Evolutionary theory makes dozens of specific, testable predictions about the natural world, from the structure of DNA to the distribution of species on islands. These predictions span genetics, anatomy, the fossil record, and observable changes in living populations. Many were made long before the technology existed to verify them, and the fact that they’ve held up is a major reason evolutionary biology is considered one of the strongest frameworks in science.
All Living Things Share a Common Ancestor
One of the broadest predictions is that every organism on Earth descends from a shared ancestor. If true, you’d expect to find a universal system for storing and reading genetic information, and that’s exactly what exists. The genetic code, the mapping of 64 DNA “words” to 20 amino acids plus start and stop signals, is shared by virtually all known life. It was first deciphered in bacteria and then found, with only minor variations, in everything from archaea to plants to humans. The main features of this code appear not to have changed since the Last Universal Common Ancestor, or LUCA, the single-celled organism from which all modern cellular life descends.
This universality isn’t just a curiosity. It’s a prediction fulfilled. If life had originated multiple times independently, you’d expect fundamentally different coding systems in different lineages. Instead, the code is essentially frozen in place, because any major change would scramble the proteins an organism needs to survive.
Transitional Forms in the Fossil Record
If species gradually changed over time, the fossil record should contain intermediate forms, organisms that bridge the gap between an ancestral body plan and a modern one. Evolutionary theory doesn’t just predict that these exist in general; it predicts specific features they should have.
Whale evolution is a textbook example. Modern whales breathe through blowholes on top of their heads, while their earliest land-dwelling ancestors (pakicetids) had nostrils at the tips of their snouts. The theory predicts an intermediate stage with nostrils partway up the skull. Fossils of Aetiocetus confirmed exactly that, with nostrils positioned in the middle of the skull. Horse evolution follows the same logic: if modern single-toed horses descended from four-toed ancestors, there should be three-toed intermediates. Fossils of Archaeohippus and Parahippus fill that gap precisely.
These aren’t cherry-picked examples. The pattern repeats across fish-to-tetrapod transitions, dinosaur-to-bird transitions, and the emergence of mammals from reptilian ancestors. In each case, the predicted intermediates were found after the prediction was made.
Island Species Should Follow Predictable Patterns
Evolutionary theory predicts that species on islands should look different from their mainland relatives, and that the degree of difference should follow geographic rules. Populations on distant islands should be more genetically divergent from their mainland source populations than those on nearby islands. Populations on large islands should also be more divergent than those on small islands, because larger islands support bigger populations with more opportunity for genetic change over time.
Published observations confirm both predictions. A larger proportion of species found nowhere else on Earth (endemic species) exist on large, remote islands compared to small, nearby ones. The Galápagos finches Darwin studied are a famous case, but the pattern holds globally, from Hawaiian honeycreepers to Madagascar’s lemurs.
DNA Should Record Evolutionary History
If species diverge from common ancestors, their DNA should function like a historical document. Closely related species should have more similar genomes than distantly related ones, and the number of genetic differences between two species should be roughly proportional to how long ago they split apart.
This idea, called the molecular clock, was formalized in the 1960s. The core logic: neutral mutations (changes in DNA that don’t help or harm the organism) accumulate at a relatively steady rate. By counting those differences and calibrating against known fossil dates, researchers can estimate when two lineages diverged. The original assumption of a perfectly constant rate turned out to be too simple. Mutation rates vary between organisms, so modern “relaxed” molecular clocks allow the rate to fluctuate within limits. Even so, the fundamental prediction holds: genetic distance tracks evolutionary time.
Human Chromosome 2
One of the most striking genomic predictions involves human chromosomes. Humans have 46 chromosomes, while chimpanzees, gorillas, and orangutans all have 48. If humans and great apes share a common ancestor, two ancestral chromosomes must have fused into one somewhere in the human lineage. The theory predicts that one human chromosome should show clear evidence of this fusion: remnants of the original chromosome tips (telomeres) stuck in the middle, and a second, deactivated centromere.
Human chromosome 2 has all of this. At the fusion site in the region called 2q13-2q14.1, two head-to-head arrays of degenerate telomere repeats sit where the ancestral chromosome ends joined together. One of the two original centromeres was inactivated. The evidence reads like a receipt for a specific evolutionary event.
Vestigial Structures Should Exist
If organisms evolve from ancestors with different body plans and lifestyles, some structures should persist even after they’ve lost their original purpose. Evolutionary theory predicts these “vestigial” features, and the human body has several.
The human appendix is the most commonly cited example. In herbivorous ancestors, a larger version of this structure likely helped digest plant material. In humans, it’s a small pouch with no essential digestive role. Goosebumps are another: the reflex that raises tiny hairs on your skin served an obvious function in fur-covered ancestors, making them look larger to scare off predators. In mostly hairless humans, the reflex remains but achieves nothing visible. Wisdom teeth, the muscles some people can use to wiggle their ears, and the tailbone (coccyx) all follow the same pattern. Each one makes sense only if the human body carries the architectural legacy of ancestors who needed those features.
Evolution Should Be Observable in Real Time
If natural selection drives change, you should be able to watch it happen in organisms with short generation times. Bacteria are the clearest test case, and they deliver exactly what the theory predicts.
In a 2025 study using E. coli, researchers exposed bacterial populations to low concentrations of various antibiotics over 60 days, roughly 120 generations. The bacteria evolved significant resistance to every antibiotic tested. Against ciprofloxacin and kanamycin, resistance increased 256-fold. This isn’t a vague trend. It’s a massive, measurable change driven by selection pressure in a controlled environment, playing out over weeks instead of millennia.
The same study also tested whether bacteria could evolve resistance to antimicrobial peptides (a different class of germ-killing molecules) at the same rate. They largely couldn’t. The rate and degree of resistance to conventional antibiotics was significantly stronger, which itself is a testable evolutionary prediction: the ease of evolving resistance depends on how many genetic pathways are available to get around a particular threat.
Coevolution Between Species
Evolutionary theory predicts that when two species depend on each other, they should evolve in tandem, each shaping the other’s traits. One of the most famous predictions in all of biology came from this logic. In 1862, Charles Darwin received a specimen of the Madagascar orchid Angraecum sesquipedale, which stores its nectar at the bottom of a spur nearly 30 centimeters long. He predicted that a moth with a tongue long enough to reach the nectar must exist, because the orchid’s reproduction depended on it.
The prediction was ridiculed at the time. Decades later, a hawk moth with a proboscis matching the orchid’s spur was discovered and eventually given the subspecies name “praedicta,” Latin for “the predicted one.” Observations later confirmed that this moth does visit the orchid and remove its pollen, proving Darwin’s reasoning correct. The prediction worked because coevolution is not optional under natural selection: if a plant’s reproduction depends on a pollinator, both species are locked into an evolutionary arms race that shapes their anatomy.
Stronger Selection Produces More Predictable Outcomes
A subtler prediction concerns how reliably scientists can forecast evolutionary change. The theory predicts that when selection pressure on a trait is strong and the genetic basis of that trait is well understood, evolutionary outcomes become more predictable. When selection is weak or many genes with small effects are involved, prediction gets harder.
Recent work in comparative genomics supports this. Accurate predictions about how a trait will change are more common when more is known about the genetic basis of the variation, when the trait was strongly affected by positive selection, and when the genes with the largest effects have been identified. Gene-environment interactions complicate things further, because the same gene can produce different traits in different environmental contexts. This means evolutionary theory doesn’t claim to predict everything, but it does specify the conditions under which its predictions should be strongest, and those conditions check out.