Demographic characteristics such as survival and life expectancy play critical roles in population dynamics and, consequently, community structure. Though increasing attention has focused on factors underlying geographic variation in a species' demographic rates, the difficulty of collecting long-term observation data for multiple species from multiple sites has limited our understanding of the influence of life form and environmental drivers. We compiled five long-term datasets from Sonoran desert, Arizona (1915-1935), Sagebrush steppe, Idaho (1923-1957), Southern mixed prairie, Kansas (1934-1972), Northern mixed prairie, Montana (1932-1945) and Chihuahuan desert, New Mexico (1915-1960), USA, to test for site-dependent demographic responses for forbs and grasses. In these datasets, all individual plants in 1-m2 permanent quadrats were annually identified and mapped, allowing us to track the lifespan of individual genets. Based on these data, we characterized the survival curves of 114 species. Pooling forb and grass species together, we found significant differences in first year survival rates across the five sites, but differences in life expectancy at year 1 and the shape of the survival curves (Type III) were not significant. For forbs, first year survival and life expectancy, but not the shape of the survival curve, differed significantly among sites. Grasses showed statistically indistinguishable patterns across all sites. In addition, we found significant differences between forbs and grasses in all three demographic rates, with forbs conforming more to Type II survivorship curves than grasses. Our results emphasize that much of the variation in survival across and within species is related to life form (forbs vs. grasses). In contrast, geographic differences in demography were relatively subtle, especially for the grasses, despite dramatic variation in climate and soils across the five sites.