Yet McGregors team has shown thatSiamese fighting fish (Betta splendens), pic-tured above, possess considerable socialnous. Males of this famously aggressivespecies defend their territories with displaysof fin-waving and gill-raising. But if thisdoesnt settle matters, things turn physical sometimes fatally so.
Like human boxers, Siamese fighting fishstudy their opponentsprevious bouts.Malespay more attention to their neighbours whenthey fight than at other times, McGregor andhis colleagues found. And after viewing suchcontests, males approach the winners morewarily than they do the losers, relying moreon visual displays and less on biting2.
Verbal abuseSome researchers have questioned whethersuch experiments prove that bystandersscrutinize the interaction between oppo-nents they might be responding to theanimals inherent toughness or weediness.McGregors team tackled this issue in greattits (Parus major) by using recordings of thebirds songs. To a male great tit, victory is aquestion of timing. A male threatens a rivalby singing over his song, and shows defer-ence by singing only in the gaps between theothers choruses. This allowed the researchersto use the same songs, regardless of anyintrinsic property they might have, to denoteattack or defence, belligerence or tact.
Setting up two loudspeakers outside amales territory, the researchers played out
Nosiness isnt nice. But in the past fewyears, behavioural biologists haveshown the trait in a more positiveand intriguing light. Animals from fish tosongbirds, they have found, can achieve suc-cess by keeping watch on their neighbourssocial lives. Such eavesdropping may also bewoven into the fabric of human societies and might even help to explain why peopleoften behave charitably.
Prying animals reap significant rewards.They know when to pick a fight and when to back down; who to mate with, and who to cuckold. Not surprisingly, perhaps,researchers have also found that animalsbehave differently depending on who iswatching or listening. Animal communica-tion, experts are coming to realize, hasevolved to fit into a social network, ratherthan being a collection of signals intendedsimply to impress a particular mate or rival1.
Eavesdropping shows how incrediblysubtle animal strategies are,says evolutionarybiologist Lee Dugatkin of the University ofLouisville in Kentucky. This subtlety explainswhy it went unnoticed until recently itstricky to design experiments to tease out theeffects on one animal of watching other ani-mals interact. Peter McGregor, a behaviouralecologist at the University of Copenhagen inDenmark, suggests that researchers may alsohave neglected such experiments becausethey underestimated animals cunning. Formost people, fish dont rate when it comes tocognitive abilities,he observes.
duets of differing structures and outcomes.They then moved either the winning or losing speaker into the birds territory, andnoted his reaction. Males sang less to losers3,perhaps because they regard them as less ofa threat, or perhaps because they are moreready to escalate contests with losers to visualdisplays or violence. A winner got the samecautious treatment as a stranger.
Playback experiments with nightingales(Luscinia megarhynchos) yield similar results except that males intensify their singingtowards winners, rather than giving losers thesilent treatment4,5. Again, it is hard to knowwhether the differences reflect a more, or less,aggressive response. The interpretation cango into hand-waving, McGregor admits. Butin each case,it is clear that eavesdropping influ-ences the animalssubsequent behaviour.
Watching a fight also changes physiology.Cichlid fish (Oreochromis mossambicus) that
see a contest experience arush of testosterone, per-haps priming them to fight6.Dugatkin believes that the next challenge is to integrate behavioural andphysiological data on eaves-dropping. There are veryfew studies looking at thephysiology and behaviourof one system, he says. Ithink that synthesis is goingto happen soon.
Males are not the only
242 | wwwNATURE | VOL 419 | 19 SEPTEMBER 2002 | www.nature.com/nature
Nosiness altersphysiology, saysLee Dugatkin.
By snooping on one anothers social lives, animalscan work out how to behave when they meet inthe future. John Whitfield listens in on the naturalworlds eavesdroppers. R.F
2002 Nature Publishing Group
ioural ecologists call extra-pair copulations.Experiments on a closely related species
lend support to this idea. Daniel Mennill ofQueens University in Kingston, Ontario,and his colleagues picked playback fightswith male black-capped chickadees (Poecileatricapillus), and then analysed the DNA of the chicks born to their mates. Theresearchers found that the female partners ofdefeated males were about five times morelikely to lay eggs fertilized by other males,compared with females who never heardtheir partner get beaten8.
Covert struggleWith such high stakes, it is likely that eaves-droppers have shaped the evolution ofanimal communication. Some behavioursseem adapted to avoid prying ears. In manysongbirds, says Mennill, the longest, mostevenly matched song duels are the quietest.Where both males are struggling to domi-nate, he suggests, they might not want tobroadcast whats going on.
The effect of an audience on animalssocial interactions is harder to study thaneavesdropping, and this work is at an earlystage. Again working with fighting fish,McGregors team has found that males displayto each other differently when a female iswatching9. They reduce their aggression, andswitch to conspicuous displays incorporatingsome of the elements used in courting,such astail waving. And in July, Michael Kidd of theUniversity of New Hampshire in Durhamtold the Animal Behavior Societys annualmeeting at Indiana University in Blooming-ton that defeated male fighting fish prefer tocourt females that didnt witness their humili-ation.They have a fairly strong preference forfemales that didnt see them lose,says Kidd.
Eavesdropping is thought to help animalsto avoid fights they cannot win.But paradox-ically, eavesdroppers might make contestsmore aggressive, according to evolutionarybiologist Rufus Johnstone of the Universityof Cambridge, UK. He used game theory toanalyse the costs and benefits of winning and
losing fights, and of backing down quicklyversus a prolonged tussle. Eavesdroppers, hefound, increase the value of victory: an ani-mal that wins its current contest will get thedeterrent benefit of a tough-guy reputation,and so is more likely to escalate a fight10.Eavesdropping can evolve to reduce the riskof fighting,but once it becomes established itpromotes aggression,says Johnstone.
Replace acts of violence with ones of chari-ty, and Johnstones model becomes similar tothose used to explain apparently selfless kind-ness. We often help people we are unlikely tomeet again. One reason might be that gooddeeds get their perpetrator a glowing reputa-tion that helps them in the future. Theoreticalmodels suggest that altruism can survive inpopulations where individuals trust those theyhave seen cooperate with others,but give noth-ing to those they have seen behave selfishly11.
Research by Manfred Milinski, a behav-ioural ecologist at the Max Planck Institute forLimnology in Pln, Germany, and his col-leagues supports this idea. In one experiment,volunteers were given money and told theycould donate some of it to the other partici-pants over a series of rounds. This benefitedthe recipients more than the donors, becausethe experimenters supplemented each dona-tion. Even though participants could notdonate to someone who had given to them,they were more generous towards those whothey had seen give to others12. In anothergame, Milinski found that people were morelikely to contribute to a public fund if theirenhanced reputation could be used to attractprivate donations from other players13.
The behavioural science of eavesdrop-ping might soon be tested in the humansocial marketplace. Milinskis research hasattracted the attention of managers trying tocontrol demands on Germanys health ser-vice. He suggests that doctors could publishlists of how many treatments they have pre-scribed and how much each has cost. Evenwithout naming names, Milinski argues,people might be so concerned about gaininga bad reputation that they will be shamed outof seeking needless medical attention. Ifpeoples reputation is at stake they are muchmore cooperative,he says. nJohn Whitfield works in Natures news syndication team.1. McGregor, P. K. & Peake, T. M. Acta Ethol. 2, 7181 (2000).
2. Oliveira, R. F., McGregor, P. K. & Latruffe, C. Proc. R. Soc. Lond. B
265, 10451049 (1998).
3. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T.
Proc. R. Soc. Lond. B 268, 11831187 (2001).
4. Naguib, M. & Todt, D. Anim. Behav. 54, 15351543 (1997).
5. Naguib, M., Fichtel, C. & Todt, D. Proc. R. Soc. Lond. B 266,
6. Oliveira, R. F., Lopes, M., Carneiro, L. A. & Canrio, A. V. M.
Nature 409, 475 (2001).
7. Otter, K. et al. Proc. R. Soc. Lond. B 266, 13051309 (1999).
8. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. Science 296, 873 (2002).
9. Doutrelant, C., McGregor, P. K. & Oliveira, R. F. Behav. Ecol. 12,
10. Johnstone, R. A. Proc. Natl Acad. Sci. USA 98, 91779180 (2001).
11.Nowak, M. A. & Sigmund, K. Nature 393, 573577 (1998).
12.Wedekind, C. & Milinski, M. Science 288, 850852 (2000).
13.Milinski, M., Semmann, D. & Krambeck, H.-J. Nature 415,
ones noting the results of their neighbourssquabbles females use the same informa-tion to help them to choose their mates.Againusing song playback, McGregors team esca-lated contests with some male great tits, whilebacking down against those on neighbouringterritories. Subsequently, the mates ofdefeated males were more likely to visit theadjacent territory7. Seemingly disenchantedwith their partner but impressed by whatthey heard coming from next door thefemales were presumably seeking what behav-
NATURE | VOL 419 | 19 SEPTEMBER 2002 | www.nature.com/nature 243
Good guys win: Manfred Milinski has shown that we help those with a charitable reputation.
Daniel Mennill found that female black-cappedchickadees will cheat on mates that lose fights.
2002 Nature Publishing Group