Studies of habitat fragmentation in marine ecosystems are few, despite a high potential for disturbance and the fragmentation of macrophytes that constitute important habitats. In this study, we investigated the relative importance of habitat fragmentation and loss on the recruitment and abundance of fishes associated with giant kelp Macrocystis pyrifera forests. We experimentally fragmented naturally occurring kelp forests at 3 sites along Santa Catalina Island, California, USA, into areas of 100 to 1600 m 2 . We also explored the relationship between kelp patch size and recruitment of fishes at smaller spatial scales by constructing kelp plots that were consistent in density but varied in area. The numerical and biomass densities of recruit, older juvenile, and adult fishes across fragment areas differed among species, resulting in linear, exponential, or asymptotic increases in numerical and biomass abundances with fragment size. Overall, fishes exhibited mostly linear relationships between abundance and fragment area, suggesting that kelp-associated fishes are more affected by habitat loss than by fragmentation. When incorporating kelp forests of larger area, however, formerly linear functions became non-linear, indicating that habitat fragmentation may become more important at larger spatial scales. At the scale of small kelp patches, the densities of Paralabrax clathratus and Oxyjulis californica were unrelated to patch size, and recruitment occurred in the smallest patches and was highly variable for both species. These results may provide resource managers with important information on the relative effects of habitat loss and fragmentation in the conservation and management of coastal fishes.

Salt marsh vegetation provides the structure that supports key wetland functions and the success of wetland restoration efforts depend on the establishment of vegetation, which can take decades. Field experiments aimed at identifying the factors that limit the colonization rate of marsh vegetation can be useful for guiding restoration efforts. We used field experiments and additional measurements to determine the causes for persistently low vegetation cover in a restored salt marsh in southern California. First, we characterized the soil properties of the marsh and found that high salinity and low soil moisture increased with elevation and improved with vegetation. Using this information, we conducted two field experiments-one in the high marsh and the other in the mid marsh to determine the effectiveness of different remedial actions (e.g. irrigation, soil decompaction, soil amendments, and planting and seeding different species) on increasing vegetation cover; the mid marsh experiment evaluated planting and seeding only. In the high marsh, plantings increased cover, and irrigation in combination with soil decompaction facilitated natural plant establishment. In the mid marsh, increases in vegetation cover varied by species planted. In both experiments, the marsh dominant (Salicornia pacifica) naturally recruited, increasing cover whereas seeding did not increase cover. In general, the magnitude and pace of vegetation growth was greater in the mid marsh relative to the high marsh that received less tidal inundation. Collectively, our results show how information gained from experiments conducted during wetland restoration can inform the processes underlying the establishment of vegetation.