Conclusions: A need for a broader range ofinformation on effects of contaminants on individuals exists among the 4 classes of terrestrial vertebrates, especially mammals, reptiles, and amphibians. Separation of contaminant effects from other effects and reduction of speculative extrapolation within and among species requires information that can be produced only by combined field and laboratory investigations that incorporate seasonal or annual cycles and important spatial and interaction conditions. Assessments of contaminant effects at the population level and higher are frequently dependent on extrapolations from a lower organizational level. Actual measurements of the effects of contaminants on populations or communities, possibly in conjunction with case studies that establish relations between effects on individuals and effects on populations, are needed to reduce the uncertainty associated with these extrapolations. Associated with these assessment levels is the need for acceptable definitions of what we mean when we refer to a 'meaningful population change' or an 'effect on communities or ecosystems.' At these higher levels of organization we are also confronted with the need for procedures useful for separating contaminant effects from effects caused by other environmental conditions. Although the bulk of literature surveyed was of the focused cause-and-effect type that is necessary for proving relations between contaminants and wildlife, community or ecosystem field assessments, as sometimes performed with reptiles and amphibians, might be a useful alternative for estimating the potential of a contaminant to cause environmental harm. Assumptions about the special usefulness of reptiles and amphibians as environmental indicators ought to be tested with comparisons to mammals and birds. Information on the effects of contaminants above the individual level is needed to generate accurate estimates of the potential consequences of anthropogenic pollution (e.g., ecological risk assessments). However, realized population, or higher, levels of effects should not be part of regulatory guidelines because the threshold of harm would be too high to be used as a catalyst for action. Measures of realized population or community effects could be used to evaluate the effectiveness of regulatory actions and assess chronic or difficult environmental problems. Some of these information needs can be satisfied with modest effort and expense, but much of the suggested work that incorporates great complexity or long duration is likely to be difficult to accomplish. Cooperation among investigators with different specialties and a willingness by government, academia, and corporate organizations to support the most challenging work will be necessary. Because we are unlikely to have the financial resources to evaluate more than a small number of contaminants for effects at the levels of population, community, or ecosystem, we might need to thoroughly study a few contaminants and then extend the findings to functionally similar contaminants. If sufficient cooperation and organizational support does not materialize, the pursuit of estimation methods will overshadow the collection of actual information on relations between contaminants and wildlife.