Modelling Natural Selection
Evolution is normally far too slow to watch, populations shift over generations, not class periods. So biologists build models: simplified, sped-up stand-ins that let us see the mechanism in action. Today you become a predator picking "prey" off a background, and watch a population's colour change before your eyes. But a model is only as good as its assumptions, and the real skill is knowing where the model is faithful and where it lies.
Orient yourself
Make a prediction, scan the goals and the key terms, and see which misconception this lesson is built to fix.
Before You Read
warm-upA model is a simplified version of something real, built to help us understand or predict it. A toy car is a model of a car, but you cannot drive to school in it.
Before reading: name one thing a "bead-and-cup" model of natural selection could show well, and one important feature of real evolution it could never capture. Why does this matter for the conclusions we draw?
Learning Intentions
goalsKnow
- What a scientific model is and why we use them
- A method for modelling natural selection (the "predation" model)
- What "validity of a model" means
Understand
- How each part of the model maps to a real biological feature
- Why the model shows the population, not individuals, changing
- The strengths and limitations that affect model validity
Can Do
- Conduct a model investigation and record results across "generations"
- Discuss the validity of the model with specific strengths and limits
- Suggest improvements that would make the model more valid
Scan these before reading
vocabMatch each term to its meaning: use the Key Terms card above.
- Model
- Validity of a model
- Generation
- Controlled variable
- Assumption
- A factor kept constant so the results are a fair test
- A simplified representation of a system, used to understand or predict it
- Something a model takes to be true in order to stay simple
- How well the model represents the real process it stands for
- One round of survival followed by the survivors reproducing
Misconceptions To Fix
watch outThe Predation Model, Method
How to model selection by predation with coloured "prey"
Core Content
The Predation Model, Method
+5 XPHow to model selection by predation with coloured "prey"
A simple, repeatable model of natural selection uses coloured paper dots as prey, a patterned background as the habitat, and you as the predator who removes the easiest-to-see prey each generation.
Materials: a sheet of patterned (e.g. green) fabric or paper as the "habitat", and 100 small paper dots in two colours, say 50 green (camouflaged) and 50 red (visible).
Variables. Independent variable: the number of generations completed. Dependent variable: the percentage of the surviving population that is camouflaged. Controlled variables: hunting time (exactly 15 s), the same background sheet, the same starting number of dots, the same lighting, the same person hunting, and the dots scattered from the same height each time.
Risk assessment. The hazards here are low but they are not zero, and stating them is part of designing an investigation (BI-11WS-02). Small paper dots are a choking hazard around young children and a slip hazard on a hard floor, so keep them contained and sweep up afterwards; scissors used to cut the dots are a cut hazard, so cut away from the body; and reaching quickly across a bench can knock equipment, so clear the bench first.
Method:
- Scatter all 100 dots randomly over the habitat sheet.
- As the "predator", pick up dots one at a time for exactly 15 seconds, you naturally grab the most visible ones first.
- Count and record how many of each colour you "ate" and how many survived.
- The survivors "reproduce": double each survivor (add one new dot of the same colour for each one left).
- Record the new population's colour ratio, then repeat steps 1–4 for several generations.
Model method: scatter coloured prey on a matching habitat, "hunt" for a fixed time (removing the most visible), record survivors by colour, let survivors reproduce, and repeat over generations. The camouflaged colour increases each generation, modelling natural selection by predation.
Pause, copy the highlighted method summary into your book before moving on.
In this model, one round of hunting followed by the survivors reproducing represents one _____.
What the Results Show
Mapping the model onto real natural selection
What the Results Show
+5 XPMapping the model onto real natural selection
We just ran the method. That raises a question: what do the changing colour ratios actually represent? This card answers it → each part of the model maps onto a real piece of natural selection.
As generations pass, the visible colour is removed faster than it reproduces, so the camouflaged colour comes to dominate, exactly mirroring how a selective pressure shifts a real population.
Text description
- Percentage of the population that is the camouflaged colour, by generation.
- Generation 1: 50%. Generation 2: 67%. Generation 3: 80%. Generation 4: 89%.
Model-to-reality map: dot colours = heritable variation; the predator removing visible dots = a selective pressure causing differential survival; doubling survivors = inheritance/reproduction; the rising camouflaged ratio = a change in allele frequency, i.e. evolution. The population (not an individual) changes.
Pause, copy the model-to-reality mapping into your book.
In this model, the act of "the predator removing the most visible dots" represents which part of natural selection?
Apply it: Run and Record the Model
Work the activity in your own book before you read on. This is the first point where you are on your own.
Activity 1
Run and Record the Model
Pattern A, Structured practical record
Carry out (or simulate) the model for four generations. In your book, make a results table with columns for Generation, Green survivors, Red survivors, and % green. Then write one sentence describing the trend and one sentence naming which controlled variable you kept constant (e.g. hunting time, background).
Discussing the Model's Validity
Strengths, limitations, and how to judge a model
Discussing the Model's Validity
+5 XPStrengths, limitations, and how to judge a model
We just saw the model give the expected result. That raises a question: does that make it a valid model? This card answers it → only partly; a valid model is judged by how well it represents reality, not by whether it "works".
To judge a model's validity we weigh how faithfully it represents the real process, listing what it captures well and what it simplifies or ignores.
Strengths (what makes it valid): it correctly shows variation, a selective pressure, differential survival, inheritance and a population-level change over generations, the core logic of natural selection is faithfully represented.
Limitations (what reduces validity):
- No mutation, no new variation can arise during the model; real populations gain variation by mutation.
- The "predator" (you) may have biases or knowledge of the experiment that a real predator would not.
- "Reproduction" is unrealistically simple (just doubling), no mating, no genetics, no chance.
- The environment is fixed, real selective pressures change over time.
- Small numbers and few generations make results prone to chance.
A model's validity is how well it represents the real process. This model validly shows variation, selective pressure, differential survival, inheritance and population change. Its validity is limited because it ignores mutation, uses unrealistic reproduction, has a possibly biased "predator", a fixed environment and small numbers. Improvements (more generations, larger numbers, a randomised predator) raise validity.
Add at least two strengths and two limitations to your notes before the check below.
A limitation of the model is that no new variation (mutation) can appear during it.
Because the model gives the expected result, it must be a fully valid representation of real evolution.
Running more generations with larger numbers would increase the model's validity.
Comparing Various Models
+5 XPThe dot point says models, plural, so you need more than one to compare
We just judged one model against reality. That raises a question: is the predation model the only way to demonstrate natural selection? This card answers it → there are three common classroom models, and each is valid for a different part of the process.
A model is judged against the specific claim it is being used to support, so two models of the same process can both be valid while being valid for different things.
Model A, the predation model (the one you just ran). Coloured prey on a background, hunted by a person. It represents variation, a selective pressure, differential survival and a population shift, and a learner can see all four in one lesson. It has no mutation, its reproduction is a bare doubling, and its "predator" knows the aim of the experiment.
Model B, the allele-frequency bead model. A bag holds beads of two colours representing two alleles. Each generation you draw a fixed number of beads to form the next generation, removing a set proportion of one colour to represent selection against it. It is more valid than Model A for showing selection as a change in allele frequency rather than in visible individuals, and it can model a heterozygote if three bead types are used. It is less valid for showing why one variant is favoured, because the "pressure" is a rule you impose rather than something the learner can see acting.
Model C, a computer simulation. A spreadsheet or program runs hundreds of generations with thousands of individuals, and can include mutation, migration and chance. It is far more valid for long-run behaviour and for showing that outcomes vary between runs, which the hand models cannot do at all. Its weakness is that the learner cannot see the mechanism, only its output, so a wrong assumption written into the code is invisible.
| Model | Most valid for | Least valid for |
|---|---|---|
| A. Coloured prey on a background | Seeing variation, a pressure and differential survival act in one place | Mutation, realistic reproduction, an unbiased predator |
| B. Allele-frequency beads | Showing selection as a shift in allele frequency across generations | Showing why a variant is favoured, since the pressure is imposed by rule |
| C. Computer simulation | Long runs, large populations, chance, mutation, variation between runs | Seeing the mechanism, since only the output is visible |
Various models of natural selection are valid for different claims. A predation model best shows the mechanism acting; an allele-frequency bead model best shows selection as a change in allele frequency; a computer simulation best shows long-run behaviour, chance and mutation. Judge every model against the specific claim it is being used to support, not against reality in general.
Copy the three-model comparison table into your book. An exam answer that discusses the validity of models, plural, needs at least two of them.
Two truths, one lie: three of these statements about judging models are correct. Which one is the lie?
Drive It: Run the Selection Model
+XP for exploringIndividuals do not change; the proportion of each variant in the population does
Use the simulation. On dark bark with selection strength 2×, run ten generations. What has changed?
Use the simulation. True or false: with selection strength 1× (random predation), the proportion of dark moths rises steadily in one direction.
See where it lands in the real world
Why this matters: models run the modern world
Why this matters: models run the modern world
Scientists model things they cannot watch directly all the time, climate decades ahead, the spread of a disease, the path of a spacecraft. Every one of these models is judged the same way you judged yours today: by how faithfully its assumptions match reality and where it breaks down. Learning to say "this model is useful for X but cannot be trusted for Y" is one of the most transferable skills in all of science, and it is exactly what this practical trains.
Describe a practical method for modelling natural selection by predation, and describe how you would process the results to show that selection has occurred.
Describe asks for characteristics and features. A method description must be detailed enough for another student to repeat it, and the second clause asks specifically about processing, not just collecting.
5 marks. Give the materials, the procedure as numbered steps, what is recorded, and how the recorded numbers are turned into evidence of selection.
Use a patterned or coloured background as the habitat, for example a sheet of green fabric, and 100 small paper dots in two colours, 50 green so that they are camouflaged against the background and 50 red so that they are visible. Scatter the 100 dots evenly and at random across the background without looking as you place them.
One student acts as the predator. In a fixed time, for example ten seconds, they pick up as many dots as they can, taking whichever they see first rather than deliberately choosing a colour. The dots taken are removed. The survivors remaining on the background then reproduce: count the survivors of each colour and add an equal number of new dots of the same colour, so each surviving dot is replaced by itself plus one offspring. Scatter the new population and repeat for four generations.
After each generation, record the number of green dots, the number of red dots, and the total. Record these in a table with one row per generation, so the raw counts are preserved as well as any figure calculated from them.
Convert each generation's counts to the percentage of the population that is camouflaged, as the number of green dots divided by the total number of dots, multiplied by 100. Plot that percentage against generation number. Typical results rise generation on generation, for example about 50 per cent at generation 1, 67 per cent at generation 2, 80 per cent at generation 3 and 89 per cent at generation 4, which corresponds to the green-to-red ratio doubling each generation from 1 to 1, through 2 to 1 and 4 to 1, to 8 to 1.
A rising percentage of camouflaged individuals across generations is the evidence that selection has occurred, because the proportion has changed rather than merely the number. In the model, the dot colours represent heritable variation, the predator removing visible dots represents a selective pressure causing differential survival, doubling the survivors represents inheritance through reproduction, and the rising camouflaged proportion represents a change in allele frequency, which is evolution.
Use a coloured or patterned background as the habitat, for example green fabric, together with 100 small paper dots in two colours, 50 green so that they are camouflaged and 50 red so that they are visible, scattered evenly and at random across the background without looking as they are placed. One student acts as the predator and, in a fixed time such as ten seconds, picks up as many dots as they can, taking whichever they see first rather than choosing a colour deliberately; the dots taken are removed. The survivors then reproduce: count the survivors of each colour and add an equal number of new dots of the same colour, so each survivor is replaced by itself plus one offspring. Scatter the new population and repeat for four generations. After each generation record the number of green dots, the number of red dots and the total, in a table with one row per generation, so the raw counts are preserved alongside anything calculated from them. To process the results, convert each generation to the percentage of the population that is camouflaged, as green dots divided by total dots multiplied by 100, and plot that percentage against generation number. Typical results rise generation on generation, for instance about 50 per cent, 67 per cent, 80 per cent and 89 per cent across four generations, corresponding to a green-to-red ratio doubling from 1 to 1 through 2 to 1 and 4 to 1 to 8 to 1. Percentage is the correct measure because the total population changes each generation as survivors reproduce, so the raw number of green dots would rise even with no selection. A rising percentage of camouflaged individuals is therefore the evidence that selection has occurred, and in the model the dot colours represent heritable variation, the predator removing visible dots a selective pressure causing differential survival, the doubling of survivors inheritance through reproduction, and the rising camouflaged proportion a change in allele frequency, which is evolution.
Assess the validity of a bead model, in which coloured beads represent alleles drawn from a gene pool, as a demonstration of natural selection.
Assess asks you to make a judgement of value, quality or outcome. The judgement is the mark-bearing part, and it must follow from the points you raise rather than being announced at the start and left unsupported.
5 marks. State what the model represents, give what it does well, give what it does badly, then judge it against a specific claim rather than in general.
In a bead model, each bead stands for an allele and the container stands for the population's gene pool. Beads are drawn to form the next generation, with the draw weighted so that one allele is removed more often than another, representing selection against it. The quantity being tracked is the proportion of each bead colour in the container, which is allele frequency.
Its main strength is that it represents the quantity that actually changes during evolution. Natural selection is defined as a change in allele frequency in a population, and this model measures allele frequency directly rather than counting visible individuals. It also represents the gene pool as a property of a population rather than of any individual, and with three bead types it can represent a heterozygote and show that a recessive allele persists in heterozygotes even under strong selection against the recessive phenotype.
Its main limitation is that it strips out the biology that gives selection its cause. Nothing in the model shows why one allele is disadvantageous: the weighting is imposed by the person running it rather than emerging from an interaction with an environment. There is no organism, no environment and no selective pressure that a student could identify, so the model can demonstrate that allele frequencies shift without demonstrating what makes them shift. It also usually assumes a constant population size and ignores mutation and migration.
The judgement therefore depends on the claim being supported. To demonstrate that natural selection changes allele frequencies across generations, and that a recessive allele is not eliminated quickly because it is sheltered in heterozygotes, the bead model is valid and is better suited than a predation model, because those are exactly the quantities it represents. To demonstrate why a variant is favoured, or that selection arises from an interaction between organism and environment, it is not valid, and a predation model with camouflage against a background is the better choice.
The model could be improved by pairing it with a predation model in the same lesson, so that one supplies the mechanism and the other the measurement, or by deriving the draw weighting from a stated survival rate rather than from an arbitrary rule, so that students can see where the selection coefficient comes from. Overall, the bead model is a valid representation of the genetic consequence of selection and an invalid representation of its cause, and it should be presented as such rather than as a complete model of natural selection.
In a bead model each bead represents an allele and the container represents the population's gene pool; beads are drawn to form the next generation with the draw weighted so that one allele is removed more often than another, and the quantity tracked is the proportion of each colour, which is allele frequency. Its principal strength is that this is the quantity that actually changes during evolution: natural selection is defined as a change in allele frequency in a population, and the model measures that directly rather than counting visible individuals. It also represents the gene pool as a property of a population rather than an individual, and with three bead types it can represent a heterozygote and show that a recessive allele persists in heterozygotes even under strong selection against the recessive phenotype, which a two-colour predation model cannot show at all. Its principal limitation is that it removes the biology that gives selection its cause: nothing in the model shows why one allele is disadvantageous, because the weighting is imposed by whoever runs it rather than emerging from an interaction with an environment, so there is no organism, no environment and no identifiable selective pressure, and it usually also assumes constant population size and ignores mutation and migration. The judgement therefore depends on the claim. To demonstrate that selection changes allele frequencies across generations, and that a recessive allele is not rapidly eliminated because it is sheltered in heterozygotes, the model is valid and better suited than a predation model, since those are precisely the quantities it represents; to demonstrate why a variant is favoured, or that selection arises from an interaction between organism and environment, it is not valid and a camouflage predation model is the better choice. It could be improved by pairing the two models so that one supplies the mechanism and the other the measurement, or by deriving the draw weighting from a stated survival rate so that students can see where it comes from. Overall the bead model is a valid representation of the genetic consequence of selection and an invalid representation of its cause, and should be presented as such.
Practise independently
Answer every question in your own words first. The model answers are for checking against, not for copying.
Multiple Choice
+5 XPA fresh set drawn from this lesson's question bank, feedback shown immediately. +5 XP per correct · +25 XP all correct
Pick your answer, then rate your confidence, that tells the system what to drill next.
Short Answer, 12 marks
+5 XPEvaluateBand 5(5 marks) 1. A class used coloured dots and a patterned background to model natural selection by predation. Discuss the validity of this model, giving two strengths and two limitations, and suggest one improvement.
2 marks: two valid strengths (features faithfully represented) · 2 marks: two valid limitations · 1 mark: a sensible improvement linked to a limitation
ApplyBand 4(3 marks) 2. In the coloured-prey model, state the independent variable, the dependent variable, and two variables that must be controlled. For one of your controlled variables, explain what would happen to the conclusion if it were not controlled.
1 mark: independent and dependent variables both correctly identified · 1 mark: two valid controlled variables named · 1 mark: explains how failing to control one of them would make the conclusion unsafe
EvaluateBand 5-6(4 marks) 3. Compare the validity of two different models of natural selection: a coloured-prey predation model and an allele-frequency bead model. For each, state one claim it supports well and one claim it does not, then judge which you would choose to demonstrate that natural selection changes allele frequencies.
1 mark: a claim the predation model supports well and one it does not · 1 mark: a claim the bead model supports well and one it does not · 1 mark: an explicit comparison rather than two separate descriptions · 1 mark: a judgement with a reason tied to the specific claim being demonstrated
Show all answers
Multiple choice
MC answers and full explanations are shown inline as you complete each question. Use the retry button to attempt a fresh set from the lesson bank.
Short Answer Model Answer
Q1 (5 marks): The model is valid in several ways: it faithfully represents heritable variation (two dot colours), a selective pressure causing differential survival (the predator removing the most visible dots), inheritance and reproduction (doubling the survivors), and a population-level change over generations (the camouflaged colour rising). These features mean the core logic of natural selection is correctly demonstrated. However, the model has limitations that reduce its validity: it includes no mutation, so no new variation can arise as it would in a real population; reproduction is unrealistically simple (mere doubling with no mating or genetics); the human "predator" may be biased or know the aim of the experiment; the environment is fixed whereas real selective pressures change; and small numbers over few generations make results prone to chance. One improvement would be to run many more generations with much larger numbers of dots and a randomised or blindfolded selection method, which addresses the limitations of small sample size and predator bias, making the model more representative of a real population.
Q2 (3 marks): The independent variable is the number of generations completed, and the dependent variable is the percentage of the surviving population that is the camouflaged colour [1]. Two variables that must be controlled are the hunting time (exactly the same number of seconds each generation) and the background sheet (the same pattern and colour throughout), and others include the starting number of dots, the lighting and the person doing the hunting [1]. If hunting time were not controlled, a longer hunt in a later generation would remove more prey of both colours, so a change in the surviving percentage could be caused by the extra hunting time rather than by selection on colour, and the conclusion that camouflage was favoured would no longer be safe [1].
Q3 (4 marks): The predation model supports the claim that variation, a selective pressure and differential survival together shift a population, because the learner can see all three acting in one place; it does not support any claim about mutation, since no new variation can arise during it, and its reproduction is a bare doubling with no genetics [1]. The bead model supports the claim that selection changes allele frequency across generations, because the beads represent alleles and the changing ratio is measured directly, and with three bead types it can even represent a heterozygote; it does not support a claim about why one variant is favoured, because the selection is a rule the experimenter imposes rather than a pressure the learner can observe acting [1]. Comparing them, the predation model is stronger on mechanism and weaker on genetics, while the bead model is the reverse, so neither is simply more valid than the other, they are valid for different claims [1]. To demonstrate specifically that natural selection changes allele frequencies, I would choose the bead model, because that is the quantity it directly represents and measures, whereas the predation model tracks visible individuals and only implies the underlying alleles [1].
Retrieve and reflect
Do the spaced-review set from memory first, then look back at the prediction you made before you read anything.
Check what actually stuck
A short spaced-review set drawn from this lesson and the lessons before it. Answer from memory first, that is what makes it stick.
Lock it in
bossYou can now model natural selection and, just as importantly, discuss the model's validity with specific strengths and limitations. Next lesson uses real examples to distinguish convergent from divergent evolution.
↩ Back to focus-area hubHow did your thinking change?
You named one thing a model could show well and one it could not. A bead or dot model shows the logic of natural selection clearly, variation, a pressure, differential survival, a shifting population. What it cannot capture is the appearance of new variation by mutation, realistic reproduction and genetics, and a changing environment. This matters because a model that gives the "right" answer can still mislead if we forget its assumptions, which is exactly why discussing validity is a required scientific skill.
If your first answer assumed a working model must be fully realistic, revisit the misconception box and Card 3.