Aim I propose and develop a new classification system to explain diversity patterns in habitat fragments, equally applicable to islands and other inherently patchy ecosystems. My primary goal is to provide an inclusive model to improve the comparability of studies and enhance future efforts to synthesize their findings, yielding a generalized basis for understanding species composition in patchy ecosystems. Results Differentiating islands from fragments and incorporating patch age and patch: matrix contrast, eight classes of patch are distinguished, spanning a range of geographical features. To compare studies of diversity patterns among and between patch types, patch biota are divided into three categories based on their origin—relict species (present before fragmentation), matrix-derived species and interpatch dispersers. Applying this novel scheme to existing data, the effects of insularization are synthesized. Direct comparisons among fragments revealed broad similarity in the long-term effects of habitat fragmentation compared with highly divergent patterns in younger landscapes (<200 years). Holding patch: matrix contrast and age constant, fragments and islands were compared. Despite initial differences in community assembly, the biota of islands and fragments converge in several properties over time, as diversities stabilize and patch biotas become distinct from the surrounding matrix. Main conclusions Although necessarily broad, this framework provides an explicit context within which to test forty-four specific predictions regarding the distribution of diversity in patchy landscapes and thereby gain a clearer understanding of the long-term biological consequences of insularization. I propose that the fragments-as-islands analogy be revisited, potentially yielding valuable insight into the long-term future awaiting anthropogenically altered ecosystems.