Beyond abundance: the direct and indirect effects of predation in a terrestrial trophic web

We need to understand the mechanism by which species interact in food webs to predict how natural ecosystems will respond to disturbances that affect species abundance, such as the loss of top predators. The study of predator-prey interactions and trophic cascades has a long tradition in ecology, and classical views have focused on the importance of lethal predator effects on prey populations (direct effects on density), and the indirect transmission of effects that may cascade through the system (density-mediated indirect interactions). However, trophic cascades can also occur without changes in the density of interacting species, due to non-lethal predator effects on prey traits, such as behaviour (trait-mediated indirect interactions). Studies of direct and indirect predation effects have traditionally considered predator control of herbivore populations; however, top predators may also control smaller predators. Due to their versatility and high predation rates, intermediate-level predators (mesopredators) may have disproportionably larger effects on lower trophic levels than top predators. Despite knowledge of the importance of indirect predation effects, and the effects of mesopredators, there is little evidence of the relative importance of density- and trait-mediated effects of predation in complex terrestrial trophic systems. My study consisted of a manipulative field experiment where the effects of top predators and mesopredators were both simultaneously and individually isolated. I manipulated the access of varanids (top predators) and skinks (mesopredators) to areas at my study site using large (200 m2) fenced study plots, and determined the effects of these predators on spiders (lower-order mesopredators), other arthropods and primary producers. To appreciate the impact that predation had on fauna, a detailed understanding of the ecology of the prey species was needed. Small skinks dominated the faunal community at my study site and thus were likely to make a large contribution to energy flow in the system, both as prey to higher trophic levels and as predators of lower trophic levels. Current knowledge of the demography, ecology and life history strategies of small, tropical skinks was inadequate. I described the population structure, survival rates, reproductive seasons and recruitment patterns of the dominant skink species at my study site, and I defined their growth patterns, age at maturity, sexual dimorphism and relationship with ectoparasites. I found that these closely-related, sympatric skinks shared similarities in many aspects of their life history, but that different mechanisms acted on juvenile skinks to cause various amounts of sexual dimorphism of adult body size. Observations of skink activity, use of microhabitat and behaviour revealed that skinks were highly active; suggesting that they were efficient and effective foragers, but also conspicuous and highly vulnerable to predators. From diet analysis, I determined that skinks were selective in their prey choice, consuming certain prey types and sizes that were in short supply in the environment. The species varied in the prey they consumed, but spiders and orthopterans were consistent prey types for all skinks. Skinks were found to have a strong direct effect on spider abundance in study plots; they reduced the density of cursorial spiders and of web-building spiders close to the ground. The number of web-building spiders was, in turn, negatively correlated with the abundance of aerial arthropods. Herbivore damage to leaves was reduced in skink access plots. The selective removal by skinks of other predators (spiders) and herbivores (orthopterans) from the trophic system was thus shown to have far-reaching indirect effects that cascaded through the system to alter the composition of the arthropod community, and the amount of damage to local vegetation. Varanids did not directly alter the abundance of skinks or arthropods, but did have strong indirect effects. Skinks significantly reduced their activity in the presence of varanids, likely leading to lower energetic requirements and less foraging. The cascading effects of predation on skink activity were evident when comparing arthropod community composition among study plots: skinks directly altered the structure of the arthropod community by consumption and in the presence of varanids, skink foraging was restricted. In the absence of varanids, however, skinks were 'released' from predation threat, and only those arthropod taxa rarely consumed by skinks were abundant. Varanids thus indirectly altered arthropod community composition by altering skink behaviour, clear evidence of a trait-mediated indirect effect of predation. This is the first study, to my knowledge, to experimentally show the effect of behaviourally mediated mesopredator release on lower trophic levels, and such behavioural predator effects may be very influential in community ecology. The results of my study reveal the strength and importance of indirect predation effects, and show that measuring more than just the abundance of individuals is vital to determine the full effects of predator removal.