Natural selection

Natural selection is the process by which heritable replicators become more or less common because some persist and reproduce more successfully than others. In this account, selection has no foresight or purpose: it begins with the persistence of stable forms, and its cumulative effects explain evolutionary change, including adaptations and patterns of behavior.

From stable molecules to evolution

Before life, ordinary physical and chemical processes could have favored stable arrangements of atoms. Molecules that lasted longer, replicated more rapidly, or copied themselves more accurately would have become a larger proportion of the primeval soup over time. The earliest natural selection was therefore selection among stable forms, without design or direction.

Once replicators existed, their copying was not perfectly accurate. Variation made evolutionary change possible, but this does not mean that replicators benefit from changing: greater copying fidelity can itself be favored because it preserves successful forms. In this account, evolution is something that happens through differential persistence, not a goal pursued by organisms or genes. Whether the original replicators should be called living is treated as a matter of terminology; they were the ancestors of life either way.

What selection acts on

Natural selection in its broadest form is the differential survival of entities. For this difference to have lasting effects, entities must occur in many copies, and at least some copies must persist across significant evolutionary time. The book therefore identifies small genetic units, rather than individuals, groups, or species, as the replicators best suited to serve as enduring units of selection. A gene is defined as chromosomal material capable of lasting through enough generations to function as a unit of selection; it is seldom divided and can pass intact from grandparent to grandchild.

This gene-centered account does not deny that organisms live or die. Individual survival and reproductive success are the immediate occasions on which selection commonly operates. Their long-term evolutionary consequence, however, is a change in the relative frequencies of genes in the gene pool. Sex and crossing-over continually shuffle genes among bodies, keeping that pool fluid; evolution is the resulting increase in some genes and decrease in others.

Genes are not independent architects of bodies. A survival machine contains many thousands of genes whose effects interact: a body part is influenced by multiple genes, and any one gene may affect different parts. A gene described as being “for” a trait does not build that trait alone. It means that, other things being equal, the gene tends to influence development in a particular direction, within an intricate web of genetic and environmental causes.

Selection, behavior, and evidence

The same logic applies to behavior. For an altruistic or selfish behavior pattern to evolve, a gene influencing the development of nervous systems in a way that makes the behavior more likely must fare better in the gene pool than alternatives. Genes exert this influence indirectly: they shape survival machines and nervous systems, while moment-to-moment decisions are made by nervous systems. Learning and simulation can give brains an increasing role in decisions without changing the evolutionary requirement that heritable variants differ in their success.

The account warns against treating vivid examples as proof of a general rule. Stories of apparent altruism illustrate what the terms mean at the level of individual behavior, but they are not serious evidence for a broad evolutionary explanation. Likewise, a behavior that looks complex does not by itself establish how it is inherited. In a honeybee study, hygienic strains removed infected grubs from their cells and discarded them; susceptible strains did not. Rothenbuhler’s breeding experiments found that first-generation hybrid hives were non-hygienic, while back-crosses produced hives that performed all, none, or only the uncapping part of the behavior. By removing cell caps himself, he showed that some apparently non-hygienic bees could then remove the larvae. The staged results supported the hypothesis that separate inherited factors contributed to uncapping and removal, illustrating how behavioral differences may be investigated rather than assumed.

Adaptation, strategy, and competing explanations

Natural selection also supplies the historical explanation for adaptations. An adaptation must rest on heritable differences among which selection can act. The book applies this reasoning even to a caddisfly’s constructed house: ancestral genes that led their carriers to build more effective protective houses could have been favored, although the genetics of caddis houses had not actually been studied. The proposed genes affect behavior and development; their effects can be described through the resulting variation in house properties.

In behavioral ecology, selection can be analyzed through costs and benefits rather than by assuming that fighting or restraint is always advantageous. Game theory gives this reasoning a formal shape. An evolutionarily stable strategy is one that cannot be bettered by an alternative once most of a population adopts it; selection penalizes deviations from such a strategy. This account treats the strategy as a pre-programmed behavioral policy, not a conscious calculation by the animal.

The same selection-centered test challenges explanations that treat sacrifice for a species or group as the general cause of altruism. A group of self-sacrificing individuals could appear likely to outlast a competing group, but a selfish individual within it might exploit the others, survive, and leave descendants. The book therefore asks what a trait does to gene frequencies, rather than inferring its evolutionary function from an apparent benefit to a group. It also treats Wynne-Edwards’s population-regulation proposal cautiously: examples of crowding behavior could fit his interpretation, but could also be explained in gene-centered terms, and the evidence presented for his theory is judged inadequate.

Across these arguments, natural selection links distinct scales without confusing them: stable molecules give way to replicators; genes persist through changing bodies; and differences among organisms’ survival and reproduction alter the gene pool. The book’s metaphors of genes making their living or pursuing their interests are useful shorthand, but the underlying account is non-intentional: selection is the accumulated consequence of differential persistence and reproduction.