Frequency-dependent selection

In our understanding of evolution, it is tempting to view natural selection as a simple contest where the "best" trait invariably wins. However, the fitness of a trait is often not absolute but deeply contextual, its value shifting based on its surrounding environment. A particularly powerful contextual factor is a trait's own prevalence. This phenomenon, where the success of a gene or strategy depends on its frequency in the population, is the core of frequency-dependent selection. It addresses the crucial question of how biological diversity is maintained and why the "fittest" is not always a fixed target.

This article delves into this dynamic evolutionary principle. We will first explore the foundational principles and mechanisms, examining how both rarity and commonness can be advantageous under different circumstances. Following this, we will journey through its widespread applications, discovering how this single concept explains a vast array of patterns across the natural world, from the evolution of sex to the endless arms race against disease.

In our everyday thinking about evolution, we often fall into a simple trap. We imagine a trait, like sharp talons or keen eyesight, as being intrinsically "good" or "bad." We picture natural selection as a straightforward process where the "better" version always wins, marching relentlessly towards some optimal design. But Nature, in her infinite subtlety, often plays a more interesting game. The fitness of a trait, its contribution to survival and reproduction, is frequently not a fixed, absolute quantity. Instead, it can be a moving target, its value dependent on the social and ecological context. And one of the most powerful contextual factors is simply how common the trait is. This is the heart of ​​frequency-dependent selection​​.

Imagine a simple game of rock-paper-scissors. Is "rock" a good move? The question is meaningless without knowing what your opponent is playing. If they are about to play "scissors," rock is a winning move. If they are about to play "paper," rock is a losing move. If the entire population of players becomes obsessed with playing "rock," the cleverest strategy is to start playing "paper." The best strategy's fitness depends entirely on the frequency of other strategies in the population.

Frequency-dependent selection comes in two main flavors, with opposite effects on the diversity of life.

Negative Frequency-Dependent Selection: The Advantage of Being Different

This is the evolutionary principle of the hipster: a trait is advantageous precisely because it is rare. As it becomes more popular, its advantage fades. The classic example comes from the world of predator and prey. Imagine a species of snail that comes in two shell patterns: banded and unbanded. A local bird species loves to eat these snails. The birds, being efficient hunters, tend to form a "search image" for the most common pattern. If banded snails are everywhere, the birds become experts at spotting them, swooping down on banded snail after banded snail, while the rare, unbanded ones often go unnoticed.

In this situation, the unbanded snails have higher fitness (they survive more) simply because they are rare. But what happens next? Because they survive better, they leave more offspring, and the unbanded pattern starts to become more common. As its frequency rises, the birds' search image may begin to shift. Suddenly, being unbanded is no longer an advantage. The birds are now keying in on this "new" common pattern, and the now-rare banded snails enjoy a period of relative safety.

This dynamic, where the fitness of a type decreases as its own frequency increases, is called ​​negative frequency-dependent selection (NFDS)​​. It acts as a powerful balancing force. It never allows one type to completely take over, because success breeds failure. Whenever a type becomes too common, selection turns against it, giving the rare types a chance to rebound. This is one of nature's most elegant mechanisms for maintaining genetic diversity within a population.

Positive Frequency-Dependent Selection: The Power of the Crowd

The opposite scenario is ​​positive frequency-dependent selection (PFDS)​​, where the motto is "strength in numbers." The more common a trait is, the fitter it becomes. Think of warning coloration in toxic animals. Many species of poisonous butterflies, like Monarchs and Viceroys, have evolved remarkably similar orange-and-black wing patterns. This is an example of Müllerian mimicry. A young, naive bird might try to eat one of these butterflies, get a mouthful of foul-tasting toxins, and quickly learn to avoid that pattern in the future.

The more butterflies that share this pattern, the faster the local predators learn the lesson. An individual butterfly with a rare, unusual warning pattern, even if it is just as toxic, is at a disadvantage. A predator might not have learned to avoid its specific pattern and may attack it. In this case, conformity is rewarded. Positive frequency-dependent selection pushes populations towards a uniform state, eliminating rare variations and reinforcing the common type. It is a force for homogeneity, not diversity.

We can move beyond these nice stories and see how this works with a bit more precision. Let's return to our game, but this time with two strategies in a population, $M$ and $N$. Individuals pair up randomly and "play" against each other, with the payoff being the number of offspring they produce. We can summarize the outcomes in a ​​payoff matrix​​:

Let's say the frequency of strategy $M$ in the population is $p$. The frequency of strategy $N$ must then be $1-p$. What is the average, or expected, fitness of an $M$ individual? It will meet another $M$ with probability $p$ (getting a payoff of 2) and an $N$ with probability $1-p$ (getting a payoff of 5). So, its fitness is:

$W_M(p) = p \cdot 2 + (1-p) \cdot 5 = 5 - 3p$

Notice something crucial: the fitness of $M$, $W_M(p)$, is not a constant! It's a function of its own frequency. As $M$ becomes more common (as $p$ increases), its fitness goes down.

Now let's do the same for strategy $N$. Its fitness is:

$W_N(p) = p \cdot 4 + (1-p) \cdot 1 = 1 + 3p$

The fitness of $N$ increases as the frequency of $M$ increases. This is a classic signature of negative frequency-dependent selection. When $M$ is very rare ($p$ is close to 0), its fitness is $W_M(0) = 5$, while $N$'s fitness is $W_N(0)=1$. $M$ has a huge advantage and will increase. But when $M$ is very common ($p$ is close to 1), its fitness is $W_M(1)=2$, while $N$'s fitness is $W_N(1)=4$. Now $N$ has the advantage, and the frequency of $M$ will decrease.

This process can't go on forever. There must be an intermediate point where the tables are perfectly balanced—an equilibrium. This happens when the fitnesses are equal: $W_M(p^*) = W_N(p^*)$.

$5 - 3p^* = 1 + 3p^*$ $4 = 6p^*$ $p^* = \frac{4}{6} = \frac{2}{3}$

At a frequency of $p = 2/3$, both strategies have exactly the same fitness. If the frequency of $M$ drifts above $2/3$, $N$ does better and brings it back down. If it drifts below $2/3$, $M$ does better and brings it back up. This is a ​​stable equilibrium​​. The population will naturally settle at a state where both strategies are maintained, with $M$ at a frequency of $2/3$ and $N$ at $1/3$.

Had the selection been positive frequency-dependent, the equilibrium would be ​​unstable​​. It would be like balancing a pencil on its tip. The slightest nudge would send the population hurtling towards one of the all-$M$ or all-$N$ states, wiping out diversity.

It's important to realize that negative frequency-dependence is not the only way Nature maintains diversity. Another well-known mechanism is ​​heterozygote advantage​​ (or overdominance), where individuals with two different versions of a gene (heterozygotes) have higher fitness than individuals with two identical copies (homozygotes). The textbook case is sickle-cell anemia in regions with malaria.

The crucial difference lies in the nature of fitness itself. In classic heterozygote advantage, the fitness of each genotype ($AA$, $Aa$, $aa$) is a fixed constant. An $Aa$ individual is always fitter, regardless of whether the $A$ or $a$ allele is common or rare. The balancing effect emerges implicitly because as an allele, say $A$, becomes common, it is more likely to be found in the less-fit $AA$ homozygote form. In negative frequency-dependent selection, the fitness values themselves are dynamic variables that change with the population's composition. The feedback is direct and explicit.

Once you learn to recognize it, you start seeing negative frequency-dependent selection everywhere. It is a fundamental organizing principle in ecology and evolution.

The Red Queen's Race

In the arms race between hosts and their parasites, NFDS is the engine of a perpetual cycle known as the "Red Queen" dynamic, from Lewis Carroll's character who had to run as fast as she could just to stay in the same place.

Consider a host with a resistance gene, $R$, that protects it from a parasite. As long as $R$ is rare, it's highly advantageous. But as it spreads through the host population, it creates an enormous selective pressure on the parasite to evolve a counter-measure—a virulence gene, $V$, that can overcome $R$. Once parasites with the $V$ gene appear and spread, the $R$ gene is no longer so useful; in fact, if carrying the $R$ gene has even a small cost, it now becomes a liability. Selection will turn against the common $R$ gene, favoring a different, rarer resistance gene (or even the original susceptible allele). This, in turn, changes the selective landscape for the parasite again. Each side is constantly adapting to the most common strategy of the other, a chase that never ends and, in the process, maintains a rich diversity of both resistance and virulence genes in the respective populations.

Disruptive Looks, Stabilizing Forces

Sometimes, the effect of NFDS can be paradoxical. Imagine a bird population where there are two types of seeds available: small and large. Birds with small beaks are good at eating small seeds, and birds with large beaks are good at eating large seeds. Birds with intermediate beaks are clumsy with both. This looks like a recipe for ​​disruptive selection​​, where the extremes are favored and the middle is selected against.

Now, let's add NFDS to the mix. If small-beaked birds become too common, they will deplete the supply of small seeds, making life harder for themselves. The now-abundant large seeds provide a feast for the rare, large-beaked birds, giving them a fitness advantage. This is NFDS at work, maintaining a stable mix of both small- and large-beaked birds. The underlying genetic force is stabilizing—it pulls the allele frequencies towards a 50/50 equilibrium. But the phenotypic pattern it produces is disruptive: a bimodal distribution of beak sizes with very few intermediates. NFDS provides a mechanism for how a population can split into different specialist morphs and stably maintain them.

An Entangled Dance of Ecology and Evolution

The line between population size (ecology) and gene frequencies (evolution) can also blur. The fitness of a genotype can depend on both the total population density and the frequency of other genotypes. Imagine two genotypes, $A$ and $B$, competing for resources. Perhaps $A$ is more sensitive to crowding than $B$ is. This means that as the total population size, $N$, increases, the fitness of $A$ drops faster than the fitness of $B$. This is ​​density-dependent selection​​. Now, what if the equilibrium population size, $N$, itself depends on the frequency of $A$ and $B$? A population full of $A$ types might support a different density than a population full of $B$ types.

In this intricate dance, the underlying density-dependent effects become translated into effective frequency-dependent selection. The evolutionary trajectory of an allele depends on its frequency, not just because of direct interactions, but because its frequency alters the ecological stage—the population density—on which the play unfolds.

If a new, advantageous allele arises but its spread is halted by negative frequency-dependent selection, it should leave a unique footprint in the organism's DNA. When a beneficial mutation sweeps to fixation, it drags along its neighboring DNA, wiping out genetic variation in a process called ​​genetic hitchhiking​​. This results in a "valley" of low diversity around the selected gene.

An "arrested sweep" due to NFDS is different. The new allele, $S'$, rises in frequency rapidly at first, so its genetic background (the haplotype it's on) is "young" and shows very little diversity. But it doesn't fix. The old allele, $s$, persists in the population at a stable frequency. The haplotypes carrying $s$ are a grab-bag of the ancestral variation, having had much more time to accumulate mutations and be shuffled by recombination. The result is a striking asymmetry: a large block of low-diversity haplotypes carrying $S'$ coexisting with a smaller, more diverse collection of haplotypes carrying $s$. Finding these signatures in genomic data is like a form of evolutionary archaeology, allowing us to uncover these dynamic stories of balancing selection long after they played out.

From simple games to the intricate dance of host and parasite, from the shape of a bird's beak to the patterns in our DNA, frequency-dependent selection reveals a profound truth: in the grand theater of evolution, fitness is rarely a monologue. It is a dialogue, a conversation where the success of any one actor depends on the entire cast.

Having grasped the fundamental principle of frequency-dependent selection—that an individual's success is not an absolute measure but a function of its rarity or commonness—we can now embark on a journey across the vast landscape of the living world. We will see how this single, elegant idea echoes from the most intimate choices of a mate to the global arms race against disease, and even into the structure of our own societies. It is a unifying theme, a simple rule that generates profound and often surprising complexity. It is the director of a grand, unceasing dance between rarity and abundance.

Let's begin with one of the most powerful forces in evolution: the choice of a mate. In a world saturated with signals, how does one stand out? Sometimes, the answer is simply to be different. Imagine a forest filled with the songs of male katydids, all vying for the attention of females. If all the males sing the same tune, the song becomes mere background noise. But a male who produces a song with an unusual pitch—be it higher, lower, or just different—cuts through the chorus. In some species, this is precisely what happens. Females exhibit a preference for the rarest song in the population, granting those males a significant reproductive advantage. This is negative frequency-dependent selection in its purest form. The popular song of one generation becomes common, and therefore less attractive, in the next. This dynamic doesn't drive the population in one direction; instead, it actively preserves a rich variety of songs, a constant bubbling of novelty.

This simple preference for the rare can have consequences far beyond maintaining a colorful palette of traits. It can be a powerful engine for the creation of new species. In the isolated crater lakes of Africa, certain cichlid fish populations display a dazzling variety of color patterns. In some cases, a fish with a rare color morph enjoys greater mating success precisely because it is rare. This advantage can be strong enough to offset other disadvantages, such as being slightly less camouflaged. A simple mathematical model can capture this trade-off, balancing a morph's baseline viability against the penalty of being common. When the rare-morph advantage is strong enough, it can maintain multiple distinct forms within a single, interbreeding population. This stable polymorphism, driven by the whims of frequency-dependent sexual selection, can be the first critical step on the path to sympatric speciation—the splitting of one species into two, not because of a geographic barrier, but because of the internal social dynamics of the population itself.

The dance of frequency extends beyond individuals to the intricate relationships between species. Consider the world of mimicry, nature's grand stage for illusion and information warfare. Some butterflies, like the harmless Batesian mimic, have evolved to wear the conspicuous warning colors of a truly noxious species. This is a brilliant deception, but its success is precarious and utterly dependent on frequency. A predator, like a bird, learns to avoid the warning pattern by having an unpleasant experience with the genuinely toxic model. As long as the harmless mimics are rare, the predator's lesson holds: "bright colors mean bad taste." But if the mimics become too common, the bird will increasingly encounter the "lie." After a few tasty, harmless meals, it unlearns the association, and the warning signal becomes meaningless. The fitness of the mimic plummets. The success of this cheating strategy is therefore subject to intense negative frequency-dependent selection; it only works if you're one of a few.

Now, consider a different scenario: two different species of noxious butterflies that have both evolved the same warning pattern. This is Müllerian mimicry, and it operates on the opposite logic. Here, the signal is an honest one, and the interaction is one of mutual benefit. Every individual, regardless of its species, that is sampled by a predator reinforces the same lesson: "avoid this pattern!" The more butterflies that share this "mutual advertising," the faster predators learn, and the lower the risk for any single individual. The fitness of any individual wearing the pattern increases as the pattern becomes more common. This is a classic case of positive frequency-dependent selection, a force that drives convergence and uniformity, rewarding conformity rather than novelty. In these two forms of mimicry, we see the two faces of frequency-dependent selection: one that preserves deception by punishing its success, and one that reinforces honesty by rewarding its abundance.

Nowhere are the stakes of frequency-dependent selection higher than in the silent, perpetual war between hosts and their pathogens. As the Red Queen explained to Alice, "it takes all the running you can do, to keep in the same place." This is the essence of host-pathogen coevolution. Pathogens, with their short generation times, are incredibly adept at evolving to overcome the defenses of their hosts. Which hosts do they target? The most common ones, of course. A host genotype that becomes prevalent is a large, stable, and profitable target for pathogens to specialize upon.

This creates a powerful selective advantage for any host with a rare resistance gene. The rare genotype is effectively invisible to the bulk of the parasite population, which is adapted to the common type. As a result, the rare host thrives, and its resistance gene spreads. But as it becomes common, it paints a target on its own back, becoming the new focus of pathogen evolution. Its fitness advantage wanes and eventually reverses. This endless cycle of pursuit and escape, driven by reciprocal negative frequency-dependent selection, is a major force maintaining vast genetic diversity in the defense genes of host populations.

This is not just a theoretical curiosity; it is etched into our very own DNA. The Human Leukocyte Antigen (HLA) system (also known as the Major Histocompatibility Complex or MHC) is a group of genes critical for our immune system's ability to recognize and fight invaders. It is, by far, the most polymorphic region of the human genome. Why? The Red Queen provides the answer. While having two different HLA alleles (heterozygote advantage) is generally good, the dramatic diversity we see is best explained by negative frequency-dependent selection. As pathogens adapt to the most common HLA types in a population, individuals carrying rare HLA alleles gain a survival advantage. Hypothetical data can illustrate how we might see this in action: as an allele's frequency shifts from common to rare, we would expect the relative fitness of individuals carrying it to increase, a clear fingerprint of frequency-dependent dynamics. This constant selective pressure for novelty ensures that our species as a whole maintains a deep reservoir of immune variability, a crucial defense against an ever-evolving world of pathogens.

The logic of frequency dependence also provides a stunningly simple explanation for one of the most fundamental features of life: the near-universal 1:1 sex ratio. Why don't populations evolve to have many females and just a few males, which would seem to maximize population growth? The answer, famously articulated by R. A. Fisher, lies not in what is good for the group, but in the selfish interests of an individual parent's genes.

Every offspring has exactly one mother and one father. Therefore, the total reproductive success of all males in a generation must equal the total reproductive success of all females. If, for example, males are rare in the population, then each individual male will, on average, have more offspring than each individual female. In this scenario, a parent who produces sons will have a greater fitness return than a parent who produces daughters. Natural selection will favor genes for producing the rarer sex. This is negative frequency-dependent selection at its finest. The selective pressure constantly pushes the population's sex ratio back towards the point where the fitness returns from producing a son or a daughter are equal—and that point is a 1:1 ratio.

This "social calculus" scales down to the level of microbes. Consider a population of bacteria, where some individuals are "cooperators" that pay a cost to produce a public good (like an enzyme to digest food), while others are "defectors" that use the good without contributing. Can cooperation survive? Frequency-dependent selection provides the key. The outcome depends on the economics of the public good. If the good provides diminishing returns (a concave benefit curve), then the first few cooperators in a group of defectors provide a huge per-capita benefit. Cooperation is most advantageous when it is rare, a case of negative frequency-dependent selection that can lead to a stable coexistence of cooperators and cheats. Conversely, if the good is synergistic, with accelerating returns (a convex benefit curve), then cooperators only do well when they are already common. This is positive frequency-dependent selection, which can lead to an all-or-nothing world of either pure cooperation or pure defection.

From the song of a cricket to the biochemistry of our immune system, frequency-dependent selection is a pervasive and powerful force. It demonstrates that in the complex, interconnected web of life, context is everything. The value of a gene, a trait, or a behavior is not fixed but is constantly being renegotiated in a dynamic feedback loop with the rest of the population. This principle even extends into the realm of human culture, where the value of a skill, the impact of a fashion, or the vulnerability of a security system can depend on its frequency in society. Understanding this dance of rarity and abundance is not just key to understanding evolution; it is a lens through which we can see the deep, unifying logic that connects the myriad forms of life on Earth.