Keystone Predation: What Is It, and Is It Supported by Empirical Evidence?

Keystone predation is the process whereby a predator indirectly facilitates weak competitors by preferentially consuming strong competitors, often affecting local diversity. The keystone predation process is therefore different from keystone species, which refers to any species with a large and disproportionately large effect relative to its abundance. Keystone predation is widely cited as a fundamental ecological process in textbooks and scientific literature, but its empirical evidence has not been reviewed. We addressed this research gap by reviewing papers that explicitly (a) tested for keystone predation with experiments or models, and/or (b) mentioned 'keystone predation' in their main text, to assess its prevalence and identify potential context dependencies. We found that almost 2000 publications mentioned keystone predation in their texts, but only 73 papers tested or discussed the process in detail. Most of the 73 studies were from North America (87.4%), temperate (59.4%) and aquatic habitats (48% freshwater, 44% marine). Just 25 publications manipulated predator abundances to explicitly test for keystone predation, of which 10, 7 and 8 publications reported consistent support (mobile invertebrates preying on sessile competitors), context-dependent support (effects varied, e.g., with temperature or nutrients) or no support (mobile competitors), respectively. The few studies that tested keystone predation make it difficult to evaluate its prevalence, its relative importance, or whether general rules exist to predict its strength. Still, the few published studies suggest that environment, species traits and dispersal can modify keystone predation. We also recognise that more papers may have tested for keystone predation, but lacked this terminology. These papers should be identified in future meta-analyses by combining wider search terms and expert knowledge. We recommend that researchers use the precise keystone predation terminology to experimentally test this process on different consumer-prey communities and under contrasting environmental conditions, to better understand its prevalence and importance in ecology.