Résumé
Service plants are primarily used in agroecosystems to provide ecosystem services that are not directly marketable. They are a promising option to promote biological pest regulation. Past studies have demonstrated their usefulness for regulating one pest category (either pathogens/parasites, herbivores or weeds). However, a multi-pest view of the role of service plants, including the potential disservices (negative impacts) that they may generate, is lacking. Such an overview is essential to meet the challenge of agroecology. This paper aims to fill this gap. Here, a trait-based approach was used to provide an overview of the potentialities of service plants, (inserted either in intercropping, in rotation with the crops, or in field edges) for regulating multiple pests, while limiting disservices. For that purpose, we first laid the foundation of a conceptual framework by synthesizing the mechanisms and service plant traits involved in the regulation of each pest category and in the mitigation of each disservice. On this basis, we analyzed (1) the compatibility in the regulation of the different pests by service plants, and (2) the compatibility between multi-pest regulation vs disservice mitigation. Our main conclusions are: (1) Despite knowledge gaps, there is good potential of service plants for multi-pest regulation; (2) The challenge lies at least as much to mitigate disservices that service plants may cause as to promote multi-pest regulation; (3) The level of incompatibility between promoting multi-pest regulation vs mitigating disservices varies with the mode of insertion of service plants, increasing with interactions with crop plants. This review shows how a trait-based approach can be used to synthesize knowledge from different disciplines and provides a tool for cross-disciplinary dialogue. It identifies priority research actions that are needed to increase synergy, genericity and adaptation of service plants to local conditions, and provides foundations for the design of service-plant based agroecosystems.