The Long Reach of the Gene

Core idea

The book has, for twelve topics, argued that genes build bodies, and that bodies are the vehicles through which selection acts on genes. This final topic asks a more radical question: why should the phenotype stop at the body's skin?

Dawkins's argument: A gene's phenotypic effects can extend beyond the body it builds — into the artifacts the body constructs (beaver dams, bird nests, caddis-fly cases), into the bodies of other organisms (parasite-induced behavioral change in the host), even into the social and ecological systems the gene operates within. The extended phenotype is the gene's full causal reach, which is often larger than the body that carries it.

This is the topic that introduced the term Dawkins would later elaborate in his 1982 book The Extended Phenotype — which he has more than once called the work he is most proud of, more so than the book containing this topic. The extended-phenotype framework is the gene's-eye view taken to its logical conclusion: the body itself is just one phenotypic effect of a gene, and there is no a priori reason to privilege it over the gene's other effects.

Why it matters

The phenotype, properly defined

A phenotype is any effect a gene has on the world. A gene for blue eyes produces blue eyes. A gene for tongue-rolling produces tongue-rolling. A gene for shyness produces shy behavior. All of these are "phenotypic effects" of the gene.

The conventional view tacitly assumes the phenotype stops at the skin. The topic shows this assumption is just convenience — a useful approximation for most situations, but not a deep principle. A gene that builds a body that builds a nest is also the gene for the nest. The dam is part of the beaver's phenotype. The case is part of the caddis fly's phenotype. The web is part of the spider's phenotype.

Beaver dams as phenotype

The clearest case is the beaver. Beavers build dams that can be tens of meters long, dramatically reshape watersheds, and create entire wetland ecosystems. Different beaver populations build subtly different dams. The dam-building behavior, the shape of the dams, even the size — all are partly heritable. So the dam is a heritable trait, varying with genes, subject to selection.

Selection has, over millions of years, tuned beaver dams. Dam-building genes that produce better dams produce more beaver offspring. The dam is part of the beaver's phenotype as much as the beaver's teeth are. The skin of the beaver is not a meaningful boundary for selection.

Host manipulation as phenotype

Even more radically, a parasite's genes can manipulate the host's body to serve the parasite's reproduction. Dawkins surveys many examples:

• Trematode worms in snails reach a stage where they need to be eaten by a bird. They infiltrate the snail's eye-stalks and cause them to swell, pulsate, and become brightly colored — making the snail conspicuous to birds. The worm's genes have built a phenotype in the snail's body.
• Nematomorph (horsehair) worms in crickets induce the cricket to seek water and drown itself — releasing the worm into its breeding habitat.
• Toxoplasma in rats causes the rat to lose its fear of cat odor — making it more likely to be eaten, completing the parasite's life cycle in the cat's gut.
• Cuckoo chicks in host nests have evolved begging calls and gaping mouths that exploit the host's nestling-recognition circuitry. The cuckoo's genes are reaching across species and across bodies to manipulate the host's parental behavior.

In each case the parasite's genes have a phenotypic effect outside the parasite's body. The host is a survival machine — but not the parasite's survival machine. The boundary of the phenotype has become permeable.

The "central theorem" of the extended phenotype

Dawkins states the central theorem this way: an animal's behavior tends to maximize the survival of the genes "for" that behavior, whether or not those genes are inside the body of the particular animal performing the behavior. A snail with swollen eyestalks is doing something that maximizes the trematode genes; the snail's own genes are losing. A rat that approaches cat-scented places is doing something that maximizes the Toxoplasma genes. The behavior is in the host; the genes that built the behavior are in the parasite.

This is the deepest version of the gene's-eye view. The body is no longer privileged. The unit of selection is the gene; the body is one of many possible phenotypic effects.

The genome as a parliament of genes

A complementary picture emerges from looking inside the body. The genes of a single organism are themselves not unified actors. Most cooperate (because they share the same exit from the body — the gametes — and thus have parallel interests). But outlaw genes (segregation distorters, mitochondrial DNA with its own agenda, transposons) act for themselves at the rest of the genome's expense.

The genome, then, is a parliament of replicators that have learned, through long co-existence, to mostly cooperate. The body is the parliament's joint product. The boundary of the organism is, even from the inside, somewhat arbitrary.

Why the extended phenotype matters

The extended-phenotype framing changes how we ask evolutionary questions. Instead of asking "what is this body for?" we ask "what are the genes for?" — and the answer often involves things outside the body.

It also changes how we understand certain biological phenomena. Many traits look maladaptive from the organism's point of view but make perfect sense from a gene-extended-phenotype perspective. The snail with swollen eyestalks is doing something obviously bad for itself — but obviously good for the trematode genes inside it.

Key takeaways

Mental model

Copy sourceZoom

Practical application

Stop privileging the body when explaining behavior

When you read about animal behavior — bird songs, courtship displays, nests, alarm calls — try the extended-phenotype reframing. The behavior is not "for" the bird; it is "for" the genes that built the behavior. Sometimes the genes are in the actor (most cases). Sometimes they are elsewhere (parasite, host, mate). The question "whose genes does this behavior serve?" is more incisive than "what does the organism want?"

See manipulation everywhere

Many behaviors that look puzzlingly self-destructive in the actor make sense as extended phenotypes of another organism's genes. Cuckoo chicks. Mind-controlled crickets. Even — at the level of cultural evolution — humans who devote their lives to spreading memes that benefit themselves not at all. The extended phenotype gives you a vocabulary for this kind of analysis.

Recognize the limits of the body as a unit

Bodies are useful units for most everyday biology — they are bounded, they reproduce as units, they are where most genes coexist. But the body is not metaphysically fundamental. For some evolutionary questions, the right unit is the gene's full causal reach, not the body. Trained biologists learn to switch between these perspectives as the question requires.

Apply to human institutions

Human organizations — corporations, religions, political movements — also have phenotypic reach that extends beyond their formal boundaries. A corporation's "phenotype" includes its products, its advertising, the behaviors it induces in customers and employees, the regulations it shapes. Asking "what serves the organization's gene equivalent (its persistence and growth)?" can illuminate behavior that looks puzzling from inside the organization.

Example

Consider the trematode worm Dicrocoelium dendriticum, which has one of biology's most baroque life cycles. Adults live in the bile ducts of sheep. Eggs leave with sheep dung. Snails eat the eggs. The larvae mature in the snail. Then a stage of the parasite must reach the next host — an ant — and from there, back into a sheep.

How does it get from ant to sheep? The parasite migrates to the ant's brain and induces it to climb to the top of a blade of grass each night and clamp its jaws onto the grass. There it waits, motionless, all night — exactly where it is most likely to be eaten by a grazing sheep at dawn. If no sheep arrives, the ant releases its grip in the morning warmth and resumes normal life — only to climb back the next night.

The grass-clamping behavior is in the ant's body. The genes that built that behavior are in the parasite. By any standard, the genes' "extended phenotype" reaches into the ant's nervous system and controls it. The ant is not a free agent acting bizarrely; it is a vehicle that is partly being driven by a gene system located in another species.

This is the topic's most vivid evidence that the body is not the right boundary for thinking about whose genes are at stake.

Related lessons