# Predicting Traits with Punnett Squares

> For one-gene crosses, Punnett squares predict chances, not exact offspring counts. You follow a hidden allele from purple-flowered parents into their offspring, build and test Punnett squares, and compare their predictions with Gregor Mendel’s real pea counts. Afterwards you can predict each genotype and phenotype for one trait and explain why real counts usually come close to the predicted ratio without matching it exactly.

Course: Life Science (Heredity). Level: grades 7–12. About 40 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/predicting-traits/

Builds on: Reproduction and Meiosis · DNA: The Code of Life

## What you'll learn

- You will explain genotype (the pair of alleles a living thing carries) and phenotype (the trait that shows), and how a dominant allele masks a recessive allele, using letters such as P and p.
- You will use a Punnett square for one trait to predict the probability and the ratio of genotypes and phenotypes among offspring, knowing that each sex cell carries either allele of a pair with a 50 percent chance.
- You will read Mendel’s real pea-plant counts and make a claim, with evidence and reasoning, about whether they fit the 3 to 1 ratio a Punnett square predicts, and why real counts are close to the prediction but rarely exact.

The first film showed how two purple-flowered Pp pea plants can have a white-flowered offspring. Could a gardener predict, before planting a single seed, what share will bloom white?

What do you think? Decide before you read.

1. Offspring are a blend of their parents, so two purple parents can only have purple offspring. Agree Disagree **Scientists disagree:** Mendel crossed pea plants and counted thousands of offspring. For the seven traits he tracked, he found no in-between forms. A trait that vanished in one generation came back unchanged in the next. A blend could not come apart again like that, so traits must pass on as separate units. What scientists think **Scientists disagree:** Mendel crossed pea plants and counted thousands of offspring. For the seven traits he tracked, he found no in-between forms. A trait that vanished in one generation came back unchanged in the next. A blend could not come apart again like that, so traits must pass on as separate units.
2. If a Punnett square says 1 in 4, then exactly one out of every four offspring will have the trait. Agree Disagree **Scientists disagree:** Mendel recorded one pea plant with 43 round seeds and only 2 wrinkled ones, far from 1 in 4. When he added up thousands of seeds from many plants, the count came much closer to the prediction. What scientists think **Scientists disagree:** Mendel recorded one pea plant with 43 round seeds and only 2 wrinkled ones, far from 1 in 4. When he added up thousands of seeds from many plants, the count came much closer to the prediction.
3. A recessive allele is weak or damaged, and it disappears when it is hidden. Agree Disagree **Scientists disagree:** In Mendel’s crosses, hidden traits came back unchanged in later generations, so the allele was still there. The word recessive only tells you whether a trait shows when a dominant allele is also present. What scientists think **Scientists disagree:** In Mendel’s crosses, hidden traits came back unchanged in later generations, so the allele was still there. The word recessive only tells you whether a trait shows when a dominant allele is also present.

## Explain: alleles can hide but do not disappear

- allele
- dominant
- recessive
- genotype
- phenotype
- Punnett square
- probability
- ratio

A pea plant gets one copy of each gene from each parent, so each body cell carries two copies of the flower-color gene. An **allele** is one version of a gene. In this one-gene model, the flower-color gene has a purple allele, P, and a white allele, p.

- A **dominant** allele shows its trait even when there is only one copy. It gets a capital letter.
- A **recessive** allele shows its trait only when both copies are recessive. It gets a small letter.

A dominant allele **masks** a recessive one: it hides it from view without destroying it.

Your **genotype** is the pair of alleles you carry, such as Pp. Your **phenotype** is the trait that shows, such as purple flowers.

| Genotype | Alleles | Phenotype |
| --- | --- | --- |
| PP | two same alleles (homozygous) | purple |
| Pp | two different alleles (heterozygous) | purple |
| pp | two same alleles (homozygous) | white |

Sex cells carry only one allele of each pair, as you saw in [that lesson](https://stemhop.com/lessons/reproduction-and-meiosis/). A PP plant can only pass on P. A pp plant can only pass on p. A Pp plant passes on P or p, with a 50 percent chance for each.

**Key idea:**

A dominant allele can hide a recessive allele, but the recessive allele is still carried and can be passed to offspring.

**Predict.** A PP plant is crossed with a pp plant. Which allele can each parent pass, and what color will the offspring be?

> **Check your prediction.** The parents pass P and p, so every offspring is Pp and purple but carries p.

## Explain: a Punnett square shows the chances

A **Punnett square**, the grid the first film built, shows how the sex cells of two parents can join. Here it is for two Pp parents.

|  | P | p |
| --- | --- | --- |
| **P** | PP | Pp |
| **p** | Pp | pp |

*Interactive on the lesson page:* Tap to enlarge one parent's sex cells (P and p)the other parent's sex cells (P and p)PPPppp one parent's sex cells (P an

**Probability** is the chance that something will happen. Every box needs one particular allele from each Pp parent. Each allele has a 1 in 2 chance. So every box has the same chance: 1/2 × 1/2 = 1/4.

One box is PP, two are Pp and one is pp, so the genotype chances are 1/4 PP, 2/4 Pp and 1/4 pp. PP and Pp are purple, while pp is white. A **ratio** compares two numbers, such as 3 purple to 1 white.

| What you predict | As fractions | As percents | As a ratio |
| --- | --- | --- | --- |
| Genotypes | 1/4 PP, 2/4 Pp, 1/4 pp | 25, 50, 25 | 1 PP : 2 Pp : 1 pp |
| Phenotypes | 3/4 purple, 1/4 white | 75, 25 | 3 purple : 1 white |

This 3 : 1 prediction needs two Pp parents.

**The one idea:**

For one-gene crosses, Punnett squares predict chances, not exact offspring counts. Real counts usually land near the predicted ratio but rarely match it exactly.

Here a one-gene cross means one gene, two alleles (one dominant) and known parent genotypes.

**Predict.** Cross a Pp plant with a pp plant. What share of the boxes are pp and white?

> **Check your prediction.** The pp parent passes p. The Pp parent passes P or p. Half the boxes are Pp and half are pp.

Four seeds from two Pp parents might all bloom purple; the chance is about 32 percent. Mendel’s rules explain how many inherited traits are passed on; later courses add other patterns.

- Two purple parents can have a white offspring when blank.
- The chance of a white offspring from two Pp parents is blank, or blank percent.

## Explore: breed purple and white pea plants

Choose PP, Pp or pp for each parent. The simulation fills the square and grows 4, 40 or 1,000 offspring. Each one is made by a random draw of one sex cell from each parent. A bar graph puts your real counts next to the predicted ones.

*Interactive on the lesson page:* Breed purple and white pea plants

**Try this.**

1. Cross two Pp plants. Predict how many of 4 offspring will be white. Grow them, then repeat three times.
2. Grow 40, then 1,000. Predict first: will the share of white plants move closer to one in four, or stay just as scattered?
3. Cross a PP plant with a Pp plant. Make the prediction below, then test it.

**Predict.** In the PP × Pp cross, can any offspring be white?

> **Check your prediction.** No. The PP parent always passes P, so every offspring gets at least one P and is purple. Some carry a hidden p. Two purple parents give white offspring only when both are Pp.

Your groups of 4 probably wandered from one in four. The second film shows why chance and count differ.

**Film transcript: Why Real Counts Wander**

### Film transcript: Why Real Counts Wander

The first film built a Punnett square for Pp parents. Why might their real offspring counts miss its predicted ratio? A square gives each new plant a chance. A real group records what happened after those separate chances played out.

Think of the square as a chance map, not a packing list. For every new offspring, one box is selected by the parents’ allele contributions. Then the process starts again. Earlier flowers do not remove boxes or force later results. The map resets without remembering the last selection. Watch a small computed example. Several selectors land, and flowers fill the short row. The colors form an uneven group. That does not make the square wrong. A small group leaves plenty of room for chance to swing the count. Another short run could lean differently. The short row has leaned toward one color. The next slot is still empty. Does the earlier run force the next flower to balance the group, or does the same chance map apply again?

Make your prediction. The same chance map applies again. The next result is not a correction chosen by the earlier results. Each offspring is a separate chance. The group may move nearer the prediction or farther away before later results add more evidence. Now extend the computed run. More flower slots stream in while the bars rescale. The observed share still wiggles, but its swings usually shrink. With more separate outcomes, an unusual short streak has less control over the whole group. The curve approaches its target unevenly, not in a straight march.

Mendel tested the pattern with real pea plants. He counted offspring for several traits, not just one small family. Place each dominant-to-recessive ratio beside the square’s target. The rows gather near the target, but they do not land exactly together. That spread is the evidence to examine. A prediction is useful when many real results gather near it, even without an exact match. No single row must sit exactly on the target. In the lesson, you will compare counts, ratios and group sizes yourself. The square remains a one-gene model. Weather, soil, lost seeds, and more complex inheritance are outside it. Within its scope, it gives a clear target for repeated outcomes.

Larger groups usually reveal that target more clearly than small groups do.

From the film

- Larger groups usually come closer to a Punnett square’s predicted ratio.
- Each offspring is a new chance, not a step in a fixed schedule.
- Small groups can land far from the predicted ratio.
- Mendel’s full pea counts came close to the predicted ratio.

## Elaborate: Mendel’s pea counts

Now check Mendel’s counts yourself. He crossed plants differing in one trait. Their first offspring all showed the dominant trait. He let those hybrids fertilize themselves and counted the next generation.

| Trait (dominant vs recessive) | What was counted | Dominant (number) | Recessive (number) | Ratio (dominant ÷ recessive) |
| --- | --- | --- | --- | --- |
| Seed shape: round vs wrinkled | seeds | 5,474 | 1,850 | 2.96 : 1 |
| Seed color: yellow vs green | seeds | 6,022 | 2,001 | ? |
| Flower color: purple vs white | plants | 705 | 224 | ? |
| Pod shape: full vs pinched | plants | 882 | 299 | 2.95 : 1 |
| Unripe pod color: green vs yellow | plants | 428 | 152 | 2.82 : 1 |
| Flower position: along the stem vs at the tip | plants | 651 | 207 | 3.14 : 1 |
| Stem length: tall vs short | plants | 787 | 277 | 2.84 : 1 |

Source: Gregor Mendel’s pea paper, read at two meetings in Brno in 1865 and printed in 1866 (English translation by Abbott and Fairbanks, 2016). The plants grew in his garden beds and in some pots. The paper does not give a year for each row. For flower color, Mendel sorted plants by seed-coat color, which goes with flower color. The ratios were calculated here.

**Predict.** If a group has 3 dominant offspring for every 1 recessive offspring, is the recessive share 1/3 or 1/4 of the total?

> **Check your prediction.** It is 1/4. The total has four parts: three dominant plus one recessive. A fraction compares one group with that total.

1. Find the two missing ratios. Divide the dominant count by the recessive count and round to two decimal places.
2. Which row is farthest from 3 : 1? Which row has the fewest offspring counted? What do you notice?
3. What percent of the flower-color plants were white? Divide 224 by the total, 929, and compare with 25 percent.

Claim, evidence, reasoning

Can a Punnett square predict the traits of pea offspring? Make a claim using Mendel’s counts.

Claim Your answer to the question, in one sentence.

Start with: Mendel’s counts show that a Punnett square blank.

Evidence The data that support your claim.

Start with: For the seven traits, the ratios ranged from blank to blank. For flower color, blank of 929 plants were white, about blank percent.

Reasoning Why the evidence supports the claim, using the science.

Start with: The counts are close to 3 : 1 because blank. They are not exactly 3 : 1 because blank.

> **Compare with a model answer.** A Punnett square predicts Mendel’s pea offspring well, but not exactly. For all seven traits, the ratio was between 2.82 : 1 and 3.15 : 1, close to the predicted 3 : 1. For flower color, 224 of 929 plants were white, about 24.1 percent, near the predicted 25 percent. The counts are close because each offspring had a 1 in 4 chance of getting two recessive alleles. They are not exact because each offspring is a separate chance. One limit is that all the data come from one scientist, one kind of plant and one garden.

Saved in this browser only.

### Design problem

Plan a bead model to test whether the number of offspring changes how close a cross comes to 3 : 1.

**Predict.** What would go wrong if you kept each drawn bead out of its bag?

> **Check your prediction.** The chances would change after each draw. Parents do not run out of an allele after passing it on, so replace both beads.

> **One way to do it.** Make two paper bags, one for each Pp parent. Put 20 purple beads (P) and 20 white beads (p) in each bag. All beads are the same size, about 1 centimetre across, so you cannot tell them apart by feel.
>
> - **Make one offspring:** without looking, draw one bead from each bag. Record the pair, then put both beads back and shake the bags.
> - **Independent variable:** the number of offspring you make: 4, 20 or 100.
> - **Dependent variable:** the percent of offspring that are pp (white).
> - **Constants:** the same bags, the same 20 and 20 beads, and putting beads back after every draw.
> - **Control:** a PP bag (40 purple beads) paired with a Pp bag. It should never give a pp offspring, which checks that your method works.
> - **Repeated trials:** make each group size 5 times and find the mean percent of white.
> - **Safety:** pick up dropped beads so no one slips, and keep beads away from your mouth.
>
> If the 100-offspring groups usually land closer to 25 percent than the 4-offspring groups, sample size matters. The bead model is not a real plant: it has one gene and no weather, soil or lost seeds.
>
> *Figure (AI-generated illustration).* The bead model of a Pp by Pp cross. Each blind draw of one bead from each bag makes one offspring, and the beads go back before the next draw. Image: two paper bags of mixed purple and white beads on a table, a hand reaching into each. One purple and one white bead, just drawn, sit in front as one offspring, with arrows showing them going back into the bags. An inset shows the control: a bag of only purple beads beside a mixed bag.

## Evaluate: check your thinking

Back to the gardener: both purple parents were Pp, so each offspring had a 1 in 4 chance of getting p from both and blooming white. Remember: for one-gene crosses, Punnett squares predict chances, not exact offspring counts.

**Science practices you used:** making and testing predictions, using a Punnett square as a model and naming its limits, calculating ratios and percents from real counts, and making a claim with evidence and reasoning.

---

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