AP BIOLOGY · UNIT 7 · NATURAL SELECTION

Evolution, counted

Evolution is a change in allele frequencies, and that can be counted. Count the alleles in a gene pool, watch chance alone move them in a small population, then use two short equations to say what a population would look like if nothing were acting on it, and test real data against that.

AP Biology 7.4, 7.5 and 7.115 lessons10 chaptersdry lab

Start learningRun the dry labTeach it

Greater prairie chickens once filled the prairies of Illinois. The prairie shrinks, most birds go, and the gene pool is redrawn from the few that are left. The flock grows back. Its alleles do not. Then birds from other states arrive.
NavigationBy the end you cantap a goal to go to its chapter

01Learn

Ten chapters and a dry lab, in three parts. What changes a gene pool: counting alleles (chapter 1), mutation (2), genetic drift (3), bottlenecks and founders (4) and gene flow (5). A population where nothing changes: the Hardy-Weinberg model (6), its two equations (7) and testing real counts against it (8). Why diversity matters: populations with little of it (9), and one allele that helps in one place and harms in another (10). Make each prediction before you open the answer.

NavigationChapterstap a chapter to jump to it
  1. 1Count the alleles, not the animals
  2. 2New alleles come one way
  3. 3Chance moves small populations most
  4. 4A few survive, or a few leave
  5. 5Alleles that move keep populations alike
  6. 6A population where nothing happens
  7. 7Two equations, two ways in
  8. 8Is this population evolving?
  9. 9Less variety, less to select from
  10. 10Good here, harmful there

7.4.C Gene pools and allele frequencies

Count the alleles, not the animals

A population of 100 diploid mice has 30 BB, 50 Bb and 20 bb. How many copies of the gene are in its gene pool?

  • A population is a group of one species living in one place and breeding with each other.
  • Its gene pool is every allele of every gene that its members carry.
  • An allele frequency is the share of all the copies of a gene in the gene pool that are one allele.
  • A change in allele frequencies over generations is evidence that the population is evolving, whatever caused it.

The detail

  • Allele frequencies are written as decimals from 0 to 1. For a gene with two alleles, the two frequencies add to 1.
  • A genotype frequency is something else: the share of individuals with one genotype (AA, Aa or aa). Keep the two apart from the start.
  • Individuals do not evolve. A beetle keeps the alleles it was born with. What changes is how common each allele is in the next generation.
  • Natural selection is one way frequencies change. Chapters 2 to 5 are the others, and the course calls them random: mutation, genetic drift and migration.

Watch out

  • A heterozygote carries one copy of each allele. It counts once for A and once for a.
  • A population can change in size without evolving, and evolve without changing in size. Look at the frequencies.

Evolution happens to populations, so you need a way to describe a whole population's genes at once. A population is a group of one species living in one place and breeding with each other. Its gene pool is every allele that its members carry. Picture each individual tipping its two copies of one gene into a jar.

An allele frequency is then easy to read: count the copies of one allele and divide by all the copies. In the first figure, ten individuals give a jar of 20 alleles, 14 of them A, so the frequency of A is 0.7. Count alleles, not individuals. A genotype frequency is a different number, the share of individuals with one genotype, and the next chapters keep the two apart.

Now the definition the rest of the guide uses. When the allele frequencies of a population change from one generation to the next, the population has evolved. The second figure runs it with selection: no beetle changes, but the survivors' gene pool holds more J than the one before. A change in allele frequencies is the evidence that evolution has happened. It does not, by itself, say what caused it.

  • Gene pool: every allele in a population. Allele frequency: one allele's share of the copies.
  • Diploid: two copies per individual, so N individuals carry 2N copies.
  • Evolution is a change in allele frequencies over generations.
Watch out The dominant allele is not always the common one. Dominance is about which phenotype a heterozygote shows. How common an allele is depends on the population's history, and a dominant allele can be rare.
Population 1: ten individualseach carries two allelesAagene pool: 20 alleles14 A and 6 aPopulation 1A 0.7a 0.314 ÷ 20 = 0.7the frequencies of A and a add to 1seaPopulation 2, cut off from population 1A 0.4a 0.6different
  1. Ten individuals, each carrying two alleles of one gene: allele A in green, allele a in pink.
  2. Tip every allele into one jar: the gene pool, 20 alleles, 14 of them A and 6 of them a.
  3. The allele frequency of A is 14 out of 20, which is 0.7. The frequencies of A and a add to 1.
  4. A second population, cut off from the first, has different frequencies: 0.4 and 0.6.
Generation 0: twenty beetlesJjtwo alleles of a jaw gene eachDrought: only tough seeds. The four jj beetles cannotcrack them, and do not breed.JjGeneration 0: 20 J, 20 jp = 0.5Jj16 survivors: 20 J, 12 jp = 0.625dashed: where the J share startedThe survivors breed: their pool is the next generation.J rose from 0.5 to 0.625: evolution.
  1. Twenty beetles, each with two alleles of a jaw gene. Across the population, J and j are equally common: p = 0.5.
  2. A drought leaves only tough seeds. The four jj beetles cannot feed and do not breed.
  3. Count the alleles of the 16 survivors: 20 J and 12 j. The frequency of J is now 0.625.
  4. The survivors breed, so their gene pool is the next generation's. The allele frequency has changed: evolution.

7.4.A, 7.4.B Mutation

New alleles come one way

Bacteria are spread on a plate that contains an antibiotic, and a few colonies grow. When did the resistance mutations in those colonies most likely arise?

  • Mutation is a random process that adds new genetic variation to a population.
  • It is the only source of new alleles. The other processes change how common alleles already are.
  • New alleles give new phenotypes, and those are what natural selection acts on.
  • Mutations are random with respect to need: they do not arise because an organism would benefit.

The detail

  • A mutation is a change in the DNA sequence: an error in replication or repair, or damage from radiation or chemicals. Unit 6 covers how they happen.
  • Only a mutation in a cell that makes gametes reaches the next generation's gene pool.
  • For any one gene, mutation is rare, so on its own it changes allele frequencies very slowly.
  • Across a whole genome and a large population, many new mutations appear every generation. That steady supply is the raw material.

Links across the course

  • Point mutations, frameshifts and the causes of mutation: see Unit 6, mutations and biotechnology.
  • Gene flow can bring an allele that is new to one population. It is not new to the species: it arose by mutation somewhere first.

How we know

  • Experiments with bacteria in the 1940s and 1950s grew many separate cultures that had never met a virus or a drug, then tested each one. Resistant cells were already present in some before any exposure, and their number varied widely from culture to culture, as it would if each mutation arose at a random moment.

Every allele in a gene pool began as a mutation, a change in the DNA sequence. Mutation is the one process that makes alleles that were not there before. Drift, gene flow and selection, in the chapters ahead, only change how common alleles already are.

Mutation is random. It does not happen where it is needed or when it is needed. A copying error can land anywhere, in any individual, at any time, and most do nothing useful. In the figure, press for the next generation: one bead in the jar changes to a new color, and the moment it happens differs every time you run it.

A new allele can give a new phenotype: a different enzyme, a darker coat, a cell that survives a drug. That is why mutation matters to natural selection. Selection cannot favor a variation that does not exist, so mutation supplies the phenotypes and selection acts on them. Mutation is random; selection is not.

  • Mutation: random, and the only source of new alleles.
  • It provides the variation, and so the phenotypes, that selection acts on.
  • Mutation is random; natural selection is not.
Watch out Natural selection is not a random process. Mutation, drift and migration are the random ones. Selection favors the phenotypes that leave more offspring in a particular environment, and that is the opposite of chance.
Generation 1Generation 4Generation 5gene poolallele Aallele anew allele
A gene pool as a jar of two kinds of bead. Press Next generation until one bead changes. New alleles come only this way, and the moment is different each run.

7.4.A, 7.4.B Genetic drift

Chance moves small populations most

The same allele starts at a frequency of 0.5 in a population of 10 and in a population of 10,000. Nothing selects for or against it. After 20 generations, where is its frequency more likely to be far from 0.5?

  • Genetic drift is a change in allele frequencies by a nonselective process: chance in which individuals breed and which alleles they pass on.
  • How strong its effect is depends on population size.
  • Drift can lose an allele altogether, or leave one allele as the only one.
  • Drift can let a small population diverge from other populations of the same species.

The detail

  • Each generation's gene pool is a sample of the last. A parent passes on one of its two alleles at random, and some individuals leave more offspring than others for reasons unrelated to their alleles: a nest flooded, a seed landed on rock.
  • A small sample can be far from the jar it came from. Toss a coin 10 times and 7 heads is common; toss it 10,000 times and 70% heads almost never happens.
  • Once an allele is lost, drift cannot bring it back. Only mutation or gene flow can.
  • Drift is not directed. It does not favor the alleles that help, and it can carry a harmful allele up or a useful one away.

Watch out

  • Drift acts in every population. In a large one each generation's sample is close to the last, so the change is tiny; in a small one it can be large.

Each generation is drawn from the one before. Which individuals breed, how many offspring each leaves and which of its two alleles each offspring gets all involve chance, and none of it has to do with which allele is better. The change this causes is genetic drift.

Run the figure. Each line is one population of the chosen size, with no selection, starting at 0.5. Try 1,000 a few times, then 10. At 1,000 the lines stay near the middle. At 10 they wander, and some hit 0 or 1 within a few generations: the allele is lost, or it is the only one left, and that population cannot change at this gene again unless a new allele arrives.

Look at two of your small populations side by side. They started the same and nothing acted on them differently, yet they ended with different gene pools. That is how drift can let a small population diverge from others of its species: by chance alone, it comes to differ from them.

  • Drift: chance changes in allele frequencies, not caused by selection.
  • The smaller the population, the bigger the swings.
  • Drift can lose alleles and make small populations diverge.
Watch out Drift does not only happen in tiny populations. It acts in all of them; it matters most in small ones, where a generation is a small sample.
Four runs at each size, no selection, starting at 0.5N = 1000.5101530generationN = 10001530generationN = 1,00001530generationallele frequency
Choose a population size and press Run. Each run adds a line: one population, no selection, the allele starting at 0.5. Run each size several times and compare.

7.4.A The bottleneck effect and the founder effect

A few survive, or a few leave

A storm leaves 6 survivors out of 600 lizards on an island. Compared with the 600, what does the survivors' gene pool most likely have?

  • The bottleneck effect is drift that happens when a population is cut to a small number of individuals for at least one generation.
  • The founder effect is drift that happens when a few individuals are separated from the rest and start a new population. Its frequencies follow the alleles the founders happened to carry.
  • Both are drift acting on one small sample of a population.
  • A population that grows back in number does not regain the alleles it lost.

Examples

  • Northern elephant seals (a bottleneck). Hunting cut them to perhaps 20 animals by the 1890s. Today there are more than 200,000, and their genetic variation is still very low.
  • The black robin of the Chatham Islands (a bottleneck). In 1980 there were 5 birds, with one breeding female. There are about 300 today, every one descended from her.
  • The Old Order Amish of Lancaster County, Pennsylvania (a founder effect). The community grew from a small group of settlers and has married mostly within itself. Ellis-van Creveld syndrome, a recessive condition whose features include extra fingers, is far commoner there than elsewhere, and its allele has been traced to one couple who arrived in the 1740s.

Telling them apart

  • Bottleneck: the population stays where it is and most of it dies. Ask "did most die?"
  • Founder: the original population carries on, and a few start again somewhere new. Ask "did a few leave?"
  • Either way, the few are a small sample, so it is drift: chance, not selection, decides which alleles they carry.

Two kinds of drift have names, because they turn up again and again. In the bottleneck effect a population is cut to a few individuals for at least one generation, by hunting, a storm, a disease or a lost habitat. The survivors are a small sample of the gene pool, so alleles are lost and frequencies shift.

In the founder effect nothing has to die. A few individuals are separated from the rest and start a new population: birds blown to an island, a few families who settle somewhere new. The new gene pool holds only what the founders carried, at their frequencies, and a rare allele that one founder happened to carry can become common.

The figure shows both, unlabeled: say which row is which before you tap. Then play the film in steps. It counts a prairie chicken gene pool before the bottleneck, at the low point and after the flock has grown back. The number of birds recovers. The number of alleles does not, because a lost allele can come back only by mutation or by gene flow.

Remember it Two kinds of drift, by their first letters

  • Bad dayBottleneck effectfew survive, the population stays put
  • Far awayFounder effectfew leave and start again elsewhere
  • Bottleneck: cut to a few for at least one generation.
  • Founder: a few separated, starting a new population.
  • Both are drift; numbers can recover, lost alleles do not.
Watch out A population that grows back has not got its diversity back. Elephant seals number in the hundreds of thousands and still carry the gene pool of the few that survived the hunting.
Row 1Bottleneck effectTap to revealRow 2Founder effectTap to reveal
  1. Two large populations, each with five alleles shown as colors. Something happens to each.
  2. A few individuals hold each gene pool now. Which row is the bottleneck effect and which the founder effect?
Two ways a few individuals end up holding a gene pool. Decide which row is the bottleneck effect and which is the founder effect, then tap each row to check.
The hero film in steps, with the counts: how many birds, and how many alleles in the gene pool, before, at the low point, after the flock grows back and after birds arrive from other states.

7.4.A, 7.4.B Migration and gene flow

Alleles that move keep populations alike

Two populations of a flower differ in the frequency of an allele for petal color. A new bridge lets pollinators cross between them every year. Over time, what will the two frequencies most likely do?

  • Migration can result in gene flow: the addition of alleles to a population, or their removal from it.
  • Gene flow changes allele frequencies when the migrants' frequencies differ from the population they join or leave.
  • Gene flow carries alleles between populations, generation after generation.
  • So gene flow prevents two populations from diverging into separate species.

The detail

  • Alleles move only when migrants breed in their new population. Animals that pass through and leave no offspring add nothing.
  • In plants, gene flow is mostly pollen and seeds: carried by wind, water or animals.
  • Gene flow can bring back an allele that drift removed, or bring one that the population never had.
  • Gene flow works against drift and against selection that differs between places: both make populations differ, and gene flow keeps mixing them.

Examples

  • Greater prairie chickens in Illinois (an example, met again in chapter 9). From 1992, birds from Minnesota, Kansas and Nebraska were released into the last small flocks, adding alleles the flocks had lost.
  • A wildlife overpass over a highway (an example). It lets animals cross between two halves of a population that the road had split, and breed on the other side.

Links across the course

  • When gene flow stops for good, the populations can go on diverging until they cannot interbreed: speciation, later in this unit.

Migration moves individuals, and when they breed it moves alleles. That movement of alleles into or out of a population is gene flow. It changes a population's allele frequencies whenever the migrants' frequencies differ from those of the population they join or leave.

The figure has two ponds of beetles that start with different allele frequencies. With the slider at zero, no beetles cross, and each pond drifts on its own: run it a few times and the two bars wander apart. Move the slider up so a few migrants cross each generation, and run it again.

That result matters beyond two ponds. Drift, and selection that differs from place to place, push populations apart. Gene flow pulls them together, because every generation each gene pool is partly mixed into the other. As long as alleles keep flowing, two populations stay one species. Stop the flow, and nothing holds them together.

  • Gene flow: alleles added or removed by migration.
  • Gene flow makes populations more alike.
  • So it prevents populations from diverging into separate species.
Watch out Gene flow needs breeding. A deer that crosses a road and is killed adds no alleles to the population on the other side.
pond 1pond 20 migrants per generation00.51010203040generation5 migrants per generation010203040generationfrequency of allele B
Two ponds, two gene pools. Set the number of migrants each generation, run it and watch the two bars.

7.5.A The Hardy-Weinberg model

A population where nothing happens

  • Hardy-Weinberg equilibrium is a model that describes and predicts allele frequencies in a population that is not evolving.
  • Its five conditions: a large population, no migration, no new mutations, random mating and no natural selection.
  • These conditions are never all met in a real population.
  • The model is still valuable, as a null hypothesis: what you would see if nothing were acting.

Each condition, and what it rules out

  • Large population rules out genetic drift (chapters 3 and 4).
  • No migration rules out gene flow (chapter 5).
  • No new mutations rules out new alleles (chapter 2).
  • Random mating: any individual is as likely to mate with one partner as another, whatever their genotypes. Mating with relatives, or with partners like themselves, gives more homozygotes and fewer heterozygotes. On its own it changes genotype frequencies, not allele frequencies.
  • No natural selection: every genotype survives and reproduces equally well.

Why the square

  • With random mating, an egg meets a sperm with no regard to its allele. A fraction p of eggs carry A and a fraction p of sperm carry A, so a fraction p × p of offspring are AA.
  • Aa can arise two ways, A egg with a sperm or a egg with A sperm, so heterozygotes are pq + pq = 2pq.

Named once

  • The model is named for Godfrey Hardy and Wilhelm Weinberg, who worked it out separately in 1908. You need the model, not the names.

Which population comes closest to meeting the Hardy-Weinberg conditions for one gene?

To tell whether a population is evolving, you need to know what it would look like if it were not. Hardy-Weinberg equilibrium is that picture: a model in which allele frequencies stay the same from one generation to the next, and genotype frequencies follow from them in fixed proportions.

The model needs five conditions, and each one switches off a process from the chapters before. A large population removes drift. No migration removes gene flow. No new mutations removes new alleles. No natural selection removes selection. Random mating means genotypes pair up only by chance, which sets the proportions in the square in the first figure.

No real population meets all five. Every population is finite, mutations always arise, and most face some selection. So why use the model? Because it is a null hypothesis: a clear statement of what no change looks like. If real data match it, nothing detectable is acting on that gene. If they do not, something is, and the way they differ is a clue to what. Try it in the tool: switch off one condition at a time and see how the population departs from the prediction.

Remember it The five conditions: Big Mice Must Run Slowly

  • BigLarge populationno drift
  • MiceNo migrationno gene flow
  • MustNo new mutationsno new alleles
  • RunRandom matinggenotypes pair by chance
  • SlowlyNo natural selectionevery genotype does equally well
  • Hardy-Weinberg: a model of a population that is not evolving.
  • Five conditions, never all met.
  • It is the null hypothesis that real populations are tested against.
Watch out "Never met" does not make the model useless. A null hypothesis is not a claim that nothing is happening; it is the baseline that lets you show that something is.
Random mating: any egg can meet any spermeggsA: p = 0.6a: q = 0.4spermAp = 0.6aq = 0.4p²AApqAapqAaq²aa0.360.240.240.16p² + 2pq + q² = 1AA 0.36+ Aa 0.48+ aa 0.16= 1If 16% show therecessive trait,q² = 0.16,so q = 0.4and p = 1 − 0.4= 0.6
  1. Egg and sperm alleles meet at random. If the frequency of A is p = 0.6 and of a is q = 0.4, each edge of the square is split in those proportions.
  2. The four areas are the four ways eggs and sperm can combine: p², pq, pq and q².
  3. AA is p² = 0.36, Aa is 2pq = 0.48 and aa is q² = 0.16. Together they make 1.
  4. Working back: if 16% show the recessive trait, q² = 0.16, so q = 0.4 and p = 0.6.
Aatwo discs: one individual100 individuals: p = 0.60, q = 0.40predicted, from pcounted36.036AAp²48.048Aa2pq16.016aaq²A recessive condition in 1 in 2500 peopleaaq² = 1 ÷ 2500 = 0.0004qq = √0.0004 = 0.02pp = 1 − q = 1 − 0.02 = 0.98Aa2pq = 2 × 0.98 × 0.02 = 0.0392, about 1 in 25.5AAp² = 0.98² = 0.9604Assumes the population meets the Hardy-Weinberg conditions.
  1. One gene, two alleles: p is the frequency of A, q of a, and p + q = 1. Here p = 0.60.
  2. Hardy-Weinberg predicts the genotypes from p and q: p² + 2pq + q² = 1. When every condition is met, the counts fit.
  3. It works backwards too. From the aa frequency, 1 in 2500: q = 0.02, p = 0.98, and carriers 2pq = 0.0392, about 1 in 26.
Set p, then run generations with every condition met. Switch off one condition at a time and say which process you have let in, and what it does to the counts.

7.5.A The Hardy-Weinberg equations

Two equations, two ways in

In a population at Hardy-Weinberg equilibrium, 1 baby in 2,500 has a recessive condition. About what fraction of people are carriers?

  • For a gene with two alleles, p is the frequency of one allele and q the frequency of the other: p + q = 1.
  • At equilibrium the genotype frequencies are p² + 2pq + q² = 1: p² for AA, 2pq for Aa and q² for aa.
  • From genotype counts, count the alleles. This always works.
  • From the recessive phenotype, q² is its frequency, so q is its square root. This works only if you can assume equilibrium.

Worked examples

  • 1. Count the alleles. 200 beetles: 98 GG, 84 Gg, 18 gg. p = (2 × 98 + 84) ÷ 400 = 280 ÷ 400 = 0.70. q = (2 × 18 + 84) ÷ 400 = 120 ÷ 400 = 0.30. Check: p + q = 1.00. At equilibrium you would expect 0.49 × 200 = 98, 0.42 × 200 = 84 and 0.09 × 200 = 18: this population fits exactly.
  • 2. From the recessive phenotype. 16% of a population of flowers is white, the recessive color. q² = 0.16, so q = 0.4 and p = 0.6. p² = 0.36 and 2pq = 0.48. Check: 0.36 + 0.48 + 0.16 = 1.00.
  • 3. Carriers. 1 baby in 2,500 has a recessive condition. q² = 0.0004, q = 0.02, p = 0.98. 2pq = 2 × 0.98 × 0.02 = 0.0392, which is 1 in 25.5: about 1 in 25.
  • 4. Hidden in carriers. If q = 0.10, then 2pq = 2 × 0.9 × 0.1 = 0.18 and q² = 0.01. There are 18 carriers for every affected person, which is why selection against a rare recessive allele is slow: most copies sit unseen in heterozygotes.
  • 5. Two samples (numbers invented). In rock pocket mice, dark fur is dominant and light fur recessive. On dark rock, 9 of 100 mice are light: q² = 0.09, q = 0.3, p = 0.7. On light rock, 81 of 100 are light: q² = 0.81, q = 0.9, p = 0.1. Both use the shortcut, so both assume each sample is at equilibrium.

Why not square roots of everything?

  • 1,000 individuals: 123 AA, 455 Aa, 422 aa. Counting gives p = (2 × 123 + 455) ÷ 2,000 = 701 ÷ 2,000 = 0.3505.
  • The square root of 0.123 is 0.351, close, but only because this population happens to be near equilibrium. In one that is not, square roots give the wrong answer. Count when you have the counts.

The model comes with two equations. Before using either, write down what p and q stand for in this question: p is the frequency of one allele and q of the other. Those are the only two alleles, so p + q = 1. That equation is about alleles.

The second is about genotypes. With random mating, the genotypes are in the proportions of the square from chapter 6: p² + 2pq + q² = 1, with p² the frequency of AA, 2pq of Aa and q² of aa. Every individual has one of the three genotypes, so the three add to 1.

There are two ways in. If you have counts of all three genotypes, count the alleles: each homozygote carries two copies of its allele and each heterozygote one of each. This needs no assumption, and it is the method for testing whether a population is at equilibrium. If you can see only one phenotype, start from the recessive one: those individuals are aa, so their frequency is q². Take the square root for q, then p = 1 − q. That shortcut assumes equilibrium, so say so. In the tool, enter 1 in 2,500 for aa and follow the working.

Remember it Which symbols are alleles and which are people

  • Piecep and qallele frequencies: one copy
  • Personp², 2pq, q²genotype frequencies: two copies each
  • p + q = 1 (alleles). p² + 2pq + q² = 1 (genotypes).
  • Genotype counts: count the alleles; always valid.
  • Recessive phenotype only: q = √q²; assumes equilibrium.
Watch out p is not "the dominant phenotype" and q² is not "the recessive allele". p and q are allele frequencies. The dominant phenotype is p² + 2pq, and the recessive allele's frequency is q, the square root of q².
Aatwo discs: one individual100 individuals: p = 0.60, q = 0.40predicted, from pcounted36.036AAp²48.048Aa2pq16.016aaq²A recessive condition in 1 in 2500 peopleaaq² = 1 ÷ 2500 = 0.0004qq = √0.0004 = 0.02pp = 1 − q = 1 − 0.02 = 0.98Aa2pq = 2 × 0.98 × 0.02 = 0.0392, about 1 in 25.5AAp² = 0.98² = 0.9604Assumes the population meets the Hardy-Weinberg conditions.
  1. One gene, two alleles: p is the frequency of A, q of a, and p + q = 1. Here p = 0.60.
  2. Hardy-Weinberg predicts the genotypes from p and q: p² + 2pq + q² = 1. When every condition is met, the counts fit.
  3. It works backwards too. From the aa frequency, 1 in 2500: q = 0.02, p = 0.98, and carriers 2pq = 0.0392, about 1 in 26.
Calculate from one genotype. Start from 1 in 2,500 affected, then try a recessive phenotype of your own and check that p², 2pq and q² add to 1.

7.5.A, 7.4.C Testing a population against the model

Is this population evolving?

A population has 90 AA, 60 Aa and 50 aa. From its own allele frequencies, the model expects 72, 96 and 32. What is the clearest reading of these numbers?

  • The null hypothesis: the population is in Hardy-Weinberg equilibrium for this gene, so any difference between observed and expected counts is due to chance.
  • Expected counts come from the population's own allele frequencies: p², 2pq and q² times the number of individuals.
  • A chi-square test says whether the difference is too big to be chance.
  • Allele frequencies that change over generations, on a graph or in a table, are direct evidence of evolution.

The steps

  • Count the alleles to get p and q.
  • Work out the expected counts: p² × N, 2pq × N and q² × N.
  • Chi-square = the sum of (o − e)² ÷ e over the three genotypes. The formula and the critical values are on the AP formula sheet.
  • Degrees of freedom: the rule used in this course is the number of genotype classes minus one, so 2 for three genotypes. At p = 0.05 the critical value is 5.99.
  • Above the critical value: reject the null hypothesis. At or below it: fail to reject.

Two worked tests

  • Does not fit. 200 individuals: 90 AA, 60 Aa, 50 aa. p = (180 + 60) ÷ 400 = 0.6, q = 0.4. Expected: 72, 96, 32. Terms: 18² ÷ 72 = 4.5; 36² ÷ 96 = 13.5; 18² ÷ 32 = 10.125. Chi-square = 28.125, far above 5.99: reject. There are too few heterozygotes.
  • Fits. 400 individuals: 150 AA, 180 Aa, 70 aa. p = (300 + 180) ÷ 800 = 0.6. Expected: 144, 192, 64. Terms: 36 ÷ 144 = 0.25; 144 ÷ 192 = 0.75; 36 ÷ 64 = 0.5625. Chi-square = 1.5625, below 5.99: fail to reject.

Watch out

  • Statisticians often take one more degree away, because p was worked out from the same counts, which gives 1 degree of freedom and a critical value of 3.84. Both tests above reach the same decision either way. Use the rule your teacher gives you, and show it.
  • Rejecting the null says the population is not in equilibrium at this gene. It does not say which condition failed: the pattern of the difference is a clue, not proof.

Chapter 7 gave the model's prediction; this chapter tests a real population against it. The null hypothesis is that the population is at Hardy-Weinberg equilibrium for the gene, so any gap between the counts you observe and the counts the model expects is due to chance. Work out p and q by counting alleles, then the expected counts: p², 2pq and q², each times the number of individuals.

Real counts never match exactly, so the question is how big a gap chance can explain. The chi-square test from Unit 5 answers it. Add up (o − e)² ÷ e over the three genotypes and compare the total with the critical value. The figure works the first example in the fold a step at a time.

The course's other kind of evidence is a graph: allele frequency against generation. The second figure has one for two populations, with invented data. Before you answer the question under it, decide what size of wobble a sample could show by chance, and what kind of change it could not.

  • Null hypothesis: in equilibrium; any difference is chance.
  • Expected counts from the population's own p and q.
  • Reject: not in equilibrium. Changing frequencies: evolving.
Watch out Fail to reject is not proof of equilibrium. These counts give no reason to doubt it, and that is all. Never write "accept".

On the second figure, population 1's frequency of allele A stays between 0.59 and 0.61 for 50 generations, and population 2's falls from 0.60 to 0.25. Which claim do the data support?

Does the population fit the model?observedexpected9072AA6096Aa5032aa200 individualsCount the allelesp = (180 + 60) ÷ 400 = 0.6q = 1 − 0.6 = 0.4Expected in 200AA 0.36 × 200 = 72Aa 0.48 × 200 = 96aa 0.16 × 200 = 32oeo − e(o − e)² / eAA9072184.5Aa6096−3613.5aa50321810.125sum20020028.125051015202530critical value 5.99reject the null hypothesisfail torejectχ² = 28.125With 1 degree of freedom the critical value is 3.84: the decision is the same.
  1. 200 individuals: the observed counts of AA, Aa and aa.
  2. Count the alleles: p = 0.6 and q = 0.4. The expected counts are 72, 96 and 32.
  3. Each genotype's (o − e)² ÷ e: 4.5, 13.5 and 10.125.
  4. Chi-square is 28.125. With 2 degrees of freedom the critical value is 5.99: reject the null hypothesis.
00.20.40.60.8101020304050population 1population 2generationfrequency of allele Ainvented data
The frequency of allele A in two populations over 50 generations (invented data). Which population is evolving, and what is your evidence?

7.11.A Genetic diversity and resilience

Less variety, less to select from

In the 1840s nearly all the potatoes in Ireland were one variety, grown from pieces of tuber. When a new disease arrived, what happened to the crop?

  • A population's ability to respond to a change in its environment depends on its genetic diversity.
  • Species and populations with little genetic diversity are at risk of decline or extinction.
  • Genetically diverse populations are more resilient: they are more likely to include individuals who can withstand a new pressure.
  • A bottleneck or a founder effect leaves little diversity, even after the numbers recover.

Examples

  • The Irish potato famine, 1845 to 1852 (an example). Most of the crop was one variety, the Lumper, grown from pieces of tuber, so the plants were clones. When Phytophthora infestans, a water mold (not a fungus), arrived, almost every plant was equally open to it. About a million people died, and a million or more emigrated.
  • Antibiotic resistance (an example). A large, varied population of bacteria often includes a few resistant cells, and those survive a drug that kills the rest: see the resistance on a plate film in Unit 6. Diversity makes the bacteria resilient, and that is the problem for medicine.
  • The course also names corn rust as an example of a crop disease that spreads where little variation exists.

The cards

  • California condor (an example). 22 birds in 1982. The last wild condor was caught in April 1987, leaving 27, every one in captivity. Bred and released, there were 561 in 2022. Lead poisoning, from bullet fragments in the carcasses they eat, is the main continuing threat.
  • Black-footed ferret (an example). Thought extinct, it was found again near Meeteetse, Wyoming, in 1981. Disease then struck, and the last 18 were taken into captivity between 1985 and 1987. Nearly every black-footed ferret alive today descends from seven of them. Since 2024 the kits of a ferret cloned from one of the other captured animals have added a little new variation.
  • Greater prairie chicken (an example). Millions in Illinois in the 1800s, about 2,000 by 1962 and fewer than 50 by the early 1990s, with many eggs failing to hatch. From 1992 birds were brought in from Minnesota, Kansas and Nebraska, and hatching success rose at first. Later studies question how lasting that rescue was.

Links across the course

  • The prairie chickens are gene flow used on purpose (chapter 5), and the hero film tells their story.

Genetic diversity is how much variety a gene pool holds. It matters because environments change: a new disease, a hotter summer, a new predator. Natural selection can only work with the variation a population already has. A population with many alleles at many genes is more likely to include some individuals who can survive a new pressure, and their offspring carry it on. That makes it resilient.

A population with little diversity has less to offer. Press "disease arrives" on the two fields in the first figure, after you have said what you expect. The clone field is the Irish potato crop of the 1840s in miniature.

Bottlenecks are how many wild species lose diversity. The second figure holds three of them, each brought back from a few dozen animals or fewer. Their numbers have risen. Their gene pools are still those of a few founders, and that leaves them at risk of decline or extinction from any pressure that those few could not withstand.

  • Genetic diversity: the variety in a gene pool.
  • More diversity, more chance some individuals withstand a new pressure: resilience.
  • Little diversity raises the risk of decline or extinction.
Watch out Recovered numbers are not recovered diversity. Hundreds of condors still carry the alleles of the 27 they came from, and only mutation or gene flow can add more.
FIELD 1one varietygrown from pieces of tuber0 of 24 standingFIELD 2six varietiesfour plants of each16 of 24 standing
One field of a single clone, one of mixed varieties. Say what you expect, then press Disease arrives.
California condor2219822719875612022birds, not to scaleWhy is it still at risk?Black-footed ferret18captured7foundersferrets, 1985 to 1987Why is it still at risk?Greater prairie chicken2,0001962under 501990sIllinois, not to scaleWhy is it still at risk?

Tap a card to read its low point, its founders and what was done.

Three species brought back from a bottleneck. Each card gives the numbers; the question under each is why the population is still at risk.

7.11.A An allele in two environments

Good here, harmful there

Where malaria is common, the sickle cell allele is kept at a higher frequency than elsewhere. In a region with no malaria, over many generations, what would its frequency most likely do?

  • An allele that is adaptive in one environment may be deleterious in another, because the selective pressures differ.
  • Whether an allele helps depends on the environment, not on the allele alone.
  • So a change of environment can reverse the direction of selection on an allele.

Examples

  • The sickle cell allele (an example). Two copies cause sickle cell disease. One copy gives some protection against severe malaria. Where malaria is common, carriers survive better than people with no copy, so the allele is kept in the gene pool; it is commonest in regions where malaria has long been common. (This is sometimes called heterozygote advantage.)
  • Fur color in rock pocket mice (an example). Dark fur hides a mouse on dark lava rock and gives it away on pale rock, where light fur hides it. Owls and other predators do the selecting. The dark allele is favored on one rock and selected against on the other.
  • Antibiotic resistance (an example). A resistance allele helps a bacterium while the drug is present. Many resistance alleles carry a cost, so where the drug is absent, cells without them often grow faster.

Links across the course

Earlier chapters showed that natural selection favors the phenotypes that do best in a particular environment. The word "particular" matters. An allele has no fixed value: it can be adaptive in one place and deleterious in another, because the selective pressures in the two places differ.

The figure shows the sickle cell allele in two regions. In each region there are three genotypes: no copy, one copy and two copies. Switch between malaria common and malaria absent, and compare the survival bars. One copy helps only where malaria is a threat. Two copies cause disease everywhere.

That is also why diversity matters. A gene pool that holds alleles that are useless today may hold the alleles that are adaptive tomorrow. Lose them in a bottleneck and that option is gone.

  • Adaptive in one environment, deleterious in another.
  • The environment decides an allele's effect on fitness.
  • Change the environment and selection on an allele can reverse.
Watch out There is no "good allele" without a place. Ask "good where, and against what?"
MALARIA COMMONMALARIA ABSENTsickle cell alleleother alleleno copyone copytwo copies010.81.01.00.2relative survival to adulthoodillustrative, not measured values
Three genotypes in two regions. Choose a region and compare the survival of people with no copy, one copy and two copies of the sickle cell allele.

02Dry lab: The gene pool bench

Set a population's size and its starting allele frequency, break one Hardy-Weinberg condition at a time, and watch p across 100 generations. Then test a generation's genotype counts against the model.

00.51020406080100generationfrequency of allele AAa10,000population0generation0.500p, frequency of AGENE POOL00.51020406080100generationfrequency of allele AAa10population100generation0.000p, frequency of AGENE POOLrunNbrokengenerationpAAAaaa
  1. The bench: population size, starting p and three switches.
  2. Every condition met at 10,000: p stays put. Every switch off at 10: runs wander apart.
Your questionWith every Hardy-Weinberg condition met, does p stay put? And when one condition is broken, which process changes p, how fast, and does it push the same way every run?
What you can changeThe population size, from 10 to 10,000; the starting frequency of allele A; and three switches, each with a strength: selection against one genotype, migrants each generation from a population with its own p, and a mutation rate. The rates are exaggerated so that a change shows within 100 generations.
What to recordFor each run: the settings, p at chosen generations and the genotype counts (AA, Aa, aa) for any generation you pick. Download CSV keeps the whole run.

Find out

  • The model. Leave every switch off, so that every condition is met, and set the population to 10,000. Run 20 generations and record p at every fifth one. What does the Hardy-Weinberg model predict, and does the bench agree?
  • Small populations. Keep every switch off but set the population to 10. Run it five times from the same starting p. Describe how the five runs differ, and explain why, when nothing differed between them.
  • One condition at a time. Back at 10,000, turn on selection, then migration, then mutation, one at a time, and run each a few times. Which changes p fastest? Which pushes p the same way every run?
  • Test a generation. Pick generation 20 of any run, take its AA, Aa and aa counts, and test them against Hardy-Weinberg as in chapter 8: count the alleles, work out the expected counts, then chi-square.
  • Back to populations. Use your runs to say why a model that no population meets is still a useful null hypothesis.

Record

  • For each run: the population size, the starting p, which switches were on and at what strength.
  • p at the generations you chose, and how far it ended from where it started.
  • The genotype counts you tested, the expected counts, your chi-square and your decision.
  • Which process changed p in each run, and the evidence from the graph. Download CSV keeps every run's table.

With every condition met and a population of 10,000, p starts at 0.50 and stays between 0.48 and 0.52 for 20 generations. What does this show?

Five runs with every switch off, a population of 10 and p starting at 0.5 end at 0.85, 0.10, 0.50, 0 and 0.65. Why are they so different?

At 10,000, with selection against aa switched on, five runs all show p rising. With migrants from a source where p is 0.1, five runs all show p falling toward 0.1. What do the two have in common?

Back to populations

Use your runs to explain why the Hardy-Weinberg model is a useful null hypothesis even though no real population meets its five conditions, and name which process changed p in each run where it changed.

03Check

A fresh paper from the exam bank on this guide's topics each time: past-paper style questions, marked as you go, and saved to your Exam practice progress when you are signed in. It opens on biologybybradford.com, so it needs the internet.

Exam practice

Exam questions on 7.4, 7.5 and 7.11, marked

Start 15 marks of exam questionsYour exam practice

04Downloads

This file is the guide, saved to run without the internet.

Saved from biologybybradford.com/guides/ap-biology-unit-7-population-genetics, where the guide is kept up to date with the notes, the revision sheet and, for the teacher, the lesson plans and slides.

05Teacher resources

The guide as five one-hour lessons: what each covers, its focus and the chapters it uses.

LessonCoversFocusChaptersLesson planSlides
1Gene pools, mutation and drift 7.4.A, 7.4.C60 min Gene pools and allele frequencies as evidence of evolution, mutation as the source of new alleles, and genetic drift by hand and on screen 1, 2, 3
2Bottlenecks, founders and gene flow 7.4.A, 7.4.B60 min The bottleneck effect and the founder effect as kinds of drift, with verified cases, and gene flow that keeps populations from diverging 4, 5
3The Hardy-Weinberg model and its equations 7.5.A60 min The five Hardy-Weinberg conditions, the process each rules out, the model as a null hypothesis, and the two equations 6, 7
4Testing a population, and the dry lab 7.5.A, 7.4.C60 min Testing genotype counts against Hardy-Weinberg with chi-square, allele frequencies over generations as evidence, and the gene pool dry lab 8, dry lab
5Variation in populations 7.11.A60 min How genetic diversity makes a population resilient, why bottlenecked species stay at risk, and an allele that is adaptive in one environment and deleterious in another 9, 10

The revision sheet and the offline copy are in Downloads, for you and the class.

PRACTICE AND REVIEW

Make it for your class, or set it

Build for your class

Already on the site for Unit 7

Card sets on splycr

Open a set to host it as a class game, or make a copy to edit.

How this was made. The course structure, the activities and the teaching choices come from my own classroom, my presentations and my years of teaching IB and AP Biology. I use Claude, an AI, to help write and build the interactive guides, lesson plans, slides and practice questions, and to draw the diagrams and code the interactives. I direct the work and review it against the IB and AP specifications before it goes live. Sources and acknowledgements.