Perfect match: tied contest gives fish no hormone rush.

A male fish produces a burst of hormones as he fights off an intruder. Now, researchers say that this surge isn't triggered simply by fighting.

"Apparently, the fish need to know whether they are winning," says Rui Oliveira of the Institute of Applied Psychology in Lisbon, Portugal.

Oliveira and his colleagues took this view after exposing each fish to a perfectly matched opponent: his image in a mirror. When a male cichlid sees his reflection, he displays increasingly menacing postures, as if he's facing off with a live intruder. However, he doesn't show a surge of male hormones, the researchers report in the Sept. 8 Nature.

Fish such as these cichlids (Oreochromis mossambicus) are valuable for studying the ways in which hormonal reactions change social behavior, says Oliveira. The male hormones, or androgens, that his lab tracks show up in similar forms in other animals, including mammals and birds.

Decades of various researchers' studies have shown that among fish, birds, and mammals, winners' androgen concentrations surge and losers' dip. These hormonal changes improve or diminish, respectively, an animal's chance of winning his next fight.

In previous work, Oliveira and his colleagues found that just watching a fight boosted androgens in spectator fish. The finding echoed a study of androgens in sports fans watching soccer or basketball. Researchers from Utah and Georgia found that spectators' hormone concentrations rose or dipped, depending on whether they had rooted for a winning or losing team.

To search for more-precise triggers for hormone responses, Oliveira says, he first asked whether the fish had to analyze information about their situation or used some simpler, more direct reaction to fighting.

He and his colleagues exposed some male fish to mirrored surfaces for 20 minutes and other fish to nonreflective glass. The fish looking at the mirrors acted aggressively toward their reflections. After the encounters, the researchers analyzed androgens in fish-urine samples taken over the next 6 hours. They found no difference between the two groups.

Both groups showed a small drop in hormone concentration typical for the morning hours, when the team conducted the tests. So, Oliveira says, that dip doesn't reflect the reduction in hormone production seen in fish that lose fights.

Ethan Clotfelter of Amherst (Mass.) College comments that he would now like to see experiments comparing fish in matches of similar and dissimilar opponents.

In his work exploring aggression, he, too, has fish fight their reflections. He says that Oliveira's results have convinced him that this common research practice is inappropriate for certain kinds of experiments. "At the very least, this will make a lot of people who study fish say, 'Wait a minute,'" he says.

John Wingfield of the University of Washington in Seattle, who developed a theoretical framework for hormonal influence on social behavior, says that it seems plausible that animals don't respond hormonally unless they can predict a fight's outcome.

"In hindsight, it seems fairly obvious," Wingfield says. "But Rui has shown it elegantly."