For several decades, evolutionary biologists have been trying to figure out the forces that set this balance. It appears that they come down to a tug of war between competing interests. Imagine a species in which a freakish mutation makes the females gives birth to lots and lots of daughters. If you're a male, suddenly your chances of reproducing look very good--certainly better than all those females. Now imagine that some of these lucky males acquire mutations that makes them father more sons than daughters. This son-favoring mutation would spread because of the advantage to being male. In time, these mutations would tip balance of the sexes over to the males. The now-common males would have less chance of producing offspring than the now-rare females. The advantage shifts to the females. Over time, these opposite forces pull the ratio back and forth until they settle down into an equilibrium.
Sometimes, though, the ratio of males to females lurches out of balance. Some insects, for example, only give birth to daughters. That's because a third player has entered the tug of war--a bacterium called Wolbachia. Wolbachia lives in animal cells, and so the only way it can survive beyond the life of a host is to get into its eggs, which then grow into adults. Since Wolbachia cannot fit into sperm, males are useless to it. And so it has evolved a number of tools that it uses to forces its female hosts to give birth only to daughters.
But biologists have also noticed that in some situations other animals--including humans--can give birth to an overabundance of sons or daughters. In the early 1970s, Robert Trivers and Dan Willard, both then at Harvard, wondered if mothers might be able to control the sex ratio of their offspring to boost their own reproductive success. They imagined a species, such as deer, in which relatively few males mated with most females. If a doe gives birth to a healthy male, he is likely to grow up into a healthy buck that has a good chance of producing lots of grandchildren for the doe. In fact, he would be able to produce far more grandchildren than even the most successful daughter. He could impregnate lots of mates, while a daughter could produce only a couple offspring a year, which she would then have to nurse. So if the doe is in good health and can give birth to strong offspring, she would do best to produce sons.
On the other hand, if this doe gives birth to a male in poor condition, he may be unable to compete with other males, and his chances of reproducing fall to zero. Since most females that live to adulthood give birth to at least some offspring, it would make more sense to give birth to a daughter rather than a son in poor condition.
Trivers and Willard speculated that if a mother could somehow gauge the prospects of her offspring, she might manipulate their sex ratio to her own evolutionary advantage. In bad times, she'd produce females, and in good times she'd produce males.
Since Trivers and Willard first published their idea in 1973, scientists have tested it in hundreds of studies. Some of the results have been quite powerful. Scientists moved a bird known as a Seychelles warbler from one habitat to another and measured the sex ratio of their chicks. In places with lots of food, they produced lots of daughters that stayed at the nest to help raise their younger chicks. In places with little food, the ratio swung in favor of sons, which flew off in search of new territory. But the results have been far from clear-cut, particularly for mammals, which has led some researchers to wonder whether this particular force is very strong in mammal evolution.
In an article in press at The Proceedings of the Royal Society of London, South African zoologist Elissa Cameron of the University of Pretoria argues that the case for adjusting sex ratios is actually very good if you look at the evidence properly. She analyzed over 400 studies of sex ratios and noticed that, depending on the study, the scientists measured the condition of mothers at different points in their pregnancy. Some took their measurements at conception, some in the middle of gestation, and some at birth. While studies during gestation and at birth provided ambiguous results, almost all the studies done around conception gave strong support to the Trivers-Willard hypothesis.
These results, Cameron argues, indicate that mothers can shift the balance of the sexes, but only right around conception. It's possible, for example, that the amount of glucose in the uterus when an egg begins to divide may trigger different responses, depending on whether the egg is male or female. For example, high glucose may signal that the mother is doing well and would do well to raise sons, while low glucose would favor females. In a paper in press at Biology of Reproduction, Cheryl Rosenfeld and Michael Roberts of the University of Missouri review some evidence that may support Cameron. (You can download it for free here.) They raised female mice on two different diets--one high in saturated fats, and one high in carbohydrates. Mothers eating a high-fat diet (which probably led to high levels of glucose) gave birth to litters with two sons for every daughter. Mothers eating high-carb diets produced about one son for every two daughters.
And here's where test tube babies come in. Humans may not have the sort of mating imbalance that you find in deer, but a lot of evidence suggests that we descend from a long lineage of primates in which a few males mated with a lot of females. It wouldn't be surprising, therefore, to find adaptations in women to favor sons or daughters. Rosenfeld and Richards survey some interesting studies on census data that suggest that this does indeed happen. And test tube babies offer some clues about the biochemistry that may be at work. When doctors fertilize a woman's eggs and let them begin to divide into a ball of cells, they keep the embryo in a solution of glucose. The farther along the doctors let the embryos develop, the more likely it is that their patients will wind up with sons rather than daughters. In this glucose-rich environment, male embryos may thrive, while females may risk failure to develop. Even in an age of reproductive technology, it seems, we are grappling with our evolutionary legacy.