Why Did It Take Evolutionary Biology 60 Years To Catch Up To Its Own Evidence?

Group Selection Was Debated For Decades. A New Review Says The Empirical Case Is Far Stronger Than Critics Claimed.

For decades, many evolutionary biologists argued that natural selection acts primarily on individuals. In other words, that organisms are rewarded or penalized one at a time, and that the idea of entire groups being shaped by evolutionary pressure is, at best, naive. For decades, the concept of group selection was treated like a bad idea that serious scientists had already put to rest.

Now, the authors of a new review argue that the dismissal was never justified. Researchers combed through nearly 3,000 peer-reviewed papers and identified 280 studies providing direct empirical evidence that natural selection operates not just on individual organisms but on groups, colonies, communities, and even genetic elements at the same time. The evidence stretches back to 1976. It covers organisms from viruses and bacteria to ants, chickens, chipmunks, and humans. It was sitting in the scientific literature the entire time the debate raged.

“Contrary to common notions, we found solid empirical support for the utility and importance of MLS in explaining natural selection and evolution,” the authors write.

Multilevel selection, or MLS, holds that natural selection can act at several levels of biological organization at once: on genes, on individual organisms, and on groups of organisms. A group of cooperating individuals might outcompete a group of more selfish ones even when selfish individuals are doing better within their own group. When that happens, selection is operating on the group itself. The idea was largely pushed out of mainstream evolutionary biology in the 1960s, when influential scientists argued it was theoretically implausible. The authors of the new review contend that those early critiques focused on theoretical plausibility, and that empirical evidence kept accumulating alongside the debate.

How Group Selection Got Buried by Evolutionary Biology

What the new review makes clear is just how lopsided the conversation became. Of the 1,829 papers that made it past the initial screening, 1,415 (77%) consisted of mathematical models, simulations, opinion pieces, and conceptual debates rather than actual experiments or field observations. Much of the published debate focused on theoretical models and conceptual arguments rather than actual experiments or field observations.

Research on MLS has accelerated sharply since 2012. Of the 280 qualifying studies, 199 were published in that final 12-year stretch. Only 81 appeared in the first 35 years of MLS empirical research combined. The sharp increase since 2012 points to growing interest in MLS as a legitimate research framework, raising questions about why it took so long.

What 280 Studies Reveal About Multilevel Selection

To build their case, researchers César Marín, Anne B. Clark, Conner S. Philson, Omar Tonsi Eldakar, and Michael J. Wade systematically combed through the scientific literature, screening nearly 3,000 journal articles published between 1900 and 2024. After filtering out theoretical work and studies that only flagged group selection as a possible explanation rather than a demonstrated one, they arrived at 280 papers with direct empirical support. About two-thirds (180) were controlled laboratory experiments; the remaining 100 measured selection in natural populations in the wild.

Some of the most persuasive evidence comes from practical applications never framed as tests of controversial theory. A 1996 poultry breeding experiment found that selecting hens based on how their entire cage performed, rather than pulling the best individual egg-layers, transformed the flock. By the sixth generation, the results were hard to argue with. Mortality from hen-on-hen aggression dropped from 68% down to 8.8%, and egg production per hen climbed from 91 to 237 eggs. Farmers, it turns out, had been running group selection experiments long before the theoretical debate was settled.

At the other end of the size scale, laboratory researchers working with snowflake yeast ran 600 rounds of artificial selection for larger group size. Within roughly a year, yeast in one treatment group had become visible to the naked eye (20,000 times larger than at the start) while maintaining a clonal, multicellular life cycle. One of evolution’s most ancient transitions, the move from single-celled to multicellular life, replayed inside a laboratory flask.

The Technique That Makes Group Selection Detectable

A key tool in the empirical case for MLS is contextual analysis, a statistical method that separates the strength and direction of selection acting on individuals from selection acting on groups: simultaneously, in the same population. The review identified 26 studies using this approach across a wide range of wild species, from jewel flowers and sea rockets to chipmunks and great tits.

What those 26 studies show is that group-level and individual-level selection do not always point the same way. In some populations, they reinforce each other. In others, they push in opposite directions. In a study of ants, group selection was stronger and moved traits against the direction of individual selection. In a study of sea rocket plants, the two levels moved in concert. No single pattern dominated, which, the authors note, is precisely what MLS theory predicts. Selection at multiple levels has to be evaluated case by case, not assumed away.

The review, published in Frontiers in Ecology and Evolution, also documents strong acceptance of MLS in anthropology and the social sciences. A survey of 175 evolutionary anthropologists found that 78.7% considered cultural multilevel selection (the idea that groups can be selected for culturally transmitted traits, not just genetic ones) to be important. A separate analysis of historical human societies found that major patterns in cultural evolution, including warfare, competition, and the spread of cooperative norms, align with what MLS theory predicts.

Undergraduate biology textbooks, meanwhile, still largely dismiss or minimize the concept. Recent analyses of biology curricula found that MLS is consistently sidelined relative to individual-level selection. That means many students are still taught evolution primarily through an individual-level lens.

The authors are not diplomatic about what needs to change: “Disregarding MLS will continue to hold the field of evolutionary biology back and prevent us from more fully understanding life on earth.”

This review argues that the empirical case for multilevel selection is far stronger than critics have acknowledged, and the field of evolutionary biology has more catching up to do than most textbooks suggest.

---

Disclaimer: The findings discussed in this article are based on a single bibliometric review. While the study draws on a broad body of existing research, science is an ongoing process and expert perspectives on multilevel selection continue to evolve. This article is intended for general informational purposes and does not constitute scientific or professional advice.

---