2Department Ofconservationbiology, Estacio Nbiolo Gicadedon Anacsic,Sevilla,Spain

Update

Whyistherevariationinbaselineglucocorticoidlevels?

NielsJ.Dingemanse1, PimEdelaar2 andBartKempenaers1

1Department ofBehavioural EcologyEvolutionaryGenetics,MaxPlanckInstitute forOrnithology,Seewiesen,Germany

2Department ofConservationBiology, Estacio´nBiolo´gicadeDon˜anaCSIC,Sevilla,Spain

Inarecent articleinTREE,Bonierand colleagues[1] reviewed the supportfor the Cort-Fitness Hypothesis, which states that baselineglucocorticoid(cort) levels are elevatedin individuals,or populations,that experience challengingconditions,thereforesignallinglowfuturefit- ness. The idea isthatthe optimallevel ofresourcesallo- catedtowardsself-maintenance(immediate survival) versus long-term survival and/or reproduction differs acrossenvironments,with selectionfavouringindividuals investing inself-maintenance when the environment becomes‘challenging.’The hypothesisstatesthatthisre- allocationofresourcescanbeachievedbyalteringbaseline levels ofcort. Theirliteraturereview did not supportthe hypothesis,asitrevealedmuchvariationinsignandshape oftherelationshipbetweencortandresidualfitnesswithin and acrossspecies,populations,and life-historystages[1]. We hereproposethat the inabilityto find unequivocal supportfor the Cort-FitnessHypothesisis a side effect ofBonieretal.’simplicitassumptionthatindividualvar- iance componentscan beignoredwhen testingevolution- ary hypotheses, a phenomenon recently coined the

‘tyrannyofthe Golden Mean’[2].

Bonieretal.assumethatanimalsreacttoenvironmen- tal change by altering their phenotype,and that this responseisadaptive.Selection would have favoureda particular relationshipbetweencort and environmental quality. Such a relationship can be viewed as a line (Figure1a)characterisedbyacertainelevationand slope, often called a‘reactionnorm’[3].Ifallindividualshad the samereactionnorm,the Cort-FitnessHypothesiscouldbe testedusinga cross-sectionalapproach,whereresidual fitnessiscompared acrossindividualsshowinglowversus high levels ofcort [1],the approachtakenin the review.

Baselinecortdoes,however,vary consistentlyacrossindi- viduals, with certainindividualsshowinghigh and others showinglow levels ofcort underany type ofcondition (Figure1b). The slope ofthe relationship betweencort and environmentalquality(i.e.plasticity)can also differ between individuals(Figure1c)and both types ofindivid- ual variationcan beheritable(see[4,5]).

Thepresenceofindividualvariationinaveragelevel(I), andplasticity(IxE),inbaselinecorthasconsequencesfor howthe Cort-FitnessHypothesisshouldbetested.When I or IxE are present,high cort levels donot necessarily signal ‘challenging’ conditions.Imagine a situation (Figure1b)in which two typesofindividualsexist, both showingadaptiveincreasesincortwhen the environment becomeschallenging,butonetypealwayshashigherlevels ofcort.Both typescouldhave exactly the samephenotype (dottedlineinFigure 1b),becauseoneexperiencesthe environmentaschallengingwhereasthe otherdoesnot.A valid testofthe hypothesiswould thereforerequirethe measurement ofindividualreactionnorms(see[5]), expecting anegativeassociationbetween cortleveland residualfitnesswithinindividuals.

Howshouldwetheninterpretthe seemingly‘inconsist- ent’Cort-Fitnessrelationship?Wesuggestthatitiscaused byheterogeneousselectionactingonthe between-individ- ual componentofcort. Recentstudieson natural popu- lationsshow thattemporaland spatialvariation existsin selectionactingonbehaviours(e.g.[6])thataregenetically correlatedwith baselinecort [7].Individualvariationin cortmightthusrepresentfundamentaldifferencesinhow individualscopewith environmental challenges[7],imply- ingthattheremightnotbeasingleoptimalresponse[8,9]. We encourage endocrinologists to try to disentangle

Figure1.Hypotheticalrelationships betweenbaselinelevelsofcortandenvironmentalquality. Eachlinerepresentsatypeofindividualpresentwithinthepopulation. Two typesofindividuals(iandj)arepresentinpopulation(b)and(c),asingletypeinpopulation(a).Fordetailsseetext.

Correspondingauthor:Dingemanse,N.J. ().

within-versusbetween-individualvariationincortlevels and toevaluatethisvariationwithin an evolutionary framework.

References

1 Bonier, F. et al. (2009) Do baseline glucocorticoidspredict fitness?

TrendsEcol.Evol.24,634–642

2 Williams,T.D.(2008)Individualvariationinendocrinesystems:moving beyond the‘tyrannyofthe Golden Mean’.Philos.Trans.R.Soc.Lond. B Biol.Sci.363, 1687–1698

3 Sarkar,S.(1999) Fromthe Reaktionsnormtothe adaptivenorm: the norm ofreaction,1909–1960.Biol.Philos.14,235–252

4 Nussey, D.H. et al. (2007) The evolutionary ecology of individual phenotypicplasticityinwild populations.J.Evol. Biol.20,831–844

5 Dingemanse,N.J., Kazem,A.J.N.,Re´ale, D. and Wright,J. (2010) Behavioural reaction norms: where animal personality meets individualplasticity.TrendsEcol.Evol. 25,81–89

6 Re´ale, D. et al. (2007) Integrating temperament in ecology and evolutionarybiology. Biol.Rev.Camb.Philos.Soc.82,291–318

7 Koolhaas,J.M. etal.(1999)Coping stylesinanimals:currentstatusin behaviorand stress-physiology.Neurosci.Biobehav.Rev.23,925–935

8 Wolf, M. et al. (2007) Life-historytrade-offsfavourthe evolutionof animalpersonalities.Nature447,581–585

9 Wolf, M. et al. (2008) Evolutionary emergence of responsiveand

unresponsive personalities. Proc. Natl. Acad. Sci. U. S. A. 105,

15825–15830

Letters Response

Clarifying theCort-Fitness Hypothesis:

aresponsetoDingemanse etal.

FrancesBonier1, PaulR.Martin2,IgnacioT.Moore3andJohnC.Wingfield4

1Queen’sUniversity,Biology Department, Kingston, Ontario K7L3N6,Canada,andVirginia Tech,Department ofBiological

Sciences,Blacksburg, VA24061,USA

2Queen’sUniversity,Biology Department, Kingston, Ontario K7L3N6,Canada

3Virginia Tech,Department ofBiologicalSciences,Blacksburg, VA24061,USA

4UniversityofCalifornia, Davis,Department ofNeurobiology,Physiology,andBehavior, Davis,CA95616,USA

Dingemanseet al. [1] suggest thatthe Cort-Fitness Hypothesisrestsupon the idea that‘theoptimallevel of resourcesallocated towardsself-maintenance(immediate survival) versuslong-term survivaland/orreproduction differs acrossenvironments,with selectionfavoringindi- vidualsinvestingin self-maintenancewhen the environ- mentbecomes ‘‘challenging’’’.This interpretationis incorrect.Instead,the Cort-FitnessHypothesisrestsupon the idea thatdifficult orchallengingconditionslead tolow fitness(i.e.lowfuturesurvivaland/orlowfuturereproduc- tive success),and that glucocorticoid(cort) levels can be used asaproxyforthe degree ofenvironmentalchallenge facing individuals and populations [2].Indeed, thispro- posed connectionbetweenhigh cort,difficult environmen- tal conditionsand lowfitness isthe mainreasonwhy conservationbiologistsuse cort as aproxy forthe health ofindividualsand populations.

The Cort-FitnessHypothesis isnot an evolutionary hypothesisand isnot aimed atmeasuringselection. Instead,the Cort-FitnessHypothesisisasimplepattern hypothesiswherebycortcanbeusedasaproxyforenviron- mentalchallengesfacing individualsorpopulations.Ulti- mately,the proposedcorrelationsbetweenlevels ofcort, environmentalchallengesand fitness,ifthey exist, could resultfrom one ormore evolutionaryselectivepressures actingon cort secretion.However,the Cort-Fitness Hy- pothesisisnotahypothesistoexplainwhycortlevelsvary withfitnessortheenvironment;itissimplyaddressingthe common usageofcortasaproxy forthe healthofindivid- ualsand populationsinthe scientificliterature.

Correspondingauthor:Bonier,F. ().

Wedonot makethe assumptionthatindividualvari- ation canbeignoredwhen examiningthe Cort-Fitness Hypothesis.Among-individualvariationincort secretion, varying reactionnormsand/or coping strategiesmight obscurerelationships betweenlevels ofcort and fitness, and we discussall ofthesepossibilitiesin our original paper(e.g.‘Geneticanddevelopmentaleffects’in[2],p.638). The inconsistentrelationships betweencort and fitness (includingpositive,negativeand norelationships)donot resultsolelyfromindividualvariationinthe cortresponse acrossenvironmentsbecausesomestudies have found individualswith the highestfitnessexhibitingboth low and high cortlevels relative tolowfitnessindividuals, dependingonthe stageofbreeding[e.g.3,4].

Overall,our conclusionremains unchanged:levels of cort cannotbe used as a proxy for environmental chal- lengesorfitnesswithoutfirst understanding the relation- ship betweencort,the environment and fitnessinthe organismofinterest.Weagreewholeheartedlythateval- uatingsources ofwithinand amongindividualvariation withinan evolutionaryframeworkcan provideimportant advancesinourunderstanding ofthesecort-fitness relationships,and weare presentlytakingthisapproach inour ownwork [e.g.4–6].

References

1Dingemanse,N.J. etal. (2010)Whyistherevariation inbaseline glucocorticoidlevels? TrendsEcol.Evol. 25,261–262

2 Bonier,F. et al. (2009) Do baseline glucocorticoidspredict fitness?

TrendsEcol.Evol. 24,634–642

3 Magee, S.E. etal. (2006) Plasmalevels ofandrogensand cortisol in relationtobreedingbehaviorinparentalmale bluegillsunfish,Lepomis macrochirus.Horm.Behav. 49,598–609