The formulas in the heredity chapter are not hard to memorize — the hard part is believing them. Why would two tall pea plants produce a short offspring at all? And why exactly one in four? This experiment needs no pea plants and no waiting for a harvest: two coins, one sheet of paper, and about fifteen minutes are enough to produce the 3 : 1 ratio with your own hands.

What it simulates is the very core of heredity: when reproductive cells form, paired genes separate — and at fertilization they recombine at random. A coin's two faces are the perfect stand-in for that randomness.

Safety first (read this before you start)

  • Coins are small objects — keep them well away from children under three and from pets to prevent swallowing, and count them back in when the experiment ends.
  • Toss low, over a tabletop or a cloth-covered surface — no high throws, and never toss toward people or anything breakable, so nothing gets hit on the bounce.
  • Circulating coins are dirty: wash your hands before and after, and keep hands away from your mouth, nose and food during the experiment.
  • Tosses are noisy and coins roll — clear cups off the table first, and mind that coins don't roll into gaps in electrical equipment.
  • There is no chemical or biological hazard; the whole experiment can be done independently, though younger students should have a parent alongside to help count.

Materials and equipment

  • 2 coins of the same size (any common coin works; you must be able to tell the two apart);
  • 2 small sticky notes or bits of masking tape, to label the coins "father" and "mother";
  • 2 opaque paper cups (optional — shaking and pouring is closer to truly random than hand-tossing);
  • record paper and a pen, ruled into three columns: DD, Dd, dd;
  • a calculator or phone (for working out ratios);
  • best done in pairs: one tosses, one records — twice the speed.
Agree on the symbols first: we use pea plant height as the model. D stands for the tall-stem gene (dominant) and maps to heads (the face-value side); d stands for the short-stem gene (recessive) and maps to tails. The two coins represent one sperm from the father plant and one egg cell from the mother plant.

Step 1: simulate Dd × Dd with 20 tosses

  1. Label the coins "father" and "mother". This step declares that both parent plants have the gene pair Dd (both tall, but each carrying one short-stem gene).
  2. Why can a single coin stand for a whole Dd plant? Because when a Dd plant forms reproductive cells, D and d separate into different gametes, producing D-carrying and d-carrying gametes in roughly equal numbers — exactly a coin's two faces.
  3. Put both coins in a cup, shake, and tip them out onto the table together. Each pour simulates one fertilization.
  4. Record the combination as it lands: both heads is DD; one head and one tail is Dd (no matter which coin shows which); both tails is dd.
  5. Repeat 20 times, keeping a tally on the record sheet.
  6. Now compute: what is the ratio of tall (DD plus Dd) to short (dd)? Most groups get numbers like 13 : 7 or 17 : 3 — a long way from 3 : 1. Don't worry yet; just write it down.

Step 2: toss past 100, and the ratio settles in

  1. Keep tossing until the running total reaches 100. Every 20 tosses, work out the current ratio and note it in the margin.
  2. Students will watch the ratio slowly "pull in": from wild early swings, it gradually steadies near 3 : 1.
  3. If you are doing this as a class, pool every group's data (often over 1,000 tosses in total) — the ratio gets closer still to the theoretical value.
  4. Finally look at the three combinations separately: DD, Dd and dd should approach 1 : 2 : 1. Dd comes out on top because it has two routes to appear: "father heads, mother tails" and "father tails, mother heads".
This step is the real point of the experiment. Mendel counted thousands of pea plants — not because he was tireless, but because the ratios only emerge at large numbers. The scatter you got from 20 classroom tosses is precisely why he had to grow that many peas.

Step 3: toss one coin only — simulating a test cross

  1. Now swap the mother plant for a short-stemmed dd. A dd plant can only produce d-carrying gametes, so the "mother" coin no longer needs tossing — fix it on tails.
  2. Toss only the "father" coin, 40 times, recording the pairings: heads with d gives Dd (tall); tails with d gives dd (short).
  3. The result will come out near 1 : 1 — half tall, half short.
  4. In plant breeding this maneuver has a formal name: the test cross. To learn whether a tall pea plant is DD or Dd, cross it with a short one. The moment any short offspring appears, the parent must be Dd.
Interactive simulation: this experiment has a drag-and-play simulator in our Interactive Lab. The lab interface is currently Chinese-only — the controls are simple sliders and switches, so it is still easy to explore.

What you'll see

Here is a typical set of real results (your numbers will differ, but the trend should match):

  • First 20 tosses: DD 7, Dd 9, dd 4 — tall : short = 16 : 4 = 4 : 1, higher than theory.
  • At 60 tosses: DD 14, Dd 32, dd 14 — tall : short = 46 : 14 ≈ 3.3 : 1, closing in.
  • At 100 tosses: DD 24, Dd 51, dd 25 — tall : short = 75 : 25 = 3 : 1; genotypes 24 : 51 : 25 ≈ 1 : 2 : 1.
  • Test cross (40 tosses): Dd 21, dd 19 — close to 1 : 1.
  • Throughout, Dd stays the most common combination, at roughly half.
  • Four or five dd results in a row — or ten tosses with no dd at all — are perfectly normal and do not mean the experiment went wrong.

Why it works

An organism's traits are controlled by genes. In body cells, the genes controlling a given trait come in pairs, one on each chromosome of a pair. Pea stem height is decided by one such pair: D for tall stems is the dominant gene; d for short stems is the recessive gene.

So there are three possible gene combinations, but only two visible traits:

  • DD — tall;
  • Dd — still tall, because when D is present, the trait d controls stays hidden;
  • dd — short. A recessive trait shows itself only when both genes in the pair are recessive.

The key thing the coins simulate is the formation of reproductive cells: the paired genes separate, each entering its own sperm or egg cell, so every gamete carries just one gene. A Dd plant therefore produces D-carrying and d-carrying gametes in roughly equal numbers — precisely the behavior of a fair coin. At fertilization, sperm and egg combine at random and genes pair up again, giving four equally likely combinations: DD, Dd, Dd, dd.

Sort those four by trait: the three containing D are all tall and only dd is short, so the trait ratio is 3 : 1; sort by gene combination and you get 1 DD : 2 Dd : 1 dd. Dd takes two shares because "father D, mother d" and "father d, mother D" are two distinct combinations — which is why Dd always leads your tally.

This is exactly what Mendel discovered. He crossed true-breeding tall peas with true-breeding short ones, and the first generation was all tall — shortness seemed to vanish. He then let that generation self-pollinate, and the short plants came back, with tall to short close to 3 : 1. Shortness had never disappeared: it rode along intact as the d inside Dd, hidden but faithfully passed down.

One more thing to remember: 3 : 1 is a probability, not a quota. Every fertilization is like a fresh coin toss, independent of all the others. So "we've had three tall ones, the fourth must be short" is wrong — the fourth still has a 3/4 chance of being tall. The "getting more accurate the more you toss" effect between Steps 1 and 2 is what statisticians call the law of large numbers, and it is why every probability-based experiment demands plenty of trials.

This machinery answers a question students often press on: if both parents are healthy, how can a child have a recessive genetic disorder? If both parents are Aa — unaffected carriers — each child has about a 1/4 chance of being aa and affected. Close relatives are far more likely than strangers to carry the same recessive disease gene, so their children's risk rises sharply. That is the biological reasoning behind laws and customs against close-relative marriage, and it is the very calculation premarital genetic counseling performs.

The four easiest mistakes

  • "One head and one tail only counts as one outcome" — the most common error in this experiment. "Father heads, mother tails" and "father tails, mother heads" are two different combinations, so Dd's chance is 2/4, not 1/3. Labeling the coins "father" and "mother" exists precisely to keep them from being lumped together.
  • "Dominant genes are stronger and better" — no. Dominance only says which trait shows when the pair is mixed; it says nothing about quality. In humans, extra fingers (polydactyly) is a dominant trait and albinism is recessive — clearly not a case of "dominant is better".
  • "The d in Dd gets eaten by the D" — it doesn't. The d survives intact and passes to the next generation as usual; it simply gets no chance to show in this one. That is why recessive traits can skip a generation.
  • "A tall plant must be DD" — not necessarily; it could be Dd. To tell them apart, run Step 3's test cross: if any offspring is short, the parent has to be Dd.

Notes for teachers and parents

  • Timing and groups: 30 minutes, in pairs; two coins per pair, with heads and tails agreed on in advance.
  • Say it before anyone touches anything: coins stay out of mouths; accompany younger children.

The step most often skipped. Have students toss 20 times first and publish their results — the groups will differ wildly, which sets up "the sample is too small" perfectly. Only when the whole class merges its data does 3 : 1 emerge. Save the Punnett square until after the tossing.

How to tell they really get it. They can distinguish traits from gene combinations, draw the Dd × Dd square, and explain that 3 : 1 is a probability, not a quota. That's the bar.

Take it further

  • Swap the two coins for two four-sided dice or drawn lots to simulate a gene with three or more forms (multiple alleles) — human ABO blood types, for instance.
  • Textbooks love earlobe, tongue-rolling and dimple examples, but most human traits are controlled by several gene pairs together and shaped by the environment too — so don't apply the simplified pea model straight to your own family and draw conclusions.
  • Try the math: if both parents are Aa, what is the probability that both of two children are aa? (Hint: multiply two independent events.)
  • Plot Step 2's data as a line graph — toss count on the x-axis, dd share on the y-axis — and watch the curve wobble its way toward 25%. The graph itself makes excellent material for an inquiry report.
  • Once you've finished, see how variation gets filtered by the environment in natural selection in a paper-chip game. Heredity supplies the material, selection sets the direction — the two articles together are the full story of evolution.