The Janzen-Connell hypothesis proposes that plant species richness is maintained by distance- and density-dependent seedling recruitment, which should produce increased dispersion of conspecifics through time. While there has been much support for increased mortality at high densities or close proximities to conspecifics, very few studies have documented the resulting spacing dynamics. Our unprecedented long-term dataset on seedling dynamics in the BCI 50-ha Forest Dynamics Plot (Panama) allowed us to track single cohorts of 50 species over a twelve-year period. From 2001 to 2013, all woody seedlings were tagged and measured every 1-2 years within 20,000 1-m2 quadrats. We ask 1) how does mortality vary as a function of distance to adult conspecific over 12 years? and 2) does the seedling density function shift away from adult conspecifics over time? We estimated the probability of seedling mortality as a function of distance from conspecific adult using mark correlation functions, and compared these results to random mortality null simulations. We then estimated seedling densities from adults using the pair correlation function for each census interval. For both analyses, we report individual species results, as well as community-wide summaries using standardized effect sizes that control for variation in abundance and mortality rate.
Results/Conclusions
We found evidence of negative distance-dependent mortality for approximately one third of the 50 species analyzed. However, wide variation existed across the community, and most species exhibited random patterns of mortality with respect to proximity to adults. Two species showed a decrease in mortality when close to large adults. When comparing the average seedling densities across all species, the peak curve was always closest to zero for all census intervals. However, when using standardized scores that provide equal weights to species, a dramatic shift in the mean peak occurred from zero to 28 meters from 2001 to 2013. Our results provide support for negative distance dependence as a stabilizing force at BCI. Importantly, ours is one of the only studies to show a shift in the peak recruitment curve over a long period of time for many species in a tropical forest community. However, reconciling observed patterns of adult aggregation with theoretical expectations from the Janzen-Connell hypothesis remains a challenge for future work. Therefore, we recommend that future studies on the Janzen-Connell hypothesis focus more explicitly on linking spatial demography to long-term community dynamics.