Sexual selection and mate choiceIntroductionEvolution of mate choiceTesting the evolution of trait and preferenceBottom-up genetic testingMate choice and sexual conflictMate choice or paternity choice?Interacting pre- and postcopulatory choicePostcopulatory sexual conflictConclusionsAcknowledgementsReferencesSexual selection and mate choiceMalte Andersson1and Leigh W. Simmons21Department of Zoology, University of Gothenburg, SE 405 30 Gothenburg, Sweden2Centre for Evolutionary Biology, School of Animal Biology (M092), The University of Western Australia, Crawley 6009, WA,AustraliaThe past two decades have seen extensive growth ofsexual selection research. Theoretical and empiricalwork has clarified many components of pre- andpostcopulatory sexual selection, such as aggressivecompetition, mate choice, sperm utilization and sexualconflict. Genetic mechanisms of mate choice evolutionhave been l ess ame nable to empirical testing, butmolecular genetic analyses can now be used for incisiveexperimentation. Here, we highlight some of thecurrently debated area s in pre- and postcopulatorysexual selection. We identify where new techniquescan help estimate the relative roles of the variousselection mechanisms that might work together in theevolution of mating preferences and attractive traits,and in sperm–egg interactions.IntroductionTwenty years ago, when sexual selection was reviewed inthe first issue of TREE [1], it was a rapidly growing field.Darwin’s idea of female preferences for male ornamentswas still controversial, although his theory had receivednew support from two directions. First, empirical studiesshowed that male ornaments are favoured by femalechoice in some fishes and birds [1]. Second, a majorproblem left open by Darwin, the reasons why femalesprefer ornamented males, was clarified when geneticmodels [2,3] verified the logical coherence of Fisherianself-reinforcing coevolution of male ornaments and femalepreferences (Box 1).These results posed as many interesting questions asthey answered, and inspired much new research. In thetwo decades since 1986, sexual selection theory has beencorroborated [4] and enriched with exciting new ideas anddiscoveries, some of which we highlight here. We alsopoint to new possibilities for testing genetic mechanismsof sexual selection in the era of functional genomics.Genetic analyses of sexual selection by mate choice haveworked so far on a mainly top-down basis, inferringgenetic causes from phenotypic patterns, based on few-locus genetics or quantitative genetics theory [5]. Theseremain excellent tools for the analysis of preference–display coevolution and for other purposes in sexualselection research. However, quantitative trait locus(QTL) id entification and sequencing combined withfunctional genomic s now provide the opportunity forbottom-up approaches, based on th e precisecharacterization of genes and their effects, from DNAsequences via protein to phenotypic expression at the levelof t he indiv idual, with possible consequences at thepopulation level and above.Evolution of mate choiceAlthough mate choice occurs in males and females [4], forconvenience we refer here to female choice of male traits.As experimental evidence accumulated, mate choicebecame widely recognized, but the genetic mechanismsunderlying its evolution remain the subject of debate(Box 1). Showing how mating preferences evolve geneti-cally is harder than showing that they exist, and theproblem is aggravated by the possibility that severalmechanisms co-occur ( Box 1). Moreover, conflicts betweenthe sexes can add further selection pressures on pre-ference and the preferred trait [6,7]. Opinions differ overthe relationships between Fisherian and genetic indicatormechanisms of mate choice (e.g. [5,8–13]). Given thatthere are qualitative differences between them, we thinkthere are good reaso ns to keep the distin ction clear[9,11,12].Costs of mate choice, such as the loss of energy andtime, might prevent Fisherian self-reinforcing coevolutionof the trait and the preference for it, but need not do so; theoutcomes of models depend on the details of severalassumptions, for instance about sex linkage [10] (reviewedin [5]). In spite of being theoretically pl ausible, theFisherian genetic mechanism has been difficult todemonstrate empirically. There is corroborating pheno-typic evidence (e.g. [14,15]), and this is also the case forindicator mechanisms (e.g. reviewed in [5,11,16]). Theseand other mechanisms (Box 1) are compatible and mightco-occur (e.g. [4,12]), and thus a challenging task is togauge their relative roles. Estimating the effects of theFisherian sexy son mechanism might help us to decidewhether it explains why male ornaments and displays areoften extrem e, and differ more than do other traitsbetween closely related species, apparently evolvingrapidly and perhaps being involved in speciation[2,17,18]. These questions have been with us since CharlesDarwin’s time and have yet to be answered satisfactorily.Testing the evolution of trait and preferenceMolecular genetic and genomic tools enable the detailedcharacterization of genes and their effects (e.g. reviewed in[19–21]). Combined with selection experiments thatquantify genetic evolution over generations, they offertools for detailed genetic analyses and tests of whetherCorresponding author: Andersson, M. ([email protected]).Available online 3 April 2006Review TRENDS in Ecology and Evolution Vol.21 No.6 June 2006www.sciencedirect.com 0169-5347/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tree.2006.03.015mating preferences and preferred traits evolve as pre-dicted by sensory bias, Fisherian, indicator, orother mechanisms.There is evidence for sensory bias in an increasingnumber of species, where phylogenetic analyses have shownthat femalepreference for a male display trait is also presentin other related species lacking the male trait (e.g. [22–24]).In such cases, it seems likely that the female preferencearose first for other reasons, such as attraction tofood (whichneeds to be tested [12]), hereafter the male trait evolved,favoured by the pre-existing female sensory bias [22].The Fisherian mechanism is, in essence, based only ongenes that influence mating preferences and preferredtraits, and it can lead to their rapid coevolution if the mostpreferred males are polygynous (e.g. [2,5,10,17]). Geneticindicator mechanisms, however, depend on overall qualityand