|
Atlantic
Cooperative Wildlife
Ecology Research
Network
|
| Canadian Wildlife Service | Acadia University | Memorial University of Newfoundland | University of New Brunswick |
| Abstracts | Publications | Students | Research Associates | About Us | Contact Info | News | Listserve |
|
IRVING ENVIRONMENTAL CENTRE, WOLFVILLE, NS (ACADIA) 2003 Determinants of herring gull
foraging activity and success within an Arctic common eider duck colony In many situations, it remains uncertain whether avian predation has a significant impact on the reproductive success of colonial nesting waterfowl. We studied the foraging ecology of herring gulls breeding within a common eider duck colony in the Canadian Arctic between 1997 and 2002. Specifically, we quantified how gull search activity and success varied with time of day, weather, tides, eider nesting density and reproductive phenology. We hypothesized variability in foraging activity would be determined by factors that influence the relative costs and risks of foraging within the colony. We found that gulls were most active in areas of high eider nest densities, during high tide, and under windy conditions >15 km/h. Herring gulls did not force incubating hens off their nest, but instead, only took eggs from unattended nests. Consequently, herring gulls were most successful during egg-laying when hen nest attendance was sporadic. Approximately 75% of eggs taken were removed singly by individual gulls. We estimate that between 30% and 40% of eggs laid by eiders were depredated by gulls. Feeding movements of Great
Black-backed (Larus marinus) and Herring (L.
argentatus ) Gulls breeding in the Bay of Fundy In the northwestern part of the Bay of Fundy, during the 2001 and 2002 breeding seasons for Great Black-backed (Larus marinus ) and Herring (L. argentatus ) Gulls, colour- and shape-coded markers were distributed in pieces of bread and fish bait at sites known to have concentrations of feeding gulls. These feeding sites included anthropogenic food sources, such as landfill sites, a fish plant, an aquaculture dump site and lobster fishing boats, and natural food sources, such as a tidal flat and a set of falls during a fish run. Islands with gull breeding colonies were later searched for any markers that may have been consumed by the gulls at the distribution sites and regurgitated back on their colonies. In this way, we gained evidence of the distances gulls will travel from their colonies for food, and from which colonies gulls travel to use specific food sources. Some preliminary results from these data will be presented. Foraging Behavior of Arctic and
Common Terns on Machias Seal Island Machias Seal Island is designated a Migratory Bird Sanctuary, and is located at the mouth of the Bay of Fundy . It is one of the largest tern colonies in North America . We will attempt to investigate the feeding habits of Arctic Terns ( Sterna paradisaea) and Common Terns (S. hirundo ) on their breeding grounds to see if there are differences between the two species and between two stages of their breeding cycle, the incubation and chick-rearing periods. We will also compare their strategies to the foraging strategies of other seabirds. Given the different constraints of nestlings, it is possible that they change their foraging patterns to accommodate nest responsibilities. Based on studies of Common Terns is expected that they ( Arctic and Common Terns) will feed fairly close to the island, but this is not assumed. We also hypothesize that the distance that they will travel to feed on fish and marine invertebrates will decrease as responsibilities at the nest increase.To test this idea we will be attaching radio tags to both Common and Arctic Terns during the second week of incubation and following their movements via a receiver and an antenna until fledging. Challenges in Estimating Seabird
Demographic Parameters: Insights and Uncertainty There are many challenges that arise when we set out to estimate demographic parameters of seabirds. Based on our studies using marked Arctic Terns and Atlantic Puffins to estimate survival and emigration rates, we will identify some of the challenges we experienced and how they might be dealt with in future studies. Our presentation will be divided into two parts addressing challenges in the field and at the analysis stage. Careful a priori planning in studies of marked birds should include model simulations to determine optimal sample size and encounter rates (resight/recapture) and selection of models relevant to hypotheses one wishes to test. However, no level of preparation can avoid all complications that can be encountered in the field. Logistics, environmental conditions and the birds themselves are certain to affect the quantity and quality of our data sets. At the analysis stage, what was not avoided by careful planning and unforeseen complications can present challenges in parameter estimation and model selection. When sample size and/or encounter rates are low, certainty provided by estimates of precision (SE and 95% CI) will decline. Heterogeneity caused by band degradation and temporary emigration result in a bias of some unknown quantity in our estimates. Another source of concern in estimates is model bias; given enough data, all models will produce estimates of parameters (e.g. survival) but if the model is not consistent with the biology of the animal, rates will be, again immeasurably, biased. In spite of these challenges, we can still gain important insights into the demographics of seabird populations using data from marked birds. Bicknell's Thrush: population status
and habitat of a pivotal species Bicknell's Thrush is one of the least known of our landbird species. The only extant bird species endemic to northeastern North America , it was originally considered a sub-species of the Gray-cheeked Thrush. It became a full species in 1995. Due to its rarety and affinity for remote and inhospitable habitats, Bicknell's Thrush has been little studied. Listing by COSEWIC in 1999 as a Species of Special Concern has elevated the profile of the species and research interest in it. Recent designation by Partners in Flight as a species of continnental priority is sure to draw further interest.This presentation will focus on the Canadian Wildlife Service work between 1995 and 2001 to survey the Maritimes for Bicknell's Thrush and will report on important habitat information, population status and conservation efforts throughout its range. Pre- and post-hatching sex
allocation biases in the Common Murre (Uria aalge)
on Great Island, NL. Parents can choose to expend more effort in rearing either sons or daughters, according to the relative costs and benefits of offspring of each sex. This selective allocation may take the form of differential production of male and female offspring, or can occur after hatching, by varying feeding effort during the period of parental dependence. This study investigates sex allocation in the Common Murre, an alcid seabird with slight sexual size dimorphism. Males are slightly larger in adulthood, and so are thought to be more costly to rear. Chick feeding rates were obtained for 30-35 marked breeding pairs on Great Island , NL from 1997 to 2001. Chicks were sexed using DNA extracted from feather samples. Feeding rates to chicks of each sex were compared for thirteen pairs producing both male and female offspring in the course of the study. Both mothers and fathers fed sons more frequently than daughters in the latter part of the chick feeding period at the nest site, evidence that males are in fact more costly to produce. Chick feeding rates varied from year to year, likely reflecting variation in foraging conditions at the colony. In years in which chicks were fed less frequently overall, there was a significantly greater proportion of the "cheaper" sex, females, produced at the study site. The use stable isotope analysis to
infer diets at different temporal scales. Stable isotopes are different stable forms of various elements found in nature, which differ in nuclear mass. Different natural fractionation processes cause differences in ratios of stable isotopes, such as 12C and 13C and 14N and 15N ratios (Hobson 1999). These stable isotopes will reveal information about consumed prey and foraging habits. Comparing these ratios in predator tissues to ratios in collected prey tissues allows us to determine which prey species are incorporated in the predators' diets. Stable Isotope Analysis (S.I.A) of carbon allows us to trace the origins of terrestrial or marine feeding sources and S.I.A. of nitrogen indicates trophic level. S.I.A allows researchers to infer diet over different time scales. Stable isotope analysis on Common Tern (Sterna hirundo) and Arctic Tern (Sterna paradisaea) diets will be used to discuss the use of this tool on various tissues; eggs, breast feathers and blood. Eggs reflect diet of the female during egg production, winter diet is inferred from feathers grown on the wintering ground, and blood allows inference about diet within the last week of consumption of prey. The application of deuterium and sulfur to inferring diets and other advantages of S.I.A to ecological studies will be reviewed. Attracting Roseate Terns (Sterna
dougallii) to Machias Seal Island, NB using social
stimulants. Roseate Terns are found on two colonies in Atlantic Canada, a population estimated between 123-149 pairs (CWS 2002). This renders the Canadian Rose ate Tern population vulnerable to extirpation through natural catastrophes such as predation, disease and pollution. Increasing the number of colonies will decrease this threat. We thought that Machias Seal Island be an ideal candidate for colony expansion because it has a history of a small number of birds breeding on the island throughout the years ( Devlin et. al. 2002). Furthermore, a population of 3551 pairs of Arctic and Common Terns ( Boyne et. al. 2002) nest on Machias Seal Island and offers a protective umbrella favoured by Rose ate Terns. Finally, there is an adequate amount of nesting habitat making it an ideal site for a Restoration project. In order to attract Rose ate terns we used social stimulants including an outdoor sound-system playing colony sounds during daylight hours, triangular mirrors set-ups, and Rose ate Tern decoys on two sites. Remote cameras were placed to record activity during daylight hours. The cameras allow the monitoring of the sites with minimal disturbance and will provide a better understanding of their behaviour as they are establishing a colony. A small number of Rose ate Terns were observed for 19 days between May 10th and August 17th but never established. This project will be continued at least for the next breeding season, hopefully providing insights on minimal requirements for the establishment of new Rose ate Tern colonies. Effects of landscape structure on Leucorrhinia
hudsonica movement in western Newfoundland What's been going on in the GGME for
the past few years Sweating the small stuff: a guide to
preparing and critiquing articles and deciding authorship In practice, working biologists spend a large proportion of their time preparing or reviewing documents. Most of us acquire what skills we have for this through experience, trial-and-error, or intellectual osmosis, often receiving little help, guidance or formal training in the process. Here I address three components of this activity: preparing a document for efficient review; critiquing drafts; and the thorny issue of co-authorship. I refer to our practices in the UNB lab of ACWERN and to published sources which I have found useful. How Changes in the Canadian Wildlife
Service May Affect ACWERN The passage of the Species at Risk Act has brought new obligations to CWS, with the need to hire new scientific and administrative staff in the species at risk program. Although we must address new bureaocratic requiremnents, we will be looking at key scientific questions relating to critical havbbitats, setting population objectives, etc. with partners such as ACWERN. There will be an increasing emphasis on managing bird populations towards numerical targets, assessing the impacts of climate change on birds and their habitats, understanding wildlife diseases such as West Nile Virus, and working with resource industries such as forestry on approaches to minimise impacts on bird populations. We hope that these initiatives will create joint research projects between ACWERN and CWS. Sources of variation in
orientation behaviour of Yellow-rumped Warblers (Dendroica
coronata) Orientation performance (i.e. direction finding) of migratory birds generates the migratory patterns that we observe over broad geographic areas. For instance, the predominant migratory direction of Atlantic region landbirds is southwest. However, differences are detected when these patterns are monitored at local scales. Migratory orientation is influenced by several factors such as: bird age, weather, genetic disposition, and availability of sources of compass information (e.g. celestial and magnetic cues). However, little is known about the influence of breeding origin on these migratory patterns, especially those observed at local scales. To assess the importance of breeding origin and explore other sources of variation on directional preferences of migratory birds, we conducted orientation registration experiments with Yellow-rumped Warblers (Dendroica coronata) captured at a stopover site during fall migration in southwest Nova Scotia. Previous work at this site has shown that birds originate from a very broad geographic area. Circular regression was used to model variation in directional preferences with the variables of interest: breeding origin (determined by deuterium levels), experimental treatment (horizon visible or shielded), amount of fat reserve (low, moderate, high), wind condition at time of trial (favourable or unfavourable) and day the trial was conducted. The final best-fit model indicates that experimental treatment, fat reserves, breeding origin and wind are important sources of variation of orientation behaviour. Birds from local regions orientated more southerly and individuals from northern latitudes oriented more westerly. Birds with greater fuel reserves, and those under experimental conditions with an extended view of the horizon, chose directions that were more concentrated in the expected migratory direction (southwest). The directions that birds chose under favourable wind conditions were more concentrated. Overall, our models suggest the importance of several variables, particularly breeding origin, to migratory patterns in the Atlantic region. Behaviour and productivity of
Harlequin Ducks breeding in Central Labrador : a test of the food
limitation hypothesis. I tested the long held paradigm that productivity of Harlequin Ducks breeding along fast moving rivers and streams is limited by variation in available epibenthic insect foods. Productivity of Harlequin Ducks breeding in central Labrador varied annually with 33.3 %, 27.3 %, 20.0% and 25.0 % of females successfully producing broods in 1999, 2000, 2001 and 2002, respectively. Relatively low proportions of time were spent feeding by females (mean: 0.369 of the ~17h of daylight, range: 0.354 + 0.016 SE – 0.389 + 0.025 SE), and proportions of time spent feeding in each year were not negatively related to productivity as predicted from the food-limitation hypothesis. Taken together, behaviour resulting in pair breakdown and reduced detections of females, physical evidence of egg passage through the cloaca, radio-telemetry of females, and lack of annual variation in body condition supported that all adult female Harlequin Ducks on the study area attempted nesting each year. There was no support for the behaviourally based paradigm that some females were constrained by lack of sufficient food on their breeding habitat and deferred breeding, i.e., facultative non-breeding. An environmentally based factor affecting productivity was supported as productivity increased when small rodent populations were higher. These data were consistent with a pure reciprocal model of population limitations where cyclic herbivore populations (hares and/or small rodents) are keystone species in north boreal ecosystems. Conserving biodiversity of remote
islands in Atlantic Canada , an urgent concern Small oceanic islands are the most threatened of the world's natural ecosystems, because they support animal and plant species found no where else, and are so sensitive to human development. This problem is most acute in tropical and sub-tropical regions, but is also an important issue related to conserving Canada 's biodiversity. Three isolated islands (Sable, Seal and St. Paul 's Islands ) off the coast of Nova Scotia are Canadian examples of small isolated islands with endemic plant, vertebrate and invertebrate biodiversity. Like many of the world's islands, these islands have been degraded by introduced mammalian predators and herbivores, loss of wetlands, toxic chemical contamination and uncontrolled development. Sable island was drastically modified during the 19th century by humans, domestic animals and rats, remains in a degraded state, yet has endemic plants and invertebrates (some extinct or threatened), endangered bird species and the potential for recolonizing seabird species. Seal Island had the largest seabird colony in Nova Scotia until Norway rats and other alien species were introduced during the 19th century, remains in a degraded state, yet has an endemic mammal and the potential for recolonizing seabird and endangered bird species. St. Paul 's Island is a remote, apparently degraded, island between Nova Scotia and Newfoundland with geography suitable for large seabird colonies, and a population of a threatened passerine. Sable, Seal and St. Paul's Islands all contain Federal lands, habitat for migratory bird species at risk, provide excellent research opportunities and urgently require restoration plans and removal of introduced mammals if their precious biodiversity is to be conserved - an opportunity for leadership from ACWERN. State-space models for ecologists:
getting more out of movement and population data The study of animal movement and behavior is being revolutionized by technology, such as satellite tags and harmonic radar, that allows us to track the movements of individual animals. However, our ability to analyze and model such data has lagged behind the sophisticated collection methods. We review problems with current methods and suggest a more powerful and flexible approach, state-space modelling, and we illustrate how these models can be posed in a meta-analytic framework so that information from individual trajectories may be combined optimally. State-space models enable us to deal with the complexity of modelling animals interacting with their environment but, unlike other methods, they allow simultaneous estimation of measurement error and process noise that are inherent in animal trajectory data. A Bayesian framework allows us to incorporate important prior information when available and also allows meta-analytic techniques to be incorporated in a straightforward fashion. Meta-analysis enables both individual and broader-level inference from observations of multiple individual pathways. Our approach is powerful because it allows researchers to test hypotheses regarding animal movement, to connect theoretical models to data, and to use modern likelihood-based estimation techniques, all under a single statistical framework. We illustrate the approach with examples using ARGOS satellite data collected from marine vertebrates such as Leatherback turtles ( Dermochelys coriacea) and radio transmitter data collected from terrestrial vertebrates such as fishers (Martes pennanti). Application to population time-series data will also be discussed. Demography of the Razorbill (Alca
torda) at the Gannet Islands , Labrador Recent advances have been made in ecological modeling in long-lived species, such as seabirds. Razorbills are a long-lived seabird, which respond to a variety of environmental perturbations. As a result, an ecological model for this species is needed. Although Razorbills (Alca torda) are a relatively well-studied species in Europe , many areas of Razorbill demography in North America remain unknown. Therefore, the goal of this project is to measure Razorbill demographic parameters to determine what factors ultimately regulate Razorbill populations. To do this, I will address 3 main objectives: (1) to quantify and examine Razorbill demographic parameters including annual survival, age of first breeding, natal philopatry, and dispersal distance, (2) to quantify and examine the effects of intraspecific kleptoparasitism on Razorbill hatchling success, and (3) to develop a model that can be used to examine and quantify the impact of hunting and arctic fox (Alopex lagopus ) predation on Razorbills. Using radio telemetry to determine
the effects of landscape on juvenile songbird movements Not much is known about the gap-crossing behaviour of forest-dependent songbirds as they move throughout fragmented landscapes. My research will be conducted in the Greater Gros Morne Ecosystem of Newfoundland, where current forestry practices are resulting in the fragmentation of an old-growth spruce-fir forest. I will use radio telemetry to characterize gap-crossing tendencies of male and juvenile songbirds during the post-fledging period as they encounter habitat edges when moving between natural, clear-cut, and selective-cut landscapes. Target species inc lude the Blackpoll Warbler, Hermit Thrush, Northern Waterthrush, and Yellow-rumped (Myrtle) Warbler. Results of this study will be incorporated into a model which may predict songbird movement across the landscape. This model will provide forest managers with pertinent information when developing plans for future harvesting operations so that connectivity across the harvested landscape will be maintained. Communication Tower Kills: A source
of data? To improve communication systems in the region of Fundy National Park, NB, a new 61m cell tower was erected in the fall of 2003, on the same location in the Park as a shorter tower, but with regulations which involve documenting the effect of the new tower on birds, particularly during migration. A 2 m high fence was erected in an attempt to prevent mamalian predators from removing bird bodies before the duty warden can retrieve them. A protocol for retrieval and collection was set up. To my knowledge only a few bodies have been collected thus far. In researching this topic, three areas have interested me: how many birds are actually being killed by towers of all sizes (this includes when, where, why and how); how can researchers make use of the bodies for studies ranging from physiology to the timing of migration and migration pathways; and how can the information be used to develop national policies with the aim of reducing tower kills? Feeding data of four species of
Atlantic seabirds Feeding data for breeding Arctic and Common Terns, Razorbills, and Atlantic Puffins have been collected on Machias Seal Island , New Brunswick since 1995. The data are collected in feeding watches, where the time of prey delivery, location, prey provider (if possible to differentiate), recipient young, prey species, and prey size in relation to the bird's culmen are recorded. The data are compiled in the MSI annual progress report and fragments of the data have been included in published papers and in graduate theses, but much of the data remains unexamined and largely uninterpreted. I am interested in carrying out a more detailed analysis of the information we have. One possibility is to examine within season and between season variation of prey species and prey species size. I also would like to compare the prey species composition in the diets of the four species of seabird, investigating feeding overlap with particular attention to how the different seabirds respond to changes in prey availability due to their different foraging strategies. I hope that I can use the ACWERN forum to discuss these ideas further and to devise other ideas for analysis of the data we have. The distribution, transmission and
ecology of West Nile Virus in migratory and resident passerines in
Eastern Canada The West Nile Virus (WNV) is a mosquito-borne virus of the genus Flavivirus transmitted primarily by mosquitoes of the genus Culex. The transmission cycle typically involves a wide range of birds serving as amplifying hosts and has been documented in 111 bird species in North America . Large-scale fatalities have been detected in Corvids although deaths in other passerines or levels of viremia in surviving individuals remain largely unknown. Humans are susceptible to infection and may have a range of symptoms ranging from ordinary headaches and fever to neurological conditions sometimes leading to death. The transmission of WNV over long distances by migratory birds is supposedly common but has not been demonstrated in North America . This project will attempt to generate data on West Nile activity through serological techniques. Ten passerine species have been selected for the assessment of WNV during fall and spring migration. Blood samples will be collected from 40-60 individuals of each species from 4 bird banding stations (1 in MB, 2 in ON, and 1 in NS). Exposure to WNV will be detected by the VecTest or by ELISA techniques, and samples testing positive will be further analyzed using Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) to ascertain the level of viremia. All the sampling will take place between May and June (Spring Migration); and between July and September (Fall Migration). Blood from selected resident species will also be sampled. On each sampling station, CO2-baited light traps will be used to collect mosquitoes from April through to November. Mosquito pools will be analyzed for viremia by using VecTest. The overall goal of the project will be to determine the prevalence, transmission patterns and ecology of West Nile virus in passerines. Fundy Coastal case Study: The
Pollution Pathway of Mercury in Herring gulls and Great black-backed
Gulls from the Bay of Fundy. Seabirds are high trophic-level organisms in the Bay of Fundy, and are therefore of special interest in studying the ecological pathways of Hg within the coastal environment. This poster provides ecological pathway details that are pertinent to two sympatric species, namely [Larus argentatus (herring gulls) and Larus marinus (great black-backed gulls)]. During the summer these birds nest in high concentrations on small and uninhabited islands. Here, atmospheric inputs of Hg are higher than on shore and further inland due to direct intrusion of frequent fog banks. The greatest source of Hg on these islands, however, would be provided by the birds themselves, in the form of guano that quickly builds up around each nest location. In summer 2003 blood, feathers, and diet samples were collected from 20 adults of each species and their chicks. Eggs were also collected from 10 nests of each species. It was found that herring gulls had higher mercury concentrations in feathers and blood than great black-backed gulls. The mercury pathway from food to adult gull to egg to chick and back into the island ecosystem was traced in terms of total mercury concentrations. Overall, there was evidence of Hg build-up in the Hospital Island soil compared to the control. This suggests gulls do raise Hg levels on their colony but this process is too slow to detect within one year. To test this would require sampling of more colonies and controls in one year. What's going on at the ABO Songbird movement in natural and
harvested landscapes Very little is known about songbird movement patterns during the breeding season or the effects of different harvesting regimes on these movements. We examined how landscape structure (composition and configuration) influences patterns of bird movement at both the local (within-territory, non-dispersal movements, associated with raising young) and the landscape (between territory movements, juvenile dispersal) level by mist-netting. Four 2 km x 2 km landscapes (2 natural, 2 clearcut) in western Newfoundland were netted from June to August 2003 in order to quantify bird movement. A total of 649 birds were captured - 333 birds in clearcut landscapes and 301 birds in natural landscapes. 129 of these birds were recaptured. Habitat around each mist net was classified in the field and habitat within each landscape will be determined using GIS data. We also investigated how different harvesting strategies influence songbird density and composition by surveying 3 landscapes (variable retention cut, control, and potential cut) using point counts. Ten points were surveyed three times in each of these areas in late June and early July. Habitat around each of the points was classified in the field and habitat in each landscape will also be determined using GIS data. Data are currently being analyzed and results from this study will help forest managers to make better decisions on how to harvest forests to minimize negative impacts on species such as songbirds. The extent of semipalmated
sandpiper, Calidris pusilla , movement during
their fall migration stopover in the upper Bay of Fundy , and factors
affecting this movement. The upper Bay of Fundy is a critical migratory stopover point for Semi-palmated Sandpipers, with an estimated 1 to 2 million birds annually visiting local mudflats in late summer to feed on Corophium volutator, the most abundant macroinvertebrate on the flats. Little is known about how these birds select foraging habitat and if they stay on single mudflats, or move within the region, during their approximate 2-week stay. Such knowledge is important from a conservation standpoint, in that the ability to use multiple mudflats should reflect birds' capacity to adapt to human induced changes to mudflats, such as the damming of tidal rivers and the subsequent removal of these barriers. We hypothesize that the main factors that could potentially influence sandpiper movements and habitat use are: 1) abundance and size distribution of C. volutator , 2) proximity of roost sites to mudflat foraging areas, and 3) predation threats by peregrine falcons and merlins. To assess movements and test the importance of these factors, in late summer of 2004, a number of sandpipers from both Shepody Bay and the Minas Basin will be fitted with radio transmitters. Their movements will be monitored during their two-week migration stopover in the area. At the same time, prey abundance and size distribution will be quantified by sampling C. volutator and counting mud snails (an indicator of amphipod abundance) on mudflats in the region. Further, predation threats will be assessed through behavioural observations at the same mudflats, and bird movements will be compared between areas where single roosting sites are near several mudflats and those where mudflats are more separated. Between-patch movements and survival
of a boreal peatland dragonfly (Leucorrhinia hudsonica
): a better method for analysing mark-recapture data Obtaining habitat-specific movement and survival rates is important for assessing the impacts of landscape change on insect populations. Such information is lacking for most Odonates, despite rising interest in their conservation. We present a rigorous modelling technique that provides maximum likelihood estimates for daily movement and survival probabilities from multi-site, mark-recapture data sets. Movement rates of adult Leucorrhinia hudsonica were high, but different methods of calculation produced varied results. Food web structure and the dynamics
of pitcher plant microfauna Various consumer-resource models predict that bottom-up and top-down factors will influence the stability of populations, but the expected direction of the effect varies, depending on the characteristics of the community. We tested for effects of bottom-up and top-down factors on the stability (persistence and temporal variability) of microfaunal populations living in pitcher plants by manipulating resource and predator density in a natural community open to colonization. Top-down (omnivorous) effects were destabilizing, decreasing the persistence time of a rotifer,H. rosa, and perhaps a microflagellate, Bodo. Bottom-up effects were more variable. Higher resource availability (more bacteria due to shredding by midges) increased the persistence of Bodo. Bottom-up effects on H. rosa populations, however, depended on the predator regime. Increasing resources by adding ants decreased persistence when predators were rare, but increased persistence where predators had colonized. Thus, differences in food web structure, caused by variation in colonization and capture rate, leads to variation in the strengths of top-down and bottom-up forces, and this variation then results in different levels of persistence for local microfaunal populations within pitcher plant leaves. |