The suppression of agricultural pests has often been proposed as an important service of natural enemy diversity, but few experiments have tested this assertion. In this study we present empirical evidence that increasing the richness of a particular guild of natural enemies can reduce the density of a widespread group of herbivorous pests and, in turn, increase the yield of an economically important crop. We performed an experiment in large field enclosures where we manipulated the presence/absence of three of the most important natural enemies (the coccinellid beetle Harmonia axyridis, the damsel bug Nabis sp., and the parasitic wasp Aphidius ervi) of pea aphids (Acyrthosiphon pisum) that feed on alfalfa (Medicago sativa). When all three enemy species were together, the population density of the pea aphid was suppressed more than could be predicted from the summed impact of each enemy species alone. As crop yield was negatively related to pea aphid density, there was a concomitant non‐additive increase in the production of alfalfa in enclosures containing the more diverse enemy guild. This trophic cascade appeared to be influenced by an indirect interaction involving a second herbivore inhabiting the system – the cowpea aphid, Aphis craccivora. Data suggest that high relative densities of cowpea aphids inhibited parasitism of pea aphids by the specialist parasitoid, A. ervi. Therefore, when natural enemies were together and densities of cowpea aphids were reduced by generalist predators, parasitism of pea aphids increased. This interaction modification is similar to other types of indirect interactions among enemy species (e.g. predator–predator facilitation) that can enhance the suppression of agricultural pests. Results of our study, and those of others performed in agroecosystems, complement the broader debate over how biodiversity influences ecosystem functioning by specifically focusing on systems that produce goods of immediate relevance to human society.

Detecting the presence of rare species has interested ecologists and conservation biologists for many years. A particularly daunting application of this problem pertains to the detection of non‐indigenous species (NIS) as they colonize new ecosystems. Ethical issues prevent experimental additions of NIS to most natural systems to explore the relationship between sampling intensity and the detection probability of a colonizing NIS. Here we examine this question using a recently introduced water flea, Cercopagis pengoi, in Lake Ontario. The species has biphasic population development, with sexually‐produced ‘spring morphs’ developing prior to parthenogenetically‐produced ‘typical’ morphs. Thus, this biphasic morphology allows distinction between new colonists (spring morphs) from subsequent generations. We repeatedly sampled Hamilton Harbour, Lake Ontario for the presence of both spring and typical morphs. Probability of detection was positively related to both the number of samples taken and animal density in the lake; however, even highly intensive sampling (100 samples) failed to detect the species in early spring when densities were very low. Spatial variation was greatest when densities of Cercopagis were intermediate to low. Sub‐sampling, which increased space between adjacent samples, significantly decreased the number of samples required to reach greater, calculated detection probabilities on these dates. Typical sampling protocols for zooplankton have a low probability (< 0.2) of detecting the species unless population density is high. Results of this study suggest that early detection of colonizing, aquatic NIS may be optimized through use of a risk‐based sampling design, combined with high sampling intensity in areas deemed most vulnerable to invasion, rather than less intensive sampling at a wider array of sites.

Antipredator defensive behaviors are a well‐studied and often crucial part of prey life histories, but little has been done to quantify how such behaviors affect natural enemies, their foraging, and their effectiveness as biological control agents. We explored how the generalist predatory coccinellid Harmonia axyridis Pallas (Coleoptera: Coccinellidae) affects the dropping behavior of the pea aphid, Acyrthosiphon pisum (Harris) (Homoptera: Aphididae), and in turn, how that defensive behavior affects the foraging efficiency of the predator. Experimental arenas that allowed or prevented pea aphid dropping were compared to determine how dropping influences the foraging of multiple life stages of H. axyridis: second instars, fourth instars, and adults. Dropping reduced predation on aphids by all ladybeetle life stages. Despite older predators inducing more dropping, aphid dropping reduced predation by approximately 40% across all ladybeetle life stages. Aphid dropping and predator consumption of aphids were both correlated with how much the predator moved, which also increased with predator life stage. We suggest that the high rates of dropping induced by H. axyridis and the subsequent decrease in H. axyridis foraging efficiency may partially explain why H. axyridis is less effective at controlling pea aphids than it is at controlling other aphid species that do not drop.

Interactions between natural enemies can be crucial for determining their overall control of pest species, yet the mechanisms that govern such interactions are often poorly understood. The risk of negative effects such as intraguild predation and the possibility of mitigating such risks are important components for ultimately determining the compatibility of biological control agents. We performed a group of experiments to determine whether the coccinellid Harmonia axyridis Pallas (Coleoptera: Coccinellidae) poses an intraguild threat to the parasitoid Aphidius ervi Haliday (Hymenoptera: Braconidae) and to see whether A. ervi is able to avoid predation by responding to the chemical tracks deposited by H. axyridis. We show that although H. axyridis does not readily consume A. ervi mummies, it preferentially consumes parasitized aphids over unparasitized aphids. We also show that A. ervi can defend against this threat by avoiding oviposition in the presence of H. axyridis chemical tracks. Aphidius ervi parasitized far fewer pea aphids Acyrthosiphon pisum (Harris) (Hemiptera: Aphididae) on plants [Vicia faba L. (Fabaceae)] with H. axyridis chemical tracks in a no‐choice environment at a single‐plant scale. Similarly, when parasitoids could move freely between plants with and without tracks, A. ervi parasitism was higher on plants without tracks. Behavioral observations of A. ervi foraging suggested that this might be because of reduced A. ervi attack rates and patch residence times in the presence of H. axyridis tracks. Despite a risk of intraguild predation by H. axyridis, our study suggests that A. ervi may be able to mitigate this risk by altering its behavior in response to chemical cues.