SPATIAL DENSITY DEPENDENCE SCALES UP BUT DOES NOT PRODUCE TEMPORAL DENSITY DEPENDENCE IN A REEF FISH.

Field experiments provide rigorous tests of ecological hypotheses but are typically of short duration and use small spatial replicates. We assessed empirically whether the results of experiments testing for density dependence applied at larger spatial domains and explained temporal population dynamics. We studied a small coral reef fish, the goldspot goby (Gnatholepis thompsoni), in the Bahamas. We assessed the effects of interactions with conspecifics and with an ecologically similar species, the bridled goby (Coryphopterus glaucofraenum). Two density manipulations on small reef patches revealed that goldspot goby mortality over one month increased as conspecifics became crowded. On five large natural reefs, we correlated the initial year-class density of both species (annual larval settlement) with the subsequent decline of goldspot goby year-classes for five years. Mortality was correlated with conspecific density among reefs for all years, but not among years for all reefs. Thus, spatial density dependence in mortality scaled up qualitatively from small patches to entire reefs but was not associated with temporal density dependence. Our results support the conclusion that field experiments may be extrapolated to larger spatial domains with care, but that using small spatial comparisons to predict temporal responses is difficult without knowing the underlying biological mechanisms.