Spatially distributed populations can exhibit source-sink dynamics. Despite long-standing interest in source-sink dynamics, identifying sources and sinks and quantifying their coupled dynamics in wild populations remains challenging. We used 26 years of data on reproduction, movement, and survival to estimate annual source-sink dynamics across the entire range of the endangered Everglade snail kite (Rostrhamus sociabilis plumbeus) during the invasion of a novel prey species. Populations underwent striking changes in source-sink dynamics over time since invasion. Our results illustrate how invasive species can cause impacts in uninvaded populations via changes in native species retention and emigration. Fluctuating emigration and retention alter long-term interpretations of source-sink dynamics through variation in local vs landscape contribution of populations to the metapopulation. Our study demonstrates that the status of ‘source’ or ‘sink’ can be highly variable through time and highlights how emigration is a key dimension of metapopulation dynamics, an insight with consequences for conservation.

Empirical investigations of source-sink dynamics are needed to relate metapopulation theory to spatially-structured, temporally varying population dynamics in the real world. The difficulty of acquiring the data needed to estimate vital rates has often constrained analyses of source-sink dynamics to static or simulated systems. We used 26 years of data on reproduction, movement, and survival to estimate annual source-sink dynamics across the entire range of the endangered Everglade snail kite (Rostrhamus sociabilis plumbeus) during the invasion of a novel prey species. Populations underwent striking changes in source-sink dynamics over time, varying with time since invasion. Both source and sink populations depended on immigration to offset high emigration rates. Source populations are often prioritized for conservation, but in a metapopulation composed largely of sinks, or sources with high emigration and immigration, an emphasis on maintaining connectivity or expanding available breeding habitat may be more important.

Adaptive habitat construction is a process by which individuals alter their environment so as to increase their (inclusive) fitness. Such alterations are a subset of the myriad ways that individuals condition their environment. We present an individual-based model of habitat construction to explore what factors might favor selection when the benefits of environmental alterations are shared by individuals of the same species. Our results confirm the predictions of inclusive fitness and group selection theory and expectations based on previous models that construction will be more favored when its benefits are more likely to be directed to self or near kin. We found that temporal variation had no effect on the evolution of construction. For spatial heterogeneity, construction was disfavored when the spatial pattern of movement did not match the spatial pattern of environmental heterogeneity, especially when there was spatial heterogeneity in the optimal amount of construction. Under those conditions, very strong selection was necessary to favor genetic differentiation of construction propensity among demes. We put forth a constitutive theory for the evolution of adaptive habitat construction that unifies our model with previous verbal and quantitative models into a formal conceptual framework.