Balanced sex ratios

Why do species keep producing equal numbers of males and females when few males suffice?

A herd of a hundred deer needs nothing like fifty stags to reproduce. Five would do. The other forty-five eat the same grass, take the same predation risk, and contribute nothing a smaller number could not supply. A species that produced ninety females and ten males would seem to grow faster than its even-handed neighbour and replace it.

Yet across the great majority of sexually reproducing species, the ratio at birth sits close to one to one, and it comes back to that value when it is pushed away. Something is holding it there against an argument that looks, on its face, sound. What?

The first move is to notice that the argument above is about the population, and selection does not act on populations. It acts on the parents making the decision. So ask a narrower question: for one parent, in one generation, is it better to make a son or a daughter?

Now count grandchildren, not children. Every individual in the generation after next has exactly one father and one mother. So the whole of that generation's autosomal ancestry passes through the males as a class, and equally through the females as a class — two channels of identical size, no matter how many bodies carry each one.

That single fact does all the work. If the two classes are equal in total but unequal in number, then the smaller class has more per head. In a population that is ten per cent male, a male's expected share of the next generation is nine times a female's. A parent producing sons is buying into a scarce channel at the same price.

So consider a mutant that biases its offspring toward sons in that female-heavy population. It gets a disproportionate return — its grandchildren are over-represented — and the son-biasing tendency spreads. As it spreads, males become less scarce, and the premium shrinks. When males and females are equally numerous, a son and a daughter are worth exactly the same, and the advantage vanishes. Overshoot into a male-heavy population and the identical argument now favours daughters, pushing back.

Notice the shape of what we have found: not an optimum for the species but a stable point of no advantage. Fisher's argument, laid out in 1930 and anticipated by Darwin's puzzling over the same problem, does not claim a 1:1 ratio is good for anyone. It claims that any departure from it creates its own correction. The rarer sex always pays better, which is exactly what makes rarity temporary.

Which lets us answer the opening objection directly. Yes, ninety females and ten males would rear more fawns per year. But no gene can stay on that arrangement, because in that population a gene that makes sons out-reproduces every gene that makes daughters. Group-level efficiency loses to individual-level advantage. The forty-five spare stags are the price of a stable equilibrium, not evidence of one.

Now the refinement that the popular version usually drops. Fisher's claim was never about equal numbers; it was about equal expenditure. What a parent trades away is resources, and the currency of the argument is investment per offspring. If sons cost twice as much to raise to independence as daughters, then equal spending buys two daughters per son — and that, not 1:1, is the equilibrium. Where the sexes cost the same, the two versions coincide, which is why the numerical form is the one everyone remembers.

That also predicts where the rule should break, and it does, in ways that confirm rather than embarrass it. When brothers compete with each other for mates — as in fig wasps, where a few offspring mature inside a single fruit and mate among themselves — producing extra sons only means sons competing with sons, and the ratios observed are strongly female-biased. Ratios can also be dragged away from equilibrium by genetic elements that distort transmission for their own benefit, against which the rest of the genome then evolves suppressors.

Three refinements hold the reasoning together.

First, "the two sexes contribute equally" is a statement about autosomal genes, and it is exact, not approximate — it follows from every offspring having one parent of each sex. It does not hold for genes with sex-linked inheritance, and mitochondria pass through mothers only, which is why the interests of different parts of a genome over sex ratio can genuinely conflict.

Second, the equilibrium is over the ratio at the end of parental investment, not at conception. If one sex dies more often during the period the parent is paying for it, the parent is spending more per survivor of that sex, and the argument pushes the conception ratio the other way. The slight male excess at birth in humans — close to 105 boys per 100 girls, though it varies by population — is often read this way, but the causal story behind it is not settled.

Third, this is a claim about equilibrium, not optimality. It says only that departures are self-correcting. A population can be far from balance at any moment, and often is, for local ecological reasons — the argument says only that if those reasons are removed, the ratio drifts back.

• How local mate competition inverts the argument, and why fig wasps are its best test case.
• Why species with environmental sex determination — turtles, some fish — sit outside this argument, and what constrains them instead.

Every term the collection defines is gathered in the glossary.