Unity through nonlinearity: a unimodal coral-nutrient interaction.

The magnitude and direction of biological effects of environmental disturbances can vary considerably, especially among studies that use presence/absence manipulations. Because nonlinearities (e.g., humped relationships) are common in biological systems, this heterogeneity in effects may arise if systems are similar in their responses but specific studies use few (e.g., two) levels, or a narrow range, of a factor. To test whether nonlinearity can explain heterogeneous responses to a common environmental disturbance, I examined the effect of nutrient enrichment on coral growth, which has been previously shown using simple (e.g., two-level) manipulations to yield positive, negative, or neutral responses. I subjected corals (Porites) to a nutrient gradient in situ for 28 days. Coral growth rate increased (2.4-fold) then decreased (2.7-fold) with enrichment, returning to near-ambient values at the highest nutrient levels. This unimodal response could explain disparities among past findings and provides a compelling case for using regression designs to understand heterogeneity within ecological interactions.