Small-scale patterns of species richness in Swedish semi-natural grasslands:: the effects of community species pools

Temperate semi-natural grasslands are known for their high plant species richness at small spatial scales. We examined the variation in small-scale species richness in a sample of 63 sites from Swedish semi-natural grasslands, located as fragments in the modem landscape dominated by forest and agricultural land. Data were obtained from two spatial scales at each site, 1 dm(2) and 4 m(2). Using an analysis based on a Monte Carlo simulation, we found support for the species-pool hypothesis: a high species richness at the 1 dm(2) scale was associated with high species richness at the 4 m(2) scale. The conclusion from this pattern analysis would, however, be considerably strengthened if we could reduce the likelihood that other mechanisms than sampling from species pools of unequal size influence the pattern of small-scale species richness. Additional analyses were made in order to identify such mechanisms. We examined whether four putative key traits: seed size, seed production, plant size and reproductive allocation were different among species at comparatively species-rich vs species-poor 1 dm(2) plots. We found only a little evidence for such differences. There was a weak tendency that species in the plots with high species richness possessed larger (and fewer) seeds than species from species-poor plots. Our results are congruent with the main prediction of the species pool model; variation in small-scale species richness (1 dm(2) scale) is basically a result of sampling from unequally sized community species pools (4 m(2) scale). Variation in species richness between the 4 m(2) sen-ii-natural grassland "patches" may thus be sought for among mechanisms operating at larger spatial scales than 4 m(2). We briefly discuss such mechanisms, based on other studies performed in the same study area.