A team of researchers at the University of Vermont just upended a fundamental assumption about how species adapt to change. We've long thought of evolution as a one-way climb up a fitness peak—organisms adapt to their environment, reach a stable state, done. But new research shows evolution doesn't work that way at all. When environments keep shifting, populations don't just climb higher; they can climb in completely different directions, or sometimes stumble backward.
The study, led by biologist Melissa Pespeni and computational scientist Csenge Petak, used a powerful computer model to simulate thousands of generations of digital organisms facing different types of environmental fluctuations. What they found was striking: the same species, facing the same kinds of challenges, could end up with wildly different evolutionary outcomes depending on which challenges it faced first.
Take temperature swings versus rainfall cycles. A population cycling between hot and cold seasons might evolve better tolerance to both extremes—a genuine advantage. But a population cycling between wet and dry years might actually get worse at handling drought. Why? Because after a long rainy period, the population essentially "resets," losing the drought adaptations it had built up. It's like forgetting how to swim because you spent too long on land.
We're a new kind of news feed.
Regular news is designed to drain you. We're a non-profit built to restore you. Every story we publish is scored for impact, progress, and hope.
Start Your News Detox"The biggest takeaway," Pespeni says, "is that a population's history shapes how high it can climb and how hard the path is to get there." This matters more than we've realized. Scientists often study a single population in a controlled environment, then assume those findings apply to the whole species. But the Vermont team's work suggests that assumption is shaky. You can't understand a species' evolutionary potential by looking at just one starting point.
This has real implications for questions we're asking right now. Can species adapt fast enough to survive climate change? How quickly do bacteria develop antibiotic resistance? These aren't abstract questions—they affect human health and food security. But if we're only studying one population in one type of fluctuating environment, we're missing the full picture.
Interestingly, the findings also speak to artificial intelligence. AI systems often struggle with something called catastrophic forgetting—they learn a new task and forget the old one. "Evolution in nature and AI training follow remarkably similar patterns," says computer scientist Nick Cheney, a co-author. A growing field called online continual learning is trying to build AI systems that learn continuously across diverse tasks, much like organisms evolving in variable environments. The parallels are striking enough that evolutionary insights might help us build smarter, more adaptable machines.
At its heart, this study reveals something humbling: evolution isn't a universal algorithm. It's deeply shaped by history. Where you start, what you've already adapted to, and the specific sequence of challenges you face—these all matter enormously. Two populations of the same species can take completely different evolutionary paths. The implication is both unsettling and oddly hopeful. It means we can't predict evolution from first principles alone. But it also means the future is far more contingent, far more shaped by the particular choices and circumstances we create, than we thought.
