Consider a lake in New Zealand where snails live side by side. Some females reproduce sexually, mating with males and producing a mix of sons and daughters. Others are clones. Every egg they lay grows into a daughter genetically identical to her mother, and no male is required. Run the arithmetic forward a few generations and the clones should take over completely.

Yet across complex life, the opposite pattern prevails. The overwhelming majority of animal and plant species reproduce sexually, at least some of the time. Fully asexual lineages do exist, but most are twigs rather than branches on the tree of life. They tend to be recently derived and, by evolutionary standards, short-lived.

The biologist Graham Bell called this the queen of problems in evolutionary biology. Solving it means finding an advantage large enough to repay a cost that looks ruinous on paper. Three lines of explanation stand out: the arithmetic of the cost itself, the arms race with parasites, and the way recombination handles mutations. Each reveals something about how selection actually operates on populations.

The sharpest statement of the problem comes from John Maynard Smith in 1978. Imagine two females, one sexual and one asexual, who produce the same number of offspring with the same survival. The sexual female devotes roughly half her offspring to sons, and sons cannot lay eggs. The asexual female produces only daughters, and every one of them will reproduce.

Now follow the lineages. If each female produces four offspring, the sexual line yields two daughters while the asexual line yields four. In the next generation, the asexual line is again twice as productive per individual. A single clonal mutant arising in a sexual population should therefore roughly double its share each generation while rare. All else being equal, it would sweep to fixation within a few dozen generations.

The same cost appears from the gene's perspective. A sexually produced offspring carries only half of its mother's genome, because meiosis dilutes her contribution with a male's. A clone transmits the entire genome. In population genetic terms, any allele that causes asexual reproduction enjoys a built-in transmission advantage before any effect on survival is counted. Further costs pile on top: time and energy spent finding mates, exposure to predators and sexually transmitted infections, and recombination itself, which breaks apart gene combinations that selection has already proven successful.

The cost is not universal. It approaches two-fold when males contribute only genes, but it shrinks where fathers provide substantial care. It largely disappears in organisms such as many algae and fungi, whose mating types produce gametes of equal size, so there are no males to pay for. For most animals, though, the cost is real and large. Whatever sex provides, the benefit must be strong enough to offset a lineage that grows twice as fast every generation.

Takeaway

A trait that persists despite a large, obvious cost points to a hidden benefit at least as large. Measuring the cost tells you how big the advantage must be.

In Through the Looking-Glass, the Red Queen tells Alice that it takes all the running you can do to keep in the same place. In 1973, Leigh Van Valen borrowed the image to describe species locked in constant evolutionary pursuit. W. D. Hamilton and others later applied it to sex, proposing that the decisive adversaries are parasites.

The logic rests on negative frequency-dependent selection. Parasites have short generation times and enormous populations, so they adapt quickly to whatever host genotypes are most common. A successful clone, by definition, becomes common, and that makes it the most attractive target. Its advantage erodes precisely because it succeeded. Sexual reproduction shuffles alleles every generation, so it keeps producing rare genotypes that parasites have not yet learned to exploit.

Field evidence comes from Curtis Lively's long-running work on the mud snail Potamopyrgus antipodarum, the same New Zealand snail where sexual and clonal females coexist. Sexual individuals are most common in shallow lake margins, where trematode parasites are abundant. Clones dominate where infection risk is low. Repeated sampling over years showed that common clones became disproportionately infected and then declined, while previously rare clones rose in their place. This is the cycling the hypothesis predicts.

Experiments added causal weight. In 2011, Levi Morran and colleagues engineered populations of the nematode Caenorhabditis elegans to be either obligately self-fertilising or obligately outcrossing. They then exposed the worms to the bacterial pathogen Serratia marcescens. When the bacteria were allowed to coevolve with their hosts, the selfing populations were driven extinct within about twenty generations, while the outcrossing populations persisted. Against a fixed, non-evolving pathogen, the difference largely vanished. The advantage of sex depended on an opponent that could adapt.

Takeaway

When your enemies evolve, being rare can matter more than being well adapted. Sex can be understood as a strategy for staying a moving target.

A second family of explanations focuses on mutations themselves. Every genome accumulates copying errors each generation. Most are harmful, and a smaller number are beneficial. Asexual lineages handle both kinds poorly, because each genome is inherited as a single, unbreakable unit.

Consider harmful mutations first. In a finite clonal population, some individuals carry the fewest mutations of anyone. If that best class is small, genetic drift can eliminate it by chance. Without recombination, it can never be rebuilt except through rare back-mutation. Hermann Muller described this process in 1964, and it became known as Muller's ratchet: each click permanently raises the minimum mutation load, and repeated clicks can drive small asexual populations toward mutational meltdown. Alexey Kondrashov later proposed a deterministic version. If each additional harmful mutation hurts more than the last, recombination concentrates them into some offspring, and selection removes those offspring efficiently, purging many mutations at once.

Beneficial mutations face a different problem. Suppose two useful mutations arise in separate asexual individuals. They cannot be combined, so their lineages must compete, and one is lost. Ronald Fisher and Muller both recognised that sex lets favourable mutations from different lineages come together in a single genome. Population geneticists call the broader problem Hill-Robertson interference: when sites are linked, selection at one site reduces the efficiency of selection at the others.

Experimental evolution supports these ideas. In baker's yeast, sex can be switched on or off genetically, and sexual populations have repeatedly adapted faster in stressful environments. A 2016 study by Michael McDonald and colleagues tracked genomes directly. It found that sex sped adaptation partly by freeing beneficial mutations from harmful ones hitchhiking alongside them. Bdelloid rotifers, which appear to have persisted for tens of millions of years without conventional sex, remain an instructive exception. They seem to compensate partly through unusual DNA repair and genes acquired from other organisms.

Takeaway

Without recombination, a genome's fate is all-or-nothing. Sex lets selection judge genes on their own merits rather than by the company they keep.

The paradox of sex is not simply that sex has costs. It is that the costs are large and immediate, while the proposed benefits accumulate over many generations. A clone wins today, so sex has to win over the long run.

These explanations are not rivals. Coevolving parasites keep shifting the selective landscape, which raises the value of new beneficial mutations. Finite populations make drift and linked selection unavoidable. Theoretical work, notably by Sarah Otto, suggests these mechanisms reinforce one another. That may be why no single hypothesis feels complete on its own.

What emerges is a picture in which variation itself is under selection. Sex persists not because it benefits any individual, but because lineages that shuffle their genes are the ones still around.