PREDATORS, THEIR PREY, AND WHAT COMES BETWEEN THEM Completing the Web in a Trophically Complex Coffee Agrosystem

Abstract

The tropics are home to two important entities: tropical ecosystems (which house an amazing amount of biodiversity) and many fast-growing export economies (which are converting those ecosystems into managed ones with alarming zeal). In many ways, coffee epitomizes this transformation (13, 2). Spurred by skyrocketing global demand, coffee production is intensifying at the expense of primary forest (13). However, shade-grown coffee, oft championed by conservationists, could be a best-of-both-worlds alternative, able to please both producers and consumers while providing viable refuge for dwindling tropical diversity (24, 17, 4, 13, 20, 2, and many others). Furthermore, shaded coffee systems might also provide a conduit for migration between forest fragments, forging a metapopulation structure that would facilitate landscape-wide conservation (23, 28, 29). This diversity then gives back in the form of ecosystem services, valuable services performed by functional, healthy ecosystems, such as biological control of pests through top-down, or predator-driven, effects (22, 27, 26, 25, 13, 39, 2). Top-down control in complex systems often transcends simple predator-prey relationships; such systems sport a web structure, one in which intraguild (in this case, between predator) predation, as well as trait mediated effects—those effects which alter the rate at which predator-prey interactions occur—blur the linearity of the one-dimensional food chain (38). While viewing the trophic structure as a web of interactions can make teasing apart the net impacts of each individual predator more difficult, doing so is often more realistic and more enlightening (31, 23, 16, 27, but see 12). Four predatory taxa that research has found critical to shade-coffee agrosystems are insectivorous birds (9, 22, 13, 39), bats (39), spiders (30), and Azteca ants (25, 26, 27). Some studies have even tried to study interactions between two or more of these groups (39, 26). However, we present the here-to-date first attempt at examining the entire predator guild at once. Once thought to be pests or liabilities, predators are now thought to be critical for maintaining low pest levels, and how factors such as predator diet breadth, prey diversity, predator habits and preferences, prey palatability and abundance, and intraguild predation alter the regulatory capacity of predators is now under increasing scrutiny (12, 33, 31). We argue that studying predators in these systems is paramount, not tangential, to conservation efforts (39). One way to successfully attach a value to top-down control is to simply remove the predator(s) and observe the response of the lower trophic levels (12). If a predator exerts a strong top down control, we would expect a trophic cascade-- a “series of alternating positive and negative effects (33)”--to precipitate down through the trophic levels in its absence (for examples, see 1, 34). After we observe (or don’t observe) these cascading effects, we can begin to characterize the role of the predator. Of course, some caveats to this approach do exist. First, when intraguild predation, omnivory, cannibalism, or non-hierarchical food webs characterize a system, predator effects can become blurry (1, 16, 13). For example, in addition to their predictable consumption of prey species, birds (31), spiders (6), and ants (26) are all known to engage in intraguild predation or trait mediated effects in this system: Birds and spiders are both known to eat ants on occasion (27, 10, but see 13, 2) and Azteca ants are thought to actively reduce bird foraging times (26, 27). Still, we feel predator removal experiments have a lot to offer studies of top-down control, especially when the study design addresses the entire predator guild. In this study, our lab excluded all possible combinations of flying insectivores (birds and bats), spiders, and Azteca ants in order to elucidate their impacts on lower trophic levels in a shaded coffee agroecosystem. Birds and bats were excluded with fish netting over individual coffee plants, spiders were physically removed from coffee plants, and Azteca instabilis, a keystone arboreal and aggressive ant species (37) was spatially segregated by the experimental design. I then charted the responses of the arthropod assemblages on these plants over time. I addressed the following questions: (1) How much does each predator individually affect arthropod taxa positively and/or negatively? (2) How does the combination of predators affect arthropod taxa? How do they complement or interfere with one another? (3) Can we observe intraguild predation in this system? (4) Do we see evidence of top-down control in this system?