The competitive exclusion principle is one of ecology's cleanest predictions: when two species compete for the same resource, the better competitor eventually wins. Complete overlap, complete replacement. Gause demonstrated it in 1934 with Paramecium in test tubes. The principle is mathematically elegant, experimentally robust, and empirically wrong — natural ecosystems maintain far more species than the number of distinct resources should allow.
Zanchetta et al. (arXiv 2502.13720, accepted in PRX Life) show that fluctuating metabolic strategies can resolve this. Their model lets species' resource uptake rates vary stochastically over time rather than remaining fixed. A species that's the best competitor right now might be mediocre in five minutes. The hierarchy exists at every instant but shuffles continuously.
The result is non-monotonic. At low fluctuation amplitude, competitive exclusion operates normally — the static hierarchy has time to grind out its verdict. At high amplitude, stochastic extinction dominates — populations can't maintain viable density when conditions swing too violently. At intermediate amplitude, biodiversity peaks. The fluctuations are strong enough to prevent any species from maintaining permanent dominance, but gentle enough that populations can track their changing fortunes without crashing.
The mechanism is niche differentiation in time. In classical ecology, coexistence requires spatial or functional separation — different habitats, different food sources, different activity periods. Temporal metabolic fluctuation creates an equivalent separation without any spatial or functional difference. Two species can have identical average strategies and still coexist, because they're never simultaneously at their average. The overlap is in the mean; the separation is in the moment.
The path integral formalism reveals a phase transition. Below a critical ratio of species to resources, the system supports diverse communities. Above it, all species go extinct regardless of fluctuation amplitude. The threshold is sharp — not a gradual decline but a collapse. There is a maximum number of species the system can carry, and noise sets that number higher than determinism does, but not infinitely higher.
What makes this structurally interesting: the noise isn't disrupting an otherwise stable system. It's creating structure that stability prevents. A perfectly stable ecosystem is a perfectly exclusionary one — the hierarchy has infinite time to enforce itself. Intermediate instability is the mechanism of coexistence, not the obstacle to it. The diversity isn't maintained despite the fluctuations. It's maintained by them.