A Punnett square is a simple grid used to predict the possible genetic outcomes when two organisms reproduce. It maps out every combination of genes that offspring could inherit from their parents, letting you see at a glance what traits are likely and how probable each one is. Developed in 1905 by British geneticist Reginald Punnett, it remains one of the most widely taught tools in biology.
How a Punnett Square Works
In its simplest form, a Punnett square is a table divided into four boxes. One parent’s possible gene contributions are listed across the top (one per column), and the other parent’s are listed down the left side (one per row). Each box inside the grid represents one possible combination a child could receive, one gene copy from each parent.
This works because of a basic rule of inheritance called the law of segregation: when a parent produces a sperm or egg cell, only one of their two copies of a gene ends up in that cell. The copy that gets passed on is random. A Punnett square simply lays out all the possibilities side by side so you can count them up.
Say both parents carry one copy of a gene for brown eyes (B) and one copy for blue eyes (b). Across the top you’d write B and b for one parent. Down the side, B and b for the other. Filling in the four boxes gives you BB, Bb, bB, and bb. Three of those four combinations include at least one B, so three out of four offspring would be expected to show the dominant trait (brown eyes), while one in four would show the recessive trait (blue eyes).
Genotypic vs. Phenotypic Ratios
A Punnett square gives you two kinds of information. The genotypic ratio describes the actual gene combinations: in a cross between two carriers (Bb × Bb), you get 1 BB : 2 Bb : 1 bb. The phenotypic ratio describes what you’d actually see in terms of physical traits: 3 showing the dominant trait to 1 showing the recessive trait.
That 3:1 phenotypic ratio is one of the most famous numbers in genetics. Gregor Mendel first observed it in the 1800s when crossing pea plants, though he didn’t have the Punnett square to visualize it. He crossed yellow-seeded plants with green-seeded plants, got all yellow in the first generation, then crossed those offspring with each other and saw yellow and green reappear at a ratio of three to one.
Expanding the Grid for Two Traits
The basic 2×2 grid tracks one trait at a time. When you want to follow two traits simultaneously, the grid expands to 4×4, giving you 16 boxes instead of four. This is called a dihybrid cross.
The reason the grid grows is that each parent can now produce four different types of reproductive cells instead of two. If a parent carries genes for both seed shape (Ss) and seed color (Yy), their sperm or egg cells could carry SY, Sy, sY, or sy. Lining up four options from each parent creates the 16-box grid. Each box still represents an equally likely outcome, so each one has a 1-in-16 chance of occurring.
The classic dihybrid cross between two parents who are both heterozygous for both traits (SsYy × SsYy) produces a phenotypic ratio of 9:3:3:1. In Mendel’s pea experiment, that broke down to 9 round yellow seeds, 3 round green, 3 wrinkled yellow, and 1 wrinkled green. This pattern holds when the two traits are inherited independently of each other, a principle called the law of independent assortment.
Reading Probabilities From the Grid
Each box in a Punnett square carries equal probability. In a 2×2 grid, each box represents a 25% chance. In a 4×4 grid, each box is about 6.25% (1 out of 16). To find the probability of a specific outcome, you count how many boxes match and divide by the total.
Two math rules make this even more flexible. The product rule (the “and” rule) says that if you need two independent events to both happen, you multiply their individual probabilities. If you want to know the chance of a child being both homozygous recessive for trait one AND homozygous recessive for trait two, you’d multiply 1/4 by 1/4 to get 1/16. The sum rule (the “or” rule) works for mutually exclusive outcomes: if either result would satisfy your question, you add their probabilities together. These two rules let you solve complex genetics problems without drawing enormous grids.
When Traits Don’t Follow Simple Dominance
The standard 3:1 ratio assumes one allele is completely dominant over the other. Not all traits work this way, but the Punnett square still applies. You just interpret the boxes differently.
In incomplete dominance, the heterozygous combination produces a blended middle phenotype. Snapdragon flowers are the classic example: cross a red-flowered plant with a white-flowered plant, and the offspring are pink. Cross two pink plants with each other, and the Punnett square predicts 1 red : 2 pink : 1 white. The phenotypic ratio matches the genotypic ratio (1:2:1) because each combination looks different.
Codominance is slightly different. Instead of blending, both alleles show up fully at the same time. Human blood types are a good example. The A and B blood type alleles are codominant with each other (producing AB blood when both are present), while both are dominant over the O allele. A Punnett square for two parents who each carry one A allele and one B allele would predict children with A, AB, and B blood types in a 1:2:1 ratio. In both cases, the grid itself works exactly the same way. What changes is how you translate genotypes into visible traits.
Limitations of the Punnett Square
Punnett squares are powerful for simple, single-gene traits, but most characteristics you can see in the mirror (height, skin color, disease risk) involve dozens or hundreds of genes plus environmental influences. No grid can realistically map all of those interactions. The tool also assumes that genes on different chromosomes sort independently, which isn’t always true. Genes that sit close together on the same chromosome tend to be inherited as a package, a phenomenon called genetic linkage.
The ratios a Punnett square predicts are also probabilities, not guarantees. A 3:1 ratio doesn’t mean that out of four children, exactly three will show the dominant trait. It means each individual child has a 75% chance of showing it. With small numbers, like the two or three children in a typical human family, the actual results can easily deviate from the predicted ratio. The larger the sample size, the closer the real numbers tend to get to the expected ones.