The level of endemism at a site may indicate species richness of the site. Nevertheless, assessing endemism levels in taxonomic groups such as plants may be difficult because the species richness of plants is high relative to species richness of other taxonomic groups (e.g., vertebrates). A major problem in determining whether plant species are endemic is the lack of standardization of the geographic extent of endemism: species are considered endemic to, for example, countries, continents, or states. We compiled a history of the concept of endemism as it applies to plants. The application of the concept to geographic distribution dates from the 19th century, when European explorers discovered many taxa exclusive to regions outside Europe. Two types of endemism, paleoendemism and neoendemism, were then acknowledged, according to evolutionary age, and these categories are still in use. In the 20th century, most of the research on endemism focused on explaining range restriction on the basis of cytological data, edaphic and geological factors, and phylogeny. This research led to a vast number of concepts, of which only edaphic endemism remains relatively well accepted. More recently, researchers suggest that competition may determine endemism in some cases. We suggest that plants be labeled as endemic only if their distribution occurs in a distinct ecological unit, such as a biome. On the basis of a literature review of the factors that cause range restriction, we categorized endemic taxa as paleoendemic, neoendemic, edaphically endemic, or suppressed endemic. For example, Schlechtendalia luzulifolia, is a rare forb that is a paleoendemic species of the granite and sandstone-based grasslands of the Pampa. Levels of endemism in southern Brazilian grasslands are poorly known. We emphasize the importance of recognizing these grasslands as a single transnational biome so that levels of endemism of species therein can be assessed correctly.