Ecology Enables Endless Evolution
A simple way to picture biological evolution is to imagine a landscape where each point on the surface represents a different combination of traits that an organism could have. The higher the point, the greater the organism’s fitness—its ability to survive and reproduce. Over many generations, populations traverse this fitness landscape, ascend peaks, and, in theory, come to rest. But evidence from the fossil record and from lab studies of microbial evolution suggests that evolution does not settle down; instead, it goes on without signs of stopping [1]. The typical explanation for this seemingly endless evolution is the so-called Red Queen hypothesis, in which a changing environment constantly reshapes the fitness landscape [2]. Specifically, once an organism gets well adapted, it becomes a prime target for parasites, forcing it to readapt in an evolutionary arms race [3]. But can anything other than host–parasite warfare fuel evolution forever? Now Aditya Mahadevan and Daniel Fisher at Stanford University have shown that ecological interactions among dozens of species can also create the ever-changing environment needed to drive evolution indefinitely [4].
Ecologists have spent decades studying how interactions between species shape highly diverse ecosystems (Fig. 1), but they have largely ignored evolution. Their justification seemed reasonable: While ecology operates quickly, evolution is slow. However, over a long enough time, evolution is inevitable [5]. Moreover, in microbial communities, high mutation rates and large population sizes make it possible to watch community evolution in action. Any theory describing microbial ecosystems must therefore account for evolution.
Yet, almost every attempt to model evolving ecological communities ran into the same problem: One organism, dubbed a Darwinian monster, evolves to be good at everything, killing diversity and collapsing the community [6]. Theorists circumvented this outcome by imposing metabolic trade-offs, essentially declaring that no species could excel at everything [7]. But that approach felt like cheating because the trade-offs in the models needed to be unreasonably strict. Moreover, for mathematical convenience, previous models assumed that ecological interactions between species were reciprocal: Species A affects species B in exactly the same way that B affects A. However, when interactions are reciprocal, community evolution ends up resembling the misleading fixed fitness landscape. Evolution is fast at first but eventually slows down and stops instead of going on endlessly.
Mahadevan and Fisher solved this puzzle by focusing on a previously neglected but ubiquitous aspect of ecological interactions: nonreciprocity. This feature occurs when the way species A affects species B differs from the way B affects A—for example, when two species compete for the same nutrient, but the competition harms one species more than the other. In their theoretical work, the researchers found that nonreciprocal interactions cause highly diverse communities to settle into a Red Queen phase, in which evolution continues seemingly forever without slowing down. In this phase, new mutants arise, establish themselves in the community, and drive other species to extinction. The entire community keeps continually changing, but the diversity stabilizes instead of being lost. The finding that nonreciprocity can generate distinct evolutionary phenomena resonates with other discoveries in physics over the past few years. In particular, scientists have shown that nonreciprocal interactions can induce exotic effects in systems ranging from mechanical metamaterials to neural networks [8, 9].
The reason such ecological interactions facilitate endless evolution is that organisms do not just experience a fixed external environment—they themselves actively reshape it. For example, when zebras graze grasses, they deplete resources and transform the environment for every other herbivore. Such collective environmental modifications drive most ecological interactions. They turn every ecosystem into a dynamic system where the actions of one species ripple through the entire community, in turn affecting that species itself. This dynamic environment, shaped by a community of organisms, is the basis for how evolution can proceed forever.
Mahadevan and Fisher also contribute to a long-debated question in ecology and evolution: Do communities assembled from many preexisting species (for example, those being repopulated in the gut microbiome after antibiotic treatment) look different from those that evolve gradually from a few species? Most scientists assumed the answer was no. After all, evolution increases species diversity, so the two approaches should yield similar ecosystems. The researchers show that this assumption is false: It matters whether diversity comes from the bottom up, as in evolution, or from the top down, as in ecological assembly. Compared with evolved communities, assembled ones have very different species abundance distributions—representations of how common or rare each species is. This finding is not just significant but testable, because ecologists can measure such distributions.
The new work provides the foundation for a general eco-evolutionary theory of diverse communities—one that does not require artificial trade-offs to maintain diversity. Theorists have chased this goal for decades, and Mahadevan and Fisher have now taken the first steps. But major questions remain. Notably, the researchers assumed that mutants bear little resemblance to their parents—a mathematically convenient but biologically unreasonable stipulation. What happens when mutants actually closely resemble their parents, as they do in nature? Will evolved communities become more diverse as mutants carve out new ecological niches when faced with increased competition? Or will these communities become less diverse as mutants directly outcompete their parents?
Lastly, in explaining how diversity can enable endless evolution, this work questions the idea of the fitness landscape itself, which has been central to evolutionary thinking. The study suggests that every time a species becomes better adapted to its environment, it changes the environment for others in the community. Moreover, the others’ actions affect how that species experiences the environment in the first place. In this way, species are not climbing a fixed fitness landscape but something akin to an undulating snowscape (Fig. 2). Each step that they take deforms the fitness landscape as if they were walking on snow, and over time there is no clear peak that they are headed toward. Evolution is thus an endless exploration across an ever-changing adaptive landscape.
References
- B. H. Good et al., “The dynamics of molecular evolution over 60,000 generations,” Nature 551, 45 (2017).
- L. Van Valen, “The Red Queen,” Am. Nat. 111, 809 (1977).
- L. T. Morran et al., “Running with the Red Queen: Host-parasite coevolution selects for biparental sex,” Science 333, 216 (2011).
- A. Mahadevan and D. S. Fisher, “Continual evolution in nonreciprocal ecological models,” PRX Life 3, 033008 (2025).
- R. Lallensack, “How warp-speed evolution is transforming ecology,” Nature 554, 19 (2018).
- L. Fant, “Eco-evolutionary dynamics lead to functionally robust and redundant communities,” arXiv:10.1101/2021.04.02.438173.
- A. Posfai et al., “Metabolic trade-offs promote diversity in a model ecosystem,” Phys. Rev. Lett. 118, 028103 (2017).
- M. Brandenbourger et al., “Non-reciprocal robotic metamaterials,” Nat. Commun. 10, 4608 (2019).
- M. Fruchart et al., “Non-reciprocal phase transitions,” Nature 592, 363 (2021).