Dispersal is considered a vital life history characteristic for insects exploiting temporary habitats, and life history theorists have often hypothesized an inverse relationship between dispersal capability and habitat persistence. Most often, this hypothesis has been tested using interspecific comparisons of dispersal capability and qualitative estimates of habitat persistence. Consequently, most assessments have failed to control for possible phylogenetic nonindependence and they also lack quantitative rigor. We capitalized on existing intraspecific variation in the dispersal capability of Prokelisia planthoppers to examine the relationship between habitat persistence and dispersal, thereby minimizing possible phylogenetic effects. Two congeneric species (Prokelisia marginata and P. dolus) occur in the intertidal marshes of North America, where they feed exclusively on cordgrasses (Spartina). Because these planthoppers exhibit wing dimorphism, flight-capable adults (macropters with fully developed wings) are easily differentiated from flightless adults (brachypters with reduced wings). Thus, dispersal capability can be readily estimated by the percentage of macropters in a population. At a regional spatial scale, we found a highly significant negative relationship between dispersal capability (percent macroptery) and habitat persistence. In this system, habitat persistence is influenced by a combination of marsh elevation, winter severity, and tidal range, which interact to determine the ability of planthoppers to endure through winter in their primary habitat for development. P. marginata develops primarily in low-marsh habitats during summer, habitats that can be subjected to pronounced winter disturbance due to ice scouring and/or extensive tidal inundation. Levels of winter disturbance of the low marsh are extreme along the Atlantic coast, intermediate along the Pacific, and low along the Gulf. Both the failure of P. marginata populations to remain through winter in this habitat, and the dispersal ability of these populations (92%, 29%, and 17% macroptery, respectively), are correlated with levels of disturbance. Thus, in regions where winter disturbance is high, levels of dispersal are correspondingly high to allow for recolonization of extirpated habitats from overwintering sites on the high marsh. Unlike P. marginata, P. dolus develops primarily in high-marsh habitats, which are much less disturbed on all coasts during winter. Consequently, this species remains year-round in its primary habitat for development, and most populations exhibit relatively low levels of macroptery (<10%). When raised under common garden conditions, many more macropters of both species were produced from Atlantic compared to Gulf populations. Thus the proportion of macropters produced from the populations used in this experiment paralleled the incidence of macroptery measured in the field, providing evidence that the geographic variation in dispersal capability in both species has in part a genetic basis. The results of this study provide strong intraspecific evidence for an inverse relationship between the dispersal capability of insects and the persistence of their habitats.