Genetic drift

The basic idea

Imagine flipping a coin ten times. You might get exactly five heads and five tails, but often you will not. With only a few flips, chance can push the result well away from a fifty-fifty split.

Genetic drift works the same way. The next generation is not a perfect copy of the current one. It is a sample. Some individuals happen to leave more offspring than others for reasons that have nothing to do with their genes. A deer is struck by lightning. A nest is flooded. A plant’s seeds land on rock. By chance alone, some genetic variants become more common and others become less common.

Over many generations those small random changes add up. A variant can drift all the way to 100 percent, which biologists call fixation, or drift down to zero and disappear for good.

Why small populations drift faster

Chance matters more when numbers are small. A simple experiment with marbles shows why.

Picture a huge jar that is exactly half red and half blue. You draw 10 marbles to build the next generation. Here is roughly what happens.

• About one draw in three gives 7 or more of one color. That is a jump from 50 percent to 70 percent in a single step.
• About one draw in 500 gives all 10 the same color. The other color is gone in one generation.

Now draw 1,000 marbles instead. About 95 percent of the time you get between roughly 470 and 530 red. Getting 700 of one color is so unlikely that you could draw every second for the age of the universe and not expect to see it.

Real populations behave the same way. In a famous 1956 experiment, Peter Buri set up 107 small populations of fruit flies, each with 8 males and 8 females. Every population started with two versions of an eye-color gene at exactly 50 percent each. No version was favored. After 19 generations, 58 of the 107 populations had lost one version completely. Some lost one, some lost the other, by chance.

There is also a neat rule for brand new variants. In a population where a new mutation has no effect at all, its chance of eventually taking over equals its starting share. In a population of 10 animals, each carrying two copies of every gene, a single new copy is 1 of 20. So it has a 1 in 20 chance of reaching fixation. In a population of 1,000 it is 1 in 2,000.

Key idea: Drift happens in every population, because every generation is a sample. The smaller the population, the bigger the random swings.

You can try this yourself with the simulator below this article. Set a small population size, sample a few generations, and watch the blue allele wander. Then try a larger population and compare.

Drift ignores usefulness

This is what makes drift different from natural selection. In selection, a variant spreads because it affects survival or reproduction in a particular environment. In drift, a variant can spread even if it has no advantage at all.

Drift can even spread variants that are mildly harmful, if the population is small enough. Selection still acts at the same time. Both processes run together, and which one wins depends on how strong selection is compared with the size of the random swings. In a large population, even a small advantage usually beats chance. In a tiny population, chance can overpower a modest advantage.

Bottlenecks: elephant seals and cheetahs

A bottleneck happens when a population suddenly becomes much smaller. The survivors are only a sample of the earlier population. Variants that none of them carried are lost, even if the population later grows back.

Northern elephant seals

Northern elephant seals live along the Pacific coast of North America. In the 1800s they were hunted heavily for the oil in their blubber. By the 1890s only a tiny group survived, probably on Guadalupe Island off Mexico.

Protection in Mexico and the United States let them recover. Today there are well over 100,000 of them. But their genes still carry the scar.

In 1993, Rus Hoelzel and colleagues compared their DNA with that of southern elephant seals, which were also hunted but never fell below about 1,000 animals. In one stretch of mitochondrial DNA, northern elephant seals had just 2 types. Southern elephant seals had 23. The team’s computer models suggested the northern bottleneck was fewer than about 30 seals, and possibly fewer than 20.

Cheetahs

Cheetahs carry unusually little genetic variation for a wild cat. In 1985, Stephen O’Brien and colleagues tested this in a memorable way. They grafted small patches of skin between unrelated cheetahs. In most mammals, the immune system rejects skin from an unrelated animal. All 14 grafts in the cheetahs were accepted, because their immune genes were so similar.

The same paper described an outbreak of a coronavirus disease in a captive breeding colony in Oregon in 1983. The virus rarely causes serious illness in domestic cats, but it spread through the cheetahs and killed many of them. Low genetic variation may make a species more vulnerable to new diseases, though it is hard to prove that link in a single case.

Most researchers think cheetahs went through at least one severe bottleneck in the past, possibly near the end of the last Ice Age. The exact timing is still debated.

Founder effects: the Amish and Pingelap

A founder effect happens when a small group leaves and starts a new population. The founders carry only part of the original variation. Rare variants that a founder happens to carry can become common in the new group.

The Old Order Amish of Lancaster County

The Amish of Lancaster County, Pennsylvania, descend from a fairly small number of families who arrived in the 1700s. Most marry within the community.

Ellis-van Creveld syndrome is a rare inherited condition that causes short limbs, extra fingers, and often heart defects. In most populations it is extremely rare, around 1 in 60,000 births or fewer. Among the Lancaster County Amish it is far more common.

In the 1960s, the geneticist Victor McKusick and his colleagues traced the family trees of affected children. Nearly every case led back to one immigrant couple, Samuel King and his wife. At least one of them carried a copy of the variant. A chance arrival, multiplied over many generations in a closed community, made a rare condition common.

Pingelap atoll

Pingelap is a tiny coral atoll in Micronesia. Around 1775, a typhoon followed by famine reduced its population to about 20 survivors. One of them, the island’s ruler, is thought to have carried a variant causing achromatopsia, a complete form of color blindness with poor vision in bright light.

Today several percent of Pingelapese people have the condition, with estimates commonly given between about 5 and 10 percent. In most of the world it affects roughly 1 in 30,000 people. In 2000, Olof Sundin and colleagues identified the cause as a mutation in a gene called CNGB3. The neurologist Oliver Sacks wrote about the island in his 1996 book The Island of the Colorblind.

Pingelap is technically both a bottleneck and a founder event. A tiny sample of people rebuilt the whole population.

Neutral theory: drift at the DNA level

In 1968 the Japanese geneticist Motoo Kimura proposed an idea that surprised many biologists. He argued that most of the changes we see when we compare DNA or proteins between species are neutral. They neither help nor hurt. They spread by drift, not by selection. Jack King and Thomas Jukes made a similar argument in 1969.

This is called the neutral theory of molecular evolution. It does not say selection is unimportant for bodies and behavior. It says that at the level of individual DNA letters, much of the change is chance.

The theory makes a useful prediction. For neutral changes, the rate at which they build up over time equals the mutation rate, no matter how big the population is. That steady build-up is one reason scientists can use DNA differences as a rough molecular clock to estimate when species split.

How much of the genome is truly neutral is still argued about. Most biologists today accept that drift explains a large share of molecular change, while selection shapes the parts that matter for function.

Drift vs natural selection

Why drift matters

Drift often removes genetic variation. When variants disappear by chance and are not replaced by new mutations, a population becomes more uniform. That matters in conservation. A species with little variation may have fewer options when a new disease or climate shift arrives. Conservation programs for small populations often try to keep as many individuals breeding as possible, and sometimes move animals between groups to add variation back.

Drift can also push separated populations apart. If two small groups are isolated, random changes will pull them in different directions over time. Combined with mutation and selection, this can contribute to speciation.

Common misunderstandings

• Genetic drift is real evolution, because the population’s genetic makeup changes over generations.
• Drift does not require a trait to be good or bad.
• Drift is not rare. It affects every population, though its effects are easiest to see in small ones.
• Drift is not the same as mutation. Mutation creates new variants. Drift changes how common existing variants are.

Drift reminds us that evolution is not only a story of adaptation. Some features of populations reflect chance, history, and sample size as much as usefulness.

Where to go next

To compare drift with the process that does produce adaptations, read natural selection. To see where new variants come from, read mutation. For more hands-on models, visit the simulations page.