Phenotypic plasticity segregates antagonistic species but undermines their coexistence

Environmental change often induces novel phenotypes in multiple cohabiting species, yet how such common-cue-induced plasticity shapes their coexistence is largely unknown. Testing this link requires disentangling inducing-environment effects from induced-phenotype effects, which remains elusive. We address this gap using two mutually antagonistic bacterial species, Staphylococcus aureus ( Sa ) and Pseudomonas aeruginosa ( Pa ). In habitual salinity, where both bacteria show unicellularity, Pa severely diminished Sa . In contrast, high salinity induced macroscopic cell clustering that spatially segregated Sa and Pa ( Sa : bottom-dwelling vertical filament; Pa : top-dwelling horizontal mat), and led to >10-fold higher coexistence than habitual salinity. Counterintuitively, physically inhibiting cell clustering (abolishing segregation) in high salinity led to a further >100-fold higher coexistence ( Sa:Pa ∷~1:1). This was because salinity-induced clustering was adaptive for Pa , but not for Sa : such adaptive plasticity enabled killer Pa to proliferate by monopolizing oxygen, which diminished Sa in a correlated manner, thereby weakening coexistence (mechanistically explained by diffusion-based killing). Hence, although the inducing environment promoted overall coexistence, the induced phenotype undermined it. Community-wide genomics revealed that the high level of coexistence achieved by abolishing segregation in high salinity allowed rapid coculture-specific evolution. Taken together, we uncover how plasticity can appear to promote antagonistic coexistence despite causally undermining it.