Australian scientists are asking people to rate different body shapes in an online survey in the hope of understanding how the rules of attraction have helped shape the human race.
The online study has drawn almost 50,000 participants worldwide in its first four months and is already offering some key insights into what, in particular, makes an attractive woman.
"We've had a few androgynous-type shapes and they haven't lasted very long," lead researcher Rob Brooks told AFP. "You need to have a certain amount of curvaceousness, a certain amount of body fat."
The extremely overweight and underweight were consistently rejected, he said, adding that proportion was everything.
"You can't simply say a narrow waist is more attractive, or a large bust or large biceps are more attractive, it really depends whether those things fit with the rest of the structure of the body," he said.
The Sydney-based study asks participants to rate the attractiveness of 120 bodies ranging from very thin to very overweight, and the most popular half are then "mated" using computer simulations to tweak their appearance.
"What the people are doing by rating the bodies is they are acting the equivalent to natural selection," said Brooks.
Heterosexual men have most enthusiastically embraced the survey, meaning the research was now up to its fourth generation of women, and only second of men.
The grey computer bodies have been stripped of hair, clothes or any other cultural or other identifiers.
Participants in the survey have to volunteer a range of information about themselves, including their nationality, ethnic origin and sexual orientation.
Brooks hopes cross-referencing the data will allow him to draw conclusions about the impact of socio-economic status, sexuality and culture on ideas of attractiveness.
"There's an idea out there, for example, that in countries where a high proportion of people are not getting enough to eat, being overweight may be more attractive than it is in countries where they have escaped from hunger and where obesity's a bigger problem," he explained.
He also hopes to challenge the notion that the Western media was "constructing preferences or tastes in some kind of sinister way."
"We should be able to test that by looking at folks from countries with different media traditions and seeing if they have different tastes or not," Brooks said.