Invasibility of Experimental Habitat Islands in a California Winter Annual Grassland

In an experimental test of plant community invasibility, we introduced seeds of a native ruderal, California poppy (Eschscholzia californica), at fixed density into experimental plots in a California winter annual grassland. Each of the 42 plots, which ranged in size from 2 m 2 to 32 m 2 , had been studied for 4 yr previous to the introduction, with the common observation that a subset of plots of each size consistently held more species than others. It was primarily in these more species—rich plots that establishment and reproduction by the experimental invader occurred. Success of the invader per plot, measured as the total number of plants germinating, producing seeds, or perennating, varied with plot size, but the statistical contribution of plot size was secondary to that of local species number. Contributing variables were the extent of small mammal disturbance (positive) and the degree to which a single resident plant species (in particular, Bromus diandrus) dominated a plot (negative). In contrast to theories of competitive exclusion via niche partitioning, species—rich plots were more invasible.