Experimental evolution
Based on Wikipedia: Experimental evolution
In February 1988, a biologist named Richard Lenski placed twelve flasks of Escherichia coli bacteria in an incubator and began what would become the longest-running evolution experiment in history. More than sixty thousand generations later—the equivalent of over a million years of human time—those bacteria are still evolving, still surprising researchers, and still teaching us things Darwin could only dream of observing directly.
This is experimental evolution: the practice of watching evolution happen in real time, under controlled conditions, in a laboratory or carefully managed field setting.
Why Watch Evolution in a Flask?
Darwin figured out how evolution works by examining its products—finch beaks shaped by island environments, orchids sculpted to attract specific pollinators, the branching tree of life visible in fossils and anatomy. But he never got to watch it happen. Evolution, as traditionally conceived, occurs over timescales that mock human impatience. We live our entire lives in what amounts to a single frame of the evolutionary movie.
Bacteria don't have this problem. E. coli can produce a new generation every twenty minutes under ideal conditions. In a single human day, bacteria can rack up more generations than humans have experienced since we diverged from our common ancestor with chimpanzees. This compression of evolutionary time means that what would require millions of years to observe in elephants or oak trees can unfold in months or years in microorganisms.
But speed isn't the only advantage. Laboratory evolution lets researchers control everything—the food source, the temperature, the presence of competitors or predators, the population size. They can freeze samples at any point, creating a living fossil record that can be thawed and revived to compete against descendants from the future. They can replay evolution from the same starting point, asking whether it takes the same path twice or finds different solutions to the same problem.
The Accidental Pioneer
The first person to deliberately evolve organisms in a laboratory wasn't a professional scientist but a Methodist minister with a passion for microscopy. William Dallinger spent seven years, from 1880 to 1886, cultivating single-celled organisms in a custom-built incubator that he could heat with extraordinary precision. He started his cultures at 60 degrees Fahrenheit—a comfortable room temperature—and began a patient campaign of thermal torture.
The key was gradualism. Dallinger raised the temperature so slowly that the organisms had time to adapt. At 73 degrees Fahrenheit, his cultures showed obvious distress—the microscopic equivalent of visible suffering. But he kept going, inching upward over months and years. By the end of his experiment, his organisms thrived at 158 degrees Fahrenheit, a temperature that would have killed their ancestors instantly.
These heat-adapted creatures had paid a price for their new ability. When Dallinger tested them at the original 60-degree temperature, they could no longer survive in the cooler conditions. Evolution had taken them in one direction, and the return path was closed.
Then disaster struck. Dallinger's incubator was accidentally destroyed in 1886, killing his uniquely adapted organisms and ending his research program. The scientific world would have to wait nearly a century before experimental evolution became a rigorous, widespread practice.
Two Paths to Adaptation
When organisms adapt to new environments in the laboratory, they can do so through two fundamentally different mechanisms, and understanding the distinction matters for grasping how evolution works.
The first path is mutation—a random change in the DNA sequence that happens to be beneficial in the new environment. A copying error during cell division might alter a protein's shape, allowing it to function at higher temperatures or digest a new food source. This is evolution at its most fundamental: genetic accident transformed into adaptation by natural selection.
The second path involves what geneticists call standing genetic variation. In any natural population, individuals differ from one another. These differences might be neutral or even slightly harmful under normal conditions, but when the environment changes, some of that existing variation can suddenly become advantageous. The genes were already there, waiting in the wings; the new environment merely changed which versions did best.
The distinction has practical implications for experimental design. Viruses and bacteria, with their enormous populations and rapid generations, frequently adapt through new mutations. You can start with genetically identical individuals and watch diversity emerge. But multicellular organisms like fruit flies or mice usually adapt through changes in the frequency of genetic variants already present in the population. They don't have time to wait for beneficial mutations to appear.
Darwin's Unintentional Experiment
Humans have been running evolution experiments for ten thousand years without realizing it. Every time our ancestors selected the largest seeds for next year's planting, or bred only the tamest wolves, or crossed the sweetest grapes with the hardiest vines, they were manipulating evolutionary trajectories.
The results are staggering. Cabbage, broccoli, cauliflower, Brussels sprouts, kale, and kohlrabi are all the same species—Brassica oleracea—sculpted by human selection into forms so different they seem unrelated. Modern corn, with its massive ears of starchy kernels, descends from teosinte, a scraggly grass whose seeds are scattered individually rather than packed onto a cob. The transformation is so complete that botanists initially refused to believe the two plants were related.
Darwin recognized this power. He opened On the Origin of Species not with finches or fossils but with a chapter on domestic animals, particularly pigeons. The diversity among pigeon breeds—the bulging throat of the pouter, the fantail's elaborate display, the tumbler's aerial acrobatics—seemed to Darwin almost impossible. An ornithologist shown these birds without context, Darwin argued, would certainly classify them as different species, possibly even different genera. Yet they all descended from the rock dove, the common pigeon of city squares and cliff ledges.
Altogether at least a score of pigeons might be chosen, which if shown to an ornithologist, and he were told that they were wild birds, would certainly, I think, be ranked by him as well-defined species.
If human breeders could produce such dramatic changes in a few centuries, Darwin reasoned, imagine what natural selection could accomplish over geological time.
The Long-Term Evolution Experiment
Richard Lenski's ongoing experiment has achieved a kind of scientific immortality. Starting with a single ancestral strain of E. coli, he established twelve identical populations and began growing them under identical conditions: the same temperature, the same nutrients, the same daily routine of dilution and growth. Every day, a sample of each population is transferred to fresh medium. Every five hundred generations, samples are frozen—living time capsules that can be revived decades later.
For years, the twelve populations evolved in broadly parallel ways, improving their ability to grow on the glucose provided in the medium. They grew faster. They used resources more efficiently. The changes were interesting but perhaps not surprising—this was adaptation in its most expected form.
Then something remarkable happened. After more than thirty thousand generations, one of the twelve populations suddenly developed the ability to metabolize citrate, a chemical that was present in the growth medium all along but that E. coli normally cannot use as a food source in the presence of oxygen. The population exploded in size, having discovered an entirely new resource.
This was evolution producing a genuinely novel capability—not just doing an old thing better, but doing something new. By reviving frozen ancestors, Lenski's team could pinpoint exactly when the crucial mutations occurred and trace the evolutionary path that led to citrate metabolism. It wasn't a single lucky mutation but a series of changes, some of which were necessary precursors that did nothing on their own but set the stage for later innovations.
Flies That Run on Empty Air
Not all experimental evolution uses microorganisms. At the University of California, San Diego, Gabriel Haddad and his colleagues raised fruit flies under progressively lower oxygen conditions—a kind of high-altitude training for insects. After two hundred generations, the flies had adapted to hypoxia in ways that went far beyond simple acclimatization.
The approach called "Evolve and Resequence" allowed researchers to compare the entire genomes of adapted flies with their ancestors. By identifying which regions of the genome had changed most dramatically, they could pinpoint the genes responsible for adaptation and begin to understand the molecular machinery underlying the evolutionary response.
This represents a revolution in experimental evolution. Where Dallinger could only observe that his organisms survived higher temperatures, modern researchers can identify the specific DNA sequences that changed, the proteins those sequences encode, and the physiological mechanisms those proteins control. Evolution has become visible at every level, from behavior to biochemistry.
Running Mice and Their Hyperactive Brains
In 1993, Theodore Garland Jr. began selecting mice for voluntary running. He wasn't forcing them to exercise; he simply measured how much they chose to run on wheels and bred from the most enthusiastic runners. After more than one hundred generations, the "High Runner" lines sprint nearly three times as many wheel revolutions per day as unselected control mice.
The interesting part is how they achieve this. The High Runner mice don't primarily run for longer periods—they run faster. They've evolved changes in muscle fiber composition, cardiovascular capacity, and bone structure that support sustained high-speed locomotion. They have elevated endurance when tested on motorized treadmills and higher maximum aerobic capacity.
But the most intriguing changes might be in their brains. The High Runner mice show alterations in motivation and reward systems, particularly in pathways involving dopamine and endocannabinoids—the brain's internal marijuana-like signaling molecules. They appear to have evolved a greater desire to run, not just a greater ability.
This has led researchers to propose the High Runner lines as a model for studying attention-deficit hyperactivity disorder in humans. Remarkably, when given Ritalin—the standard treatment for human ADHD—the hyperactive mice reduce their running to approximately the levels of control mice. Whatever drives their excessive activity responds to the same drugs that help human children sit still in classrooms.
The Soviet Aphid Experiments
In the 1950s, the Soviet biologist Georgy Shaposhnikov conducted experiments that demonstrated something rarely observed: the origin of reproductive isolation through adaptation to new conditions.
Shaposhnikov took aphids from the genus Dysaphis and transferred them to plant species they could barely survive on—hosts that were normally unsuitable for these particular aphids. The transplanted populations struggled at first, but over many parthenogenetic generations (aphids can reproduce without mating, cloning themselves rapidly), they adapted to their new food source.
The remarkable finding came when Shaposhnikov tried to cross the adapted aphids with members of their original population. They could no longer successfully interbreed. In evolutionary terms, they had become reproductively isolated—the first step toward becoming separate species. Shaposhnikov had watched speciation begin in his laboratory, driven not by geographic separation but by adaptation to different ecological niches.
Parasites Losing Their Tails
Some of the most detailed recent work comes from studying Leishmania donovani, a parasitic protozoan that causes visceral leishmaniasis, a deadly disease affecting hundreds of thousands of people annually. When researchers cultured these parasites in laboratory conditions for 3,800 generations—about thirty-six weeks—they documented the progressive rewriting of the organism's operating system.
In their natural environment, Leishmania needs flagella—whip-like tails that propel the parasite through its hosts. But in a culture dish, swimming is pointless. The parasites progressively reduced expression of twenty-three genes involved in building flagella. Why waste energy on locomotion when you're floating in nutrient broth?
The energy saved on flagella was reinvested in growth machinery. Expression of ribosomal proteins—the cellular factories that build other proteins—increased, as did production of specialized RNA molecules that regulate ribosome function. The parasites had streamlined themselves for their new environment, trading mobility for multiplication speed.
These adaptations weren't primarily driven by new mutations but by changes in gene copy number and complex interactions between genes. The parasites' infamous genomic instability, often seen as a puzzling feature, emerged as a sophisticated system for rapid adaptation through adjustable gene expression.
Evolving the Holobiont
The relationship between an organism and its internal microbes—its microbiome—has emerged as a frontier in evolutionary biology. We are not individuals but ecosystems, our bodies hosting trillions of bacteria, viruses, and other microscopic partners whose genes vastly outnumber our own. Selection on the host inevitably shapes the community within.
Researchers at Jagiellonian University in Poland have been exploring this territory using bank voles, small rodents selected for either high aerobic capacity, predatory behavior, or the ability to survive on low-quality plant food. The voles selected for herbivorous capability—those that can maintain body weight on a diet diluted with dried grass—have evolved altered gut microbiomes.
The bacteria in their intestines have shifted toward species better suited for breaking down plant fiber, suggesting that selection on the vole has produced evolutionary change in its bacterial partners. This is holobiome evolution: the animal and its microbes adapting together as an integrated system.
When Circuits Break
Synthetic biology has opened new experimental frontiers. Researchers can now insert precisely designed genetic circuits into organisms and watch how evolution treats these artificial additions. The results are often humbling.
When synthetic genetic circuits—sequences of DNA designed to perform specific functions like producing a fluorescent protein in response to a chemical signal—are inserted into bacteria or yeast, they tend to degrade. Given enough generations without strong selection maintaining their function, these engineered circuits accumulate mutations that break them. Evolution is a powerful optimizer, but it optimizes for current conditions, not for the intentions of engineers.
With appropriate selection, however, these systems can reveal how organisms recover lost functions. If a broken circuit reduces fitness, evolution will find ways to restore capability—sometimes by repairing the original damage, sometimes by evolving entirely new solutions to the same problem.
This work has implications far beyond academic biology. Cancer cells evolving resistance to chemotherapy are, in effect, running their own experimental evolution. Understanding how populations adapt to selection pressures—and how to predict and prevent unwanted adaptations—could inform treatment strategies that stay ahead of evolving tumors.
The Control Problem
One methodological lesson from experimental evolution deserves special attention. In traditional biology experiments, you typically have one control group and one treatment group, and you test whether they differ. But evolutionary trajectories are inherently random. Even populations experiencing identical conditions will diverge from each other through genetic drift—the statistical noise of finite populations.
This means that comparing a single experimental population to a single control tells you almost nothing. The differences might reflect your treatment, or they might reflect the random divergence that would have happened anyway. Lenski's experiment uses twelve populations for this reason. The four High Runner lines and four control lines in the mouse experiment serve the same purpose. You need multiple replicates to distinguish signal from noise.
This insight matters for understanding antimicrobial resistance and cancer drug resistance. When we ask whether a particular drug selects for resistance, we need to know not just whether treated populations become resistant, but whether untreated populations remain susceptible. Without proper controls, we might mistake inevitable evolutionary drift for drug-induced adaptation.
What We Learn by Watching
Experimental evolution has transformed our understanding of how adaptation works. We now know that evolution can be remarkably fast when selection is strong and population sizes are large. We've seen that different populations facing identical challenges often find different solutions—evolution is creative, not deterministic. We've watched organisms gain genuinely new capabilities, not just refine existing ones. We've traced evolutionary changes from behavior through physiology down to individual DNA bases.
Perhaps most importantly, experimental evolution has made the abstract concrete. Darwin knew that natural selection should produce adaptation, but he had to infer past evolution from present patterns. We can now watch adaptation unfold, freeze it in time, replay it from any point, and dissect its molecular mechanisms. Evolution has become an experimental science, as amenable to hypothesis testing as chemistry or physics.
Dallinger's heat-loving microbes are long dead, casualties of a laboratory accident more than a century ago. But their descendants—spiritual if not biological—continue evolving in incubators and culture flasks around the world, answering questions their Victorian ancestor could barely have imagined asking.