The living arrangements of parents at the time a baby is conceived may play a role in determining its sex, research suggests.
Are boys more likely in stable relationships?
A US study found parents who were married or living together before conception were slightly more likely to have a boy than those who were not.
The study, by the US National Bureau of Economic Research, is based on data from 86,436 births.
Details are published in Proceedings of The Royal Society.
Overall, the study found that 51.5% of babies born to couples living together at the time of conception were boys, compared to 49.9% among parents who were not.
Although this might seem like a small difference, it is actually statistically highly significant when considered across a whole population.
When the researchers looked at brothers and sisters, they found that couples who were living together before conception were 14% more likely to have a male child than when they were not.
The researchers say their finding could explain the fall in the proportion of male births in some developed countries over the past 30 years.
Previous research has suggested that women who are not in stable, monogamous relationships might be less likely to give birth to boys.
There are reports dating back to the 19th Century of a lower percentage of boys being born to women who were not married.
And studies in modern Kenya have found a similar trait among polygynously married women.
Male embryos are less robust than their female counterparts, and so require a greater degree of nurturing through pregnancy if they are to survive to full term.
It may be that a woman who is in a stable relationship may be in a better position to provide this care.
Researcher Dr Karen Norberg told BBC News she found the same effect in five separate US samples, representing births spanning a 40 year period of time and a great diversity of ethnic and social backgrounds.
She said: "The finding supports a biological theory that predicts that offspring sex might vary according to environmental conditions, if these conditions have different effects on sons and daughters.
"There are several possible mechanisms that could explain the effect. Factors operating at conception could include the mother's or father's hormone status, or the timing or frequency of intercourse; factors operating later in pregnancy could result in sex biases in risk of miscarriage."
Professor Andrew Reid, spokesman for the British Society for Human Genetics, told BBC News that in the general population it was thought that roughly 106 boys were born for every 100 girls.
"It is thought this is nature's way of compensating for the fact that boys are more likely to die in infancy because there are genetic diseases such as muscular dystrophy which almost entirely affect boys," he said.
Professor Reid said it was possible that the reason for the difference highlighted in the study was that people who do not live together tended to have intercourse at different times to those who do.
It is known that children who are conceived from intercourse that occurs exactly on the day of ovulation are more likely to be sons than children who are conceived from intercourse two to four days before ovulation.