Fish logic surprises researchers
Jan 26,2007 00:00 by Stanford University and World

A male fish can size up po­ten­tial ri­vals, rank­ing them from strongest to weak­est, just by watch­ing how they per­form in fights with oth­er ma­les, ac­cord­ing to a new stu­dy.

The re­search­ers say their finding pro­vides the first di­rect ev­i­dence that fish, like peo­ple, can use log­ic to learn their place in a peck­ing or­der. The stu­dy, pub­lished in the Jan. 25 edi­tion of the re­search jour­nal Na­ture, in­volved ci­ch­lids (SIK-lids), small ter­ri­to­rial fish from Af­ri­ca. 

A dominant male A. bur­toni cichlid. (Credit: Logan Grosenick)
“Male ci­ch­lids are con­stant­ly try­ing to as­cend so­cially by beat­ing each oth­er up,” said study co-author Rus­sell D. Fer­nald of Stan­ford Uni­ver­si­ty in Stan­ford, Ca­lif. “It would be real­ly val­u­a­ble for them to know in ad­vance who to pick a fight with.”

The sci­en­tists aimed to find out wheth­er ter­ri­to­rial fish use a basic type of rea­son­ing, tran­si­tive in­fer­ence, in which known re­la­tion­ships are used to fi­g­ure out un­fa­mil­iar ones. 

“It’s some­thing that kids gen­er­al­ly fig­ure out by age four or five: Mary is taller than Fred, Fred is taller than Pete, there­fore Mary is taller than Pete. It’s been dem­on­strat­ed in pri­ma­tes, rats and some bird spe­cies, but how and why it evolved in an­i­mals is a mat­ter of de­bate.” 

The researchers used a pop­u­lar lab­o­ra­to­ry fish called As­ta­toti­lapia bur­toni. The ma­les, ex­treme­ly ter­ri­to­rial, reg­u­lar­ly en­ter ag­gres­sive jousts whose out­come de­ter­mines ac­cess to food and mates. Males that re­peat­ed­ly lose, can’t hold ter­ri­to­ries and thus drop in sta­tus, the au­thors wrote.

When they fight, it’s easy to spot the win­ner. Ma­ture ma­les have a men­ac­ing black stripe, or eye­bar, on their face. Af­ter a bout, the win­ner re­tains his showy ap­pear­ance, but the los­er’s eye­bar tem­po­rar­i­ly fades away as he tries to flee.

The re­search­ers staged a se­ries of one-on-one com­bats be­tween ma­les of equal size. Fish that lost their eye­bar were de­clared the los­er, sep­a­rat­ed from their op­po­nent and put back in their orig­i­nal tank. With­in min­utes, the los­er’s eye­bar re­turned, and he looked like all the oth­er dom­i­nant ma­les again. 

The fights were staged in a tank di­vid­ed in­to com­part­ments. A cu­bi­cle in the cen­ter con­tained lone male “by­stander.” Around him were five smaller com­part­ments, each with one male ri­val iden­ti­fied as A, B, C, D and E. Re­search­ers made sure that the by­stand­er had nev­er met any of these po­ten­tial ri­vals.

The by­stand­er was al­lowed to watch a se­ries of fights be­tween ri­val pairs: A vs. B, B vs. C, C vs. D, and D vs. E. Re­search­ers ma­nip­u­lat­ed the fights so that A would dom­i­nate B, B would dom­i­nate C, and so on. 

Tak­en to­geth­er, the fights im­ply a dom­i­nance hi­er­ar­chy with A on top, fol­lowed by B, C, D and E in that or­der. Did the by­stand­er grasp this peck­ing or­der? And could he use that knowl­edge to make log­ical de­ci­sions about the same fish paired in new re­la­tion­ships?

To find out, eight dif­fer­ent by­stand­ers were tested in the fa­mil­iar square tank and in a new set­ting: a rec­tan­gu­lar aquar­i­um with three ad­ja­cent com­part­ments. In each test, a by­stand­er was placed in the mid­dle com­part­ment be­tween two sets of ri­vals that he had nev­er seen to­geth­er—A and E on the one hand, and B and D on the oth­er. All ri­vals had re­cov­ered from ear­li­er losses, so their phys­i­cal ap­pear­ance was sim­i­lar. 

Us­ing a vid­e­o cam­era, re­search­ers recorded which ri­val the by­stand­er ap­proached first, and the over­all time he spent next to each of them. Pre­vi­ous tests had shown that by­stand­ers pre­fer to spend more time near the ri­val they per­ceive as weaker, the au­thors ex­plained.

The re­sults were dra­mat­ic, ac­cord­ing to the re­search­ers: vir­tu­al­ly all by­stand­ers swam to the weaker ri­val first and stayed near him for sig­nif­i­cantly long­er. In the A-E tests, by­stand­ers pre­ferred E, the wimp­i­est fish. In the more sub­tle B-D tests, most by­stand­ers chose D. So “fish do, in fact, use tran­si­tive in­fer­ence to fig­ure out where they rank,” Fer­nald said. “I was amazed that they could do this through vi­car­i­ous ex­pe­ri­ence, just by watch­ing oth­er ma­les fight.”

In Lake Tangan­yika in east­ern Af­ri­ca, the ci­ch­lids’ hab­i­tat, con­di­tions change con­stant­ly and “it would be ad­van­ta­geous for a male to know who the new boss is go­ing to be and who his weak­est ri­vals are,” he added. “Our ex­per­i­ment shows that male ci­ch­lids can ac­tu­al­ly fig­ure out their odds of suc­cess by ob­ser­va­tion alone. From an ev­o­lu­tion­ary stand­point, tran­si­tive in­fer­ence saves them val­u­a­ble time and en­er­gy.”

Fish might have the ru­di­men­ta­ry brain cir­cuit­ry for tran­si­tive in­fer­ence that ap­peared lat­er in birds and mam­mals, he con­tin­ued. “Any an­i­mal that has evolved a so­cial sys­tem that re­quires com­bat among ma­les will have some kind of eaves­drop­ping ca­pa­bil­i­ty al­low­ing them to sur­rep­ti­tiously draw in­fer­ences about their so­cial rank,” Fer­nald said. “Ca­pac­i­ties that evolved in fish may con­trib­ute to hu­man tran­si­tive in­fer­ence, or per­haps this ca­pac­i­ty evolved in­de­pend­ent­ly. The ques­tion re­mains un­re­solved.”