonids Atlantic salmon (Salmo salar) and brown trout (Salmo trutta). In contrast, the abundance of 1+ salmon and brown trout appeared to be largely unresponsive to the same flows. I demonstrated that short-term flow cessation had a negative impact on the abundance of 1+ brown trout in the following spring, but that recovery was rapid with negligible longer-term consequences. I further established that flow restoration in upland streams impacted by water abstraction provided limited short-term benefits to salmonid abundance when compared with changes at control locations. However, while benefits to salmonid abundance were limited, I detected important benefits to the mean growth rates of 0+ and 1+ brown trout from flow restoration. I discuss the implications of my findings for salmonid management and conservation and propose a more evidence-based approach to fishery management based on robust quantitative evidence derived using appropriate statistical models. The current approach to flow management for salmonids requires revision and I recommend an alternative approach based on quantitative evidence.

In this thesis, population level variation is elucidated for Atlantic salmon living in the chalk streams of southern England – a unique and unusual habitat – as well as in immediately surrounding regions. Salmon in these chalk streams have yet to be robustly investigated, despite individual populations standing out from neighbouring populations in several previous studies. This thesis attempts to identify how different they are and the reasons for it. Then, this thesis also investigates the effect of this distinction on their internal population structure, as well as the current and future trajectory. A panel of microsatellite markers from the SALSEA-merge project were used to complete four studies of population structure in Atlantic salmon. In the first study, which served primarily, as a training exercise, a multi-national baseline was used to identify the origins of salmon recolonising the river Mersey in northwest England. Fish entering the Mersey originated from multiple sources, with the greatest proportion (45–60%) assigning to rivers in the geographical region just north of the Mersey, including Northwest England and the Solway Firth. The number of fish originating from proximal rivers to the west of the Mersey was lower than expected. The results suggested that the recolonisers were straying in accordance with the predominantly clockwise gyre present in the eastern Irish Sea. In the second study, the relationship of salmon in the chalk streams of southern England to salmon outside this region was elucidated. Salmon from all five chalk streams in southern England with major salmon populations were found to all be genetically distinct from these neighbours and statistically less genetically diverse than salmon in southwest England and France. The reasons for this were relatively low immigration and a history of low effective population size. In the third study, the extent of population structure of salmon between the chalk streams and within one chalk stream, the river Frome, was explored. The results suggested these salmon were divided into three groups, i.e. 1) the Frome & Piddle, 2) the Avon and 3) the Test & Itchen. A significant pattern of isolation by distance between salmon in these five rivers was also identified. Historic samples from the Avon were assigned to the contemporary three groups. Surprisingly, most of these fish assigned to the Frome and Piddle group. Within the river Frome, further sub-structure was identified over two separate years of sampling. Salmon from 2009 comprised three genetic groups, and salmon in 2011 comprised just two. In the fourth study, historic scale samples were used to assess the current trajectory of genetic diversity and effective population size of salmon populations across Scotland, England, Wales and France. The majority of samples greater than 30 years old proved ineffective using the SALSEA panel. However, data was compiled from samples from eight rivers ranging from the Tweed in Scotland to the Scorff in France and from 1972 to 2012. Contrary to our hypothesis, most populations showed increases in allelic richness. Populations from one chalk stream show the steepest temporal decline in genetic diversity, which we speculate is partly due to the low immigration into the region. Effective population size proved difficult to determine using a number of methods and no robust pattern was identified. Together these studies indicate that low immigration of salmon into the chalk streams appears to be key to their low genetic diversity and genetic distinction. Low immigration may also have enabled marked within-river population structure and the current negative trajectory of genetic diversity. The implications for general understanding of Atlantic salmon population structure across their range, and for the conservation of this species are discussed.