Combining at random
The Punnett square says three tall to one dwarf. Sow four seeds and you may well get four tall plants. The ratio is not broken — it was never a promise about four seeds.
What happens at fertilisation
A flowering pea plant makes an enormous number of pollen grains and a much smaller number of ova, and the ones that meet do so by chance.
- Half the pollen of a Tt plant carries T and half carries t.
- The same is true of its ova.
- Which pollen grain reaches which ovum is not controlled by anything. Any of the four combinations is equally likely.
Equally likely is the whole assumption
The 3 : 1 ratio follows from one claim only — that a T gamete is exactly as likely to be involved as a t gamete. That is what "at random" means here, and it is why the prediction is a probability rather than a certainty.
Like tossing a coin
Toss a coin four times and you will not reliably get two heads. Toss it four hundred times and you will get close to two hundred. The more offspring counted, the closer the result sits to the predicted ratio — which is why Mendel counted thousands of plants and not four.
This is why Mendel needed the pea
A plant that gives a handful of seeds would have produced numbers too small to see a ratio in. Large numbers of offspring are not a convenience here; they are the only way a chance process shows its underlying rule.