Ecologists are under pressure to scale up their science to deal with larger areas and longer spans of time.  At the same time, ecologists and their constituency demand rigor and mechanistic understanding.  The goals of large scale ecology are a natural tension point because of the inherent difficulties attendant to linking large scale patterns with underlying mechanisms.  My research is focused on this interface.

Most of my work has centered on pond-breeding amphibians.  These organisms are an excellent case study in the difficulties and promise confronted by ecologists today.  By the late 1980’s amphibians were held up as a model system in ecology.  As a result of more than two decades of steady effort, some of the most rigorous, complex and informative experiments in community ecology had been completed using larval amphibians.

Nevertheless, in 1990 when reports of disappearing populations became widespread, amphibian ecologists were left entirely flatfooted.  We had very little concrete to say about the situation.  More than a decade later, things are only marginally better.  This has happened, in large part, because like most of their colleagues, amphibian ecologists have a generally poor knowledge of the patterns and mechanisms of large scale distributions. I have been working toward filling this gap via a combination of experimentation and observation.  Below, I briefly sketch five areas of research:


    The textbook description of amphibian community structure has been based on keystone predation.  This concept relies on some sort of gradient over which the intensity of predation varies, encouraging reduction or elimination of dominant competitors in some locations and allowing dominant competitors to exclude inferior competitors in others.

    Our long-term research in Michigan (since 1988) and Connecticut (since 1996), largely conflicts with this view[abstract].  Instead of assemblages of strong interactors, it appears that larval amphibians in natural populations typically do not compete intensely and there is little evidence that predators result in exclusion from nonpermanent wetlands.  In the language of metacommunity theory, it appears much more like these amphibians may exist in layers of single species metapopulations.

    Nevertheless, amphibian distributions are strongly patterned.  Where do these patterns come from?  The permanence of a wetland and its light and thermal environment appear to be key variables.  Repeated cohort loss from drying can lead to local extinction of a species.  The time constraint imposed by drying can be enhanced when vegetation grows up around wetlands.  Shading leads to lower temperatures and slower developmental rates.  Vegetation also evapotranspires.  In spite of any contrary intuitions, a given wetland heavily shaded by forest vegetation will dry faster than the same which have little vegetation near the basin.  The synergistic action of these factors can cause populations to go extinct.  Similarly, it appears that disturbances that remove vegetation and alter hydrology can promote population establishment.  The likelihood of establishment can be related to the proximity of source populations and the structure of the intervening terrestrial environment.

    This research is important because, as with rocky intertidal systems, further research has shown that “post-settlement” mechanisms initially used to describe community structure now need to be revised and expanded.  For amphibians, linkages among breeding ponds and the dynamics of the terrestrial ecosystems that surround breeding ponds have demonstrable impacts on amphibians that may often trump the impacts of species interactions initially used as a basis for understanding amphibian communities.


    The ecology of infectious disease is poorly understood.  Parasites and pathogens are undoubtedly more ubiquitous and more important to ecological patterns than is currently appreciated.  Our laboratory has focused on two lines of disease related research.  In the first, we are working to understand how human development of landscapes affects the rates and consequences of infection by macroparasites.  Most of this work has focused on echinostomes, a trematode parasite with a complex life cycle that encysts in the kidneys of amphibians, appears to emerge in developed contexts and is associated with high rates of mortality in a field experiment.

    One related goal in this line of research is to develop new tools that can be used to diagnose infection noninvasively in preserved and live organisms.  A grant from the National Science Foundation is allowing us to develop the use of high resolution ultrasound technology to this end [abstract]. A preliminary goal is to use ultrasound to evaluate historical infection patterns captured in museum specimens.  We hope to contribute understanding to longstanding questions about the role of human development in the frequency and intensity of infection and disease.

    A second line of disease research is focused on limb deformities in amphibians.  The leading suspected cause of deformities involves infection by Ribeiroia ondatrae, a trematode.  Exhaustive sampling in Vermont has failed to turn up any evidence of Ribeiroia demonstrating that high frequencies of deformities can occur over a large area in the absence of infection by Ribeiroia ondatrae.  These findings leave open the issue of what is causing deformities in Vermont and other regions where Ribeiroia is absent.  One possibility, exposure to chemical pollutants, is supported by an association between the risk of deformities and the proximity to agricultural sites.  Continuing work in collaboration with Gunter Wagner’s group in Yale’s Ecology & Evolutionary Biology Department will focus on identifying potential chemical agents of deformities and evaluating the mechanisms by which they cause abnormal development.


    Ecology is being transformed by the recognition that ecological and evolutionary timescales are not easily differentiated.  A 1999 review of evolutionary rates by Andrew Hendry and Mike Kinnison (The pace of modern life: measuring rates of contemporary microevolution.Evolution 53:1637-1653) provided the striking conclusion that rates of contemporary evolution are much faster than generally appreciated.  The fact that genotype frequencies can be expected to change as the environment changes is profoundly important to general ecological understanding as well as our approach to dealing with issues such as global climate change.

    Most of our research on rapid evolution has focused on the response of wood frogs to changing canopy gradients.  Our work reveals that a number of traits including critical thermal maximum, embryonic development rate, and thermal preference behavior all show variation consistent with local adaptation that occurs on the scale of decades and tens of meters.  These findings offer a startlingly different picture of interactions between organisms and their environment prompting us to rethink, in larger sense, how we should conceive of ecological assemblages.


    Ecology is evolving from an exploration of what can happen to a science focused on what does, or will happen.  As it has changed there has been a rising premium for developing the role of prediction in ecology.

    Distributional patterns are at the core of ecological science and are central to my research.  Despite a tradition of study extending over 100 years, we still have relatively rudimentary capabilities in predicting the pattern of a species’ distribution in space and how that distribution will change over time.  Improved predictive ability is important on at least two fronts.  First, as critics of ecology have pointed out, prediction is the acid test of our science.  If we cannot predict distributional patterns then ecologists have little right to claim that they understand how ecological systems work.  Second, prediction is a critical task for ecologists involved in conservation.  Many biologists contributing to conservation efforts are pressed to predict what will happen to species or system X if perturbation Y is allowed to happen.

    Our efforts are focused on developing more effective predictions that are based on the types of data that are most likely to be available.  Consequently, for some of this work, we have focused on modeling patterns of presence and absence from spatially distributed surveys.  Highlights from our results include the development of rule based models to evaluate competing hypotheses underlying distributional change, and the discovery that low cost, widely available climate data can be equal in predictive ability compared with much more expensive GIS based vegetation models (e.g., GAP Analysis).

  • Amphibian declines and developmental abnormalities:  It has been over a decade since widespread declines and extinctions of amphibians were first reported.  Nevertheless, ecologists still have little concrete idea why amphibian populations have declined and, in many cases, gone extinct.  We are using long term observational data in conjunction with field experiments to unravel the causes for amphibian population extinctions.

    We are also interested in the mechanisms leading to outbreaks of amphibian deformities.  Ongoing work in Vermont is aimed at understanding the causes for limb abnormalities in leopard frogs in the Lake Champlain basin. The leading suspected cause of deformities involves infection by Ribeiroia ondatrae, a trematode.  Exhaustive sampling in Vermont has failed to turn up any evidence of Ribeiroia demonstrating that high frequencies of deformities can occur over a large area in the absence of infection by Ribeiroia ondatrae.  These findings leave open the issue of what is causing deformities in Vermont and other regions where Ribeiroia is absent.  One possibility, exposure to chemical pollutants, is supported by an association between the risk of deformities and the proximity to agricultural sites.

    Interactions between human and non-human components of watersheds:  This project is a collaborative effort among several faculty at the School of Forestry & Environmental Studies.  We have been working to document the associations between patterns of abundance and distribution in a number of non-human taxa, physicochemical conditions, and patterns of distribution, behavior, and attitudes among human residents.  The study has been conducted in 18 watersheds of ca. 1000 ha spread across the greater New Haven region.  We have found that human attitudes, beliefs, and behavior are strongly linked to landscape structure, which in turn, is a strong predictor of the biological condition of stream watersheds.

    Habitat Conservation Planning:  Graduate Students from the School of Forestry & Environmental Studies participated in a working group on Habitat Conservation Plans (HCPs) at the National Center for Ecological Analysis and Synthesis in Santa Barbara.  Our task was to complete a comprehensive evaluation of the content and quality of HCPs developed to protect endangered species on non-Federal land.  The analysis involved faculty and students from 8 different Universities around the U.S.  We reviewed most of the 200-odd plans in existence at the time the project began and looked at 43 of those plans in much greater depth.  Among the more critical findings was the determination that use of natural history information varies greatly and that relevant ecological theory has been applied sporadically.  We concluded our report with a set of recommendations to improve the development and implementation of HCPs that have, in large measure, been adopted by the U. S. Fish and Wildlife Service.