Context: Metapopulation theory makes useful predictions for conservation in fragmented landscapes. For randomly distributed habitat patches, it predicts that the ability of a metapopulation to recover from low occupancy level (the “metapopulation capacity”) linearly increases with habitat amount. This prediction derives from describing the dispersal between two patches as a function of their features and the distance separating them only, without interaction with the rest of the landscape. However, if individuals can stop dispersal when hitting a patch (“habitat detection and settling” ability), the rest of habitat may modulate the dispersal between two patches by intercepting dispersers (which constitutes a “shadow” effect). Objectives: We aim at evaluating how habitat detection and settling ability, and the subsequent shadow effect, can modulate the relationship between the metapopulation capacity and the habitat amount in the metapopulation. Methods: Considering two simple metapopulation models with contrasted animal movement types, we used analytical predictions and simulations to study the relationship between habitat amount and metapopulation capacity under various levels of dispersers’ habitat detection and settling ability. Results: Increasing habitat detection and settling ability led to: (i) larger metapopulation capacity values than expected from classic metapopulation theory and (ii) concave habitat amount–metapopulation capacity relationship. Conclusions: Overlooking dispersers’ habitat detection and settling ability may lead to underestimating the metapopulation capacity and misevaluating the conservation benefit of increasing habitat amount. Therefore, a further integration of our mechanistic understanding of animals’ displacement into metapopulation theory is urgently needed.