# Natural Selection in Ecological Context

> Natural selection is a difference in reproductive success among variants that already exist, caused by an agent in an environment; inheritance carries the sorting forward, so it becomes evolution only when that variation is inherited. This lesson builds the argument from four parts (heritable variation, a measured fitness component, a named agent and an environment that can change), measures selection as a fitness ratio in peppered moths and birds and turns it into a change in morph share, and then reads how that ratio changes along a trait, the mode of selection, in field and experimental studies of gall flies, sticklebacks, guppies and widowbirds. Afterwards you can name the agent and the mode of selection a data set shows, separate selection from evolutionary response, and say what a study measured and what it did not.

Course: Ecology and Evolution (Evidence and evolutionary reasoning). Level: core university (third year). About 17 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/selection-in-ecological-context/

Builds on: How to Test an Evolutionary Explanation

## What you'll learn

- Define natural selection as a difference in relative reproductive success among heritable variants, caused by a named ecological agent in a stated environment, and explain why a change in environment can change which variant is favoured.
- Separate selection within a generation from evolutionary response across generations, and say what inheritance evidence the second needs.
- Interpret field and experimental evidence for directional, stabilizing, disruptive, negative frequency-dependent and sexual selection, and recognize when one trait carries more than one component at once.
- Read a relative selection coefficient, such as the peppered moth's per-day survival ratio, with its units and scope, and turn it into a change in morph share.

Birds taking more dark moths than pale ones is not yet evolution: the first film ended on the four parts a selection argument needs (variation, inheritance, a fitness difference, and an agent in an environment). This text adds how to measure the fitness difference, what it does to a population, and how to read selection on a trait that comes in a range of sizes.

> **The one idea.** Natural selection is a difference in reproductive success among variants that already exist, caused by an agent in an environment; inheritance carries the sorting forward. You measure selection as a fitness ratio between variants, and the way that ratio changes along a trait names the mode of selection.

## Measuring the fitness difference

**Fitness** is reproductive success relative to other variants in a given environment. **Relative fitness** is one variant’s fitness divided by the average, or by a reference variant’s. You rarely measure lifetime reproduction directly, so you measure a **fitness component**: survival over a stated period, mating success, or fecundity (number of eggs or young).

**Selection** is a difference in fitness among phenotypes (the observable forms of a trait) within one generation. **Evolutionary response** is the change across generations, and it happens only to the extent that the selected variation is inherited. A trait can be strongly selected and not change at all if its variation is environmental. The size of the response belongs to [population genetics](https://stemhop.com/lessons/population-genetics/) and [complex traits](https://stemhop.com/lessons/complex-traits-and-gwas/).

## Peppered moths and birds

The first film introduced the peppered moth, *Biston betularia*, and its pale typica and dark carbonaria morphs. (Intermediate insularia morphs also exist; the experiment below used only typica and carbonaria.) It did not show how differently birds see them: Walton and Stevens (2018) measured the match with models of bird vision: typica is a close match to lichen, while carbonaria is poorly matched to lichen and closer to plain bark in lightness.

### How big is the difference, and what does it do?

Earlier experiments had been criticised for how moths were released, so Michael Majerus repeated the work near Cambridge over six years. He released live moths of both morphs and counted a moth as eaten if it was gone four hours after sunrise; only about a quarter of those disappearances were seen as bird attacks. Cook, Grant, Saccheri and Mallet (2012) analysed his records.

On each day, a dark moth’s chance of surviving was about 0.91 times a pale moth’s: the first film’s dial, now with a number. The **selection coefficient**, *s*, states it as a shortfall from 1: s = 1 − 0.909 = **0.091 per day** against carbonaria. It compares the two morphs; it is not the share that died, not an annual decline and not a figure for all of Britain. It holds for one site, over those years, with birds as the agent.

**Predict.** Start with equal numbers of each morph on lichen-covered trunks. After five days of this selection, are dark moths about 45%, 38% or 5% of the survivors?

> **Check your prediction.** About 38%. Work in **odds**, the number of dark moths per pale moth. Equal numbers give odds of 1 : 1. One day multiplies them by 0.909, to 0.909 : 1. Five days give 0.9095 = 0.62 : 1. To turn odds into a share, divide by the total: 0.62 / 1.62 = 0.38. Only the ratio mattered, not the number of moths or a pale moth’s survival. A common slip is 0.5 × 0.62 = 0.31, which treats the pale moths as if all survived and mixes survivors with starting numbers; the share must be taken among the survivors.

In general:

> **odds of dark : pale after t days = starting odds × (1 − s)t**

Each day multiplies the odds by the ratio of the two morphs’ fitnesses, wdark / wpale = 1 − s. The rule assumes two morphs, a ratio steady over the interval, and no births or migration, and it gives shares among one generation’s survivors: still selection, not evolution. Turning the 38% into a change in allele frequency needs the dominance model in [population genetics](https://stemhop.com/lessons/population-genetics/), because carbonaria is dominant and a dark moth can carry one or two copies.

The moth ratio pointed one way, at one site. Most traits come in a range of sizes, and then the ratio changes along the trait: the shape of the curve of relative fitness against trait size names the mode. The second film builds that curve, in a simple model, for a gall caught between two enemies.

**Film transcript: Two Enemies, One Gall**

### Film transcript: Two Enemies, One Gall

In the first film, birds made one moth morph survive better than the other. Now take a gall. If wasps attack small galls and birds attack large ones, which gall is safest? The goldenrod gall fly lays an egg in a bud of tall goldenrod. The plant grows a round gall around the larva. The larva spends the winter inside. The gall is made by insect and plant together, and galls vary widely in size. Line the galls up by size, small on the left, large on the right. Above them we will draw relative fitness. That is how well larvae in galls of each size survive, compared with the average. A point above the dashed line means better than average. From the left come the wasps.

A parasitoid develops on or in a host and ultimately kills it. This parasitoid wasp attacks larvae in small galls more than in large ones. Many small galls go dark. From the right come the birds. Birds more often peck open the large galls and take the larva inside. Many large galls go dark too. So the middle survives best, with two named enemies pressing from opposite sides. Take a simple model where the two enemies are equally strong. The fitness curve rises, peaks over intermediate galls, and falls again.

Favouring the middle over both extremes is called stabilizing selection. Below, the spread of gall sizes narrows. Now, in the model, make wasp attack heavier than bird attack. Does the average surviving gall stay in the middle?

No. In the model, the peak slides toward larger galls, away from the stronger enemy. The spread still narrows, but the average moves up. Favouring one end like this is directional selection. So one trait can carry a stabilizing and a directional component at once. A field study found this pattern in central Pennsylvania, across twenty fields over two years. Parasitoid attack was more frequent than bird attack in those fields. Intermediate galls survived best, and selection also favoured larger galls. Now flip the hump into a dip. The middle does worst, and both ends do better. That shape is called disruptive selection.

Theory predicts it when individuals crowd onto the commonest resource, and those that use something different do better. One end, the middle, or both ends. Directional, stabilizing and disruptive are shapes of one curve. They are not boxes a trait must fit. So read the curve first, then name the mode.

From the film

- For a trait with a range of values, plotting relative fitness against the trait shows the mode of selection.
- Favouring one end is directional, favouring the middle is stabilizing, and favouring both ends is disruptive.
- In goldenrod galls, wasps took more larvae in small galls and birds more in large ones, so intermediate galls survived best.
- Because wasp attack outweighed bird attack in those fields, selection also favoured larger galls: one trait carried two components.

## Gall flies: two modes in one trait

Abrahamson, Sattler, McCrea and Weis (1989) found the second film’s pattern, **stabilizing** selection with a **directional** component toward larger galls, in galls of the goldenrod gall fly, *Eurosta solidaginis*, whose parasitoid wasp was *Eurytoma gigantea*. The directional component’s mean intensity was 0.21 phenotypic standard deviations: the surviving galls’ mean sat about a fifth of the usual spread in gall size above the mean of all galls, a shift in the trait, not a death rate.

**Predict.** Two other enemies, a second parasitoid and a beetle, attacked galls of all sizes. Suppose one of them killed the same share of larvae at every gall size. Would it change the shape of the fitness curve?

> **Check your prediction.** No, as long as that enemy acts on its own, multiplying survival at every size by the same factor. Relative fitness divides that factor out, just as the pale moths’ own survival dropped out of the moth calculation. An enemy changes the curve’s shape only if it takes some sizes more than others, or if its effect depends on what the other enemies do.

## Disruptive: crowded sticklebacks

The second film ended on the dip, **disruptive selection**, predicted when individuals crowd onto the commonest resource. Bolnick (2004) tested that by making crowding the only difference. In each of five lakes he set a pair of enclosures of threespine sticklebacks, *Gasterosteus aculeatus*, one stocked at low density and one at high. The trait was **gill raker length**; gill rakers are bony projections on the gill arches that help sieve food, so their length relates to what a fish eats. Consistently across the lakes, the curve dipped more in the middle at high density. Because only crowding differed within each pair, crowding became a tested cause, within the limits of the fitness measure: body size and relative gonad mass, stand-ins for how many offspring a fish could produce, not lifetime reproduction.

When reading such a curve, check that the dip lies inside the measured data, not past its edge (see Going deeper).

## Frequency-dependent: rare guppies

**Negative frequency-dependent selection** favours a phenotype when it is rare. Because the fitness ratio changes with the share, it can keep several variants in a population. A test must separate rarity from the phenotype. Olendorf and colleagues (2006) did this with male guppies, *Poecilia reticulata*, in three natural populations in Trinidad streams. They set male colour-pattern morphs at about 3 common to 1 rare, and reversed which morph was rare between replicate pools. After about two weeks (15 or 17 days, depending on the site), rare males had higher estimated survival, whichever pattern was rare. Recapture stands in for survival only if missing fish died rather than moved. Predators forming a search image for the common morph is one possible explanation, not a demonstrated one. The next generation was not measured, so this is selection, not an evolutionary response.

*Figure (AI-generated illustration).* Reversing which pattern was rare between pools, with male guppies set at about 3 common to 1 rare, let rarity be told apart from the pattern itself. The drawing simplifies the study's two tail-colour patterns to plain and spotted. Image: two stream pools side by side with four male guppies each. In the left pool three are plain and one is spotted; in the right pool three are spotted and one is plain. The rare fish in each pool is ringed in gold. An inset shows the plain and spotted patterns once each.

### Design problem

Students propose that birds take rare colour morphs of a prey less often than common ones. Design a field test.

> **One way to do it.** Use several comparable pools or plots, with morph A rare in some and common in others (for example 1 A to 3 B in half and 3 A to 1 B in the rest), so each morph is rare somewhere. Keep total density, habitat and season the same.
>
> Mark every individual, release, and recapture after a set interval, such as two weeks. State the assumption that an individual not recaptured has died rather than emigrated. Reduce emigration with barriers, and detect it by searching neighbouring habitat. Search each pool until you have caught every individual you can, so both morphs are equally likely to be found.
>
> The prediction is that whichever morph is rare survives better, in both sets of replicates. One colour doing better everywhere shows directional selection on colour but does not rule out frequency dependence too, so compare each colour’s survival when rare and when common. To test a mechanism, record whether predators take the common morph more than its share. A search image would produce that pattern, but so could other predator behaviours, so it supports a search image without proving one.
>
> Keep the conclusion to its scope: one season of survival in these sites shows frequency-dependent selection, not an evolutionary response, which would need evidence that colour is inherited and data on the next generation.

## Sexual selection: the widowbird tail

**Sexual selection** is selection through mating success, by competition for mates or by mate choice. Mating success is the fitness component here: a male with more females nesting on his territory can father more young. Male long-tailed widowbirds, *Euplectes progne*, carry very long breeding tails. Males that happen to differ in tail length may also differ in age or territory quality, so Andersson (1982) changed the tails of males in Kenya. Males given longer tails attracted more females to nest on their territories than males with normal or shortened tails.

**Predict.** Suppose the long-tailed males had also held their territories longer. What could you no longer conclude?

> **Check your prediction.** That the tail acted through female choice. A territory held longer could gather more nests by itself, tangling the two effects.

In fact, males with shortened tails held their territories as long as the others, so the tail affected mating, not territory holding, which points to female choice. The experiment does not show why females prefer long tails, or a long-tailed male’s net lifetime fitness.

*Figure (AI-generated illustration).* Random assignment made the tail the one planned difference: in Andersson's experiment in Kenya, males matched in groups of four had tails shortened, left normal (two control groups), or lengthened with added feather. Image: three male long-tailed widowbirds on grass stems, identical except for the tail: the left has a shortened tail, the middle a normal very long tail, and the right a longer tail whose added section is drawn in gold.

## Going deeper

### The moth numbers, worked

**Worked problem.** With invented numbers (1,000 moths of each morph, typica daily survival 0.80, no migration, reproduction or change in rates), carbonaria survives at 0.80 x 0.909 = 0.7272. One day leaves 800 typica and 727.2 carbonaria, so the dark share is 727.2 / 1,527.2 = 47.6%, down from 50%. The 0.80 and the 1,000s cancel: 0.909 / 1.909 gives the same 0.476.

> **carbonaria share after t days = 0.909t / (1 + 0.909t) (equal starting numbers)**

The odds form works from any start: from 10% dark (odds 1 : 9), five days give 0.62 : 9, about 6.5% dark. A fractional survivor such as 727.2 is an expectation, not a moth.

*Interactive on the lesson page:* Moth morph share: relative daily survival

### How the moth estimate was made

Majerus released 4,864 moths near Coton, Cambridgeshire, over six years. Moths gone at the four-hour check were presumed eaten, although a missing moth may have flown away or been taken by a predator other than a bird; only about 26% of disappearances were witnessed bird attacks, and the rest were inferred. The likelihood 95% interval on s was 0.028 to 0.157; at its ends, five days from equal numbers leave about 46% or 30% dark, not 38%. The inheritance of colour has a known locus: van’t Hof and colleagues (2016) traced carbonaria to a transposable-element insertion in the *cortex* region. Later work in butterflies suggests that non-coding RNAs at the same locus, rather than cortex itself, control melanism there.

### Reading a fitness curve

Studies usually fit relative fitness to a trait by regression. A significant positive quadratic term means the curve bends upward, but alone it does not show a fitness minimum among the trait values sampled. Picture a curve still falling, or just flattening, when the data run out: its low point lies outside the data, which show accelerating directional selection, not a dip. Martin and Pfennig (2012), working on Mexican spadefoot tadpoles, *Spea multiplicata*, from 15 ponds in Arizona and New Mexico, found that disruptive selection was the commonest form of curvature across the populations they studied, and strongest where competition within the species was most intense. They applied the established caution: look at the fitted curve, not only the coefficient.

### Selection over time

The fitness ratio need not keep pointing one way. On Daphne Major in the Galápagos, Peter and Rosemary Grant measured selection on the beaks and bodies of two finch species from 1972 to 2001, and it ranged from consistently one-directional to oscillating ([testing evolutionary explanations](https://stemhop.com/lessons/testing-evolutionary-explanations/)).

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/selection-in-ecological-context/
