Dummy caterpillars as a simple method to assess predation rates on invertebrates in a tropical agroecosystem

Abstract

Attention to the role ecosystem services play in main-taining the world’s habitats and their inhabitants isincreasing as both our understanding and opportunitiesto estimate the economic value of these services increase.Generally, these services are of immense importance,not only from an anthropogenic viewpoint; they are alsonecessary to maintain the health and functioning of eco-systems (MA, 2005). The recently completed Millen-nium Ecosystem Assessment highlighted the need formethods to evaluate these services, but existing knowl-edge already indicates that many ecosystem services arenot in good shape (MA, 2005).Agricultural management is very important from thisperspective (Tilman et al., 2002; Dale & Polasky, 2007).Ecosystem services tend to be more important in devel-oping than in developed countries (Mertz et al., 2007),mainly (but not only) because in those agroecosystems,external resources are often not available to substitutefor ecosystem services, especially in rural communities(Mertz et al., 2007). Biological control⁄natural pest con-trol through predation and parasitism by an array ofnaturally occurring organisms (native natural enemies)is one of the recognised ecosystem services (de Grootet al., 2002). However, our knowledge about the level ofbiological control provided by ecosystems is not toodetailed, and especially lacking from developing coun-tries in the tropics.Studies on natural enemies often assume density as ameasure of natural enemy importance or intensity offunction (Waage & Mills, 1992), arguing that a doublingin natural enemy densities equates to a two-fold increasein predation pressure. This is not necessarily so, becauseseveral factors, including predator satiation, intraguildpredation, and intra- and interspecific competition cancomplicate the picture (Schmitz, 2007). Therefore, thereis a need for functional biological control studies whichmeasure levels of natural enemy pressure on pests underfield conditions. Natural enemy activity in general is dif-ficult to detect, especially by invertebrates. Assessingparasitism rates is less problematic because the hosts ofparasitoids often become immobile and are easily recog-nised. However, predation often leaves no trace, or onlyfragments of the consumed prey can be found. A preda-tion event usually happens quickly, predators often hidewhile consuming prey, and many of them are active atnight (Crawley, 1992). Such factors make field assess-ments of predation intensity difficult. These inherentproblems have curtailed our knowledge about the kindand extent of predation in different habitats by differentnatural enemies (Crawley, 1992). Consequently, we haveseveral examples of quantitative parasitoid food webs(Mu¨ller et al., 1999), but fewer of predator food webs.Predation can be studied by video recording (Varleyet al., 1994), by indirect means, using dyes, markers,and analysis of predators for prey remains, or by usingsentinel prey (Jervis & Kidd, 1996). The last methodmonitors the rate of disappearance of prey provided bythe experimenter, and is the easiest one to obtain quan-titative data on predation pressure. Immobile (eggs,pupae) or immobilised stages of arthropods are oftenused as sentinel prey. This approach has several inherentproblems, such as generating sufficient numbers (whichoften requires mass rearing), putting them out into thefield, and recording⁄identifying the natural enemiesresponsible for prey disappearance. Several of the abovedifficulties can be avoided or circumvented if artificial