Most marine habitats are generated by the presence of habitat-modifying species. However, little is know about many aspects of this process, such as how individual- and population-level traits of habitat modifiers affect their ability to reduce environmental stress and thus facilitate other species. An important habitat modifier in New England is the intertidal grass Spartina alterniflora which facilitates the establishment and persistence of cobble beach plant communities by reducing wave-related disturbance. The objectives of this study were to (1) quantify the modification of cobble beach habitats by S. alterniflora, (2) determine how this process is related to S. alterniflora bed traits, and (3) determine why small patches of S. alterniflora generally remain unoccupied by cobble beach plants. Our results demonstrate that S. alterniflora substantially reduces flow-related physical disturbance on cobble beaches. Behind S. alterniflora, mean flow velocity was reduced by 40-60% and substrate stability was dramatically increased compared to portions of the shoreline not bordered by this species. These comparative results were supported by a S. alterniflora shoot removal experiment, which resulted in a 33% increase in average flow velocity and an 85% increase in substrate instability relative to control areas. There was a strong inverse logarithmic relationship between bed length and both average flow velocity and substrate instability behind S. alterniflora. Most S. alterniflora beds were small and bed length was significantly related to the presence of one or more cobble beach plant species. Only 13% of beds <25 m and 40% of beds 30-40 m in length were occupied, in contrast to an occupancy rate of 87% for beds >40 m long. Seeds of two annual cobble beach species (Suaeda linearis and Salicornia europaea) were added to plots behind large (>100 m in length) and small S. alterniflora (<25 m) beds with and without a substrate stabilization manipulation. Seedlings of both species only emerged and established behind small beds when the substrate was stabilized. These results indicate that smaller S. alterniflora patches are usually unoccupied because they do not stabilize the substrate to a degree that meets the establishment requirements of seedlings. Thus, both habitat modification and facilitation by S. alterniflora are patch-size dependent. The conditionality of this facilitation appears to generate a pattern of patchy yet predictable population and community distribution at a landscape spatial scale.

B. 2004. Landscape-scale patterns of biological invasions in shoreline plant communities. Á/ Oikos 107: 531 Á/540.Little is known about the patterns and dynamics of exotic species invasions at landscape to regional spatial scales. We quantified the presence (identity, abundance, and richness) and characteristics of native and exotic species in estuarine strandline plant communities at 24 sites in Narragansett Bay, Rhode Island, USA. Our results do not support several fundamental predictions of invasion biology. Established exotics (79 of 147 recorded plant species) were nearly indistinguishable from the native plant species (i.e. in terms of growth form, taxonomic grouping, and patterns of spatial distribution and abundance) and essentially represent a random sub-set of the current regional species pool. The cover and richness of exotic species varied substantially among quadrats and sites but were not strongly related to any site-level physical characteristics thought to affect invasibility (i.e. the physical disturbance regime, legal status, neighboring habitat type, and substrate characteristics). Native and exotic cover or richness were not negatively related within most sites. Across sites, native and exotic richness were positively correlated and exotic cover was unrelated to native richness. The colonization and spread of exotics does not appear to have been substantially reduced at sites with high native diversity. Furthermore, despite the fact that the Rhode Island strandline system is one of the most highly-invaded natural plant communities described to date, exotic species, both individually and as a group, currently appear to pose little threat to native plant diversity. Our findings are concordant with most recent, large-scale investigations that do not support the theoretical foundation of invasion biology and generally contradict small-scale experimental work.

Summary 1The New England cobble beach plant community is an intertidal assemblage of halophytic forbs found exclusively behind fringing beds of the grass Spartina alterni¯ora. The purpose of this study was to determine the life stage and factors that limit the upper (landward) distribution of cobble beach plants in Narragansett Bay, Rhode Island, USA. 2 Seed traps and soil samples above the community border contained large numbers of seeds and buried seedlings of three cobble beach species, indicating that seed supply and germination do not limit vertical plant distribution. Experimentally added seeds of four species germinated above the community border, but seedlings did not emerge suggesting that seedling emergence is the proximate life stage limiting population and community distribution. 3 Measures of wave disturbance (cobble movement and change in cobble depth) indicated that substrate instability is substantially greater above than within the community, probably due to the lack of bu ering by the S. alterni¯ora bed at higher tidal heights. 4 A second seed addition experiment demonstrated that seedlings of Suaeda linearis, a common cobble beach forb, are only able to emerge and grow into reproductive adults above the community border when the substrate there is arti®cially stabilized. Seedling transplants and glasshouse manipulations demonstrated that neither herbivores nor soil quality limited seedling emergence. 5 Overall, the vertical restriction of habitat modi®cation by S. alterni¯ora appears to result in substrate instability at higher tidal elevations su cient to prevent seedling emergence and limit the vertical distribution of cobble beach plants.