A study of insects known as water striders has yielded what researchers say is the first evidence of an anatomical arms race between the sexes.
The researchers spotted evolutionary changes that made it easier for males to mate or for females to avoid unnecessary mating.
Scientists have long theorized that such evolutionary competition between the sexes takes place in many species. But that has been difficult to prove because such changes take place over many generations.
Evolutionary changes, however, are more pronounced in water striders than in birds or mammals.
Researchers studied 15 species of water striders and found in some that one sex had evolved to get the upper hand in mating, at least temporarily.
In some species, male water striders developed a flatter abdomen and longer clasping genitalia that made them more successful in overcoming resisting females.
In others, females developed body shapes that were more effective at warding off unwanted suitors.
As females became more resistant, males became more persistent, and so on and so on, and "off you go into this arms race," said Locke Rowe, a researcher at the University of Toronto and one of the study's authors.
"It's kind of an endless co-evolutionary spiral," Rowe said.
The work by Rowe and Goeran Arnqvist of the University of Uppsala in Sweden appears in Thursday's issue of the journal Nature.
After collecting water striders in North America and Europe, the researchers examined them and then observed their mating habits.
The findings support the contention that female water striders resist mating once they have fertilized their eggs. Excessive mating is thought to take a physical toll on the females.
Sergey Gavrilets, an associate professor of ecology and evolutionary biology at the University of Tennessee, said the arms-race approach has been theorized for some time.
"What's been missing, however, is a clear demonstration of effects and power of sexual conflict in natural populations," Gavrilets said. "It's actually not that easy to demonstrate the effects of natural selection because of the long time-scale involved. Arnqvist and Lowe's paper fills that gap."