A sweeping new review of five decades of scientific literature finds robust evidence that natural selection operates simultaneously across multiple levels of life, from genes to groups to entire communities, and not just on individual organisms.
The study, co-authored by Michael Wade, Distinguished Professor Emeritus of Biology in the College of Arts and Sciences at Indiana University Bloomington, addresses a stubborn misconception that has lingered in the field, namely, that natural selection acts almost exclusively on individual organisms competing for survival and reproduction.
Distinguished Professor Emeritus, Biology, Michael Wade
The landmark new study, led by researchers from Chile, the Netherlands and the United States and published the journal Frontiers in Ecology and Evolution, gathers scholarly evidence to refute this view.
Employing a systematic bibliometric review, that is, a rigorous, structured survey of published scientific literature, the study authors identified 280 peer-reviewed studies spanning nearly 50 years that document "multilevel selection" (MLS) at work in nature and in the laboratory. Multilevel selection is the phenomenon in which natural selection operates simultaneously at two or more levels of biological organization, meaning, MLS shapes not only individual organisms' evolution, but also the groups, colonies, or communities to which they belong.
The findings carry implications well beyond academic debate, touching fields as varied as cancer research and agriculture, and may shape society's fundamental understanding of how life on Earth evolves.
"The paper is significant because it surveys field and experimental research around the world on multilevel selection, based on experimental work that I've done, as well as the ensuing research of my students, and those influenced by this research," said Professor Wade.
The review traces its intellectual roots to Wade's own laboratory experiments with flour beetles in the 1970s. By controlling which groups of beetles were able to reproduce, he found that the populations changed quickly and substantially within just a few generations. This presented early evidence that selection acting on groups, not just on individuals, was a real and measurable phenomenon.
This new study traces the explosion of research that followed. With lead author César Marín of Universidad Santo Tomás in Valdivia, Chile, the research team searched databases covering scientific literature from 1900 through 2024, ultimately screening nearly 3,000 articles. The team was deliberately strict about what they counted. They excluded theoretical models, opinion pieces, and any study that merely suggested group selection as a possible explanation, keeping only studies with direct, observable evidence. That rigorous filtering left 280 studies that documented natural selection operating at more than one level of biological organization at the same time.
This included research spans a remarkable breadth of life. Organisms studied include viruses, bacteria, fungi, social insects such as ants and bees, wild birds, mammals, crop plants, and humans. About 90 percent of the studies examined how natural selection impacts groups of organisms, including animal colonies, local wildlife populations, or clusters of single-celled organisms living together. The other 10 percent looked beyond that, at selection acting on individual cells, genes, or entire communities composed of multiple species.
Wade points to two domains in particular where multilevel selection has already changed how scientists and practitioners approach real-world problems.
"On the applied side, for several decades now, cancer biologists have described the origin and spread of malignant tumors as a multilevel selection process acting among cells in the human body," he explained. "Moreover, animal breeders have turned to multilevel selection to increase things like egg yield and diminish feather pecking, a revolution in the way domestic animals and plants are reared."
That agricultural application is one of the study's more striking examples. For example, a series of studies beginning in the 1990s by William Muir, a professor of animal sciences at Purdue University, showed that when egg-laying hens were selected as groups for breeding, meaning, the most productive cages for egg-laying rather than the most productive individual birds, the results were dramatic. Mortality from aggressive pecking plummeted from 68 percent in early generations to under nine percent by the sixth generation, while egg production more than doubled.
Also striking were experiments with yeast where, after 600 rounds of artificial selection for larger group size, the yeast responded by evolving into clusters visible to the naked eye, mirroring one of the most ancient transitions in the history of life: the leap from single-celled to multi-celled organisms.
The new review also found that MLS research has accelerated sharply in recent years. Of the 280 qualifying studies, 199 were published between 2012 and 2024, suggesting the framework is gaining traction across disciplines.
Still, the authors acknowledge limits. Publication bias likely means some studies finding no group-level effects went unpublished and uncounted. Moreover, the 280-study total is almost certainly an undercount, in that many studies whose results are consistent with multilevel selection, the authors note, avoid using the terminology altogether, including, sometimes, out of deference to historical skepticism.
But that skepticism, the study authors assert, has outlived its justification. This scholarship points toward a future in which researchers study multilevel selection as a standard analytical lens, rather than through the prism of controversy, applicable wherever organisms live, compete, and reproduce in social groups. With new molecular tools, genomic methods, and long-term field datasets now available, the authors suggest that many more ground-breaking discoveries are likely in the coming decades.
The research was funded in part by Chile's National Research and Development Agency.
