Population introductions provide rare opportunities to observe ecological and evolutionary dynamics from a known starting point. Yet we still struggle to predict which populations will establish, how species interactions will change following founding, and whether subsequent adaptation will follow repeatable trajectories. To address these questions, we introduced more than 12,000 threespine stickleback from multiple, ecologically divergent source populations into nine lakes in Alaska and followed the resulting populations through time. I will first show that early source-population success was remarkably consistent among recipient lakes. Intrinsic differences among source populations generally outweighed environmental matching, challenging the assumption that populations originating from environments similar to their destinations will necessarily perform best following introduction. I will then examine how population founding altered host–parasite dynamics, generating rapid changes in infection and heritable immune traits. Finally, I will describe ongoing genomic and phenotypic time-series analyses testing whether replicate populations converge as they adapt to their new environments or retain persistent signatures of source history and contingency. By connecting establishment, species interactions, and genomic evolution within a replicated whole-ecosystem experiment, this work provides a window into the processes governing population resilience in novel environments. More broadly, it suggests that predicting responses to environmental change requires understanding not only the environments populations encounter, but also the ecological and evolutionary histories they carry with them.