Species Interactions in a Successional Grassland. I. Seed Rain and Seedling Recruitment

Abstract

SUMMARY (1) Seed rain and recruitment for the five most abundant species of grasses were quantified in a California coastal grassland undergoing succession after release from sheep grazing. (2) Species distributions were very patchy, and four patch types accounted for about 950/o of total cover. Three patch types ('perennial patches') were dominated, respectively, by one of three species of perennial grasses, Anthoxanthum odoratum, Holcus lanatus or Deschampsia holciformis. The fourth type ('annual patches') was dominated by annual grasses, including the most abundant annual Vulpia bromoides, the perennial grass Rytidosperma pilosum and forbs. (3) Dispersal limited the seed rain to the species dominating the local vegetation. Species not present in patches contributed little to the seed rain and nothing to recruitment in the interiors of those patches. (4) Species differed significantly in the densities of their seed rain in the patches they dominated, ranging from 2300 m-2 for R. pilosum to 82 300 m-2 for H. lanatus. The density of seed rain was patchy on all spatial scales examined (from cm to km), but was not significantly correlated with densities of recruits on any scale for the perennial species. However, V. bromoides recruitment correlated positively with the density of its seed rain on small spatial scales (up to 1 mi2) in annual patches. The poor relationship between seed rain and recruitment for the perennials probably reflects the over-riding importance of adult interference as an influence on seedling establishment in the perennial vegetation. (5) A. odoratum had relatively high recruitment (30 m-2) in the patches it dominated. Other perennials had very low recruitment in their own stands (less than 5 m-2). V. bromoides had the highest recruitment (904 m-2) in annual patches. (6) The seed bank contribution to recruitment was estimated using seed exclosures. Only those species abundant in the seed rain and in the local vegetation recruited inside the seed exclosures, i.e. there was no evidence of a persistent, functional seed bank of other species. The seed bank could account for no more than 30/o of V. bromoides recruitment in annual patches. However, establishment in the seed exclosures overestimated recruitment from the seed bank for A. odoratum and possibly for other perennial species (7) The implications of the results for population dynamics and succession are discussed.