Fragmentation of natural vegetation is one of the most pervasive changes in terrestrial ecosystems across the Earth. Developing a general understanding of how fragmentation affects plant and animal populations is essential to meet the pressing need for guidelines for the management of fragmented systems. Nevertheless, this general understanding has to take account of differences in ecosystem types and different biogeographic, evolutionary and ecological backgrounds against which fragmentation impacts are played out in different parts of the world. Here, we examine fragmentation impacts on plant populations by considering the processes underlying fragmentation. We suggest that it is critical to focus on the key processes that are important in particular situations, rather than assuming that the same factors are likely to be important everywhere. In other words, there are inevitable limits to generalisation because of the idiosyncratic nature of the geography, history and biota of different regions. Studies on the effects of fragmentation on plant populations have focused on a limited subset of plant types and have concentrated heavily on reproductive output rather than actual regeneration success. These studies have indicated a clear impact of fragmentation on fecundity, but there is no clear signal in terms of the actual importance of this in relation to population viability. Other factors including local habitat conditions, disturbance and competition from weeds may be just as important as the classical biogeographical impacts of fragmentation. Generalisations based on a clear assessment of key life-history processes may be valuable tools in developing management responses to ecosystem fragmentation, but this requires considerably more emphasis on factors affecting successful recruitment as well as factors affecting fecundity.