Introduction

Our assessment of the importance of trait-mediated indirect effects (TMIEs) relative to density-mediated indirect effects (DMIEs) in ecological communities continues to rise. It seems only yesterday that the landmark experiment of Beckerman et al. (1997) was published, showing that grasshoppers, in the presence of ‘disabled’ spiders incapable of consumption, were intimidated enough to reduce their feeding activity and thereby maintain strong trophic cascades to the producer level. The first comprehensive review, by Werner and Peacor, appeared 6 years later and summarized numerous studies of TMIEs, including some that were able to calculate the relative importance of trait versus density effects; they concluded ‘trait effects are often as strong or stronger than density effects’. A short time before, the same authors published an experiment (Peacor and Werner 2001) on a community consisting of an odonate predator and two competing anuran prey species; the predator consumed and intimidated one of the prey species, reducing its competitive effect on the second anuran species. The impact via the trait effect was 76–86%, as compared to 14–24% via the density effect. A second review by Peacor and Werner (2004) concluded that the effect of predators on prey traits modifies the magnitude of an effect farther along the pathway by 20–90%. In the same year, Schmitz et al. (2004) published a review of TMIEs and DMIEs in trophic cascades, whose title gives the conclusion ‘Trophic cascades: the primacy of trait-mediated indirect interactions’. Shortly after, Preisser et al. (2005) did a meta-analysis of trait and density effects, finding for simple predator–prey interactions that certain trait effects (which they call ‘intimidation’) were at least as strong as density effects (which they call ‘direct consumption’), being 63% versus 51%, respectively; the density effect actually declined along tritrophic cascades, while the trait effect climbed to 85% of the total effect. More recently again, Creel and Christianson (2008) reviewed risk effects and concluded that they ‘can be large, sometimes substantially larger than direct effects’. One of their examples was an elegant experiment by Pangle et al. (2007) showing that the risk effects of predatory water fleas on the population growth of their zooplankton prey were more than seven times larger than the consumptive (direct predation) effects.