I’m teaching introductory evolution this quarter, and am using as a textbook Doug Futuyma’s Evolution (second edition, Sinauer). Today’s lecture will be on the maintenance of genetic variation via natural selection (heterosis, etc.), and in the textbook under “frequency dependent selection,” I see this on page 319:
Why is the sex ratio about even (1:1) in many species of animals? This is quite a puzzle, because from a group-selectionist perspective, we might expect that a female-biased sex ratio (i.e., production of more females than males) would be advantagesous because such a population could grow more rapidly. [JAC: such a sex-ratio-biased group would then outcompete other groups and predominate]. If sex ratio evolves by individual selection, however, and if all females have the same number of progeny, why should a genotype producing an even sex ratio have an advantage over any other?
The answer, first realized by Ronald Fisher in 1930, is that there is individual “frequency-dependent” selection that enforces an even sex ratio. Consider, for instance, a population in which females predominated, and males were rare. In that population, a female who produced more males would have more grandchildren than other females, for the average reproductive success of her offspring will be higher. (Imagine if there were only one male and elebenty gazillion females in a population. A mutant female producing mostly or all males would have huge numbers of grandchildren, for her male offspring would inseminate most of the females. Evolutionarily, whatever genes gave her that male-biased sex ratio would increase in the population.)
The reverse would be the case if males predominated in the population: any mutant individual producing more females would leave more grandchildren.
In this case, then, the rarer sex always has a reproductive advantage, and any variant individual producing the rarer sex would have an evolutionary advantage. The upshot is that the sex ratio will reach equilibrium only when there IS no rarer sex, i.e., when there are equal numbers of males and females. In such a case no new mutant individual will have a reproductive advantage. This has been tested experimentally by varying sex ratios in species which have three sex chromosomes, and populations always settle down at the 50/50 sex ratio.
Although there are some exceptions to a 50/50 sex ratio in animals, most conform to the 50/50 value. This is precisely what is expected if sex ratio is a result of individual and not of group selection. Ergo, when the two are in evolutionary conflict, as they are here, individual selection wins. And evolution is the answer to a question you’ve probably never asked yourself: why are there as many females as males?
I still know of no adaptation in nature that is explained more plausibly by group selection than by individual or kin selection; but there are plenty of adaptations, like sex ratio, easily explained by individual or kin selection.
It’s time for biologists to stop banging on about group selection until we find evidence that it has actually operated in nature. We don’t have time to waste on theoretically plausible but infrequent mechanisms for which there’s no evidence.
sub
In humans we might expect that group selection would instead favour groups with more and more aggressive males, as these groups would be more likely to kill the other groups.
The trouble is you would get a stage when all groups are aggressive as the characteristic spread. As it is aggression can have drawbacks e.g. liklihood of death or injury reducing the chance of finding a mte before your demise.
All fitness criteria are local, even including aggression. ‘More likely to kill other groups’ may be a positive individual adaptation some times and in some ecologies, but won’t be in others. Particularly not in situations where the individual can be expected to gain from peaceful interaction with other groups (mate acquisition, trade, etc.). So no, personally I would not expect this in humans.
Just to clarify, this was a response to DF/JAC’s point that group selection would imply a female-biased sex ratio.
I’m just pointing out that in humans, with their propensity for killing each other, it might instead imply a male-biased sex-ratio.
This makes hermaphrodites totally understandable. The recent article on the BBC website about sea slugs – on the radio a sea slug expert said as they are solitary & not found in vast numbers, being a hermaphrodite makes sense as it ensures every encounter with another sea slug could lead to mating. What are the stages that lead from hermaphrodites to having two separate sexes?
Here is the sea slug link.
http://www.bbc.co.uk/news/science-environment-21431678
Also today fab pictures of White Sea jellyfish on the Guardian website…
I can’t answer your question, but I can mention that I’m reading Ursula K. Le Guin’s “The Left Hand of Darkness,” which is about a hermaphroditic society; when two individuals start to have sex, hormonal changes cause (at random) one individual to become male and one female. Hence, an individual can be, and usually is, both a father and a mother during its life. Needless to say, this has major effects on the structure and nature of society.
That would I venture to guess, lead to eventual differentiation into two sexes, as it is usually not advantageous to be female & have to give birth & require the additional resources that requires – hence the habit of some slugs to chew the penis off the other…
Since you bring it up, in your opinion, is there a wholly satisfactory individual selection explanation of sexual reproduction itself? I thought that was a bit of a problem.
IIRC clone populations can be less stable. They crash more, because when they encounter something bad for one of them, its bad for all of them. In terms of individual selection, what this means is that the individual who gets half their genes into two offspring (sexual reproduction) will generally have more of those genes survive than the individual who gets all of their genes into a single offspring (asexual reproduction). Obviously not true in all cases, as many single celled critters are asexual. But relevant to many species; sexual reproduction will give a net gain in expected genetic offspring in ecologies where crashes like this are a regular or significant issue.
Several years ago Olivia Judson had an interesting NYT article about the occasional incidence of male cloning. Seems nonsensical at first, but it occasionally happens that a mutation crops up allowing males to completely biologically co-opt reproduction, in essence eliminating the female genetic contribution in the womb. Upshot – it never lasts. If I find the article I’ll link to it here (but I may not be able to find it).
Here it is:
Evolving the single daddy.
Here’s another one on the subject of sexual and asexual reproduction: Ciliates have up to 100 sexes and reproduce through a combination of asexual and sexual reproduction.
From your first to your third third sentence you transition from group selection to individual selection.
The issue of group vs. individual-level selection is not a question of mutually incompatible alternatives. The question is, which makes more sense? If we -can- reduce a group-level explanation to an individual-level explanation, is this always the best thing to do from the standpoint of communication and comprehension?
In other words, why not just stop at your first sentence? If the explanation is obvious and straightforward at one level of analysis, do we gain anything by pursuing reductionism below that level?
The same issue arises with individual- vs. gene-level explanations, but for some reason this is consistently overlooked by opponents of group selection. Any individual-level explanation can be reduced to a gene-level explanation. So why do we ever bother talking about individuals as units of selection? Presumably because doing so typically results in simpler and more comprehensible explanation. So, why are we OK with rejecting reductionism in favor of comprehensibility in this case… but reject comprehensibility in favor of reductionism when anything above this particular level of analysis is proposed? Let’s at least choose one consistent guideline, rather than drawing an arbitrary line and declaring–“Above this level of analysis, we prefer reductionism; below this level of analysis, we prefer comprehensibility!”
Or, if we do choose to work with that kind of inconsistent and arbitrary set of rules, we can at least admit it… and admit that group selection isn’t -wrong-, it’s just a kind of explanation we’ve decided to reject by arbitrary fiat.
Confused about whose 1st through 3rd sentences you are referring to. To answer your points: I agree with your general point but that’s precisely why I think most of us prefer an inclusive fitness model: I find group selection LESS comprehensible and an impediment to clear communication. If it were more useful, I would definitely adopt it. As for the genes–>indviduals and individuals–> group equivalency you suggest, I don’t think that works. There are plenty of reasons to talk about individuals and genes if you have a gene’s eye perspective: Individuals are the vehicles through which genes realize their strategies. Individuals are always quite discrete entities, unlike groups which have a pretty fuzzy definition in my mind. For the record, we “opponents of group selection” also have no problem considering the group when it is clear what a group is (like an ant colony), but prefer to consider the ways in which said group affects the genes inside it. That’s what sex ratio theory is all about: how does what the group does affect what’s best for an individual gene to do.
Can of worms. Basically, there are a bunch of theories for why “sexual reproduction” (not differentiation of sexes, but outcrossing with recomnbination) occurs. But no clear way of knowing which of them are involved.
I’ve worked in this area, and believe me, you don’t want to ask for a quick explanation.
What are the explanations for exceptions to the 50/50 sex ratio?
Mainly, unequal parental investment in offspring of the different sexes. If, say, females are twice as big as males at weaning and therefore require twice as much feeding by both parents, the equilibrium sex ratio is 1:2, i.e. equal investment in each sex.
If the parents are not both involved in provisioning the young, this leads to conflict (and usually compromise) as they have different optimum ratios. In social hymenoptera (which are haplodiploid, unfertilized eggs becoming males, those fertilized inside the queen with stored sperm becoming females), queens get to pick the sex ratio they want because males have no control. Then there’s another layer of conflict as sterile workers get to decide how many female larvae will get the special nurture that makes them fertile new queens.
Thank you John.
That’s a question I have. The human gender ratio (at birth) is not exactly 1:1. It’s close but, I believe, consistently slightly biased in favor of males.
One reason may be that mortality amongst males is higher (we engage in riskier behaviour) so there would be an evolutionary advantage to producing slightly more male children.
Mike.
The human sex ratio at birth is slightly biased in favor of males. Because males have higher juvenile mortality, the sex ratio is 1:1 at about age 20, which is generally the age of highest reproductive value (a measure of the expected number of offspring, which also includes the probability of survival). I’m not sure how rigorously this empirical pattern has been tested, but the frequency dependence argument would lead to a prediction of 1:1 at the time of reproduction, not at birth.
I’m not sure how rigorously this empirical pattern has been tested, but the frequency dependence argument would lead to a prediction of 1:1 at the time of reproduction, not at birth.
No, the prediction of 1:1 is at the time of sex ratio control by the parent. A doubling of the mortality rate of males doubles the reproductive value of the males that do survive: they will have twice the number of mates. These effects exactly cancel. The equilibrium sex ratio is therefore independent of sex-specific mortality after the period of parental investment.
Your assertion is incorrect. If it were 1:1 at the time of control, it would be 1:1 period, with no allowance for differences in investment. Anything that happens after successful conception cancels out, according to your claim.
But that’s not what happens, of course. Anything which affects reproductive success feeds back into the ratio control. That includes parental investment and sex-specific mortality.
You’re right. My argument assumes there are no differences in sex-specific investment after sex ratio control. I was trying to point out that it’s not true in general that sex ratio selection favors a 1:1 sex ratio at sexual maturity. Investment after successful conception does play a role, as you point out correctly, but it doesn’t imply that the sex ratio at sexual maturity should be 1:1. Sex-specific mortality after parental investment and before sexual maturity has no effect on the equilibrium sex ratio. Agreed?
Agreed that selection on parents leads to equality of investment in the sexes at the cessation of parental investment (not birth). Fisher also argued that, once parental investment ceased, selection on the now independent offspring (not parents) would also tend toward equality, barring other factors: “Any great differential mortality in this [post-investment] period will, however, tend to be checked by Natural Selection, owing to the fact that the total reproductive value of either sex, being, during this period, equal to that of the other, whichever is the scarcer, will be the more valuable, and consequently a more intense selection will be exerted in favour of all modifications tending towards its preservation,” which is a continuation of his frequency-dependence argument. In Fisher’s human data the peak reproductive value was at 18.5 years, so that the end of parental investment and the onset of reproduction would not be far apart in time. (Quotation from the Genetical Theory of Natural Selection, p. 143.)
Cultural evolution. Such as what’s happening in China and India where males are the preferred offspring.
That’s a cultural fad potentially operating as a strong selection pressure, if there is available variation in genes affecting sex ratio.
This is not what the term ‘cultural evolution’ usually means. It’s really a case of artificial selection, but may be as ineffective as the selection for female-biased sex ratio in domestic fowl and dairy cattle which must have been going on for thousands of years.
There is a nice long discussion of sex ratios in chapter 4 of Matt Ridley’s “The Red Queen.
Thank you Mark.
Fascinating! Thanks for the explantion.
I love evolution (even if it is all witchcraft created by the devil in order to destroy god).
That’s precisely *why* we love it!
Could someone give examples of the exceptions to the 50/50-rule? At what stage is there discrimination (gamete-embryo-fetus-newborn)?
What is the biological mechanism behind such discrimination?
Jerry, thank-you for the clear explanation. If I understand you correctly, this is an example of negative selection. In other words, all those mutations that would alter the sex ratio will be eliminated from the population. Is that right?
Is this strong negative selection? It would have to have pretty severe negative effects on fitness or else we might expect populations to harbor considerable variation among individuals with respect to sex ratio.
I don’t think there’s real positive selection for equal numbers of XX and XY sperm during meiosis since that’s what you would normally expect during cell division. It’s the default ratio.
Is there any evidence for mutations that simply alter the sex ratio in vertebrates but have no other effect? If so, how frequently do these mutations arise in a population? Are there any vertebrate populations where some individual lineages have more males than females or vice versa? (They would be in the process of elimination but if selection were weak, we would expect some variation.)
Don’t you need to have evidence that there are actual mutations being selected against in order to invoke “frequency-dependent selection”?
It seems to me that a species that did, say, temperature-based sex determination would be an excellent avenue for such questions. While the XY and ZW systems would default to 50%, that certainly wouldn’t be the case for a given temperature to implicitly lead to one gender or the other.
You might be interested in a paper by me and my colleagues on a species of lizard that has populations with temperature-based sex determination and populations with genetic sex determination.
http://www.sciencemag.org/content/250/4987/1556.abstract
Don’t social Hymenopteras have female biased sex ratios of about 3:1 with lots of variable plasticity? I think at one time group selection was put forward as one of a few possible explanations but more recently the prevailing theories all have to do with kin selection.
Hi Dr. Coyne,
As you mention, there are some glaring examples to the 50:50 ratio in nature, notably the eusocial hymenoptera. A new paper in the American Naturalist gets at this split sex ratio and why there is a preponderance of haplodiploid eusocial systems:
Haplodiploidy, Sex-Ratio Adjustment, and Eusociality (2013). Garnder A, and Ross L. The American Naturalist.
In sum, queens may exploit the common sex (females) and make them more likely to help since each additional female’s direct reproduction becomes less valuable as the number of females increases.
Reblogged this on DownHouseSoftware and commented:
Although this post is about individual vs group selection theories, I would like to focus on the example put forward. – Why the sex ratio is bound to remain approximately 1:1 in many organisms? I discuss this often in my own general bio class when we come to stabilizing pressure on populations, but I think I will start using this extreme example to make the point even clearer.
“a population in which females predominated, and males were rare. In that population, a female who produced more males would have more grandchildren than other females, for the average reproductive success of her offspring will be higher.”
That doesn’t sound quite right to me. The limiting factor in reproduction (in animals with primarily maternal care, at least) is not normally the number of males, but the number of viable females. Adding more males (which in mammals, at least, is not up to the female or her genetics) would not necessarily increase the relative population. Because males are relatively unlimited in the number of offspring there is not a strong selective pressure to increase their numbers as there would be for females.
A female with a majority of sons over daughters in a population dominated by females would have many more grandchildren because her sons would have many more opportunities to reproduce than he daughters would.
That is the benefit of being male.
The male favoring mutation would come to dominate in the population until the population it’s self reached 50/50.
this is what I’m questioning a bit.
First, since a female does not have a significant ‘say’ in the gender of her offspring (in mammals at least), I’m not sure how that would be a heritable trait from her perspective.
Secondly, in a reasonable female dominant population, a female with sisters would be part of a genetic advantage, but a male with brothers would be much less of an advantage. His brothers would be competitors for available females and he would be at a genetic advantage without them. Without them he’d get more successful matings.
The issue is not male-male competition or female-female support, the issue is who has more kids, males or females. With (illustrative example only) 10 females for every male, males can be expected to average more offspring than females because every female has to partner with some male to have a kid, and so some males must be doing double (triple, dectouple) duty compared to the females.
to follow my point a bit further:
(assuming a numerically female colony)
A female with 3 sisters will (as a group) produce about 4 times as many closely related offspring.
A male with 3 brothers might produce some more offspring, but it won’t be linear, since it’s the number of females that sets the limit. As a group they’d only get a few extra mating opportunities over a single individual.
Hence with a numerical difference in selective advantage, we would see a natural balance between males and females, but it would not seem to work out to 1:1. My gut feeling would be about 1 male to 4-6 females.
You sure it’s your gut you’re thinking with? 🙂
Any increase (or decrease) in population size is a red herring, likely to set your thinkng along vaguely group-selectionist or ‘good-of-the-species’ lines. The tendency to higher variation in male reproductive output is also an unnecessary distraction, as it just cancels out in the population average. Ditto for talk of brothers and sisters, as kin selection isn’t involved in the solution.
Think instead of the fact that (for a normal diploid species) each individual has one male and one female parent. So necessarily, if there were 6 females for every male in the original population, each male would have (on average) 6 times the fitness of each female. The optimum strategy for a gene in a parent’s body (assuming male and female offspring cost the same to produce) is therefore to make as many males as possible.
Look at it from the point of view of two grandmothers. Granny A has one son and one daughter, while Granny B has two sons. Suppose females outnumber males by ten to one. So each male will on average fertilize ten females.
The upshot is that Granny A has 11 grandchildren (one through her daughter and ten through her son), whereas Granny B has 20 grandchildren (ten through each son). Granny B’s genes (including the two-sons gene) therefore outnumber Granny A’s in the second generation.
though you’re assuming an unlimited number of females is available. The question is the optimal number of males vs females.
Suppose you are at 50/50 (or even 40m/60f). Adding more males will not increase the reproductive success much at all (success per male would be reduced), adding more females has a drastic effect. As a parent, throwing more females into the mix will greatly increase your longterm success.
The example given in the article justifies stable sex ratios. It does NOT however justify the magic 1:1, which would only occur (under that causation) if the relative advantages and disadvantages of sex ration change were EXACTLY equal between the genders. That’s a mighty strong assumption, and I believe, unlikely to be generally true.
“Suppose you are at 50/50 (or even 40m/60f). Adding more males will not increase the reproductive success much at all (success per male would be reduced), adding more females has a drastic effect. As a parent, throwing more females into the mix will greatly increase your longterm success.”
No, that’s not true. Assuming we are talking about humans. A female can produce a maximum of one offspring per 9 months. A male doesn’t pay the cost of gestation. His ability to produce offspring is only slowed by access to females who are fertile.
In a population of 100 (for ease of calculation) with a single male, that male could produce 99 offspring every 9 months. With two males, each male could produce 49 offspring every 9 months. At 40m:60f a given male still has slightly greater potential reproductive success than a
female because he could still produce more than 1 offspring every 9 months.
When males are more numerous, a given male will on average produce less than one offspring every 9 months. Thus, to maximise the reproductive output of your offspring, it pays to produce the rarer sex assuming all other variables are equal.
“The example given in the article justifies stable sex ratios. It does NOT however justify the magic 1:1, which would only occur (under that causation) if the relative advantages and disadvantages of sex ration change were EXACTLY equal between the genders.”
This is also not true. The fitness distribution does not need to be symmetric. It just needs to decline either side of 1:1.
“The limiting factor in reproduction (in animals with primarily maternal care, at least) is not normally the number of males, but the number of viable females.”
You seem to be confusing what limits population growth with what determines fitness. An individual’s fitness is not determined by the population size in the next generation, but the proportion of the population that carries its genes.
If the same element that pushes a female biased sex ratio can also reduce the reproductive success of non-carriers mating with carriers, does this alone allow the ratio persist? I’m thinking specifically of Wolbachia
For me, great evidence for the argument that the 50/50 sex ratio comes from individual selection comes from the exceptions. We see sex ratios at birth very strongly favoring females in some species. In these species, brother/sister matings are the norm. One brother can inseminate many sisters, so the mother’s reproductive success (production of grandchildren) is greatest if she produces mostly daughters and just enough sons to fertilize them all. This happens in some insects.
Yeah, but the Wilsons (D.S. and E.O.) consider this a good example of group selection! Go figure. See their Q Rev Biol paper.
Link to Wilsons paper
Darwin was intrigued by the 50/50 sex ratio. He was unable to solve this problem and found it safer to leave its solution for the future. I wonder if this was because he could not see with sufficient clarity at what level (levels?) natural selection acts. Now we know. It is DNA. The only true replicator is DNA. So in the end this is what is selected from one generation to the next. Modify DNA and the modification will be inherited by future generations. This is the basis of the hereditary changes which are critical to evolution by natural selection. DNA closely orchestrates the development of the embryo. In turn the embryo’s development will give you the specie. Speciation and adaptation require modifications to DNA. A change like amputation does not reappear at the next generation. Why? Quite simply, DNA is left unchanged by such an operation.
A gene X must establish its “worth” (or luck) by invading a “group” before group selection (within a given specie) can even come into play. If so why stop at the group and not continue little by little the already successful march and see the gene pool capitulate to the invasion of gene X. All that is needed is the survival and the (exponential) multiplication of the “fittest gene”.
Since there seems to be no convincing model or evidence for adaptation through group selection it is hard to see, for a non biologist like me, what the fuss is all about. I have read Gould’s “The Structure of Evolutionary Theory.” I found it confusing and devoid of convincing evidence for group selection. I will happily read any book that can provide solid evidence of the truthfulness of group selection. In the meantime I will remain unconvinced.
Couple of comments. The idea that more male than female human babies are produced has been explained by the greater survivorship to sexual maturity of females. However there was a paper in Nature some years ago which found a few populations where females are preferred over males. In those populations there is a predominance of female babies. So the neat biological explaination is not universal.
The self fertilizing hermaphrodite fish, Cryptolebias mormoratus can occur in very dense populations. It is also good at pioneering.
I read that female red deer can control the sex of their offspring. When times are hard, female offspring predominate. Red deer do harems, so a runty male has little chance to reproduce, but a runty female is OK.
I wonder about situations where temprature, alkalinity, etc. affect sex ratios in various fish species. I also wonder about initial sex rations in fishes which have harems. Seems like a female would have an easier time joining a harem than a male in creating one.
Hi Jerry,
I taught this just the other day as an example of why thinking about adaptations on a population level can be seriously misleading. I also often teach extraordinary sex ratios and sex ratio distortion [selfish X’s etc] in the same lecture.
However, Fisher was not the first to offer this explanation for the 50/50 sex ratio, with the earliest math being done by Dusing in 1884
http://www.ncbi.nlm.nih.gov/pubmed/11120652
http://www.jstor.org/discover/10.1086/286141?uid=3739560&uid=2&uid=4&uid=3739256&sid=21101737975481
in perhaps ones of the earliest mathematical arguments in evolutionary biology.
Graham
Robert K. Colwell (1981) Group selection is implicated in the evolution of female-biased sex ratios. Nature 290: 401-404.
http://www.nature.com/nature/journal/v290/n5805/abs/290401a0.html
sub