Esta tese foca no papel dos misídeos em ecossistemas costeiros, considerando aspectos de sua distribuição, biologia e ecologia, tais como: distribuição zoogeográfica no Atlântico Sudoeste; dinâmica populacional dos misídeos no estuário de Cananéia, explorando sua distribuição espacial e temporal em relação às variáveis ambientais e abundância do zooplâncton; traços reprodutivos, produção anual e tolerância a gradientes de salinidade; taxas de alimentação, seletividade e respostas funcionais; predação inter- e intraespecífica; e predação sobre os misídeos. Atenção particular foi dada ao papel trófico dos misídeos no contexto da teoria da Biodiversidade e Funcionamento do Ecossistema em ecologia. Através de abordagens experimentais foram avaliados os efeitos da riqueza de espécies da comunidade de misídeos sobre sua seletividade alimentar e taxas de consumo. O efeito da riqueza de espécies da comunidade de predadores dos misídeos também foi considerado. Misídeos geralmente compõem a epifauna bêntica dominante em águas costeiras, tendo assim papel chave nas cadeias alimentares de regiões marinhas costeiras. No entanto, poucos estudos focaram esses organismos na América do Sul: apenas 31 de 1131 espécies de misídeos descritas no mundo foram registradas no Atlântico Sudoeste. No estuário de Cananéia, salinidade e temperatura foram parâmetros importantes na distribuição espacial e temporal dos misídeos. Maiores abundâncias e produção ocorreram na primavera, em decorrência de condições termohalinas mais favoráveis e a alta disponibilidade de alimento. Metamysidopsis elongata atlantica foi a espécie dominante numericamente, provavelmente por ser a única espécie de misídeo que consegue colonizar áreas de menor salinidade de forma efetiva, onde o alimento é mais abundante e as outras espécies de misídeos não ocorrem, evitando assim competição interespecífica e predação intraguilda. Metamysidopsis e. atlantica teve alta produção anual, fato que confirma a importância dos misídeos no fluxo de carbono em águas costeiras de regiões tropicas e subtropicais. Chlamydopleon dissimile, M. e. atlantica e Mysidopsis coelhoi tiveram comportamento alimentar oportunista sobre a comunidade de zooplâncton natural, mas houve evidência de que eles evitaram presas medindo entre 300 e 400 µm, compostas principalmente por Oithona spp. As taxas de ingestão tiveram um aumento não linear com o aumento da concentração de alimento, sugerindo resposta funcional do tipo II para as três espécies de misídeos. Houve tanto efeitos positivos (complementaridade) quanto negativos da combinação de espécies sobre a resposta alimentar das mesmas. Interações negativas estiveram provavelmente ligadas à predação intraguilda, resultando na redução da predação dos misídeos sobre o zooplâncton. Um aumento no número de espécies de predadores de misídeos resultou em um aumento na predação de misídeos em relação ao desempenho médio das monoculturas de predadores, mas não em comparação ao predador mais efetivo. Efeitos positivos da diversidade de predadores aumentaram com a heterogeneidade espacial, pois esta permitiu que efeitos de complementaridade interespecífica entre os predadores fossem melhor expressados. Além disso, a morte de predadores foi reduzida nos tratamentos com maior heterogeneidade espacial, provavelmente porque esta amenizou a ocorrência de predação intraguilda e/ou interações comportamentais negativas entre os predadores. E por último, efeitos de cascata trófica não foram observados, provavelmente em decorrência dos misídeos serem onívoros generalistas. / This thesis focuses on the role of mysids in coastal ecosystems. It considers aspects of their distribution, biology, and ecology, such as: zoogeographical distribution in the Southwest Atlantic; their population dynamics in the Cananeia estuary, exploring spatial and temporal distribution in relation to environmental variables and zooplankton abundance; reproductive traits, annual production and tolerance to salinity gradients; feeding rates, selectivity, and functional responses; inter- and intraspecific predation; and the suppression of mysid abundance by predators. Particular attention was put to the trophic role of mysids in the context of the Biodiversity and Ecosystem Functioning theory in ecology (BEF). Experimental approaches were employed to evaluate the effects of species richness of the mysid assemblage on their feeding selectivity and consumption fluxes. Also, the effect of species richness of the mysids predador assemblage was taken into account. Mysids are usually the dominant benthic epifauna in coastal waters, and thus they play a key role in marine coastal food webs. However, few studies focused on these organisms in the South American region: only 31 out of 1131 mysid species described worldwide have been registered in the Southwest Atlantic. In the Cananeia estuary, salinity and temperature were important factors influencing mysid spatio-temporal distribution. Higher abundance and production occurred in spring, when thermohaline conditions were more favorable and food availability was high. Metamysidopsis elongata atlantica was the numerically dominant species, probably because it is the single mysid species that effectively colonizes low salinity areas, where food is more abundant and other mysid species are absent, thus avoiding interspecific competition and intraguild predation. Metamysidopsis e. atlantica had high annual production values, confirming the importance of mysids in the carbon flux of tropical and subtropical coastal waters. Chlamydopleon dissimile, M. e. atlantica, and Mysidopsis coelhoi had opportunistic feeding behavior on the natural zooplankton assemblage, but there was evidence of avoidance of medium-size prey (300-400 µm), mostly comprised by Oithona spp. Ingestion rates showed a non-linear increase with increasing food concentrations, suggesting a type II functional response for the three mysid species. There were both positive (complementarity effect) and negative effects of species combinations on their feeding response. Negative interactions were probably linked to intraguild predation, resulting in a reduction of mysid predation over the zooplankton. An increase in the number of mysid predator species enhanced mysid suppression relative to the mean performance of predator monocultures, but not in comparison to the most effective predator. Positive effects of predator diversity increased with spatial heterogeneity, as it allowed interspecific complementarity effects between predators to be expressed. Moreover, mortality of predators was reduced in spatially heterogeneous treatments, most likely because it dampened intraguild predation and/or negative behavioral interactions between predators. Finally, trophic cascades were not observed, probably due to the omnivorous - generalist character of mysids.

An outbreak of Deidamia inscriptum (Lettered Sphinx Moth) caterpillars was noted in northeast Tennessee where Oxydendrum arboreum (Sourwood) trees were defoliated. Nearly all published literature and online resources list only plants in the grape family (Vitaceae) as larval food plants. Food-plant preference trials using fresh leaves of 3 woody plant species showed that Deidamiacaterpillars from this region had a preference for Sourwood over Parthenocissus quinquefolia(Virginia Creeper), and rejected Acer rubrum (Red Maple), a non-host species. Ursus americanus(Black Bear) were feeding on the caterpillars as evidenced by bent and broken Sourwood saplings bearing claw marks and by abundant sphingid remains in bear scat.

Greater sage-grouse (Centrocercus urophasianus; hereafter, sage-grouse) are a species of conservation concern in the rangelands of western North America due to their dramatic decline over the last half century. Effective conservation and management of sensitive species requires an understanding of how species respond to management actions. We examined two aspects of the reproductive phases of sage-grouse: nest predation, and habitat selection by female sage-grouse with chicks. In Chapter 1, we developed resource selection functions to assess the influence of mechanical treatments of mountain big sagebrush (Artemisia tridentata vaseyana) on habitat selection by greater sage-grouse with chicks. Post-treatment sage-grouse showed stronger selection for treatments and treatment edges than did pre-treatment sage-grouse. This altered pattern of selection by sage-grouse with broods suggests mechanical treatments may be a suitable way to increase use of mountain big sagebrush during the brooding period. In Chapter 2, we assessed the effect of habitat edges on nest predation of sage-grouse. The "edge effect" hypothesis states that habitat edges are associated with reduced nest success for birds. We tested the edge effect hypothesis using 155 nest locations from 114 sage-grouse. We derived edge metrics for 11 habitat cover types to determine which variables may have affected nest predation. We found support for the edge effect hypothesis in that nest predation increased with increasing edge density of paved roads. We provide evidence that the edge effect hypothesis may apply to greater sage-grouse and their habitats. Based on our results, we recommend minimizing disturbances that fragment critical nesting habitat of greater sage-grouse.

Lethal biotic interactions strongly influence the potential for aquatic non-native species to establish and endure in habitats to which they are introduced. Predators in the recipient area, including native and previously established non-native predators, can prevent establishment, limit habitat use, and reduce abundance of non-native species. Management efforts by humans using methods designed to cause mass mortality (e.g., trapping, biocide applications) can reduce or eradicate non-native populations. However, the impacts of predator and human induced mortality may be mitigated by the behavior or population-level responses of a given non-native species. My dissertation examined the responses of non-native aquatic species to the risk of predation by novel (i.e., no previous exposure) predators in the recipient community and indicators of potential compensatory responses by non-native populations to increased mortality resulting from removal efforts. My dissertation addresses four primary questions. 1) Can first generation, naïve invaders recognize and defend against predators found within the region of invasion through the expression of inducible defenses? 2) Can the overcompensatory potential of a population be predicted through examinations of intraspecific interactions of individuals from the population? 3) What is the relationship between removal effort outcome (i.e., successful or unsuccessful reduction of the target population) and compensatory population responses? 4) Is there a relationship between characteristics of removal efforts that are typically available to managers (e.g., target area size, target area connectivity, removal methodology) and compensatory population responses that could indicate the relative likelihood of compensation resulting from removal efforts? An invading species should be more likely to establish if it can successfully identify and defend against predators in the recipient range, such as through the expression of inducible defenses. Inducible defenses are behavioral or physiological changes that reduce an organism's susceptibility to predation. Through a series of laboratory experiments, I tested whether inducible defenses, in the form of increased burrowing depth, may have benefited the early stage of invasion of Nuttallia obscurata (purple varnish clam), an established Northeast Pacific invader. Specimens of N. obscurata were collected from introduced populations in the Northeast Pacific and from a native population in Japan. The clams were exposed to chemical and physical cues from Northeast Pacific crab predators, including the native Metacarcinus magister (Dungeness crab), an abundant and frequent predator of N. obscurata. While introduced N. obscurata increased their burrowing depth in the physical presence of M. magister, clams collected from their native range showed no such response. This lack of increased burrowing depth by naïve clams in response to a predator native to the newly invaded range, but a significant increase in depth for clams from populations established in the range suggests that while inducible defenses likely did not contribute to the initial establishment of N. obscurata in the Northeast Pacific, they may contribute to their continued persistence and expansion in their introduced range. Some efforts to reduce invasive populations have paradoxically led to population increases. This phenomenon, referred to as overcompensation, occurs when strong negative density-dependent interactions are reduced through increased mortality within a population, resulting in an increase in the population's recruitment rate sufficient to increase the population's overall abundance. Increases in a population's recruitment rate can result from reduced cannibalism of juveniles resulting in lower mortality of new recruits, from increased adult reproductive output, which increases the number of potential recruits, or from reductions in size and/or age at maturity of the unharvested population, which increases the number of reproductive individuals. I predicted the overcompensatory potential of a population of Carcinus maenas (European green crab) in Bodega Harbor, California, using a series of laboratory and field experiments examining intraspecific pressures of adults on juveniles in the population. This measure of intraspecific pressure was used to predict the overcompensatory potential of the population in response to increased mortality from ongoing removal efforts. This prediction was then assessed using pre- and post-removal surveys of juvenile recruitment in Bodega Harbor compared to nearby populations, testing for evidence of overcompensation. While adult C. maenas in Bodega Harbor had limited negative impacts on juveniles, I concluded it was unlikely to result in overcompensation. Relative juvenile abundance did not statistically increase in removal compared to reference populations, consistent with my conclusion from the experiments. Increases in recruitment rates can occur as a result of efforts to remove non-native species. This increase in recruitment can result in overcompensation, but more commonly results in compensation, where recruitment rates increase relative to pre-removal recruitment but does not result in in the population's abundance exceeding pre-removal levels. However, a detailed and accurate prediction of the response of a population to harvest is time consuming and data intensive. This is not feasible for most efforts to eradicate non-native species, which have the greatest chance of success when enacted rapidly after detection. For my final chapter, I performed a literature review and accompanying statistical analysis to determine if typically available information related to the removal effort (site size, site connectivity, and removal technique) could be used to determine increased risk of compensation for a given effort to remove aquatic invasive species. Compensation was closely linked to unsuccessful removal efforts and was observed only among efforts utilizing physical removal methods. However, the frequency with which compensation occurred varied with the exact technique employed, occurring most frequently in removal utilizing electrofishing. Additionally, evidence of compensation was more frequent among larger removal areas with variable connectivity. While other predictors (temperature, effort, etc) might add to the predicative power, the findings of the review provide criteria for managers to determine the relative risk of compensation prior to the start of removal. Further understanding of how invasive species respond to lethal biotic interactions, including anthropogenically mediated control measures, can aid in assessing the risk of invasion for a given species and inform managers of the risk of complications resulting from removal efforts. While inducible defenses may contribute to the long-term success of an introduced species in their recipient range, my findings did not support the idea that inducible defenses triggered by predator cues contributed to their initial introduction in this case. However, research on other non-native species and offspring of previously naïve prey would allow for a clearer picture of the role of inducible defenses in the invasion process. Compensation resulting from removal efforts does not guarantee failure, and certain characteristics of removal efforts seem to indicate increased risk of compensation. Together these components help identify how biotic interactions surrounding mortality risk of an invading species help shape the trajectory of invasion.

Coastal anthropogenic infrastructure has significantly modified nearshore environments. Because these structures often have a strong association with shipping as would be found in ports and harbors, they have been identified as invasion hotspots. Due to propagule pressure from shipping and recreational boating and suitable uncolonized substrate that provides a refuge from native predators, a greater number of non-native species have been found on these structures compared to nearby natural substrate. The mechanisms that limit the spread of non-native species from anthropogenic structures to natural substrate have been explored for several taxa at a species-specific level, but less so from an overall community perspective. Predation has been identified as one of the biotic interactions limiting invasion success. In addition to predation, dispersal ability may also prevent the spread of non-native species from anthropogenic structures to natural substrate. This thesis addresses how these two mechanisms interact to limit the spread of non-native species from anthropogenic structures to natural substrate and how that alters overall community composition. I aimed to explore differences between communities inside and outside of a marina and determine the extent to which predator and dispersal limitation were structuring these communities. I used a three-factor design, deploying seven unglazed ceramic tiles per each treatment combination of 1) inside versus outside a marina in Yaquina Bay, Oregon; 2) cage keeping out predators greater than the mesh size, no cage, or partial cage; 3) fixed near the substrata (benthic) versus suspended 1 meter below the surface. I also transplanted caged, suspended tiles of either adults or recruits from inside the marina to benthic and suspended caging treatments outside of the marina. These tiles allowed me to examine predation when dispersal limitation was not a factor for the community inside the marina, i.e. what happens to both recruits and adults if they can get outside of the marina. I found that the communities inside and outside of the marina were different and the data suggest that both predation and dispersal limitation interact to limit the spread of non-native species. Additionally, I found that mesopredators that could fit through the caging may be influencing predation results and community structure. This research addresses gaps in scientific knowledge regarding the mechanisms that prevent or facilitate the spread of non-native species. Future work could include the further exploration of mesopredation as an important factor in limiting the spread of non-native species and exploring dispersal limitation more in depth as well as broadening the geographic scope to see if the same trends hold true across bays and bioregions.

I studied aspects of the ecology of the Murray River turtle, Emydura macquarii, to determine the impact of the introduced red fox, Vulpes vulpes. The fox is one of Australia's worst vertebrate pests through its predation on livestock and native mammals, but their impact on reptilian communities is not known. I conducted a large-scale mark-recapture study to evaluate population growth of E. macquarii in the Albury region of the upper Murray River by determining growth, reproduction and survival. The study was conducted downstream of the first, and largest, impoundment on the Murray River, Lake Hume. Emydura macquarii predominantly inhabit the lagoons in the upper Murray River, as the mainstream and Lake are possibly too cool to maintain metabolic processes. They are easily captured in hoop traps and the use of live decoys maximises trap success. Over 2000 hatchling turtles were marked and released into two lagoons between January 1997 and January 1998. Growth of these individuals is rapid over the first few years but declines towards maturity, and is indeterminate after maturity. Although growth annuli are not well defined, even on young individuals, the von Bertalanffy model describes the growth of both male and female E. macquarii. Male turtles mature at 5-6 years and females mature at 10-12 years. Female turtles may maximise reproductive potential by delaying maturity and producing one relatively large clutch (mean = 21 eggs) per year, which is positively correlated with body size (PL). Although primarily related to body size, clutch size varies annually because of environmental conditions. If winter and summer rainfalls are below average and temperatures are above average, E. macquarii may reduce clutch size to increase the chance of the eggs surviving. Nesting predominantly occurs during the first major rain-bearing depression in November. Habitat variables, including distance from water, nearest nest, and tree, and soil type were measured for each nest to determine characteristics that attract predators. Nests close to the shoreline and trees are heavily preyed on, and nests constructed in sand are less likely to be destroyed by predators. Foxes detect nests through a combination of chemical cues from eggs and slight soil disturbances, whereas birds only destroy nests observed being constructed during the day. Female turtles alter nesting behaviour and construct nests much further away from water when foxes were removed and as a result, nests are less dense and away from trees. Thus in high predation risk areas, turtles minimise emergence and search times to reduce the risk of direct predation by foxes. Predation is reduced when nests are in lower densities and away from trees, because predators increase search efforts when nests are in higher densities and birds are more likely to destroy nests close to trees. Reproductive success is further reduced in high predation risk areas because more nests are constructed in sandy substrates where clutch success is reduced compared to incubation in more dense substrates. Where predators are a significant source of mortality, prey may use indirect methods, such as chemical recognition, to avoid encounters. Nesting turtles did not avoid areas where fox odour was present, suggesting that they assess predation pressure from foxes by other mechanisms, such as visual recognition. However, an innate response occurs to the odour of a once common predator on the Murray River, the eastern quoll (Dasyurus viverrinus), whereby turtles recognise and avoid nesting in areas where quoll odour is present. Therefore nesting turtles show a similar avoidance response to two different predators, using different mechanisms of detection. Similarly, predation risk may influence hatching times and nest emergence. The rate of embryonic development of E. macquarii may increase or eggs may hatch early so that the clutch hatches synchronously, thereby reducing the risk of predation through group emergence from the nest. Emydura macquarii reach densities of over 100 turtles.ha-1, with the majority of the population consisting of sexually mature individuals. Emydura macquarii has a Type III survival curve where mortality is extremely high in the egg stage (93% nest predation), remaining high over the hatchling stage (minimum survival rate- 10%), but decreasing rapidly throughout the juvenile stage (~70% juvenile survival). Adult survival is extremely high, with greater than 95% of adults surviving each year. Foxes through nest predation cause most mortality but a small proportion (~3%) of nesting adult females are killed by foxes each year. A removal program evaluated the impact of foxes. In 1996, fox numbers were monitored around four lagoons by spotlighting and non-toxic bait uptake. Foxes were removed from around two of the lagoons throughout 1997 and 1998, using spotlight shooting and 1080 bait poisoning. Fox numbers were continually monitored around all four lagoons during the study. Nest predation rates remained around 90% in all sites where foxes were present, but fell to less than 50% when foxes were removed. At the same time, predation on nesting female turtles was eliminated where foxes were removed. Demographic models using staged based survival schedules, together with growth and fecundity values for E. macquarii show a decline of 4% per year in these populations. Elasticity analyses shows that survival of adult female E. macquarii has the major influence on population stability and a reduction of nest predation alone is unlikely to address the population decline. Management options, such as reducing foxes prior to nesting around key lagoons, will stabilise the population decline, and eliminating foxes completely from certain areas with high dispersal potential, will promote recruitment of juvenile E. macquarii.

<p>In this thesis I investigate predator-induced effects on behavioural and life-history characteristics of prey. At any moment a given predator is capable of attacking a small number of prey. However, the mere presence of a predator may impact a much larger number of individuals, as prey implement various behavioural and developmental mechanisms to reduce the risk of predation. It has become increasingly clear that predator induced responses have the potential to affect patterns of species abundance and distribution as well as individual fitness of prey. I study these responses by incorporating field surveys, semi-field experiments and laboratory experiments. All experiments were done in an aquatic environment using fish or large odonate larvae as predators and damselfly-or diving beetle larvae as prey.</p><p>My work highlights the importance of monitoring prey behaviour when studying life-history characteristics. I show that fish presence is an important factor for determining species abundance and distribution of odonates, and that prey behaviour may be a good predictor for fish vulnerability. Larval damselflies react behaviourally to predator presence by reducing activity and/or restricting habitat use. I confirm that such anti-predator responses have positive effects on prey survival in the presence of a predator but negative effects on growth and development of prey. In addition, my results suggest that the increase in per capita food resources for surviving prey following a predation episode (i.e. thinning) can have a stronger positive effect on prey growth and development than the negative effect of anti-predator responses. I also show that the strength of an anti-predator response is dependent on resource availability of the prey, with prey responding less strongly when resources are scarce. My results also indicate that the strength of the anti-predator response of damselfly larvae depends on predator diet and larval age. Predators feeding on prey conspecifics induce a stronger behavioural response in young larva than predators that feed on prey heterospecifics do. This diet-effect was not found in larvae late in ontogeny, due to an increased activity of larva where predators consumed damselflies. Such increased larval activity can be explained as a reaction to a time-constraint. Finally, I found that activity of damselfly larvae is genetically determined and that this has lead to a behavioural syndrome that might limit larval plasticity to a certain activity-range. This phenomenon may have implications for how well larvae are able to react to both biotic and abiotic changes in the environment.</p>

<p>This thesis focuses on a black-tailed godwit (<i>Limosa limosa</i>) population, consisting of 35-40 pairs, that breeds on a grazed shore meadow on SE Gotland, Sweden. The small size of the population makes it more prone to extinction due to chance events, than a larger population.</p><p>The godwits showed microhabitat preferences when choosing nest sites. Godwit nests had higher vegetation over the nest cup, lower surrounding (1-3 m) vegetation and different plant species composition, as compared to random sites. Breeding near conspecifics or other wader species, especially lapwings (<i>Vanellus vanellus</i>) and further away from potential predator perches were the most important factors in decreasing nest predation. A comparison between different shore meadows along the east coast of the island revealed that large, open areas suffered less from nest predation. Thus, shore meadows suitable for breeding godwits should be large and without trees or other predator perches and have a grazing regime that favours variation in vegetation height.</p><p>Over 80% of previously ringed adults returned each year, but very few birds ringed as chicks were recovered. Hatching success was 55-60% for all observed nests. To predict the future of the current population, demographic data were used in an ecological risk analysis. The simulations showed that the Gotlandic population will not survive the coming 40 years without immigration.</p><p>Black-tailed godwits are divided into three subspecies. Genetic analyses (mtDNA) revealed that all subspecies had unique haplotypes and there was a clear geographic structure among subspecies. Within the <i>limosa</i> subspecies, godwits on Gotland and Öland showed a high proportion of rare haplotypes, but no genetic variation was found in Dutch birds. These results imply that black-tailed godwits on Gotland and Öland have high conservation value.</p>

This thesis focuses on a black-tailed godwit (Limosa limosa) population, consisting of 35-40 pairs, that breeds on a grazed shore meadow on SE Gotland, Sweden. The small size of the population makes it more prone to extinction due to chance events, than a larger population. The godwits showed microhabitat preferences when choosing nest sites. Godwit nests had higher vegetation over the nest cup, lower surrounding (1-3 m) vegetation and different plant species composition, as compared to random sites. Breeding near conspecifics or other wader species, especially lapwings (Vanellus vanellus) and further away from potential predator perches were the most important factors in decreasing nest predation. A comparison between different shore meadows along the east coast of the island revealed that large, open areas suffered less from nest predation. Thus, shore meadows suitable for breeding godwits should be large and without trees or other predator perches and have a grazing regime that favours variation in vegetation height. Over 80% of previously ringed adults returned each year, but very few birds ringed as chicks were recovered. Hatching success was 55-60% for all observed nests. To predict the future of the current population, demographic data were used in an ecological risk analysis. The simulations showed that the Gotlandic population will not survive the coming 40 years without immigration. Black-tailed godwits are divided into three subspecies. Genetic analyses (mtDNA) revealed that all subspecies had unique haplotypes and there was a clear geographic structure among subspecies. Within the limosa subspecies, godwits on Gotland and Öland showed a high proportion of rare haplotypes, but no genetic variation was found in Dutch birds. These results imply that black-tailed godwits on Gotland and Öland have high conservation value.

In this thesis I investigate predator-induced effects on behavioural and life-history characteristics of prey. At any moment a given predator is capable of attacking a small number of prey. However, the mere presence of a predator may impact a much larger number of individuals, as prey implement various behavioural and developmental mechanisms to reduce the risk of predation. It has become increasingly clear that predator induced responses have the potential to affect patterns of species abundance and distribution as well as individual fitness of prey. I study these responses by incorporating field surveys, semi-field experiments and laboratory experiments. All experiments were done in an aquatic environment using fish or large odonate larvae as predators and damselfly-or diving beetle larvae as prey. My work highlights the importance of monitoring prey behaviour when studying life-history characteristics. I show that fish presence is an important factor for determining species abundance and distribution of odonates, and that prey behaviour may be a good predictor for fish vulnerability. Larval damselflies react behaviourally to predator presence by reducing activity and/or restricting habitat use. I confirm that such anti-predator responses have positive effects on prey survival in the presence of a predator but negative effects on growth and development of prey. In addition, my results suggest that the increase in per capita food resources for surviving prey following a predation episode (i.e. thinning) can have a stronger positive effect on prey growth and development than the negative effect of anti-predator responses. I also show that the strength of an anti-predator response is dependent on resource availability of the prey, with prey responding less strongly when resources are scarce. My results also indicate that the strength of the anti-predator response of damselfly larvae depends on predator diet and larval age. Predators feeding on prey conspecifics induce a stronger behavioural response in young larva than predators that feed on prey heterospecifics do. This diet-effect was not found in larvae late in ontogeny, due to an increased activity of larva where predators consumed damselflies. Such increased larval activity can be explained as a reaction to a time-constraint. Finally, I found that activity of damselfly larvae is genetically determined and that this has lead to a behavioural syndrome that might limit larval plasticity to a certain activity-range. This phenomenon may have implications for how well larvae are able to react to both biotic and abiotic changes in the environment.