Understanding species' responses to habitat loss is a major challenge for ecologists and conservation biologists, who need quantitative, yet practical, frameworks to design landscapes better able to sustain native species. I here develop one such framework by synthesizing two ecological paradigms: scale-dependence and constraint-like interactions in biological phenomena. I develop a model and apply it to birds around Tucson, USA, investigating the manner in which spatial scales interact to constrain species distributions in fragmented urban landscapes. Species' responses vary in interesting ways. Surprisingly, most show situations in which habitat at one spatial scale constrains the influence of habitat at another scale. I discuss the implications of this work for conservation in human-dominated landscapes, and the need to recognize constraint-like interactions among processes and spatial scales in ecology.