The evolutionary stability of antagonistic plant facilitation across environmental gradients

O_LIPlant interactions, understood as the net effect of an individual on the fitness of a neighbor, vary in strength and can shift from negative to positive as the environmental conditions change in time and space. Evolutionary theory questions the stability of non-reciprocal interactions in which one plant has a positive net effect on a neighbor, which in return has a negative net impact on its benefactor. This type of interaction is known as antagonistic facilitation. C_LIO_LIWe develop a spatially explicit consumer-resource model for below-ground plant competition, including plants able to mine resources and make them available for any other plant in the community, termed ecosystem engineers. We use the model to assess whether and under which environmental conditions antagonistic facilitation via soil resource engineering is evolutionarily stable. C_LIO_LIWe find that antagonistic facilitation is stable in highly stressful conditions, which supports the theory of ecosystem engineers as drivers of primary succession and provides a theoretical ground to investigate facilitation mechanistically in the context of the stress gradient hypothesis. C_LIO_LIAmong all potential causes of stress considered in the model, the key environmental parameter driving changes in the interaction between plants is the proportion of the limiting resource available to plants without mining. This finding represents a challenge for empirical studies, which usually measure the resource input or loss in the system as a proxy for stress. We also find that the total root biomass and its spatial allocation through the root system, often used to measure the nature of the interaction between plants, do not predict facilitation reliably. C_LIO_LISynthesis. Antagonistic facilitation established between an ecosystem engineer nurse plant and neighbor opportunistic individuals can be evolutionarily stable in stressful environments where ecosystem engineers self-benefits from mining resources outweigh the competition with opportunistic neighbors. These results align with theories of primary succession and the stress gradient hypothesis as they show that antagonistic facilitation is stable under environmental stress, but it evolves into mutual interference in milder environments. However, using inaccurate parameters to measure facilitation and stress gradients in empirical studies might mask these patterns. C_LI