Punnett Square Calculator

How to use the punnett square calculator

1. Type Parent 1's genotype using gene letters: a capital and lowercase letter for a heterozygote (Bb), two capitals or two lowercase letters for homozygotes (BB or bb). prefill →
2. Add a second gene by concatenating pairs, e.g. BbRr for a dihybrid parent. Parent 2 must use the same number of genes. prefill →
3. Choose complete dominance for standard Mendelian traits, or the codominance option when heterozygotes show their own phenotype (blood types, roan coats).
4. Read the grid: rows are Parent 1's gametes, columns are Parent 2's. Each cell is one possible offspring genotype.
5. Use the ratio lists beneath the grid for probability statements: e.g. 'a 25% chance of bb'.

Punnett Square: the formula & its source

Each parent produces gametes by the law of segregation: for every heterozygous locus, half the gametes carry the dominant allele and half carry the recessive allele. A parent with n heterozygous loci produces 2ⁿ equally frequent gametes. The cross is the Cartesian product of the two parental gamete sets, so a one-gene cross is a 2×2 grid (4 cells), a two-gene cross is a 4×4 grid (16 cells), and every cell is equally probable.

Genotype ratios are the counts of each allele combination divided by the total number of cells. Phenotype ratios collapse genotypes that express the same trait: under complete dominance, Bb and BB map to the same phenotype, giving the classic 3:1 ratio for a monohybrid heterozygous cross.

Punnett Square: worked example

Worked example: Bb × Bb. Both parents are heterozygous, so each makes two gametes: B and b. The 2×2 grid yields BB, Bb, Bb, bb. Genotypes are 1 BB : 2 Bb : 1 bb (25% / 50% / 25%). Phenotypes are 3 dominant : 1 recessive, a 75% chance the offspring shows the dominant trait and 25% the recessive trait.

Example B: test cross against bb

Example B: test cross against bb. Set Parent 1 to Bb and Parent 2 to bb with complete dominance. The grid halves to four cells, two of them Bb, so half the offspring show the dominant trait and half show the recessive one. This is how breeders read the hidden genotype of a dominant-looking parent, and the same reasoning works for a carrier test in a human pedigree. A litter of four can still come back all dominant, so score dozens of offspring before you conclude anything.

How the grid builds the ratio

Each parent carries two alleles at a locus, and segregation hands exactly one of them to each gamete. A heterozygote therefore produces two equally frequent gamete types, while a homozygote produces one. Heterozygous loci multiply: a parent with n of them makes 2n gamete kinds, which is why BbRr contributes B, b, R and r in all four combinations.

The square is simply the full pairing of one parent's gametes against the other's. One gene fills a 2 by 2 grid of four equally likely cells; a second gene grows each edge to four and fills sixteen. Because every gamete is common and the loci combine freely, each cell carries the same weight: one quarter on the small grid, one sixteenth on the large one. Tally the cells and the ratios fall out on their own.

Rows and columns follow a fixed convention: Parent 1's gametes label the rows and Parent 2's label the columns, so any cell is read as one allele from each side. A two-gene parent writes its pairs concatenated, as in BbRr, and the calculator splits them back into loci before pairing anything, which is why both sides must carry the same number of pairs. Because the grid is the full Cartesian product of the two gamete sets, every combination appears exactly once and nothing appears twice. That single property is what lets a count of cells stand in for a probability: four cells means one quarter each, sixteen cells means one sixteenth each, and the ratio lists underneath are just those cell counts grouped by phenotype.

What each cell is worth

Genotype ratios count allele combinations directly. Phenotype ratios merge the genotypes that look identical, which is what hides the 1:2:1 genotype split inside the classic 3:1 phenotype result. Switch the inheritance setting to codominance and the heterozygote gets its own label, so both ratios stay visible side by side.

Every figure on the page is a long-run frequency rather than a promise about one family. Two carrier parents (Aa × Aa) hold a one-in-four chance per pregnancy of an affected child, a half of a carrier child and a quarter of a child who inherited neither copy. Those are the same numbers a counselling session quotes, and they apply whether you imagine four offspring or four hundred, provided the sample is large enough for the counts to settle near the expectation.

The probability lists beneath the grid exist for careful wording. Where the output reads 25% bb, the safe sentence is that a quarter of many such offspring would be bb; the unsafe one is that the fourth child will be. For Bb × Bb the four cells read BB, Bb, Bb and bb, giving 25%, 50% and 25% genotypes and a 75% against 25% phenotype split. Switching the inheritance setting to codominance leaves those counts untouched and only relabels the middle class, which is what you want for blood-type style questions where the heterozygote is its own phenotype rather than a copy of the dominant one.

Assumptions

The calculator runs on Mendel's two rules. Segregation keeps the alleles of a locus apart so every gamete receives one, and independent assortment (or calculate linked genes with our recombination frequency calculator) lets genes on different chromosomes pair without preference. It also assumes every gamete is functional, every fertilisation succeeds and every genotype survives to be scored, so all cells start out equally likely.

Dominance is treated as complete unless you select the codominance option, where heterozygotes keep their own label. Genotypes are parsed as letters in gene order, both parents must carry the same number of allele pairs, and each locus is assumed to hold two alleles only. Sex linkage, imprinting and maternal effects sit outside the model, since the grid always treats both parents as equivalent contributors.

Biological limits

Real litters drift away from the printed ratios whenever an assumption fails. Genes on the same chromosome travel together, so linked loci push counts toward the parental combinations. Lethal genotypes delete cells after fertilisation, incomplete penetrance hides phenotypes the genotype predicts, and unequal embryo or gamete viability means the survivors no longer stand for the original grid.

Sample size draws a hard boundary too. A one-in-four expectation swings widely across litters of four or five, so a single family rarely lands on the ratio. Past two genes the grid reaches 64 cells and then 256, and beyond that the ratio tables carry the same information far more clearly than a wall of cells. Use the square for one or two loci, then fall back on observed counts when biology starts pushing back.

What your punnett square result means

Each cell is a probability, not a promise. A 25% recessive outcome means that across many offspring a quarter will be recessive. In a single litter of four you could still see zero, one, or four. Ratios converge on the expected values as sample size grows.

For humans the same grid applies to single-gene conditions: two carrier parents (Aa × Aa) have a 25% chance per pregnancy of an affected child, a 50% chance of a carrier, and 25% chance of a non-carrier: the numbers behind every genetic-counselling session for autosomal recessive traits.

Common mistakes with the punnett square

• Describing the trait ("tall") instead of entering its alleles: Write gene letters such as Tt; the grid parses allele pairs only, so phenotype wording gives it nothing to cross.
• Expecting a 25% outcome in every four-offspring litter: Treat ratios as long-run frequencies. Small litters scatter widely, so compare counts across dozens of offspring first.
• Feeding it genes that sit on the same chromosome: Linked loci break independent assortment, so read the output as a starting point and judge it against observed counts.
• Giving the two parents a different number of genes: Match the gene count on both sides, for example BbRr against bbRr, so every locus lines up with its partner.

Limits of the punnett square calculator

This calculator assumes independent assortment, equal gamete viability, and no lethal genotypes. Real data drifts from 3:1 when alleles are linked on the same chromosome, when some genotypes die before birth, or when penetrance is incomplete. For three or more genes, extend the same logic manually. A 64-cell grid is rarely worth automating.

Research note from Dr. Dinesh A. Nagegowda, PhD: Punnett square grids provide visual clarity for tracking allele combinations, but their output represents probabilistic expectations across infinite progeny. In actual experimental crosses with small sample sizes, random segregation variance produces natural deviations from the expected 3:1 or 1:2:1 ratios. We instruct students and researchers to perform chi-square testing on raw counts rather than percentages to verify genetic models.

Sources

1. Campbell Biology (Pearson), Mendel's principles of inheritance (used for: law of segregation and independent assortment behind every grid)
2. OpenStax Biology 2e, Laws of Inheritance (used for: monohybrid 3:1 and dihybrid 9:3:3:1 ratio derivations)
3. Edwards (2012), Punnett's square, Studies in History and Philosophy of Science (used for: history of the grid method this calculator automates)
4. Punnett, Mendelism, full text (Project Gutenberg) (used for: original presentation of genotype ratios from paired crosses)

Punnett Square: frequently asked questions

More punnett square tools & guides

• What Is a Punnett Square, and How Do You Read One? (the guide that explains this formula in depth)
• Dihybrid Cross Calculator (Two genes at once: that page expands the same logic to sixteen cells with an expected-count column.)
• Blood Type (Child) Calculator (ABO typing is a codominant cross, so that calculator adds allele notation and the Rh system this grid does not parse.)