Document Type



Virginia Institute of Marine Science

Publication Date



Ecology Letters





First Page


Last Page



Despite the importance of consumers in structuring communities, and the widespread assumption that consumption is strongest at low latitudes, empirical tests for global scale patterns in the magnitude of consumer impacts are limited. In marine systems, the long tradition of experimentally excluding herbivores in their natural environments allows consumer impacts to be quantified on global scales using consistent methodology. We present a quantitative synthesis of 613 marine herbivore exclusion experiments to test the influence of consumer traits, producer traits and the environment on the strength of herbivore impacts on benthic producers. Across the globe, marine herbivores profoundly reduced producer abundance (by 68% on average), with strongest effects in rocky intertidal habitats and the weakest effects on habitats dominated by vascular plants. Unexpectedly, we found little or no influence of latitude or mean annual water temperature. Instead, herbivore impacts differed most consistently among producer taxonomic and morphological groups. Our results show that grazing impacts on plant abundance are better predicted by producer traits than by large‐scale variation in habitat or mean temperature, and that there is a previously unrecognised degree of phylogenetic conservatism in producer susceptibility to consumption.


Coral reef, crustacean, grazing, herbivory, latitudinal gradient, macroalgae, marine, meta‐analysis, mollusc, plant‐animal interaction, primary production, rocky reef, sea urchin, seagrass

Poore_et_al-2012-Ecology_Letters.sup-1.pdf (216 kB)
Appendix S1.The distribution of experiments among habitats and grazer types

Poore_et_al-2012-Ecology_Letters.sup-2.pdf (101 kB)
Appendix S2. Statistical analyses of effect sizes using linear mixed models (Table S2) and multiple regression modelling (Table S3)

Poore_et_al-2012-Ecology_Letters.sup-3.pdf (207 kB)
Appendix 3.Variation in grazer effect sizes with experimental methods.

Poore_et_al-2012-Ecology_Letters.sup-4.pdf (62 kB)
Appendix 4

Poore_et_al-2012-Ecology_Letters.sup-5.pdf (61 kB)
Appendix 5

Poore_et_al-2012-Ecology_Letters.sup-6.pdf (61 kB)