• Three theories have been put forward to explain the existence of adaptation: supernatural creation, Lamarckism, and natural selection. Only natural selection works as a scientific theory.

• Natural selection is not the only process that causes evolution, but is the only process causing adaptation.

• Natural selection, at least in principle, can explain all known adaptations. Examples of coadaptation and useless incipient stages have been suggested but they can be reconciled with the theory of natural selection. The vertebrate eye could have evolved rapidly by small advantageous steps.

• Adaptation can either be defined historically or by current function: we can say either that an adaptation is a character that evolved by natural selection to perform its modern function, or one that evolved by natural selection whether or not its modern function is the same as the one it first evolved to perform.

• For many characters, it is not obvious how (or whether) they are adaptive.

• Adaptation can be studied by comparing the observed form of an organ with a theoretical prediction, by experimentally altering the organ, and by comparing the form of the organ in many species.

• Not all the effects of an organ will have evolved as adaptations by natural selection. Some will be inevitable consequences of the laws of physics.

• Adaptations cannot be simultaneously optimal for all the levels of organization in life. What is optimal for the organism may not be optimal for its population.

• Adaptations may be imperfect because of time-lags: a species may be adapted to its past environment because it takes time for natural selection to operate.

• Adaptations are imperfect because the mutations that would enable perfect adaptation have not arisen. The imperfections of living things are due to genetic, developmental, and historical constraints, and to trade-offs between competing demands.

• For particular characters, adaptation and constraint can be alternative explanations. Likewise, differences in the form of a character between species may be due to adaptation to different conditions or to constraint. Forms that are not found in nature may be absent because they are selected against or because a constraint renders them impossible.

• Adaptation and constraint can be tested between by several methods: by the use of predictions from a hypothesis of adaptation or constraint, by direct measures of selection, by seeing whether the character is variable, and whether the variation is heritable and can be altered by artificial selection, and by examining comparative trends.

• The methods of analysing adaptation are valid when applied to adaptive characters and inter-specific trends; they might be misleading for non-adaptive characters and trends.

• Biologists disagree about how exact, and how widespread, adaptation is in nature.

• There are criteria to distinguish adaptive from non-adaptive characters. Measurement of selection provides an objective criterion, but is not always practical; other criteria, such as non-randomness and purposiveness are often useful, but may become subjective in borderline cases.