ISSUES IN AGRICULTURAL AND FOREST ENTOMOLOGY Are the ecological concepts of assembly and function of biodiversity useful frameworks for understanding natural pest control? Andrew Wilby and Matthew B. Thomas NERC Centre for Population Biology, and CABI Bioscience, Silwood Park, Ascot, Berks. SL5 7PY, UK Introduction At a time when biodiversity is being lost at an unprecedented rate due to human activity, much research effort has been spent on assessment of the importance of biodiversity for the functioning and stability of ecosystems and for the delivery of ecosystem services. Pest control has been identified on numerous occasions as a valuable ecosystem service delivered by biodiversity (Mooney et al., 1995a, b; Naeem et al., 1999; SchlaÈpfer et al., 1999), and one that is at risk from human activity (Naylor & Ehrlich, 1997). There is considerable evi- dence that as agricultural production systems are intensified by increased use of external inputs to increase yield, they tend to lose biodiversity and become destabilized, with increased frequency and extent of pest outbreaks (Altieri, 1991; Swift et al., 1996). However, we know relatively little about the ecological mechanisms that result in this destabilization, or how important natural enemy diversity is in maintaining pest-control functioning. Numerous studies exist which consider the ecological or theoretical basis of biological control (e.g. see Hawkins & Cornell, 1999) but relatively little attention has been given to understanding pest control as a natural ecosystem function and how changes in agricultural practice may affect function through impacts on biodiversity. This applies particularly to the non-equilibrium conditions that exist in short duration or seasonal cropping systems. As such, and whilst some notable exceptions exist, it can generally be argued that patterns of pest emergence remain poorly explained and attempts to improve natural pest con- trol through modifications to management practices have limited ecological foundation. The aim of this paper is to explore how current ecological thinking can help to improve on this situation and to propose some ecological frameworks that may facilitate investigation of the importance of biodiversity and the impact of management practices on it. Characterization of the relationship between biodiversity and ecosystem functioning has been a major research goal of ecology over the past decade. Numerous theoretical and empirical studies have been undertaken on many ecosystem properties, including biomass production (producers, con- sumers and decomposers), nutrient uptake and retention, decomposition, soil pH, soil water and organic matter content and community respiration (SchlaÈpfer & Schmid, 1999), although little has been published on the importance of diversity in natural enemy assemblages for pest control functioning (but see RodrõÂguez & Hawkins, 2000). The majority of these studies have observed a saturating positive relationship between diversity and ecosystem functioning (Schwartz et al., 2000). One mechanism whereby species diversity influences eco- system functioning is resource-use complementarity. That is, different species occupy different niches allowing com- plementary functioning of species in heterogeneous or tem- porally variable environments. The slope of the relationship between species diversity and ecosystem functioning is partly dependent on the extent to which species overlap functionally, with respect to the particular ecosystem process of concern. If all species have an equal and non- overlapping effect on a particular process (perfect comple- mentarity), then a linear reduction in the rate of the process would occur as species richness decreased (Fig. 1). At the other extreme, if there is no complementarity (species are `redundant' with respect to the function concerned, see Walker, 1992; Lawton & Brown, 1993), then remaining species would be able to compensate for lost species and a sudden and complete loss of functioning would occur as the final functional species was lost. Another mechanism resulting in a positive relationship between species diversity and ecosystem functioning has been termed the `sampling effect' (Aarssen, 1997; Huston, 1997). This effect occurs if the probability of including a species with a strong influence on ecosystem functioning increases as species diversity increases. The sampling effect is viewed by some as a true diversity effect (Tilman et al., 1997), but by others as an experimental artefact, because its expression is dependent on the probability of inclusion of species, which are under the control of the experimenter and usually do not reflect natural probabilities. The sampling effect is one expression of species composition effects, i.e. effects that arise from the fact that species differ in their functional capacity. More generally, compositional effects are expressedas variance inthe relationship between diversity Correspondence: A. Wilby. Tel.:  44 (0)2075942528; fax:  44 (0)1344873173; e-mail: a.wilby@ic.ac.uk Agricultural and Forest Entomology (2002) 4, 237±243 # 2002 The Royal Entomological Society