Eye colour is the Punnett square everyone actually does in class, and the awkward part is that two different courses teach it two different ways. One treats it as a single gene where brown dominates blue, giving 3 brown : 1 blue. The other uses two separate genes, one for brown pigment and one for green pigment, giving 9 brown : 3 hazel : 3 green : 1 blue. Both appear in exam papers, so this does both and you pick the one your syllabus is asking for.
Eye colour is the Punnett square everyone actually does in class, and courses split on how many genes to use. Some teach it as one gene, brown over blue. Some teach two genes, where brown and green pigment are separate and you get 9 brown : 3 hazel : 3 green : 1 blue. This does both, so you can use whichever one your syllabus is asking for.
Type Bb for a heterozygous parent, BB for homozygous dominant, bb for homozygous recessive. Upper case is dominant, so a parent is Bb or BB, never BG.
One honest caveat before the maths
Real human eye colour is polygenic. It depends on several genes, mostly in two regions of chromosome 15 and one on chromosome 2, and the pigment itself sits in the iris in a structure that is not a simple dominant/recessive switch. If you inherit the genes for a green iris, that is not a case of green dominating brown. It is a different pigment, in a different place, sitting in front of a different one.
That is why the models below are simplifications. They are the simplifications your course uses, which is what gets the marks, and they are useful because they let you reason about ratios. Just do not walk out believing eye colour is a single gene with an uppercase and a lowercase version.
Model 1: the single gene
One gene, B for brown pigment, dominant over b for none. Any B gives brown eyes, because brown is dominant. Only bb gives blue.
The classic cross is Bb × Bb, which gives a 2×2 square:
- BB, Bb, Bb — three brown
- bb — one blue
So 3 : 1, with 75% brown and 25% blue. The 2×2 is the smallest useful Punnett square there is: one parent contributes two gametes, the other contributes two gametes, and the four boxes are all four possible combinations. A BB × bb cross is the other one worth knowing, and every child is Bb, so every child has brown eyes.
The useful thing to take from this: a child with blue eyes must have inherited a b from both parents. If a parent has brown eyes, that tells you nothing about whether they carry a b. Which is the whole reason eye colour is a bad predictor of anything else.
Model 2: two genes
Now split it. B/b makes brown pigment. G/g makes green pigment. Each dominant allele switches on its own pigment, and brown reads in front, so:
- Any B with any G — brown, 9 out of 16
- B with only gg — hazel, 3 out of 16
- bb with any G — green, 3 out of 16
- bbgg — blue, 1 out of 16
That is 9 : 3 : 3 : 1, and it comes from BbGg × BbGg, which is a 4×4 square with 16 boxes. The reason it is 16 is that each parent has two genes, so each parent can pass on four different gametes: BG, Bg, bG and bg. Four by four is sixteen.
Where 9 : 3 : 3 : 1 actually comes from
It is worth being able to derive it, because that is what the question is testing. Treat the two genes separately and multiply the answers.
For B alone, Bb × Bb gives 3 brown to 1 not. For G alone, Bb × Bb, that is Gg × Gg, gives 3 with pigment to 1 without. The four combinations, multiplied out:
- Brown pigment (3/4) with green pigment (3/4) = 9/16 brown
- Brown pigment (3/4) with no green (1/4) = 3/16 hazel
- No brown (1/4) with green pigment (3/4) = 3/16 green
- No brown (1/4) with no green (1/4) = 1/16 blue
Nine, three, three, one. Nothing clever, it is just two independent monohybrid ratios multiplied together. If you can do the single gene, you can do this.
Linkage breaks the 9 : 3 : 3 : 1
That multiplication only works if the two genes assort independently, which means they sit on different chromosomes and can be inherited separately. If they are linked, on the same chromosome, they travel together and the rarer combinations never get made.
Take BbGg × BbGg with the genes linked. The gametes are no longer BG, Bg, bG and bg. A parent with the genes on one chromosome together, in cis, can only pass on BG or bg. The square becomes 2×2, and the answer collapses to 3 brown : 1 blue. Hazel and green disappear entirely.
This is the single most useful thing on the page. The 9 : 3 : 3 : 1 answer is only correct because the genes assort independently, and a question that mentions linkage, or one chromosome, is telling you that assumption has been removed. Set the tool to linked and watch the middle two phenotypes vanish.
How to use it
Pick your model, type the two parent genotypes, and the square builds itself. For a single gene that is a 2×2. For two genes it is a 4×4, with all four gametes listed so you can check your own working against them.
Type genotypes in upper and lower case, because the case is carrying the dominance. A heterozygous parent is Bb, never BG. If your question gives you phenotypes instead of genotypes, be careful: a brown-eyed parent could be BB, Bb, BBGg, BbGg and several other combinations, and each gives a different square.
If your course uses different letters, the logic is identical. Swap B for whatever they call the brown gene and G for the green one and the ratios do not change.
Worked examples
- Bb × Bb, single gene — 3 brown : 1 blue, 75% and 25%.
- BB × bb, single gene — all four children Bb, so all brown.
- bb × bb, single gene — all blue. The only way to guarantee blue eyes in this model.
- BbGg × BbGg, two genes, unlinked — 9 brown : 3 hazel : 3 green : 1 blue.
- BbGg × BbGg, two genes, linked — 3 brown : 1 blue, no hazel or green at all.
- BbGg × bbgg, unlinked — all four phenotypes, one each.
Every figure here comes from the simplified Mendelian model, on the assumption of independent assortment unless told otherwise. For anything to do with real inherited eye colour in a real family, a genetics textbook or a clinical genetics source is the place to go, because the biology is more complicated than a 4×4 square.
Frequently asked questions
Is eye colour really a single gene?
No, and this is the most important thing on the page. Real human eye colour is polygenic, involving several genes mainly on chromosome 15 and one on chromosome 2, and the pigment sits in the iris in a way that is not a simple on/off switch. The single-gene and two-gene models here are the simplifications your course uses, not the biology.
What is the difference between the one-gene and two-gene models?
The one-gene model has a single B/b gene where brown is dominant, giving 3 brown to 1 blue. The two-gene model splits it into B/b for brown pigment and G/g for green pigment, giving 9 brown to 3 hazel to 3 green to 1 blue. Courses pick one or the other, so check which your syllabus is using.
Why does the two-gene model give 9:3:3:1?
Because it is two independent monohybrid ratios multiplied together. For B alone, Bb x Bb gives 3 with pigment to 1 without. For G alone, Gg x Gg gives the same 3 to 1. Multiplying the four combinations gives 9/16, 3/16, 3/16 and 1/16. Nothing clever, just two simple ratios composed.
What does a child with blue eyes tell you about their parents?
In the single-gene model, a blue-eyed child must be bb, which means both parents passed on a b. A brown-eyed parent tells you nothing, because BB and Bb both look brown and a blood test or a glance cannot tell them apart. That is why eye colour is a poor predictor of anything inherited.
Does linkage change the 9:3:3:1 ratio for eye colour?
It does, and dramatically. If the two genes are linked, on the same chromosome, a parent can only pass on BG or bg rather than all four gametes. The square shrinks to a 2×2 and the ratio collapses to 3 brown to 1 blue, with hazel and green disappearing entirely. Set the tool to linked to watch it happen.
Why is a genotype written in upper and lower case?
The case carries the dominance. An uppercase letter is the dominant allele and a lowercase one is recessive, so a heterozygous parent is written Bb rather than BG. Writing BG would imply the two genes are different, which is a different question entirely and this tool will reject it.
Can I use different letters for the same model?
Yes. Swap B for whatever your course calls the brown gene and G for the green one and the ratios do not change at all. What matters is which allele is dominant, not which letter it is.