A bright red band on a butterfly’s wing is a warning: eat me and you’ll regret it. That same red band shows up on dozens of butterfly species scattered across Central and South America, and on at least one moth whose lineage split from butterflies at an estimated divergence of roughly 120 million years ago. Now, a peer-reviewed study published in PLOS Biology in 2025 reveals that all of these species arrived at their warning colors by tweaking the same two genes, called ivory and optix. The finding suggests that evolution does not wander freely through an unlimited space of genetic possibilities. Instead, it returns to the same small toolkit, over and over, across vast stretches of time.
Two genes, dozens of species, an estimated 120-million-year divergence
The research team compared the genomes of multiple neotropical butterfly lineages whose common ancestors are estimated to have diverged between roughly one million and 120 million years ago, based on molecular clock analyses rather than direct observation. Despite that enormous gap, the bold red and yellow warning bands that signal toxicity to predators consistently traced back to the same two genetic addresses. The changes were not in the protein-coding portions of ivory and optix themselves but in nearby noncoding regulatory regions, the molecular switches that control when and where those genes turn on. Evolution, in other words, was not inventing new hardware. It was reprogramming existing equipment.
The gene optix was already known to play this role. A 2011 study published in Science showed that optix repeatedly drove wing-pattern mimicry across Heliconius butterfly species, establishing that a single gene could be reused by natural selection in distantly related lineages. The new study builds on that foundation by confirming ivory as a second hotspot and by stretching the documented timeline to an estimated 120 million years, a span covering most of the order Lepidoptera.
The ivory signal maps to or near the region also known as the cortex locus; the two names may refer to the same or closely overlapping genomic region, and the precise relationship between them is still being clarified. The cortex locus was previously identified as a major wing-patterning region across both butterflies and moths in a study published in Nature. Separate work on enhancer shuffling, also in PLOS Biology, has demonstrated that rearranging these regulatory switches can generate novel color patterns without creating new genes. Together, these findings from independent research groups, using different species and different methods, keep pointing to the same small set of genetic control points.
A moth locked into the same pattern
The study’s most striking extension goes beyond butterflies entirely. A day-flying moth that mimics toxic butterflies shows a similar genetic signature: its color variation also maps to regulatory regions near ivory and optix. Supplementary genomic data from the same PLOS Biology study, archived in PubMed Central, describe a roughly 1.018-megabase chromosomal inversion in the moth near the ivory locus. Chromosomal inversions act like molecular locks: they suppress recombination and keep clusters of adaptive gene variants inherited together as a package.
That architecture mirrors what earlier research documented in Heliconius numata butterflies, where inversions maintain a supergene controlling multiple mimicry traits at once. The parallel is suggestive, but the moth’s inversion has not yet been tested with direct gene-editing experiments. The evidence so far is structural and statistical, not functional. Proving that the moth’s inversion works through the same molecular mechanism as the butterfly supergene will require further experimental work.
What this actually says about randomness
The headline claim, that this changes what “random” means in evolution, needs careful unpacking. Mutations themselves remain random in the sense that they do not arise because an organism “needs” them. What the study challenges is the assumption that evolutionary outcomes are equally random. If dozens of species, separated by an estimated divergence of up to 120 million years, independently land on the same two regulatory regions to solve the same adaptive problem, then the menu of likely genetic solutions is far shorter than classical models assumed.
Think of it this way: mutations can strike anywhere in the genome, but when they happen to hit the regulatory switches near ivory or optix, natural selection strongly favors them because those switches sit at control points for wing pigmentation. The result is that adaptation gets funneled through a narrow set of genetic pathways. Predicting which changes will spread through a population becomes more tractable than a purely random model would suggest.
This pattern is not unique to butterflies. In three-spined sticklebacks, repeated loss of pelvic spines across independent freshwater populations traces to regulatory changes near the Pitx1 gene. In vertebrates more broadly, the Mc1r gene has been linked to pigmentation shifts in species as different as mice, birds, and lizards. The Lepidoptera study adds to a growing body of evidence that evolution’s toolkit, while not deterministic, is far more constrained than once thought.
Where the boundaries are
The study’s scope is limited to Lepidoptera, and the authors do not claim their findings apply directly to warning coloration in poison frogs, coral snakes, or other aposematic animals. Extending the “limited toolkit” conclusion to those groups would require separate genomic work. The interpretive step from “the same genes keep showing up in butterflies and moths” to “evolution is broadly constrained” is reasonable but still bounded by the taxa and traits studied.
It is also worth noting that no direct quotes from the lead researchers about redefining randomness appear in the primary peer-reviewed text. The framing around what “random” means is an inference drawn from the data’s pattern of repeated convergence, not a claim the authors themselves have made in sweeping terms. That inference is well supported by the evidence, but readers should understand it as a logical extension of the findings rather than a pronouncement from the research team.
Why repeated genetic convergence reshapes evolutionary prediction
If evolution repeatedly channels adaptation through a small number of genetic control points, the implications reach well beyond entomology. For conservation biologists, it could mean that predicting how species will adapt to environmental change, or fail to, becomes more feasible. For medical geneticists studying how pathogens evolve drug resistance, the principle that certain genetic pathways are disproportionately favored by selection is already a working assumption; the Lepidoptera data offer one of the clearest demonstrations of that principle operating across deep evolutionary time.
As of June 2026, the core findings rest on strong, multiply corroborated genomic evidence from independent teams working over more than a decade. The functional details of the moth’s chromosomal inversion remain an open question, and the broader claim about constrained randomness will sharpen or soften as researchers test whether the same pattern holds in other animal groups. For now, the data point in a clear direction: when it comes to building bright warning colors, evolution does not start from scratch. It reaches for the same switches it has used for an estimated 120 million years.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
