Background/Question/Methods Many community and food web patterns are driven by movements of organisms between ecosystems. Specifically, variation in movement, caused by behavioral habitat selection of animals with complex life histories, has been shown alter assembly, structure and diversity of aquatic communities. Yet, measured variation in habitat selection in response to specific stressors in the water (e.g. predators, competitors, pesticides) could be due to 1) baseline differences in behavior or 2) feedbacks between the assembling community and the stressor of interest. Here we experimentally test the hypothesis that the community assembly process modifies the effect of a top predator on colonization and ovipoistion of aquatic habitats through time. We accomplished this by manipulating colonization patterns over 2 months in the presence or absence fish (Enneacanthus gloriosus) in experimental ponds. Results/Conclusions During colonization, most taxa avoided fish but some taxa (e.g., chironomids) were attracted to fish pools. The direction of the response of taxa to fish was likely related to the relative risk of predation by fish versus by predatory macroinvertebrates (especially, aquatic beetles) post-colonization. However, in some cases past colonization reduced (e.g., dytscid beetles) or even reversed (e.g., hylid frog) top predator effects on habitat selection. In these cases, lower numbers of large macroinvertebrates in colonist-removal pools may have increased the risk of fish consumption for some aquatic beetles and frogs. Finally, the responses of colonists to fish varied temporally across the experiment, but in only one case did we see the predicted monotonic change over time in response to fish as communities assembled. Our results suggest that community assembly can itself modify or add to the taxon-specific impacts of top predators on the colonization of lentic habitats over time by organisms with complex life histories. Variation in assembly in past studies may be partially responsible for generating measured differences in the strength of habitat selection across taxa.