The threat of infants being killed by unrelated males is the key driver of monogamy in humans and other primates.
The study, by academics from UCL, Oxford University, the University of Manchester and the University of Auckland, is the first to show that this evolutionary pathway explains the emergence of pair living.
The team also found that following the emergence of monogamy, males are more likely to care for their offspring. Where fathers care for young, not only can they protect infants from other males, but they can also share the burden of childcare.
Lead author Dr Kit Opie of UCL said: 'This is the first time that the theories for the evolution of monogamy have been systematically tested, conclusively showing that infanticide is the driver of monogamy. This brings to a close the long running debate about the origin of monogamy in primates.'
Until now, a number of hypotheses have been proposed to explain the evolution of monogamy among mammals. These include the benefit of paternal care when the cost of raising offspring is high; the ability to guard solitary females from rival males; and protection against the risk of infanticide from rival males.
To uncover the evolutionary pathway that led to monogamy, the team gathered data across 230 primate species. These were then plotted on a family tree of the evolutionary relationships between those species.
Computational modelling methods were used to re-run evolution millions of times across the family tree to discover whether different behaviours evolved together across time and, if so, which behaviour evolved first.
This 'Bayesian' approach allowed the team to determine that male infanticide is the cause of the switch from a multi-male mating system to monogamy in primates. They showed that two-parent care and solitary ranging by females are a result of monogamy, not the cause.
Professor Robin Dunbar of the Department of Experimental Psychology at Oxford University explained: 'To decide between the suggested causes for monogamy, we used the same Bayesian methods that Nate Silver used to predict Obama's win in the last US presidential election. But instead of predicting the future, we used the methods to predict the past, finding that it was infanticide risk that provided the strongest evolutionary incentives for pairing up.'
The findings are published in the journal PNAS.
Infants are most vulnerable when they are fully dependent on their mother, because females delay further conception while nursing slowly developing young. This leads to the threat from unrelated males, who can bring the next conception forward by killing the infant.
Sharing the costs of raising young between both parents shortens the period of infant dependency and can allow females to reproduce more quickly.
An additional benefit of sharing the burden of care is that females can then have more 'costly' young.
Many primate species have large brains that have resulted from the considerable cognitive requirements of living in complex societies. Growing a big brain is expensive and requires that offspring mature slowly. Caring fathers can help alleviate the burden of looking after young with long childhoods and may explain how large brains could evolve in humans.
Humans, uniquely among primates, have both very long childhoods and mothers that can reproduce quickly relative to other great apes.
Dr Susanne Shultz from the University of Manchester said: 'What makes this study so exciting is that it allows us to peer back into our evolutionary past to understand the factors that were important in making us human. Once fathers decide to stick around and care for young, mothers can then change their reproductive decisions and have more, brainy offspring.'