Abundant empirical evidence of multilevel selection revealed by a bibliometric review

Natural selection is based on the notion of differential reproduction between entities, often characterized as a struggle between individual organisms. However, natural selection can act at all levels of biological organization, thus being termed ‘multilevel selection’ (MLS). A common misconception is that MLS lacks empirical support. To address this, we conducted a bibliometric review of 2,950 Web of Science/Scopus-indexed scientific articles. Our goal was documenting the range of taxa/systems, levels, and research topics/tools where MLS has been used to understand natural selection across levels. We found 280 studies providing empirical support for MLS: 100 were performed in situ, 180 were laboratory experiments. The studies span a vast range of organisms, from viruses to humans and eusocial insects. While 90.4% of studies focused on some form of organismal group (demes, colonies, aggregates), the remaining 9.6% explored selection at other levels (communities, cells, nuclei). We classified these 280 studies into research categories such as artificial selection, breeding through group selection, indirect/social genetic effects, and contextual analysis, among others. In contextual analysis studies, the strength of selection was comparable across levels. Contrary to common notions, there is solid empirical support for the utility and importance of MLS in explaining natural selection and evolution.