For humans and other animals that traditionally have just one baby at a time, twins are a gamble. Pregnancy is riskier for the mother and the fetuses. If the twins make it to birth, they're likely to be undersized. And even if she has two healthy babies, a mother must find twice as much food as usual to keep them that way--and must keep twice as many helpless, chubby morsels away from the lions. But if both kids survive to adulthood, the parents will have doubled their genetic contribution to the next generation.
Presumably because it's an evolutionary trade-off with so many variables, the rate of twinning has evolved differently across human populations. This variation applies only to "dizygotic," or fraternal, twins. Identical (or "monozygotic") twins are more rare, and their frequency of about 0.4% stays fairly constant across populations. One embryo splitting into two seems to be a random accident. Dizygotic twinning, though, seem to result from genetic tendencies; some mothers are more likely to release two eggs simultaneously.
One potential explanation involves hormones called insulin-like growth factors (IGFs). The IGF system may affect a mother's chances of releasing multiple eggs. It's been linked to twinning in a study of cattle (another mammal that usually sticks to one calf at a time). IGF hormones have also been linked to the growth of a fetus. This led a group of researchers in the UK to ask whether mothers who give birth to twins, thanks to the effect of IGF hormones, also give birth to larger non-twin babies.
The researchers, led by Ian Rickard, used data collected from Gambian women between 1978 and 2009. These women had an average of about 7 children each. Researchers divided the women into those who gave birth to twins at some point, and those who never had twins. They found that "singleton" (non-twin) babies born to twinning mothers were significantly larger, on average, than babies born to non-twinning mothers. In other words, twinning mothers--when they're not having twins, which are usually small--give birth to heavier babies. This effect was independent of the mothers' BMI or height. It even held true for singleton babies that were born before their twin siblings: these babies were heavier at birth than singleton babies born to mothers who never had twins.
The only time the effect didn't appear was when mothers were exposed to the "hungry season," the furthest time from the harvest, during their third trimester. With more limited resources during their pregnancies, twinning mothers gave birth to babies that were roughly as small as non-twinning mothers' babies.
Though the dataset doesn't include information on IGFs or other hormones, it seems that some underlying factor causes mothers both to have twins (because they release two eggs at once) and to have large singleton babies. The researchers also don't know how many of these sets of twins were actually identical, or monozygotic (the less-common variety). But since monozygotic twinning seems to happen randomly, mothers of identical twins would presumably dilute the effect seen here. So if anything, the birthweight advantage in twinning mothers might be even greater than it seems.
Whatever is underlying it, the study shows that there's an advantage to twinning--or, at least, that twinning is a side effect of an advantageous trait that makes for sturdier singleton babies. This might help to explain why twinning has evolved to be more common in some human populations than in others. Each population has its own pressures, and must strike its own balance between factors such as birthweight and survivorship. Thanks to the intricate mathematics of natural selection, twins are never just a twofer.
Rickard, I., Prentice, A., Fulford, A., & Lummaa, V. (2011). Twinning propensity and offspring in utero growth covary in rural African women Biology Letters DOI: 10.1098/rsbl.2011.0598