People have been running evolution experiments for 15,000 years
Nobody in ancient Mexico or Ice Age Europe said, “let us conduct a long-term selection experiment.” But that is exactly what happened.
Whenever people kept the seeds of the best plants, or bred the calmest animals, they decided who got to reproduce. Do that for generation after generation and the population changes. The animals and plants around you today are the results.
Scientists call this artificial selection. Almost everything on your dinner plate is evidence of it.
Dogs from wolves
Every dog alive, from a two-kilogram chihuahua to a great dane taller than a child, descends from wolf ancestors.
The oldest widely accepted dog remains come from Bonn-Oberkassel in Germany and are roughly 15,000 years old. That dog was buried alongside two people. Its bones and its DNA both separate it from local wolves, and there is evidence it was cared for through an illness. Genetic estimates push the start of domestication further back still, and the exact wolf population involved appears to be extinct.
Here is what makes dogs such a powerful teaching example. Registries do not agree on an exact count, because each one uses its own rules. The largest international body, the Fédération Cynologique Internationale, lists over 350 breeds. Whatever the number, the range of body shapes inside this one species is greater than in almost any other mammal.
Yet most of the breeds you can name were not shaped over 15,000 years. They were shaped over roughly the last 200, once dog shows and formal breed registries created strong, narrow selection.
Two centuries. That produced the difference between a chihuahua and a great dane. Nobody had to invent new genes to do it. Breeders simply picked, from existing variation, the individuals closest to what they wanted, and repeated the choice.
Corn from a grass you would walk past
Pick up an ear of corn. Now imagine its wild ancestor: a scrawny Mexican grass called teosinte, with many spindly branches and a tiny “ear” carrying about a dozen kernels. Each kernel is sealed inside a casing hard enough to be difficult to eat.
Genetic and archaeological evidence points to domestication beginning about 9,000 years ago in the Balsas River valley of southern Mexico.
The transformation is enormous. A modern corn ear carries hundreds of kernels, all naked and edible, on a single strong stalk. And a surprising amount of that change traces to a small number of genes with large effects.
Two are famous. The gene tb1, short for teosinte branched1, controls branching. Changes to it turned a bushy, many-branched grass into a plant that puts its energy into one main stalk. The gene tga1, teosinte glume architecture1, governs the hard casing. Changes to it left the kernels exposed.
Other genes with smaller effects contributed as well, so this was not a single magic switch. But it does show something important: a few changes at the right control points can reshape an organism dramatically.
One more detail matters. Modern corn cannot survive without us. Its kernels are locked onto a cob wrapped in husks, so they cannot scatter. We depend on corn, and corn now depends on farmers. Selection built a partnership.
Six vegetables, one plant
This one surprises almost everyone.
Cabbage, broccoli, cauliflower, kale, brussels sprouts and kohlrabi are not six species. They are all Brassica oleracea, a single species whose wild form grows on sea cliffs along the Mediterranean and the Atlantic coast of Europe.
Growers took that one plant and pushed on different parts of it:
- select for big leaves, and you get kale and collards
- select for a tight main bud, and you get cabbage
- select for swollen buds along the stem, and you get brussels sprouts
- select for a thick stem, and you get kohlrabi
- select for chunky flower stalks and buds, and you get broccoli
- select for a dense head of unopened flowers, and you get cauliflower
Domestication of this group goes back thousands of years. Nothing was added from outside. All that variation was latent in one wild plant, and different farmers pulled it in different directions.
If you want a picture of how one ancestral population can branch into very different-looking descendants, that vegetable aisle is the clearest one you will find.
The fox experiment: selection you can watch
Most artificial selection happened long ago, without records. In 1959, a Soviet geneticist named Dmitry Belyaev decided to run it deliberately, from the beginning, and write everything down.
At the Institute of Cytology and Genetics in Novosibirsk, Siberia, Belyaev’s team started with about 130 silver foxes from fur farms. These animals had lived near humans for generations and still snapped and panicked at the sight of one.
The team selected on exactly one thing: calmness toward people. Each generation, only the foxes least afraid of a human hand were allowed to breed. Nothing else was chosen for. Not ears, not color, not tails. His colleague Lyudmila Trut ran the project for decades after him.
Tameness increased, which is what they expected. What they did not expect was everything else that came with it. The tame line began showing floppy ears, curled tails, white patches in the coat, shorter and rounder snouts, and dog-like sounds. Biologists call this cluster the domestication syndrome.
Why does it happen? Because genes are not independent switches. The systems that control fear and stress also affect development more widely. Push on one, and other traits move too.
One honest caution. Some later researchers have argued that the founding foxes came from farms where they were already partly tame, and already carried some coat variation. So the size of the effect is debated. The core lesson survives that debate: selecting for one trait can change others you never asked for.
The same process, a different filter
Now the key point.
Natural selection needs three things: individuals differ, some of those differences are inherited, and some individuals leave more offspring than others.
Artificial selection needs exactly the same three things. Only the third one changes hands. Instead of the environment deciding who reproduces, a person decides.
That is the whole difference. Not a different mechanism. A different filter.
Charles Darwin knew this, and he used it. When he published On the Origin of Species in 1859, he did not open with finches or fossils. He opened with Chapter 1, “Variation under Domestication,” and spent it on pigeon breeders and farm animals.
The move was deliberate. His readers had seen breeders reshape pigeons within their own lifetimes. Nobody disputed that it worked. Darwin’s argument was: you already accept that selection can transform a population, because you have watched people do it. Now consider what happens across far longer stretches of time, with the environment doing the choosing.
Starting from something familiar made a strange idea reachable. It still does.
The costs are real
Artificial selection is not a happy ending in every case, and honesty about that matters.
Dogs pay a price. Many breeds passed through severe genetic bottlenecks. Breed standards require mating within a closed group, so inbreeding is built in. That raises the chance of inheriting two copies of a harmful gene variant, and it is a major reason particular breeds carry particular inherited disorders. Selecting hard for a flat face in pugs, bulldogs and French bulldogs produced brachycephalic obstructive airway syndrome: dogs that struggle to breathe, overheat easily, and sleep badly. The trait people chose was a look. The consequence was a welfare problem.
Crops pay a price too. Selecting for uniform, high-yielding fields means growing millions of nearly identical plants. Identical plants share identical weaknesses.
In 1970, a fungal disease called southern corn leaf blight swept the United States. Most hybrid seed that year had been produced using one particular genetic background, known as Texas male-sterile cytoplasm. A new form of the fungus attacked exactly that background. Roughly 15 percent of the national corn crop was lost in a single season. Seed producers switched away from that background, and the epidemic ended the following year.
Low genetic diversity is efficient right up until the moment something arrives that the whole population is vulnerable to. That is a lesson about evolution, not just farming: variation is what lets a population respond to a change nobody predicted.
What to take away
Artificial selection is evolution you can hold in your hand. A dog, an ear of corn, a head of broccoli.
None of it required a new mechanism. Populations vary. Some of that variation is inherited. Somebody or something decides who breeds. Repeat for enough generations and the descendants stop looking much like the ancestors.
Humans have been doing this for thousands of years, mostly by accident, and got wolves into chihuahuas and a scraggly grass into corn. The natural world has been doing the same thing, without a chooser, for far longer.
To see the process in the wild, read natural selection. To see change measured as it happens, read observed evolution.