The major aim of this thesis project was to determine the variables that most explain the elevated mercury concentrations in lake trout (<i>Salvelinus namaycush</i>), a predatory aquatic fish species in some lakes in northwestern Canada. High mercury concentrations in lake trout in other regions have been associated with the biological features of the fish and various chemical and physical aspects of their aquatic ecosystems. Data including lake trout age, length, weight, and stable isotope values, water chemistry, latitude, and lake and watershed area were collected, compiled and then included in statistical analyses of the factors affecting mercury concentration in the muscle of lake trout from a series of lakes from the Mackenzie River Basin (MRB) in the Northwest Territories (NT), Canada. These results are reported in Chapter 2. Fish age and lake surface area were the most important variables affecting mercury concentrations. However mercury concentration in muscle also was significantly (p < 0.05) related to: fish length, weight, and δ13C; watershed area to lake area ratio; and to total mercury concentration in zooplankton and water. These variables were run through best subsets analyses and multiple regressions in order to determine the regression equation most efficiently capable of predicting mercury concentration in lake trout in unstudied lakes in the MRB region. The resulting equation was: log Hg = 0.698 (0.0156 × latitude) + (0.0031 × age) + (0.000535 × length) (0.245 × log lake area) + (0.00675 × watershed area/lake area ratio), r2 = 0.73<p>Small lakes located in the southern NT and dominated by large and/or old lake trout are most likely to have lake trout whose mean mercury concentrations exceed 0.5 μg/g; the guideline for the commercial sale of fish. Latitude may be linked to mean annual temperature (and variables such as duration of ice cover, summer water temperature) while fish age and length may be related in part to fishing pressures and growth rates on these lake populations. In chapter 3, a more in-depth study was undertaken to investigate of role of feeding and relative tropic level in the bioaccumulation of mercury in lake trout. This was accomplished by comparing MRB lake trout population characteristics with those from a series of lakes in northern Alberta and Saskatchewan (NAS). The two population groups were compared with respect to size, age, growth rates, and mercury concentrations. In addition, trophic and mercury biomagnification relationships, as inferred from stable carbon and nitrogen isotope analyses, for the two lake trout populations were compared. Lake trout from the NT exhibited significantly higher mercury concentrations than those from the NAS lakes (p < 0.001). Mercury concentrations in biota (including lake trout, forage fish, benthic invertebrates and zooplankton) were positively and significantly correlated to δ15N values in all lakes in both of the study areas (p < 0.001). Mercury biomagnification in the NT lakes, as estimated from the slope of δ15N versus mercury concentration, was lower than in the NAS lakes. Thus, mercury biomagnifies more slowly in NT lake trout, but because of their greater mean age, reaches higher values than in NAS lakes. Northwest Territory lake trout generally exhibited more negative δ13C values, indicating more pelagic feeding habits than in NAS lakes: higher mercury concentrations previously have been associated with more pelagic feeding. <p>Finally, the relationship between mercury levels and growth rates in lake trout was investigated by comparing NAS and NT lake trout populations. These results are reported in chapter 4. Lake trout from the NT lakes grew at a slower rate (10.4 mm per year) than those from the NAS lakes (35.1 mm per year). Log mercury concentration was inversely correlated (p < 0.001) with growth rate for both lake trout populations; however, growth rate explained more of the variation in mercury level in the NT lakes than in the NAS lakes (NT, r2 = 0.11, p < 0.001; NAS, r2 = 0.03, p = 0.024). However, the correlation between mercury concentration and growth rate in the NAS study area improved when Reindeer Lake, possibly affected by anthropogenic inputs, was removed from the analyses (r = 0.13, p = 0.001). Therefore, lower mercury levels in lake trout are associated with higher growth rates through growth dilution. The higher mercury concentrations in NT lake trout are due not only to the old age of the fish, but to slower growth rates as well.

The major aim of this thesis project was to determine the variables that most explain the elevated mercury concentrations in lake trout (<i>Salvelinus namaycush</i>), a predatory aquatic fish species in some lakes in northwestern Canada. High mercury concentrations in lake trout in other regions have been associated with the biological features of the fish and various chemical and physical aspects of their aquatic ecosystems. Data including lake trout age, length, weight, and stable isotope values, water chemistry, latitude, and lake and watershed area were collected, compiled and then included in statistical analyses of the factors affecting mercury concentration in the muscle of lake trout from a series of lakes from the Mackenzie River Basin (MRB) in the Northwest Territories (NT), Canada. These results are reported in Chapter 2. Fish age and lake surface area were the most important variables affecting mercury concentrations. However mercury concentration in muscle also was significantly (p < 0.05) related to: fish length, weight, and δ13C; watershed area to lake area ratio; and to total mercury concentration in zooplankton and water. These variables were run through best subsets analyses and multiple regressions in order to determine the regression equation most efficiently capable of predicting mercury concentration in lake trout in unstudied lakes in the MRB region. The resulting equation was: log Hg = 0.698 (0.0156 × latitude) + (0.0031 × age) + (0.000535 × length) (0.245 × log lake area) + (0.00675 × watershed area/lake area ratio), r2 = 0.73<p>Small lakes located in the southern NT and dominated by large and/or old lake trout are most likely to have lake trout whose mean mercury concentrations exceed 0.5 μg/g; the guideline for the commercial sale of fish. Latitude may be linked to mean annual temperature (and variables such as duration of ice cover, summer water temperature) while fish age and length may be related in part to fishing pressures and growth rates on these lake populations. In chapter 3, a more in-depth study was undertaken to investigate of role of feeding and relative tropic level in the bioaccumulation of mercury in lake trout. This was accomplished by comparing MRB lake trout population characteristics with those from a series of lakes in northern Alberta and Saskatchewan (NAS). The two population groups were compared with respect to size, age, growth rates, and mercury concentrations. In addition, trophic and mercury biomagnification relationships, as inferred from stable carbon and nitrogen isotope analyses, for the two lake trout populations were compared. Lake trout from the NT exhibited significantly higher mercury concentrations than those from the NAS lakes (p < 0.001). Mercury concentrations in biota (including lake trout, forage fish, benthic invertebrates and zooplankton) were positively and significantly correlated to δ15N values in all lakes in both of the study areas (p < 0.001). Mercury biomagnification in the NT lakes, as estimated from the slope of δ15N versus mercury concentration, was lower than in the NAS lakes. Thus, mercury biomagnifies more slowly in NT lake trout, but because of their greater mean age, reaches higher values than in NAS lakes. Northwest Territory lake trout generally exhibited more negative δ13C values, indicating more pelagic feeding habits than in NAS lakes: higher mercury concentrations previously have been associated with more pelagic feeding. <p>Finally, the relationship between mercury levels and growth rates in lake trout was investigated by comparing NAS and NT lake trout populations. These results are reported in chapter 4. Lake trout from the NT lakes grew at a slower rate (10.4 mm per year) than those from the NAS lakes (35.1 mm per year). Log mercury concentration was inversely correlated (p < 0.001) with growth rate for both lake trout populations; however, growth rate explained more of the variation in mercury level in the NT lakes than in the NAS lakes (NT, r2 = 0.11, p < 0.001; NAS, r2 = 0.03, p = 0.024). However, the correlation between mercury concentration and growth rate in the NAS study area improved when Reindeer Lake, possibly affected by anthropogenic inputs, was removed from the analyses (r = 0.13, p = 0.001). Therefore, lower mercury levels in lake trout are associated with higher growth rates through growth dilution. The higher mercury concentrations in NT lake trout are due not only to the old age of the fish, but to slower growth rates as well.

Biodiversity is a fundamental part of the functioning of ecosystems and their ability to provide ecosystem services. It has been shown that a high biodiversity increases the robustness of an ecosystem according to the insurance hypothesis. I propose that a similar effect can be seen on a higher scale, where a high diversity of ecosystem types might stabilize the ecological functionality of a region. By comparing eleven network characters in 70 tropical and temperate ecosystems, their diversity was measured as Euclidean distance between the systems in the 11-dimensional room defined by these characters.  The diversity of ecosystems was shown to be significantly higher in tropical latitudes than in temperate. A possible explanation to this result could be that the higher species diversity in the tropics allows for more types of ecosystems. A higher diversity of ecosystems in a region might indicate a larger amount and variation of possible ecosystem goods and services as well as provide the region with an increased robustness. The measurement of ecosystem diversity between regions might also be of importance in a conservation perspective, where unique and vulnerable ecosystems can be discovered and protected.

Trophic interactions within aquatic ecosystems are complex, with many different pathways facilitating transfer of energy and nutrients among trophic levels and many different mechanisms that influence energy and nutrient transfer. This is illustrated in the “top down” and “bottom up” regulatory effects on aquatic food webs, through which primary producer biomass and, therefore, herbivore and carnivore densities, are influenced by both nutrient availability (bottom up) and densities of consumers at higher trophic levels (top down). In an aquatic food web, planktivore presence can directly alter zooplankton density via consumption, while indirectly shaping phytoplankton biomass via reduced herbivore abundance and the release of nutrients due to excretion, egestion, and decomposition. Novel species introduced into an established food web may have important consequences. An invasive species may impact an invaded food web through competition, predation, alteration of nutrient cycling, or, potentially, through facilitation of native species or other invasives. For example, an invasive planktivore may shift zooplankton density or community composition, thereby facilitating phytoplankton blooms. Such a planktivore may also compete with and, potentially, replace native species. Moreover, an invasive species that reaches high densities within its invaded range may serve as an important nutrient sink as it consumes a high biomass of native species or a nutrient source via excretion or decomposition. Two such invasive species with the capacity to dramatically alter native food web dynamics are bighead (Hypophthalmichthys nobilis) and silver carp (H. molitrix; collectively, bigheaded carp). Bigheaded carp are large-bodied, planktivorous fishes that were introduced into the United States in the 1970s and have since spread throughout much of the Mississippi River and its tributaries. These species currently threaten the Great Lakes, where they may constitute a threat to native planktivores such as gizzard shad (Dorosoma cepedianum) and commercially important species such as walleye (Sander vitreus), although there remains a great deal of uncertainty surrounding their potential ecosystem impacts. Consumption of both zooplankton and phytoplankton has been observed in bigheaded carp, although their impact on primary producer biomass is not well understood. Although field observations suggest that condition and abundance of native planktivores, including gizzard shad and bigmouth buffalo (Ictiobus cyprinellus), as well as zooplankton density, have declined following the bigheaded carp invasion, there is little direct, experimental evidence of bigheaded carp food web impacts. Therefore, I sought to examine the effects of bigheaded carp on native ecosystems through a series of mesocosm experiments at the Southern Illinois University pond facility. My primary objectives were to 1) observe potential competition between bigheaded carp and the native gizzard shad, 2) evaluate effects of bigheaded carp predation on zooplankton and phytoplankton communities, 3) assess impacts of bigheaded carp decomposition on nitrogen and phosphorus availability, and 4) measure the rate at which bigheaded carp excrete nitrogen and phosphorus. In order to elucidate the impacts of bigheaded carp on gizzard shad growth and survival, zooplankton and phytoplankton densities, and nitrogen and phosphorus availability in the pelagic and benthic pools and to determine whether gizzard shad experience a diet shift in response to bigheaded carp presence, I performed two mesocosm experiments with three treatments: gizzard shad only, gizzard shad, bigheaded carp, and fishless control (Chapter 1). I predicted that bigheaded carp would reduce zooplankton densities but that gizzard shad, which are both detritivorous and planktivorous, would be unaffected due to their ability to use detritus as an alternative food source. Additionally, both predator release via zooplankton consumption and increased nutrient availability from bigheaded carp excretion would stimulate phytoplankton. I found that gizzard shad survival was reduced by bigheaded carp presence but that surviving gizzard shad did not experience a decline in growth in the bigheaded carp plus gizzard shad treatments. This may have been due to the ability of gizzard shad to consume detritus, as foreguts of sampled gizzard shad in Experiment 2 contained mostly detritus. Moreover, phytoplankton density declined in the presence of silver carp in Experiment 2, suggesting silver carp herbivory. In addition, nitrogen and phosphorus availability in either the pelagic or benthic pools did not appear to be impacted by bigheaded carp presence. After demonstrating experimentally the overall negative impact of bigheaded planktivory on native food webs, I focused my remaining two chapters on the effects of silver carp on nutrient availability. In Chapter 2, I outline a decomposition experiment testing for potential changes in pelagic and benthic nitrogen and phosphorus availability and, in turn, phytoplankton, zooplankton, and macroinvertebrate densities in response to silver carp decomposition. Although silver carp die offs have been reported throughout the Midwest, little is known about the magnitude of those die offs and their consequences for the ecosystem. In this study, silver carp decomposition did not appear to alter nutrient availability or densities of phytoplankton or invertebrates. However, in comparison to northern streams in which salmon spawning and decomposition provide an important nutrient subsidy, the mesocosms used in this study have relatively higher background nutrient concentrations. Thus, silver carp decomposition, at least at the densities studied, may have little importance to in-stream nutrient availability. Lastly, because I am interested in how bigheaded carp, particularly silver carp, alter nutrient dynamics in invaded food webs, it is necessary to calculate silver carp nitrogen and phosphorus excretion rates, as well as body nitrogen and phosphorus content (Chapter 3). Nutrient stoichiometry theory predicts a balance between the relative consumption of nutrients by an organism and the extent to which the organism retains nutrients in its tissues or excretes them. Thus, it is a useful tool in determining how an invasive species may alter nutrient availability via consumption and excretion. In Chapter 3, I describe the body and excretion N:P ratios for silver carp, which exhibit a lower body N:P ratio than excretion N:P, suggesting that these organisms may serve as a sink for phosphorus. Moreover, silver carp body excretion N:P ratios were higher than those reported for gizzard shad, suggesting that, in regions where silver carp may replace gizzard shad or lower gizzard shad population density via competition (Chapter 1), silver carp may alter nutrient cycling processes in aquatic ecosystems by shifting the overall available N:P ratio. Bigheaded carp may pose a significant threat to invaded ecosystems through their potential to compete with native species, reduce plankton densities, and alter nutrient availability. However, although bigheaded carp are expanding in range and approaching the Great Lakes, the full extent of their ecosystem impacts remain uncertain. Through my work on bigheaded carp food web impacts, particularly the influence of silver carp on native species and nutrient cycling processes, I have found that bigheaded carp have the capacity to negatively impact invaded ecosystems overall by reducing zooplankton, phytoplankton, and forage fish densities. Moreover, as bigheaded carp in particular continue to reach high densities as they expand in range, their capacity to alter relative nitrogen and phosphorus availabilities must be monitored to understand the extent of their influence. Due to their ability to disrupt top down and bottom up processes in freshwater ecosystems, bigheaded carp constitute a critical environmental issue in the Great Lakes area and throughout the Midwest and, thus, it is imperative to continue to experimentally assess how bigheaded carp interact with native species to the detriment or benefit of U.S. freshwater communities.

This thesis deals with the role of species traits in predator-prey interactions and food web structure. I conducted laboratory experiments with predatory aquatic insects and their prey to reveal the traits determining who eats whom in small standing waters. I also focused on the possibility of incorporating the observed dependence of predator-prey interactions on body mass into existing food web models. Further, I developed a simple simulation model to explore the consequences of body mass dependent feeding and dispersal for food web assembly. Last, I show that four common methods for sampling aquatic insects differ in their selectivity, especially on the basis of body mass of sampled insects. In conclusion, I combined laboratory experiments, field work and mathematical models to evaluate the importance of body mass and other species traits, such as foraging behaviour and microhabitat selectivity, in predator-prey interactions and explored selected food web level consequences.

The boundaries between freshwater and terrestrial ecosystems can be areas of important subsidy transfers. These subsidies, such as leaf litter inputs to streams or aquatic emerging insects into riparian zones, link food webs and provide benefits to consumers in the form of nutrients and energy. Subsidies from aquatic systems tend to have high levels of essential long chain polyunsaturated fatty acids (LC-PUFAs), such as eicosapentaenoic acid (EPA; 20:5n-3) and docosahexaenoic acid (DHA; 22:6n-3) that are only produced by certain forms of aquatic algae. These LC-PUFAs are highly important in growth, development, and other metabolic functions across animal groups (Brett and Muller-Navarra 1997, Gladyshev et al. 2009).

Anthropogenic global change is altering food web dynamics. Global change comprises factors, like temperature, sound, light and more. In this dissertation, I evaluate how two factors, sound and temperature, alter prairie communities. In Chapter 1, I test if sound influences grasshopper respiration rate, thereby altering diet. Some sound frequencies increased grasshopper respiration rate while others decreased respiration rate. Frequencies that elevated respiration rate led to grasshoppers consuming more carbohydrate-rich foods compared to protein-rich foods. This diet change stems from a carbon deficit due to increased respiration rate and could act as a mechanism by which sound pollution indirectly alters plant communities. In Chapter 2, I test the hypothesis that sound can alter grasshopper movement. In response to sound-induced stress grasshoppers may alter their movement in one of two ways: in situ and displacement. I found no evidence that grasshoppers, nor non-Orthopteran insects alter their movement in the presence of sound. This chapter provides foundational methods to evaluate sound for applications in conservation and management. Further research will improve techniques for grassland or agricultural systems. Temperature is another driver of community change. What is less understood is how warming influences predator-pollinator relationships. In Chapter 3, I ask if warming alters a spider that consumes pollinators in a prairie system. My results indicate that pollinators benefit when spiders are not on the flower. Warming shifts spiders down the plant, thus positively impacting pollinator-plant interactions. In addition, warming may benefit plants two-fold if spiders shift their diet to herbivores. This requires additional research, but it is evident that warming generates a positive indirect effect on plants. These chapters contribute to a growing understanding of how global change is restructuring ecosystems. While global change may alter population dynamics or lead to evolutionary change over longer time scales, behavioral responses happen rapidly and can drive ecological dynamics in the short term. My dissertation demonstrates that sound and temperature alter animal behavior that cascades to lower trophic levels. Thus, in addition to demonstrating the indirect effects of global change, these experiments contribute to growing literature on the importance of top-down control in shaping ecosystems.

Forested streams are closely linked to terrestrial catchments which affects their biogeochemical cycling and carbon inputs. Catchment disturbances alter stream water quality and food webs, including changes in productivity. Such changes in stream conditions can potentially alter consumers’ reliance on autochthonous (in-stream) or allochthonous (terrestrial) sources and mercury bioaccumulation. A recent outbreak of the spruce budworm (SBW) that feeds on spruce and fir trees has provided the unique opportunity to examine stream food web responses across watersheds experiencing a range of defoliation in the Gaspé Peninsula, Québec. This project compares streams in twelve watersheds which were selectively sprayed to control SBW and create a gradient in defoliation. Food web samples (food sources, invertebrates, fish) were analyzed for stable isotopes of carbon and nitrogen in 2019 and 2020 to characterize food web structure, and algal productivity was measured in 2019. Hierarchical partitioning models were used throughout the study to compare the contributions of various local and landscape conditions to stream responses. Models indicated that watershed defoliation contributed to increasing autochthonous production, although some invertebrates were more allochthonous in heavily defoliated watersheds, and brook diets were unaffected by defoliation. Next, food web samples were analyzed for methylmercury (food sources, invertebrates) or total mercury (fish) and trophic magnification slopes were determined for each stream food web. Mercury levels in carnivorous invertebrates and brook trout were driven by dissolved organic carbon (DOC), but not consumer autochthony or watershed defoliation. Additionally, rates of trophic magnification were not related to defoliation severity or DOC. This study found that defoliation contributed to increasing autochthonous production and invertebrate consumer allochthony. However, this disturbance did not increase consumer mercury levels or biomagnification in stream food webs. These findings suggest that intervention to reduce defoliation would mitigate algal responses and dietary shifts, but not mercury cycling as it is influenced by DOC levels in the streams of this region. / Thesis / Master of Science (MSc) / A recent spruce budworm outbreak is causing widespread defoliation of spruce and fir trees, but the impacts to stream environments, including primary production, its consumption, and contaminant levels, are largely unknown. Streams are sensitive to conditions in the surrounding terrestrial environment, as such changes can affect the diets of stream invertebrates and fish and are also linked to increased contaminant levels in aquatic organisms. Specifically, mercury is a metal that is transferred through diet and can reach toxic levels in fish. This study found that defoliation is contributing to increased algal production in streams in the Gaspé Peninsula, Québec. However, some stream invertebrates consumed more terrestrial material in streams that had heavier defoliation. Furthermore, defoliation and algal diets did not increase levels of mercury in aquatic organisms, but this contaminant was affected by increasing concentrations of dissolved organic carbon in the streams. These findings suggest that forest defoliation can alter organisms’ diets but not mercury levels.

Examining how ecological processes are influenced by spatial variation can provide valuable insights into how communities are formed and how they may change in dynamic landscapes. In this thesis I address three objectives surrounding the spatial and temporal variation in species’ recruitment and predation, the influence of habitat isolation on consumer-resource relationships, and the influence of habitat fragmentation on a multi-trophic system. I used marine invertebrates, specifically crustaceans, bivalves, and sessile species as a model system. First, I address the spatial and temporal variation in local and regional processes in a multispecies assemblage of marine sessile invertebrates. Using diverse communities of marine sessile invertebrates as a model system I tested the hypothesis that spatial and temporal variation in recruitment and predation would shape local communities, and that both recruitment and predation would have significant effects on the abundance and structure of adult communities. I found that both recruitment and predation vary through time and space leading to the emergence of regional community divergence. I also address how habitat isolation interacts with top-down and bottom-up processes in seagrass ecosystems. Spatial structure of the habitat may mediate top-down and bottom-up controls of species abundances through decreased habitat connectivity and increased habitat isolation. I manipulated top down and bottom up processes by excluding mesograzers, adding resources, or altering both factors in isolated and contiguous patches of artificial seagrass. I then measured epiphyte recruitment, epiphyte abundances, and macroalgae abundance. I paired this with epiphyte sampling from isolated natural seagrass patches. I found that habitat isolation significantly decreased the abundance of epiphytes settling on seagrass blades due to dispersal limitation for epiphytic invertebrates. I found that consumers had strong effects on epiphyte biomass in continuous habitats, but not isolated habitats. Resource additions increased macroalgae cover and epiphyte biomass only in isolated habitats. The results suggest that isolated habitats may be nutrient limited and that top-down effects are stronger in continuous habitats, while bottom-up effects may dominate in isolated habitats. In my third objective, I address how habitat fragmentation may alter marine food webs. I examined whether predation rates, prey, and predator behavior differed between continuous and fragmented seagrass habitat in a multi-trophic context at two sites in Barnegat Bay, NJ. I hypothesized that blue crab predation rates and foraging would decrease in fragmented seascapes, due to a reduction in adult blue crab densities, increasing survival rates of juvenile blue crabs and hard clams. I expected hard clams to exhibit weaker predator avoidance behavior in fragmented habitats because of decreased predation. I found that species’ responses to fragmentation were different based on trophic level. Clams experienced higher predation and burrowed deeper in continuous habitats at both sites. Densities of blue crabs, the primary predator of hard clams, were higher in continuous habitats at both sites. Predation on juvenile blue crabs was significantly higher in fragmented seagrass at one site. Our results suggest that in fragmented seascapes, the impact of fragmentation on higher trophic level predators may drive predation rates and prey responses across the seascape, which may lead to trophic cascades in fragmented habitats. / Biology

The potential relationship between riparian arachnids and aquatic insect productivity was assessed in forest streams throughout the central South Island of New Zealand. Initially, a survey was conducted of thirty seven, first-third order forest streams. Streams were selected to represent a range of benthic invertebrate standing crops (as a surrogate measure of "productivity") from Banks Peninsula streams with relatively high benthic invertebrate densities to acid mine drainage streams near Reefton that were almost devoid of aquatic life. At each site benthic invertebrate densities and biomass were measured in riffle habitats and adjacent gravel bars were sampled for terrestrial invertebrates. At a sub-set of 16 sites, a 20 metre longitudinal web-building spider survey was conducted along each bank of the stream. As an additional component, a 20 metre transect starting at the stream margin and running perpendicularly into the forest was used to survey the density of web-building spiders with increasing distance from the stream. Results from the survey of in-situ stream insect biomass and gravel bar invertebrates showed a strong relationship between aquatic insect biomass and the biomass of riparian arachnids (R2 = 0.42, P < 0.001) having accounted for potentially confounding factors such as stream size, elevation, substrate and disturbance. The 20 metre longitudinal survey showed that streams with the highest in-situ insect biomass had significantly higher densities of web-building spiders along their banks (R2 = 0.28, P < 0.05), having accounted for potential confounding variables of elevation, habitat architecture and stream and channel width. The stream to forest survey showed a strong exponential decay in web-building spider densities with increasing distance from the stream (R2 = 0.96, P < 0.0001). Regardless of stream productivity web-building spiders were most abundant at the stream margins and rapidly declined to very low densities 20 metres from the stream. In order to further test the relationship between riparian web-building spider densities and stream insect productivity, a stream fertilization experiment was conducted on six first-second order streams in the Maimai experimental catchment, Reefton. Three streams were enriched by the addition of a fertiliser solution mainly consisting of sodium nitrate for seven months, and the other three streams were used as controls. Water chemistry, benthic invertebrate communities, emerging aquatic adults, and the densities of web-building spiders along the stream corridor and in the forest were monitored in three seasons (spring, summer and autumn) over the course of the nutrient-addition. By the end of the experiment, conductivity was significantly higher in nutrient-addition streams than in the control streams (F = 80.5, P < 0.001), but chlorophyll concentrations showed no significant differences between treatments. Both benthic mayfly densities (F = 6.15, P < 0.05) and the biomass of adult aquatic dipterans (Chironomidae, Simuliidae) (F = 9.25, P < 0.01) were significantly higher in nutrient-addition streams in the last sampling round. Spiders recorded from intercept traps indicated that by the end of the experiment spider activity was significantly higher within 2.5 metres of the nutrient-addition streams (F = 5.70, P < 0.01). However, seasonal densities of web-building spiders along the stream margin and in the forest decreased with no significant differences observed between nutrient-addition and control streams. The results from these studies indicate that adult insects emerging from streams represent an important source of prey that could influence the biomass and abundance of riparian arachnids. Additionally, the results imply that stream productivity and size could mediate the strength of the interaction between riparian and stream habitats. Moreover, feedback mechanisms present in both systems could have implications for such interactions. The elevated densities of web-building spiders observed at the stream margin led to the proposal of the "Highway Robber" hypothesis. This hypothesis suggests that such higher densities of spiders are the result of increased insect activity along the stream corridor: the emergence of adult aquatic insects was predicted to vary less over temporal and spatial scales than that of terrestrial insects due to the poorly synchronized life histories in many New Zealand stream insects. I conclude by suggesting that there are numerous anthropocentric perturbations such as loss of heterogeneity, introduced species, pollution and habitat degradation that could undermine and decouple the intimate linkages between aquatic and terrestrial ecosystems.