|Tammar wallaby, with a joey|
There are several reasons why this might be useful, but, for the moment, lets just look at two of the more popular ones. The Trivers-Willard hypothesis applies to species where males normally mate with several females, and states that healthy mothers should tend to have more sons than unhealthy ones. The assumption is that less fit mothers will also tend to have less fit offspring, and that they want to have as many grandchildren as possible. For female offspring that's not much of an issue; they're going to mate with somebody at some point anyway.
But for males offspring, it makes quite a difference - if male offspring aren't fit, they won't get the chance to mate, because the bigger males grab all the females first. So, if you're healthy, have lots of sons, because they will give you lots of grandchildren, but if you're not so fit, concentrate on daughters, because your sons won't have much luck. Its worth noting that 'healthy' in this context, doesn't necessarily refer to disease (although it could), but to anything that affects the chances of males mating - genes for larger antlers, perhaps, or for a more sexually dominant attitude.
The "local resource competition" hypothesis, on the other hand, suggests that mothers living in areas of dense population should have more male offspring than those living elsewhere. This is because, in most mammal species, males tend to wander off in search of females when they reach adulthood, while the females stay around in the area they were born - and, as a result, ensure they don't end up mating with their brothers. (There are a few species where the opposite happens, in which case this theory predicts that they will want female offspring when the population is high). The reason for this is that there is only so much food to go round - if you have lots of daughters, they will hang around and leave less food for you. If the population is high and food is scarce, you want sons that will wander off and find less inhabited places where they aren't pestering you, but, if not, you can afford lots of daughters.
There's no particular reason, of course, why both of these explanations can't be true at the same time, which can make it very difficult to tell what's going on in a particular species.
But all this raises the question of exactly how you ensure you have more children of one sex than the other. Perhaps the most obvious reason why you wouldn't expect this in mammals is that whether your child is male or female depends on their genetics. Males have the XY chromosome pattern, and females the XX, and the two should be produced in reasonably equal numbers. (It's worth noting, incidentally, that the XX/XY system is a peculiarly mammalian thing and isn't true, for example, of reptiles). There is some evidence that rhinoceroses, among others, might be able to influence the survival of embryos in the womb, although the details are not fully clear.
Another problem is that, compared with other animals, mammals spend a lot of time looking after their offspring. Even by the time you've finished being pregnant, and discovered that your offspring isn't the sex you wanted, you've already put a lot of biological resources into raising them, and its a bit of a waste to abandon them now. On the other hand, mammals do at least have the option that they could do that if they really wanted, by giving the offspring less milk. That won't work for mammals that give birth in litters, but it might for those that only have one at a time. In fact, a study conducted on tammar wallabies last year did seem to demonstrate this. The researchers swapped baby wallabies between different mothers, and found that mothers who had given birth to sons did a better job of raising their foster children, even when the foster children were (unbeknownst to the wallaby) female.
Indeed, it has recently been suggested that marsupials are exactly where we should look for sex selection in mammals. Although the authors give a number of reasons for this, perhaps the most apparent is that pregnancy is really very brief in marsupials. That means that, compared with placental mammals, putting less effort into raising offspring after birth wouldn't be quite such a waste of energy. The authors point out that strong biases, mostly in favour of males, do occur in many marsupial species, to the extent that this can be a nuisance for anyone trying to breed them in captivity. A particularly interesting study from 2009, supporting the Trivers-Willard hypothesis, showed that Tasmanian devils infected with a nasty face-eating disease had more daughters than those that were healthy.
Most mammal species do seem to produce roughly equal numbers of sons and daughters, as humans do. But with those that don't, being able to understand how they manage it could have important implications for preserving endangered populations.
[Picture from Wikimedia Commons]