Seed weight and environment in Mediterranean-type grasslands in California and Israel

I tested the hypotheses that competition, shade, and drought favor large seeds, in annual grasslands. Most seed ecologists have employed invalid methodology, often confusing mass and number, and drawn insupportable conclusions. There is good empirical evidence that shade and competition favor large seeds, little that drought does so. I determined mean seed weight (MSW: Baker, 1972) variation along productivity, rainfall, grazing, soil depth, slope, and shade gradients, with species weighted equally or by importance.; MSW correlated with productivity and shade and inversely with grazing pressure. Correlation with rainfall was weaker, because productivity did not always correlate with rainfall; but at two of three sites where dry and wet years could be compared, MSW was greater in the wet year. Within taxa, larger seeds were consistently produced in wetter sites.; Distribution patterns of California natives suggested that invading aliens replaced annuals, not perennials. Grazing cannot explain the dominance of introduced species, which seem more fit than natives.; Large-seeded annual grasses became dominant in a Galilee exclosure: perennials became restricted to rocks or shade. Grazing may favor perennial establishment under summer-dry conditions, while annuals can be "climax". Traditional "succession" theory is thus at best tautological. Temporal changes in microsite preference suggested that rock outcrops contribute disproportionately to diversity by serving as a refuge and providing high microhabitat density. Irregularly fluctuating disturbance regimes may foster diversity better than no or regular disturbance.; A majority of the world's largest seeded grasses are annuals, growing where there is severe seasonal drought and good wet season moisture; they appear adapted to competition or shade. Correspondence between agricultural origin centers and regions supporting large seeded annuals is excellent; the first domesticates tended to be the largest seeded grasses available. Such species are surprisingly similar to their domesticated derivatives.; The results suggest that in annuals large seed weight is selected where there is competition or shade, not drought. Seed weight in annuals is severely constrained by productivity, so it may not be possible to evolve increased weight in response to drought. The notion that annuals are ruderals even in seasonally dry climates must be abandoned.