Menu Close

Natural history of the present

Boys, war and the prescient mosquitofish mother

Some aspects of the natural world are so commonplace that we mistake them for essential truths. This is especially true of the fact that the animals we know best, including our own species, usually give birth to near-indistinguishable numbers of sons and daughters.

In high-school genetics we learn that half a father’s sperm carries his Y-chromosome, and the other half carry his X. So the 50-50 sex ratio seems like the consequence of thousands of chance conceptions.

But while chance plays a part in each conception, the ratio of sons and daughters at birth results from some of the strongest and most consistent natural selection known to science. The fact that, in humans, the sex ratio at birth leans ever so slightly in a male direction is no accidental quirk.

When nature takes its course, 106 boys are born for every 100 girls. This slight male bias is a consequence of differences in how taxing boys and girls are to raise and how likely they are to survive to reproductive age.

Selection on sex ratios is so strong because when one sex becomes rare, individuals of that sex grow more valuable as mates. So parents who conceive offspring of that rarer sex rather than the more common sex will eventually have more grandchildren through those offspring. Which favours behaviours and physiologic traits that bias conception.

In most species, this effect of competition keeps the sex ratio close to even. But under some conditions selection should favour departures from an even sex ratio.

I have written in this column before about the Trivers-Willard effect, possibly the most surprising idea yet to spring from evolutionary theory.

In 1973, Trivers and Willard predicted that parents who can afford to invest heavily in caring for offspring should bear offspring of the sex that benefits most from lavish parental care.

This prediction has been upheld by hundreds of tests. Wasps lay female eggs on large caterpillars (thus bearing large, fecund daughters) and prefer sons when they have caught a puny prey item to feed the larva. A hind or female antelope in good condition is more likely to have a son because they can afford the extra milk that will set him on the perilous path to becoming a successful stag. In poor condition, she is more likely to bear a daughter.

Evidence consistent with the Trivers-Willard effect has also been reported in humans, although the quality of the data tends to vary. Approximately 60% of the children born to billionaire families are sons.

Junior wives in polygynous Rwandan marriages bear far more daughters than higher-status first wives or women in monogamous relationships. And female infanticide in early-colonial North-West India was peculiar to the highest-born castes, whose daughters suffered very poor prospects of marrying upward (downward marriage being unthinkable).

The case that Trivers-Willard effects can lead some parents to bias conception or care toward one sex and other parents to bias investment toward the other sex has reasonably solid support – although much quibbling surrounds many purported cases. But what about other factors skewing the birth sex ratio?

What happens, for example, if large-scale death of one sex distorts the mating market? In humans, warfare does exactly that, because young men are far more likely to die in warfare than any other group.

Interestingly, after major wars there is often a spike in the birth of boys, as there was after World War I in Britain.

But other conflicts, such as the Iran-Iraq war in the 1980s have precipitated the birth of more girls than boys.

Another problem with the idea of sex-ratios responding to warfare is that the men who die in warfare are already old enough to be sexually active, and so newborn babies won’t really replace those men in the mating market. Unless the conflict runs for decades.

But if parents could anticipate a future shortfall of one sex and then bear offspring of the rare sex, those offspring would enjoy an enormous advantage when they grew up and mated.

The Eastern Mosquitofish, Gambusia holbrooki, is a livebearer from the southeastern USA that has established ferally in many parts of the world, including near Canberra. Andrew Kahn

A paper published in Nature Communications by ANU researchers this week shows that, in one species at least, parents can anticipate a future shortfall of one sex and respond accordingly. They made use of a curious biological quirk in the live-bearing mosquitofish, Gambusia holbrooki.

PhD student Andrew Kahn who led the study told me:

We first suspected something interesting was going on in the previous summer. I was out collecting fish for a behavioural study. One week there were heaps of males around, the next week I couldn’t find any! So we knew there were some strange, sex-specific patterns of mortality happening.

We then collected a pilot sample of females in that autumn, just to see if there was anything interesting going on with their birth sex ratio. A couple of months later, we had a lot of little female fish swimming around the lab, and knew we’d stumbled on something unusual.

Mosquitofish breeding peaks in spring and then again in autumn. Males and females born in spring tend only to breed in autumn and then die. Autumn-born males breed in spring and then die, but many autumn-born females breed in spring and live long enough to do so again the following autumn.

Kahn, together with professors Hanna Kokko and Michael Jennions, spent some time considering and rejecting various explanations for why sex ratios in several lakes around Canberra skew toward sons in spring and daughters in autumn.

They then struck upon the idea that generation overlap might change the seasonal incentives to have sons and daughters and explored how this idea would work by building a mathematical model.

If the birth sex ratio were always 50:50, the number of breeding females in autumn would far exceed the number of breeding males - due to the presence of the older females born the previous autumn.

To capitalise on this rather odd piece of biology, parents should bear more sons in spring (to exploit the female bounty in autumn) and slightly more daughters in autumn (because these daughters enjoy two chances to mate).

They tested this idea, painstakingly sampling the fish in each lake every fortnight for a year. They also took pregnant females from each lake back to the lab and observed the sex ratios when they gave birth.

The strength of the sex-ratio skews was tied to the proportion of autumn-born females who survived a year and bred twice. In one lake hardly any females bred twice and birth sex ratios hovered around even.

In another population, old, large females bore more than half of each Autumn’s young. And in this lake the sex-ratio skews were strongest. Generational overlap and sex-ratio skews in the last lake, were in-between.

What can we generalise from this study?

First, it shows that parents can skew the numbers of sons and daughters they produce in anticipation of a future shortfall of one sex. Admittedly the anticipation in these mosquitofish populations is seasonal and reasonably neat.

But, in theory, parents of many species including ours may be making much more strategic anticipatory decisions than we have yet conceived.

Seldom does a population succumb so easily to such a neat and simple model. Even so, the authors assure me this was a frustratingly complex problem to work out.

I am hoping this paper will open up the theoretic modelling and empirical study of how parents anticipate future mating-market opportunities in other, more complicated populations.

Then, perhaps, these ideas may help us better understand what happens when human sex-ratios vary catastrophically. This tends to happen after war, some famines, and is particularly evident due to sex-biased abortion in northwest India and China right now.

In generating new and previously never-imagined ideas and exposing them to test, science demonstrates its greatest strength.


Disclosure: Mike Jennions and Hanna Kokko are friends and collaborators of mine, although I have no connection with this study

Want to write?

Write an article and join a growing community of more than 180,900 academics and researchers from 4,919 institutions.

Register now