Darwin Was Wrong

By Joan Roughgarden

I would like to plant the corrosive seed of doubt in your mind that maybe Darwin’s theory of sexual selection is totally incorrect, and that you might actually be better off junking it lock, stock, and barrel.

Darwin’s writings are an enunciation of essentialist gender roles. As Darwin wrote in 1871, “Males of almost all animals have stronger passions than the females,” and “the female . . . with the rarest of exceptions is less eager than the male . . . she is coy.”1

Notice the phrases “less eager,” “with the rarest of exceptions,” and “almost all.” Darwin’s statements are understood to be empirical claims, potentially subject to revision with empirical information. The supposed explanation for these empirical claims is that females choose mates who are more attractive, vigorous, and well armed, “just as man can improve the breed of his game-cocks by the selection of those birds which are victorious in the cock-pit.” The supposition here is that the female peahen chooses to sleep with the cocks that have beautiful tails. Through this successive action of female choice, females breed the males to have the desired traits. This is the mechanism, according to Darwin, for the production of these ornaments.

Even as the facts increasingly reveal that not all males are passionate and not all females are coy, making the empirical justification for this argument increasingly suspect, the argument itself is retained as true. It is as though this third proposition—about female choice for the good genes that males possess—has itself an attractiveness that is coded as truth.

You might say that it’s not fair to criticize Darwin’s writing from 1871, but there are plenty of contemporary works which reveal that this is still the central narrative of sexual selection theory. For example, Jerry Coyne, a geneticist at the University of Chicago, states (I believe arrogantly): “We now understand . . . males, who can produce many offspring with only minimal investment, spread their genes most effectively by mating promiscuously.”2

Here we have the “promiscuous” male instead of the “passionate” male. Notice the reference to “minimal investment.” This idea involves deriving gender role from gamete size. Sperm are cheap, hence males can go attempt to mate with anything that walks. Female reproductive output is far more constrained by the metabolic costs of producing eggs or offspring, and thus a female’s interests are served more by mate quality—and by that he means genetic quality—than by mate quantity. So females making expensive eggs therefore must be coy and choosy, so as not to waste their investment on sperm carrying poor genes.

As you can see, the basic rationale for sexual selection theory today and the central narrative of male-female social dynamics is still basically the Darwinian narrative. This narrative began innocently enough with reference to explaining ornaments like the peacock’s tail and armaments like the antlers on red deer, but it has been elaborated into what I term an “evolutionary system of sex, gender, and sexuality.”3

Over the years, evolutionary biologists have developed a whole collection of inter-locking propositions that pertain to these issues, including the origin of sexual repro-duction and the origin of the male-female binary (separate issues). There are also sub-theories or correlative theories about the characteristics of sexual reproduction, such as how much mating should take place out-side the pair bond. If you look at two birds at a nest, it frequently occurs that the two birds also mate with adjacent birds who are not members of that nest, and there is extensive literature on that phenomenon.

These additional theories have been necessary because of the direct contradictions to sexual selection, such as sex role reversal. It turns out that as the phenomenology be-comes elaborated, there are many species doing all sorts of things that are contrary to the theory. So if you’re going to retain the sexual selection theory as central, but you have all of this diversity, it means that you have to accrete to that theory all kinds of subsidiary theories and peripheral narratives that amount to “special case” explanations. This whole package is the evolutionary system of sexual selection. Because it contains all of these interlocking pieces, the system is impossible to falsify at this point.

Yet, as you might have discerned already, the feeling of evolutionary biologists about this is not one of despair. It’s similar to going to someone’s messy house and they have things stuffed everywhere, so that it’s almost impossible to walk through their living room. But when you say, “I could come on the weekend and help clean it up for you,” they say, “No! Don’t do that. I know where everything is!” They like the mess. That’s exactly what we have in sexual selection theory today: a mess.

Let me briefly touch on some of the problematic cases. First, there’s the absence of a very clear sex binary. For a biologist, sex refers to the size of the gamete that is produced. It is true that there’s a nearly universal male-female binary at the gametic level, in the sense that if a species makes more than one gamete size, it makes only two—it doesn’t make three, or four, or five—and when there is more than one gamete size, one is tiny and the other is big.

The problem, however, is that the binary defined at the gametic level doesn’t extend to the whole-body level. At the whole-body level, you find species that go from making eggs to making sperm, so they change sex, or species that go from making sperm to making eggs, or that make both eggs and sperm at the same time. In fact, the most com-mon body plan among organisms, including plants, is for an animal or plant to make both eggs and sperm, either at the same time or at different times during their life. This means that the binary is a derived condition which is outside the sexual selection narrative.

Another huge issue is that there are multiple templates per sex. I call each one of these templates a gender. So in the case of the ruff, for example, which is a sandpiper-like bird from Northern Europe, there are three completely different types of males. There’s one which has a black collar, from which it gets the name ruff; there’s one that has a white “ruff”; and there’s one with no ruff at all. The female also has no ruff. The black ruff male actually solicits the white ruff male to join him at the breeding territory and they court the females together. The mating occurs by both males of the eggs produced by the female in the nest tended by the large male. The ruff-less male was only discovered about two years ago, so its social role isn’t clear, but there are instances of same-sex mating between black ruff males and ruff-less males. This is fundamentally a problem for the sexual selection theory, because the Darwinian narrative sets up one template per sex. You can’t just fix the template.

The problem of sex role reversal is also very important, illustrated by the common seahorse and also by pipefish. In these species, the males carry the eggs in a pouch and the female has to deposit eggs in the pouch of the male. Because it’s possible for the females to produce eggs faster than the males can “graduate” them, or hatch them, there could be a limitation to the number of males who are eligible to carry eggs. This has been termed “sex role reversal.” Some scientists have argued that this example sustains sexual selection theory. In my opinion, this illustration is a restating of the problem, because none of the theories adequately ex-plain when this circumstance should occur. Simply reversing the theory to be applicable to whichever sex is more available is merely renaming the phenomenon, not explaining the phenomenon. In fact, this evidence falsifies any connection between gamete size and sex role. The males in these species are making tiny sperm—that’s why they’re males—nonetheless they’re providing the primary parental investment. Sex role reversal actually refutes the observation of Coyne and others that males are naturally promiscuous because sperm are cheap. These species make sperm, and their sperm are clearly not cheap.

Finally, a huge issue is the prevalence and commonness of same-sex sexuality. There are documented cases of both homosexual and homogenderal matings for over 300 species of vertebrates. (One example is Bonobos, who often walk bipedally, so they’re very human-looking.) According to the Darwinian narrative, this is a complete waste of time, and yet there’s a lot of time being put into it. Unless these species have a lot of free time, something else is going on that bears investigation and explanation.

My criticism of sexual selection theory has been derided as emphasizing exceptions and as not addressing the supposed commonness of the great majority of species which are supposed to be following the Darwinian templates and to be explained by the Darwinian narrative. I dispute this characterization completely. It’s far from clear that there are any species that actually obey the Darwinian template for the Darwinian reason. More and more studies are coming out to this effect. All of the “poster species” that are supposed to be classic examples of Darwinian sexual selection do not to hold up under close examination.

For example, a reputable 24-year study was done on the collared flycatcher on an island off of Sweden with over 8,000 marked birds.4 There’s a white spot on the head of male collared flycatchers which females are supposed to want so that their male offspring will be endowed with the spot, too. In evolutionary biology, that’s called the “sexy son hypothesis.” But what did the extensive study reveal? The spot is slightly heritable, with a heritability of .3 (which is low), but the variance in fitness has zero heritability. So if you look at a bunch of different males and some of them sire a lot of young and some of them don’t, it has nothing to do with their genes. Therefore, there’s no reason for a female to choose a male on the basis of his genes because the choice is useless. There is zero genetic correlation be-tween preference and fitness of the young. The choice of a mate is based on present-day circumstances, not on genes.

Time after time, if you actually look into a case that is purported to exemplify sexual selection, it collapses. Thus, we have the so-called “exceptions” not being explained, and the cases that sexual selection is supposed to explain decomposing before our eyes. Furthermore, there are a lot of internal contradictions to sexual selection, and logical ones as well, such as the proposition that there would be genetic variation for quality of male after generations of female choice. If females are continually weeding out males for bad genes, after about 20 generations of this, there should be no bad genes left. This is called the “paradox of the lek.” It’s a logical flaw in sexual selection theory which has engendered a literature that’s intended to resolve the paradox. My claim is that it’s not resolvable because it is truly a grave contradiction, and the theory of sexual selection itself is wrong.

I’ve suggested and elaborated an alternative to sexual selection called social selection. Sexual selection interprets social behavior that takes place during reproduction, like courtship, as being all about mating, about how to get mates. Therefore, natural selection arises from differences in mating success. In this account, females are the limiting resource for males. Males compete for mating opportunities, and females choose males for their genes.

As I see it, behavior is an offspring-rearing system. The focus is on the production and rearing of offspring, not on the acquisition of mates. Natural selection, there-fore, arises from differences in offspring-rearing success rather than mating success. Mating is only, at best, one component of  evolutionary success. The problem with sexual selection is that is has elevated one component of evolution into a goal of its own.

Males and females in this account negotiate bargains and side payments to maximize offspring production and control, and to control the offspring-rearing social infrastructure. Offspring are not just raised, they’re raised in a social system. A lot of the behavioral interactions that are going on have the function of managing the social infrastructure within which the offspring are raised. We’ve begun to elaborate in my lab an entire alternative system to the theory of sexual selection that addresses every one of these items, from the origin of sex down to the implications for humans.

After considering the empirical evidence, I invite you to consider the outrageous possibility that we need to rethink the evolutionary biology of sex, gender, and sexuality from scratch.5

This talk is from the panel discussion “Evolution, Cooperation, and Gender,” held at Harvard Divinity School on April 12, 2007.


  1. Charles Darwin, The Descent of Man, and Selection in Relation to Sex, facsimile edition (Princeton University Press, 1871).
  2. Jerry Coyne, “Charm Schools,” Times Literary Supplement, July 30, 2004.
  3. I echo here the anthropologist and gender theorist Gayle Rubin’s phrase “sex-gender system.”
  4. Anna Qvarnström, J. E. Brommer, and L. Gustafsson, “Testing the Genetics Underlying the Co-evolution of Mate Choice and Ornament in the Wild,” Nature 441 (May 4, 2006): 84–86.
  5. A longer article reviewing all of this in greater detail was recently published in Daedalus: Joan Roughgarden, “Challenging Darwin’s Theory of Sexual Selection,” 136, no. 2 (2007): 23–26.

Joan Roughgarden is Professor of Biological Sciences and of Geophysics at Stanford University. Her most recent book is Evolution and Christian Faith: Reflections of an Evolutionary Biologist. This talk is from the panel discussion “Evolution, Cooperation, and Gender,” held at Harvard Divinity School on April 12, 2007.

Please follow our Commentary Guidelines when engaging in discussion on this site.