A Punnett square is a simple diagram used to predict the possible genetic combinations that can result when two parents reproduce. It is one of the most useful tools for learning how traits are inherited because it turns the rules of genetics into a visual model.
Punnett squares do not tell you exactly what a child will inherit. Instead, they show the possible combinations of gene variants, called alleles, and—when the underlying assumptions are appropriate—the probability of each combination.
What a Punnett square shows
To understand a Punnett square, it helps to distinguish three related ideas: genes, alleles, and traits.
A gene is a segment of DNA that contributes to a biological characteristic or function. A gene can exist in different versions, called alleles. For example, a gene associated with a particular characteristic might have two different alleles.
For many genes in humans, a person inherits one allele from each biological parent. The pair of alleles makes up the person’s genotype, meaning their genetic combination for that gene. The observable characteristic associated with that genotype is the phenotype.
A Punnett square represents the alleles that each parent could pass to an offspring. Each box in the square combines one possible allele from one parent with one possible allele from the other.
How to make a basic Punnett square
Consider a simplified example in which a gene has two alleles, represented by A and a. Suppose both parents have the genotype Aa.
Each parent can pass either A or a to an offspring. Place one parent’s possible alleles across the top of the square and the other parent’s along the side:
| A | a | |
|---|---|---|
| A | AA | Aa |
| a | Aa | aa |
The four boxes represent four equally likely allele combinations under the assumptions of this simple model.
The possible offspring genotypes are therefore:
- AA: 1 out of 4, or 25%
- Aa: 2 out of 4, or 50%
- aa: 1 out of 4, or 25%
These percentages describe probabilities across many comparable inheritance events. They do not mean that, in a family with four children, exactly one must be AA, two must be Aa, and one must be aa. Each pregnancy is a separate genetic event.
Dominant and recessive alleles
Many introductory Punnett-square problems involve dominant and recessive alleles.
A dominant allele is one whose associated phenotype can be expressed when only one copy is present. A recessive allele generally produces its associated phenotype only when two copies are present in the relevant genotype.
In the example above, if A is dominant and a is recessive, both AA and Aa would have the dominant phenotype, while aa would have the recessive phenotype.
That produces a different set of probabilities at the phenotype level:
- 75% dominant phenotype
- 25% recessive phenotype
This distinction matters because genotype and phenotype are not the same thing. Two people can have different genotypes but the same phenotype when one allele is dominant over the other.
A dominant allele is not necessarily a stronger, healthier, or more common allele. “Dominant” describes how an allele affects phenotype in a particular genetic context.
Homozygous and heterozygous genotypes
Two terms commonly appear when reading Punnett squares.
A homozygous genotype contains two copies of the same allele. AA and aa are homozygous.
A heterozygous genotype contains two different alleles. Aa is heterozygous.
These terms describe the genotype, not whether a trait is dominant or recessive. A heterozygous individual, for example, can carry a recessive allele without showing the recessive phenotype in a simple dominant-recessive system.
What the percentages really mean
The probability in a Punnett square describes the likelihood of a particular genetic outcome for an individual offspring.
If two heterozygous parents have a 25% probability of producing an offspring with genotype aa, that does not mean every fourth child will have that genotype. Random inheritance can produce different results in actual families.
The same principle applies to repeated events. If a particular outcome has a 25% probability for each pregnancy, the probability remains 25% for the next pregnancy, assuming the same parental genotypes and genetic circumstances. Previous children do not change the alleles available in the parents’ germ cells for a future pregnancy.
Why alleles are separated during reproduction
Punnett squares are based on a fundamental feature of sexual reproduction. During the formation of eggs and sperm, the two copies of a gene carried by an individual are separated so that each mature reproductive cell generally receives one allele.
When an egg and sperm combine during fertilization, their alleles come together again, restoring the pair in the resulting offspring.
For a parent with genotype Aa, this means that a basic Punnett-square model treats A and a as possible alleles that the parent can contribute. The square then combines those possibilities with the corresponding possibilities from the other parent.
This is rooted in the principle known as the law of segregation, one of the inheritance patterns described by Gregor Mendel.
Punnett squares with two different genes
Punnett squares can also be used to examine inheritance involving two genes at once. These dihybrid crosses require a larger grid because each parent can produce several possible allele combinations in their reproductive cells.
For example, a parent with genotype AaBb can potentially contribute combinations such as AB, Ab, aB, or ab when the two genes assort independently.
The larger square combines each possible contribution from one parent with each possible contribution from the other. This can reveal probabilities for particular combinations of genotypes or phenotypes.
However, independent assortment is an important assumption. Genes located close together on the same chromosome may be inherited together more often than a simple independent-assortment model predicts because of genetic linkage.
Punnett squares do not work the same way for every trait
The classic dominant-recessive Punnett square is useful, but it is a simplified model. Many human traits do not follow a single-gene, two-allele pattern with complete dominance.
In incomplete dominance, neither allele completely determines the phenotype in a heterozygote, producing an intermediate phenotype in the relevant system.
In codominance, both alleles are expressed in the phenotype of a heterozygote. The ABO blood-group system provides a familiar example: the A and B alleles can both be expressed in a person with type AB blood.
Some traits are influenced by multiple genes, environmental factors, or both. Other inheritance patterns involve sex-linked genes, mitochondrial DNA, multiple alleles, or interactions among genes. In such cases, a basic four-box Punnett square may be inadequate or misleading.
Sex-linked inheritance requires a different setup
Some genes are located on the sex chromosomes rather than on the non-sex chromosomes, which are called autosomes. When inheritance involves a gene on a sex chromosome, the sex chromosome carried by each parent becomes part of the calculation.
For an X-linked gene, for example, the alleles carried on an X chromosome can have different consequences depending on the chromosome combination of the offspring. A Punnett square can still be useful, but the alleles must be labeled in a way that shows their location on the chromosome.
This is one reason a Punnett square should be viewed as a model rather than a universal formula for human inheritance.
A Punnett square predicts possibilities, not certainties
The greatest value of a Punnett square is that it makes inheritance probabilities visible. It helps answer questions such as: What genotypes could an offspring have? What proportion of offspring would be expected to have a particular genotype? Under a simple dominant-recessive model, what phenotypes could result?
It cannot predict the exact genetic makeup of a particular future child. It also cannot, by itself, account for every biological factor that influences a person’s traits.
For simple inheritance patterns, however, the method is remarkably effective. Start with the parents’ genotypes, identify the alleles each parent can pass on, combine those possibilities in the grid, and then interpret the resulting genotypes according to the inheritance pattern involved.
That is the central logic behind a Punnett square: possible allele from one parent + possible allele from the other parent = possible offspring genotype.
