Assessing ground predation by invertebrates in crops: Camera observations clarify the relevance of aphid cards and coleopteran sentinel prey

International audienceRedesigning cropping systems to promote natural pest control requires a sound understanding of the regulation potential of natural enemies. Sentinel preys are useful to quantify biological control, but the use of aphid cards (aphids glued to a sandpaper card) is controversial for their lack of realism. The objectives here were: (1) to compare aphid cards to non-fixed beetle larvae resembling coleopteran pests, with respect to the consumers attracted; and (2) to ascertain whether predation on coleopteran prey larvae responds to different invertebrates communities. We used cameras to identify the consumers of two types of sentinel prey on the ground, and compared aphid cards to an original sentinel prey, bruchid beetle larvae. Aphids glued to predation cards attracted opportunistic consumers and scavengers with very long handling times, and even scavengers, including harvestmen (33% of consumption events), spiders and slugs, and a very few carabid beetles. Unattached bruchid beetle larvae were consumed largely by carabids (64%), and secondarily by rove beetles, ants and chilopods. As the sentinel prey were not consumed by the same organisms, these findings demonstrate the importance of the choice of organisms exposed to estimate pest predation. Knowing the identity of the bruchid larvae consumers from the images, we investigated whether the assemblages of epigeic arthropods could explain their predation rates under contrasting field conditions. Accounting for all the expected consumers, sentinel prey disappearance rates could be explained statistically with a satisfactory goodness-of-fit. Slugs and ants contributed more strongly to the disappearance of sentinel prey than carnivorous carabids and chilopods. Finally, alternative prey (especially Collembola) reduced consumption by diverting consumers from their usual prey. Camera observations were valuable to assess the relevance of two different sentinel preys and of the consumers involved. With this knowledge, it is possible to estimate the pest control potential from the ground-dwelling communities present