A hospital treats a patient's infection with an antibiotic. Within days the bacteria are gone, then a few survivors regrow, and the next dose does nothing. Nobody designed that resistance. A farmer in Mexico saves the biggest kernels of corn each year for nine thousand years, and a grass with a handful of hard seeds turns into maize. Nobody designed that either, and the farmer never knew a gene was involved.
Both are the same process, and it is the one Darwin and Wallace described in 1858: living things vary, some of the variation is inherited, and some variants leave more offspring than others. Given enough time, that is enough to produce eyes, wings and rainforests. As Theodosius Dobzhansky titled a 1973 essay, "nothing in biology makes sense except in the light of evolution". This piece is about why a biologist says that, and what the idea is not.
Charles Darwin spent twenty years collecting evidence, and in 1858 learned that Alfred Russel Wallace, working in the Malay Archipelago, had reached the same idea. Their joint papers were read at the Linnean Society of London that July. Darwin's On the Origin of Species followed in November 1859. The logic needs three things, and if all three are present, change follows:
This is selection, and it needs no foresight. Darwin did not know how heredity worked; Mendel's laws and DNA supplied it. The size of the response can be predicted. Breeders and biologists use the breeder's equation:
Heritability here means what Personality & Individual Differences (35/26) says it does: the share of variation in this population that tracks genes. If it is zero, selecting changes nothing next generation, however strong the selection. Take a trait with standard deviation 1 and let only the top 20% reproduce. Then S is about 1.4 standard deviations, and with h² = 0.5 the next generation's average moves up about 0.7. Try it.
Simulation of 1,000 individuals over 10 generations, with a seeded random generator: the trait is genes plus environment, parents are chosen by their trait value, and offspring inherit the midparent value plus shuffling. Units are standard deviations of the starting population. Predicted line: R = h² × S, with S from the normal curve. The simulation usually falls a little short over many generations, because selection itself uses up some of the variation it works on.
Every crop and farm animal is the result: maize from the Mexican grass teosinte (domesticated roughly nine thousand years ago), rice in the Yangtze valley, wheat in the Fertile Crescent, potatoes in the Andes, sorghum in Africa; cabbage, kale and broccoli are one species. Darwin opened the Origin with pigeon breeders for the same reason: artificial selection is natural selection with a person choosing.
Real populations do not move in straight lines. On the small Galápagos island of Daphne Major, Peter and Rosemary Grant and their students measured every finch for decades. In the 1977 drought, small, soft seeds ran out and mostly large, hard ones were left, and birds with deeper beaks survived better; their offspring had deeper beaks. Selection plus heredity, watched in a lifetime.
Not all change is selection. When a population is small, which alleles get passed on is partly luck, like a coin flipped few times. This is genetic drift. In the Wright–Fisher model each new generation is drawn by sampling the previous one, and a variant at frequency p will, over time, either vanish or take over (fixation). A neutral variant's chance of eventually fixing is just its starting frequency, p. What differs is how fast: in a population of 2N gene copies, genetic variety shrinks each generation by a factor of (1 − 1/2N).
The bigger the population, the weaker drift. That is why small populations lose variety, why cheetahs and island species are fragile, and why founders matter in human history: a few people who leave a population carry only part of its diversity. Mutation, selection, drift and migration are the four forces. Try a coin-flip population.
Wright–Fisher model: N diploid individuals, so 2N gene copies sampled with replacement each generation; no selection. Ten replicate populations, 100 generations, seeded. Expected variety left after t generations: (1 − 1/2N)t.
A species is, in Ernst Mayr's classic 1942 definition, a group of interbreeding natural populations reproductively isolated from others. It works well for birds and badly for bacteria, plants that hybridise, and fossils, so biologists use several definitions. What matters is the process: when two populations stop exchanging genes, they drift apart.
The commonest route is geographic. A mountain, a sea or a river splits a population (allopatric speciation). Each side meets different environments and different mutations. Over time, changes accumulate that do not work together when the groups meet again: the Dobzhansky–Muller picture, in which each side's new alleles are fine at home but clash in hybrids. Closure of the Isthmus of Panama, a few million years ago, split marine populations into Atlantic and Pacific halves that are now related species. Salamanders in California form a ring around the Central Valley in which neighbours interbreed but the two ends, where the ring closes, do not. And hundreds of cichlid fish species in Africa's Great Lakes arose in lakes that are young in geological terms.
It can also happen without a barrier. Plants often double their chromosome number in a hybrid, which isolates them in a single generation; bread wheat is such a hybrid of three grasses, formed in the Fertile Crescent roughly eight to ten thousand years ago. Boundaries stay fuzzy: modern humans outside Africa carry roughly 1 to 2% of DNA from Neanderthals, a sign that populations long separated still interbred when they met.
The case for common descent comes from independent lines that agree. A few:
DNA also gives a clock. In 1962 Emile Zuckerkandl and Linus Pauling saw that the differences between species' haemoglobins grew with the time since the species split. Motoo Kimura's neutral theory (1968) explained why: most changes that persist at the DNA level are neutral, and for neutral changes the rate of substitution equals the mutation rate, whatever the population size. Then the time since two species split is the genetic difference divided by twice the rate (both lineages change).
Approximate single-letter differences: human–chimpanzee about 1.2% (Chimpanzee Sequencing Consortium, 2005), human–gorilla about 1.6%, human–orangutan about 3%. The rate must be calibrated from fossils or from pedigrees, and the two methods disagree: that is why published human–chimp split dates range from roughly 6 to 13 million years.
Moving the rate slider is the honest lesson. A molecular clock is a ratio, and its answer depends on a calibration. Fossils and DNA must agree across hundreds of cases, and they do; at the edges, they argue.
Evolution is not only about the past. In 1988 Richard Lenski began growing twelve populations of E. coli in his Michigan lab, transferring a sample to fresh food every day. By 2020 the populations had passed 70,000 generations. All twelve got faster at growing in the flask. In one, around generation 31,500, bacteria arose that could use citrate, a molecule in the broth that E. coli normally cannot eat: a new ability, noticed and traced. The frozen samples let the team replay history.
Resistance is the same story with higher stakes. Alexander Fleming warned in his 1945 Nobel lecture that careless use could breed resistant microbes. Within a few years of methicillin's introduction, resistant staphylococci (MRSA) were reported, in Britain in 1961. A 2022 analysis in The Lancet (Murray and colleagues) estimated 1.27 million deaths in 2019 directly attributable to bacterial resistance, and 4.95 million associated with it. How does it happen? There is a small number of resistant mutants in any large bacterial population before the drug arrives; the drug kills the rest. Resistance usually carries a cost, which is why resistant strains fade when the drug is absent.
Illustrative expected-value model, not a clinical one: 109 bacteria regrow to the same total between eight treatments; resistant cells survive each drug exposure and grow at the stated relative speed. Real resistance varies by drug, species and patient; follow a clinician's advice on antibiotics.
| Common idea | What the science says |
|---|---|
| "Survival of the fittest" means the strongest win. | Herbert Spencer coined the phrase (1864); Darwin adopted it from 1869. Fitness means how many offspring that survive to reproduce, which can favour small, quiet or cooperative. |
| Evolution is "only a theory". | In science a theory is a tested, explanatory framework, like germ theory or plate tectonics. It is not a guess. |
| Humans came from monkeys. | Humans and living apes share common ancestors; we are apes, not descended from any living species. |
| Evolution is random. | Mutation is random with respect to need; selection is not. The combination is not a lottery. |
| Individuals evolve to meet needs. | Individuals do not change their genes. Populations change over generations as variants spread. |
| Evolution is a ladder toward us. | It is a branching tree with no goal. Bacteria alive today have as long a history as we do. |
| Evolution explains how life began. | Evolution explains change in living populations; the origin of the first life is a separate research question. |
Beliefs matter here. Many scientists are religious, and many religious communities accept evolution. The Catholic Church, for example: Pope John Paul II told the Pontifical Academy of Sciences in 1996 that evolution is "more than a hypothesis". Others disagree, and that is their right. What science can offer is the evidence and the method, and anyone can check both.
In the 1930s and 1940s, R. A. Fisher, J. B. S. Haldane and Sewall Wright showed mathematically that Mendelian genes and gradual selection fit together, and Dobzhansky, Mayr, Simpson and others tied this to experiments, species and fossils. This modern synthesis is the base. Since then, DNA, development and genomes have added to it, and real arguments continue:
Variation, inheritance and unequal success are enough for populations to change; the breeder's equation says how fast. Drift adds luck, isolation makes new species, and fossils, anatomy, DNA and geography all agree on common descent. Resistance shows it in real time. The arguments that remain, between neutral and adaptive, slow and fast, narrow and extended, are among scientists who agree on the core.
Try this: next time you hear "survival of the fittest" or "evolved to", ask whose offspring, in what environment, and over how many generations.
Sources. Darwin and Wallace, Linnean Society (1 July 1858); Darwin, On the Origin of Species (1859); Dobzhansky, American Biology Teacher (1973); Fisher (1930), Haldane (1932), Wright (1931) and the modern synthesis; Mayr, Systematics and the Origin of Species (1942); Kimura (1968); Zuckerkandl and Pauling (1962); Eldredge and Gould (1972); Laland and others, Nature 514 (2014) and Wray and others, same issue; Hamilton (1964); Margulis (1967); Daeschler, Shubin and Jenkins, Nature (2006) on Tiktaalik; Boag and Grant, Science (1981) on the 1977 drought; Blount, Borland and Lenski, PNAS (2008); Murray and others, Lancet 399 (2022); Jevons (1961) on MRSA; Fleming, Nobel lecture (1945); Chimpanzee Sequencing Consortium, Nature (2005); Green and others, Science (2010) on Neanderthal DNA; John Paul II, address to the Pontifical Academy of Sciences (22 October 1996). Dates for domestication and speciation events are approximate. Describes the science and is not medical advice.