When biologists refer to the theory of evolution, they use the word "theory" as it is used throughout science. It does not mean a mere speculation or an unsupported hypothesis. Rather, as The Oxford English Dictionary puts it, "a hypothesis that has been confirmed or established by observation or experiment, and is propounded or accepted as accounting for the known facts; a statement of the general laws, principles, or causes of something known or observed" (our italics). The complex body of principles that explain evolutionary change is a theory in the same sense as "quantum theory" in physics or "atomic theory" in chemistry: it has been developed from evidence, tested, and refined, and it accounts for literally thousands of observations made throughout the entirety of biological science and paleontology.

Like all scientific theories, the theory of evolution is a current best explanation. It has withstood innumerable tests and attempts to disprove it, but it is still being refined, modified in the light of new knowledge, and extended to account for newly discovered phenomena. The theory of genetics has had such a history, progressing from Mendel's simple early principles to the complex body of molecular principles that constitute today's theory of inheritance, and it is constantly being refined and modified, even though its core principles have remained valid for a century. So it is with the theory of evolution.

Is evolution also a fact? All but the most trivial facts begin as untested hypotheses--such as the hypothesis that the earth revolves around the sun. They acquire "facthood" as more and more evidence accrues in their favor, and as they withstand attempts to refute them. The evidence and attempt at refutation may take many forms besides simple observations; indeed, the most powerful evidence is not mere observations, but conformity to predictions that the hypothesis makes about what we should see if the hypothesis is true or false. We do not observe the earth making a circuit around the sun; we accept this hypothesis because of the numerous, verified astronomical observations--and more recently observations from spacecraft--that conform to the predictions of the hypothesis. So Copernicus's hypothesis is now a fact--a statement supported by so much evidence that we use it as if it were true.

Biologists accept as fact that all organisms, living and extinct, have descended, with innumerable changes, from one or at most a few original forms of life. For Darwin in 1859, this was a hypothesis, for which he provided abundant evidence from comparative anatomy, embryology, behavior, agriculture, paleontology, and the geographic distributions of organisms. Since that time, all of the many thousands of observations in each of these areas have supported Darwin's core hypothesis. To these observations has been added copious evidence that Darwin could hardly have dreamed of, especially from paleontology and molecular biology. A century's accumulation of such evidence establishes descent, with modification, from common ancestors as a fact of science. How we explain this fact--what the principles and causes of it may be--is the theory of evolutionary process, parts of which are subject to various amounts of scientific debate, modification, and extension.

To claim evolution as a fact is to confront controversy, for probably no claim in all of science evokes as much emotional opposition as biological evolution. Nonetheless, no scientific hypothesis other than common descent with modification can account for and make predictions about the unity, diversity, and properties of living organisms. No other hypothesis of the origin of biological diversity is supported by such overwhelming evidence, and no competing hypothesis spawns such richness of scientific study and has as many implications for the biological sciences and their applications to societal needs.

How is it studied

We can note only a few of the most general, commonly used methods.
Phylogenetic inference methods are used to estimate relationships among species (living and extinct). Recent advances in logical and computational methods have greatly enhanced the confidence with which this can be done. Greatly oversimplified, the underlying principle of these methods is that species that share a greater number of derived ("advanced") features stem from a more recent common ancestor than species that share fewer such features. It is obvious, then, that rats, whales, apes, and other mammals share a more recent common ancestor with each other than with birds or lizards, since the mammals possess many unique, derived features (e.g., milk, hair, a single lower jawbone). It is less obvious, but nonetheless increasingly likely as new data accumulate, that chimpanzees are more closely related to humans than to gorillas. These conclusions are based not only on improved methods of analyzing data, but also on a virtually inexhaustible trove of new data: long sequences of DNA, which reveal far more similarities and differences among species than can be found readily in their anatomy. The same methods used to infer the genealogy of species can be used to infer the genealogy of the genes themselves. Thus, for example, molecular evolutionary studies can use DNA sequences to estimate how recently variants of a gene carried by different people arose from a single ancestral gene.

Origins of modern humans

Paleontological databases. Evolutionary paleontology is founded on systematics, including phylogenetic inference, because it is necessary to classify and determine the relationships of fossilized organisms before anything else can be done with them. Once this is done, fossils can be used for two major kinds of evolutionary study. One is tracing evolutionary changes in the characteristics of lineages through geologic time, such as those that occurred during the descent of mammals from reptilian ancestors. The other is determining the times and rates of origination and extinction of lineages and relating such changes to other events in earth history. For instance, each of five great mass extinctions--one of them evidently due to an asteroid impact--was followed by a great increase in the rate of origination of species and higher taxa, providing evidence that diversification of species is stimulated by the availability of vacated resources.  Studies of fossil biodiversity rely on computerized databases of the geologic and geographic occurrence of thousands of fossil taxa, data accumulated by thousands of paleontologists throughout the world over the course of two centuries.

Characterizing genetic and phenotypic variation. Characterizing variation is one of evolutionary biology's most important tasks. The statistical methods used to do this can be applied to data of many different kinds. Quantitative genetic analysis, which is also used extensively in the breeding of crops and domestic animals, is an important tool for measuring and distinguishing between genetic and nongenetic variation in phenotypic characteristics. One method of making this distinction involves measuring similarities among relatives, which requires knowledge of the relationships among individuals within natural populations. Molecular genetic markers can often provide such information. Recent advances in DNA-based molecular technologies have made it feasible to construct detailed genetic maps for a wide range of species, and to identify specific DNA regions that control or regulate quantitative characters.

Inference from genetic patterns. Many evolutionary changes (though not all) take immense amounts of time, so the processes involved are often inferred from existing patterns of variation rather than observed directly. Many hypotheses about evolutionary processes can be tested by comparing patterns of genetic and phenotypic variation with those predicted by evolutionary models. For instance, the "neutral theory" of molecular evolution by genetic drift holds that molecular variation within species should be greater, and divergence among species more rapid, for genes in which most mutations have no effect on organisms' fitness than for those in which most mutations have a strong effect. According to this model, genes that encode unimportant proteins or which do not encode functional proteins at all, should display more nucleotide variation than genes that encode functionally important proteins. Studies of DNA variation have abundantly confirmed this model. This model is so powerful that molecular biologists now routinely use the level of sequence variation among species as a clue to whether or not a newly described DNA sequence has an important function.

Observing evolutionary change. Some important evolutionary changes happen fast enough to document within one or a few scientific lifetimes. This is especially likely when, due to human activities or other causes, a population's environment changes, or a species is introduced into a new environment. For example, changes in food supply due to drought in the Galápagos Islands caused substantial, although temporary, evolutionary change in the beak size of a finch, within just a few years; a virus introduced to control rabbits in Australia evolved to be less virulent in less than a decade (and the rabbit population became more resistant to it); rats evolved resistance to the poison warfarin; hundreds of species of crop-infesting and disease-carrying insects have evolved resistance to DDT and other insecticides since World War II; and the rapid evolution of resistance to antibiotics in pathogenic microorganisms poses one of the most serious problems in public health 4, 42.
Evolution of pesticide resistance

Experimentation. Evolutionary studies often involve experiments, such as placing populations in new environments and monitoring changes or selecting directly on a particular character of interest. Among the most common experiments are those that analyze evolutionary change in manipulated populations, either under natural conditions or in the laboratory, using organisms with short generation times that can evolve rapidly. For example, experimenters have used laboratory populations of bacteria to monitor the course of adaptation to high temperatures, novel chemical diets, antibiotics, and bacteriophage (viruses that attack bacteria), and have characterized the new mutations underlying these adaptations 16. One group of researchers predicted the evolutionary changes in life history characteristics (e.g., rate of maturation) that guppies should undergo if they were subjected to a certain species of predatory fish. They introduced guppies into a Trinidad stream where this predator lived, and found that after about six years, the introduced guppies differed from the ancestral population just as they had predicted 50.

The comparative method. Convergent evolution is the independent evolution, in different lineages, of similar characteristics that serve the same or similar functions. For example, several unrelated groups of fishes that inhabit turbid waters have independently evolved the capacity to generate a weak electric field that enables them to sense nearby objects. Convergent evolution is so common that it can often be used to test hypotheses. If we hypothesize a certain function for a feature, then its occurrence or condition should be correlated with specific environments or ways of life. For example, evolutionary ecologists predicted that, irrespective of their phylogenetic relationships, plant species that inhabit environments poor in light, water, or nutrients, and which therefore cannot readily replace tissues lost to herbivores, should contain greater quantities of defensive chemicals than species that grow in richer environments. By comparing many species of plants that grow in different environments, evolutionary ecologists have found considerable evidence supporting this prediction