AA Aa Aa aa
Biology

Punnett square

Single-trait Punnett square for two diploid parents.

01Inputs
Two letters; uppercase = dominant, lowercase = recessive (e.g. Bb).
Use the same letter as parent 1 for consistency.
02Results
Genotype distribution
Dominant phenotype
Recessive phenotype
03How it works

Why this calculation

The Punnett square is the visual workhorse of high-school and introductory college genetics. Drawn by Reginald Punnett in 1905, it shows in a 2 × 2 grid (or 4 × 4 for two traits) every possible combination of alleles a child can inherit from two diploid parents. The simplicity is the point: by the time a student leaves the classroom, they should be able to reason about why two brown-eyed parents can have a blue-eyed child, why cystic fibrosis can skip a generation in carriers, why a cross between a true-breeding red flower and a true-breeding white can yield 100 % pink offspring under incomplete dominance. The same square is used in animal breeding, in family-planning genetic counselling, and in any argument about heritable traits where a probabilistic answer is required from a finite set of input alleles. This calculator handles the canonical single-trait, two-allele case and shows both the genotype distribution (the four cell contents) and the phenotype split (dominant versus recessive) computed from them.

The formula

For a single trait with two alleles per parent, the four cells of the Punnett square are obtained by combining each allele from parent 1 with each allele from parent 2. If parent 1 is Aa and parent 2 is Aa, the cells are AA, Aa, Aa, aa — yielding a genotype ratio of 1 : 2 : 1 (homozygous dominant : heterozygous : homozygous recessive). Under classic complete dominance, both AA and Aa express the dominant phenotype, so the phenotype ratio is 3 : 1. The convention is to write the dominant allele in uppercase and the recessive allele in lowercase; the genotype label sorts dominant first (Aa, never aA). For two traits with two alleles each, the square is 4 × 4 with 16 cells and produces the famous 9 : 3 : 3 : 1 phenotype ratio under independent assortment — but this calculator handles only the single-trait case, which covers the vast majority of classroom scenarios.

How to use it

The panel takes two inputs: parent 1 alleles and parent 2 alleles, each entered as a two-character string (e.g. Bb, AA, aa). Uppercase letters are dominant alleles, lowercase letters are recessive alleles, and the calculator does not enforce that you use the same letter for both parents — it simply combines what you give it. Defaults are Bb × Bb, the heterozygous cross that produces the canonical 1 : 2 : 1 ratio. The result panel shows the 2 × 2 grid with the four offspring genotypes, the genotype distribution as fractions of four (and percentages), and the dominant-versus-recessive phenotype split, also as fractions of four.

Worked example

A pea plant cross: TT × tt (true-breeding tall × true-breeding dwarf, the original Mendel experiment). All four cells contain Tt — every offspring is heterozygous and, under complete dominance, all four express the tall phenotype. Now cross two of those Tt offspring: Tt × Tt. The square gives TT, Tt, Tt, tt — a 1 : 2 : 1 genotype ratio and a 3 : 1 phenotype ratio (three tall to one dwarf). This is the famous 3 : 1 rediscovery that founded modern genetics. A second example: cystic fibrosis is recessive, so two non-affected carrier parents are both Cc. Their cross is Cc × Cc, identical structure to the previous: 1 CC (unaffected, non-carrier), 2 Cc (unaffected carriers), 1 cc (affected). The probability that any single child has cystic fibrosis is 25 %, the probability they are a carrier is 50 %, and the probability they are entirely free of the recessive allele is 25 %. A third example: an autosomal-dominant disorder where one parent is Hh and the other is hh gives 50 % affected offspring (Hh) and 50 % unaffected (hh) — the textbook autosomal-dominant inheritance pattern.

Common pitfalls

First, the square gives the probability of each outcome per child, not the guaranteed distribution in a small family. A 3 : 1 ratio across four children does not mean every set of four siblings will have exactly three tall and one dwarf — independent draws from a 0.75 / 0.25 distribution can produce any combination. Second, the model assumes independent segregation of alleles, which fails for linked genes on the same chromosome: alleles close together cross over less than 50 % of the time and the offspring distribution skews. Third, it assumes complete dominance. Many real traits show incomplete dominance (heterozygotes show an intermediate phenotype), co-dominance (both alleles express simultaneously, as in AB blood type), or multi-allelic systems (the ABO blood group has three alleles total, not two). Fourth, sex-linked genes on the X chromosome need a different treatment because sons inherit only one X allele while daughters inherit two — a single Punnett square does not capture this asymmetry. Fifth, polygenic traits (height, skin colour, intelligence) cannot be modelled with a Punnett square at all because they involve many genes with small effects.

Variations & context

The dihybrid cross (two traits, two alleles each, independent assortment) extends the same logic to a 4 × 4 square: parent gametes are AB, Ab, aB, ab (each with probability 0.25 from a AaBb × AaBb parent), and the resulting 16-cell grid produces the 9 : 3 : 3 : 1 phenotype ratio Mendel observed in pea seed colour and shape. Trihybrid and beyond use 8 × 8, 16 × 16 grids and are usually replaced by the branching diagram or the multiplication rule of independent probabilities. Hardy-Weinberg equilibrium generalises the Punnett logic to whole populations, predicting the steady-state allele frequencies under random mating, no selection, no mutation, and no migration. Pedigree analysis complements the Punnett square for tracking known traits across multiple generations. Modern genetics adds a layer the early 20th century could not see: epigenetic effects, in which the same genotype produces different phenotypes depending on environmental signals during development. None of those replace the Punnett square as a first-pass tool; they extend the conversation it starts.

Related calculators