# How to Use a Punnett Square: Monohybrid, Dihybrid and Trihybrid Crosses

A Punnett square is a grid that lists the gametes of one parent along the top and the gametes of the other parent down the side. Each cell in the grid represents one equally likely offspring genotype, under the assumptions that every gamete type is produced at equal frequency and that fertilization is random. The method was popularized by Reginald C. Punnett in the first decade of the 1900s as a compact way to enumerate the gametic combinations predicted by Mendel's theory [1][2].

For a lab report or a genetics problem set, the square does two jobs. It forces you to write out the gametes explicitly, which catches errors in segregation and independent assortment, and it converts those gametes into genotype and phenotype ratios you can compare against observed counts. This article covers the mechanics for one, two and three genes, the ratios each cross produces, and the point at which drawing the grid by hand stops being sensible.

## Quick Answer

- A monohybrid cross Aa x Aa fills a 2 x 2 grid (4 cells) and gives a genotype ratio of 1 AA : 2 Aa : 1 aa and a phenotype ratio of 3 dominant : 1 recessive under complete dominance.
- A dihybrid cross AaBb x AaBb fills a 4 x 4 grid (16 cells) and gives a phenotype ratio of 9:3:3:1 when both genes show complete dominance and assort independently.
- The dihybrid result is the product of two independent 3:1 ratios: (3:1) x (3:1) = 9:3:3:1. The product rule lets you calculate any single class without drawing the grid, for example P(aabb) = 1/4 x 1/4 = 1/16.
- In general, n heterozygous genes give 2^n gamete types per parent and a 4^n-cell grid: 4 cells for one gene, 16 for two, 64 for three.
- A Punnett square gives a probability for each offspring, not a guarantee for a litter or a family.

## Setting Up the Square

Define your symbols before you draw anything. An allele is one of two or more versions of a DNA sequence at a given genomic location [3]. By convention, the dominant allele gets a capital letter and the recessive allele gets the lowercase version of the same letter. A dominant allele is expressed when one copy is present, while a recessive trait appears only when both copies are the recessive allele [4].

The steps are the same regardless of how many genes you track:

1. Write the parental genotypes.
2. Determine the gametes each parent can produce. A parent heterozygous at n loci produces 2^n gamete types.
3. Place one parent's gametes across the top and the other parent's gametes down the side.
4. Fill each cell by combining the row and column gametes.
5. Count cells by genotype, then collapse genotypes into phenotypes using the dominance relationships.

Two assumptions sit underneath the whole procedure. Each gamete type must be produced at equal frequency, and fertilization must be random. When genes are linked, the first assumption fails for the combinations of alleles at those loci, and the observed ratios depart from the expected ones.

## Monohybrid Cross

A monohybrid cross tracks one gene with two alleles. Cross two heterozygotes, Aa x Aa, and each parent makes two gamete types, A and a, at equal frequency.

| | A | a |
|---|---|---|
| **A** | AA | Aa |
| **a** | Aa | aa |

Four cells, three genotypes. The genotype ratio is 1 AA : 2 Aa : 1 aa. Because A is completely dominant, AA and Aa look the same, so the phenotype ratio is 3 dominant : 1 recessive.

Mendel's own F2 seed-shape counts were 5,474 round and 1,850 wrinkled, a ratio of 2.96 to 1. Against the expected 3:1 split of 7,324 seeds (5,493 round : 1,831 wrinkled), chi-square = 0.263 with 1 df and p = 0.61, so the data fit 3:1. That is what a good fit looks like: the observed ratio is close to 3:1 but not exactly 3:1, and the deviation is within what sampling noise produces.

## Worked Example: Dihybrid Cross

Take two parents heterozygous at both loci, AaBb x AaBb, with complete dominance at each gene and independent assortment between them. Each parent produces four gamete types: AB, Ab, aB and ab. That gives a 4 x 4 grid with 16 cells.

| | AB | Ab | aB | ab |
|---|---|---|---|---|
| **AB** | AABB | AABb | AaBB | AaBb |
| **Ab** | AABb | AAbb | AaBb | Aabb |
| **aB** | AaBB | AaBb | aaBB | aaBb |
| **ab** | AaBb | Aabb | aaBb | aabb |

Counting the 16 cells by genotype gives nine genotypes in the ratio 1:2:1:2:4:2:1:2:1. Collapsing to phenotypes under complete dominance gives the familiar 9:3:3:1 ratio: 9 A_B_, 3 A_bb, 3 aaB_, 1 aabb.

You can reach the same numbers without the grid. The dihybrid result is the product of two independent 3:1 ratios, (3:1) x (3:1) = 9:3:3:1, and the product rule gives any single class directly. The probability of aabb is 1/4 x 1/4 = 1/16, which matches the single aabb cell in the grid. If you want to check a full set of expected values against observed counts, the [Punnett Square Calculator](/tools/punnett-square-calculator) generates the grid and the ratios for you, and the [Chi-Square Calculator](/tools/chi-square-calculator) tests whether your observed numbers fit.

A test cross is the practical use of this logic. Cross AaBb x aabb and you get four phenotype classes in a 1:1:1:1 ratio. That is how a breeder checks whether a dominant-looking individual is heterozygous: if any recessive offspring appear, the parent carried the recessive allele.

## Trihybrid Cross and the Product Rule

A trihybrid cross, AaBbCc x AaBbCc, has three heterozygous genes. Each parent makes 2^3 = 8 gamete types, so the grid has 8 x 8 = 64 cells. The phenotype ratio is 27:9:9:9:3:3:3:1, which is (3:1)^3 expanded.

The pattern generalizes cleanly:

| Genes (n) | Gamete types per parent (2^n) | Grid cells (4^n) |
|---|---|---|
| 1 | 2 | 4 |
| 2 | 4 | 16 |
| 3 | 8 | 64 |
| 4 | 16 | 256 |

Beyond two genes, hand-drawing the grid is slow and error-prone. Use the product rule instead. For AaBbCc x AaBbCc, the probability of aabbcc is 1/4 x 1/4 x 1/4 = 1/64, and the probability of the fully dominant phenotype A_B_C_ is 3/4 x 3/4 x 3/4 = 27/64. Those two numbers alone account for the 27 and the 1 in the 27:9:9:9:3:3:3:1 ratio. A calculator or a short script handles the rest.

## When the Ratios Change

The 3:1 and 9:3:3:1 ratios depend on complete dominance and independent assortment. Change the dominance relationship and the phenotype ratio changes while the genotype ratio stays the same. Break independent assortment and both ratios change.

Incomplete dominance: a red x white snapdragon F1 is pink, and the F2 phenotypes follow the genotype ratio of 1 red : 2 pink : 1 white. The heterozygote is distinguishable, so phenotype and genotype ratios coincide.

Codominance: both alleles are expressed, as in the human ABO blood group, where IA IB individuals have type AB.

Epistasis: one locus masks another. In Labrador retrievers the E locus masks the B locus, so ee dogs are yellow regardless of B. A BbEe x BbEe cross gives 9 black : 3 chocolate : 4 yellow instead of 9:3:3:1.

Linkage: genes close together on the same chromosome do not assort independently. Observed dihybrid ratios depart from 9:3:3:1, and recombinant classes are rarer than expected.

Sex-linked genes: X-linked loci need separate grids for sons and daughters, because males receive only one X chromosome. Mendelian inheritance patterns include autosomal dominant, autosomal recessive and X-linked patterns, and these are described for human genetic conditions by MedlinePlus [5][6].

## Common Mistakes

- **Forgetting to write out the gametes first.** Skipping straight to the grid leads to duplicated or missing gamete types. List them before you draw.
- **Writing gametes with two alleles of the same gene.** A gamete carries one allele per gene, such as Ab. A gamete listed as AA or Bb means segregation was skipped.
- **Confusing genotype ratio with phenotype ratio.** The monohybrid genotype ratio is 1:2:1; the phenotype ratio is 3:1. Report the one your question asks for.
- **Assuming the ratio guarantees the outcome.** A 3:1 expectation can easily give 4:0 or 2:2 in four offspring. For four offspring the chance that all four show the dominant phenotype is (3/4)^4 = 0.316.
- **Applying 9:3:3:1 to linked genes.** Check for independent assortment before you use the ratio. If the loci are linked, the expected ratio does not hold.
- **Using one grid for X-linked traits in both sexes.** Sons and daughters inherit X-linked alleles differently, so they need separate squares.

## Limitations

The square assumes equal gamete frequencies and random fertilization. Both can fail. Linkage violates independent assortment for genes on the same chromosome. Meiotic drive, selection acting on gametes, and small sample sizes all push observed ratios away from expectation.

The square also says nothing about penetrance or expressivity. A genotype can be present without producing the expected phenotype, and the same genotype can produce a range of phenotypes. For a single offspring, the square gives a probability, not a prediction. Teaching studies use simulations of monohybrid crosses precisely to connect Punnett-square probabilities with the variation seen in real samples [7].

Finally, the grid grows exponentially. With four heterozygous genes you are looking at 256 cells, and the arithmetic invites mistakes. Past two genes, use the product rule or a calculator.

## Frequently Asked Questions

### How do I do a Punnett square for a monohybrid cross?

Write the parental genotypes, list the two gametes each parent can make, and place one parent's gametes across the top and the other's down the side. Fill the four cells by combining row and column gametes. For Aa x Aa you get 1 AA : 2 Aa : 1 aa, which is 3 dominant : 1 recessive under complete dominance.

### What is the phenotype ratio for a dihybrid cross?

For AaBb x AaBb with complete dominance at both loci and independent assortment, the phenotype ratio is 9:3:3:1. It comes from multiplying two independent 3:1 ratios: (3:1) x (3:1) = 9:3:3:1. The genotype ratio across the 16 cells is 1:2:1:2:4:2:1:2:1.

### How many cells does a trihybrid cross need?

A trihybrid cross AaBbCc x AaBbCc has 8 gamete types per parent, so the grid has 8 x 8 = 64 cells. The phenotype ratio is 27:9:9:9:3:3:3:1. In general, n heterozygous genes give 2^n gamete types and a 4^n-cell grid, which is why calculators or the product rule beat hand-drawn grids beyond two genes.

### Can I skip the grid and use a formula?

Yes, for single classes. The product rule gives the probability of any combination directly: for AaBb x AaBb, P(aabb) = 1/4 x 1/4 = 1/16. Use the grid when you need the full distribution or when you are learning the method, and switch to the product rule once the logic is clear.

### Does a 3:1 ratio mean three out of every four offspring will show the dominant trait?

No. The ratio is a probability for each offspring, not a guarantee for a group. In four offspring you can see 4:0 or 2:2 and still be consistent with a 3:1 expectation. The chance that all four show the dominant phenotype is (3/4)^4 = 0.316.

## References

1. [Edwards AWF. Punnett's square. Studies in History and Philosophy of Biological and Biomedical Sciences, 2012](https://doi.org/10.1016/j.shpsc.2011.11.011)
2. [Wimsatt WC. The analytic geometry of genetics: part I. Archive for History of Exact Sciences, 2012](https://doi.org/10.1007/s00407-012-0096-7)
3. [NHGRI Talking Glossary: Allele](https://www.genome.gov/genetics-glossary/Allele)
4. [NHGRI Talking Glossary: Dominant Traits and Alleles](https://www.genome.gov/genetics-glossary/Dominant-Traits-and-Alleles)
5. [NHGRI Talking Glossary: Mendelian Inheritance](https://www.genome.gov/genetics-glossary/Mendelian-Inheritance)
6. [MedlinePlus Genetics: What are the different ways a genetic condition can be inherited?](https://medlineplus.gov/genetics/understanding/inheritance/inheritancepatterns/)
7. [Butler N. Simulating Monohybrid Punnett Squares. The American Biology Teacher, 2025](https://doi.org/10.1525/abt.2025.87.1.55)

## Related Articles

- [Leveraging Mendelian Inheritance in Germline Variant Calling: Trio Analysis and De Novo Mutation Detection](/knowledge/bioinformatics/leveraging-mendelian-inheritance-in-germline-variant-calling-trio-analysis-and-de-novo-mutation-dete)
- [How to Use gnomAD Allele Frequencies for Variant Filtering: Thresholds, Subpopulations, and Best Practices](/knowledge/bioinformatics/how-to-use-gnomad-allele-frequencies-for-variant-filtering-thresholds-subpopulations-and-best-practi)
- [Genotype-Phenotype Correlation in Antimicrobial Resistance](/knowledge/bacteria/general/genotype-phenotype-correlation-in-antimicrobial-resistance-why-some-resistance-genes-do-not-predict)