The fragmented Quercus oleoides (tropical live oak) forests in northwestern Costa Rica are nearly entirely contained within the Área de Conservación Guanacaste (ACG).  Former cattle pastures are no longer grazed but live oak seedlings have not been found recolonizing the pastures.  Our goal was to determine the factors limiting seedling regeneration and to test for local adaptation of seedlings from different elevations.  We established 18 transects across the forest to pasture transition at six sites and planted oak seedlings at distances up to 20 meters into the forest and pasture.  Additionally, we characterized the abiotic changes across this transition, and we grew seedlings in a light, water, and a grass competition factorial treatment experiment for two years.  We measured seedling growth and survival to determine if pasture conditions were limiting regeneration.  Finally, we tested for local adaptation among populations by planting seedlings from maternal families from different elevations in lowland and upland common gardens.  We compared estimates of fitness in the home environment versus the away environment with a conditional likelihood method (aster models) that appropriately models  combined growth and survival data.

Results/Conclusions
Preliminary data indicated that seedlings survived equally well along the forest to pasture transect during the first year.  Survival decreased in the second year, however, and mortality appeared to be greater in pastures.  This result might be due to unexpected belowground herbivory at some sites.  Pasture seedlings that did survive had greater biomass than forest seedlings indicating a possible growth/survival trade-off.  In the factorial treatment experiment there was a significant effect of site on biomass but no significant effect from the treatments after one year.  Combined these results indicate that the abiotic environment and competition from pasture grass do not limit establishment of oak seedlings in pastures.  Lack of regeneration in the pastures may be better explained by dispersal limitation.  In the reciprocal transplant experiment, seedlings of maternal families from the more severe lowland environment (longer dry season and wetter wet season) had a higher estimated fitness in both the lowland and upland garden.  We hypothesize that evolution of increased plasticity in the lowland population may allow the maintenance of a fitness advantage in both sites. Collectively, these results suggest that a more active approach to the restoration of these forests will be necessary, and attention to seed source may be important in the successful establishment of live oak seedlings.