Punnett Square Calculator
Punnett squares plus the family odds — two carriers with four children have a 68% chance of an affected child, not a certainty, and 32% of none.
One gene, two alleles, capital for dominant. Every birth is an independent draw, which is what the family table below the square is showing. Real traits are frequently more tangled than this — several genes, incomplete penetrance, sex linkage — so a single-gene square is a model of the arithmetic rather than a clinical answer. Nothing is uploaded.
Every cross, and the one that hides a trait completely
Two carriers give the familiar quarter affected and half carriers. The row worth a second look is AA × aa: an affected parent and a clear one produce no affected children at all — but every single child is a carrier. The trait skips a generation entirely and then reappears, which is exactly how recessive conditions surprise families that thought they had no history of one.
| Cross | Affected | Carriers | Genotypes | |
|---|---|---|---|---|
| AA × AA | 0% | 0% | 1 | both unaffected non-carriers |
| AA × Aa | 0% | 50% | 2 | one carrier |
| AA × aa | 0% | 100% | 1 | non-carrier and affected |
| Aa × Aa | 25% | 50% | 3 | both carriers |
| Aa × aa | 50% | 50% | 2 | carrier and affected |
| aa × aa | 100% | 0% | 1 | both affected |
Only the carrier-by-carrier cross produces all three genotypes; every other pairing collapses to one or two. And a carrier with an affected partner faces double the risk of two carriers — a half rather than a quarter.
"One in four" is not "one of our four"
A quarter per child is a per-birth chance, and births are independent. Two carriers with four children have a 68.4% chance of at least one affected child — not a certainty. Exactly one, the "expected" outcome, happens only 42.2% of the time, and 31.6% of four-child families have none at all. The other direction fails too: after one affected child the remaining three are not spent, and carry a 57.8% chance of another.
| Children | None affected | Exactly one | At least one |
|---|---|---|---|
| 1 | 75.00% | 25.00% | 25.00% |
| 2 | 56.25% | 37.50% | 43.75% |
| 3 | 42.19% | 42.19% | 57.81% |
| 4 | 31.64% | 42.19% | 68.36% |
| 5 | 23.73% | 39.55% | 76.27% |
| 6 | 17.80% | 35.60% | 82.20% |
Risk accumulates with every child and never reaches certainty; it passes even odds at the third. Track two genes at once and the same independence gives the 9:3:3:1 ratio — 56.25% showing both dominant traits and 6.25% showing both recessive ones, out of sixteen parts rather than four.
How to use
- Pick the first parent genotype.
- Pick the second parent genotype.
- Enter how many children to work out odds for.
- Read the square, then the family probabilities beneath it.
Frequently asked questions
Two carriers have four children — will one be affected?
Probably, but far from certainly. The chance of at least one affected child is 68.4%, exactly one is 42.2%, and 31.6% of four-child families have none at all. One in four is a per-birth chance, not a quota that four children fill.
Does an affected child use up the risk?
No. Every birth is an independent draw, so three further children still carry a 57.8% chance of another affected one. The reassurance that the quota is spent is the single most common misreading of a Punnett square.
What do two carriers actually produce?
A quarter affected, half unaffected carriers and a quarter carrying nothing. It is the only cross of the six that produces all three genotypes; every other pairing collapses to one or two.
Can a trait skip a generation entirely?
Yes, and the cleanest example is AA by aa. An affected parent and a clear one have no affected children whatsoever, but every single child is a carrier — so the trait vanishes for a generation and can reappear in the next.
What are the odds with one affected parent?
A carrier with an affected partner faces a half rather than a quarter, which is double the risk of two carriers. Two clear non-carriers face none at all, and two affected parents pass it on every time.
What is the 9:3:3:1 ratio?
Two genes tracked at once, each a carrier cross. Because the genes are independent the chances multiply: 56.25% show both dominant traits and 6.25% show both recessive ones, out of sixteen parts rather than four.
Does a real trait work like this?
Often not. This is one gene with two alleles and full dominance. Many traits involve several genes, incomplete penetrance or sex linkage, so treat a square as a model of the arithmetic rather than a clinical answer.
Does this send anything anywhere?
No. Every figure is computed in your browser, and nothing is uploaded or stored.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.