Community composition and diversity arise from limitations in propagule supply (i.e. propagule or seed limitation) and propagule establishment after arriving at a site (i.e. establishment or microsite limitation). Recent meta-analyses suggest that the degree of propagule limitation depends on local abiotic and biotic conditions, which in turn are likely to vary spatially and temporally. Nevertheless, seed addition studies testing propagule limitation are rarely replicated in multiple locations and years and often lack experimental manipulations of critical determinants of propagule limitation, such as the density and species richness of the propagule pool and disturbance and resource supply of recipient site. The invasion of California (USA) grasslands by exotic annual species from the Mediterranean region is unique in its scope (over 9 million ha) and persistence (more than 150 years). This invasion provides an exciting context in which to test the role of spatial and temporal variability in mediating propagule limitation and ultimately the potential for restoration. Here I present the results of native-grass, seed-addition trials conducted along a 500 km gradient in California in three consecutive years spanning a wide range of environmental conditions and initial conditions (e.g. rainfall, seeding density and disturbance). While native grasses were able to establish at many locations, per capita seed survival was low suggesting that the fate of these species is governed by interplay between propagule and establishment limitation. Establishment primarily varied spatially, with the most successful establishment occurring at fertile locations with low resident species richness. While recruitment was highly variable among years initially, there was no difference among seedling trials after three years.