thèse - Université Toulouse III
Transcription
thèse - Université Toulouse III
THÈSE En vue de l'obtention du DOCTORAT DE L’UNIVERSITÉ DE TOULOUSE Délivré par l'Université Toulouse III - Paul Sabatier Discipline ou spécialité : Ecologie Comportementale Présentée et soutenue par Jérémie CORNUAU Le 04-12-12 Titre : Signaux multiples et choix du partenaire chez le triton palmé Lissotriton helveticus JURY Thierry Lengagne (Chargé de recherche, Rapporteur) Bruno Faivre (Professeur, Rapporteur) Claude Miaud (Directeur EPHE, Examinateur) Christophe Thébaud (Professeur, Examinateur) Dirk Schmeller (Ingénieur de Recherche, Directeur de thèse) Adeline Loyau (Post Doctorante, Directeur de thèse) Ecole doctorale : SEVAB Unité de recherche : Station d'Ecologie Experimentale du CNRS à Moulis USR 2936 Directeur(s) de Thèse : Adeline Loyau, Dirk S. Schmeller Rapporteurs : Thierry Lengagne (Chargé de recherche) & Bruno Faivre (Professeur) ● Les sabots de Bethmale "Le soir de Noël, le fiancé offre à sa fiancée une paire de sabots à longues pointes, habillée de cuir et richement décorée de clous dorés dessinant un cœur. Il offre aussi une quenouille rouge et un fuseau, le tout fabriqué avec tout son amour. La tradition veut que plus la pointe des sabots est longue, plus l’amour est ardent..." Autrefois le Couserans 2012 (St Girons/Ariège/France, Photo : Audrey Tourat). Les hommes et les femmes portent des sabots de Bethmale dont la longueur reflète l’amour du bien-aimé. REMERCIEMENTS Août 2012, après une marche difficile entre brouillard et pluie, et accompagné de deux amis, j’arrive en haut de la Pic Rouge de Bassiès. Juste avant le sommet, nous passons au-dessus des intempéries accueillis par un soleil réconfortant. Je place ma pierre en haut du petit édifice construit au fur et à mesure des passages de randonneurs. Je ne sais pas si « l’après thèse » me procurera autant de plaisir que les pâtes chaudes qui m’ont accueilli au retour de ce périple. Par contre ce qui est certain c’est que cette ascension marque la fin de ma thèse et l’heure pour moi de remercier toutes les personnes qui m’ont permis de poser un premier grain de sable dans le monde scientifique. Je voudrais tout d’abord remercier Thierry Lengagne et Bruno Faivre d’avoir accepté de juger le travail de trois ans de ma vie. Je voudrais également remercier Claude Miaud et Christophe Thébaud, membres du jury, d’avoir accepté de partager les derniers moments de cette aventure. Mes remerciements vont également à l’Ecole Doctorale SEVAB et à son comité pour m’avoir accordé un contrat doctorale. Je tiens également à remercier Jean Clobert, directeur de la station d’Ecologie Expérimentale de Moulis de m’avoir accueilli au sein de son établissement. Merci de m’avoir permis de vivre la renaissance d’un laboratoire qui je l’espère continuera de grandir et accueillera longtemps des embryons de chercheurs. Merci également d’avoir toujours laissé votre porte ouverte. Mes remerciements vont maintenant à Adeline et Dirk de m’avoir choisi pour cette thèse et de m’avoir financé un tas de gadgets (étudiants compris) pour les expériences tritons. Merci également de m’avoir accordé beaucoup de temps même dans les moments difficiles. Je souhaite à ce sujet remercier la communauté féline de Cazaux de m’avoir prêté Adeline et Dirk durant cette thèse. Je n’oublierai pas les nombreuses soupes chaudes à Bassiès après un terrain glacial à la recherche de petites bêtes, et la gastronomie Cazausoise. Aussi je ne vous remercierai jamais assez d’avoir montré à Audrey qu’un barbecue et une raclette se conjuguent aussi en mode végétarien! Parce qu’on ne remercie jamais assez les personnes qui nous ont formés, je voudrai ici remercier tout ce qui au cours de ma formation m’a permis d’être ce que je suis aujourd’hui. Merci donc à tout le personnel des écoles qui m’ont accueilli. Merci entre autre de m’avoir laissé regarder les colombidés et autres bestioles par la fenêtre durant les maudites heures de classe. Je tiens également à remercier tous ceux qui depuis les premiers cours sur « les lumières de l’évolution » m’ont transmis leur savoir sur l’écologie comportementale. Merci à Karine, Jérôme, Frank et F-X. Je voudrai ici remercier tout le personnel de la Station d’Ecologie Expérimentale de Moulis que j’ai croisé durant ces trois années de thèse. À Michelle de rester si longtemps le soir pour veiller sur nous. À Sabine et les nombreuses visites de la grotte. À la dream team du Moulis Football Club et à Fabien et Alexis fervent garant du beau jeu Moulisien. Cela a été un réel plaisir de pouvoir jouer à un niveau « international ». À Michel pour le Moulis Tennis Club. À Vincent pour le Moulis Marathon Club. À Olivier et Elodie, deux brillants naturalistes qui m’ont beaucoup aidé durant cette thèse et qui ont rendu le « terrain » encore plus passionnant. Une station en construction c’est forcement beaucoup de changements de bureau. Merci à Jocelyne et Kim de m’avoir accueilli dans le bureau–secrétariat. Mes remerciements vont également à tous les habitants de la salle thésard: Hamed, Sylvain l’habitant nomade, Audrey mon soutient contre les galères administratives de fin de thèse, Julien et les discussions scientifiques et comiques. Elvire danseuse étoile sur terrain de foot, Sophie une hirondelle qui fait le printemps. Merci à tous les stagiaires qui ont participés aux programmes NEWTS. À Margaux d’avoir commencé avec moi cette aventure (merci aussi pour LUND). J’espère avoir un jour moi aussi une vue sur l’océan. À Adelaïde d’avoir toujours su voir le Panda Roux qu’il y avait en moi. Je ne peux m’empêcher de dire « Que vive la recherche à Moulis et que vive la vie… ». A Romain, tueur de cloportes, amoureux de la cagouille et pourtant si bon avec les tritons. A Géraldine, Thomas et Laurence, de m’avoir supporté et d’avoir regardé des vidéos de tritons pendant des heures. Merci également à Fanny et les autres étudiants pour le terrain à Bassiès. Merci à Philippe et Dominique pour leur accueil à Bassiès aussi bien pour nous que pour nos tritons. J’espère pouvoir revenir un jour (mais sans triton) ! J’ai ici une pensée pour tous mes amis que je n’ai pas assez vus suite à mon entrée au monastère scientifique. J’espère pouvoir corriger cela une fois que « thèse » sera passée. À Chonchon, p’tite Fred, Matthieu, Aurélia, Julien, Didie, Mylène, Nono et Gaëtan. Très bon soutien en période de thèse. À ma famille et les noëls réparateurs. À ma mère. Certainement le plus puissant des effets maternels. C’est vraiment dommage que l’évolution ne sélectionne pas un peu plus les personnes altruistes ! Tu étudie aussi des trucs dont tout le monde s’en fou... Tu te poses des questions étranges le matin du genre « Y-a-t-il une opposition entre la philosophie des lumières et la notion de hasard en génétique »… Il y a dans ta bibliothèque des livres sur l’évolution et sur le comportement des animaux… Ce n’est peut être pas un hasard si j’en suis là aujourd’hui !!! ☺ À mes deux frères, bien plus que des exemples pour moi, votre soutien m’a été précieux. À Audrey, l’amour de ma vie. Si les tritons ne sont pas monogames c’est parce que tu n’as pas d’équivalent dans leur monde. Je n’aurai jamais assez de mots pour te remercier de tout ce que tu as fait pour moi. Des expériences tritons en passant par les révisions TOIEC jusqu’au soutien pendant les périodes de stress intense, tu as toujours été présente pour moi. Cette thèse et moi te devons énormément. En biologie évolutive on dit que “Nothing in biology makes sense except in the light of evolution”. Je crois encore plus que “Nothing in my live makes sense except in the light of your eyes” Merci pour tout, I love you… PRÉAMBULE Cette thèse traite de la théorie des signaux multiples dans le cadre du choix du partenaire sexuel. La compréhension du « pourquoi » et du « comment » les individus utilisent plusieurs signaux pour choisir leurs partenaires peut apporter des réponses intéressantes dans le cadre de la conservation des espèces, de la gestion des espèces utilisées comme bien de consommation par les Hommes, et de la réponse des organismes aux changements globaux tels que les changements d’environnement, les nuisances sonores ou bien les maladies émergentes. De plus, cela permet d’aborder des questions autour de l’utilisation de plusieurs sources d’information et des mécanismes de prise de décision inhérente à toute type de communication et donc transposable à beaucoup de situations rencontrées dans l’espèce humaine. Finalement, et c’est certainement la cause qui m’anime le plus, la théorie des signaux multiples est, à l’instar des sujets d’écologie comportementale, l’occasion d’apprendre quelque chose sur les mécanismes évolutifs qui conduisent à l’élaboration des comportements. Pour traiter ces questions autour des signaux multiples, j’ai utilisé un modèle biologique tout aussi fascinant que les questions théoriques, le triton palmé, Lissotriton helveticus. Comme pour la plupart des urodèles la parade des tritons a fasciné plus d’un naturaliste. C’est pourquoi il existe un grand nombre de description des parades, plus ou moins détaillé et plus ou moins anthropomorphique. Pour la plupart des gens, les tritons sont méconnus voire insignifiants. Pourtant j’ai l’intime conviction que la « connaissance » peut changer cette vision. En effet lorsqu’ils sont mis devant le fait accompli que dans des zones humides, il existe un petit animal qui réalise des parades complexes pour séduire « sa femelle »; les gens, dans un premier lieu un brin amusés, peuvent aussi devenir plus réceptifs au problème de l’influence de l’homme sur son environnement. Cette thèse comporte 5 articles dont je suis le premier auteur, et est subdivisée en quatre grands chapitres. Le premier chapitre sert d’introduction. Il est composé d’une courte introduction sur la sélection sexuelle, puis d’un article de synthèse sur les signaux multiples dans le cadre du choix du partenaire [article 1]. Le deuxième chapitre est composé de deux articles visant à comprendre le choix des femelles sur plusieurs signaux [article 2 et 3]. Dans le troisième chapitre, j’ai testé la condition dépendance de signaux multiples [article 4 et 5]. Le quatrième et dernier chapitre est une discussion générale des travaux réalisés durant la thèse suivi de perspectives. SOMMAIRE REMERCIEMENTS .................................................................................................................. 2 PREAMBULE............................................................................................................................ 4 SOMMAIRE .............................................................................................................................. 5 I. INTRODUCTION ................................................................................................................. 7 I. 1. Sélection sexuelle .......................................................................................................... 7 I. 2. Le Choix du partenaire sexuel ....................................................................................... 9 I. 3. Vers un modèle plus dynamique du choix du partenaire sexuel.................................. 12 I. 4. Signaux multiples......................................................................................................... 15 I. 5. Objectifs de la thèse ..................................................................................................... 16 II. SIGNAUX MULTIPLES CHEZ LE TRITON PALMÉ..................................................... 22 II. 1. Etat de l’art de la théorie des signaux multiples ........................................................ 22 II. 2. Signaux multiples et accès à la reproduction chez les tritons palmés ........................ 24 II. 3. Etude de la parade du triton........................................................................................ 25 III. SIGNAUX MULTIPLES, SÉLECTION SUR LE CONTENU ET HÉTÉROGÉNÉITÉ DE L’ENVIRONNEMENT ..................................................................................................... 27 III. 1. Introduction ............................................................................................................... 27 III. 2. Condition dépendance de signaux multiples ............................................................. 28 III. 3. Effet de l’hétérogénéité de l’environnement sur l’expression de signaux multiples 29 IV. CONCLUSION GÉNÉRALE............................................................................................ 32 IV. 1. Choix des femelles sur critères multiples ................................................................. 32 IV. 2. Prise en compte de l’hétérogénéité de l’environnement ........................................... 35 IV. 3. Paradoxe du lek et conséquences sur les processus de sélection .............................. 37 IV. 4. Conséquences au niveau populationnel .................................................................... 38 IV. 5. Conséquences pour la biologie de la conservation ................................................... 40 IV. 6. Perspectives............................................................................................................... 42 BIBLIOGRAPHIE ................................................................................................................... 45 ARTICLES............................................................................................................................... 55 Article 1............................................................................................................................... 56 Cornuau JH, Rat M, Schmeller DS, Loyau A. 2012. Complex love decisions: why and how animals use several signals in mate choice. En préparation. Article 2............................................................................................................................. 100 Cornuau JH, Rat M, Schmeller DS, Loyau A. 2012. Multiple signals in the palmate newt: ornaments help when courting. Behavioral Ecology and Sociobiology, 66:1045-1055. Article 3............................................................................................................................. 112 Cornuau JH, Courtoix EA, Jolly T, Schmeller DS, Loyau A. It takes two to tango : Do both male and female identities influence complex courtship display in a newt species? En préparation. Article 4............................................................................................................................. 138 Cornuau JH, Pigeault R, Schmeller DS, Loyau A. Condition-dependance of multiple sexual traits in the palmate newt (Lissotriton helveticus). En préparation. Article 5............................................................................................................................. 174 Cornuau JH, Sibeaux A, Tourat A, Schmeller DS, Loyau A. Impacts of varying environmental conditions sexual traits and consequences for mate choice in the Palmate newt. En préparation. SUMMARY ........................................................................................................................... 214 RESUME................................................................................................................................ 215 I. Introduction. I. 1. Sélection sexuelle. Pour les individus pratiquant une reproduction sexuée, la capacité à transmettre une copie de leur génome dépend fortement de la capacité à trouver un partenaire sexuel. La sélection exercée sur les individus pour l’accès à la reproduction est appelée sélection sexuelle (Darwin 1871). A l’instar de la sélection naturelle dont elle peut être vue comme une sous-catégorie, la sélection sexuelle est un processus de tri entre variants phénotypiques. La spécificité de ce que l’on regroupe sous le terme de sélection sexuelle, est que le tri est effectué en fonction de l’accès différentiel aux partenaires sexuels (Danchin et al. 2008). La sélection sexuelle est ainsi une sélection sociale, c'est-à-dire basée sur des interactions intraspécifiques et qui utilise un système de communication du type « émetteur-receveur » (Lyon & Montgomerie 2012, Box 1, Figure 1). Box 1 : Communication sexuelle La communication sexuelle se fait à partir d’indices (cues) appelés signaux s’ils ont évolué pour la communication (Candolin 2003). Les signaux sont émis par l’émetteur (le plus souvent mâle) et traduit par le receveur (le plus souvent femelle). Un signal peut être défini comme « une structure ou une action de l’émetteur qui est sélectionnée pour ses effets sur le comportement du receveur via son système nerveux » (Bro-Jørgensen 2010). Il est important de rappeler que les deux protagonistes agissent pour leurs « propre bien » c'est-à-dire optimiser leur propre fitness. "Influencer le receiver” receveur" “Influencing Signal Emetteur “Information extraction” "Extraire de l’information" Receveur Figure 1. Représentation d’une communication entre un mâle (émetteur) et une femelle (receveur). Trois conditions sont nécessaires pour enclencher un mécanisme de sélection sexuelle (Danchin et al. 2008). La première requiert de la variation sur un trait sexuel. La deuxième repose sur le fait que cette variation du trait sexuel soit liée/se traduise par une variation sur l’aptitude phénotypique (fitness) c'est-à-dire sur la capacité d’un individu à produire plus de descendants matures par rapport aux autres individus. La troisième condition est que le trait sexuel et/ou l’aptitude phénotypique soit héritable c'est-à-dire transmissible à la descendance. Pour les espèces sexuées, une des principales différences entre les mâles et les femelles est l’investissement dans les gamètes. En effet, les mâles produisent de petits gamètes, considérés comme peu coûteux, contenant principalement le matériel génétique (spermatozoïdes). Au contraire, les femelles produisent une plus faible quantité de gros gamètes plus coûteux composés également de substance nutritive (ovocytes). Cette anisogamie dans la taille des gamètes qui conduit à des taux potentiels de reproduction différents, fait que, les femelles investissent généralement davantage que les mâles dans la descendance (Bateman 1948, Trivers 1972, mais voir Clutton-Brock 2009). Ainsi, pour une même quantité d’énergie, un mâle peut produire beaucoup plus de gamètes qu’une femelle. Il existe donc à un moment donné dans la population une grande quantité de gamètes mâles contre une faible quantité de gamètes femelles. Les femelles étant la ressource limitante, leur succès reproducteur va dépendre principalement de la qualité du partenaire sexuel (Box 2). Ainsi, nous pouvons nous attendre à ce que la sélection favorise les femelles ayant la capacité à choisir les mâles de meilleures qualités (partie B). A l’inverse le succès reproducteur des mâles va dépendre principalement de l’accès aux femelles réceptives. Ainsi, les mâles entrent généralement en compétition pour accéder aux femelles et investissent donc fortement dans la production de traits sexuels leurs facilitant cet accès. Deux grands types de stratégies existent pour accéder aux femelles et sont appelées sélection intra- et inter-sexuelle. La sélection intra-sexuelle, ou compétition intra-sexuelle, repose sur une compétition directe entre mâles pour l’accès aux femelles et conduit à l’évolution de caractères sexuels secondaires appelés dans ce cas « armements ». La sélection inter-sexuelle repose sur la capacité d’un mâle à être choisi par une femelle et conduit à l’évolution des caractères sexuels secondaires préférés par les femelles (ornements). Il existe de nombreux liens entre les armements et les ornements et faire la distinction entre la sélection intra- et inter-sexuelle n’est pas toujours aisée. En effet les femelles peuvent choisir (inter) les mâles ayant la capacité de gagner des combats (intra) allant même jusqu'à solliciter les combats (Berglund et al. 1996). Néanmoins il existe aujourd’hui un grand nombre de preuves empiriques d’évolution de caractères sexuels secondaires uniquement causée par le choix du partenaire sexuel effectué par les femelles (Andersson 1994, Danchin et al. 2008). Dans cette thèse nous avons surtout abordé l’évolution des caractères sexuels liés aux choix des femelles. Box 2 : La qualité du partenaire sexuel L’expression "qualité" d’un individu paraît au premier abord suffisamment claire pour que nous puissions nous affranchir de la définir. Cependant une lecture rapide de la littérature montre que ce terme est utilisé pour désigner des choses très différentes et peut donc rendre le concept parfois ambigu (Bergerons et al. 2011, Lailvaux & Kasumovic 2011). Ainsi, dans cette thèse, nous avons choisi de définir la qualité d’un individu comme « un axe de variation hétérogène entre les individus qui est positivement corrélé avec la fitness » (Wilson & Nussey 2010). Lorsque la qualité d’un mâle est d’ordre génétique nous parlons de bons gènes (good genes). Cependant, la qualité d’un mâle peut être d’ordre nongénétique, telle que la possession d’un territoire de bonne qualité. Un individu de meilleure qualité est donc un individu qui peut avoir une meilleure condition corporelle, être avantagé dans l’accès aux ressources, avoir acquis de meilleures ressources protéiques ou lipidiques, posséder une meilleure résistance aux parasites, une meilleure compatibilité génétique, un meilleur territoire, ou encore avoir une haute fertilité. Lorsque les qualités permettent un bénéfice pour les femelles s’exprimant seulement dans leur progéniture on parle de bénéfice indirect et lorsque les bénéfices s’expriment lors de la reproduction en cours nous parlons de bénéfice direct. I. 2. Le choix du partenaire sexuel. Le choix du partenaire sexuel est un « comportement où un individu influence la probabilité de se reproduire avec un individu particulier, plutôt qu’avec d’autres individus, suivant un ou plusieurs traits chez l’individu du sexe opposé » (Bateson 1983, Bradbury & Andersson 1987). L’appariement et/ou l’accouplement entre deux individus ne sont donc pas aléatoires. Ainsi, le choix du partenaire repose sur des règles de décision, sur des préférences internes à chaque individu et sur des contraintes environnementales telle que la disponibilité de différentes options (Wagner 1998). Plusieurs hypothèses ont été développées pour expliquer l’évolution des préférences sexuelles (Jones & Ratterman 2009, Tableau 1). Ces hypothèses peuvent être classées en fonction de l’information potentiellement contenue dans le trait des mâles et des bénéfices que les femelles peuvent espérer d’un tel choix (Jones & Ratterman 2009). En effet, le trait exprimé par les mâles peut être vu comme un "badge" signalant une information spécifique sur la qualité du mâle pour les femelles. Ce badge doit être résistant à la tricherie et donc être un signal honnête de la qualité du mâle (Maynard Smith & Harper 2003). La fiabilité du trait peut être maintenue par des contraintes, notamment par le coût trop important de la tricherie ou encore par le coût de la production et/ou de la maintenance du trait (Searcy & Nowicki 2005). Par exemple, la taille des individus peut être un bon indice de la capacité de croissance des mâles car seuls les mâles ayant une bonne capacité de croissance ont pu effectivement grandir. Les ornements des mâles sont généralement coûteux à produire et donc seuls les mâles de bonne qualité (Box 2) sont capables de les produire et de les maintenir. Ce principe est appelé « principe du Handicap » et indique que l’honnêteté du signal (et donc sa résistance à la tricherie) est scellée par le coût du signal (Zahavi 1975). Par exemple chez les Fou à pieds bleus Sulanebouxii, seuls les mâles de bonne qualité, c'est-à-dire ayant eut récemment accès à du poisson frais, peuvent maintenir la coloration bleue de leurs pattes et donc être attirants pour les femelles (Velando et al. 2006). D’autres formes d’information peuvent être extraites par les femelles sans pour autant être coûteuses pour les mâles. De nombreux exemples récents montrent que les femelles peuvent baser leurs choix sur des critères de compatibilité génétique (Mays & Hill 2004), et/ou comportementale (Schuett et al. 2010). Par exemple, chez l’épinoche à trois épines Gasterosteus aculeatus, les femelles utilisent l’odeur des mâles pour choisir un partenaire permettant de maximiser la diversité génétique de leurs progénitures (Aeschlimann et al. 2003, Milinski et al. 2005). De même, voir un mâle s’accoupler avec une femelle peut aussi être considéré comme un signal attestant de la qualité de ce mâle (copiage du choix de partenaire, Danchin et al. 2011). Dans ce cas, les femelles utilisent de l’information sociale qui, sans forcement être coûteuse pour le mâle, donnent aux femelles observatrices une indication indirecte sur la qualité des mâles, telle que cette qualité est perçue par les autres femelles (Witte & Ryan 1998, 2002). La préférence des femelles peut aussi avoir évolué sans réelle information véhiculée par le trait et sans bénéfice direct pour les femelles. Un modèle classique est le trait de Fisher (Fisher 1915, 1930, Lande 1981, Kirkpatrick 1982). Ce modèle invoque une préférence chez les femelles héritable de mère en fille et un trait phénotypique chez les mâles héritable de père en fils. Dans ce modèle, les femelles exprimant une préférence pour le trait du mâle engendreront des filles qui exprimeront la même préférence que leurs mères et des fils qui exprimeront le même trait phénotypique que leur père et qui seront donc préférés à leur tour. Ce modèle repose sur une corrélation génétique entre le trait des mâles et la préférence de la femelle. Il permet d’expliquer le maintien d’une préférence chez les femelles et même son renforcement, mais n’explique pas l’apparition de la préférence chez les femelles (Jones & Ratterman 2009). L’hypothèse du biais sensoriel permet d’expliquer l’apparition d’une préférence chez les femelles pour un trait chez les mâles. En effet, dans ce modèle la préférence des femelles existerait avant l’apparition du trait chez les mâles (Basolo 1990, Ryan & Rand 1990, Fuller et al. 2005, Kokko et al. 2003). Cette préférence pourrait être d’origine non sexuelle et maintenue par la sélection utilitaire. Un exemple connu est la préférence exercée par les femelles guppys Poecilia reticulata pour la coloration orangée des mâles (Rodd et al. 2002). Rodd et al. (2002) ont en effet montré, en utilisant plusieurs populations de guppy, que la force d’attraction des femelles pour la coloration orangée dans un contexte non sexuel (ici la nourriture) explique 94% de la variation inter-population de la préférence sexuelle des femelles pour cette coloration. Un dernier type de modèle, basé sur le conflit sexuel, peut expliquer l’évolution des préférences des femelles (Hollande & Rice 1998, Cameron et al. 2003, Parker 2006). Les intérêts des mâles et des femelles divergent souvent lorsqu’il s’agit de reproduction (voir box 1). Or, la sélection sexuelle favorise les mâles ayant la capacité d’accéder à la reproduction. Ainsi, toute mutation qui permettrait à un mâle d’être avantagé dans l’accès à la reproduction (par exemple en utilisant un biais sensoriel), lui procurerait une meilleure descendance et donc la mutation serait sélectionnée, même si celle-ci n’apporte aucun bénéfice à la femelle. Cependant, si cela induit une reproduction suboptimale de la femelle, il est attendu que toute mutation provoquant une résistance des femelles pour le trait sera sélectionnée. Suivant le modèle de sélection antagoniste, il s’en suit une nouvelle sélection sur l’habilité des mâles à séduire les femelles (Hollande & Rice 1998). Il existe donc une multitude de théories expliquant l’évolution de la préférence des femelles (Tableau 1). Toutes ces théories ne sont pas mutuellement exclusives (Kokko et al. 2003). Ceci est particulièrement vrai pour les traits de Fisher ou les biais sensoriels qui peuvent agir conjointement avec toutes les autres théories (Jones & Ratterman 2009). Tableau 1. Exemples de modèles expliquant l’évolution de la préférence des femelles. Les modèles sont souvent classés selon le potentiel informatif du trait pour la femelle. Adapté de Jones & Ratterman (2009). Mécanisme d’évolution de la préférence des femelles Trait informatif Bénéfice direct Particularités Références Le bénéfice s’exprime lors de la reproduction en cours Nakatsuru & Kramer 1982 Le bénéfice s’exprime uniquement dans la progéniture Welch et al. 1998 Déséquilibre de liaison entre la préférence des femelles et le trait chez les mâles Rowe 2001 Biais sensoriel La préférence des femelles précède l’apparition du trait chez les mâles Basolo 1990 Conflit sexuel Evolution d’une résistance des femelles contre des traits manipulateurs chez les mâles Demary & Lewis 2008 Bénéfice indirect Trait non-informatif Processus de Fisher I. 3. Vers un modèle plus dynamique du choix du partenaire sexuel ? L’approche de classification de l’évolution des préférences des femelles pour un trait chez les mâles peut avoir l’inconvénient d’admettre de manière implicite que la communication sexuelle est robuste aux variations spatiale et temporelle (Jones & Ratterman 2009). En effet, les études expérimentales et les modèles théoriques ont eu tendance à considérer les préférences sexuelles comme fixes (Wagner 1998). Or, comme tout trait phénotypique, le choix du partenaire et l’expression des traits sexuels peuvent être expliqués par des gènes, par l’environnement et par l’interaction entre les gènes et l’environnement (Greenfield & Rodriguez 2004). En effet, les gènes peuvent s’exprimer différemment dans des environnements différents ; ce phénomène est connu sous le nom de plasticité phénotypique (Schlichting & Pigliucci 1998). L’expression d’un génotype suivant un gradient d’environnement est appelé norme de réaction (Schlichting & Pigliucci 1998). Lorsque les normes de réaction de différents génotypes ne sont pas parallèles, une interaction gène-environnement se produit (GEI, Fig. 2, Greenfield & Rodriguez 2004). a) e py to 1 né h P c) 1 b) 2 1 2 d) 2 1 Environnement 2 Figure 2. Expression du phénotype (trait sexuel, choix du partenaire) dans différentes conditions environnementales. Le trait plein représente un génotype A, le trait en tiret représente un génotype B. a) Les génotypes A et B ont des niveaux d’expression qui ne diffèrent pas en fonction de l’environnement (phénotype = génotype), b) Les génotypes A et B ont des niveaux d’expression qui diffèrent en fonction de l’environnement (plasticité) mais la norme de réaction est parallèle (phénotype = génotype + environnement). c) et d) Les génotypes A et B ont des niveaux d’expression qui diffèrent en fonction de l’environnement et les normes de réaction ne sont pas parallèles (phénotype = génotype + environnement + génotype × environnement). En c) l’expression du génotype A est toujours supérieure à celui de B mais contrairement à la situation d) où les normes de réaction se croisent. Ce phénomène couramment étudié pour des traits soumis à la sélection utilitaire a fait l’objet de rares études dans le cadre du choix du partenaire (Cornwallis & Uller 2010). Étant donné que la plupart des modèles d’évolution de la préférence pour un partenaire se focalise sur le choix pour un trait plus ou moins informatif (Jones & Ratterman 2009), des études plus récentes se sont focalisées sur l’effet de l’hétérogénéité de l’environnement sur le choix du partenaire en lui-même et sur la condition dépendance des traits sexuels (Cotton et al. 2006). Ces études montrent que le choix des femelles peut varier rapidement dans sa sélectivité ainsi que dans sa direction selon l’environnement et donc affecter grandement le régime de sélection (Jennions & Petrie 1997). Plusieurs sources environnementales peuvent influencer le choix des femelles. On distingue des facteurs externes telles que la qualité de l’habitat, la prédation, la densité et la sex-ratio et des facteurs internes tels que l’âge de la femelle, son expérience, sa condition physique, sa personnalité et son statut social (Jennions & Petrie 1997, Cotton et al. 2006). Un exemple classique est la variation dans le choix causée par une variation de la réceptivité des femelles. Par exemple, chez le triton ponctué Lissotriton vulgaris, la sélectivité des femelles est très faible lorsqu’elles ne se sont pas encore reproduites et leur choix apparait comme aléatoire suivant la qualité du mâle (Gabor & Halliday 1997). En revanche, après une première reproduction, les femelles deviennent plus sélectives et montrent une préférence pour les mâles de meilleure qualité c'est-à-dire les mâles ayant une plus grande condition corporelle (Gabor & Halliday 1997). En effet, il parait plus avantageux de se reproduire avec un mâle sans avoir fait un choix au préalable que de ne pas se reproduire (Jennions & Petrie 2000). Par contre, il devient avantageux pour les femelles en terme de fitness d’être de plus en plus sélectives à mesure que les épisodes de reproduction se succèdent (Jennions & Petrie 2000). Des études récentes suggèrent également un fort effet de l’hétérogénéité de l’environnement sur la condition-dépendance des ornements (Robinson et al. 2007, 2012, Scordato et al. 2012). Il a notamment été montré, chez le Gobemouche à collier Ficedula albicollis, que les femelles en couple avec un mâle exprimant une forte ornementation (coloration blanche connue pour être dépendante de la condition et soumise à la sélection sexuelle) avaient une meilleure fitness durant les périodes sèches (Robinson et al. 2012). Cependant cette relation était inversée en période humide. Ainsi les femelles en couple avec des mâles fortement ornementés avaient une fitness plus faible que les femelles en couple avec des mâles moins ornementés (Robinson et al. 2012). Ces exemples montrent que l’honnêteté d’un signal est relative à l’environnement et amène des questionnements sur le caractère adaptatif du choix des femelles en fonction des environnements. Ces questions seront abordées dans les articles 1 et 5. Bien que la prise en compte des effets de l’environnement apparaisse comme essentielle pour la compréhension des mécanismes de sélection sexuelle, peu d’études incluent ce facteur dans la compréhension du choix du partenaire (Bussière et al. 2008, Cornwallis & Uller 2010). Il convient donc de l’inclure dans de futures études (article 1, article 5). 1. 4. Signaux multiples. Les exemples que nous venons de développer dans les parties précédentes de l’introduction font appel à une préférence des femelles pour un trait chez les mâles. Dans la partie suivante nous avons intégré le fait que les femelles expriment souvent des préférences multiples et que les mâles possèdent plusieurs traits potentiellement sélectionnés par les femelles. Cependant, toutes les théories développées précédemment pour expliquer l’évolution d’une seule préférence, ainsi que l’effet de l’hétérogénéité de l’environnement sur le choix et l’expression des traits sexuels demeurent valables dans un contexte de préférences multiples. La possibilité que la communication sexuelle soit basée sur plusieurs traits, résultant en plusieurs critères de choix du partenaire, a reçu une attention particulière en sélection sexuelle (Guilfort & Dawkins 1991, Møller & Pomiankowski 1993). En effet, des observations ont montré que les mâles exhibaient de nombreux ornements et que les femelles pouvaient exprimer des préférences pour plusieurs traits (Guilfort & Dawkins 1991, Møller & Pomiankowski 1993). On parle ainsi de signaux multiples, signaux complexes, ou signaux à multi-composants (Candolin 2003). Les recherches sur les signaux multiples dans le cadre du choix de partenaires se sont intensifiées (Candolin 2003, Hebets & Papaj 2005) et continuent d’être un axe de recherche important en sélection sexuelle (Lozano 2009, Bro-Jørgensen 2010). Ces recherches se concentrent principalement sur deux grands types de questions : (1) Pourquoi les individus utilisent-ils des signaux multiples dans la communication sexuelle ? (2) Comment utilisent-ils et hiérarchisent-ils les multiples signaux ? La première question vise à comprendre les mécanismes évolutifs qui permettent l’apparition et le maintien de l’utilisation de plusieurs signaux. La seconde vise à comprendre les règles de décision utilisées par les femelles lorsque plusieurs signaux sont disponibles. Dans cette thèse, nous avons réalisé une synthèse sur les signaux multiples dans le cadre du choix du partenaire. Ces signaux multiples sont également utilisés dans le cadre plus large de la sélection sociale, tel que la communication parent-enfant (Jacob et al. 2011), les receveurs multiples (Pryke et al. 2001, Andersson et al. 2002), la compétition intra-sexuelle (Chaine & Lyon 2008) ou la dominance (Chaine et al. 2011). Cependant, nous avons choisi de nous restreindre spécifiquement au choix du partenaire. Au cours de ce travail bibliographique, nous avons synthétisé les facteurs ultimes et proximaux du choix du partenaire basés sur des signaux multiples en utilisant des exemples récents. Nous avons également mis en lumière la nécessité d’inclure certains signaux (notamment l’information sociale et la compatibilité) jusque-là peu considérés dans le cadre théorique des signaux multiples. Ensuite, nous avons développé deux grands axes d’avenir pour l’étude des signaux multiples. Le premier est la prise en compte de l’hétérogénéité de l’environnement sur les signaux multiples. Le second est la nécessité de coupler les théories développées sur les signaux multiples avec les connaissances apportées par l’écologie cognitive afin de mieux comprendre les prises de décision basées sur des signaux complexes. Cette synthèse permet d’avoir une vision générale de la théorie des signaux multiples tant sur les recherches passées et actuelles, que sur ses futurs développements. Pour plus d’informations sur les signaux multiples, je vous invite à lire ce travail de synthèse que vous trouverez dans l’article 1 “Complex love decisions: why and how animals use several signals in mate choice?”. I. 5. Objectifs de la thèse. L’objectif général de cette thèse était l’étude des signaux multiples dans le cadre du choix du partenaire sexuel. En effet, si les facteurs ultimes et proximaux de l’utilisation de signaux multiples sont maintenant bien étudiés, il reste encore à apporter les preuves expérimentales sur des taxons variés (article 1). De plus, bien qu’il soit admis que l’hétérogénéité de l’environnement puisse avoir un effet important sur les signaux multiples et éventuellement expliquer leur évolution, les preuves expérimentales demeurent encore relativement faibles (article 1). Nous avons choisi le triton palmé Lissotriton helveticus (Box 3) comme modèle biologique à notre étude sur les signaux multiples dans le cadre du choix du partenaire. Cette espèce rassemble les qualités adéquates pour notre étude. En effet les mâles possèdent de nombreux caractères sexuels secondaires tels que des palmures au niveau des pattes arrières, une crête dorsale, un filament caudal, une coloration ventrale orangée, ainsi que des comportements de parade complexes (Halliday 1975, Griffiths & Mylotte 1988, Figures 3, 4 et 5). Chez cette espèce de triton, les interactions de compétition directe entre les mâles semblent très faibles (Wells 2007), même si une compétition par exploitation n’est pas à exclure. La présence de ces traits complexes et la faible compétition entre les mâles suggèrent un rôle actif du choix des femelles dans l’évolution de ces traits sexuels multiples (article 1). Dans une première partie nous avons étudié l’effet des signaux multiples du mâle sur le choix des femelles chez le triton palmé (article 2 et 3). Nous avons tout d’abord examiné quels traits sexuels avaient le potentiel pour être sélectionnés par les femelles, et ceci en nous focalisant à la fois sur les traits sexuels morphologiques et des traits comportementaux des mâles (article 2). Dans le but de mieux comprendre l’importance des traits comportementaux identifiés grâce à l’étude rapportée dans l’article 2, nous avons examiné s’ils étaient répétables et s’ils étaient liés à des caractéristiques phénotypiques (article 3). Comme nous l’avons expliqué dans l’introduction, les traits sexuels des mâles peuvent être révélateurs de leur qualité. Cette condition-dépendance des traits sexuels est primordiale pour distinguer les hypothèses d’évolution des préférences multiples (article 1). Dans une deuxième partie nous avons donc exploré le rôle des signaux multiples des mâles en tant que signaux de qualité individuelle chez le triton palmé (article 4 et 5). Nous avons également testé l’effet de l’hétérogénéité de l’environnement sur l’honnêteté des traits sexuels et les conséquences que cela pouvait avoir sur le choix des femelles (article 5). Box 3 : Le Triton palmé Le triton palmé Lissotriton helveticus est un amphibien de la famille des Salamandridae. Cette espèce est largement répandue en France autant en basse altitude qu’en montagne (observation jusqu'à 2400 mètres). Cette espèce est également présente en Allemagne, Grande-Bretagne, Suisse, Belgique, Luxembourg, ainsi que dans le nord de l’Espagne et du Portugal. Le triton palmé est une espèce protégée, classée comme « considérations mineures » par l’IUCN. Les tritons palmés font partis des plus petits tritons d’Europe puisque la taille totale des mâles ne dépasse que rarement 8 cm pour les mâles et 10 cm pour les femelles. Ils requièrent peu d’espace lorsqu’ils sont maintenus en captivité, ce qui en fait un modèle intéressant en laboratoire. Les tritons palmés peuvent vivre jusqu'à 8 ans à 11 ans dans la nature (Amat et al. 2010). Il existe deux grandes phases de vie du triton durant l’année. Durant la phase dite terrestre, le triton vit à terre, proche des zones humides et dépend de la présence de couvert boisé. La phase dite aquatique, correspond à la période de reproduction. Lorsqu’il est en phase aquatique, le triton palmé vit dans des milieux aquatiques stagnants ou à très faible courant tels que des petites flaques d’eau, des mares peu profondes, ou des lacs (Denoël & Lehmann 2006). La migration prénuptiale marque la fin de la phase terrestre et le début de la phase aquatique. Les mâles se dirigent vers leur point d’eau et s’ornent de caractères sexuels secondaires, inexistants en période terrestre, tels que les palmures, le filament caudal, et la crête dorsale (Figure 5). C’est également en phase aquatique que les mâles expriment des comportements de parade envers les femelles (Figure 3 et 4). La reproduction se fait par transfert de spermatophore que le mâle place sur le sol et que la femelle prend au niveau de son cloaque. La femelle peut prendre plusieurs spermatophores de plusieurs mâles différents au cours de la saison de reproduction. Il est important de noter qu’il n’existe pas de comportement d’amplexus chez le triton palmé et qu’il n’existe pas non plus de comportement agressif physique du mâle envers la femelle (Wells 2007). En revanche, les mâles passent une grande partie de leur phase aquatique à parader devant les femelles et à les suivre. Les femelles pondent des œufs (jusqu'à 400) dans l’eau en enroulant chaque œuf individuellement dans une feuille. Ce geste, important pour la survie des œufs, constitue le seul soin parental que la progéniture reçoit (Miaud 1990). A la fin de la saison de reproduction, les mâles et les femelles retournent en phase terrestre. Une fois le développement de l’œuf achevé (après environ 1 mois), les larves de triton croissent dans l’eau, durant environ 2 mois, avant de gagner la terre. Les individus reviendront dans l’eau une fois devenus sexuellement mature c'est-à-dire pas avant l’âge de 3 à 4 ans. reniflement Le mâle se place devant la femelle Coup de fouet Vague Orientation poursuite Parade « statique » éventail éventail Parade « de retrait » Coup de fouet Devancer Trembler Phase du transfert de spermatophore Touche queue Dépôt Avancer encore Appât distal, Queue frémissante interrompre repousser Figure 3. Représentation des séquences principales du comportement de cour chez le triton palmé Lissotriton helveticus. Le mâle est représenté en noir et la femelle en blanc. Eventail = fan, coup de fouet = whip, trembler = quiver. D’après Arntzen & Sparreboom (1989) et Denoël (1999). Figure 4. Trois comportements principaux de parade chez le triton palmé Lissotriton helvetivus. En haut, le comportement de vague (wave). Au milieu le comportement de coup de fouet (whip). En bas le comportement d’éventail (fan). D’après Halliday (1977) et Wells (2007). Figure 5. Traits morphologiques du mâle triton palmé Lissotriton helveticus. En haut sont représentées les extensions de la peau (crête, filament, palmure). En bas sont représentés les patterns de coloration ventrale. (Photographies : Jérémie Cornuau). II/ Signaux multiples chez le triton palmé. II. I. Etat de l’art de la théorie des signaux multiples. La théorie des signaux multiples stipule que plusieurs caractères sexuels secondaires sont importants pour expliquer le succès reproducteur des mâles (article 1). L’expression de ces traits peut être plus ou moins corrélée et l’influence d’un trait par rapport à l’autre peut être différente suivant la présence ou non d’autres traits (article 1). Pour de nombreuses espèces, le succès des mâles dépend d’une combinaison de traits comportementaux (parade) et de traits morphologiques (Tableau 2). Par exemple, chez la rainette arboricole Hyla arborea, la couleur et le chant des mâles expliquent le choix des femelles et les deux traits semblent avoir la même importance (Richardson & Lengagne 2010, Richardson et al. 2010). A l’inverse chez l’épinoche à cinq épines Culaea inconstans, alors que la couleur nuptiale et la taille des mâles peuvent être tous deux des critères de choix, la parade est le critère le plus important pour les femelles (Ward & McLennan 2008). Tableau 2. Liste non exhaustive d’espèces pouvant utiliser à la fois des traits morphologiques et des traits comportementaux au cours du choix du partenaire sexuel. Les trais indiqués en gras sont les traits prédominants dans le choix des femelles. Espèce Grillons des champs Araignée Loup Rainette arboricole Lézard à collerette commun Guppy Traits morphologiques Traits comportementaux Références Gryllus campestris Schizocosa uetzi Hyla arborea Taille de la harpe Taux de gazouillement Scheuber et al. 2004 Couleur des pattes Schamble et al. 2009 Coloration du sac vocal Fréquence de coups de patte Chant Crotaphytus collaris Taille et couleur corporelle Comportements territoriaux Bairs et al. 2008 Poecilia Coloration Sigmoïde reticulata Epinoche à Gasterosteus Taille et coloration Nage en zigzag trois épines aculeatus corporelle Gobie Rhinogobius Nageoire dorsale* Nage* brunneus Epinoche à Culaea Taille et coloration Intensité de parade cinq épines inconstans corporelle Junco ardoisé Junco Couleur de la queue Chant hyemalis Colin de Callipeplaga Coloration corporelle Parade Gambel mbelii Siffleur doré Pachycephala Coloration de la gorge Chant pectorialis *l’importance donnée à chaque trait dépend de la condition de la femelle. Richardson & Lengagne 2010 Kodric-Brown & Nicoletto 2001 Künzler & Bakker 2001 Suk & Choe 2008 Ward & McLennan 2009 Hill et al. 1999 Hagelin & Ligon 2001 Van Dongen & Mulder 2008 De nombreuses études se sont focalisées sur les signaux multiples chez les anoures (Poole & Murphy 2007, Akre & Ryan 2010, Richardson & Lengagne 2010, Richardson et al. 2010, Taylor et al. 2011), alors que des études équivalentes sont rares chez les urodèles. Ainsi à notre connaissance la seule étude suggérant l’implication potentielle de signaux multiples chez un urodèle a été conduite chez le triton palmé Lissotriton helveticus (Haerty et al. 2007). Pourtant, il est généralement admis que chez les urodèles les comportements de parade ainsi que les traits morphologiques et certainement la combinaison des deux influencent grandement le choix des femelles (Wells 2007). Parmi les traits morphologiques, les extensions de la peau telle que la crête dorsale semblent affecter le choix des femelles chez de nombreuses espèces de tritons (Wells 2007). Ainsi il a été montré chez le triton crêté Triturus cristatus et le triton ponctué Lissotriton vulgaris que les mâles possédant une large crête étaient avantagés dans l’accès aux femelles par rapport aux mâles possédant une crête moins développée (Green 1991, Gabor & Halliday 1997). La taille du mâle peut également être un critère de choix pour les femelles (Haerty et al. 2007). Enfin la coloration du mâle pourrait avoir une importance chez les tritons (Denoël & Dollen 2010, Secondi et al. 2012), ce qui n’a été jusque-là que peu explorée chez les amphibiens (Richardson & Lengagne 2010, Richardson et al. 2010). Les comportements de parade pourraient également être des éléments clefs du choix des femelles (Wells 2007). En effet, chez le Pléthodonte nord américain Desmognathus ocoee, les mâles ayant des fréquences de parade élevée ont un succès plus important dans l’accès aux femelles par rapport aux males à faible fréquence de parade (Vinnedge & Verrell 1998). En revanche à notre connaissance aucune étude n’a étudié l’influence de la combinaison de la parade et des traits morphologiques sur l’accès à la reproduction chez les urodèles. Comme nous l’avons précédemment décrit, les tritons palmés mâles possèdent des traits morphologiques et comportementaux potentiellement importants dans le choix des femelles (Haerty et al. 2007, Wells 2007, Figure 3, 4, 5). Nous avons réalisé des expériences, synthétisées sous la forme de deux articles (article 2 et 3), dont le but était de mieux appréhender le rôle des traits morphologiques et comportementaux dans le cadre du choix du partenaire chez le triton palmé. II. 2. Signaux multiples et accès à la reproduction chez les tritons palmés. [article 2] Haerty et al. (2007) ont mesuré le choix des femelles de triton palmé pour le filament et pour la taille, deux traits fortement indépendants l’un de l’autre. Leurs résultats montrent que les femelles passaient d’avantage de temps dans le compartiment contenant des mâles de petite taille et possédant un filament plus développé (Haerty et al. 2007). Cependant, le protocole utilisé dans cette étude ne permettait pas aux mâles de parader librement et cela avait pour conséquence de retirer à l’appréciation des femelles un trait potentiellement important pour elles. Afin de tester l’importance des signaux multiples dans le choix du partenaire chez le triton palmé, nous avons pris en compte à la fois l’importance du filament mais également d’autres traits tels que la coloration ventrale, la taille corporelle et les comportements de parade (article 2). Nous avons utilisé un protocole expérimental ouvert, c'est-à-dire que les individus interagissaient librement, et les femelles avaient le choix entre deux mâles qui différaient pour la taille de leur filament (un grand versus un petit filament). Nos prédictions étaient un meilleur accès à la reproduction pour les mâles possédant un grand filament, une fréquence de parade forte, une coloration orangée intense et une taille réduite. Afin de tester l’importance du filament seul, nous avons réalisé une seconde expérience où nous avons expérimentalement inversé la taille relative des filaments des mâles proposés au choix des femelles. Ainsi, les mâles à grand filament étaient devenus des mâles à petit filament. Nos prédictions étaient une diminution du succès reproducteur des mâles dont la longueur du filament avait été modifiée. Les résultats que nous avons obtenus sont détaillés dans l’article 2. Ils montrent que l’accès des mâles à la reproduction est fortement influencé par la taille du filament. En effet les mâles à grand filament avaient un meilleur accès à la reproduction. De plus, lorsque le filament avait été réduit, le succès des mâles l’avait également été. En revanche, une inversion de la taille relative du filament n’avait pas complètement inversé le succès des mâles. Ce résultat suggère soit que d’autres traits corrélés au filament sont potentiellement importants, soit que sous un certain seuil de longueur de filament, la taille du filament n’a plus d’importance dans l’accès à la reproduction. Nos résultats ont également montré que la parade est un élément déterminant pour expliquer l’accès à la reproduction. En effet, les mâles qui avaient une activité de parade plus importante avaient un meilleur accès à la reproduction. Ainsi nous montrons pour la première fois chez un urodèle l’importance conjointe des traits morphologiques et comportementaux dans le cadre du choix de partenaire. II. 3. Etude de la parade du triton palmé. [article 3] Dans l’étude précédente, nous avons montré que l’accès à la reproduction des mâles était partiellement lié à la variation dans leur temps de parade (article 2). Bien que moins étudiée, la variation individuelle dans le temps de parade n’en est pas moins importante (Bell et al. 2009). La question est alors de savoir si les comportements de parade sont constants dans le temps et donc répétables entre les individus. De nombreuses études ont montré que les comportements de parade des mâles sont globalement répétables mais cette répétabilité reste souvent assez faible, une grande partie de la variation restant inexpliquée (Bell et al. 2009). Des études récentes mettent en avant que cette faible constance des comportements de parade pourrait être due à une forte labilité en fonction des situations et notamment des caractéristiques des femelles (Ruiz et al. 2008, Lehtonen et al. 2011). En effet, alors qu’un triton palmé mâle ne peut pas ajuster la taille de son filament en fonction de la femelle qu’il rencontre, il peut ajuster son comportement de parade. Ainsi, un mâle donné pourrait choisir de parader peu pour une femelle qu’il juge de mauvaise qualité afin, par exemple, de limiter ses dépenses énergétiques. Cependant peu d’études se sont intéressées à l’importance relative de l’identité des mâles et des femelles sur la variation des comportements de parade (Lehtonen et al. 2011). A notre connaissance une seule étude a étudié la répétabilité des comportements de parade chez les tritons (Michalak 1996, voir Wells 2007 et Bell et al. 2009). Cette étude a montré que, globalement, les comportements des mâles ne sont pas répétables (Michalak 1996). Nous avons donc examiné l’importance relative de l’identité des mâles et des femelles et nous avons testé l’influence relative des caractéristiques du mâle et de la femelle sur la variation dans les comportements de parade tels que les éléments de la parade statique, les comportements trembler (quiver) et touche queue (touch tail) (Figures 3 et 4). Pour cela nous avons mis en place un protocole expérimental où le comportement de chaque individu (mâle et femelle) était mesuré trois fois, avec un intervalle d’un jour puis de quatre jours. Nous avons mis en évidence que le temps passé en parade statique était en partie expliqué par l’identité du mâle et pas du tout expliqué par l’identité de la femelle. Les comportements de trembler (quiver) semblaient être expliqués à la fois par l’identité du mâle et de la femelle. Enfin les comportements de touche queue (touch tail) étaient uniquement expliqués par l’identité de la femelle. Nous n’avons trouvé aucun lien entre les comportements et les traits morphologiques, et ceci à la fois pour les mâles et les femelles. Notamment, le temps de parade n’était pas lié à la taille du filament du mâle. De même, nous n’avons pas trouvé de lien entre la constance dans les comportements de parade et les traits morphologiques des mâles : les mâles de grande taille n’étaient pas plus répétables que les mâles de petite taille. Contrairement à ce que les travaux de Michalak (1996) suggéraient, nous avons montré pour la première fois chez un urodèle que les comportements de la parade sont significativement répétables. Nous montrons également l’importance de prendre en compte les caractéristiques des mâles et des femelles pour comprendre l’expression des comportements de parade. III. Signaux multiples, sélection sur le contenu et hétérogénéité de l’environnement chez le triton palmé. III. 1. Introduction. Les signaux multiples, ainsi que les potentielles interactions entre les différents signaux s’étudient en évaluant la réponse des femelles pour chaque signal pris séparément ainsi que la combinaison de plusieurs signaux (Partan & Marler 1999, 2005, article 1). Bien que les résultats de telles études soient nécessaires et extrêmement utiles pour comprendre l’importance relative de chaque signal, ils ne renseignent pas directement sur l’information véhiculée par les signaux (Hebet & Papaj 2005, tableau 1, article 1). Or, une grande partie des hypothèses expliquant l’utilisation de signaux multiples reposent sur l’information extraite par les femelles des traits des mâles (article 1). Møller & Pomiankowski (1993) ont émis trois grandes hypothèses permettant d’expliquer l’évolution des signaux multiples qui sont les signaux non-informatifs (les signaux ne sont pas liés à la condition des mâles), les messages redondants (les signaux contiennent la même information sur la condition) et les messages multiples (les signaux renseignent sur des aspects différents de la condition) (article 1). Une manière de tester le contenu informatif des signaux est d’évaluer leur conditiondépendance (Cotton et al. 2004). Nous avons réalisé des expériences, synthétisées sous la forme de deux articles (article 4 et 5), dont le but était de mieux appréhender le rôle des traits morphologiques et comportementaux en tant que signaux de qualité individuelle chez le triton palmé. III. 2. Condition dépendance des signaux multiples. [Article 4]. Dans cette expérience nous avons testé les hypothèses de Møller & Pomiankowski (1993) sur l’évolution du filament, des palmures, de la crête et de la parade chez le triton palmé. Pour tester ces hypothèses, nous nous sommes basés sur deux grands types de qualité du mâle qui sont le statut immunitaire et la condition corporelle, un aspect peu étudié pour la classe des urodèles. Si l’on se réfère à la synthèse de Cotton et al. (2004) sur la condition dépendance, la seule preuve expérimentale de condition dépendance chez les amphibiens est une expérience réalisée par Green (1991) chez le triton ponctué Lissotriton vulgaris. Dans cette expérience Green (1991) a montré, en utilisant 14 tritons mâles, que les tritons du groupe soumis à un stress de nourriture avaient une crête moins développée que le groupe contrôle de tritons nourris. Il a aussi montré une relation positive entre la condition corporelle et la taille de la crête sur 70 mâles capturés dans la nature (Green 1991). A ce jour et à notre connaissance, il n’existe en revanche aucune étude ayant utilisé la réponse à un challenge immunitaire comme indice de la qualité des mâles chez les urodèles. Nous avons donc soumis des mâles à un challenge immunitaire afin de voir si leurs traits sexuels étaient corrélés à leur statut immunitaire. Pour cela, nous avons injecté des fragments de bactéries (LPS) connues pour activer le système immunitaire chez une grande variété d’animaux. Nous avons également testé si les traits sexuels co-varient avec la condition corporelle des mâles. Lors de cette expérience, nous n’avons mesuré aucun effet du challenge immunitaire sur l’expression des traits sexuels. Une des explications possible à ce résultat, et que ce protocole ne permet pas de vérifier, est que le traitement n’a pas activé le système immunitaire des tritons. Cependant avec un protocole similaire d’injection, des expériences précédentes ont montré que le système immunitaire des anoures était activé, même si ce type de preuve n’existe pas chez les urodèles. Nous avons mesuré une co-variation positive entre trois caractères sexuels (la taille du filament, la surface de palmure, l’indice de développement de la crête), ainsi qu’une covariation positive de ces trois traits avec l’indice de condition corporelle des mâles. En revanche la parade ne semblait aucunement liée à la condition corporelle du mâle. Globalement nos résultats suggèrent que les traits morphologiques (filament, palmure, crête) sont des signaux redondants de la qualité du mâle, renseignant sur la condition corporelle des mâles. De son côté, la parade semble être un trait non informatif sur la condition corporelle du mâle. III. 3. Effet de l’hétérogénéité de l’environnement sur l’expression de signaux multiples [Article 5]. Bien qu’elle soit encore relativement peu étudiée, l’hétérogénéité de l’environnement peut avoir des effets importants sur la sélection sexuelle (Greenfield & Rodriguez 2004, Cornwallis & Uller 2010). En particulier, l’hétérogénéité de l’environnement peut être un facteur expliquant l’évolution du choix des femelles pour des signaux multiples (article 1, Bro-Jørgensen 2010). Les signaux multiples pourraient ainsi permettre aux femelles de choisir les meilleurs mâles dans des environnements fluctuants (article 1). En effet l’hétérogénéité de l’environnement pourrait avoir un effet sur l’honnêteté des signaux, effet qui pourrait être différent en fonction de la labilité des traits (Bro-Jørgensen 2010). Les traits phénotypiques condition-dépendants indiquent la qualité des mâles au moment où ils sont produits. Lorsque les traits sont relativement statiques il peut alors exister un écart important entre l’expression du trait et la qualité de l’individu au moment de l’évaluation du trait. Par exemple alors que la qualité des individus peut varier rapidement dans le temps et l’espace, l’expression de certains traits morphologiques reste constante (Figure 6, Bro-Jørgensen 2010). Dans ce cas les traits morphologiques indiqueraient la qualité du mâle uniquement au moment de leurs productions. Lorsque l’environnement change les traits morphologiques peuvent ainsi devenir des signaux non fiables de la qualité des mâles (Figure 6, Robinson et al. 2012). Dans ces cas, les femelles devraient utiliser en priorité des signaux plus flexibles comme certains traits comportementaux qui indiquent la qualité présente du mâle (Figure 6). Cependant ces traits flexibles peuvent également être plus sensibles à un ajustement bref des mâles (notamment en cas de variation dans la "motivation"). En effet les mâles pourraient investir fortement dans l’expression de ces traits indépendamment de leur qualité afin d’accéder à la reproduction. Il existe très peu de preuves empiriques de ces hypothèses (Bro-Jørgensen 2010 mais voir Scheuber et al. 2003a,b, Whattam & Bertram 2011) particulièrement dans le cadre de changements d’environnement (Candolin & Wong 2012). Nous avons ici testé cette hypothèse chez le triton palmé, une espèce où à la fois les traits morphologiques et comportementaux sont importants pour expliquer l’accès à la reproduction (article 2). Pour cela, nous avons soumis des mâles à quatre traitements différents suivant un plan factoriel. Nous avons créé une hétérogénéité de l’environnement en faisant varier la disponibilité en nourriture. Les mâles ont été soumis à un traitement de quatre semaines avec deux traitements de nourriture (présence ou absence) suivi d’une semaine de traitement de nourriture (présence ou absence). 1 Environnement 2 b) Expression du trait Qualité du mâle Expression du trait Qualité du mâle a) 1 Environnement 2 Figure 6. Expression du phénotype (trait sexuel et qualité associée au trait) sur des traits multiples a-trait statique, b-trait dynamique. Le trait plein représente une qualité I associé à un trait I, le trait en tiret représente une qualité II associé à un trait II. Dans les deux cas a et b la qualité des mâles change suivant l’environnement. En a- l’expression du trait ne suit pas la qualité du mâle (signaux statiques). En b- l’expression du trait suit la qualité du mâle (signaux dynamiques). Notre première prédiction était que lorsque la disponibilité en nourriture varie rapidement (changement d’environnement) les traits comportementaux de parade devraient indiquer la qualité du mâle dans le nouvel environnement et non la qualité du mâle dans son précédent environnement. A l’inverse, les traits morphologiques devraient refléter la condition environnementale lors de leur production du fait de leur faible labilité. Notre seconde prédiction était que dans les environnements stables, le succès des mâles devrait être expliqué par les traits morphologiques dû à un choix de la femelle pour ces traits qui seraient alors plus informatifs tandis que dans les environnements fluctuants les traits comportementaux devraient être favorisés. En accord avec notre première prédiction, nos résultats montraient que les traits morphologiques étaient relativement fixes et pouvaient indiquer la qualité passée du mâle mais ne prédisaient pas sa qualité présente. Ceci était vrai pour des changements relativement brefs. Par contre sur des pas de temps plus importants, les traits morphologiques semblaient être plus flexibles et pourraient constituer de bons indicateurs de la qualité du mâle. Le comportement de parade "trembler" (quiver) était relativement flexible et influencé par la condition présente du mâle. Nous avions trouvé que les comportements de parade statique étaient constants dans le temps (répétables) dans des conditions stables de laboratoire (article 3). Contrairement à notre première prédiction, les comportements de parade statiques n’ont pas été impactés par la variation de la disponibilité en nourriture et semblaient donc relativement fixes. L’ensemble de ces éléments suggère que les traits comportementaux pourraient être fixés génétiquement et peu coûteux à produire. En accord avec notre deuxième prédiction, nous avons montré que les traits morphologiques étaient contre sélectionnés en milieu fluctuant, mais curieusement n’étaient pas sélectionnés en environnement stable. Les traits comportementaux étaient sélectionnés quelque soit l’environnement. Globalement nos résultats montrent que l’hétérogénéité de l’environnement peut avoir des impacts sur le choix du partenaire pour des signaux multiples en impactant la condition dépendance des traits. IV. Conclusion générale L’objectif général de cette thèse était l’étude des signaux multiples dans le cadre du choix du partenaire sexuel. Dans une première partie, nous avons réalisé une synthèse sur les signaux multiples dans le cadre du choix du partenaire. Cette synthèse nous a servi à établir nos hypothèses de travail (article 1). Dans une deuxième partie, nous avons étudié comment les ornements du mâle pouvaient influencer le choix des femelles chez le triton palmé (articles 2 et 3). Enfin dans une troisième partie, nous avons étudié le rôle des signaux multiples en tant que signaux de qualité individuelle (articles 4 et 5). IV. 1. Choix des femelles sur critères multiples Il apparait que, chez le triton palmé, l’accès à la reproduction est en grande partie expliqué par des traits morphologiques telle que la longueur du filament et des paramètres comportementaux tels que le temps passé en parade et le nombre de "trembler" (articles 2 et 5). Nos travaux de thèse contribuent à étayer l’hypothèse selon laquelle ces deux types de traits contribuent grandement à l’accès à la reproduction chez les tritons (Wells 2007). Bien que la parade semblait fixe pour un individu donné (article 3 mais voir article 5), elle n’était pas révélatrice de la qualité du mâle évaluée par la condition corporelle (article 4 et 5). Les expressions des traits morphologiques (longueur du filament, surface de palmure et surface de la crête) étaient positivement corrélées entre elles (articles 2, 4 et 5). De plus les traits morphologiques co-variaient positivement avec la qualité du mâle toujours mesurée par la condition corporelle (articles 4 et 5). Cela confirme l’hypothèse d’une condition dépendance des ornements chez les tritons en ajoutant un nouvel exemple avec une nouvelle espèce. En effet, il n’existait auparavant que deux exemples : le triton crêté Triturus cristatus (Baker 1992) et le triton vulgaire Lissotriton vulgaris (Green 1991). Globalement chez le triton palmé l’accès à la reproduction est expliqué par des traits multiples composés de traits morphologiques et comportementaux et seuls les traits morphologiques sont révélateurs de la condition du mâle (Figure 7). TRAIT INFORMATION crête palmure condition corporelle filament taille corporelle * ? coloration parade Figure 7. Association entre les traits potentiellement utilisés par les femelles pour choisir leurs partenaires sexuels et l’information contenue dans ces traits. Les flèches en tiret continu représentent les traits qui influencent le choix des femelles. Les flèches en tiret représentent les traits dont l’influence sur le choix n’est pas avérée. * D’après Haerty et al. (2007). Cette première conclusion de nos travaux montre que les signaux multiples peuvent être vus comme une mosaïque de traits regroupant à la fois des signaux de type non fiables, redondants et multiples, le tout au sein d’une même espèce (Figure 7, Doucet & Montgomerie 2003). Cela montre que les différentes hypothèses expliquant l’évolution des signaux multiples ne sont pas mutuellement exclusives (article 1). Les travaux sur les signaux multiples chez le triton palmé étant encore très minces, et cette thèse étant le début d’un projet sur les signaux multiples, il convient maintenant de les compléter par l’ajout d’autres signaux potentiellement sélectionnés par les femelles (développés plus en détail dans la partie perspective). Nos résultats peuvent également être interprétés dans le cadre du conflit sexuel. En effet les bénéfices divergent souvent entre les mâles et les femelles lorsqu’il s’agit de reproduction (Box 1). Alors que les mâles sont sélectionnés pour accéder aux femelles, les femelles sont sélectionnées pour choisir les meilleurs mâles (Font & Carazo 2010). Il est possible que la parade ait évolué comme moyen utilisé par les mâles pour accéder aux femelles alors que les ornements (comme le filament) aient évolué comme moyen utilisé par les femelles pour choisir le meilleur mâle (articles 2, 5, Figure 8). Bien que le conflit sexuel puissent expliquer l’existence de signaux multiples (Holland & Rice 1998, Lozano 2009) les preuves sont relativement minces (Candolin 2003). Chez les tritons la complexité de la parade semble avoir co-évolué avec le degré d’expression des ornements (Wiens et al. 2011). Cependant les traits comportementaux ne sont peut être pas sans lien avec la qualité du mâle (Figure 8). En effet il convient maintenant de tester quels autres types d’information peuvent être contenus dans les signaux (en dehors de la condition corporelle). La parade et les ornements pourraient être liés à la qualité immunitaire du mâle, les réserves en spermatophores, l’âge, les conditions durant l’ontogénèse… etc. (article 4). Cela permettrait de compléter la "carte du choix du partenaire" pour des signaux multiples élaborée au cours de cette thèse (Figure 7). Mâle « motivé »: une information sur la qualité ? Traits comportementaux « influencer la femelle », « manipulation » morphologiques « Extraction d’information » Figure 8. Interprétation de l’effet des traits comportementaux et morphologiques sur le choix des femelles en lien avec le conflit sexuel. Les colorations font parties des traits souvent utilisés par les femelles dans le cadre du choix du partenaire et elles sont largement utilisées dans le cadre des signaux multiples (article 1). L’examen d’un rôle de la coloration comme un critère du choix du partenaire chez les amphibiens connait d’ailleurs un certain succès ces dernières années (Rosenthal et al. 2004, Vasquez & Pfennig 2007, Maan & Cumming 2009, Gomez et al. 2009, 2010, 2011a,b, Doucet & Mennill 2010, Richardson et al. 2010, Sztatecnsy et al. 2010). Chez les amphibiens, la coloration orangée est généralement produite par des pigments caroténoïdiques (Richardson et al. 2009) ou des ptérines. Il est reconnu que ces deux pigments (caroténoïdes et ptérines) jouent un rôle important dans les fonctions immunitaires, et que leur fonction antioxydante permet une protection des cellules contre les radicaux-libres. Or, les ptérines sont synthétisées par l’organisme tandis que les animaux ne peuvent pas synthétiser les caroténoïdes et doivent donc les trouver dans leur alimentation. La disponibilité en caroténoïdes n’étant pas infinie, les mâles font face à un compromis entre une allocation pour des signaux colorés ou une allocation pour le bien être de leur organisme (Baeta et al. 2007). Il a donc été proposé que les mâles pouvant se permettre d’allouer une forte quantité de caroténoïdes dans leur trait sexuel devaient être des mâles de bonne qualité (Faivre et al. 2003). Dès lors, les femelles devraient baser leur choix sur des mâles très colorés (Hill 2006). Nos résultats montrent au contraire que la coloration n’est pas un critère important chez le triton palmé et ce résultat est confirmé par d’autres études (Secondi et al. 2012, Théry communication personnelle, article 2). Durant cette thèse nous avons réalisé une autre expérience, non présentée dans ce manuscrit, visant à comprendre plus finement le rôle potentiel de la coloration. Dans un dispositif tel que celui décrit dans l’article 2, nous avions proposé à des femelles le choix entre des mâles très orangé et des mâles blanchâtres. Les résultats de cette seconde expérience ne nous permettent pas de proposer une conclusion car aucun des protagonistes ne semblait réceptif, nous n’avons mesuré aucune prise de spermatophore et nous avons observé seulement un très faible nombre d’événements de parade. Même si nos premiers résultats ne suggèrent pas un rôle prédominant de la coloration (article 2), il conviendrait peut être de réaliser une expérience manipulant l’apport en caroténoïdes chez les mâles afin d’explorer l’impact de cette manipulation sur leur coloration et le choix des femelles. En effet, un apport d’antioxydants sous forme de caroténoïdes pourrait améliorer l’état de santé général des individus et ceci pourrait être perçu par les femelles, même en l’absence d’impact visible sur la coloration individuelle, par des indices olfactifs par exemple. IV. 2. Prise en compte de l’hétérogénéité de l’environnement La prise en compte de l’effet de l’hétérogénéité de l’environnement sur la sélection sexuelle est relativement récente, particulièrement dans le cadre des signaux multiples (Cornwallis & Uller 2010, Candolin & Wong 2012, article 1). Les conclusions développées dans le paragraphe précédent ainsi que la Figure 7 doivent être interprétées dans un cadre dynamique car le choix des femelles et la condition dépendance des traits sexuels peuvent varier aussi bien en direction qu’en intensité (articles 1, 2 et 5). Ainsi le choix des femelles semblaient potentiellement variable. Alors que les résultats des expériences présentées dans l’article 2 montrent que le succès des mâles dépendait grandement de la taille de leur filament, les résultats de l’expérience réalisée pour l’article 5 montrent que le filament ne prédisait pas ou prédisait inversement le succès des mâles. L’ensemble de ces éléments suggère que le choix des femelles peut varier spatialement et/ou temporellement. Ainsi, plusieurs paramètres de l’environnement pourraient causer cette variation, telles que la prédation, la disponibilité des mâles, la sex ratio opérationnelle ou la qualité de l’habitat (Ah-King 2010). Le choix du partenaire pourrait également varier suivant des facteurs internes propres à la femelle tels que sa taille, son statut parasitaire, sa condition, son âge ou son expérience passée (Ah-King 2010). Ces paramètres influant le choix du partenaire ont souvent été identifiés sur la base d’un critère unique, mais ont rarement été pris dans un contexte de signaux multiples (article 1). Dans cette thèse nous avons également montré que lors d’un changement rapide de la qualité de l’environnement (comme la disponibilité en nourriture), les traits morphologiques pouvaient perdre leurs liens avec la condition corporelle du mâle. Nous montrons ainsi, à travers les résultats présentés dans l’article 5, l’importance de l’environnement sur un élément phare en sélection sexuelle : la condition dépendance des traits sexuels. L’article 5 donne des résultats contrastés quant à l’effet de l’hétérogénéité de l’environnement sur la condition dépendance des traits sexuels. En effet sur une échelle aussi courte qu’une semaine, les ornements peuvent perdre leur lien avec la condition. Cependant sur une échelle plus longue d’un mois, il semble que le lien entre traits sexuels et qualité du mâle puissent être rétabli. Cela met en avant l’importance de l’échelle temporelle dans l’étude de la sélection sexuelle (Cornwallis & Uller 2010). Dans la continuité de l’article 5, il serait intéressant de reproduire une expérience similaire à celle que nous avons réalisée mais où les traitements expérimentaux viseraient à la fois les mâles et les femelles. Ainsi les femelles seraient également assignées à des environnements différents et variables (contrairement à l’article 5 où seuls les mâles subissaient le traitement expérimental). Cette expérience permettrait de comprendre en quoi l’hétérogénéité de l’environnement affecte également le choix des femelles. Les tritons vivent dans des habitats très différents au cours de l’année (terrestre puis aquatique). Ces habitats peuvent présenter des sources de stress et des pressions de sélection très variables pour les tritons en termes de qualité/quantité de nourriture et autres types de stress. Dans notre expérience, nous avons constaté que les conditions initiales (proches de la phase terrestre) pouvaient influencer l’expression des traits sexuels (article 5). Il serait dès lors très intéressant de tester l’effet de la phase terrestre ainsi que de la phase aquatique sur l’expression des traits sexuels, la condition dépendance des traits sexuels et le choix des femelles (Figure 9). Nourriture + Nourriture + Nourriture - Nourriture - Nourriture + Nourriture - Mesure du choix des femelles Phase aquatique Mesure des traits des mâles Phase terrestre Figure 9. Design expérimental pour l’étude de l’effet de l’hétérogénéité de l’environnement sur le choix des femelles et la condition dépendance des traits sexuels chez les mâles tritons palmés. IV. 3. Paradoxe du lek et conséquences sur les processus de sélection La prise en compte de l’effet de l’hétérogénéité de l’environnement sur le choix du partenaire peut avoir des conséquences importantes sur le maintien de la variation phénotypique et de la variabilité génétique observées dans la nature (Greenfield et al. 2012). Or, la variation sur un trait est une des trois conditions nécessaire pour qu’il y ait sélection sexuelle et donc évolution des ornements (introduction partie a). Le maintien de la variation est source de nombreux débats notamment concernant l’évolution de la préférence des femelles pour des bénéfices indirects et entourant le « paradoxe du lek » (Kokko et al. 2003). Le paradoxe du lek provient de la situation où les femelles choisissent leur(s) partenaire(s) uniquement sur la base d’un trait unique corrélé à la qualité génétique de ceux-ci, où les bénéfices des femelles sont indirects et où la force de préférence co-varie avec la quantité de variation génétique entre les mâles. Pour une espèce répondant à ces critères, plus la préférence des femelles est forte, plus le bénéfice est important en terme de fitness, mais plus la variation en terme de qualité génétique sera faible à la génération suivante (Kokko et al. 2003, Tomkins et al. 2004, Bussière et al. 2008). Il en découle que la sélectivité est diminuée au fur et à mesure des générations successives. Au final, la variation génétique qui favorise la préférence des femelles est diminuée par cette préférence (Tomkins et al. 2004) et devrait tendre à disparaître, d’où le paradoxe. La variation dans le choix des femelles tout comme la variation dans l’honnêteté du signal peut réintroduire de la variation et contribuer à résoudre le paradoxe (Bussière et al. 2008). En effet, selon l’environnement, les femelles vont être plus ou moins sélectives et donc choisir les mâles de plus ou moins bonne qualité (Greenfield & Rodriguez 2004). De plus, l’honnêteté des traits peut varier selon les environnements, ce qui conduit les femelles à plus ou moins se tromper dans la quête du mâle de meilleure qualité (Greenfield & Rodriguez 2004). En conséquence, l’hétérogénéité de l’environnement réduit la manière dont le choix des femelles peut éroder la variation. Il pourrait être intéressant d’examiner si les populations naturelles faisant face à des environnements fluctuants possèdent des variations génétiques plus importantes que les populations établies en milieu stable. Dans cette thèse, nous avons établi que l’honnêteté des traits peut être plus ou moins forte suivant les environnements et que les femelles ne choisissent pas forcement toujours les mâles de meilleure qualité (articles 1 et 5). Nous avons mesuré que, sur une échelle d’une semaine, les changements dans les conditions environnementales pouvaient impacter la qualité des mâles sans impacter les traits sexuels, diminuant ainsi la condition dépendance et donc l’honnêteté des ornements des mâles (article 5). En conséquence, nos expériences suggèrent que, chez le triton palmé, la variation peut être maintenue par les variations dans le choix des femelles ainsi que par les variations dans l’honnêteté des traits. Sur une échelle de temps d’un mois, il semble que l’environnement puisse moduler la qualité des mâles sans pour autant changer la condition dépendance des traits (article 5). En d’autres termes, même si les femelles utilisent toujours le même trait sexuel pour faire leur choix et que celui-ci est condition dépendant, la variation peut être maintenue car ce ne sont pas les mêmes individus qui sont sélectionnés selon l’environnement. IV. 4. Conséquences au niveau populationnel La prise en compte de l’effet de l’hétérogénéité de l’environnement sur le choix du partenaire peut avoir des conséquences au niveau populationnel. En effet, la variabilité du choix des femelles et la variabilité de l’honnêteté des ornements (article 5) montrent que l’on peut passer rapidement d’un modèle d’évolution de la préférence des femelles à un autre (Tableau 1). Par exemple, le fait qu’un trait puisse être condition dépendant dans un environnement et pas dans un autre (article 5) montre que l’on peut passer d’un modèle de trait dit "informatif" à un modèle de trait dit "non-informatif" (Tableau 1). Or, ces modèles ne reposent pas sur les mêmes hypothèses et n’ont pas non plus les mêmes conséquences au niveau populationnel (Jones & Ratterman 2009). Il est couramment admis que la sélection sexuelle a un coût pour la population car l’investissement dans les traits sexuels peut se faire au détriment de la survie (Doherty et al. 2003). En effet l’énergie investie dans les ornements ne peut être allouée pour d’autres traits et les ornements peuvent augmenter les risques de prédation (Andersson 1994). Cependant, suivant les modèles d’évolution de la préférence des femelles (Jones & Ratterman 2009), la sélection sexuelle pourrait aussi fournir des bénéfices au niveau populationnel (Candolin & Wong 2012). Les modèles reposant sur des traits dits informatifs peuvent avoir pour effet d’augmenter la fréquence des individus de bonne qualité à l’intérieur de la population, et ce par le choix des femelles, ce qui peut avoir pour conséquence d’augmenter la viabilité des populations (Møller & Alatalo 1999, Jarzebowska & Radwan 2010, Candolin & Wong 2012). De plus, le coût supposé des traits sexuels (notamment selon le principe du Handicap) et le choix des femelles peuvent permettre l’élimination des individus les moins bien adaptés à l’environnement, laissant plus de ressources aux individus restants (Candolin & Wong 2012). Dans nos expériences, les traits morphologiques semblent être de tels traits (article 4 et 5). Au contraire, les modèles basés sur des traits dits non-informatifs prédisent des conséquences différentes à l’échelle de la population car ils ne stipulent pas que les femelles choisissent les mâles de meilleure qualité. Les traits comportementaux chez le triton palmé semblent être non-informatifs (articles 4, 5). Certains modèles, tel que le conflit sexuel, (chapitre IV. 1) dans lesquels les mâles et les femelles ont des intérêts reproductifs divergents, où certains traits sexuels mâles imposent un coût accru aux femelles et où certains gènes sont favorables seulement à l’un des deux sexes pourraient même avoir des conséquences négatives sur la population (Candolin & Wong 2012). En effet, ce qui est favorable au niveau individuel ne l’est pas forcement pour la population. Or les mâles et les femelles sont sélectionnés pour maximiser leur propre succès reproducteur, et par exemple pour influencer la femelle contre le choix du meilleur mâle (Box 1, Figure 1 et 8). En favorisant l’utilisation de traits non-informatifs attractifs pour les femelles (dans notre cas les comportements de parade), la sélection sexuelle peut empêcher les femelles d’utiliser les meilleurs critères pour choisir leur partenaire (dans notre cas les traits morphologiques). Ceci pourrait conduire à une tragédie des communs où les bénéfices individuels conduisent à des conséquences tragiques pour la population (Rankin et al. 2011). Cependant les conséquences populationnelles de la sélection sexuelle restent encore sujettes à caution (Candolin & Heuschele 2008). En effet ces conséquences dépendent grandement des modèles d’évolution de la préférence des femelles dont les frontières semblent floues au vu de nos résultats et des récents travaux sur l’effet de l’hétérogénéité de l’environnement sur la sélection sexuelle (Cornwallis & Uller 2010, articles 1 et 5). IV. 5. Conséquences pour la biologie de la conservation Une question d’actualité en biologie de la conservation est l’adaptation des animaux aux changements environnementaux. Nos résultats montrent une certaine résistance des tritons aux pressions de l’environnement car même exposés à un environnement défavorable, les tritons semblent pouvoir se reproduire (article 5). Cependant, nos résultats montrent que des changements environnementaux peuvent affecter les processus de sélection sexuelle, notamment la fiabilité des traits (article 5). Cela pourrait rendre le choix du partenaire mal adaptatif (article 5) pouvant conduire à des problèmes au niveau populationnel (chapitre IV. 3 et IV. 4). Chez les épinoches à trois épines Gasterosteus aculeatus, la pollution entraîne une diminution de la qualité de l’eau qui elle-même engendre une visibilité réduite. Cela a pour effet de diminuer la compétition entre les mâles et l’honnêteté des signaux comportementaux (Wong et al. 2007). Pour qu’un choix qui était adaptatif dans un environnement donné continue d’être adaptatif dans un nouvel environnement, il faut que les femelles basent leurs choix sur un critère de la qualité du mâle qui demeure fiable dans ce nouvel environnement. Dans notre expérience le trait sélectionné par les femelles (i.e., le comportement de parade) ne semble pas indiquer la condition du mâle (article 2, 4 et 5). Un changement d’environnement pourrait donc avoir des conséquences négatives pour la population (Candolin & Wong 2012). Cependant une autre possibilité est que l’honnêteté du trait s’ajuste rapidement au nouvel environnement par plasticité. Par exemple, dans notre expérience, bien que la condition dépendance des traits morphologiques puisse diminuer suite à un changement d’environnement rapide, elle semble pouvoir se rétablir sur un temps un peu plus long (article 5). Cela contraste avec les résultats obtenus chez le mouton Ovis aries et chez le Gobemouche à collier Ficedula albicollis, où les ornements ne co-variaient pas avec la qualité du mâle (Robinson et al. 2007, 2012). Nos résultats demandent cependant à être confirmés par une étude à plus long terme (Figure 9) et également par des mesures en milieu naturel. Les maladies émergentes peuvent être comptées parmi les changements environnementaux qui impactent les populations (Daszak et al. 2000, Smith et al. 2009, Fisher et al. 2012). Chez les amphibiens, le champignon pathogène Batrachochytrium dendrobatidis est responsable d’une maladie infectieuse émergente, la chytridiomycose, susceptible de modifier la dynamique des populations (Fisher et al. 2012). Depuis sa découverte en 1998, la chytridiomycose a été identifiée comme l’une des causes du déclin mondial des populations d’amphibiens (Fisher et al. 2012). Les résultats de cette thèse peuvent apporter des pistes de réflexion importante sur ce sujet. Par exemple, nous montrons que le choix des femelles tritons est, sous certaines conditions, basé sur des traits reflétant la qualité du mâle (article 2, 4 et 5). Or, comme nous l’avons décrit précédemment le choix du partenaire pourrait agir comme un effet de purge en éliminant les génotypes mal-adaptés (chapitre IV. 4). Nos résultats laissent penser que les ornements reflètent la qualité du mâle chez le triton palmé (articles 4 et 5). Ils pourraient donc potentiellement refléter la résistance à la chytridiomycose. Les individus n’étant pas tous résistants de la même façon à la maladie (Kilpatrick et al. 2010), les génotypes résistants pourraient être favorisés par le choix des femelles ce qui peut avoir des effets positifs pour la population (Candolin & Heuschele 2010, Rankin et al. 2011). Cependant nous avons clairement montré que, chez cette espèce, le choix du partenaire est également basé sur des traits non informatifs (i.e., le comportement parade, articles 2 et 5), ce qui limiterait fortement un éventuel effet de purge (Candolin & Heuschele 2010). En effet, nos expériences ont révélé que les mâles de moins bonne qualité pouvaient accéder à la reproduction aussi bien que les mâles de meilleure qualité (article 5). Des expériences doivent donc maintenant établir le lien entre chytridiomycose, qualité des mâles, expression des traits sexuels et succès reproducteur. Nos résultats montrent également que l’hétérogénéité de l’environnement peut impacter la condition dépendance des traits et l’importance des critères morphologiques dans l’accès au femelles (article 5). Une conséquence est que l’hétérogénéité de l’environnement réduit la manière dont le choix des femelles peut éroder la variation (chapitre IV. 3). Or la variation est une composante essentielle de l’adaptation à des changements environnementaux par sélection naturelle. Dans le cadre de la chytridiomycose, il a été montré que les populations présentant une variation génétique plus importante étaient plus à même de résister à la maladie (Luquet et al. 2012). Il est donc possible que les populations faisant face à des environnements fluctuants résistent mieux à l’arrivée d’une maladie émergente telle que la chytridiomycose. Enfin, les amphibiens sont des espèces menacées notamment par les changements environnementaux et les des maladies émergentes (Stuart et al. 2004, Fisher et al. 2012). Une attention particulière doit donc être attribuée aux populations naturelles, notamment par un suivi précis de leurs abondances et de leurs qualités. Ceci peut être réalisé à travers des suivis de population ou des études d’impact lors de changement d’origine anthropique. Les résultats de cette thèse suggèrent à travers la condition dépendance des traits morphologiques (article 4 et 5) que des paramètres tels que les ornements peuvent être pris en compte pour juger de la qualité d’une population (Hill 1995). Par exemple, le suivi des ornements des mâles a été utilisé comme indicateur de la qualité des mâles dans le cadre de pollution lié aux marées noires (Pérez et al. 2010). Pérez et al. (2010) ont en effet mesuré que la pollution liée aux marées noires avait des impacts sublétaux sur les mâles pouvant être mesurés à travers l’expression des ornements. Chez les tritons palmés, il a été montré que l’expression des ornements était liée à la qualité de l’environnement en termes de turbidité (Secondi et al. 2007), concentrations en nitrates (Secondi et al. 2009) et disponibilité en ressources alimentaires (article 5). Des suivis, comprenant la mesure des ornements comme indicateur d’effet sublétal, pourraient être très utiles dans le cadre des études sur l’impact de la chytridiomycose sur les populations d’amphibiens. Ceci est également vrai pour les études d’impacts réalisées dans le cadre des suivis de population suite à des pressions d’origine humaine. IV. 6. Perspectives Il existe de nombreuses perspectives à cette thèse. Certaines perspectives ont déjà été développées dans les chapitres précédents. Par exemple il conviendrait de compléter la carte du choix des femelles en ajoutant des traits potentiellement préférés par les femelles et les informations potentiellement véhiculées par ces traits (Figure 7). En outre, il serait nécessaire de poursuivre le travail sur l’effet de l’hétérogénéité de l’environnement sur la sélection sexuelle (chapitre IV. 1 à 5, Figure 9). J’ai développé ici sous forme d’une liste quelques autres perspectives pouvant faire suite à ce travail de thèse. Les critères relatifs, telle que la compatibilité génétique, peuvent être extrêmement importants pour expliquer le choix des femelles. En effet, deux études montrent que les femelles exerceraient un choix pour des individus les moins apparentés chez le triton alpestre Ichthyosaura alpestris et le triton ponctué Lissotriton vulgaris (Garner & Schmidt 2003, Jehle et al. 2007). Or, nous ignorons ce qu’il en est chez le triton palmé. De plus, le rôle de l’information chimique semble important chez les urodèles et nous savons que la compatibilité génétique peut être évaluée en jugeant des odeurs (Denoël 1999, Aeschlimann et al. 2003, Milinski et al. 2005, Houck 2009). Un axe de recherche très intéressant pourrait être d’examiner l’existence d’une signature individuelle des composés chimiques des mâles. Une première étape pourrait être de réaliser un échantillonnage, sur différentes populations de tritons palmés, de composés chimiques afin de voir s’il existe un effet population, sexe, et pourquoi pas individuel sur la variation des composés chimiques. Cette technique prometteuse a déjà été utilisée sur des modèles oiseaux et mammifères (Bonadonna et al. 2007, Scordato et al. 2007) et pourrait apporter des réponses pertinentes quant aux choix des femelles sur des critères chimiques. Enfin, l’effet de l’information sociale sur le choix du partenaire n’a jamais été testé chez les amphibiens. Pourtant que ce soit pour les anoures formant des cœurs ou les urodèles tels que le triton palmé, les rencontres entre mâles et femelles se déroulent souvent sur des sites où les femelles ont l’opportunité d’observer le choix d’autres femelles et donc de collecter des informations sociales qui pourraient influencer leurs choix. Les théories sur le choix du partenaire font généralement le présupposé que les femelles sont capables de classer les différentes options présentées sur une échelle de valeur. Il est attendu que les femelles soient rationnelles dans leur choix. Or ce présupposé n’a fait l’objet que de peu d’études dans le contexte du choix du partenaire (Ryan et al. 2007). Le choix pour la taille du filament chez le triton palmé peut être un moyen de tester ce présupposé. Pour cela l’étude de la répétabilité, de la transitivité et de la répétabilité de la transitivité du choix peut être mise en place dans un design similaire à l’article 2. La théorie des signaux multiples, comme toute théorie évolutive, repose sur le caractère adaptatif (en terme de fitness) de l’utilisation des signaux multiples par rapport à l’utilisation d’un signal unique de communication (article 1). Or ce caractère adaptatif est rarement testé et démontré. Dans cette thèse, nous n’avons pas réalisé de mesure directe de la fitness (nombre de descendants viables et fertiles) de nos tritons. Ceci pourrait être réalisé avec les tritons, ou de manière plus réaliste avec des animaux à cycle de vie plus court. Un volet très important dans le cadre de la théorie des signaux multiples est l’effet de chaque signal pris indépendamment sur le choix de la femelle par rapport à l’effet de l’ensemble des signaux combinés (article 1). Nous nous sommes approchés d’une telle démarche dans l’article 2 mais les difficultés que nous avons rencontrées dans la manipulation expérimentale du phénotype des individus font qu’il est difficile de tester l’interaction des différents signaux sur le choix des femelles. Pourtant une telle approche est cruciale pour comprendre comment les femelles jugent les traits multiples tels que le filament, la crête, les palmures, la taille ou la parade. Dans cette thèse, nous avons souvent abordé la notion de qualité des individus. Comme nous l’avons dit, cette notion est relativement floue (Box 2). Dans cette thèse nous avons utilisé l’indice de condition corporelle comme proxy de la qualité des mâles. Cet indice semble d’ailleurs fiable comme indicateur de la disponibilité en ressource alimentaire (article 5). Bien que cet indice soit couramment utilisé notamment chez les amphibiens, il conviendrait de tester de manière plus approfondie si un mâle en meilleure condition corporelle est effectivement un mâle de meilleure qualité en termes de survie ou de fitness par exemple. De même il conviendrait de tester la co-variation entre les traits sexuels et la qualité du mâle en utilisant d’autre proxy de la qualité du mâle. Nous avons essayé d’utiliser l’activation du système immunitaire sans réussite (article 4). En effet nous ne pouvons pas affirmer que le système immunitaire de nos tritons a bien été affecté. De plus, nous avons utilisé un antigène sans grande réalité évolutive. L’utilisation direct d’un pathogène tels que la chytridiomycose, l’amphibiocystidium ou les bactéries responsables des phénomènes de « red leg » pourraient être une piste de recherche très intéressante. Bibliographie. Aeschlimann PB, Häberli MA, Reusch TBH, Boehm T, Milinski M. 2003. Female sticklebacks Gasterosteus aculeatus use self-reference to optimize MHC allele number during mate selection. Behavioral Ecology and Sociobiology 54:119−126. Ah-King M. 2010. Flexible mate choice. in Encyclopedia of Animal Behavior, Elsevier. Edited by Janice Moore and Michael D. Breed. Amat F, Orami N, Sanuy D. 2010. Body size, population size, and age structure of adult palmate newts (Lissotriton helveticus) in pyrenean lakes. Journal of Herpetology 44:313−319. Akre, KL, Ryan MJ. 2010. Complexity increases working memory for mating signals. Current Biology 20:502−505. Andersson M. 1994. Sexual Selection.PrincetonUniversity Press. Andersson S, Pryke SR, Örnborg J, Lawes MJ, Andersson M. 2002. Multiple receivers, multiple ornaments, and the trade-off between agonistic and epigamic signaling in a widowbird. American Naturalist, 5:684–691. Arntzen JW, M Sparreboom. 1989. A phylogeny for the Old World newts, genus Triturus: Biochemical and behavioural data. Journal of Zoololgy of London 219:645–649. Baeta R, Faivre B, Motreuil S, Gaillard M, Moreau J (2008) Carotenoid trade-off between parasitic resistance and sexual display: an experimental study in the blackbird (Turdus merula). Proceedings of the Royal Society B 275:427–434. Baird TA, Hranitz JM, Timanus DK, Schwartz AM. 2007. Behavioral attributes influence annual male mating success more than morphology in collared lizards. Behavioral Ecology 18:1146–1154. Baker JMR. 1992. Body condition and tail height in great crested newts, Triturus cristatus. Animal Behaviour 43:157–159. Basolo, AL. 1990. Female preference predates the evolution of the sword in the swordtail fish. Science, 250:808–810. Bateman, AJ. 1948. Intra-sexual selection in Drosophila. Heredity 2:349–368. Bateson P. 1983. Mate choice. Cambridge University Press. Bell AM, Hankison SJ, Laskowski KL. 2009. The repeatability of behaviour: a meta-analysis. Animal Behaviour, 77:771–783. Bergeron P, Baeta R, Pelletier F, Réale D, Garant D. 2011. Individual quality: tautology or biological reality. Journal of Animal Ecology, 80:361–364. Berglund A, Bisazza A, Pilastro A. 1996. Armaments and ornaments: an evolutionary explanation of traits of dual utility. Biological Journal of the Linnean Society, 58:385–399. Bonadonna F, Miguel E, Grosbois V, Jouventin P, Bessiere JM. 2007. Individual odor recognition in birds: An endogenous olfactory signature on petrels' feathers? Journal of Chemical Ecology 33:1819–1829. Bro-Jørgensen J. 2010. Dynamics of multiple signalling systems: animal communication in a world in flux. Trends in Ecology & Evolution, 25:292–300. Bussière LF, Hunt J, Stölting KN, Jennions MD, Brooks R. 2008. Mate choice for genetic quality when environments vary: suggestions for empirical progress. Genetica, 134:69–78. Cameron E, Day T, Rowe L. 2003. Sexual conflict and indirect benefits. Journal of Evolutionnary Biology 16:1055–1060. Candolin U, Heuschele J. 2008. Is sexual selection beneficial during adaptation to environmental change? Trends in Ecology and Evolution, 23:446–452. Candolin U, Wong BBM. 2012. Sexual selection in changing environments: consequences for individuals and populations. pp201-215. In Behavioural responses to a changing world mechanisms and consequences. Edited by Candolin U, Wong BBM; Oxford University Press, Oxford. Candolin, U. 2003. The use of multiple cues in mate choice. Biological Reviews, 78:575–595. ChaineAS, Lyon BE. 2008. Intrasexual selection on multiple plumage ornaments in the lark bunting. Animal Behaviour, 76: 657–667. ChaineAS, Tjernell KA, Shizuka D, Lyon BE. 2011. Sparrows use multiple status signals in winter social flocks. Animal Behaviour, 81: 447–453. Clutton-Brock T. 2009. Sexual selection in females. Animal Behaviour, 77:3–11. Cornwallis CK, Uller T. 2010. Towards an evolutionary ecology of sexual traits. Trends in Ecology and Evolution, 25:145–152 Cotton S, Fowler K, Pomiankowski A. 2004. Do sexual ornaments demonstrate heightened condition-dependent expression as predict by the handicap hypothesis? Proceedings of the Royal Society B 271:771–783. Cotton S, Small J, Pomiankowski A. 2006. Sexual selection and condition dependent mate preferences. Current Biology, 16:755–765. Danchin E, Charmantier A, Champagne FA, Mesoudi A, Pujol B, Blanchet S. 2011. Beyond DNA: integrating inclusive inheritance into an extended theory of evolution. Nature Review Genetics. 12:475–486. Danchin E, Giraldeau L-A, Cézilly F. 2008. Behavioral Ecology. Oxford. Darwin C. 1859. On the origine of species by means of natural selection, or the preservation of favoured races in the struggle for life. John Murray. Darwin C. 1871. The descent of man and selection in relation to sex. John Murray. Daszak P, Cunningham AA, Hyatt AD. 2000. Emerging infectious diseases of wildlife threats to biodiversity and human health. Science, 287:443–449. Demary KC, Lewis SM. 2007. Male courtship attractiveness and paternity success in Photinusgreeni fireflies. Evolution 61:431–439. Denoël M. 1999. Le comportement social des urodèles. Cahiers d’Ethologie 19:21–258. Denoël M, Lehmann A. 2006. Multi-scale effect of landscape processes and habitat quality on newt abundance: Implications for conservation. Biological Conservation, 130:495–504. Denoël M, Dollen J. 2010. Displaying in the dark : light-dependent alternative mating tactics in the Alpine newt. Behavioral Ecology and Sociobiology 64:1171–1177. Doherty PFJ, Sorci G, Andrew Royle J, Hines JE, Nichols JD, Boulinier T. 2003. Sexual selection affects local extinction and turnover in bird communities. Proceedings of the National Academy of Sciences, 100, 5858–5862. van Dongen WFD, Mulder RA. 2008. Male and female golden whistlers respond differently to static and dynamic signals of male intruders. Behavioral Ecology 19:1025–1033. Doucet SM, Montgomerie R. 2003. Multiple sexual ornaments in satin bowerbirds: ultraviolet plumage and bowers signal different aspects of male quality. Behavioral Ecology 14:503–509 Doucet SM, Mennill DJ. 2010. Dynamic sexual dichromatism in an explosively breeding Neotropical toad. Biology Letters 6:63–66. Faivre B, Grégoire A, Préault M, Cézilly F, Sorci G. 2003. Immune activation rapidly mirrored in a secondary sexual trait. Science 300:103. Fisher RA. 1915. The evolution of sexual preference. Eugenics Review 7:184–192. Fisher RA. 1930. The Genetical Theory of Natural Selection. (Clarendon Press, Oxford). Fisher MC, Henk DA, Briggs CJ, Brownstein JS, Madoff LC, McCraw SL, Gurr SJ. 2012. Emerging fungal threats to animal, plant and ecosystem health. Nature, 484:186–194. Font E, Carazo P. 2010. Animals in translation: why there is meaning (but probably no message) in animal communication. Animal Behaviour, 79:e1–e6. Fuller RC, Houle D, Travis J. 2005. Sensory bias as an explanation for the evolution of mate preferences. American Naturalist 166:437–446. Gabor CR, Halliday TR. 1997. Sequential mate choice by multiply mating smooth newts: females become more choosy. Behavioral Ecology 8:162–166. Garner TWJ, Schmidt BR. 2003. Relatedness, body size and paternity in the alpine newt, Triturus alpestris. Proceedings of the Royal Society B 270:619–624. Gomez D, Richardson C, Lengagne T, Plénet S, Joly P, Léna JP, Théry M. 2009. The role of noctural vision in mate choice : females prefer conspicuous males in the European tree frog (Hyla arborea). Proceedings of the Royal Society B 276:2351–2358. Gomez D, Richardson C, Lengagne T, Derex M, Plenet S, Joly P, Léna JP, Théry M. 2010. Support for a role of colour vision in mate choice in Hyla arborea. Behaviour 147:1753– 1768. Gomez D, Théry M, Gauthier AL, Lengagne T. 2011a. Costly help of audiovisual bimodality for female mate choice in a nocturnal anuran (Hyla arborea). Behavioral Ecology, 22:889– 898. Gomez D, Richardson C, Théry M, Lengagne T, Léna JP, Plénet S, Joly P, 2011b. Multimodal signals in male European treefrog (Hyla arborea) and the influence of population isolation on signal expression. Biological Journal of the Linnean Society 103:633–647. Green AJ. 1991. Large male crests, an honest indicator of condition, are preferred by female smooth newts, Triturus vulgaris (Salamandridae) at the sperm mass transfer stage. Animal Behaviour 41:367–369. GreenfieldMD, Danka RG, Gleason JM, Harris BR, Zhou Y. 2012. Genotype x environment interaction, environmental heterogeneity and the lek paradox. Journal of Evolutionary Biology, 25:601–613. Greenfield MD, Rodriguez RL. 2004. Genotype-environment interactions and the reliability of mating signals. Animal Behaviour 68:1461–1468. Griffiths RA, Mylotte VJ. 1988. Observations on the development of the secondary sexual characters of male newts, Triturus vulgaris and Triturushelveticus. Journal of Herpetology, 22:476–480. Guilford T, Dawkins MS. 1991. Receiver psychology and the evolution of animal signals. Animal Behaviour, 42:1–14. Haerty W, Gentilhomme E, Secondi J. 2007. Female preference for a male sexual trait uncorrelated with male body size in the palmate newt (Triturushelveticus). Behaviour 144:797–814. Hagelin JC, Ligon JD. 2001. Female quail prefer testosterone-mediated traits, rather than the ornate plumage of males. Animal Behaviour 61:465–476. Halliday TR. 1975. On the biological significance of certain morphological characters in males of the Smooth newts Triturusvulgaris and of the Palmate newt Triturushelveticus (Urodela: Salamandridae). Zoological Journal of Linnean Society, 56:291–300. Halliday TR. 1977. The courtship of European newts. An evoutionary perspective. In: The reproductive biology of amphibians. (eds. Taylor DH, Guttman SI). Plenum, New York. Hebets EA, Papaj DR. 2005. Complex signal function: developing a framework of testable hypotheses. Behavioral Ecology and Sociobiology, 57:197–214. Heisler, I. L., Andersson, M. B., Arnold, S. J., Boake, C. R., Borgia, G., Hausfater, G., Kirkpatrick, M., Lande, R., Maynard Smith, J., O’Donald, P., Thornhill, A. R. &Weissing, F. Bradbury JW, Andersson MB. 1987. Sexual Selection: Testing the Alternatives. John Wiley. Hill GE. 1995. Ornamental traits as indicators of environmental health. BioScience, 45:25–31. Hill GE. 2006. Female mate choice for ornamental coloration. In: Bird coloration. Function and evolution (Ed. by Hill, GE & McGraw KJ). London: Harvard University Press. Holland B, Rice WR. 1998. Perspective: Chase-away sexual selection: Antagonistic seduction versus resistance. Evolution 52:1–7. Houck LD (2009) Pheromone communication in amphibians and reptiles. Annual Review of Physiology 71:161–176 Jacob S, Rieucau G, Heeb P. 2011. Multimodal begging signals reflect independent indices of nesting condition in European starlings. Behavioral Ecology, 22:1249–1255. Jarzebowska M, Radwan J. 2010. Sexual selection counteracts extinction of small populations of the bulb mites. Evolution, 64:1283–1289. Jehle R, Sztatecsny M, Wolf JBW, Whitlock A, Hödl W, Burke T. 2007. Genetic dissimilarity predicts paternity in the smooth newt (Lissotriton vulgaris). Biology Letters 3:526–528. Jennions MD, Petrie M. 1997. Variation in mate choice: a review. Biological Reviews, 72:283–327. Jennions MD, Petrie M. 2000. Why do females mate multiply ? A review of the genetic benefits. Biological Reviews, 75:21–64. Jones AG, Ratterman NL. 2009. Mate choice and sexual selection : what have we learned since Darwin ? Proceedings of the National Academy of Sciences, 106:10001–10008. Kilpatrick AM, Briggs CJ, Daszak P. 2010. The ecology and impact of chytridiomycosis: an emerging disease of amphibians. Trends in Ecology and Evolution, 25:109–118. Kirkpatrick M. 1982. Sexual selection and the evolution of female choice. Evolution 36:1–12. Kodric-Brown A, Nicoletto PF. 2001. Female choice in the guppy (Poeciliareticulata): the interaction between male color and display. Behavioral Ecology and Sociobiology 50:346– 351. Kokko H, Brooks R, Jennions MD, Morley J. 2003. The evolution of mate choice and mating biases. Proceedings of the Royal Society of London B, 270:653–664. Kokko H, Brooks R, Jennions MD, Morley J. 2003. The evolution of mate choice and mating biases. Proceedings of the Royal Society B, 270:653–664. Künzler R, Bakker TCM. 2001. Female preferences for single and combined traits in computer animated stickleback males. Behavioral Ecology 12:681–685. Lailvaux SP, Kasumovic MM. 2011. Defining individual quality over lifetimes and selective contexts. Proceedings of the Royal Society B, 278:321–328. Lande R. 1981. Models of speciation by sexual selection on polygenic traits. Proceedings of the National Academic of Science USA 78:3721–3725. Lehtonen TK, Svensson PA, Wong BBM. 2011. Both male and female identity influence variation in male signaling effort. BMC Evolutionary Biology, 11:233. LozanoGA. 2009. Multiple cues in mate selection: The sexual interference hypothesis. Bioscience Hypothesis, 2:37–42. Luquet E, Garner TWJ, Léna J-P, Bruel C, Joly P, Lengagne T, Grolet O, Plénet S. 2012. Genetic erosion in wild populations makes resistance to a pathogen more costly. Evolution, 66:1942–1952. Lyon BE, Montgomerie R. 2012. Sexual selection is a form of social selection. Philosophical Transactions of the Royal Society B, 367:2266–2273. Maan ME, Cummings ME. 2009. Sexual dimorphism and directional sexual selection on aposematic signals in a poison frog. Proceedings of the National Academy of Science USA 106:19072–19077. Maynard Smith J, Harper D. Animal signals. Oxford University Press. Mays HL, Hill GE. 2004. Choosing mates: good genes versus genes that are a good fit. Trends in Ecology & Evolution, 19:554–559. Miaud C. 1990. La dynamique des populations subdivisées: étude comparative chez trios Amphibiens Urodèles (Triturus alpestris, T. helveticus, T. Cristatus). Thèse, Université Claude Bernard, Lyon, France. Michalak P. 1996. Repeatability of mating behaviour in Montandon’s newt, Triturus montandoni (Caudata Salamandridae). Ethology Ecology & Evolution 8:19–27. Milinski M, Griffiths S, Wegner KM, Reusch TBH, Haas-Assenbaum A, Boehm T. 2005. Mate choice decisions of stickleback females predictably modified by MHC peptide ligands. Proceedings of the National Academy of Sciences USA 102:4414–4418. Møller AP, Alatalo RV. 1999. Good-genes effects in sexual selection. Proceedings of the Royal Society of London B, 266:85–91. Møller AP, Pomiankowski A. 1993. Why have birds got multiple sexual ornaments? Behavioral Ecology and Sociobiology, 32:167–176. Nakatsuru K, Kramer DL. 1982. Is sperm cheap? Limited male fertility and female choice in the Lemon Tetra (Pices, Characidae). Science 216:753–755. Parker GA. 2006. Sexual conflict over mating and fertilization: An overview. Philosophical Transactions of the Royal Society of London, Serie B, 361:235–259. Partan SR, Marler P. 2005. Issues in the classification of multimodal communication signals. American Naturalist 166:231–245. Partan SR, Marler, P. 1999. Communication Goes Multimodal. Science 283:1272–1273. Poole KG, Murphy CG. 2007. Preferences of female barking treefrogs, Hylagratiosa, for larger males: univariate and composite tests. Animal Behaviour, 73:513−524. Pérez C, Munilla I, Lόpez-Alonso M, Velando A. 2010. Sublethal effects on seabirds after the Prestige oil-spill are mirrored in sexual signals. Biology Letters, 6:33−35. Pryke SR, Andersson S, Lawes MJ. 2001. Sexual selection of multiple handicaps in the redcollared widowbird: female choice of tail length but not carotenoid display. Evolution, 55:1452–1463. Rankin DJ, Dieckmann U, Kokko H. 2011. Sexual conflict and tragedy of the commons. American Naturalist 177:780–190. Richardson C, Popovici J, Bellvert F, Lengagne T. 2009. Conspicuous coloration of the vocal sac of a nocturnal chorusing treefrog : carotenoid-based ? Amphibia-Reptilia, 30:576-580. Richardson C, Gomez D, Durieux R, Théry M, Joly P, Léna JP, Plénet S, Lengagne T. 2010. Hearing is not necessarily believing in nocturnal anurans. Biology Letters 23:633−635. Richardson C, Lengagne T. 2010. Multiple signals and male spacing affect female préférence at cocktail parties in treefrogs. Proceedings of the Royal Society of London B, 277:1247−1252. Robinson MR, Pilkington JG, Clutton-Brock TH, Pemberton JM, KruukLEB. 2007. Environmental heterogeneity generates fluctuating selection on a secondary sexual trait. Current Biology, 18:751–757. Robinson MR, van Doorn GS, Gustafsson L, Qvarnström A. 2012. Environment-dependent selection on mate choice in a natural population of birds. Ecology Letters, 15:611–618. Rodd HF, Hughes KA, Grether GF, Baril, CT.2002. A possible non-sexual origin of mate preference: are male guppies mimicking fruit? Proceedings of the Royal Society of London B, 269:475–481. Rosenthal GG, Rand AS, Ryan MJ. 2004. The vocal sac as a visual cue in anuran communication: an experimental analysis using video playback. Animal Behaviour 68:55–58. Rowe LV, Evans MR, Buchanan KL. 2001. The function and evolution of the tail streamer in hirundines. Behavioral Ecology, 12:157–163. Ruiz M, Davis E, Martins EP. 2008. Courtship attention in sagebrush lizards varies with male identity and female reproductive state. Behavioral Ecology, 19:1326–1332. Ryan MJ, Rand AS. 1990. The sensory basis of sexual selection for complex calls in the tungara frog, PhysalaemusPustolosus (sexual selection for sensory exploitation). Evolution, 44:305–314. Ryan MJ, Akre KL, Kirkpatrick M. 2007. Mate choice. Current Biology 17:R313–R316. Scheuber H, Jacot A, Brinkhof MWG. 2003a. Condition dependence of a multicomponent sexual signal in the field cricket Gryllus campestris. Animal Behaviour 65:721–727. Scheuber H, Jacot A, Brinkhof MWG. 2003b. The effect of past condition on a multicomponent sexual signal. Proceedings of the Royal Society of London B 270:1779– 1784. Scheuber H, Jacot A, Brinkhof MWG. 2004. Female preference for multiple condition– dependent components of a sexually selected signal. Proceedings of the Royal Society of London B 271:2453–2457. Schlichting CD, Pigliucci M. 1998. Phenotypic evolution: a reaction norm perspective. Sunderland, Massachusetts:Sinauer. Schuett W, Trengenza T, Dall, SRX. 2010. Sexual selection and animal personality. Biological Reviews, 85:217–246. Scordato ES, Dubay G, Drea CM. 2007. Chemical composition of scent marks in the ringtailed lemur (Lemur catta): Glandular differences, seasonal variation, and individual signatures. Chemical Senses 32:493–504. Scordato ESC, Bontrager AL, Price TD. 2012. Cross-generational effects of climate change on expression of a sexually selected trait. Current Biology, 22:78–82. Secondi J, Aumjaud A, Pays O, Boyer S, Montembault D, Violleau D. 2007. Water turbidity affects the development of sexual morphology in the palmate newt. Ethology, 113:711–720. Secondi J, Hinot E, Djalout Z, Sourice S, Jadas-Hécart A. Realistic nitrate concentration alters the expression of sexual traits and olfactory male attractiveness in newts. Functional Ecology, 23:800–808. Secondi J, Lepetz V, Théry M. 2012. Male attractiveness is influenced by UV wavelengths in a newt species but not in its close relative. Plos One, 7:e30391. Shamble PS, Wilgers DJ, Swoboda KA, Hebets EA. 2009. Courtship effort is a better predictor of mating success than ornamentation for male wolf spiders. Behavioral Ecology 20:1242–1251. Smith, KF, Acevedo-Whitehouse K, Pedersen AB. 2009. The role of infectious diseases in biological conservation. Animal Conservation, 12:1–12. Stuart SN, Chanson JS, Cox NA, Young BE, Rodrigues ASL, Fischman DL, Waller RW. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783– 1786. Suk HY, Choe JC. Dynamic female preference for multiple signals in Rhinogobius brunneus. Behavioral Ecology and Sociobiology 62:945–951. Sztatecsny M, Strondl C, Baierl A, Ries C, Hodl W. 2010. Chin up: are the bright throats of male common frogs a condition-independent visual cue? Animal Behaviour 79:779–786. Taylor RC, Klein BA, Ryan MJ. 2011. Inter-signal interaction and uncertain information in anuran multimodal signals. Current Zoology 57:153−161. Tomkins JL, Radwan J, Kotiaho JS, Tregenza T. 2004. Genic capture and resolving the lek paradox. Trends in Ecology and Evolution, 19:323–328. Trivers, RL. 1972. Parental investment and sexual selection. In Sexual Selection and the Descent of Man 1871–1971 (ed. B. Campbell), pp. 136–179. Heinemann, London. Vasquez T, Pfennig KS. 2007. Looking on the bright side: females prefer coloration indicative of male size and condition in the sexually dichromatic spade foot toad, Scaphiopus couchii. Behavioral Ecology and Sociobiology 62:127–135. Velando A, Beamonte-Barrientos R, Torres R. 2006. Pigment-based skin colour in the bluefooted booby: an honest signal of current condition used by females to adjust reproductive investment. Oecologia, 149:535–545. Vinnedge B, Verrell P. 1998. Variance in male mating success and female choice for persuasive courtship displays. Animal Behaviour 56:443–448. Wagner WE. 1998. Measuring female mating preferences. Animal Behaviour, 55 :1029–1042. Ward JL, McLennan DA. 2009. Mate choice based on complex visual signals in the brook stickleback, Culaeainconstans. Behavioral Ecology20:1323–1333. Wedekind C. 2002. Sexual selection and life-history decisions: implications for supportive breeding and the management of captive populations. Conservation Biology, 16:1204–1211. Welch AM, Raymond DS, Gerhardt HC. 1998. Call duration as an indicator of genetic quality in male gray tree frogs. Science, 280:1928–1930. Wells KD. 2007. The ecology and behavior of amphibians. The University of Chicago Press. Whattam EM, Bertram SM. 2011. Effect of juvenile and adult condition on long-distance call components in the Jamaican field cricket, Gryllus assimilis. Animal Behaviour, 81:135–144. Wiens JJ, Sparreboom M, Arntzen JW. 2011. Crest evolution in newts: implications for reconstruction methods, sexual selection, phenotypic plasticity and the origin of novelties. Journal of Evolutionary Biology, 24:2073–2086. Wilson AJ, Nussey DH. 2010. What is individual quality? An evolutionary perspective. Trends in Ecology & Evolution, 25:207–214. Witte K, Ryan MJ. 1998. Male body length influences mate-choice copying in the sailfin molly Poecilialatipinna. Behavioral Ecology, 9:534–539. Witte K, Ryan MJ. 2002. Mate choice copying in the sailfin molly, Poecilialatipinna, in the wild. Animal Behaviour, 63:943–949. Wong BBM, Candolin U, Lindström K. 2007. Environmental deterioration compromises socially enforced signals of male quality in three-spined sticklebacks. American Naturalist, 170:184–189. Zahavi A. 1975. Mate selection: A selection for a handicap. Journal of Theoretical Biology 53:205–214. ARTICLES Article 1……………………………………………………………………………….. Cornuau JH, Rat M, Schmeller DS, Loyau A. Complex love decisions: why and how animals use several signals in mate choice. En préparation. Article 2………………………………………………………………………………… Cornuau JH, Rat M, Schmeller DS, Loyau A. 2012. Multiple signals in the palmate newt: ornaments help when courting. Behavioral Ecology and Sociobiology, 66:1045-1055. Article 3………………………………………………………………………………… Cornuau JH, Courtoix EA, Jolly T, Schmeller DS, Loyau A. It takes two to tango : Do both male and female identities influence complex courtship display in a newt species? En préparation. Article 4………………………………………………………………………………… Cornuau JH, Pigeault R, Schmeller DS, Loyau A. Condition-dependance of multiple sexual traits in the palmate newt (Lissotriton helveticus). En préparation. Article 5………………………………………………………………………………… Cornuau JH, Sibeaux A, Tourat A, Schmeller DS, Loyau A. Impacts of varying environmental conditions on sexual traits and consequences for mate choice in the Palmate newt. En préparation. ARTICLE 1 Complex love decisions: why and how animals use several signals in mate choice. Cornuau JH, Rat M, Schmeller DS, Loyau A. En préparation 1 Complex love decisions: why and how animals use several signals in mate choice? 2 3 Jérémie H. Cornuaua,*, Margaux Rata,b, Dirk S. Schmellera,c, Adeline Loyaua,c 4 5 a Station d’Ecologie Expérimentale du CNRS à Moulis, 09200 Saint Girons, France. 6 b Percy FitzPatrick Institute, DST/NRF centre of Excellence, University of Cape Town, 7 7701 Rondebosch, South Africa. 8 c 9 Biology, Permoser Strasse 15, 04138 Leipzig, Germany, both senior authors Helmholtz Center of Environmental Research – UFZ, Department of Conservation 10 11 * Correspondence: J.H. Cornuau, Station d’Ecologie Expérimentale du CNRS à Moulis, 12 09200 Saint Girons, France. 13 E-mail address: [email protected] (J.H. Cornuau) 14 15 Short title: Multiple signals and mate choice 16 1 17 Abstract: 18 Ten years after the first review about the multiple signal theory, this area of research 19 continues to be flourishing. Here we proposed to revisit and add current research on the 20 different hypotheses that explain the used of multiple signals focusing on mate choice 21 decision. We first proposed that compatibility and social information need to be more 22 included in the multiple signal theory. Then, we used recent examples that now give 23 experimental support to previous hypothesis, develop in previous reviews, in order to 24 understand why and how female used multiple signals to make their choice. Afterwards, 25 we highlighted current research, particularly the influence of heterogeneous environment 26 to explain the used of multiples signals in mate choice. We particularly focus on the 27 necessity to take account of both external and internal environment to explain why and 28 how female make their choice. Next we pointed conceptual issue which make different 29 hypothesis little distinguishable. We showed with example that the previous 30 categorizations of the hypothesis that explain the used of multiple signals are often more 31 blurred than previously described. Finally we discussed the necessity to incorporated 32 recent advance in cognitive ecology in the multiple signals theory. Indeed we think that 33 the multiple signals theory cannot be fully understood without a description of their 34 cognitive consequence. Globally this review, in supplement to previous review, provide a 35 comprehensible framework for scientist that want investigate why and how animal used 36 multiples signals in mate choice decision. 37 38 39 40 Keywords: multiple signals, complex signals, multimodal, mate choice, animal 41 communication, information processing. 2 42 Introduction 43 Despite considerable advances over the last thirty years, the questions of why and 44 how animals make decisions about with whom to mate continue to puzzle behavioural 45 ecologists. Mate choice represents any pattern of behaviour, shown by members of one 46 sex, which leads to their tendency to mate non-randomly with respect to one or more 47 varying traits in members of the opposite sex (Halliday 1983, Heisler et al. 1987, Fig. 1). 48 As any other decision mechanism, mate choice involves costs in terms of signal 49 assessment, sorting and processing (Mery & Kawecki 2005). If costs associated to each 50 signal are additive (Roberts et al. 2007), overall costs may be higher when mate choice 51 involves multiple signals. In this case, selection should favour only one signal providing 52 the highest fitness benefits, i.e. an honest signal with a low assessment cost (Schluter and 53 Price 1993; Iwasa and Pomiankowski 1994, Fawcett & Johnstone 2003, but see 54 Johnstone 1995). However, it is now well established that females often base their choice 55 on multiple signals, and theories explaining the concomitant use of multiple signals for 56 mate choice are continuing to flourish (Guilford & Dawkins 1991, Møller & 57 Pomiankowski 1993, Rowe 1999, Candolin 2003, Hebets and Papaj 2005, Partan & 58 Marler 2005, Lozano 2009, Bro-Jørgensen 2010). These theories provide a strong 59 evolutionary framework allowing a better understanding of sexual preferences. They may 60 even contribute to resolve ongoing discussions about the lack of empirical evidence 61 and/or the apparent discrepancies sometimes found in studies of mate choice 62 (Roughgarden & Akçay 2010, Carranza 2010, Clutton-Brock 2010, Shuker 2010), 63 because where ones may see inconsistencies between studies, others see spatial and 64 temporal flexibility in mate choice in which a single, but different, signal is used each 65 year (Loyau et al. 2008, Chaine & Lyon 2008, Bro-Jørgensen 2010). 3 66 In this review, we propose to examine and discuss recent research on the issue of 67 multiple signalling in the context of mate choice. First, we point out the importance of 68 recently identified cues of mate choice, such as cues containing social information, and 69 we propose that social cues should be included in multiple signalling theories as any 70 other kind of signal. Second, we review different hypotheses put forward to explain the 71 evolutionary pressures underlying the evolution and maintenance of multiple signals in 72 mate choice. We particularly focused on recent theory that includes dynamic process of 73 selection. Third, we integrate multiple signals theory in the light of cognitive ecology. 74 Our aim is to provide a general and integrative framework for research on multiple 75 signals theory in the context of mate choice. For convenience, we will mostly refer to 76 female choice for male traits or signals, although mate choice occurs in both males and 77 females (Clutton-Brock 2007). 78 79 SEXUAL SIGNALS USED IN MATE CHOICE 80 Since Darwin (1871), considerable efforts have been invested to identify traits and 81 signals used for mate choice and to understand how these traits and signals have been 82 selected and maintained by evolution (Andersson 1994, Gibson & Langen 1996, 83 Andersson & Simmons 2006). Various types of traits have been recognized, including (i) 84 morphological ornaments such as an elongated tail or vivid colours, (ii) behavioural traits 85 such as elaborate courtship displays or songs, and (iii) extended phenotypes like nest 86 construction (Andersson 1994, Schaedelin & Taborsky 2009, Byers et al. 2010, Cornuau 87 et al. 2012). Note that each of these traits may be a multi-component trait that can be 88 further unbundled into multiple signals, even though this level of complexity has been 89 poorly studied so far (examples for coloration: Grether et al. 2004, extended phenotypes: 90 Schaedelin & Taborsky 2009). Morphological traits, behavioural traits and extended 4 91 phenotypes are commonly included in the multiple signals framework, whereas choice 92 for mate compatibility (or heterozygosity) and social information are less considered. 93 Although, the consideration of relatedness, genetic heterozygosity and/or compatibility 94 have recently increased, as witnessed by several theoretical reviews (Zeh & Zeh 2003, 95 Neff & Pitcher 2005, Puurtinen et al. 2005, Mays et al. 2008, Roberts & Little 2008, 96 Puurtinen et al. 2009) the integration of this cue in a multiple cue framework has rarely 97 been done (Mays & Hill 2004). Only few studies of mate choice investigated the impact 98 of several signals including compatibility cues on mate choice decision (Mays & Hill 99 2004). These studies focus particularly on the potential conflict between choice for 100 compatible partner or choice for the most ornamented partner. Indeed it is expected that 101 the most compatible partner is on average not the most ornamented partner (Mays & Hill 102 2004). Thus it’s expected that animal should trade-off between this two traits. For 103 example in mice, although both mate choice for genetic compatibility and ornamental 104 quality occurs, genetic quality seems prefer only when available partner were similar in 105 ornament (Roberts & Gosling 2003). In the cichlid Pelvicachromis taeniatus females give 106 more importance to genetic compatibility than in ornament whereas the opposite was 107 found for males (Thünken et al. 2012). These two experiments show how compatibility 108 can be included in multiple signals theory. 109 110 Individuals can also gather social information by observing conspecifics interacting 111 with each others and use this information to make more accurate adaptive decisions 112 (McGregor 1993, Otter et al. 1999, Mennill et al. 2002, Danchin et al. 2004, Valone 113 2007). The use of social information in the context of mate choice can be broadly divided 114 into “eavesdropping” and “mate choice copying”. Females can “eavesdrop” on male-male 115 competitive interactions (such as song contests) to assess male quality and compare 5 116 potential partners (McGregor 1993, Otter et al. 1999, Mennill et al. 2002, Desjardins et 117 al. 2010). Females can also observe and then “copy” the choice of other females for a 118 given male or phenotype (Valone 2007). It is thought that copying or eavesdropper 119 individuals obtain information more easily and at lower costs, as it reduces the time they 120 spend looking for a mate (Templeton & Giraldeau 1996). Only few studies of mate 121 choice investigated the impact of several signals including social cues (Dugatkin 1996, 122 Dugatkin 1998). It is still unknown whether individuals would favour one type of 123 information (e.g. collected through eavesdropping) at the expense of another one (e.g. 124 observation of mate choice) and how choosy individuals deal with receiving 125 contradictory social information. So far, most research on social cues which considered 126 several signals, dealt with the respective roles played by private (non-social) versus 127 public (social) cues. For example, in the zebra finch Taeniopygia guttata, private 128 information (mate attractiveness manipulated using tarsus-band colours) can offset mate 129 choice copying (Drullion & Dubois 2008). This example demonstrates that social cues 130 should definitively be considered into multiple signals theories. The authors further stress 131 the importance of neural mechanisms and brain functions implied by mating choice, since 132 using social cues for mate choice is likely to require cognitive abilities such as learning 133 and memory. 134 135 FROM SEVERAL SIGNALS TO MULTIPLE SIGNALING 136 Møller & Pomiankowski (1993) pioneered the formulation of verbal models that 137 explain the used of multiple signals, by proposing three hypotheses based on the 138 information content in complex signals (content-based hypotheses sensus Guilford & 139 Dawkins 1991, Hebets & Papaj 2005, Table 1). First, the multiple message hypothesis 140 states that each trait provides particular information on the sender. Assessing several 6 141 signals thus provides several pieces of information about various aspects of the sender’s 142 location, species, sex, and quality (reviewed in Hebets & Papaj 2005). For example in the 143 field cricket Gryllus assimilis, song pulse rate honestly reflects juvenile condition 144 whereas dominant frequency reflects adult condition (Whattam & Bertram 2011). 145 Secondly, the redundant signal hypothesis (also called back-up signals hypothesis) 146 postulates that each signal provides similar information on the sender but, because each 147 signal may contain a certain degree of error or female assessment of each signal is error- 148 prone, the use of several signals decreases this error (Møller & Pomiankowski 1993, 149 Johnstone 1996). In the barking treefrog, Hyla gratiosa, Poole & Murphy (2007) found 150 that three parameters in male call give information about male size and that the 151 probability for a female to choose the largest male is higher when she can assess together 152 the three parameters, rather than each parameter separately. Alternatively, the unreliable 153 hypothesis (Møller & Pomiankowski 1993) proposed that multiple traits may not inform 154 about male quality. Either they have lost their connection to quality or they have evolved 155 through a runaway process or a receiver bias (Fisher 1930, Ryan 1990). 156 157 In addition to selection on the content, multiple signals could be also explained by a 158 selection on the transmission, reception or information processing by the receiver (the 159 “efficacy-based hypothesis”, Guilford and Dawkins 1991, Hebets & Papaj 2005, Table 1, 160 Fig. 1). For instance, amplifiers are signals that exist solely because they increase the 161 value of other signals (Hasson 1997, Harper et al. 2006). Abdominal patches of the 162 jumping spider Plexippus paykulli provide an example of amplifier because they 163 highlight the abdominal width linked to body condition (Taylor et al. 2000). In agreement 164 with “the amplifier hypothesis”, Rowe (1999) brought into light that the way the receiver 165 perceives and processes signals can explain the evolution of multiple signalling, as 7 166 several signals can be more effective to catch the receiver’s attention than a single one. 167 This hypothesis is based on the efficiency of the signal on the receiver psychology. 168 Multiple signals can affect receiver psychology in several ways acting on their potential 169 of detectability, discriminability or memorability (Rowe 1999). For example multimodal 170 signalling of male wolf spiders in the genus Schizocosa increases their detection by 171 females (Uetz et al. 2009). In the tree frog Hyla arborea, it was demonstrated that the 172 ability of female to discriminate attractive calls (and so good quality mates) increased 173 when the number of call components was high (Richardson & Lengagne 2010). Recently, 174 Akre & Ryan (2010) tested whether the complexity of the signal (3 chucks vs. 1 chuck) 175 can influence the memorability of female tùngara frog (Physalaemus pustulosus). Using a 176 dichotomous mate choice design, they showed that females remember the position of the 177 male only when he produces 3 chucks. 178 179 Overall, efficacy-based and content-based hypotheses all assume that using several 180 signals is better than a single one. However the adaptive function of multiple signals is 181 often untested (but see Lancaster et al. 2009, Rundus et al. 2010). Notably, only few 182 studies examined whether senders and receivers using multiple signals have better fitness 183 than individuals using only one signal. Though, this is of crucial importance for 184 understanding multiple signals evolution. 185 186 187 COMPLEX CHOICE IN VARIABLE ENVIRONMENT 188 Evolutionary processes are dynamic in time and space, and so are the morphological 189 and behavioural phenotypes resulting from them. The expression of phenotypes is often 190 plastic and depends of genes, environmental condition experienced by the individuals and 8 191 the interaction between genotype and environment (GEI, see Greenfield & Rodriguez 192 2004). In a recent meta-analysis on the temporal dynamics of selection on 89 studies, 193 Siepielski et al. (2009) argued that the strength of selection, including selection acting on 194 mate choice, can vary from year to year and suggested that this phenomenon is very 195 common in nature (but see Morrissey & Hadfield 2012). Thus, the rules of thumb used by 196 animals to make their choice should be flexible and context-dependent to finely track the 197 temporal and spatial changes in trait expression and remain adaptive in any condition 198 (Qvarnström 2001, Cotton et al. 2006, Cornwallis & Uller 2010). This point of view is a 199 major advance in the study of mate choice because it leaves a static vision of adaptive 200 preference for a dynamic vision (Bro-Jørgensen 2010). An elegant demonstration of this 201 idea was performed by Chaine & Lyon (2008) in the lark bunting Calamospiza 202 melanocorys. Female preference was dramatically flexible over a 5-year period and even 203 reversible across years in the direction of the preference for a given trait. Importantly, 204 changes in female preference were adaptive as they allowed the females to maximize 205 their fitness outcome. Chaine & Lyon (2008) concluded that variation in the physical or 206 social environment may therefore explain the existence of multiple signals. Unexpected 207 variation in mate choice may be more common than previously recognised, since it could 208 also occur along finer time scales, such as a reproductive season or even over one day 209 (Oh & Badyaev 2006, Jacot et al. 2008). 210 211 Bro-Jørgensen (2010) called “fluctuating environments hypothesis” the hypothesis 212 that adaptively explains the presence of multiple signals by fluctuations in ecological 213 conditions (Table 1). Multiples signals have evolve to maximize fitness of senders and/or 214 receivers in various fluctuating contexts by affecting both the information content 215 (content-based selection), and the transmission (efficacy-based selection) of the signal 9 216 (Bro-Jørgensen 2010). The environmental heterogeneity could have profound effects on 217 signal transmission. For example if one signal A is adapted to one specific environment 218 E1, and one signal B is adapted in another environment E2, senders and receivers who 219 use both the signal A and B have better success whatever the environment E1 or E2 220 (“Multiple sensory environments” sensus Candolin 2003). For example in the three- 221 spined stickleback, Gasterosteus aculeatus, female based their choice on visual signals in 222 clear water, while olfactory signal was used in turbid water (Heuschele et al. 2009). 223 Female that used both visual and olfactory signal could maximise their fitness whatever 224 the environment. This environmental variation (turbidity of water can change quickly) 225 can explain why visual and olfactory are both selected by mate choice in this species 226 (Heuschele et al. 2009). 227 228 The environmental heterogeneity could have also profound effect on the information 229 content of the signal (Greenfield & Rodriguez 2004, Lindström et al. 2009). For example 230 acoustic signals indicate male condition in field crickets Gryllus camperstris, but these 231 signals are more reliable in mornings and afternoons than evenings (Jacot et al. 2008). In 232 the collard flycatchers Ficedula albicollis conspicuous white forehead patch signals male 233 quality during dry summer conditions whereas the opposite was found in wetter years 234 (Robinson et al. 2012). According to the fluctuating environments hypothesis it is 235 expected that females use another signal when signal is unreliable ie in evenings when 236 acoustic signals failed to indicate male condition in field crickets and in wetter years 237 when white forehead did not match with male quality in the collard flycatchers. 238 239 These examples described in the previous section show that senders and receivers 240 could maximise their fitness in various, fluctuating contexts using multiple signals. Yet, 10 241 the rule(s) used by the receivers to adapt their choice to varying environment remains an 242 open issue. Indeed, in the stickleback example (Heuschele et al. 2009), shifts in the 243 signals used by the receiver could result from a perception constraint (i.e. the female does 244 not receive both signals because she cannot see in turbid water) or from an adaptation of 245 the receiver to external information (i.e. the female perceives both signals but prioritize 246 one of them according to environment condition). In the sand gobies Pomatoschistus 247 minutus, females seem able to adjust their choice according to external information. 248 Females adapt their decision rule for male size according to the intensity of male-male 249 competition (Lehtonen & Lindström 2009). When male-male competition is high, 250 females prefer big males because they benefit from the production of offspring able to 251 win competition, but not when competition is low. Note that according to the fluctuating 252 environments hypothesis it is expected that when male-male competition is low, females 253 should base their choice on another signal such as nest quality (Lehtonen et al. 2007). 254 Future research needs to investigate the rule used by females to adapt their mate choice 255 according to varying environments. According to previous models receivers are expected 256 to prioritize signals that are more reliable and easier to evaluate in each given 257 environment (Schluter & Price 1993, Iwasa & Pomiankowski 1994). To identify the most 258 reliable signal, one possibility is that receivers prospect several potential mates and base 259 their choice on the signal with the highest variance. Another assumption is that receivers 260 should use flexible signals in flexible environments and more static signals in less 261 flexible environments (Bro-Jørgensen 2010). 262 263 In addition to variation in external environment (ecological, social), the receiver 264 himself can be a source of variation (“receiver variability” sensus Hebets & Papaj 2005, 265 Fig. 1). While the classical view states that mate choice is concordant among receivers in 11 266 a population, recent work shows that preference is not always absolute in a population 267 (Jennions & Petrie 1997, Mays & Hill 2004). Indeed variation in mate choice among 268 females is common, even in fixed environmental conditions (reviewed by Jennions & 269 Petrie 1997), and could be largely explained by female heterogeneity (Bakker et al. 1999, 270 Kodric-Brown & Nicoletto 2001, Forstmeier & Birkhead 2004, Burley & Foster 2006, 271 Ilmonen et al. 2009, Hebets et al. 2008, Baugh & Ryan 2009, see also the “perceptual 272 variability hypothesis” in Hebets & Papaj 2005). This inter-individual variability could 273 also explain the evolution of multiple signals (Coleman et al. 2004, Suk & Choe 2008). 274 Indeed, we can expect that in one population in which 50% of the females assess an 275 ornament A and 50% assess an ornament B, males exhibiting both A and B ornaments 276 have a higher probability to find a sexual partner than males with only one of the A and B 277 ornaments. Experimental evidence of this paradigm was established in the satin 278 bowerbird Ptilonorhynchus violaceus in which an internal feature (age of females) can 279 drive the evolution of complex male displays such as blue decoration and display 280 intensity (Coleman et al. 2004). However evidences are scarce (but see Suk & Choe 281 2008). While it is now clear that animals have different personalities that influence their 282 decision rule in various contexts such as dispersal (Cote et al. 2010), foraging tactic 283 (David et al. 2011) or information use (Kurvers et al. 2010), only few studies investigated 284 a link between personality and mate choice (but see: Schuett et al. 2010). One future line 285 of research would be to explore how individual personality influences mate choice and 286 whether the coexistence of multiple personalities in a population of receivers can favour 287 the evolution of multiple signals in senders. 288 289 290 12 291 EMPIRICAL TESTS AND CONCEPTUAL ISSUES OF MULTIPLE SIGNALS 292 The number of hypotheses put forward to explain the evolution of multiple signals 293 offers great potential for testing them empirically (Hebets & Papaj 2005). However, 294 researchers may be confronted to a conceptual issue: the difficulty to empirically and 295 theoretically disentangle between some hypotheses (Johnstone 1996, Kokko 2001, Kokko 296 et al. 2002, Van Doorn & Weissing 2004). The distinction is not always obvious between 297 hypotheses differing on the content of the information conveyed by the signals. Signals 298 can be “fisherian” or condition-dependant. However, Kokko and colleagues (2002) 299 showed that the two apparently distinct processes (“fisherian” and good genes) are 300 actually the two ends of the same continuum in a model of preference for traits that 301 indicate indirect benefits. In agreement, Marcias Garcia & Ramirez (2005) demonstrated 302 that male signals exploiting receiver biases can become honest condition-dependant 303 signals. 304 305 The multiple messages and redundant message hypotheses are commonly mixed up 306 in many empirical studies (Candolin 2003, Jawor et al. 2004, Van Doorn & Weissing 307 2004). A first reason is that two definitions of redundant signals coexist. One definition is 308 based on the content-based information that receivers can glean about senders (Møller & 309 Pomiankowski 1993), whereas the second definition is based on the effect of the signals 310 on the receiver behaviour (Partan & Marler 1999, Partan & Marler 1995). According to 311 Partan & Marler (1999), signals are redundant or non-redundant if receivers exhibit 312 respectively similar or dissimilar responses to isolated signals. A second reason is that it 313 is difficult to empirically distinguish between the multiple messages and redundant 314 message, because information on individual condition can be either considered as 315 identical or as a different kind of information, depending on the authors and the accuracy 13 316 of the method and/or experiment (Møller & Pomiankowski 1993, Johnstone 1996, Van 317 Doorn & Weissing 2004). Specifically, two traits can signal parasitic resistance (i.e. 318 redundant messages), but one trait can signal resistance against ecto-parasites and the 319 other one can signal resistance against endo-parasites (i.e. multiple messages) (e.g. Jawor 320 et al. 2004). However, it is important to point out that validation of the redundant 321 message hypothesis do not require only correlations between several traits and one index 322 of sender quality but also an experimental confirmation that the probability to choose a 323 high quality mate increases with the number of traits used, as defined by Møller & 324 Pomiankowski (1993) (see e.g. Poole & Murphy 2007). 325 326 A unifying view of multiple signals theories is to consider that all these theories are 327 not necessarily mutually exclusive and that multiple signals may constitute a mosaic of 328 various types of signals. In the Satin bowerbird, Ptilonorhynchus violaceus, different 329 signals (bower quality and ultraviolet plumage coloration) indicate different aspects of 330 condition (respectively ecto-parasite load and feather growth). In that sense they support 331 the multiple message hypothesis. Thought, these signals also indicate the same aspect of 332 condition (body size, redundant signal hypothesis) (Doucet & Montgomerie 2003). Male 333 bower and plumage in this species provide both multiple information and redundant 334 information on the sender’s quality. Multiple other scenarios are conceivable: let’s 335 suppose a situation in which three signals A, B and C convey two types of information a 336 and b (i.e. multiple messages and redundant message), transmission of A, B and C 337 altogether could be favoured by a sensory bias, which in turn would allow a better 338 memorability of a and b information by the receiver (i.e. receiver psychology). As these 339 mosaics of signals may be the norm rather than an exception, this point of view should 14 340 therefore be integrated into future mathematical models investigating the evolution of 341 multiple signalling. 342 343 344 HOW INDIVIDUALS CHOOSE THEIR MATE(S) WHEN FACING MULTIPLE 345 SIGNALS? 346 One important proximate question still largely unsolved is how receivers use several 347 signals to choose their partner(s). It is still unclear how signals are perceived and 348 processed, and more generally what are the rules explaining mate choice when several 349 signals come to the receiver’s brain. For example, when an individual uses a pool- 350 comparison tactic with multiple signals, it implies that this individual receives several 351 kind of information about each potential partner and treats all information to finally 352 assign one value to each partner (Bateson & Healy 2005). In a threshold-criterion tactic 353 with multiple signals, each signal contributes (to a various extent) to reach the threshold 354 of the behavioural answer. This perspective adds a new level of complexity to the rule of 355 choice. How do prospective individuals assign one general value to each potential partner 356 when they received many signals? What is the thumb of rule used by receivers? Several 357 decision processes have been theoretically proposed, and empirically identified to some 358 extent. 359 360 -When animal use priority rules: 361 In priority rule systems, the receiver evaluates the multiple signals one by one, in a 362 hierarchical order, and one signal B is assessed only if another signal A has exceeded a 363 certain threshold, or when the choosy individual cannot discriminate between potential 364 mates only using the signal A (Scheuber et al. 2004, Table 2). This ruling system has been 15 365 named as hierarchical, sequential or priority system (Wittenberger 1983, Scheuber et al. 366 2004). One trait could be prioritized according to assessment cost or the abundance of 367 desirable partners (Fawcett & Johnstone 2003). In mice and humans, both complementary 368 genes and good genes are used for mate choice, but complementary genes are evaluated 369 only when potential partners share the same quality in terms of good genes (Roberts & 370 Gosling 2003, Saxton et al. 2009). Hierarchical rule is also known for private and social 371 information. Witte & Ryan (1998) found that the Sailfin Molly Poecilia latipinna use 372 social information only when there is no difference in private information (in terms of 373 male sizes). 374 375 -When individuals use weighting rules: 376 Individuals can evaluate a potential mate thanks to a combination of several signals in 377 which one signal does not affect the value of another signal via an additive assessment 378 process (Table 2, “simultaneous hypothesis”, Burke & Murphy 2007). In other words, the 379 strength of the preference increases linearly with the addition of signals and there is no 380 interaction between them. When the signals A and B are perceived together, the strength 381 of the resulting preference is equal to the sum of the preference when A is perceived 382 alone plus the preference when B is perceived alone (“A+B”= “A” + “B”). To our 383 knowledge, only few studies have clearly demonstrated this pattern of rule. An example 384 of additive equal preference with two signals is provided by the butterfly Bicyclus 385 anynana, in which the individuals seem to devote as much importance to visual as to 386 chemical signals in mate choice (A=B, Costanzo & Monteiro 2007). In the stickleback, 387 Gasterosteus aculeatus, females can use redness, courtship and body size to evaluate 388 males (Künzler & Bakker 2001). Using virtual stickleback, Künzler & Bakker (2001) 389 found that adding signals increases female preference. The preference for 16 390 “red+courtship+body size” exceeds the preference for “red+courtship” that exceeds the 391 preference for “red” male (A+B+C>A+B>A), without one signal influencing the weight of 392 other signals (Künzler & Bakker 2001). 393 394 Contrary to additive assessment process, multiplicative one occurs when the strength 395 of the preference is higher than a simple sum of signals (Table 2). Preference increases 396 non-linearly with the number of signals used. Namely, the strength of preference when 397 two signals A and B are present is higher than the sum of the preference when A is 398 perceived alone plus the preference when B is perceived alone (“A×B”>“A”+“B”). A 399 recent empirical study in the Sand Goby, Pomatoschistus minutes, reveals multiplicative 400 assessment process with male size and nest size as signals (Lehtonen et al. 2007). In this 401 species female choice was poorly explained by male size and nest size alone but strongly 402 explained by the interaction between them (Lehtonen et al. 2007). Amplifiers represent a 403 special kind of multiplicative assessment process. An amplifier is a signal that increases 404 the attractiveness of another signal (Hansson 1997, Rowe 1999, Harper et al. 2006), acts 405 as an “attention grabber” (Rowe 1999, “alerting signal”: Brø-Jorgensen 2010), improves 406 detection or discrimination, that changes the way the receiver processes signals (Hebets 407 & Papaj 2005). In a larger sense, multiplicative interaction includes also “emergence”, 408 which is defined by Partan & Marler (2005) as the combination of two signals that elicits 409 an entirely new result (A+B=C; see Rowe & Guilford 1996 for an empirical example 410 with aposematic signals). Emergence represents a small fraction of multiplicative 411 interaction in which only an important change in the answer, such as being chosen or 412 being rejected as a mate, represents an “entirely new result”. In this case the two signals 413 A and B do not elicit an answer alone but together (Table 2), as previously mentioned 414 with the sand goby experiment of Lehtonen and colleagues (2007). 17 415 -How to distinguish between decision processes? 416 Only few experimental designs offer the possibility to disentangle between the different 417 decision rule(s) used by animals during mate choice (Scheuber et al. 2004, Burke & 418 Murphy 2007). Moreover as explained in the section “complex choice in variable 419 environment”, the rule of thumb used by females to make their choice could be dependent 420 on the environmental conditions and/or internal features of the prospecting females. Thus 421 even the decision rule(s) should be envisaged in a dynamic context. Recent advances in 422 technology include robotic technology, audio, video recording and engineering. These 423 advances offer promising and exciting research perspectives in the study of mate choice 424 by providing the possibility to create new signals, control existing signals or design any 425 kind of signal variation and test the impact of each signal on the strength of preference 426 (e.g. Rosenthal et al. 2004, Poole & Murphy 2007, Taylor et al. 2008, Gomez et al. 2009). 427 In addition, robotic technology will also allow testing for the effects of internal and 428 external features on decision rule used by female, while controlling for signal variations. 429 430 431 TOWARD A NEUROBIOLOGICAL PROCESSING OF MULTIPLE SIGNAL 432 THEORY? 433 Complexity in signal expression is necessarily reflected in signal perception and 434 processing. This complexity likely explains why it is still unclear how choosing 435 individuals integrate and treat all multiple signals, potential interactions between signals, 436 and information on environmental conditions to make their choice (“information 437 processing” in Fig. 1). A rewarding area for future studies is cognitive ecology, i.e. 438 animal information processing that leads to decision-making with fitness consequences 439 (Dukas 1998). Here, cognition is considered in its broader sense, namely “the sum of 18 440 processes by which information is processed, stored, retrieved and used in a flexible and 441 adaptive sense. Such a definition refers therefore to implicit faculties like perception, 442 attention, representation, expectation, planning, learning, recognition, memory and 443 knowledge organization” (Giurfa 2003). This view contradicts the common misleading 444 image of animals as mainly directed by pre-programmed reflex mechanisms, but does not 445 imply conscious levels of decision making. 446 447 Earlier work on “receiver psychology” (Rowe 1999) has highlighted the importance 448 of considering neuronal processes and brain functions to understand the implication of 449 multiple signals in mate choice. Therefore recent advances in neurology open new 450 avenues to explore mate choice more accurately when animals are facing two or more 451 signals. Indeed, modern methods now allow to visualize brain regions that are activated 452 when performing certain tasks, such as attractiveness assessment and mate decision, or 453 while being stimulated by various stimuli including sexual signals. Two main types of 454 techniques that track neural activity are currently available and are increasingly being 455 used: Immediate Early Gene expression (IEG) and functional resonance magnetic 456 imaging (fMRI). 457 458 IEGs are genes that are dynamically regulated in the brain by neural activity. They 459 are rapidly expressed upon neural stimulation (usually 20-90 minutes) and function as 460 transcription factors or code for structural proteins. Classical experiments involve one 461 group of individuals which performs a behavioural task or is exposed to a stimulus, while 462 the control group is placed in control conditions. Because IEG induction does not require 463 preparatory procedures, individuals can virtually be exposed to any kind of stimulus, 464 allowing the exploration of a variety of behavioural contexts (Farivar et al. 2004). After 19 465 sacrifice, brain sections are marked to detect either mRNA or proteins, with a resolution 466 as precise as the cellular level (Guzowski et al. 2001, Farivar et al. 2004). Using IEG 467 expression, two studies identified the forebrain regions involved in the detection and 468 discrimination of preferred male songs. In the European starling (Sturnus vulgaris), 469 females exposed to preferred longer songs showed higher expression of the ZENK gene 470 in the ventral-medial neostriatum. Similarly, female canaries showed higher expression 471 level of the same gene but in the caudomedial mesopallium when they were exposed to 472 songs containing sexy syllables (Gentner et al. 2001, Leitner et al. 2005). These two 473 examples demonstrate that IEG techniques yield promising results when females are 474 exposed to single, “simple”, sexual stimuli; however these techniques can also be 475 rewarding in more complex contexts, such as varying social context, or when females use 476 complex social cues to choose their mates (Desjardins et al. 2010). Moreover, Farivar et 477 al. (2004) proposed a new procedure called “dual activity mapping” in which two 478 different stimuli A and B are successively applied to a given individual, and the neurons 479 that are activated by A only, B only, or both A and B can be identified. Although it has not 480 been used to investigate how multiple signals are processed so far, this new approach is 481 shows potential and surely will yield important results in the next few years. 482 483 In contrast to IEGs, neuroimaging has made it possible to visualize dynamic brain 484 function without invasion. Indeed, IEG techniques can only provide a post-mortem 485 picture of how the brain reacted to a specific stimulus or event, whereas neuroimaging 486 allows a direct observation of brain function at the time of stimulation and time series of 487 observations required for longitudinal studies. Neuroimaging comprises several 488 techniques, including electroencephalogram (EEG) signals analyzed in term of event- 489 related brain potentials (ERPs) (e.g. Schacht et al. 2008), positron emission tomography 20 490 (PET) (e.g. Bales et al. 2007) and functional resonance magnetic imaging (fMRI). Both 491 PET and fMRI measures change in blood flow, oxygen consumption and glucose uptake 492 that are directly related to neural activation. fMRI is now preferred to PET because it 493 provides a high-resolution three-dimension colored picture of the brain and does not 494 entail the injection of radioactive substances that may lead to overexposure to radiation 495 (Logothesis 2008). Investigations of the brain regions activated by sexually attractive 496 stimuli have mainly been undertaken in humans so far. It is largely admitted that human 497 mate choice is reciprocal, based on multiple signals, including facial and body 498 attractiveness 499 information, and that internal information can modulate the strength of preference given 500 to each signal (Little et al. 2008, Currie & Little 2009, Saxton et al. 2009, Lass- 501 Hennemann et al. 2010). Yet physical attractiveness plays an important role in both men 502 and women mate choice, and asking experimental subjects to watch and rank photos 503 according to their attractiveness is a simple procedure. Nakamura and colleagues (1998) 504 used fMRI and identified higher activity in the left frontal regions of men brain when 505 subjects assessed a woman face as attractive. Similarly, Rupp and colleagues (2009a,b) 506 examined the brain regions associated with women’s sexual decision making with fMRI 507 while women were watching photos of men’s faces presented as potential sexual partners. 508 The researchers found a stronger activation in a number of brain regions, namely the 509 anterior cingulated cortex, posterior cingulate cortex, midbrain, parietal sulcus, superior 510 temporal gyrus, precentral gyrus, and inferior parietal lobule. In both men and women, 511 PETs revealed a higher activity in the caudate nucleus and other parts of the brain reward 512 system linked to attractive female bodies (Holliday et al. 2011). In general, most brain 513 regions that were identified by these studies are known to be involved in the brain reward physical attractiveness, self-similarity, sexual 21 dimorphism social 514 system and it is likely that an observer’s preference for an attractive photo is mediated by 515 its reward value (Fisher et al. 2006). 516 517 Clearly, behavioural ecology coupled to neurobiology has only begun to establish 518 relationships between gene expression, brain function and male sexual traits, and the 519 powerful IEG and fMRI techniques have still a lot to contribute to the understanding of 520 mate choice. The experiments described above call for future research into the effect of 521 multiple signals on brain function. More interestingly we can expect that the results of 522 such studies may provide answers on how each signal is processed, which interactions 523 occur (additive, multiplicative, hierarchical) and finally what are the rules of thumb used 524 to make a choice. More precisely we can expect that when animal use hierarchical rule (B 525 is used only after evaluation of A), the neurons processing information of A suppress the 526 neurons processing information of B. Conversely when animal use multiplicative 527 interaction, we can expect that neurons processing information of A and B activate each 528 other to increase neural response which exceeds the simple sum of response when A is 529 perceived alone plus the preference when B is perceived alone (“A×B” > “A”+“B”). We 530 can predict more connections between neurons processing information of A and B in 531 multiplicative interaction that in additive interaction. 532 533 Organisms with small brains may serve as prime model systems for identifying 534 neuronal processes and circuitry, and neural substrates involved in mate preferences for 535 multiple signals (Dickson 2008). These organisms have indeed fewer neurons and 536 simpler brain architecture than large-brained species, making research easier. Moreover 537 artificial neural network analyses typically show that the minimum number of neurons 538 necessary to perform a variety of cognitive tasks is exceedingly small. Drosophila 22 539 appears to be a great model because this species shows complex mate choice with 540 multiple signals, mating rule is affected by internal and external features, while having 541 well-known simple brain architecture (Dickson 2008). On the other hand, several 542 vertebrate species might be good candidates to study cognitive ecology: anurans such as 543 the túngara frog, fishes such as poeciliid species or human (Fisher et al. 2006, Wilczynski 544 & Ryan 2010, Cummings 2012). 545 546 Advances in the field of neurobiology will provide new insights about the proximate 547 causes of multiple signalling, allowing a better understanding on how individuals choose 548 their mate(s). Despite tremendous advances in the field of cognitive ecology, we are still 549 far from unravelling how multiple signals emitted in the context of sexual communication 550 can elicit answers such as pair-bonding or mating (Fig. 1). The field of cognitive ecology 551 is currently undergoing major advances (Dikson 2008, Wilczynski & Ryan 2010). The 552 association between neurobiology and behavioural ecology could provide powerful 553 results and better understanding of multiple signal theory. Future research should 554 investigate how receiver brain varies with the attractiveness of conspecifics senders 555 including multiple and multimodal signals (Fisher et al. 2006, Dickson 2008, Wilczynski 556 & Ryan 2010). Moreover more studies should aim at exploring receiver variability in 557 perception and information processing at the neural level (Hebets & Papaj 2005). 558 559 560 CONCLUSION 561 Since first interrogations about the use of multiple signals, it has become clear that 562 mate choice is by far more complex than a simple relation between one trait and one 563 preference. An adaptive reproductive decision commonly requires the detection and 23 564 processing of multiple, and sometimes multimodal, signals. The senders should convey 565 adaptive signals that allow them to reproduce with the receivers and the receivers 566 integrate all these signals to make an appropriate decision (behaviour) to choose the best 567 senders. This decision has a great influence on fitness of the two protagonists and natural 568 selection has selected various sender and receiver strategies using multiple signals. 569 Recent research works have revealed a large variability in the kind of signals used by 570 receivers but less conventional signals and cues (such as social cues) should be taken into 571 account more frequently in multiple signals theory. 572 573 Many theories have been developed to explain the evolution and maintenance of 574 multiple signals. These different theories are frequently considered separately, yet recent 575 empirical results strongly suggest that multiple signals can be the result of multiple 576 evolutionary pressures, and therefore should be more integrated in a common framework. 577 Importantly, recent studies highlighted the importance to take into account both internal 578 characteristics and environmental conditions that may influence the production, 579 transmission, reception and processing of multiple signals. This internal and external 580 environmental heterogeneity could explain the use of multiple cues and the use of 581 different decisions rules to process multiples signals. The use of robot technology, new 582 advances in cognitive ecology can help research on multiple signals and the 583 understanding of how signals are perceived by individuals, are processed by brains, how 584 these signals interact and lastly how internal and external information affect signal 585 processing, all in a dynamical world. This knowledge will allow understanding the 586 evolution of mate choice using multiple signals and surely will provide a better and more 587 realistic estimate of the strength of sexual selection. 588 24 589 Acknowledgements 590 This work was supported by the Ministère de la Recherche (PhD fellowship to JC and 591 two CNRS grants to AL and DS). This work was supported by the "Laboratoire 592 d’Excellence (LABEX)" entitled TULIP (ANR -10-LABX-41). 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 25 614 References 615 Akre, K.L. & Ryan M.J. 2010. Complexity increases working memory for mating 616 signals. Current Biology, 20, 502-505. 617 Andersson M. 1994. Sexual selection. Princeton: Princeton University Press. 618 Andersson, M. & Simmons, L. W. 2006. Sexual selection and mate choice. Trends in 619 620 621 Ecology & Evolution, 21, 296-302. Bakker, T. C. M., Künzler, R. & Mazzi, D. 1999 Condition-related mate choice in sticklebacks. Nature, 401, 234. 622 Bales, K. L., Mason, W. A., Catana, C., Cherry, S. R. & Mendoza, S. P. 2007 Neural 623 correlates of pair-bonding in a monogamous primate. Brain Research, 1184, 245- 624 253. 625 626 627 628 629 630 631 632 633 634 635 636 637 638 Bateson, M. & Healy, S. D. 2005. Comparative evaluation and its implication for mate choice. Trends in Ecology & Evolution, 20, 659–664. Baugh, A. T. & Ryan, M. J. 2009. Female túngara frogs vary in commitment to mate choice decisions. Behavioral Ecology, 20,1153-1159. Bro-Jørgensen, J. 2010. Dynamics of multiple signalling systems: animal communication in a world in flux. Trends in Ecology & Evolution, 25, 292-300. Burke, E.J. & Murphy, C.G. 2007. How female barking treefrogs, Hyla gratiosa, use multiple call characteristics to select a mate. Animal Behaviour, 74, 1463-1472. Burley, N. T. & Foster, V. S. 2006. Variation in female choice of mates: condition influences selectivity. Animal Behaviour, 72, 713-719. Byers, J., Hebets, E. & Podos, J. 2010. Female mate choice based upon male motor performance. Animal Behaviour, 79, 771-778. Candolin, U. 2003. The use of multiple cues in mate choice. Biological Reviews, 78, 575-595. 26 639 640 Carranza, J. 2010. Sexual selection and the evolution of evolutionary theories. Animal Behaviour, 79 (3), e5–e6. 641 Castellano, S. & Rosso, A. 2007. Female preferences for multiple attributes in the 642 acoustic signals of the Italian treefrog, Hyla intermedia. Behavioral Ecology and 643 Sociobiology, 61, 1293-1302. 644 645 646 647 648 649 650 651 Chaine, A., Lyon, B. 2008. Adaptive plasticity in female mate choice dampens sexual selection on male ornaments in the lark bunting. Science, 319, 459-462. Clutton-Brock, T. 2007. Sexual selection in males and females. Science, 318, 1882– 1885. Clutton-Brock, T. 2010. We do not need a sexual selection 2.0 – nor a theory of genial selection. Animal Behaviour, 79 (3), e1–e4. Coleman, S. W., Patricelli, G. L. & Borgia, G. 2004. Variable female preferences drive complex male displays. Nature, 428, 742-745. 652 Cornuau, J. C., Rat, M., Schmeller, D. S. & Loyau, A. Multiple signals in the palmate 653 newt : ornaments help when courting. Behavioral Ecology and Sociobiology, 66, 654 1045-1055. 655 656 Cornwallis, C. K. & Uller, T. 2010. Towards an evolutionary ecology of sexual traits. Trends in Ecology and Evolution, 25, 145-152. 657 Costanzo, K. & Monteiro, A. 2007. The use of chemical and visual cues in female 658 choice in the butterfly Bicyclus anynana. Proceedings of the Royal Society of 659 London, Series B, 274, 845-851. 660 Cote, J., Clobert, J., Brodin, T., Fogarty, S., & Sih, A. 2010. Personality-dependent 661 dispersal: characterization, ontogeny and consequences for spatially structured 662 populations. Philosophical Transactions of the Royal Society B: Biological 663 Sciences. 365, 4065-4076. 27 664 665 Cotton, S., Small, J. & Pomiankowski, A. 2006. Sexual selection and condition dependent mate preferences. Current Biology, 16, R755-R765. 666 Cotton, S., Rogers, D. W., Small, J., Pomiankowski, A., Fowler, K. 2006. Variation in 667 preference for a male ornament is positively associated with female eyespan in the 668 stalk-eyed fly Diasemopsis meigenii. Proceedings of the Royal Society B, 273, 669 1287-1292. 670 671 Cummings, M. E. 2012. Looking for sexual selection in the female brain. Philosophical Transactions of the Royal Society B, 367, 2348-2356. 672 Currie, T. E. & Little A. C. 2009. The relative importance of the face and body in 673 judgments of human physical attractiveness. Evolution and Human Behavior, 30, 674 409-416. 675 676 677 678 679 680 Danchin, E., Giraldeau, L., Valone, T. & Wagner, R. 2004. Public information: from nosy neighbours to cultural evolution. Science, 305, 487-491. Darwin, C. 1871. The Descent of Man, and Selection in Relation to Sex. London: Murray. David, M., Cézilly, F. & Giraldeau, L. A. 2011. Personality affects zebra finch feeding success in a producer-scrounger game. Animal Behaviour, 82, 61-67. 681 Desjardins, J. K., Klausner, J. Q. & Fernald, R. D. 2010. Female genomic response to 682 mate information. Proceedings of the National Academy of Sciences, U.S.A., 107, 683 21176-21180. 684 685 Dickson, B. J. 2008. Wired for sex: the neurobiology of Drosophila mating decisions. Science, 322, 904-909. 686 Doucet, S. M. & Montgomerie, R. 2003. Multiple sexual ornaments in satin bowerbirds: 687 ultraviolet plumage and bowers signal different aspects of male quality. Behavioral 688 Ecology, 14, 503-509. 28 689 Drullion, D. & Dubois, F. 2008. Mate-choice copying by female zebra fiches, 690 Taeniopygia guttata: what happens when model females provide inconsistent 691 information? Behavioral Ecology and Sociobiology, 63, 269-276. 692 Dugatkin, L. A. 1996. The interface between culturally-based preferences and genetic 693 preferences: female mate choice in Poecilia reticulata. Proceedings of the National 694 Academy of Sciences, U.S.A., 93, 2770–2773. 695 696 697 698 Dugatkin, L. A. 1998. Genes, copying, and female mate choice; shifting thresholds. Behavioral Ecology, 9, 323-327. Dukas, R. 1998. Cognitive ecology: the evolutionary ecology of information processing and decision making. Chicago: The University of Chicago. 699 Farivar, R., Zangenehpour, S. & Chaudhuri, A. 2004. Cellular-resolution activity 700 mapping of the brain using immediate-early gene expression. Frontiers in 701 Bioscience, 9, 104-109. 702 703 Fawcett, T. W. & Johnstone, R. A. 2003. Optimal assessment of multiple cues. Proceedings of the Royal Society B, 270, 1637-1643. 704 Fisher, H. E., Aron, A. & Brown, L. L. 2006. Romantic love: a mammalian brain 705 system for mate choice. Philosophical Transactions of the Royal Society B, 361, 706 2173-2186. 707 708 709 710 711 712 Fisher, R. A. 1930. The genetical theory of natural selection. Princeton: Princeton University Press. Font, E. & Carazo, P. 2010. Animals in translation: why there is meaning (but probably no message) in animal communication. Animal Behaviour, 79, e1-e6. Forstmeier, W. & Birkhead, T. R. 2004.Repeatability of mate choice in the zebra finch: consistency within and between females. Animal Behaviour, 68, 1017-1028. 29 713 Gentner, T. Q., Hulse, S. H., Duffy, D. & Ball, G. F. 2001. Response biases in auditory 714 forebrain regions of female songbirds following exposure to sexually relevant 715 variation in male song. Journal of Neurobiology, 46, 48-58. 716 717 Gibson, R. M. & Langen, T. A. 1996. How do animals choose their mates? Trends in Ecology & Evolution, 11, 468–470. 718 Gomez, D., Richardson, C., Lengagne, T., Plenet, S., Joly, P., Léna, J. P. & Théry, 719 M. 2009. The role of noctural vision in mate choice: females prefer conspicuous 720 males in the European tree frog (Hyla arborea). Proceedings of the Royal Society 721 B, 276, 2351-2358. 722 723 724 725 726 727 728 729 730 731 Greenfield, M.D., Rodriguez, R.L. 2004. Genotype-environment interaction and the reliability of mating signals. Animal Behaviour, 68, 1461–1468. Grether, G. F. & Kolluru, G. R. 2004. Individual colour patches as multicomponent signals. Biological Reviews, 79, 583-610. Guilford, T. & Dawkins, M. S. 1991. Receiver psychology and the evolution of animal signals. Animal Behaviour, 42, 1-14. Guilford, T. and M. S. Dawkins. 1991. Receiver psychology and the evolution of animal signals. Animal Behaviour 42: 1-14. Guirfa, M. 2003. Cognitive neuroethology: dissecting non-elemental learning in a honeybee brain. Current Opinion in Neurobiology, 13, 726-735. 732 Guzowski, J. F., McNaughton, B. L., Barnes, C. A., Worley, P. F. 2001. Imaging 733 neural activity with temporal and cellular resolution using FISH. Current Opinion 734 in Neurobiology, 11, 579-584. 735 736 Halliday, T. R. 1983. The study of mate choice. In: Mate choice, (Bateson P, ed), Cambridge: Cambridge University Press; 3-32. 30 737 738 739 740 741 742 743 744 Harper, D. G. C. 2006. Maynard Smith: amplifying the reasons for signal reliability. Journal of Theoretical Biology, 239, 203-209. Hasson, O. 1997. Towards a general model of biological signaling. Journal of Theoretical Biology, 185, 139-156. Hebets, E. A. & Papaj, D. R. 2005. Complex signal function: developing a framework of testable hypotheses. Behavioral Ecology and Sociobiology, 57,197-214. Hebets, E. A., Wesson, J. & Shamble, P. S. 2008. Diet influences mate choice selectivity in adult female wolf spiders. Animal Behaviour, 76, 355-363. 745 Heisler, I. L., Andersson, M. B., Arnold, S. J., Boake, C. R., Borgia, G., Hausfater, 746 G., Kirkpatrick, M., Lande, R., Maynard Smith, J., O’Donald, P., Thornhill, 747 A. R. & Weissing, F. J. 1987. The evolution of mating preferences and sexually 748 selected traits: group report. In: Sexual Selection: Testing the Alternatives (Ed. by 749 J. W. Bradbury & M. B. Andersson), pp. 96–118. Chichester: John Wiley. 750 Heuschele, J., Mannerla, M., Gienapp, P. & Candolin, U. 2009. Environment- 751 dependent use of mate choice cues in sticklebacks. Behavioral Ecology, 20, 1223- 752 1227. 753 Holliday, I. E., Longe, O. A., Thai, N. J., Hancock, P. J. B. & Tovée, M. J. 2011. 754 BMI not WHR modulates BOLD fMRI responses in a sub-cortical reward network 755 when participants judge the attractiveness of human female bodies. Plos One, 6, 756 e27255. 757 758 759 760 Ilmonen, P., Stundner, G., Thoss, M. & Penn, D. J. 2009. Females prefer the scent of outbred males: good-genes-as-heterozygosity? BMC Evolutionary Biology, 9, 104. Iwasa, Y. & Pomiankowski, A. 1994. The evolution of mate preference for multiple sexual ornaments. Evolution, 48, 853-867. 31 761 Jacot, A., Scheuber, H., Holzer, B., Otti, O. & Brinkhof, M. W. G. 2008. Diel 762 variation in a dynamic sexual display and its association with female mate- 763 searching behaviour. Proceedings of the Royal Society B, 275, 579–585. 764 Jawor, J. M., Gray, N., Beall, S. M. & Breitwisch, R. 2004. Multiple ornaments 765 correlate with aspects of condition and behaviour in female northern cardinals, 766 Cardinalis cardinalis. Animal Behaviour, 67, 875-882. 767 768 769 770 Jennions, M. D. & Petrie, M. 1997. Variation in mate choice: a review. Biological Reviews, 72, 283-327. Johnstone, R. A. 1995a. Honest advertisement of multiple qualities using multiple signals. Journal of Theoretical Biology, 177, 87-94. 771 Johnstone, R. A. 1996. Multiple displays in animal communication: ‘‘backup signals’’ 772 and ‘‘multiple messages’’. Philosophical Transactions of the Royal Society of 773 London, Series B, 351, 329-338. 774 775 776 777 778 779 780 781 782 783 Kodric-Brown, A. & Nicoletto, P. F. 2001. Age and experience affect female choice in the guppy (Poecilia reticulata). American Naturalist, 157, 316-323. Kokko, H. 2001. Fisherian and "good genes" benefits of mate choice: how (not) to distinguish between them. Ecology Letters, 4, 322-326. Kokko, H., Brooks, R., McNamara, J. M. & Houston, A. I. 2002. The sexual selection continuum. Proceedings of the Royal Society B, 269, 1331-1340. Kokko, H., Brooks, R., Jennions, M. D., Morley, J. 2003. The evolution of mate choice and mating biases. Proceedings of the Royal Society B, 270, 653–664. Künzler, R. & Bakker, T. C. M. 2001. Female preferences for single and combined traits in computer animated stickleback males. Behavioral Ecology, 12, 681-685. 784 Kurvers, R. H. J. M., Oers, K. V., Nolet, B. A., Jonker, R. M., Van Wieren, S. E., 785 Prins, H. T. H. & Ydenberg, R. C. 2010. Personality predicts the use of social 32 786 information. Ecology Letters, 13, 829-883. 787 Lancaster, L. T., Hipsley, C. A. & Sinervo, B. 2009. Female choice for optimal 788 combinations of multiple male display traits increases offspring survival. 789 Behavioral Ecology, 20, 993-999. 790 Lass-Hennemann, J., Deuter, C. E., Kuehl, L. K., Schulz, A., Blumenthal, T. D. & 791 Schachinger, H. 2010. Proceedings of the Royal Society B, 277, 2175-2183. 792 Lehtonen, T. K. and Lindstrom, K. 2009. Females decide whether size matters: plastic 793 mate preferences tuned to the intensity of male-male competition. Behavioral 794 Ecology, 20, 195-199. 795 Lehtonen, T. K., Rintakoski, S. & Lindström, K. 2007. Mate preference for multiple 796 cues: interplay between male and nest size in the sand goby, Pomatoschistus 797 minutus. Behavioral Ecology, 18, 696-700. 798 Leitner, S., Voigt, C., Metzdorf, R. & Catchpole, C. K. 2005. Immediate early gene 799 (ZENK, Arc) expression in the auditory forebrain of female canaries varies in 800 response to male song quality. Journal of Neurobiology, 3, 275-284. 801 Lindström, J., Pike, T. W., Blount, J. D., Metcalfe, N. B. 2009. Optimization of 802 resource allocation can explain the temporal dynamics and honestly of sexual 803 signals. American Naturalist, 174, 515-525. 804 Little, A. C., Burriss, R. P., Jones, B. C., De Bruine, L. M. & Caldwell, C. A. 2008. 805 Social influence in human face preference : men and women are influenced more 806 for long-term than short-term attractiveness decisions. Evolution and Human 807 Behavior, 29, 140-146. 808 809 Logothesis, N. K. 2008. What we can do and what we cannot do with fMRI. Nature, 453, 869-878. 33 810 811 812 813 814 815 816 817 818 819 Loyau, A., Petrie, M., Saint Jalme, M. & Sorci G. 2008. Do peahens not prefer peacocks with more elaborate trains ? Animal Behaviour, 76, e5-e9. Lozano, G. A. 2009. Multiple cues in mate selection: The sexual interference hypothesis. Bioscience Hypothesis 2, 37-42. Marcia Garcia, C. & Ramirez, E. 2005. Evidence that sensory traps can evolve into honest signals. Nature, 434, 501-505. Mays, H. L., Albrecht, T., Liu, M. & Hill, G. E. 2008. Female choice for genetic complementarity in birds: a review. Genetica, 134, 147-158. Mays, H. L., Jr. & Hill, G. E. 2004. Choosing mates: good genes versus genes that are a good fit. Trends in Ecology & Evolution, 19, 554-559. 820 McGregor, P. K. 1993. Signalling in territorial systems: a context for individual 821 identification, ranging and eavesdropping. Philosophical Transactions of the Royal 822 Society of London, Series, 340, 237-244. 823 824 825 826 827 828 829 830 Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mery, F. & Kawecki, T. 2005. A cost of long-term memory in Drosophila. Science, 308, 1148. Møller, A. P. & Pomiankowski, A. 1993. Why have birds got multiple sexual ornaments? Behavioral Ecology and Sociobiology, 32, 167-176. Morrissey, M. B. & Hadfield, J. D. 2012. Directional selection in temporally replicated studies is remarkably consistent. Evolution, 66, 435–442. 831 Nakamura, K., Kawashima, R., Nagumo, S., Ito, K., Sugiura, M., Kato, T., 832 Nakamura, A., Hatano, K., Kubota, K.., Fukuda, H., Kojima, S. 1998. 833 Neuroanatomical correlates of the assessment of facial attractiveness. NeuroReport, 834 9, 753-757. 34 835 836 Neff, B. D. & Pitcher, T. E. 2005. Genetic quality and sexual selection: an integrated framework for good genes and compatible genes. Molecular Ecology, 14, 19-38. 837 Oh, K. P. & Badyaev, A. V. 2006. Adaptive genetic complementarity in mate choice 838 coexists with selection for elaborate sexual traits. Proceedings of the Royal Society 839 B, 273, 1913-1919. 840 Otter, K. A., McGregor, P. K., Terry, A. M. R., Burford, F. R. L., Peake, T. M. & 841 Dabelsteen, T. 1999. Do female great tits (Parus major) assess males by 842 eavesdropping? A field study using interactive song playback. Proceedings of the 843 Royal Society of London, Series B, 266, 1305-1309. 844 845 846 847 Partan, S. R. & Marler, P. 1999. Communication Goes Multimodal. Science, 283, 1272-1273. Partan, S. R. & Marler, P. 2005. Issues in the classification of multimodal communication signals. American Naturalist, 166, 231-245. 848 Poole, K. G. & Murphy, C. G. 2007. Preferences of female barking treefrogs, Hyla 849 gratiosa, for larger males: univariate and composite tests. Animal Behaviour, 73, 850 513-524. 851 852 853 854 855 856 Puurtinen, M., Ketola, T. & Kotiaho, J. S. 2005. Genetic compatibility and sexual selection. Trends in Ecology & Evolution, 20, 157-158. Puurtinen, M., Ketola, T. & Kotiaho, J. S. 2009. The Good-Genes and CompatibleGenes Benefits of Mate Choice. American Naturalist, 174, 741-752. Qvarnström, A. 2001. Context-dependent genetic benefits from mate choice. Trends in Ecology & Evolution, 16, 5-7. 857 Richardson, C. & Lengagne T. 2010. Multiple signals and male spacing affect female 858 preference at cocktail parties in treefrogs. Proceedings of the Royal Society of 859 London, Series B, 277, 1247-1252. 35 860 861 862 863 864 865 Roberts, S. C. & Gosling, L. M. 2003. Genetic similarity and quality interact with mate choice decisions by female mice. Nature Genetics, 35, 103-106. Roberts, S. C. & Little, A. C. 2008. Good genes, complementary genes and human mate choice. Genetica, 132, 309-321. Roberts, J. A., Taylor, P. W. & Uetz, G. W. 2007. Consequences of complex signaling: predator detection of multimodal cues. Behavior Ecology, 18, 236-240. 866 Robinson, M. R., van Doorn G. S., Gustafsson L. & Qvarnström A. 2012. 867 Environment-dependent selection on mate choice in a natural population of birds. 868 Ecology Letters, 15, 611-618. 869 Rosenthal, G. G., Rand, A. S. & Ryan, M. J. 2004. The vocal sac as a visual cue in 870 anuran communication: an experimental analysis using video playback. Animal 871 Behaviour, 68, 55-58. 872 873 874 875 876 877 Roughgarden, J. & Akçay, E. 2010. Do we need a Sexual Selection 2.0? Animal Behaviour, 79 (3), e1–e4. Rowe, C. 1999. Receiver psychology and the evolution of multicomponent signals. Animal Behaviour, 58, 921-931. Rowe, C. & Guilford, T. 1996. Hidden colour aversions in domestic chicks triggered by pyrazine odours of insect warning displays. Nature, 383, 520-522. 878 Rundus, A. S., Sullivan-Beckers L, Wilgers D. J. & Hebets, E. A. 2010. Females are 879 choosier in the dark: environment-dependent reliance on courtship components and 880 its impact on fitness. Evolution, 65, 268-282. 881 Rupp, H.A., James, T. W., Ketterson, E. D., Sengelaub, D. R., Janssen E. & Heiman 882 J. R. 2009a. The role of the anterior cingulated cortex in women’s sexual decision 883 making. Neuroscience Letters, 449, 42-47. 36 884 Rupp, H.A., James, T. W., Ketterson, E. D., Sengelaub, D. R., Janssen E. & Heiman 885 J. R. 2009b. Neural activation in women in response to masculinized male faces: 886 mediation by hormones and psychosexual factors. Evolution and Human Behavior, 887 30, 1-10. 888 Ryan, M. J., Fox, J. H., Wilczynski, W. & Rand, A. S. 1990. Sexual selection for 889 sensory exploitation in the frog Physalaemus pustulosus. Nature, 343, 66-67. 890 Saxton, T. K., Little, A. C., Rowland, H. M., Gao, T., & Roberts, C. 2009. Trade-offs 891 between markers of absolute and relative quality in human facial preferences. 892 Behavioral Ecology, 20, 1133-1137. 893 894 895 896 Schacht, A., Werheid, K. & Sommer W. 2008. The appraisal of facial beauty is rapid but not mandatory. Cognitive, Affective, & Behavioral Neuroscience, 8, 132-142. Schaedelin, F. C. & Taborsky, M. 2009. Extended phenotypes as signals. Biological Reviews, 84, 293-313. 897 Scheuber, H., Jacot, A. & Brinkhof, M. W. G. 2004. Female preference for multiple 898 condition-dependent components of a sexually selected signal. Proceedings of the 899 Royal Society of London, Series B, 271, 2453-2457. 900 901 902 903 904 905 Schluter, D. & Price, T. 1993. Honesty, perception and population divergence in sexually selected traits. Proceedings of the Royal Society B, 253, 117-122. Schuett, W., Trengenza, T. & Dall, S.R.X. 2010. Sexual selection and animal personality. Biological Reviews, 85, 217-246. Shuker, D. 2010. Sexual selection: endless forms or tangled bank? Animal Behaviour, 79 (3), e11–e17. 906 Siepielski, A. M., DiBattista, J. D. & Carlson S. M. 2009. It’s about time: The temporal 907 dynamics of phenotypic selection in the wild. Ecology Letters, 12, 1261-1276. 37 908 909 Suk, H., Choe, J. 2008. Dynamic female preference for multiple signals in Rhinogobius brunneus. Behavioral Ecology Sociobiology, 62, 945-951. 910 Taylor, P. W., Hasson, O. & Clark, D. L. 2000. Body postures and patterns as 911 amplifiers of physical condition. Proceedings of the Royal Society of London, 912 Series B, 267, 917-922. 913 Taylor, R. C., Klein, B. A., Stein, J. & Ryan, M. J. 2008. Faux frogs: multimodal 914 signalling and the value of robotics in animal behaviour. Animal Behaviour, 76, 915 1089-1097. 916 Templeton, J. & Giraldeau, L. A. 1996. Vicarious sampling: the use of personal and 917 public information by starlings foraging in a simple patchy environment. 918 Behavioral Ecology and Sociobiology, 38, 105-114. 919 Thünken T., Meuthen D., Bakker, T.C.M, Baldauf, SA. 2012. A sex-specific trade-off 920 between mating preferences for genetic compatibility and body size in a cichlid fish 921 with mutual mate choice. Proceedings of the Royal Society of London, Series B, 922 279, 2959-2964. 923 Uetz, G. W., Roberts, J. A. & Taylor, P. W. 2009. Multimodal communication and 924 mate choice in wolf spiders: female response to multimodal versus unimodal 925 signals. Animal Behaviour, 78, 299-305. 926 Valone, T. J. 2007. From eavesdropping on performance to copying the behavior of 927 others: a review of public information use. Behavioral Ecology and Sociobiology, 928 62, 1-4. 929 930 van Doorn, G. S. & Weissing, F. J. 2004. The evolution of female preferences for multiple indicators of quality. American Naturalist, 164, 173-186. 38 931 Whattam, E. M. & Bertram, S. M. 2011. Effects of juvenile and adult condition on 932 long-distance call components in the Jamaican field cricket, Gryllus assimilis. 933 Animal Behaviour, 81, 135-144. 934 935 936 937 938 939 940 941 Wilczynski, W. & Ryan, M. 2010. The behavioral neuroscience of anuran social signal processing. Current Opinion in Neurobiology, 20, 754-763. Witte, K. & Ryan, M. J. 1998. Male body length influences mate-choice copying in the sailfin molly Poecilia latipinna. Behavioral Ecology, 9, 534-539. Wittenberger, J.F. 1983. Tactics of mate choice. In: Mate Choice (Ed. By P. Bateson), pp. 435-448. Cambridge: Cambridge University Press. Zeh, J. A. & Zeh, D. W. 2003. Toward a new sexual selection paradigm: polyandry, conflict and incompatibility (Invited article). Ethology, 109, 929-950. 39 942 Table 1. Existent hypotheses proposed to explain the presence of multiple signals. Hypothesis Multiple signals Example Multiple message Reflect different information about condition Whattam & Bertram 2011 Redundant message Reflect the same information about condition Poole & Murphy 2007 Ureliable signals No reflect condition MØller & Pomiankowski 1993 Amplifiers Amplify other signal Taylor et al. 2000 Detectability Increase detectability of potential mate Rowe 1999 Discriminability Increase discriminability between potential mate Richardson & Lengagne 2010 Memory Are more efficient to allow memorability Akre & Ryan 2010 Fluctuating environments Using several signals sender could “seduce” receiver according (external feature) Heuschele et al. 2009 receiver external feature Receiver variation (internal Using several signals sender could “seduce” receiver according Coleman et al. 2004 feature) receiver internal feature 943 944 945 40 946 Table 2. Different types of treatment of several signals by the receiver. Strength of preference Interaction Without interaction A B with interaction Example A*B Hierarchical Witte & Ryan 1998 Additive Künzler & Bakker 2001 Multiplicative Castellano & Rosso 2007 Amplifier Harper et al. 2006 Emergent Lehtonen et al. 2007 947 948 949 950 951 952 953 41 External feature3 predation, social environment, environment Manipulation Internal feature3 personality, self condition, age, experiment production / content signals production signal design Sender transmission efficacy 954 receiver information processing1 * decision rule choice2 reception Information extraction 955 Figure 1. Schematic view of sexual communication with multiple signals. Conventionally, mate choice implies active sampling of potential 956 mates or assessment of varying traits, but mating biases can sometimes result from passive acceptance of mates (see Kokko et al. 2003). Varying 957 traits that directly affect mating decision are also called “cues”. Traits or “cues” which have been modified by natural selection for the purpose of 958 communication are called “signals” (Candolin 2003). In communication terminology, signals are emitted by a sender and processed by a 42 959 receiver. The receiver perceives and processes signals before adjusting its behaviour according to their information content or their stimulating 960 power and their decision rule. Selection favours senders whose signals affect the behaviour of receivers to their own advantage but also favours 961 receivers who are able to extract useful information from the senders (Font & Carazo 2010). 1-Information processing/decision rule could be 962 hierarchical, additive or multiplicative (see text), 2-Correspondance to behaviour (interest or not, copulated or not, pair bond or not), 3-external 963 and internal information which could modulate signal production, signal design and information processing by changing the assessment of 964 signals. 43 ARTICLE 2 Multiple signals in the palmate newt: ornaments help when courting. Cornuau JH, Rat M, Schmeller DS, Loyau A. 2012. Behavioral Ecology and Sociobiology, 66:1045-1055. Behav Ecol Sociobiol (2012) 66:1045–1055 DOI 10.1007/s00265-012-1355-y ORIGINAL PAPER Multiple signals in the palmate newt: ornaments help when courting Jérémie H. Cornuau & Margaux Rat & Dirk S. Schmeller & Adeline Loyau Received: 12 December 2011 / Revised: 28 March 2012 / Accepted: 29 March 2012 / Published online: 18 April 2012 # Springer-Verlag 2012 Abstract There is increasing evidence that female mate choice is often based on the assessment of multiple male traits, involving both morphology and behavior. We investigated female mate choice for multiple male traits in the palmate newt, Lissotriton helveticus, including male tail filament length, hind foot web size, crest development, body size, ventral coloration, and courtship display activity. Observations of courtship display in the field revealed that females spent more time in front of males with longer tail filaments. Laboratory experiments revealed a more detailed relationship between filament length and courtship display. We found that females took more sperm masses from males with both longer filaments and greater display activity. Experimental shortening of the tail filament length substantially decreased the number of male sperm masses transferred. However, when we experimentally reversed relative filament length between two males in mating trials, male mating success was explained by courtship activity and not by filament length. Our results show that female palmate newts value multiple traits during mate choice, including both morphological ornaments and reproductive behaviors in males. Our results further suggest that, when filament length Communicated by K. McGraw Dirk S. Schmeller and Adeline Loyau share senior authorship in this study. Electronic supplementary material The online version of this article (doi:10.1007/s00265-012-1355-y) contains supplementary material, which is available to authorized users. J. H. Cornuau (*) : M. Rat : D. S. Schmeller : A. Loyau Station d’Ecologie Expérimentale du CNRS USR 2936, 2-4 Route du CNRS, 09200 Moulis, France e-mail: [email protected] is below a certain threshold, females may value the information content of courtship activity over that of filament length. Keywords Sexual selection . Sexual communication . Courtship . Filament . Amphibian . Lissotriton helveticus Introduction In many species, males express complex courtship displays involving elaborate ornaments combined with conspicuous behaviors. The evolution and maintenance of such morphological and behavioral traits are commonly explained by their signaling role in intersexual communication (Møller and Pomiankowski 1993; Candolin 2003; Bro-Jørgensen 2010). Multiple signals that potentially could be used for mate choice were empirically identified before the first theoretical explanations had been proposed. Indeed, theoretical models of sexual selection generally predict that females should evolve a preference for a single signal because of the costs of assessing multiple signals (Schluter and Price 1993; Johnstone 1996; Fawcett and Johnstone 2003; but see van Doorn and Weissing 2004). However, the idea that females do use multiple signals to choose their mate is supported by an increasing number of empirical studies, leading to a growing interest in the evolution and maintenance of such signals (Candolin 2003; Hebets and Papaj 2005; Bro-Jørgensen 2010). Several hypotheses have been put forward to explain why and how females use numerous signals: multiple messages (different signals convey information on different properties or temporal aspects of male qualities), redundant signal hypothesis (different signals convey the same information), unreliable signal hypothesis (signals not linked to quality), or receiver psychology (several signals are better detected than a single one, or a signal amplifies another signal; 1046 Guilford and Dawkins 1991; Møller and Pomiankowski 1993; Rowe 1999; Hebets and Papaj 2005). Disentangling these hypotheses mainly requires unraveling the information content of multiple signals (Candolin 2003). To date, less work has been undertaken to examine the relative use of each signal, their prioritization, and how signals are processed by choosy females (Witte and Ryan 1998; Künzler and Bakker 2001; Candolin 2003; Scheuber et al. 2004). Recently, it was proposed that a fluctuating environment is likely to drive the evolution of multiple signals because few signals may be sufficiently flexible to remain reliable when the environment changes (Chaine and Lyon 2008; Bro-Jørgensen 2010; Cornwallis and Uller 2010). Morphological and behavioral signals commonly differ in their temporal flexibility. Some morphological signals, such as an elongated tail or elaborate plumage, are relatively fixed (static sensu Hill et al. 1999) since they cannot be extensively modified on a short time scale by the signaler. Morphological signals are generally produced once per year before or at the beginning of the breeding season and remain unmodified until the next year (Hill et al. 1999; Loyau et al. 2005a, b; Baird et al. 2007; van Dongen and Mulder 2008). Morphological signals are thus expected to better reflect genetic quality or long-term viability (Kokko et al. 1999; Scheuber et al. 2004). However, if the bearer’s quality or the environmental conditions change between production and assessment, morphological signals may become unrepresentative of the signaler’s quality (Bro-Jørgensen 2010). On the contrary, behavioral signals such as vigorous displays or song rates can be finely tuned by the signaler depending on the signaler’s current quality, motivation, or environmental conditions. Behavioral signals are typically more labile (dynamic; sensu Hill et al. 1999) than morphological ones and thus are expected to accurately reflect signaler condition at the time of assessment (Hill et al. 1999; Loyau et al. 2005a; Baird et al. 2007; van Dongen and Mulder 2008). Given that individual quality and environmental conditions may vary on a short time scale, it has been predicted that females should use a mixture of static morphological and dynamic behavioral signals to adaptively choose their mates (Bro-Jørgensen 2010). Females are further expected to differ in the degree to which they pay attention to these two types of signals during mate choice. This potential difference can only be assessed when both signals are contradictory, e.g., when a male exhibits an ornament of high quality while expressing poor courtship behaviors or vice versa. However, it is still disputed how females deal with such contradictory information (Hill et al. 1999; Loyau et al. 2005a; Baird et al. 2007; van Dongen and Mulder 2008). One can predict that females should value the information content of behavioral signals over morphological signals because behavioral signals provide more accurate information on current male condition (Hill et al. 1999; Loyau et al. Behav Ecol Sociobiol (2012) 66:1045–1055 2005a; Baird et al. 2007; Bro-Jørgensen 2010). This is particularly the case when females prefer mating with healthy individuals not only to obtain indirect benefits, in terms of good genes for their offspring or more viable sperm, but also to avoid contracting infectious diseases, since a recent decline in male condition could indicate a challenged immune system (Sheldon 1993; Loyau et al. 2005a; Chargé et al. 2010). On the contrary, it has been suggested that the expression of dynamic (behavioral) signals is sensitive to very short term, motivational influences, social circumstances, or stochastic environmental factors, and thus these signals may not be as valuable to conspecifics as fixed signals (Green et al. 2000; van Dongen and Mulder 2008). We wished to test this set of predictions in the palmate newt, Lissotriton helveticus, an amphibian with strong sexual dimorphism (Griffiths and Mylotte 1988). Males are smaller and thinner than females. Males express a yellowto-dark-orange ventral coloration whose role remains unclear. During the reproductive period, males exhibit both morphological and behavioral traits that may be assessed by females when choosing a mate. Morphological male traits that appear to have a sexual function are the hind foot web, a caudal filament, and a low dorsal caudal crest. In the crested newt Triturus cristatus and the smooth newt Lissotriton vulgaris, females prefer males with high crests (Hedlund 1990a; Green 1991; Gabor and Halliday 1997; Malmgren and Enghag 2008). In the palmate newt, Haerty et al. (2007) investigated female visual preference for filament length and body size and found that females spent more time close to males with longer filaments and smaller bodies. In addition to morphological traits, male L. helveticus express a conspicuous courtship behavior called “fan display”, which mainly consists of displaying the tail and the filament to the female in a rapid vibrating fan movement (Halliday 1975a). This display behavior is very variable between individuals (Halliday 1975a, b; 1976). If the female is receptive to the displaying male, she touches his tail with her snout, which elicits spermatophore deposition by the male. The female can then pick up the sperm mass in her cloaca. We used an approach combining: (1) a correlative study in the field; (2) a correlative study in the laboratory; and (3) a manipulative study of a morphological signal in the laboratory. We first examined the impact of tail filament length on the maximum time spent by the female watching a male fan display in the field. We then examined whether female palmate newts use multiple morphological and behavioral traits to choose their mates, focusing particularly on the role played by both filament length and courtship activity, while controlling statistically for potential additional morphological signals (including crest development, hind foot web, body size, and coloration). Finally, we manipulated filament length of pairs of males so that the male with the shorter filament became the male with the relatively longer filament, while courtship Behav Ecol Sociobiol (2012) 66:1045–1055 activity was not manipulated. We predicted that, if females only value filament length, female preference should be reversed. Materials and methods Correlative study in the field Our study was designed to investigate whether females spent more time watching courtship displays by males with longer tail filaments. In amphibians, female preference has rarely been observed in the wild, and such observations are likely to provide valuable insights into the evolutionary processes involved in sexual selection and are complementary to laboratory experiments. Our behavioral observations of newts were conducted in two small ponds near the Refuge des Etangs de Bassiès in the Pyrenean Mountains (Ariège, France), at 1,655 m above sea level, between 30 June and 09 July 2010, during the peak of the reproductive season. These study ponds are ideal for behavioral observations because they are shallow with crystal clear water, were only a few meters apart, and had little vegetation and few rocks. The behavioral observations were performed for 5 days between 07:00 am and 07:00 pm. The average male courtship fan display lasted 237.5±25.5 s and was not significantly different between observation days (Kruskal–Wallis tests χ²01.95, df04, P00.75), between morning and afternoon (Wilcoxon rank sum test W0150, P00.39), and between the two ponds (W0121, P00.62). Reproductive behaviors of newts were well described by Halliday (1975a). The male first approaches the female. He then smells the female cloaca, follows her when she moves away, and orients himself in front of the female. After this orientation phase, courtship activity begins, defined by “fan” movements of the male tail and filament in front of the female. The female either stays still, seemingly watching male fan display or moves away. When the female moves, the male stops the fan movement. A new sequence of orientation and fan display can then occur. Sperm mass transfer is rarely observed in field conditions (Hedlund 1990b). We used the time spent by females watching courtship fan display as a proxy for female preference (Wagner 1998). Behavioral observations were conducted by one observer (JC) who walked around the two ponds (at 1-m distance) until seeing a male in either the orientation or displaying phase. During behavioral observations, we recorded the time between the start of the male display and the end, due to departure of the female or the male surfacing to breathe. Taking a breath is a critical moment in courtship by male newts, as they are likely to lose visual contact with the female to whom they were displaying (JC, personal observation). Therefore, males increase the time between breathing episodes when the females are receptive (Halliday and Sweatman 1047 1976; Halliday 1977). We obtained several displaying periods for each male. We used length of the longest bout of the male display to a female for statistical analysis. Males were caught to measure snout–vent length (SVL, from the snout tip to the posterior end of the cloaca) and filament length with a caliper (accuracy, 0.01 mm). We recorded individually distinctive features such as shape, black spots, or coloration by drawing the male on paper. These drawings and the two measures (SVL and filament length) allowed us to ensure that a given male was not included twice in our data set. Each male was then immediately released at the site of capture. A total of 33 males (24 and 9 per pond) were observed. Laboratory experiments Newts caught in the wild were maintained in captivity at the Station d’Ecologie Expérimentale du CNRS à Moulis (Ariège, France) at 18 °C (±0.5 °C) and a natural light/dark cycle of 10:14 h (ZooMed ReptiSun 2.0 fluorescent bulb). Unisex groups of ten individuals were placed in opaque tanks (52×33.5×29.5 cm) with de-chlorinated water and some plants collected in their native environment. The water depth was about 5 cm, corresponding to about 9 l of water. Each tank was provided with a shelter (a perforated clay brick with 17 holes). The top of each brick was just above the water level to allow newts to get out of the water. Newts were fed ad libitum with chironomid larvae, daphnia, tubifex, and earthworms. Time between capture and release back into the wild was less than a month. Behavioral experiments were designed to offer two males to a single female and were conducted in the animal housing facility between 0800 and 1500 hours local time. Experimental tanks were divided into three compartments by two removable transparent perforated Plexiglas separations, defining a central compartment (20×33.5×29.5 cm) and two external compartments (16×33.5×29.5 cm). We followed a protocol used by Haerty et al. (2007) with the palmate newt, to be able to compare our results with the previous results (data not shown). The experiment comprised four steps. In step 1, a focal female was placed in the central compartment where she could move freely for 5 min, allowing the female to explore her new environment. In step 2, the experimentalist placed the female in the center of her compartment in a circular opaque PVC tube for 5 min, and a male was placed in each external compartment. The tube was then removed so that the two males and the female could move freely in their compartments and explore their new environment for 15 min (step 3). Finally, in step 4, separations were removed and the three individuals could interact freely for 40 min. At the end of the experiment, we checked for the presence of sperm mass in the tank and we examined the female’s cloaca for sperm mass transfer. Each individual was used only once in the experiment and the experimental tank was thoroughly 1048 Behav Ecol Sociobiol (2012) 66:1045–1055 cleaned with tap water between trials. Four trials were performed simultaneously in four experimental tanks. Each trial was recorded using a video camera (Panasonic HDCTM60, full HD, 1080p) placed above the experimental tank, totaling 65.3 h of video observation. Morphological traits were measured immediately after the experiment. We recorded body mass (accuracy, 0.01 g), SVL (snout–vent length), and filament length as previously described using an electronic caliper. The crest is small in this species, so we used tail size as an index of crest development (Green 1991). Tail area and the left and right hind foot web areas were measured by taking a photo on a millimeter paper background (grid 1 mm²) with a digital camera (Nikon D5000). Photo analysis was conducted with ImageJ 1.43 (http://rsb.info.nih.gov/ij; developed by Wayne Rasband; Abramoff et al. 2004). With a subsample of 60 males, we tested the reliability of two measurements with ImageJ. Filament length, hind foot web area, and crest development were all highly repeatable (respectively, r²0 0.99, r²00.98, and r²00.96). Ventral coloration was measured using a spectrometer (USB2000, Ocean Optics Inc., Dunedin, FL, USA) with a Xenon light source (PX-2) and a R400-7-UV/VIS fiber optic cable and probe at a 45° angle to the skin surface. Five spectrometer measures were taken at the middle of the belly. We used Endler’s segment classification method on percent reflectance between 300 and 700 nm to obtain indexes of hue, chroma, and brightness (Endler 1990). We tested the repeatability of these indexes with Kendall’s coefficient of concordance (Loyau et al. 2007) with a subsample of 100 individuals. Hue, chroma, and brightness were highly repeatable (respectively, W0 0.93, χ²0361, P<0.001; W00.95, χ²0471, P<0.001; and W00.73, χ²0361, P<0.001). The videos were analyzed with The Observer 7.0 software (Noldus Information Technology). For each male, we recorded the time spent displaying the fan (courtship activity) and the number of sperm masses transferred to the female. Experiment 1: female preference of filament length This experiment was designed to test whether palmate newt females prefer males with longer tail filaments. In total, 76 males and 38 females were used to perform 38 replicate trials. Each trial involved three individuals: one female, one male with a long filament (M+), and one male with a shorter filament (M−). The pair of competitor males was chosen so that males significantly differed in filament length (paired t test, t 014.78, df 037, P < 0.001) and this difference remained stable over trials (Table 1). SVL, body mass, hue, chroma, brightness, and tail area (index of crest development) did not significantly differ between M+ and M− males (paired t test, all P>0.05). However, hind foot web area and filament length were correlated (r00.44, df074, P< 0.001), and M+ males had more developed hind foot web areas than M− males (paired t test, t02.8, df037, P00.008). Experiment 2: manipulation of filament length In this experiment, we manipulated filament length to test whether a decrease in filament length decreased the number of male sperm masses transferred. Manipulation of a trait (phenotypic engineering) allowed us to avoid potentially confounding factors (Sinervo and Basolo 1996). Filament length could not be elongated because of technical difficulties. Many other studies have used only sexual trait reduction in their experimental manipulation for similar reasons (e.g., Bischoff et al. 1985; Petrie and Halliday 1994; Tomkins and Simmons 1998; Pryke and Andersson 2005). In total, 120 males and 60 females were used in experiment 2 to perform 60 replicate trials. One week before the start of each experimental trial, male filament length was measured to rank the males according to their filament length and to choose the pairs of male competitors as in experiment 1. However, unlike experiment 1, male filaments were immediately cut after measurement, so that the male with the longer filament in the pair of males became the male with the shorter filament (Table 1). Both males in each trial had their filament length shortened to control for a possible effect of the cut itself (Table 1). Filament length was not measured immediately after cutting to avoid further stress to the individuals. The individuals were left un-manipulated until trial. Filament length was measured a second time just after the trial, and that measurement was used for statistical analyses. As in experiment 1, females had the choice between a male with a Table 1 Measures of filament length (millimeters) between experiments (means±SE) Exp 1 M+ males M− males |M+ −M−| 6.39±0.10 4.40±0.19 1.99±0.14 Exp 2 (BS) Before shortening Exp 2 (AS) After shortening P value 6.25±0.12 4.46±0.09 1.79±0.09 2.02±0.07 3.75±0.09 1.74±0.09 0.001 0.001 0.31 Equalities Exp 10exp 2 (BS) Exp 10exp 2 (BS) All “|M+ −M− |” represents the difference in filament length between the males with longer filaments (M+ male) and the males with smaller filaments (M− male) for a given experiment. P values correspond to the results of the linear model that explains filament length by experiment. Equalities show that filament length was not significantly different between experiments (Tukey’s tests) Behav Ecol Sociobiol (2012) 66:1045–1055 longer filament (M+ male) and a male with a shorter filament (M− male). The SVL, body mass, hue, chroma, brightness, and tail area did not differ between M+ and M− males (paired t test, all P>0.05). Again, hind foot web area and filament length (before cut) were positively correlated (r0 0.44, df0118, P<0.001), as a consequence M− males had more developed hind foot web areas than M+ males (paired t test, t03. 9, df059, P<0.001). We performed a study involving five newts to ensure that shortening the male filament would not result in distress. We used the same design describe previously (see Fig. S1). We did not observe any behavioral differences for a male before, immediately after, or 1 week after shortening the filament (Fig. S1). Moreover all males used in experiment 2 behaved normally after filament shortening. Statistical analysis In the field experiment, the effect of male SVL and filament length on the time a female spent watching male display (log-transformed) was assessed with a model II regression using an ordinary least squares method. For the laboratory experiments, we calculated indexes of hind foot web and crest developments as the residuals of a regression of the square root of hind foot web area or tail area on the SVL (Baker 1992). Because morphological traits were strongly correlated, we performed a principal component analysis (PCA) on male traits (SVL, body mass, hue, chroma, and brightness) using the dimdesc function of the FactoMineR package in R (R Development Core Team 2010). We did not include filament length, crest development, and hind foot web size in the PCA to fully explore their explanatory potential with model selection. Including or not, filament length, crest development, and hind foot web size in the PCA did not change the overall conclusions of the study. The PCA resulted in principal components of body size (PCbody size, loading on SVL, and body mass), of coloration (PCcoloration, loading on hue, and chroma), and of brightness (PCbrightness, loading on brightness; Table S1). The effects of sexual traits on sperm mass transfer were assessed using generalized linear mixed models with the lmer function of the lme4 package in R (R Development Core Team 2010). Morphological traits (filament length, hind foot web size, crest development, PCbody size, PCcoloration, and PCbrightness) and the behavioral trait (courtship activity) were set as fixed factors, and tank, day-of-trial, and trial identity were included as random factors. To investigate the effects of male traits on the number of sperm mass transferred to the females, we used models with a Poisson distribution of error terms and the log link function. We analyzed data and made biological inferences using the information theoretic approach that identifies the best set of models according to several 1049 competing hypotheses via information criteria (Burnham and Anderson 2002; Garamszegi et al. 2009; Burnham et al. 2011; Grueber et al. 2011). We evaluated a set of candidate models transferred to combinations of all morphological and behavioral variables and all interaction terms using Akaike’s Information Criterion corrected for small sample size (AICc; Burnham and Anderson 2002).We used the stdz.model and top.models functions of the arm and MuMIn packages in R to calculate AICc, AICc weight, and other parameters needed for the analysis (Grueber et al. 2011). We present only models that were well supported by the data (models with a ΔAICc <2, Burnham and Anderson 2002). Because several equally likely models were identified and no model had w >0.90, we model-averaged parameter estimates (“θ”) and the associated variances from the 95 % confidence interval (CI) set of candidate models using the model.avg function of the MuMIn package in R (Grueber et al. 2011). We considered the effect of one variable as having an important predictive value only when the 95 % CI did not include zero (Garamszegi et al. 2009; D’Alba et al. 2010). We calculated the coefficient of determination (percentage of deviance explained) of the best models as follows: (deviance of the null model − deviance of the model)/deviance of the null model. We additionally used chi-square tests in experiment 1 and experiment 2 to compare the global sperm mass transfer of males with longer and shorter filaments with and without cuts. Results Correlative study in the field In our correlative study in the field, male size and filament length were uncorrelated (N033, r00.16, P00.36). The time spent by a female watching a male display was explained by male filament length (N 033, r 00.42, P 0 0.02, Fig. 1) and not by male size (N033 r00.04, P00.81). Laboratory experiments During experiment 1, 89.5 % (34/38) of females accepted at least one sperm mass (60.5 % of females accepted a sperm mass from a M+ male, 18.4 % from a M− male, and 10.5 % from both males). Among males, 71.1 % of M+ and 28.9 % of M− males successfully transferred a sperm mass to a female. The latency between the beginning of the experiment and the first courtship display did not differ significantly between M+ and M− males (Wilcoxon signed rank test, V0392, N076, P00.98). Four models were retained to explain the number of sperm masses transferred in experiment 1, among which the best had a coefficient of determination of 33.3 % (Table S2). Filament 100 200 300 400 500 600 Behav Ecol Sociobiol (2012) 66:1045–1055 0 Time spent by female watching male displays (s) 1050 0 2 4 6 Filament length (mm) Fig. 1 Relationship between the time spent by a female watching male display and male filament length (N033, y025.5x+50.15, r00.42, P0 0.02). The variable “Time spent watching” represents the length of the longest bout of the male display to a female length, hind foot web size, PCbody size, and courtship activity (but not PCbrightness and crest development) were included in at least one model (Table S2). The number of sperm masses transferred to a given female was best explained by both male filament length and courtship activity, with females taking more sperm masses from males having longer filaments and higher courtship activity (Table S2, Table 2, Fig. 2). In experiment 2, the reduction of filament length decreased the number of sperm masses transferred. During the entire experiment, only 53.3 % (32/60) of females accepted at least one sperm mass, and this number was significantly smaller than in experiment 1 (χ 2 012.2, df 01, P 00.001). The number of sperm masses transferred of M− males (after filament shortening) was significantly smaller compared to M+ males in experiment 1 (Wilcoxon rank sum test, W 01,666.5, N0 98, P < 0.001, Fig. 3), while the number of sperm masses transferred of M+ males (after filament shortening) was similar compared to M− males in experiment 1 (Wilcoxon rank sum test, W01,163, N 098, P00.831, Fig. 3). In experiment 2, the latency between the beginning of the experiment and the first courtship did not differ between M+ and M− males (Wilcoxon signed rank test, W0832, N0120, P 00.544). Five models were selected by AICc and all contained courtship activity (Table S2). The remaining variables included in at least one model were hind foot web size, filament length, and PCcoloration (but not PCbrightness and crest development; Table S2). The top ranked model explaining male mating success had a coefficient of determination equal to 26 %. Model-average parameter estimates showed that the number of sperm masses transferred increased with courtship activity (Table 2, Fig. 4). However, contrary to experiment 1, filament length had very little influence on the number of sperm masses transferred (Table 2). Discussion Our approach combining behavioral observations in the field and in the laboratory showed that female palmate newts value both a morphological and a behavioral trait Table 2 Effects of morphological and behavioral sexual traits on the number of sperm mass transfers for experiment 1 (exp1) and experiment 2 (exp2) Parameter exp1 (no. of sperm mass transfers) exp2 (no. of sperm mass transfers) Estimate Unconditional SE Confidence interval Relative importance Intercept Hind foot web PCbody size Filament length −0.52 0.42 −0.36 1.46 0.18 0.25 0.26 0.44 (−0.87, −0.17) (−0.08, 0.93) (−0.88, 0.16) (0.61, 2.32) 0.55 0.44 1 Courtship activity Intercept PCcoloration Filament length Hind foot web Courtship activity 1.06 −1.53 −0.33 −0.13 −0.53 2.48 0.35 0.25 0.32 0.33 0.34 0.47 (0.38, 1.74) (−2.02, −1.04) (−0.96, 0.31) (−0.78, 0.51) (−1.2, 0.14) (1.55, 3.41) 1 0.32 0.12 0.61 1 In exp 1 female had the choice between two males (respectively with long and short filament length). In exp 2, we reversed the relative length of filament before trial. The bold values indicate the variables selected after model averaging (i.e., for which the 95 % confidence interval did not include zero). Fixed effects present in the full models: filament length, hind foot web, crest development, PCbody size, PCcoloration, PCbrightness, and courtship activity. For details on the PCA see Table S4 2.5 2.0 1.5 1.0 0.5 0.0 Number of sperm mass transfers a 1051 3.0 Behav Ecol Sociobiol (2012) 66:1045–1055 1 2 3 4 5 6 7 2.5 2.0 1.5 1.0 0.5 0.0 Number of sperm mass transfers b 3.0 Filament length (mm) 0 200 400 600 800 1000 Time in courtship (s) Fig. 2 Relationship between the number of sperm masses transferred and a filament length and b the time spent in courtship during experiment 1. The females had the choice between a male with a long filament and a male with a short filament length. Filament length was not manipulated when choosing a mate: the length of the tail filament and courtship display activity. It further suggests that, when filament length is experimentally reduced, females may value the information content of courtship activity over filament length. Female mate choice for males with long filaments Direct observations in the field revealed that females spent more time watching males that had longer filaments. Our result is congruent with previous work suggesting that filament elongation may be the consequence of female preference in this species (Haerty et al. 2007). In a second step, we experimentally tested female preference for filament length in the laboratory. In this experiment, males with longer filaments transferred more sperm masses to females than males with shorter filaments, as predicted by our observations in the wild. However, we could not reject the hypothesis that female preference was directed to an unknown factor correlated with filament length (e.g., hind foot web size). We therefore experimentally reduced filament length, while other traits were not manipulated. This reduction decreased the number of male sperm masses transferred compared to what was observed in the previous experiment. Altogether, our three experiments provided strong evidence that female palmate newts value the development of a morphological ornament, the filament length, to choose their mates. Several not mutually exclusive hypotheses can explain why female palmate newts evolved a preference for males with long filaments. First, such a preference could result from a sensory bias, so that a higher length of filaments catch a female’s attention better (Basolo 1990; Ryan and Rand 1990), as filaments strongly resemble the worms that represent a major food resource for (female) newts. Filament length could also be a reliable indicator of male condition, so that only males in good condition can produce and maintain long filaments, as previously reported for interactions between filament length and environmental stress (Secondi et al. 2007, 2009). In our study, we observed a pre-copulatory female preference for males with longer filament length. However, female newts can accept many sperm masses from the same male and from different males, likely to allow sperm competition and cryptic post-copulatory female choice (Jones et al. 2002; Garner and Schmidt 2003; Jehle et al. 2007). Females of the alpine newt Mesotriton alpestris and smooth newt L. vulgaris show cryptic mate choice, by favoring sperm of particular males over others after accepting sperm masses (Garner and Schmidt 2003; Jehle et al. 2007). Further work is therefore needed to evaluate true fitness consequences of the preference we observed in terms of number of offspring produced. Female mate choice for males with high courtship activity Our results failed to support a direct role of coloration and/ or body size on female preference, although there is increasing evidence of the use of coloration for mate choice in amphibians (Rosenthal et al. 2004; Vasquez and Pfennig 2007; Maan and Cumming 2009; Gomez et al. 2009; Doucet and Mennill 2010; Sztatecsny et al. 2010), including newts (Davis and Grayson 2008). Male filament length Fig. 3 Number of sperm masses transferred during experiments 1 and 2 (without or with filament shortening). M+ male with longer filament length, M− male with shorter filament length, exp 1 without filament shortening, and exp 2 after filament shortening. The M− males from this experiment had longer filaments before the cut and the M+ had smaller filaments before the cut (see Table 1) Behav Ecol Sociobiol (2012) 66:1045–1055 Number of sperm mass transfers 1052 *** 1.8 1.6 1.4 1.2 1 NS 0.8 0.6 0.4 0.2 0 M+ (exp 1) was, however, not sufficient to explain all of the variation in male sperm masses transfer in our study, and courtship activity also strongly predicted male sperm mass transfer. This result indicates that a male with a small filament can acquire access to reproduction if this male has a high courtship activity and that a male with a long filament but with a very low courtship activity has a low likelihood to breed. The importance of courtship to explain reproductive success raises two conclusions. First, males increased their likelihood to breed by increasing their courtship activity, and second, females selected males with the higher courtship activity. Newt courtship activity is well described (Halliday 1975a; 1976), yet its potential to contribute to female mate choice has received surprisingly little attention (but see Vinnedge and Verrell 1998). Females of many species prefer males exhibiting vigorous courtship displays, and measurements of courtship energy expenditure show that this activity may be costly (Vehrencamp et al. 1989; Kotiaho et al. 1998; Usherwood 2008; Hasselquist and Bensch 2008; Byers et al. 2010; but see Dearborn et al. 2005). Courtship can thus reveal many aspects of male quality, such as the ability to cope with an immune challenge (Loyau et al. 2005a), body condition (Holzer et al. 2003), ejaculate quality (Matthews et al. 1997; Weir and Grant 2010), or heterozygosity (Drayton et al. 2010). Female mate choice for multiple traits Our study demonstrates that female palmate newts value multiple traits during mate choice. Why males produce and females assess several traits, and not a single one despite the associated costs, remains an open issue in evolutionary ecology (Møller and Pomiankowski 1993; Candolin 2003; Bro-Jørgensen 2010). During courtship display, males present and agitate their filament in front of females, a behavior likely improving the perception and facilitating the assessment of the M-(exp 2) M-(exp 1) M+ (exp 2) filament length (Rowe 1999). Amplifiers may entail costs by increasing conspicuousness to predators, thereby potentially revealing their ability to escape predators (“amplifying handicaps” Hasson 1990; 1997; Berglund 2000; Castellano and Cermelli 2010). In line with this idea, male palmate newts can suffer from a high rate of filament predation in habitats with high density of odonates and other insect larvae (JC, personal observation). Because ornamentation acts in conjunction with visual display, it is not clear whether the exhibition of an ornament during courtship should be considered as several signals or a multi-component signal, and whether a visual display could be viewed as an amplifier of the ornament (Hasson 1989; 1990; 1991; Taylor et al. 2000) or whether the filament length may amplify the expression of the behavior (Hebets and Uetz 2000). Morphological ornamental traits, such as the filament, and behavioral traits (e.g., the courtship) vary in their flexibility and may also vary in their reliability, depending on the stability of the ecological and/or social environment (Hill et al. 1999; Loyau et al. 2005a, b; Baird et al. 2007; van Dongen and Mulder 2008; Bro-Jørgensen 2010). For example in field crickets Gryllus campestris, females based their mate choice according to carrier frequency (link to harp size development) and chirp rate, which provide reliable information about past and present condition (Scheuber et al. 2003, 2004). Assessing both filament length and courtship activity, female palmate newts could therefore gather information about male past and present conditions to adaptively mate with high-quality males (multiple message hypothesis, Møller and Pomiankowski 1993). However, to date, experimental evidence about the relative flexibility of filament length and courtship activity is lacking in the palmate newt. In newts, morphological sexual traits could indicate male body condition (Green 1991). Moreover, the link between behavioral sexual traits and current male quality remains unresolved in newts (but see Halliday and 2.5 2.0 1.5 1.0 0.5 0.0 Number of sperm mass transfers a 1053 3.0 Behav Ecol Sociobiol (2012) 66:1045–1055 1 2 3 4 5 2.5 2.0 1.5 1.0 0.5 0.0 Number of sperm mass transfers b 3.0 Time in courtship (s) 0 200 400 600 800 1000 1200 Time in courtship (s) Fig. 4 Relationship between the number of sperm masses transferred and a filament length and b the time spent in courtship during experiment 2. The females had the choice between a male with a longer filament length and a male with a shorter filament length. Filament length was shortened before trial, so that, in the pair of males, the male with the longer filament length became the male with the shorter filament Houston 1978). Additional work with newts and other species is needed to clarify if morphological and behavioral sexual traits could give information about past and current male condition. Female prioritization of behavioral traits over morphological ones? While evidence is accumulating that females often use multiple traits to choose their mate, relatively few studies have explored how these traits are processed and integrated. Choosers may assess multiple traits altogether to get a global idea of the overall quality of potential mates, or choosers may value one piece of information over another, depending on the degree to which the information conveyed is error-prone or particularly useful (Hebets and Papaj 2005). Because morphological and behavioral traits strongly differ in their temporal sensitivity to individual condition, females may weigh morphological and behavioral traits differently in the mate choice process. Our results show that when filaments were experimentally reduced and their relative length was reversed, male mating success was explained by courtship activity and not by filament length, suggesting that, in that case, females base their mate choice on courtship activity. In this experiment, the difference in filament length (2 mm) between the two males was comparable to that in experiments without filament length manipulation. According to the Weber–Fechner law, this difference should be easier to assess for smaller filament length than for longer filaments (Shettleworth 1998). Recently Akre et al. (2011) elegantly demonstrated the relevance of this law during the mate choice process. The authors showed that, for the same discrepancy in ornament size, females were better able to detect the difference in two small ornaments than in two large ornaments. Thus, there is little likelihood that, in our experiment, female newts were not able to detect the difference in filament length after filament manipulation. Instead, our manipulation may have reduced filament length below a certain stimulation threshold, so that females may not have been stimulated by either of the two males. Another interpretation is that females may have ignored an ornament (the filament) as a cue for mate choice because females detected a lack of concordance in the development of two ornaments (the filament length and the hind foot web size) that are normally correlated. Indeed, females may commonly use several signals to avoid mistakes (Johnstone 1996). In a recent experiment in the tree frog Hyla arborea, females were provided with the choice between two conflicting sexual traits, attractive call and unattractive vocal sac coloration or unattractive call and attractive vocal sac coloration. This trial resulted in random female mating with respect to the acoustic and visual traits manipulated (Richardson et al. 2010). Finally, our results are consistent with the hypothesis that, when the morphological signals are expressed below a threshold, females may value the information content of behavioral traits over morphological ones because behavioral traits provide more accurate and reliable information on current male condition (Hill et al. 1999; Loyau et al. 2005a; Baird et al. 2007; Bro-Jørgensen 2010). To conclude, we found that female palmate newts use multiple signals to choose their mates. Hence, in that species, 1054 females value both a morphological trait (the filament) and a behavioral trait (courtship activity). Future investigation should establish whether these two traits that may differ in their temporal flexibility provide different kinds of information to the female. Acknowledgments We are grateful to Sandrine Plénet and Claude Miaud for discussing the results; Bob Montgomerie, László Z. Garamszegi, and Philipp Heeb for discussions, statistical advice, and comments on the manuscript; Géraldine Domisse and Thomas Jolly for help in gathering data from experiment 2; and Olivier Calvez for technical assistance. We are grateful to Philippe and Dominique Dupui, keepers of the Refuge des Etangs de Bassiès for their help. The catching permits were no. 200913 (Ariège) and no. 2009-12 (Haute Garonne). This work was supported by the Ministère de la Recherche (PhD fellowship to JC and two CNRS grants to AL and DS), the BioDiversa-project RACE, and the Observatoire Homme-Milieu Pyrénées Haut-Vicdessos. Ethical standards The experiments comply with the current laws of the country in which they were performed. References Abramoff MD, Magelhas PJ, Ram SJ (2004) Image processing with ImageJ. Biophoton Int 11:36–42 Akre KL, Farris HE, Lea AM, Page RA, Ryan MJ (2011) Signal perception in frogs and bats and the evolution of mating signals. Science 333:751–752 Baird TA, Hranitz JM, Timanus DK, Schwartz AM (2007) Behavioral attributes influence annual male mating success more than morphology in collared lizards. Behav Ecol 18:1146–1154 Baker JMR (1992) Body condition and tail height in great crested newts, Triturus cristatus. Anim Behav 43:157–159 Basolo AL (1990) Female preference predates the evolution of the sword in swordtail fish. Science 250:808–810 Berglund A (2000) Sex role reversal in a pipefish: female ornaments as amplifying handicaps. Ann Zool Fenn 37:1–13 Bischoff RJ, Gould JL, Rubenstein DI (1985) Tail size and female choice in the guppy (Poecilia reticulate). Behav Ecol Sociobiol 17:253–255 Bro-Jørgensen J (2010) Dynamics of multiple signaling systems: animal communication in a world in flux. Trends Ecol Evol 25:292–300 Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a practical information–theoretic approach. Springer, New York Burnham KP, Anderson DR, Huyvaert (2011) AIC model selection and multimodel inference in behavioral ecology: some background, observation, and comparisons. Behav Ecol Sociobiol 65:23–35 Byers J, Hebets E, Podos J (2010) Female mate choice based upon male motor performance. Anim Behav 79:771–778 Candolin U (2003) The use of multiple cues in mate choice. Biol Rev 78:575–595 Castellano S, Cernmelli P (2010) Attractive amplifiers in sexual selection: where efficacy meets honesty. Evol Ecol 24:1187–1197 Chaine AS, Lyon BE (2008) Adaptive plasticity in female mate choice dampens sexual selection on male ornaments in the lark bunting. Science 319:459–462 Chargé R, Saint Jalme M, Lacroix F, Cadet A, Sorci G (2010) Male health status, signalled by courtship display, reveals ejaculate quality and hatching success in a leeking species. J Anim Ecol 79:843–850 Cornwallis CK, Uller T (2010) Towards an evolutionary ecology of sexual traits. Trends Ecol Evol 25:145–152 Behav Ecol Sociobiol (2012) 66:1045–1055 D’Alba L, Shawkey MD, Korsten P, Vedder O, Kingma SA, Komdeur J, Beissinger SR (2010) Differential deposition of antimicrobial proteins in blue tit (Cyanistes caeruleus) clutches by laying order and male attractiveness. Behav Ecol Sociobiol 64:1037–1045 Davis AK, Grayson KL (2008) Spots of adult male red-spotted newts are redder and brighter than in females: evidence for a role in mate selection? Herpetol J 18:83–89 Dearborn DC, Anders AD, Williams JB (2005) Courtship display by great frigatebirds, Fregata minor: an energetically costly handicap signal? Behav Ecol Sociobiol 58:397–406 Doucet SM, Mennill DJ (2010) Dynamic sexual dichromatism in an explosively breeding Neotropical toad. Biol Lett 6:63–66 Drayton JM, Milner RNC, Hunt J, Jennions MD (2010) Inbreeding and advertisement calling in the cricket Telegryllus commodus: laboratory and field experiments. Evolution 64:3069–3083 Endler JA (1990) On the measurement and classification of color in studies of animal color patterns. Biol J Linn Soc 41:315–352 Fawcett TW, Johnstone RA (2003) Optimal assessment of multiple cues. Proc R Soc Lond B 270:1637–1643 Gabor CR, Halliday TR (1997) Sequential mate choice by multiply mating smooth newts: females become more choosy. Behav Ecol 8:162–166 Garamszegi LZ, Calhim S, Dochtermann N, Hegyi G, Hurd PL, Jørgensen C, Kutsukake N, Lajeunesse MJ, Pollard KA, Schielzeth H, Symonds MRE, Nakagawa S (2009) Changing philosophies and tools for statistical inference in behavioral ecology. Behav Ecol 20:1363–1375 Garner TWJ, Schmidt BR (2003) Relatedness, body size and paternity in the alpine newt, Triturus alpestris. Proc R Soc Lond B 270:619–624 Gomez D, Richardson C, Lengagne T, Plenet S, Joly P, Léna JP, Théry M (2009) The role of noctural vision in mate choice: females prefer conspicuous males in the European tree frog (Hyla arborea). Proc R Soc Lond B 276:2351–2358 Green AJ (1991) Large male crests, an honest indicator of condition, are preferred by female smooth newts, Triturus vulgaris (Salamandridae) at the sperm mass transfer stage. Anim Behav 41:367–369 Green DJ, Osmond HL, Double MC, Cockburn A (2000) Display rate by male fairy-wrens (Malurus cyaneus) during the fertile period of females has little influence on extrapair mate choice. Behav Ecol Sociobiol 48:438–446 Griffiths RA, Mylotte VJ (1988) Observations on the development of the secondary sexual characters of male newts, Triturus vulgaris and Triturus helveticus. J Herpetol 22:476–480 Grueber CE, Nakagawa S, Laws RJ, Jamieson IG (2011) Multimodel inference in ecology and evolution: challenges and solutions. J Evol Biol 24:699–711 Guilford T, Dawkins M (1991) Receiver psychology and the evolution of animal signals. Anim Behav 42:1–14 Haerty W, Gentilhomme E, Secondi J (2007) Female preference for a male sexual trait uncorrelated with male body size in the palmate newt (Triturus helveticus). Behaviour 144:797–814 Halliday TR (1975a) An observational and experimental study of sexual behaviour in the smooth newt, Triturus vulgaris (Amphibian: Salamandridae). Anim Behav 23:291–322 Halliday TR (1975b) Sexual behaviour of the smooth newt, Triturus vulgaris (Urodela, Salamandridae). J Herpetol 8:277–292 Halliday TR (1976) The libidinous newt. An analysis of variations in the sexual behaviour of the male smooth newt, Triturus vulgaris. Anim Behav 24:398–414 Halliday TR (1977) The effect of experimental manipulation of breathing behavior on the sexual behaviour of the smooth newt, Triturus vulgaris. Anim Behav 25:39–45 Halliday TR, Houston A (1978) The newt as an honest salesman. Anim Behav 26:1273–1274 Halliday TR, Sweatman HPA (1976) To breathe or not breathe; the newt’s problem. Anim Behav 24:551–561 Behav Ecol Sociobiol (2012) 66:1045–1055 Hasselquist D, Bensch S (2008) Daily energy expenditure of signaling great reed warblers Acrocephalus arundinaceus. J Avian Biol 39:384–388 Hasson O (1989) Amplifiers and the handicap principle in sexual selection—a different emphasis. Proc R Soc Lond B 235:383–406 Hasson O (1990) The role of amplifiers in sexual selection—an integration of the amplifying and the Fisherian mechanisms. Evol Ecol 4:277–289 Hasson O (1991) Sexual displays as amplifiers—practical examples with an emphasis on feather decorations. Behav Ecol 2:189–197 Hasson O (1997) Towards a general theory of biological signalling. J Theor Biol 185:139–156 Hebets EA, Papaj D (2005) Complex signal function: developing a framework of testable hypotheses. Behav Ecol Sociobiol 57:197–214 Hebets EA, Uetz GW (2000) Leg ornamentation and the efficacy of courtship display in four species of wolf spider (Araneae: Lycosidae). Behav Ecol Sociobiol 47:280–286 Hedlund L (1990a) Factors affecting differential mating success in male crested newts, Triturus cristatus. J Zool 220:33–40 Hedlund L (1990b) Courtship display in natural population of crested newts, Triturus cristatus. Ethology 85:279–288 Hill JA, Enstrom DA, Ketterson ED, Nolan V Jr, Ziegenfus C (1999) Mate choice based on static versus dynamic secondary sexual traits in the dark-eyed junco. Behav Ecol 10:91–96 Holzer B, Jacot A, Brinkhof WG (2003) Condition-dependent signaling affects male sexual attractiveness in field crickets, Gryllus campestris. Behav Ecol 14:353–359 Jehle R, Sztatecsny M, Wolf JBW, Whitlock A, Hödl W, Burke T (2007) Genetic dissimilarity predicts paternity in the smooth newt (Lissotriton vulgaris). Biol Lett 3:526–528 Johnstone RA (1996) Multiple displays in animal communication: “backup signals” and “multiple messages”. Phil Trans R Soc Lond B 351:329–338 Jones AG, Arguello JR, Arnold SJ (2002) Validation of Bateman’s principles: a genetic study of mating patterns and sexual selection in newts. Proc R Soc Lond B 269:2533–2539 Kokko H, Rintamaki PT, Alatalo RV, Hoglund J, Karvonen E, Lundberg A (1999) Female choice selects for lifetime lekking performance in black grouse males. Proc R Soc Lond B 266:2109–2115 Kotiaho JS, Alatalo RV, Mappes J, Nielsen MG, Parri S, Rivero A (1998) Energetic costs of size and sexual signalling in a wolf spider. Proc R Soc Lond B 265:2203–2209 Künzler R, Bakker TCM (2001) Female preferences for single and combined traits in computer animated stickleback males. Behav Ecol 12:681–685 Loyau A, Saint Jalme M, Cagniant C, Sorci G (2005a) Multiple sexual advertisements honestly reflect health status in peacocks (Pavo cristatus). Behav Ecol Sociobiol 58:552–557 Loyau A, Saint Jalme M, Sorci G (2005b) Intra- and intersexual selection for multiple traits in the peacock (Pavo cristatus). Ethology 111:810–820 Loyau A, Gomez D, Moureau B, Théry M, Hart NS, Saint Jalme MS, Bennett ATD, Sorci G (2007) Iridescent structurally based coloration of eyespots correlates with mating success in the peacock. Behav Ecol 18:1123–1131 Maan ME, Cummings ME (2009) Sexual dimorphism and directional sexual selection on aposematic signals in a poison frog. P Natl Acad Sci USA 106:19072–19077 Malmgren JC, Enghag M (2008) Female preference for male dorsal crests in great crested newts (Triturus cristatus). Ethol Ecol Evol 20:71–80 Matthews IM, Evans JP, Magurran A (1997) Male display rate reveals ejaculate characteristics in the Trinidadian guppy Poecilia reticulata. Proc R Soc Lond B 264:695–700 Møller AP, Pomiankowski A (1993) Why have birds got multiple sexual ornaments. Behav Ecol Sociobiol 32:167–176 Petrie M, Halliday T (1994) Experimental and natural changes in the peacocks (Pave cristatus) train can affect mating success. Behav Ecol Sociobiol 35:213–217 1055 Pryke SR, Andersson S (2005) Experimental evidence for female choice and energetic costs of male tail elongation in red-collared widowbirds. Biol J Linn Soc 86:35–43 Richardson C, Gomez D, Durieux R, Théry M, Joly P, Léna JP, Plénet S, Lengagne T (2010) Hearing is not necessarily believing in nocturnal anurans. Biol Lett 6:633–635 Rosenthal GG, Rand AS, Ryan MJ (2004) The vocal sac as a visual cue in anuran communication: an experimental analysis using video playback. Anim Behav 68:55–58 Rowe C (1999) Receiver psychology and the evolution of multicomponent signals. Anim Behav 58:921–931 Ryan MJ, Rand AS (1990) The sensory basis of sexual selection for complex calls in the túngara frog, Physalaemus pustulosus (sexual selection for sensory exploitation). Evolution 44:305–314 Scheuber H, Jacot A, Brinkhof MWG (2003) The effect of past condition on a mutlicomponent sexual signal. Proc R Soc Lond B 270:1779–1784 Scheuber H, Jacot A, Brinkhof MWG (2004) Female preference for multiple condition-dependent components of a sexually selected signal. Proc R Soc Lond B 271:2453–2457 Schluter D, Price T (1993) Honesty, perception and population divergence in sexually selected traits. Proc R Soc Lond B 253:117–122 Secondi J, Aumjaud A, Pays O, Boyer S, Montembault D, Violleau D (2007) Water turbidity affects the development of sexual morphology in the palmate newt. Ethology 113:711–720 Secondi J, Hinot E, Djalout Z, Sourice S, Jadas-Hécart A (2009) Realistic nitrate concentration alters the expression of sexual traits and olfactory male attractiveness in newts. Funct Ecol 23:800–808 Sheldon BC (1993) Sexually transmitted disease in birds: occurrence and evolutionary significance. Phil Trans Roy Soc Lond B 339:491–497 Shettleworth S (1998) Cognition, evolution and behavior. Oxford University Press, New York Sinervo B, Basolo AL (1996) Testing adaptation using phenotypic manipulations. In: Rose MR, Lauder G (eds) Adaptation. Academic, New York, pp 149–185 Sztatecsny M, Strondl C, Baierl A, Ries C, Hodl W (2010) Chin up: are the bright throats of male common frogs a condition-independent visual cue? Anim Behav 79:779–786 Taylor PW, Hasson O, Clark DL (2000) Body postures and patterns as amplifiers of physical condition. Proc R Soc Lond B 267:917–922 Tomkins JL, Simmons LW (1998) Female choice and manipulations of forceps size and symmetry in the earwig Forficula auricularia. Anim Behav 56:347–356 Usherwood JR (2008) Collared doves Streptolelia decaocto display with high, near-maximal muscle powers, but at low energetic cost. J Avian Biol 39:1–19 van Dongen WFD, Mulder RA (2008) Male and female golden whistlers respond differently to static and dynamic signals of male intruders. Behav Ecol 19:1025–1033 van Doorn GS, Weissing FJ (2004) The evolution of female preferences for multiple indicators of quality. Am Nat 164:173–186 Vasquez T, Pfennig KS (2007) Looking on the bright side: females prefer coloration indicative of male size and condition in the sexually dichromatic spadefoot toad, Scaphiopus couchii. Behav Ecol Sociobiol 62:127–135 Vehrencamp SL, Bradbury JW, Gibson RM (1989) The energetic cost of display in male sage grouse. Anim Behav 38:885–896 Vinnedge B, Verrell P (1998) Variance in male mating success and female choice for persuasive courtship displays. Anim Behav 56:443–448 Wagner WE (1998) Measuring female mating preferences. Anim Behav 55:1029–1042 Weir LK, Grant JWA (2010) Courtship rate signals fertility in an externally fertilizing fish. Biol Lett 6:727–731 Witte K, Ryan MJ (1998) Male body length influences mate-choice copying in the sailfin molly Poecilia latipinna. Behav Ecol 5:534–539 ARTICLE 3 It takes two to tango : Do both male and female identities influence complex courtship display in a newt species? Cornuau JH, Courtoix EA, Jolly T, Schmeller DS, Loyau A. En préparation 1 It takes two to tango: Do both male and female identities influence complex courtship 2 display in a newt species? 3 Jérémie H. Cornuaua,*, Elodie A Courtoixa, Thomas Jolly a, Dirk S. Schmellera,b, Adeline 4 Loyaua,b 5 6 a Station d’Ecologie Expérimentale du CNRS à Moulis, 09200 Saint Girons, France. 7 b Department of Conservation Biology, Helmholtz Centre for Environmental Research – UFZ, 8 Permoser Strasse 15, 04318Leipzig, Germany. 9 10 * Correspondence: J.H. Cornuau, Station d’Ecologie Expérimentale du CNRS à Moulis, 11 09200 Saint Girons, France. 12 E-mail address: [email protected] (J.H. Cornuau) 13 14 15 16 17 18 19 20 1 21 Keywords: repeatability, behavioural consistency, sexual selection, mate choice, Lissotriton 22 helveticus, palmate newt 23 24 Abstract In many species, females base their mate choice on male courtship display. 25 Although courtship display is consistent in term of behavioral sequence and structure within a 26 given species, there is commonly important variation in the intensity of courtship display 27 between and within males. Interestingly, only a few studies have investigated the different 28 factors that can affect both between and within variability in courtship display intensity. 29 Female traits can be an important source of within male variation. Therein, it is unclear what 30 part of the variation in courtship display can be explained by males or females traits. In this 31 experiment we tested whether several phases of a complex courtship display could be 32 explained by both male and female identities in the palmate newts Lissotriton helveticus. 33 Overall, we found that both male and female identities influence complex courtship behavior 34 but the relative influence of each sex depended on the courtship phase considered. Male 35 identity explained variation in fan and quiver display whereas female identity explained 36 variation in quiver and touch tail display. The time spent in fan display appeared repeatable 37 between males and we found no evidence of an effect of female identity. Finally we did not 38 find any link between courtship display, and behavioral consistency and morphological traits. 39 Our results provide, to our knowledge, the first example of significant inter-variation of 40 courtship display in newts showing their potential importance to signal male identity. 41 Additional experiments are now needed to determine whether courtship display indicates any 42 aspect of male quality. 43 44 2 45 46 Extravagant courtship displays consist of series of complex stereotyped behaviours 47 that have evolved for sexual communication (Andersson 1994, Wachtmeister 2001, Mowles 48 & Ord 2012). Courtship display is largely assumed to be costly to the male in terms of time, 49 energy and exposion to higher predation risks (Andersson 1994). Generally, only males in 50 good condition or highly motivated can maintain high courtship display activity (Zahavi 1975, 51 Bradbury & Vehrencamp 2011, but see Lehtonen 2012, Mowles & Ord 2012). Therefore 52 courtship display could be used by females during mate choice as an honest signal of male 53 quality, and consequently courtship display behaviours could affect the regime of sexual 54 selection (Jaquiéry et al. 2009, Schneider & Lesmono 2009, Loyau & Lacroix 2010). 55 Although courtship display is consistent in terms of behavioural sequence and structure within 56 a given species, there is commonly important variation in the intensity of courtship display 57 between males (Bradbury & Vehrencamp 2011). Accordingly, the influence of variation 58 among males in courtship display intensity on female mate choice has been the subject of 59 much research (e.g. Künzler & Bakker 2001, Loyau et al. 2005a,b, Ward & McLennan 2009, 60 Shamble et al. 2009). Another important source of variation, yet far less explored, is the 61 behavioural plasticity of courtship display that can lead to important variation within males 62 (Bell et al. 2009). The outcome of sexual selection could be strongly affected by such 63 variation, especially when a poor quality male could momentarily be able to express 64 behavioural display comparable to that of a high quality male (Lindström et al. 2009). 65 Interestingly only a few studies have investigated the different factors that can affect both 66 between and within male variation in courtship display intensity (Bell et al. 2009). 67 Environmental heterogeneity is one major source of variation within male. Such 68 variation can occur in response to environmental pressure such as predation (Fowler-Finn & 69 Hebets 2011), environmental noise (Lengagne & Slater 2002, Brumm & Slater 2006, Wilgers 3 70 & Hebets 2011), or spatial and temporal partner availability (Lindström et al. 2009). Less well 71 studied, although equally important, is the influence of male and female individual 72 characteristics as a source of within male variation (Patricelli et al. 2002, 2006). For example 73 male desert gobies spent more time in courtship display towards larger females (Wong & 74 Svensson 2009). In the wolf spider Schizocosa rovneri males increase courtship display (i.e. 75 number of body bounces) when they receive a specific positive female feedback cue 76 (Sullivan-Beckers & Hebets 2011). In contrast males can increase courtship display after a 77 negative female feedback in an attempt to catch the female attention, and males can discount 78 courtship display after a positive feedback, as in the convict cichlid Amatitlania nigrofasciata 79 (Santangelo 2005). Within male variation can also differ among males. Indeed, in the wolf 80 spider Schizocosa the males reduce their courtship display in presence of predator odor but 81 this response to predation is stronger for more ornamented male, showing that behavioral 82 plasticity can also depend on male characteristics (Fowler-Finn & Hebets 2011). Although 83 courtship display behaviours can be driven by both male and female individual characteristics, 84 it is not yet clear which part of the display is influenced by which sex and to which extent 85 (Ruiz et al. 2008, Lehtonen et al. 2011). In this study, we extended recent findings on the 86 importance of courtship display on female mate choice in the palmate newt Lissotriton 87 helveticus (Cornuau et al. 2012, Cornuau et al. unpublished), and we examined which factors 88 influence between and within male variation of the courtship display, including both male and 89 female individual characteristics. We have previously found that the courtship display was 90 variable between males but was not explained by male body condition (Cornuau et al. 91 submitted, Cornuau et al. unpublished). 92 To investigate behavioural plasticity of courtship display a widely adopted approach is 93 to assess, in a given population, the temporal variation between and within males by 94 performing successive mate choice trials (Bell et al. 2009). An advantage to this approach is 4 95 that it allows calculating the repeatability of courtship display (Bell et al. 2009). Repeatability 96 is defined as the part of variation in a trait that is explained by differences between individuals 97 (Bell et al. 2009). A highly repeatable behaviour (e.g. male courtship display) is repeatable 98 when it shows high variation between males associated to low variation within males, in other 99 words when individuals behave differently from each other while behaving consistently 100 through time (Widemo & Saether 1999, Bell et al. 2009). In contrast, a low repeatability could 101 mean that the behaviour is highly variable within a male, or that there is a low variation 102 between males associated to a low variation within males (Widemo & Saether 1999). 103 Although many studies have described the behavioural displays of newts (see Wells 2007), 104 only one study has investigated the repeatability of newt display behaviour (Michalak 1996, 105 see Bell et al. 2009), and so far no study have performed experiment to investigated the 106 relative influence of male and female characteristics on behavioural display. The aim of this 107 study was to 1- investigate the importance of both male and female characteristics on the 108 elaboration of complex courtship displays, 2- investigate the influence of both male and 109 female morphological features on behavioral consistency. 110 111 METHODS 112 Courtship display in the palmate newt 113 The palmate newt is a suitable model species to investigate the influence of male and 114 female characteristics on behavioural display because newts use complex, stereotyped, 115 elaborated courtship display that is known to influence female mate choice (Cornuau et al. 116 2012). The internal fertilization and the use of spermatophore in this species require a close 117 coordination between males and females. The complex courtship display of palmate newts 118 combines multimodal cues; including visual, chemical and tactile cues. As in other European 5 119 newt species the courtship display is composed of four phases (Wells 2007). During the first 120 phase, named orientation, the male follows and smells the female before moving to her front. 121 The second and the third phases are static and retreat displays. During these two phases the 122 male newt performs fan, whip and wave behaviours (Table 1). Fan behaviour represents the 123 longest and the most frequent behaviour in palmate newt display. The last phase is the 124 spermatophore transfer, in which the coordination between the male and the female is 125 necessary to ensure spermatophore transfer from the male to the female. During this phase the 126 male creep before the female and performs quiver behaviour (Table 1). This behaviour 127 corresponds to an offer to deposit a spermatophore to the female. The female then touches the 128 tail of the male so that he deposes a spermatophore on the ground. To transfer the 129 spermatophore, the female follows the male until her cloaca is above the spermatophore, 130 allowing the female to take it. 131 132 Laboratory experiment 133 All individuals were originally captured by dip netting in one pond near to Caumont, 134 France, during the breeding season (04-03-2012). We collected 24 male and 24 female newts 135 (catching permits no. 2009-13). This sample size was big enough to measure repeatability 136 (Bell et al. 2009). Newts were brought to the animal facility at the Laboratoire d’Ecologie 137 Expériementale du CNRS de Moulis (France). The day following capture, the individuals 138 were individually marked with subcutaneous implants of colored visible elastomers (VIE, 139 Northwest Marine Technology, Washington, Shaw Island, WA, USA) at the basis of the legs. 140 Newts were maintained in by unisex groups of six in tanks one week before the start of the 141 experiment (52×33.5×29.5 cm) with approximately half water collected in their native pond 142 and half tape water. A clay brick perforated with holes was placed in each tank to ensure 143 shelter. Newts were maintained at 18°C (±1) under fluorescent tubes (ReptiSun 2.0, ZooMed) 6 144 to simulate natural light with a cycle of 12 hours light/ 12 hours dark. Newts were fed ad 145 libitum by daily provision of living daphnia collected in a semi natural pond. Newts were 146 released in their native ponds immediately after the experiment (04-19-2012). Their captive 147 time represented a very small part of their breeding season (two weeks on approximately three 148 month). 149 Measurement of morphological traits 150 In male palmate newts, morphological sexual traits are composed of one filament at 151 the end of the tail, two hind foot webs and a small crest on the back (Cornuau et al. 2012). 152 These ornaments are only expressed during the reproduction period which corresponds to the 153 aquatic phase of newt life. Individuals were measured the day after being caught. We 154 recorded body mass using scale (accuracy: 0.01 g). We used Snout-Vent-Length (SVL), from 155 the snout tip to the posterior end of the cloaca as body size index. We measured SVL and 156 filament length by taking photos of the individuals on a millimeter paper background (grid 1 157 mm2). We obtained sizes using the free software IMAGEJ v.1.28 (http://rsb.info.nih.gov/ij/). 158 These measures are reliable and repeatable (Cornuau et al. 2012). We calculated body 159 condition index of male and females using the residuals of the linear regression of the root 160 cute of weight on SVL for each sex separately (Báncilá et al., 2010). 161 162 Experimental design 163 We aimed to investigate the consistency of males and females behaviour and the 164 relative contribution of both males and females on several aspects of palmate newt courtship. 165 Courtship display behaviours were assessed during dyadic encounters composed of one male 166 and one female. Pairs of individuals were randomly made up, i.e. randomly with respected to 167 morphological variables. Behaviours were observed three times: at day 1, the day after (day 7 168 2), and finally six days after the first measure (day 6), for all individuals, resulting in a total of 169 72 mate choice trials. However, all encounters were unique, i.e. each male could display to 170 three different females, and each female encountered three different males. At the end of the 171 experiment all 24 males and 24 females have performed three mate choice trials (one at day 1, 172 one at day 2 and once at day 6). 173 The mate choice trials were performed in an experimental tank (26×33.5×29.5 cm), 174 from 08:00 to 12:00 am (Cornuau et al. 2012). Six mate choice trials were simultaneously 175 performed, each in a separate tank, so that all individual 24 trials were conducted the same 176 day. Each experimental tank contained 3 cm of sand and 8 cm of water. Before the trial, 177 experiment tanks were divided in two parts, one for each sex, with a visual barrier. The male 178 and the female were placed in each part of the experimental tanks during 5 min of 179 acclimatization before the visual barrier was removed and individuals could interact freely 180 during 40 min and this time was enough for the observations of sexual behaviours (Cornuau et 181 al. 2012). 182 183 Measurement of behaviour 184 Each mate choice trial was recorded using a video camera placed above the 185 experimental tank. The videos were analyzed with the software The Observer 7.0 (Noldus 186 Information Technology). As explained in the introduction male palmate newts perform a 187 complex display composed of several distinct behaviours, which are described in Table 1. We 188 recorded the number of fan, whip and wave events performed by a given male over the 40 189 minutes of the trial (static and retreat display, Table 1). Fan behaviour can last several 190 minutes, so we also recorded the time spent in fan (Table 1). We counted the number of 191 quivers and touch tail during the spermatophore transfer phase (Table 1). To avoid changes in 8 192 female receptivity over the course of the experiment, we did not allow the females to take 193 male spermatophores. Indeed female mate choice and female choosiness can change after a 194 first mating (Gabor & Halliday 1997). To prevent spermatophore transfer, we stopped the 195 female after touch tail using a glass rod. Both male and female still performed courtship 196 behaviour immediately after this action (JHC personal observation). As we did not allowed 197 spermatophore transfers to occur, we used the number of touch tail as a proxy of female mate 198 preference. Touch tail is a good indicator of female receptiveness as it always precedes 199 spermatophore transfer (Table 1). We identified this proxy using a personal data base 200 containing 40 mate choice trials between one male and one female from the same population 201 (JHC personal data). We found that touch tail was a great predictor of willingness to accepted 202 a spermatophore (GLM, Z = 4.65, P < 0.001). Finally performing courtship in water may 203 impose some costs to male newts so we counted the number of time the male surfaced to 204 breath. 205 Statistical analyses 206 All statistical analyses were performed using R 2.13.1 software. First we tested 207 whether the three static display behaviours (fan, whip and wave) performed by males during 208 the experiment were correlated using Spearman rank correlations. Following this analysis we 209 decided to use only the time spent in fan display as a proxy of the static display (Table 2). The 210 time spent in fan corresponds to the most important courtship display behaviour in palmate 211 newts (Cornuau et al. 2012). 212 Secondly we investigated for each courtship display (time spent in fan display, number 213 of quivers and number of touch tail) the relative influence of the day (day 1, 2 and 6), and 214 both male and female identity. The time spent in fan display did not match with normality so 215 we chose to work with non parametric statistics for this data set. Thus we used Kendall rank 9 216 correlations to assess the effects of both male and female identities on the time spent in fan 217 display. We used Kruskall Wallis tests to assess the effect of the day on the time spent in fan 218 display. We used Poisson distribution of error terms both for the number of quivers and the 219 number of touch tail behaviours. To estimate the variation in the number of quivers and the 220 number of touch tail associated to male identity or female identity, we used general linear 221 mixed effects models (package “lme4”). All models were fitted using the day (day 1, day 2 or 222 day 6) as a fixed factor and both male identity and female identity as random factors. We 223 assessed the statistical significance of fixed and random factors using log-likelihood tests by 224 comparing a full model with reduced models without the fixed or random factor tested, and 225 tested every possible model (Crawley 2007, Zuur et al. 2009). We used maximum likelihood 226 (ML) to fit the models when models differed in their fixed factor. 227 Thirdly we investigated the effect of morphology on the courtship display. We used 228 Spearman rank correlations to test the link between time spent in fan display and male 229 morphological characteristics (SVL, filament length, BCI, number of breathing events) and 230 female characteristics (SVL and BCI). Then we assessed the influence of male morphological 231 characteristics (SVL, filament length, BCI, number of breath event) and female characteristics 232 (SVL and BCI) on the number of quivers and the number of touch tail using generalized 233 linear mixed models with both male and female identity as random factors and morphological 234 traits as fixed factors. Again we assessed the statistical significance of fixed and random 235 factors using log-likelihood tests by comparing a full model with reduced models without the 236 fixed or random factor tested (Crawley 2007, Zuur et al. 2009). 237 Fourthly we investigated whether the variation in flexibility between individual was 238 linked to morphological parameters. We used for each males and females the variance 239 between the three tests as an estimate of the behavioral flexibility. Thus we tested whether the 10 240 variance was explained by morphological parameters using an analysis of variance 241 (ANOVA). 242 243 RESULTS 244 The correlation between filament length and male body size was not significant (r = 245 0.24, df = 22, P = 0.28). The different behaviours performed by the males during the static 246 displays were strongly correlated altogether (Table 2). Thus we subsequently used only the 247 time spent in fan as a proxy of the static display. 248 The number of males that did not perform at least one courtship display did not vary 249 significantly between days (day 1: 1/24, day 2: 7/24, day 6: 3/24; χ2 = 1.14, df = 1, P = 0.28). 250 We found that neither the variable day nor the female identity significantly explained the time 251 spent in fan display (day: χ2 = 3.79, df = 2, P = 0.15; female identity: Wt = 0.34, χ2 = 23.7, df 252 = 23, P = 0.42). However the male identity significantly explained the time spent in fan 253 display (Wt = 0.56, χ2 = 38.9, df = 23, P = 0.02). We did not detect any effect of the male 254 body size (rho = 0.11, P = 0.37), filament length (rho = -0.04, P = 0.76) or male BCI (rho = 255 0.16, P = 0.17) on the time spent on fan display. Moreover we did not detect any effect of the 256 female body size (rho = -0.03, P = 0.80) and the female BCI on the time spent in fan display 257 (rho = 0.16, P = 0.19). However we found a positive link between the time spent in fan 258 display and the number of breath events (rho = 0.27, P = 0.024). The variance in the time 259 spent in fan display by males was not explained by the filament length (F1,20 = 0.79, P = 0.38), 260 male body size (F1,20 = 0.01, P = 0.92) or male BCI (F1,20 = 0.59, P = 0.45). The variance in 261 the time fan display was performed to a female was not explained by female body size (F1,21 = 262 1.11, P = 0.31) or female BCI (F1,21 = 0.23, P = 0.63). 11 263 We found a significant contribution of the day on the number of quivers (model 1 vs. 264 model 4, χ2 = 19.95, df = 1, P < 0.001). Indeed the number of quivers decreased between day 265 1 and day 2 (Paired Wilcoxon test: V = 246, df = 22, P = 0.001) and between day 1 and day 6 266 (Paired Wilcoxon test: V = 148, df = 22, P = 0.007) but was constant between day 2 and day 6 267 (Paired Wilcoxon test: V = 70.5, df = 22, P = 0.33). We found a significant contribution of 268 both female and male identities on the number of quivers (female identity: model 1 vs. model 269 4, χ2 = 5.02, df = 1, P = 0.025; male identity: model 1 vs. model 4, χ2 = 22.03, df = 1, P < 270 0.001; Table 3). Male identity represented 28.3%, day 10.09% and female 7.77% of the total 271 variance in the number of quivers. The residual variance, unexplained by the model, 272 represented 53.84%. The number of quivers carried out by the males was not explained by 273 male size (χ2 = 0.02, df = 1, P = 0.89), filament length (χ2 = 0.03, df = 1, P = 0.87) nor male 274 BCI (χ2 = 2.36, df = 1, P = 0.12). The number of quivers performed to females was not 275 explained by female body size (χ2 = 0.39, df = 1, P = 0.53) nor female BCI (χ2 = 0.53, df = 1, 276 P = 0.47). The variance in the number of quivers by males was not explained by filament 277 length (F1,20 = 0.82, P = 0.37), male body size (F1,20 = 0.02, P = 0.89) or male BCI (F1,20 = 278 0.07, P = 0.79). The variance in the number of quivers that were performed to a female was 279 not explained by female body size (F1,21 = 0.47, P = 0.50) or female BCI (F1,21 = 0.33, P = 280 0.57). 281 We found a significant contribution of the day on the number of touch tail (model 1 282 vs. model 4, χ2 = 27.26, df = 1, P < 0.001). Indeed the number of touch tail decreased between 283 day 1 and day 2 (Paired Wilcoxon test: V = 233.5, df = 22, P < 0.001) and between day 1 and 284 day 6 (Paired Wilcoxon test: V = 168.5, df = 22, P = 0.003) but was constant between day 2 285 and day 6 (Paired Wilcoxon test: V = 44.5, df = 22, P = 0.23). We did not found a significant 286 contribution of male identity on the number of touch tail (model 1 vs. model 4, χ2 = 1.03, df = 287 1, P = 0.31, Table 3). However we found a significant contribution of female identity on the 12 288 number of touch tail (χ2 = 20.33, df = 1, P < 0.001, Table 3). This means that whatever the 289 male identity, the number of touch tail was consistent within a female. The female identity 290 represented 34.12%, the day 24.21% and the male identity 3.51% of the total variance of the 291 number of touch tail. The residual variance, unexplained by the model, represented 38.16%. 292 The number of touch tail done by females was not explained by their body size (χ2 = 0.031, df 293 = 1, P = 0.86) nor their body condition (χ2 = 0.24, df = 1, P = 0.62). The number of touch tail 294 received by males was not explained by their body size (χ2 = 0.007, df = 1, P = 0.93), filament 295 length (χ2 = 0.96, df = 1, P = 0.33) or their BCI (χ2 = 1.97, df = 1, P = 0.16). The variance in 296 the number of touch tail done be the females was not explained by female body size (F1,21 = 297 0.02, P = 0.89) or female BCI (F1,21 = 0.14, P = 0.71). The variance in the number of touch 298 tail that males received was not explained by male body size (F1,20 = 0.13, P = 0.72), male 299 BCI (F1,20 = 0.39, P = 0.54) or male filament length (F1,20 = 0.26, P = 0.61). 300 301 DISCUSSION 302 The combined results of our study provide evidence of the influence of both male and 303 female identity on the expression of complex courtship display behaviours. We found that 304 both male and female identities influenced complex courtship behaviours but the relative 305 influence of both sex depends on the courtship phase. More precisely we found that the time 306 spent in fan display was explained by male and not by female identity. The number of quiver 307 displays was explained by both male and female identity, but the influence of female identity 308 was smaller than that of the male. Finally the number of touch tail behaviours was explained 309 by female and not male identity. We found also no significant influence of male or female 310 morphological traits (body size, body condition, and filament length) that could explain the 311 individual consistency. Altogether our results are one of the few that dissected the different 13 312 sources of variation that can affect behavioural repeatability of courtship display and provides 313 to our knowledge the first example in newts of significant within male variation in courtship 314 display. 315 In our study, although females varied greatly in their body size, body condition and 316 reproductive receptivity (measured by the number of touch tail), the time spent in fan display 317 was only explained by male identity. Thus the time spent in fan display seems to reflect male 318 ability to perform this behaviour and to not be influenced by any female characteristic we 319 measured. One possibility is that male palmate newts use a fixe strategy in term of the time 320 spent in fan display and not a strategy involving behavioral plasticity. Other sources of 321 variation not measured in this study could yet occur. Indeed courtship display could be 322 flexible according to environmental conditions, unmeasured in this study because laboratory 323 conditions remained constant over the course of the experiment. For example fan display of 324 the Alpine newts, Ichthyosaura alpestris, was affected by light condition and water 325 temperature (Denoël et al. 2005, 2010). Further investigations on the relative plasticity of 326 male fan display are necessary, especially in heterogeneous environment. 327 Contrary to the time spent in fan display; quiver display appears to be driven by both 328 male and female identify. Quiver behaviour seems more labile and male could adjust the 329 frequency of this behaviour according to female characteristics. A quiver represents a male 330 offer to deposit a spermatophore, and it has been suggested that this behaviour is influenced 331 by both male and female characteristics although the role of each sex remains unclear 332 (Halliday 1975). Our results show that both male and female identity influence the number of 333 quivers, and the influence of male identity still appeared to be more important that of female 334 identity. During the courtship display the male could respond to the female cue signaling her 335 receptivity, as the female moves towards the male, by the quiver behaviour that could lead to 336 spermatophore transfer (Halliday1975). Moreover we also observed in a previous study that 14 337 when males and females were separated for a long time more quivers probably due to an 338 increase in male motivation and/or a low selectivity of the females (JHC, personal 339 observation). It is probably adaptive to adjust the quiver behaviour to female receptivity to 340 access more rapidly to the spermatophore transfer phase. In this study we did not identify any 341 male or female features, such as body condition or size, that could explain the number of 342 quivers. However in a previous experiment in which we conducted an experimental diet 343 limitation we have found that males in lower condition tend to make more quivers (JHC, DS, 344 AL unpublished data). Additional work is clearly necessary to further determine which both 345 male and female features may explain the number of quivers. 346 We found, in agreement with our previous experiment (Cornuau et al. 2012), an 347 important variation in the proxy of male attractiveness (measured here by the number of touch 348 tail). Surprisingly, we did not detect any effect of male identity caused by important intra 349 individual variation in their attractiveness to females. Yet we found a significant effect of 350 female identity that explained the number of touch tail and which suggest an inter individual 351 variation in female receptivity or choosiness. Altogether these two results show that, firstly, 352 female mate preference seems not to be directed toward the same males. Indeed if all females 353 would have shown a preference for the same males or same male characteristics, we would 354 have detected a positive effect of male identity. Secondly female mate choice is better 355 explained by female identity than male identity, an argument in agreement with free female 356 mate choice in this species. In this study, the positive effect of female identity to explain the 357 number of touch tail event was not affected by female characteristics such as size or body 358 condition index. However we could not take into account the past history of female and the 359 number of mate encounters in the pond before the experiment that could create difference in 360 female selectivity and explain the observed influence of female identity (Gabor & Halliday 361 1997). 15 362 We found a relative consistency of male courtship display and female receptivity. 363 Within individual constancy in behaviour and particularly in courtship is well documented in 364 several taxa but rarely in salamander (Bell et al. 2009). Yet, within individual constancy in 365 behaviour is crucial in evolution because such inter individual variation give the opportunity 366 to selection especially via female mate choice (Jennions & Petrie 1997). Moreover within 367 individual constancy in behaviour is also crucial to such behaviour to be informative about 368 male quality (Jennions & Petrie 1997). In a recent meta-analysis of repeatability, Bell et al. 369 (2009) found that in average 35% of the variation among individuals is attributed to 370 individual differences, this result being consistent with what we found in this study. As fan 371 display was known to be a key factor that explain reproductive success in palmate newts 372 (Cornuau et al. 2012) our study highlights the opportunity to sexual selection via female mate 373 choice to shape courtship signal. 374 We found a significant effect of the day on the number of quivers, the number of touch 375 tail and not on the time spent in fan display. The influence of the day was however lower than 376 the influence of male identity and that of female identity to explain respectively the number of 377 quivers and the number of touch tail. The influence of the day might reflect an increase of 378 both male and female selectivity after the first trial. Indeed, in species with multiple mating 379 empirical evidence shows that individual selectivity increases after first a encounter especially 380 after a first mating with a high quality partner (Jennions & Petrie 2000). This strategy has 381 been found in newts (Gabbor & Halliday 1997). However in this experiment we did not 382 allowed spermatophore transfer to the female to avoid this problem. Yet our repeated design 383 might have created habituation or stress that affected every individual differently. However, 384 whatever the cause of variations observed for quiver and touch tail between days, these 385 variations were taken into account in our models and should not overestimate but may rather 386 underestimate our measure of behavioural constancy. 16 387 To conclude we found, for the first time in a newt species, a significant inter- 388 individual variation of courtship effort. Our study shows also that both male and female 389 identity influence in various ways the different phases of the complex displays of newts. As 390 the morphological traits we measured (body size, body condition and development of a sexual 391 trait) did not contribute to explain within male and within female variation in behaviour, 392 individuals characteristics explaining these variations remainted to be identified by future 393 investigations. 394 395 Ethical standards The experiments comply with the current laws of the country in which 396 they were performed. 397 398 Acknowledgements We are grateful to Laurence Walder for her help to collected data and 399 Adelaïde Sibeaux for her valuable and accurate comments on the manuscript. Financial 400 support was provided by the Ministère de la Recherche (PhD fellowship to JC and two CNRS 401 grants to AL and DS), the BioDiversa-project RACE, and the Observatoire Homme-Milieu 402 Pyrénées Haut-Vicdessos. This work was supported by the "Laboratoire d’Excellence 403 (LABEX)" entitled TULIP (ANR -10-LABX-41). 404 405 406 407 408 17 409 TABLES 410 Table 1 Courtship display behaviours of male palmate newts. Display Phase static and retreat display Fan Description The male folds his tail against his flank and the distal portion of the tail vibrates. The male rapidly brings back his tail against his flank. The male holds his tail to be presented with a full view to the female. Whip Wave Phase spermatophore transfer Quiver 411 After creep and follow, the male quivers his tail. This behaviour could be assimilated to an offer to deposit a spermatophore to the female. Touch tail Female move toward the male and touch the tail of the male. Male put spermatophore on the floor (deposition). Spermatophore transfer* Female following male’s tail and take spermatophore. *this behaviour was not assessed in this study (reason explained in the method section) 412 413 Table 2 Results from Spearman correlations between behavioural measures in male palmate 414 newts. *** : p<0.001, nb : number. Fan (tps) Fan (nb) Whip (nb) Wave (nb) 0.78*** 0.51*** 0.51*** Whip (nb) 0.74*** 0.99*** Fan (nb) 0.73*** 415 416 417 418 18 419 Table 3 The results of linear mixed effects models of number of quivers in male palmate 420 newts and number of touch tail in female palmate newts. The ΔAIC was the difference 421 between the best model and the model of interest. 422 Response Model Fixed factors Random factors AIC ΔAIC Quiver 1 Day Male ID, Female ID 170.9 0 4 Day Male ID 173.9 3 Male ID, Female ID 188.8 17.9 2 Touch tail 3 Day Female ID 190.9 20 3 Day Female ID 155.3 0 1 Day Male ID, Female ID 156.3 1 4 Day Male ID 174.6 19.3 Male ID, Female ID 181.6 26.3 2 423 424 425 426 427 428 429 430 431 432 19 433 FIGURE Time spent in fan display a) 2000 1600 1200 800 400 0 35 36 37 38 39 40 Male body size (SVL, mm) 41 42 434 Time spent by in fan display b) 2000 1600 1200 800 400 0 38 40 42 44 Female body size (SVL, mm) 46 435 436 Figure 1. Time spent in fan display by males plotted against a) male body size and b) female 437 body size. Successive measures of a given male are connected by a solid line. Gray and dark 438 dots represent respectively two and three superposed dots. 439 440 441 20 442 References 443 Andersson M. 1994. Sexual selection. Princeton: Princeton University Press. 444 Băncilă, R. I., Hartel, T., Plaiasu, R., Smets, J. & Cogalniceanu D. 2010. Comparing three 445 body condition indices in amphibians: a case study of yellow-bellied toad Bombina 446 variegate. Amphibia Reptilia, 31,558–562 447 448 449 450 451 452 Bell, A. M., Hankison, S. J. & Laskowski, K. L. 2009. The repeatability of behaviour: a meta-analysis. Animal Behaviour, 77, 771-783. Bradbury, J. W. & Vehrencamp, S. L. 2011. Principles of Animal Communication. Sunderland, Massachusetts: Sinauer Brumm, H. & Slater, P. J. B. 2006. Animal can vary signal amplitude with receiver distance: evidence from zebra finch song. Animal Behaviour, 72, 699-705. 453 Cornuau, J. C., Rat, M., Schmeller, D. S. & Loyau, A. 2012. Multiple signals in the 454 palmate newt: ornaments help when courting. Behavioral Ecology and Sociobiology, 455 66, 1045-1055. 456 Crawley, M. J. 2007. The R book. Chichester: John Wiley & Sons Ltd. 457 Denoël, M., Mathieu, M. & Poncin, P. 2005. Effect of water temperature on the courtship 458 behavior of the Alpine newt Triturus alpestris. Behavioral Ecology and Sociobiology, 459 58, 121-127. 460 461 Denoël, M. & Doellen J. 2010. Displaying in the dark: light-dependant alternative mating tactics in the Alpine newt. Behavioral Ecology and Sociobiology, 64, 1171-1177. 462 Fowler-Finn, K. D. & Hebets, E. A. 2011. The degree of response to increased predation 463 risk corresponds to male secondary sexual traits. Behavioral Ecology, 22, 268-275. 464 Gabor, C. R., Halliday, T. R. 1997. Sequential mate choice by multiply mating smooth 465 newts: females become more choosy. Behavioral Ecology, 8,162–166. 21 466 Halliday, T. R. 1975. An observational and experimental study of sexual behaviour in the 467 smooth newt, Triturus vulgaris (Amphibia: Salamandridae). Animal Behaviour, 23, 291- 468 322. 469 Jaquiéry, J., Broquet, T., Aguilar, C., Evanno, G. & Perrin, N. 2009. Good genes drive 470 female choice for mating partners in the lek-breeding European treefrog. Evolution, 64, 471 108-115. 472 473 474 475 476 477 Jennions, M. D. & Petrie, M. 1997. Variation in mate choice and mating preferences: a review of causes and consequences. Biological Review, 72, 283-327. Jennions, M. D. & Petrie, M. 2000. Why do females mate multiply? A review of the genetic benefits. Biological Review, 75, 21-64. Künzler, R. & Bakker, T. C. M. 2001. Female preferences for single and combined traits in computer animated stickleback males. Behavioral Ecology, 12, 681-685. 478 Lengagne, T. & Slater, P. J. B. 2002. The effects of rain on acoustic communication: Tawny 479 owls have good reason for calling less in wet weather. Proceedings of the Royal Society 480 B, 269, 2121-2125. 481 482 483 484 Lehtonen, T. K. 2012. Signal value of male courtship effort in a fish with paternal care. Animal Behaviour, 83, 1153-1161. Lehtonen, T. K., Svensson, P. A. & Wong, B. B. M. 2011. Both male and female identity influence variation in male signaling effort. BMC Evolutionary Biology, 11, 233. 485 Lindström J., Pike, T. K., Blount, J. D. & Metcalfe, N. B. 2009. Optimization of resource 486 allocation can explain the temporal dynamics and honestly of sexual signals. American 487 Naturalist, 174, 515-525. 488 Loyau, A. & Lacroix, F. 2010. Watching sexy displays improves hatching success and 489 offspring growth through maternal allocation. Proceedings of the Royal Society B, 277, 490 3453-3460. 22 491 Loyau, A., Saint Jalme, M., Cagniant, C. & Sorci, G. 2005a. Intra- and intersexual 492 selection for multiple traits in the peacock (Pavo cristatus). Ethology, 111, 810-820. 493 Loyau, A., Saint Jalme, M., Cagniant, C. & Sorci, G. 2005b. Multiple sexual 494 advertisements honestly reflect health status in peacocks (Pavo cristatus). Behavioral 495 Ecology Sociobiology, 58, 552-557. 496 497 498 499 Michalak, P. 1996. Repeatability of mating behaviour in Montandon’s newt, Triturus montandoni (Caudata Salamandridae). Ethology Ecology & Evolution, 8, 19-27. Mowles, S. L. & Ord, T. J. 2012. Repetitive signals and mate choice: insights from theory. Animal Behaviour, 84, 295-304. 500 Patricelli, G. L., Albert, J., Walsh, G. & Borgia, G. 2002. Male displays adjusted to 501 female’s response - Macho courtship by the satin bowerbird is tempered to avoid 502 frightening the female. Nature, 415, 279-280. 503 Patricelli, G. L., Coleman, S. W. & Borgia, G. 2006. Male satin bowerbirds, 504 Ptilonorhynchus violaceus, adjust their display intensity in response to female startling: 505 an experiment with robotic females. Animal Behaviour, 71, 49-59. 506 507 508 509 Ruiz, M., Davis, E. & Martins, E. P. 2008. Courtship attention in sagebrush lizards varies with male identity and female reproductive state. Behavioral Ecology, 19, 1326-1332. Santangelo, N. 2005. Courtship in the monogamous convict cichlid; what are individuals saying to rejected and selected mates? Animal Behaviour, 69, 143-149. 510 Schneider, J. M. & Lesmono, K. 2009. Courtship raises male fertilization success through 511 post-mating sexual selection in a spider. Proceeding of the Royal Society B, 276, 3105- 512 3111. 513 Shamble, P. S., Wilgers, D. J., Swoboda, K. A. & Hebets, E. A. 2009. Courtship effort is a 514 better predictor of mating success than ornamentation for male wolf spiders. Behavioral 515 Ecology, 20, 1242-1251. 23 516 Sullivan-Beckers, L. & Hebets, E. A. 2011. Modality-specific experience with female 517 feedback increases the efficacy of courtship signaling in male wolf spider. Animal 518 Behaviour, 82, 1051-1057. 519 520 521 522 Wachtmeister, C.-A. 2001. Display in monogamous pairs: a review of empirical data and evolutionary explanations. Animal Behaviour, 61, 861-868. Ward, J. L. & McLennan, D. A. 2009. Mate choice based on complex visual signals in the brook stickleback, Culaea inconstans. Behavioral Ecology, 20, 1323-1333. 523 Widemo, F. & Saether, S. A. 1999. Beauty is in the eye of the beholder: causes and 524 consequences of variation in mating preferences. Trend in Ecology and Evolution, 14, 525 26-31 526 527 Wilgers, D. J. & Hebets, E. A. 2011. Complex courtship displays facilitate male 528 reproductive success and plasticity in signaling across variable environments. Current 529 Zoology, 57, 175-186. 530 531 532 533 534 535 536 537 Wong, B. B. M. & Svensson, P. A. 2009. Strategic male signaling effort in a desert-dwelling fish. Behavioral Ecology and Sociobiology, 63, 543-549. Wells, K. D. 2007. The ecology and behavior of amphibians. Chicago and London: Chicago University Press. Zahavi, A. 1975. Mate selection –a selection for a handicap. Journal of Theoretical Biology, 53: 205-214. Zuur, A. F., Ieno, E. N., Walker, N. J., Saveliev, A. A., Smith, G. M. 2009. Mixed effects models and extensions in ecology with R. Springer Science. 24 ARTICLE 4 Condition-dependance of multiple sexual traits in the palmate newt (Lissotriton helveticus). Cornuau JH, Pigeault R, Schmeller DS, Loyau A. En préparation 1 Condition-dependence of multiple sexual traits in the palmate newt (Lissotriton 2 helveticus) 3 4 Jérémie H. Cornuaua • Romain Pigeaulta • Dirk S. Schmellera,b • Adeline Loyaua,b 5 a 6 09200 Saint Girons, France 7 b 8 Permoser Strasse 15, 04318 Leipzig, Germany Station d’Ecologie Expérimentale du CNRS à Moulis, 2-4 Route du CNRS, Moulis Village, Department of Conservation Biology, Helmholtz Centre for Environmental Research – UFZ, 9 10 Correspondence to J. Cornuau. 11 Email: [email protected] 12 Phone: + 33 5 61 04 03 60 13 Fax : + 33 5 61 96 08 51 14 15 16 17 18 1 19 Abstract Why females evaluate more than one sexual trait to choose their mates has received 20 increasing attention in recent years. The redundant messages, the multiple messages and the 21 unreliable signals hypotheses are the three main theories put forward to explain the evolution 22 of preference for multiple signals. To test these hypotheses, we examined the co-variation of 23 multiple traits with two aspects of male condition in the palmate newt Lissotriton helveticus. 24 We measured body condition and experimentally injected males with an antigen (LPS) to 25 manipulate health status. The development of three morphological sexual traits (filament 26 length, hind-feet-web size, and crest size) was positively correlated (suggesting a relationship 27 and no trade-off between these traits), and were related to the body condition. These results 28 support the redundant messages hypothesis. In contrast, the development of the sexual traits 29 was not affected by the experimental treatment (LPS vs. saline solution). Courtship frequency 30 was not influenced by any aspect of male condition we measured, rather providing some 31 support for the unreliable signal hypothesis. Our results suggest that females likely obtain 32 redundant information on male condition when evaluating filament length, hind-feet-web size, 33 and crest size during mate choice, and may obtain unreliable information when assessing 34 courtship activity alone. Our results further suggest that complex, multiple traits may evolve 35 because redundant messages, multiple messages and unreliable signals can coexist. 36 37 38 39 Keywords Sexual selection • Multiple signal theory • Redundancy • Back-up signal • LPS • 40 Lissotriton helveticus 41 42 43 2 44 Introduction 45 In many species, females assess several sexual traits during mate choice such as 46 coloration, morphological structures, calls and behavioral displays (Candolin 2003, Grether et 47 al. 2004, Hebets and Papaj 2005). Usually, females prefer mating with the males that are able 48 to express the most extravagant, conspicuous, sexual traits. This preference has been 49 explained by the cost of expression and maintenance of the traits, so that only males in good 50 condition can afford these costs. The development of a trait then reliably signals male quality 51 to the choosing females by co-varying with condition (Zahavi 1975, Johnstone et al. 2009). 52 For example, females could base their choice on male coloration because coloration is a good 53 indicator of male body condition (McGlothlin et al. 2007, Griggio et al. 2011, Taylor et al. 54 2011) or male immune abilities to cope with parasites and pathogens (Faivre et al. 2003; 55 Alonso-Alvarez et al. 2004; Baeta et al. 2008). In contrast, it has been argued that an 56 exaggerate male trait can evolve even if it is unreliable, as long as the trait is heritable and 57 female preference for this trait is heritable too (Fisher 1930). 58 If females can assess male attractiveness or quality using only one single trait, why do 59 females commonly use several, multiple, traits during mate choice? This question has 60 received growing consideration over the two past decades and continues to be an active 61 research area (Candolin 2003, Hebets and Papaj 2005, Bro-Jørgensen 2010). To date, several 62 theories have been developed to explain the evolution and maintenance of multiple sexual 63 traits and have been explored with mathematical models (reviewed by Candolin 2003). First, 64 the unreliable signal hypothesis proposed that the costs of secondary sexual traits and the 65 costs of assessing multiple traits should favor a single most-revealing trait, and thus other 66 additional traits should be unreliable (Møller and Pomiankowski 1993, Iwasa and 67 Pomiankowski 1994). Therefore, the expression of only one trait should co-vary with an 68 aspect of male condition. Secondly, unreliable traits may have evolved or be maintained 3 69 because they enhance the probability that the reliable signal is perceived, making it easier to 70 assess male attractiveness (amplifiers, Hebets and Papaj 2005). Thirdly, it has been shown 71 that the evolution of multiple traits can also be based on multiple condition-dependent sexual 72 traits (van Doorn and Weissing 2004, 2006), with each trait being related to a different aspect 73 of male condition (the multiple message hypothesis). Alternatively each condition-dependent 74 trait can be linked to the same aspect of male condition with a certain error (the redundant 75 signal hypothesis; Møller and Pomiankowski 1993). In this case, the redundant traits should 76 be correlated to each other, should co-vary with a given measure of individual condition, and 77 mate choice should be more accurate when several traits are used compared to only one trait 78 (Møller and Pomiankowski 1993, Partan and Marler 1999, 2005). Finally, an empirical work 79 suggested that the unreliable signal, multiple message and redundant signal hypotheses are 80 not necessarily mutually exclusive (Doucet and Montgomerie 2003). Indeed, these authors 81 showed that, in the Satin bowerbird Ptilonorhynchus violaceus, both the multiple message and 82 the redundant signal hypotheses contribute to explain the evolution of multiple sexual traits. 83 Despite the fact that Doucet and Montgomerie (2003) predicted that future studies will reveal 84 additional examples of such a “mosaic” of multiple signals, to our knowledge, no further 85 example has been identified so far. 86 Several approaches have been proposed to experimentally challenge the unreliable 87 signal, the multiple message and the redundant signal hypotheses (Partan and Marler 1999, 88 2005; Candolin 2003; Hebets and Papaj 2005). One approach consists in examining the 89 intensity of the receiver’s response to each trait separately versus altogether (Partan and 90 Marler 1999, 2005). Similar responses are interpreted as redundant traits whereas different 91 responses are interpreted as non-redundant traits. However, it is not always possible to 92 manipulate each trait separately with non-invasive methods. Another approach focuses on the 93 relationships between the traits and the information content of the traits, with similar 4 94 information content interpreted as redundant traits (Candolin 2003; Hebets and Papaj 2005). 95 Interestingly, these two approaches could lead to opposite conclusions about trait redundancy. 96 Indeed, two traits could provide similar responses (i.e. redundant) while reflecting different 97 pieces of information (i.e. non-redundant). We investigated the evolution of multiple traits in 98 the palmate newt using the information-content approach. This second approach entails 1) 99 examining whether the expression of each trait co-varies with male condition to explore the 100 information content of the trait (reliable or unreliable trait), 2) testing whether multiple traits 101 co-vary and if so positively or negatively (redundant or multiple traits), and 3) investigating 102 correlations and trade-offs with several aspects of male condition (redundant or multiple 103 traits). 104 We applied this framework in an amphibian, the palmate newt Lissotriton helveticus. 105 Male palmate newts display multiple conspicuous sexual traits that females assess during 106 mate choice (Haerty et al. 2007, Cornuau et al. 2012). These traits include both morphological 107 traits (a caudal filament at the end of the tail, a small crest on the upper part of the tail and 108 hind-foot-webs) and courtship display behaviors (Cornuau et al. 2012). Additional sexual 109 traits may also play an important role in determining female mate choice, such as pheromones 110 or genetic relatedness, as they do in closely related newt species (Garner & Schmidt 2003, 111 Jehle et al 2007, Houck 2009). Males of the palmate newt with more conspicuous 112 morphological and behavioral sexual traits transfer more sperm masses to the females 113 (Cornuau et al.2012). However, the information content of these multiple traits is still 114 unraveled. To address this issue we have tested whether the expression of sexual traits 115 assessed by females co-vary altogether and with male condition. To test the condition- 116 dependence of the multiple sexual traits we used two classical methods: a measure of body 117 condition and the response to an experimental injection of an antigen (gram-negative bacterial 118 lipopolysaccharides, LPS; Cotton et al. 2004). 5 119 120 Material and methods 121 Reproductive behavior in the Palmate newt 122 In the palmate newt after the terrestrial phase, males and females gather in aquatic ponds for 123 the breeding season (Griffiths and Mylotte 1988). In this species, courtship display is 124 composed of several phases (Halliday 1975, Cornuau et al. 2012). In the first phase named 125 orientation, the male searches and follows the female, using his hind-foot-webs and tail 126 (Halliday 1975). Then, the male places himself in front of the female and presents his tail 127 (with crest and filament) in full view to the female and produces three different behaviors 128 named whip, fan and wave with his tail, crest and filament (Halliday 1975). A whip represents 129 a rapid movement in which the male snaps his tail against his flank before making the fan. 130 The fan is a rapid vibration of the distal part of the tail including the filament that creates a 131 water flow towards the female’s head. During fan, the two hind-foot-webs could be crucial for 132 the male to maintain the upper part of his body immobile. Wave is a movement of the tail of 133 the male where he presents his tail including crest and filament in full view to the female. At 134 any time the female can stay or move away according to her interest for the male (Cornuau et 135 al. 2012). Courtship display in front of a female can last from a few seconds to more than 10 136 minutes (JHC personal observation). A female can watch several males over a short period of 137 time and can easily compare males according to their morphological and behavioral features. 138 The spermatophore transfer phase begins with the male’s offer to deposit a spermatophore to 139 the female with a small movement with his tail named quiver. Again the female has a full 140 view on the male tail. The female can accept or refuse spermatophore deposition by the male, 141 depending on whether she touches the extremity of the male tail or not. After spermatophore 142 deposition, the male places his body perpendicular to the female body and the female is 6 143 guided by the male so that her cloaca is on the spermatophore. Finally the female can transfer 144 the sperm mass in her cloaca. 145 146 Laboratory experiment 147 Palmate newts were captured in April 2011 during aquatic phase in their natural breeding site 148 (Etang Bouteve, Mourtis, France, altitude: 1682 m). Before the experiments, the individuals 149 were individually marked with subcutaneous implants of colored visible elastomers (VIE, 150 Northwest Marine Technology, Washington, Shaw Island, WA, USA) at the basis of one of 151 the four legs. During their entire captivity newts were fed ad libitum with larvae of 152 chironomids, daphnia and tubifex. A total of 123 males were included in our experiment. 153 To explore the trade-offs between sexual traits and male condition we used two 154 common methods (Cotton et al. 2004); 1) we measured natural variation in body condition 155 (Baker 1992, Băncilă et al. 2010), and 2) we injected newts with bacterial-wall 156 lipopolysaccharide (LPS). The use of the body condition index was probably the most 157 widespread and validated measure of male quality (Cotton et al. 2004). In amphibians the use 158 of the body condition index as an index of male quality is widespread particularly to study the 159 condition dependence of sexual traits (Green 1991, Baker 1992, Judge & Brook 2001). This 160 measure reflects the lipid content in newts (Denoël et al. 2002) and body condition in other 161 species (Cotton et al. 2004; Schulte-Hostedde et al. 2005) and was recently recommended for 162 the monitoring of amphibian populations (Băncilă et al. 2010). 163 The LPS is an antigen that mimics a bacterial attack and is readily recognized by a 164 host’s immune system (Llewellyn et al. 2011). Its constituents are among the most potent 165 activators of both innate and humoral immune processes of a vertebrate host, including 166 amphibian hosts (Table S1). We conducted a literature search to gather information regarding 167 the use of LPS in different types of experiments and on the largest number of species 7 168 possible. The results of this research revealed that only three amphibian species, all anurans, 169 were already injected with LPS (Bufo marinus, B. paracnemis and Xenopus laevis) and that 170 the response time to the immune challenge was found to be 1 to 30 days (Table S1). It also 171 allowed us to determine the concentration, volume and location of the LPS-injection for our 172 experiment (see Table S1). Newts were injected intraperitoneally with 0.02 mL of a solution 173 composed either of LPS (Escherichia coli serotype 055: B5, Sigma, Lyon) in saline solution 174 (0.9%), or of saline solution only. Thirty three newts were challenged with 0.14 mg of 175 LPS/ml (low dose, which approximately corresponds to a 2 mg of LPS/kg for an average 176 sized-individual of 1.4g as by Llewellyn et al. 2011 in B. marinus and Bicego et al. 2002 in B. 177 paracnemis, Table S1), 30 newts with 0.7 mg of LPS/ml (high dose, which approximately 178 corresponds to a 10 mg of LPS/kg for an average sized-individual, Table S1), and 60 newts 179 with a saline solution (control). Due to the small body mass of newts and the low resolution of 180 the standard tuberculin syringe used, we could not adjust the injected volume to SVL or body 181 mass. However, we took this individual difference into account by including SVL as a 182 covariate in the models. After injection, each individual was placed in an opaque tank 183 together with 11 other individuals. Tank groups were composed of 6 males injected with LPS 184 (both males injected with low and high doses) plus 6 males injected with the saline solution 185 (except for three tanks in which 7 males were injected with a low dose of LPS). Tanks 186 contained 10 liters of water, plants collected in the native environment of the newts, one 187 shelter perforated with 17 holes, and an air-pump to oxygenate the water. The temperature 188 was kept constant at 18°C (±1) and we used daylight lamps (Reptisun 2.0, ZooMed) to 189 simulate the natural light spectrum with a natural light/dark cycle (12h:12h). 190 Morphological sexual traits were measured at the start of the experiment, and 32 days 191 later at the end of the experiment. The length of the experiment was determined by a 192 preliminary experiment (JHC, data on demand). We recorded body mass with a digital scale 8 193 (accuracy: 0.01 g) and measured Snout-Vent-Length (SVL) using a photo taken on a 194 millimeter paper background (grid 1 mm²). In the same way, we measured filament length, 195 hind-foot-web size and crest size (Green 1991). The photos were analyzed using the ImageJ 196 software (http://rsbweb.nih.gov/ij/), a reliable and repeatable method for morphological 197 measures as shown earlier (Cornuau et al.2012). The body condition index (BCI) was 198 calculated for each individual using the residual of the linear regression of the cube root of 199 body mass on SVL, as recommended for amphibians (see Baker 1992 and Băncilă et al. 200 2010). The development of morphological traits was calculated as the value of the trait 201 measured at the start of the experiment minus the value of the trait measured at the end of the 202 experiment (Δ trait= traitend - traitstart). The development of morphological sexual traits (Δ 203 trait) did not differ between tanks (Kruskal-Wallis rank sum test, Δ filament length: χ² = 8.66, 204 df = 9, P = 0.47; Δ crest size: χ² = 8.12, df = 9, P = 0.52; Δ hind-foot-web size: χ² = 1.35, df = 205 9, P = 0.99). After the experiment, the individuals were released in their original breeding 206 site. 207 208 Outdoor experiment 209 Newts (60 males and 45 females) were collected during their reproductive period (10 May 210 2011) in a natural pond (Caumont, France, altitude: 438 m). They were individually marked 211 with VIE (Visible Implant fluorescent Elastomer) and measured for morphological sexual 212 traits (SVL, weight and filament length) as described above for our laboratory experiment. 213 The individuals were then placed in three outdoor tanks of a volume of 1000 liters by groups 214 of 20 males and 15 females per tank. Each tank was covered by a net to avoid flight and 215 predation, was filled with water, and contained aquatic plants from the pond of origin and 216 shelters. In this experiment, courtship frequency of each male was measured by counting the 217 number of times a given male was observed in courtship. Three sessions of observations (S0, 9 218 S1 and S2) were performed during this experiment. One session corresponded to three days of 219 observation and each session was separated by three days. In the first session of three days 220 (S0), behavioral observations were conducted continuously from 6:00 am to 7:00 pm. It 221 allowed to determine the time of day where most males interacted with females, which was 222 between 6:00 am and midday and between 5:00 pm and 7:00 pm. In the following sessions S1 223 and S2, the observer did perform subsequent observations during the activity peaks. 224 During each session (S0, S1 and S2) the observations were performed by one observer (RP) 225 walking around the three tanks until seeing a courting male. Daily observations corresponded 226 to a succession of 15 min of focal observations during which one tank was observed at a time 227 and all individuals in the tank observed altogether. Every 15 min, a different tank was 228 observed, and the tank order was changed everyday. Globally all tanks were similarly 229 observed, so the likelihood to observe male courtship was constant between tanks and there 230 was no effect of tank identity on the total courtship frequency observed in the tank (F1,54 = 231 1.24, P = 0.27). VIE marking did not allow remote identification of individuals. The courting 232 male was therefore caught using a net, identified and immediately released in the tank. We 233 counted the number of times each newt was seen in courtship activity during each session 234 over three days. If a male was seen two times in courtship it scored 2, whereas if a male was 235 seen twenty times in courtship it scored 20. Thus we obtained an index value of courtship 236 frequency of each male and were unable to assess courtship duration or any other measure of 237 intensity of display effort. 238 Male newts were assigned to the experimental treatment the night before S1 and the 239 night before S2. In each tank, at around 11:00 pm, 10 males were injected with a saline 240 solution and 10 males with LPS (10mg/kg) and immediately released back in their initial 241 outdoor tank. Courtship frequency was recorded the day following injection, and for a total of 242 three days (observation session S1). Then, females were removed from the tank for a three- 10 243 day period without observation, and placed back before the second injection of the males. 244 Males were injected a second time and a second observation period (S2) was performed under 245 similar conditions as described above. Post-injection courtship frequency was defined as the 246 total number of courtship observed during the two observational periods S1 and S2. At the 247 end of the experiment, morphological traits were measured again, and all newts were released 248 in their original breeding site. 249 250 Statistical analysis 251 Statistical analyses were performed using R (R development core team 2009, version 2.9.2). 252 First, we checked if male morphological traits (filament length, hind-foot-web size, crest size, 253 SVL and weight for the laboratory experiment; filament length, SVL, weight and courtship 254 frequency for the outdoor experiment) were homogenously distributed among experimental 255 treatments (saline, low and high doses of LPS) using a MANOVA test (laboratory 256 experiment: F2,10 = 0.88, P = 0.56; outdoor experiment: F1,4 = 0.10, P = 0.81). 257 Then, we used linear mixed models (LMMs) because assumptions for parametric 258 analyses were met (all LMMs had a Gaussian distribution of error terms), the link function 259 was identity. We built LMMs to explore relations between pairs of Δ traits (Δ filament length, 260 Δ crest size and Δ hind-foot-web size). In these models, we included the experimental 261 treatment (0, 2 and 10 mg/kg of LPS), Δ SVL and two-factor interactions as fixed variables, 262 and tank identity as a random variable. Including Δ SVL as a covariate controlled for 263 potential effects of Δ body size (Cotton et al. 2004). We evaluated the effect of the random 264 factor (tank identity) by testing whether its removal from the model caused a significant 265 decrease in the model fit and found that this random effect was never significant (all P > 266 0.05).We used linear mixed models (LMMs) to investigate the impact of the experimental 267 treatment (high dose vs. saline, low dose vs. saline, and low vs. high dose) on Δ traits (Δ 11 268 filament length, Δ hind-foot-web size and Δ crest size). We investigated whether Δ traits co- 269 vary with Δ BCI using LMMs with BCI, experimental treatment and two-factor interactions 270 as fixed variables, and tank identity as a random variable. Again, this random effect was never 271 significant. For the outdoor experiment, examined the relationship between S1 and S2 using a 272 LMM, with S2 as a fixed variable and tank identity as a random variable. We used LMMs to 273 test the effect of the LPS treatment and the body condition on courtship frequency, with tank 274 identity as a random variable. We investigated also, with LMMs the link between courtship 275 frequency, filament length and SVL. 276 277 Results 278 Laboratory experiment 279 Over the experiment, the size of the morphological traits did not remain constant. There was a 280 significant increase in SVL (paired t-test, t122 = 9.95, P < 0.001), body weight (paired t-test, 281 t122 = 15.95, P < 0.001), filament length (paired t-test, t122 = 3.63, P < 0.001) and hind-foot- 282 web size (paired t-test, t122 = 2.40, P = 0.018). In contrast, crest size showed a significant 283 decrease (paired t-test, t122 = -3.82, P < 0.001). Delta filament length, Δ hind-foot-web area 284 and Δ crest area were positively related (Table 1 and Table S2), and these relations were not 285 affected by Δ SVL and LPS treatment (0, 2 and 10 mg/kg; Table 1). 286 Contrary to our prediction Δ filament length, Δ hind-foot-web size and Δ crest size 287 were not affected by the injection of neither a low dose (Δ filament length: F 1,61 = 0.55, P = 288 0.46, Δ hind-foot-web size: F 1,61 = 0.01, P = 0.91 and Δ crest size: F 1,61 = 0.007, P = 0.94), 289 nor a high dose of LPS (Δ filament length: F 1,58 = 0.28, P = 0.60, Δ hind-foot-web size: F 1,58 290 = 0.001, P = 0.97 and Δ crest size: F 1,58 = 0.30, P = 0.59). Generally, there was no difference 291 between low and high dose treatments (Δ filament length: F 1,61 = 3.43, P = 0.07, Δ hind-foot- 292 web size: F 1,61 = 0.53, P = 0.47 and Δ crest size: F 1,61 = 0.84, P = 0.36). However, each 12 293 morphological sexual trait was significantly and positively influenced by the variation in BCI 294 (Table 2, Fig. 1). An increase in male Δ BCI marked an increase in Δ filament length, Δ crest 295 area and Δ hind-foot-web area, and was not affected the experimental treatment (Table 2, Fig. 296 1). 297 298 Outdoor experiment 299 Males that performed more courtships in the first observation session (S1) performed 300 also more courtships in the second session (S2) (F 1,54 = 232.33, P < 0.001). Post-injection 301 courtship frequency (S1+S2) was not explained by the experimental treatment (F1,52 = 0.98, P 302 = 0.33), BCI (F1,52 = 0.09, P = 0.76) or their interaction (F1,52 = 0.01, P = 0.91). However, 303 courtship frequency was significantly related to the SVL at the end of the experiment (F1,52 = 304 6.62, P = 0.013) and marginally to filament length (F1,52 = 3.39, P = 0.071) but not to their 305 interaction (F1,52 = 0.08, P = 0.78). Our analysis indicates that the longer the filament length 306 and the lower the SVL, the higher the courtship frequency (Fig. 2). 307 308 Discussion 309 In the palmate newt, both morphological (filament length, crest size, hind-foot-web size) and 310 behavioral (courtship activity) traits influence female mate choice (Cornuau et al. 2012, see 311 also Haerty et al. 2007). In this study, we aimed at understanding the information content of 312 these traits by investigating their condition-dependence to 1) natural variation in body 313 condition and 2) response to an experimental injection of LPS. Here, our results obtained on 314 morphological sexual traits provide support for the redundant signal hypothesis (Møller & 315 Pomiankowski 1993, Candolin 2003). Indeed, the development of filament length, crest area 316 and hind-foot-web area co-vary positively and all these traits co-vary with male condition as 317 measured by the body condition index BCI (even though these traits were not affected by the 13 318 experimental injection of LPS). Conversely, courtship frequency was not influenced by male 319 morphological sexual traits or male condition and was not affected by the LPS-injection. Our 320 study supports the idea that courtship frequency may not be a reliable signal of male condition 321 (unreliable signal hypothesis, Møller & Pomiankowski 1993). 322 The development of filament length, crest area and hind-foot-web area was positively 323 correlated under the conditions of our experiment, and this relation was stronger than a simple 324 allometric relationship with body size. This is a first evidence to validate the redundant 325 message hypothesis (Hebet and Papaj 2005). It further shows that there is no energy or 326 physiological constraint for the production and maintenance of the filament length, the hind- 327 foot-web size and the crest size, contrasting to the multitasking hypothesis stating that the 328 signaler’s ability to produce one sexual trait is constrained by the ability to produce other 329 sexual traits (Hebet and Papaj 2005). However, in our experiment, newts were fed ad libitum 330 that could reduce the likelihood to find trade-offs among traits. Moreover, we cannot exclude 331 the existence of individual strategies in the development of these sexual traits. For example, 332 some males could invest more in filament length whereas others invest more in crest size 333 maybe linked to internal factors or environmental conditions. Additional experiments are 334 therefore needed to determine the relationship between morphological sexual traits at the 335 individual level more precisely. 336 A second assumption of the redundant message hypothesis is that all traits should co- 337 vary with one aspect of the sender’s condition (Hebet and Papaj 2005). In agreement with this 338 assumption, filament length, crest size and hind-foot-web size were all good predictors of 339 male condition, measured as body condition index. Our experimental study therefore provides 340 one additional example to the few examples of redundant signals found so far (birds: 341 Birkhead et al. 1998, Jawor et al. 2004, Roulin et al. 2011, frogs: Vásquez and Pfennig 2007, 342 lizard: Martἰn and Lόpez 2010, and spiders: Gibson and Uetz 2008). Our correlative data 14 343 showed that the better the body condition (result of food intake) the better developed the 344 sexual traits were. This is in line with previous correlative studies showing that crest height 345 was linked to body condition in the crested newt Triturus cristatus, and the smooth newt 346 Triturus vulgaris (Green 1991, Baker 1992), suggesting that condition-dependence of 347 morphological sexual traits is ubiquitous in newts. These traits grow at the start of the 348 breeding season in newts, when males go into the water (Griffiths and Mylotte 1988). They 349 could therefore provide information about the food intake during this critical period, and 350 females probably obtain a more reliable picture of the overall male quality when they assess 351 several ornaments related to male condition, which increases the likelihood to select a good 352 quality male (Møller and Pomiankowski 1993). 353 A third rarely tested assumption of the redundant message hypothesis is that “the 354 probability of selecting a truly high condition mate will be improved when females choose by 355 examining two or more male traits, if the errors in different traits are uncorrelated” (Møller 356 and Pomiankowski 1993, see also Partan and Marler 1999, 2005). For example in the barking 357 tree frog Hyla gratiosa, female preference for larger males was stronger when females had 358 access to all call characteristics than when only one was available (Poole and Murphy 2007). 359 In this study we did not challenge this third assumption that need to examine how female 360 react to each trait separately (here filament, hind-foot-web and crest) that is impossible for the 361 moment without important damage for newts. Yet interestingly, we have recently explored 362 female mate choice in palmate newts by manipulating filament length. When the relative 363 filament length between two males was experimentally reversed (higher became smaller), the 364 choice of females was not totally reversed. This result suggests that, in addition to filament, 365 hind-foot web and crest development may be necessary to females to choose a mate (Cornuau 366 et al. 2012). However, further experiments with independent manipulation of the three sexual 367 traits are needed to confirm this hypothesis. 15 368 Male condition is a broad and complex concept that may relate to capacities of energy 369 storage and fat reserves, parasitic load or health status. In our experiment, we used body 370 condition as one indicator of male condition, and we also injected newts with an elevated dose 371 of an antigen (LPS) to experimentally mount an immune response that mimic an alteration of 372 health status. LPS has been shown to efficiently activate the immune system in a large range 373 of taxa including insects, birds, mammals, reptiles and anuran amphibians (Table S1). To our 374 knowledge, no LPS challenge has been reported so far in urodele species, such as newts. 375 More specifically, LPS has an impact on the metabolism and activity of frogs (Table S1) and 376 the newt immune system is not dramatically different from frogs (Todd 2007; Raffel et al. 377 2009). We could therefore expect that the LPS injection may yield the same immune 378 activation in newts as in frogs, despite being unable to verify it at the physiological level. In 379 our study, we did not find any effect of LPS on morphological and behavioral sexual traits. In 380 the absence of physiological, morphological and behavioral evidence that the injected 381 individuals actually mounted an immune response, we cannot exclude a non-activation of the 382 immune system by LPS in the palmate newt. Such non-activation is surprising given the 383 similarities between anuran and urodele immune systems, but not impossible. Given that an 384 activation of the immune system of anurans by LPS could sometime take up to 30 days 385 (Llewellyn et al. 2011, Table S1), the immune reaction of newts in our study may have been 386 slow and still below the threshold of detectability after 3 days. Another possibility is that LPS 387 did activate the new immune system but this activation did not have a negative impact on the 388 expression of morphological and behavioral sexual traits. Similar results were found in the 389 zebra finch Taeniopygia guttata in which sexual traits appear condition-dependent but not 390 linked to immune capacity (Birkhead et al. 1998). There could be an absence of a relationship 391 between sexual traits and health status in male palmate newts, whatever the environmental 392 conditions experienced, or environmental conditions may hide trade-offs between condition 16 393 and sexual traits due to ceiling or floor effects. Indeed, good environmental conditions may 394 hide trade-offs between condition and sexual traits that are sometimes revealed in harsh 395 environmental conditions (e.g. Jacot et al. 2005, Leman et al. 2009). Stress of captivity or 396 high initial parasitic load in the newt population, on the other side, may prevent mobilization 397 of energy to mount an immune response. In the absence of evidence of LPS-driven immune 398 activation, only one aspect of male condition was measured with certainty in this study, and 399 thus we cannot draw conclusions on the multiple message hypothesis. 400 In palmate newts, males with higher courtship activity have higher reproductive 401 success (Cornuau et al. 2012); however, courtship activity was not influenced by either body 402 condition or the experimental treatment that supports a low cost of this behavioral sexual trait. 403 Courtship display could obviously contain information on aspects of male quality that were 404 not assessed in this study, such as sperm quality (Chargé et al. 2010, Weir & Grant 2010), 405 locomotor ability (Byers et al. 2010) or heterozygosity (Drayton et al. 2010). However, the 406 cost of courting has been evaluated in a related species, the Plethodontid salamander 407 Desmognathus ochrophaeus, and the results were consistent with our findings: courtship 408 activity was not subject to any energetic limitations (Bennett and Houck 1983). This low cost 409 of courting supports the unreliable signals hypothesis. 410 In conclusion, while we were not able to specifically test for the multiple message 411 hypothesis, our results provide support for the redundant messages hypothesis for 412 morphological sexual traits (filament length, hind-food-web size and crest development), and 413 the unreliable signal hypothesis for the behavioral sexual trait (courtship frequency). Our 414 results are in line with previous work suggesting that the unreliable, multiple message and 415 redundant messages hypotheses are not necessarily mutually exclusive (Doucet and 416 Montgomerie 2003). Future experiments should manipulate the three morphological sexual 417 traits to further investigate their relative effect on female mating decisions. 17 418 419 Acknowledgments 420 Adelaïde Sibeaux is thanked for technical assistance. We also thank four anonymous 421 reviewers for insightful comments that greatly improved the manuscript. Catching permits 422 n°2009-13 (Ariège) and n°2009-12 (Haute Garonne). This work was supported by the 423 Ministère de la Recherche (PhD fellowship to JHC), two CNRS grants to AL and DS, the 424 BioDiversa-project RACE and the Observatoire Homme-Milieu Pyrénées Haut-Vicdessos. 425 426 Ethical standards 427 The experiments comply with the current laws of the country in which they were performed. 428 429 References 430 Alonzo-Alvarez C, Bertrand S, Devevey G, Gaillard M, Prost J, Faivre B, Sorci G (2004) An 431 experimental test of the dose-dependent effect of carotenoids and immune activation on 432 sexual signals an antioxidant activity. Am Nat 164:651–659 433 Baeta R, Faivre B, Motreuil S, Gaillard M, Moreau J (2008) Carotenoid trade-off between 434 parasitic resistance and sexual display: an experimental study in the blackbird (Turdus 435 merula). Proc R Soc Lond B 275:427–434 436 437 Baker JMR (1992) Body condition and tail height in great crested newts, Triturus cristatus. Anim Behav 43:157–159 438 Băncilă RI, Hartel T, Plaiasu R, Smets J, Cogalniceanu D (2010) Comparing three body 439 condition indices in amphibians: a case study of yellow-bellied toad Bombina variegate. 440 Amphib Reptil 31:558–562 441 442 Bennett AF, Houck LD (1983) The Energetic Cost of Courtship and Aggression in a Plethodontid Salamander. Ecology 64:979–983 18 443 444 Bicego KC, Steiner AA, Antunes-Rodrigues J, Branco LGS (2002) Indomethacin impairs LPS-induced behavioral fever in toads. J Appl Physiol 93:512–516 445 Birkhead TR, Fletcher F, Pellatt EJ (1998) Sexual selection in the zebra finch Taeniopygia 446 guttata: condition, sex traits and immune capacity. Behav Ecol Sociobiol 44:179–191 447 Bro-Jørgensen J (2010) Dynamics of multiple signaling systems: animal communication in a 448 449 450 world in flux. Trends Ecol Evol 25:292–300 Byers J, Hebets E, Podos J (2010) Female mate choice based upon male motor performance. Anim Behav 79:771–778 451 Candolin U (2003) The use of multiple cues in mate choice. Biol Rev 78:575–595 452 Chargé R, Saint Jalme M, Lacroix F, Cadet A, Sorci G (2010) Male health status, signalled by 453 courtship display, reveals ejaculate quality and hatching success in a leeking species. J 454 Anim Ecol 79:843–850 455 456 Cornuau JH, Rat M, Schmeller DS, Loyau A (2012) Multiple signals in the palmate newt: ornaments help when courting. Behav Ecol Sociobiol 66:1045–1055 457 Cotton S, Fowler K, Pomiankowski A (2004) Do sexual ornaments demonstrate heightened 458 condition-dependent expression as predicted by the handicap hypothesis? Proc R Soc 459 Lond B 271:771–783 460 461 462 463 464 465 Denoël M, Hervant F, Schabetsberger R, Joly P (2002) Short- and long-term advantages of an alternative ontogenetic pathway. Biol J Linn Soc 77:105–112 van Doorn GS, Weissing FJ (2004) The evolution of female preferences for multiple indicators of quality. Am Nat 164:173–186 van Doorn GS, Weissing FJ (2006) Sexual conflict and the evolution of female preferences for indicators of male quality. Am Nat 168:742–757 19 466 Doucet SM, Montgomerie R (2003) Multiple sexual ornaments in satin bowerbirds: 467 ultraviolet plumage and bowers signal different aspects of male quality. Behav Ecol 468 14:503–509 469 Drayton JM, Milner RNC, Hunt J, Jennions MD (2010) Inbreeding and advertisement calling 470 in the cricket Telegryllus commodus: laboratory and field experiments. Evolution 471 64:3069–3083 472 473 Faivre B, Grégoire A, Préault M, Cézilly F, Sorci G (2003) Immune activation rapidly mirrored in a secondary sexual trait. Science 300:103 474 Fisher RA (1930) The genetical theory of natural selection. Clarendon Press, Oxford. 475 Garner TWJ, Schmidt BR (2003) Relatedness, body size and paternity in the alpine newt, 476 Triturus alpestris. Proc R Soc Lond B 270:619–624 477 Gibson JS, Uetz GW (2008) Seismic communication and mate choice in wolf spiders: 478 components of male seismic signals and mating success. Anim Behav 75:1253–1262 479 Green AJ (1991) Large male crests, an honest indicator of condition, are preferred by female 480 smooth newts, Triturus vulgaris (Salamandridae) at the spermatophore transfer stage. 481 Anim Behav 41:367–369 482 483 Grether GF, Kolluru GR, Nersissian K (2004) Individual colour patches as multicomponent signals. Biol Rev 79:583–610 484 Griffiths RA, Mylotte VJ (1988) Observations on the development of the secondary sexual 485 characters of male newts, Triturus vulgaris and T. helveticus. J Herpetol 22:476–480 486 Griggio M, Valera F, Cassas-Crivillé A, Hoi H, Barbosa A (2011) White tail markings are an 487 indicator of condition and affect mate preference in rock sparrows. Behav Ecol Sociobiol 488 65:655–664 20 489 Haerty W, Gentilhomme E, Secondi J (2007) Female preference for a male sexual trait 490 uncorrelated with male body size in the palmate newt (Triturus helveticus). Behaviour 491 144:797–814 492 Halliday TR (1975) On the biological significance of certain morphological characters in 493 males of the Smooth newt Triturus vulgaris and of the Palmate newt Triturus helveticus 494 (Urodela: Salamandridae). Zool J Linn Soc 56:291–300 495 496 497 498 499 500 501 502 Hebets EA, Papaj DR (2005) Complex signal function: developing a framework of testable hypotheses. Behav Ecol Sociobiol 57:197–214 Houck LD (2009) Pheromone communication in amphibians and reptiles. Annu Rev Physiol 2009:161–176 Iwasa Y, Pomiankowski A (1994) The evolution of mate preferences for multiple sexual ornaments. Evolution 48:853–867 Jacot A, Scheuber H, Kurtz J, Brinkhof MWG (2005) Juvenile immune status affects the expression of a sexually selected trait in field crickets. J Evol Biol 18:1060–1068 503 Jawor JM, Gray N, Beall SM, Breitwisch R (2004) Multiple ornaments correlate with aspects 504 of condition and behaviour in female northern cardinals, Cardinalis cardinalis. Anim 505 Behav 67:875–882 506 Jehle R, Sztatecsny M, Wolf JBW, Whitlock A, Hödl W, Burke T (2007) Genetic 507 dissimilarity predicts paternity in the smooth newt (Lissotriton vulgaris). Biol Lett 3:526– 508 528 509 510 511 512 Johnstone RA, Rands AS, Evans MR (2009) Sexual selection and condition-dependence. J Evol Biol 22:2387–2394 Judge KA, Brooks RJ (2001) Chorus participation by male bullfrogs, Rana catesbeiana: a test of the energetic constraint hypothesis. Anim Behav 62:849–861 21 513 Leman JC, Weddle CB, Gershman SN, Kerr AM, Ower GD, St John JM, Vogel LA, Sakaluk 514 SK (2009) Lovesick: immunological costs of mating to male sagebrush crickets. J Evol 515 Biol 22:163–171 516 Llewellyn D, Brown GP, Thompson MB, Shine R (2011) Behavioral responses to immune- 517 system activation in an Anuran (the Cane toad, Bufo marinus): field and laboratory 518 studies. Physiol Biochem Zool 84:77–86 519 Martἰn J, Lόpez P (2010) Multimodal sexual signals in male ocellated lizards Lacerta lepida: 520 vitamin E in scent and green coloration may signal male condition in different sensory 521 channels. Naturwissenschaften 97:545–553 522 McGlothlin JW, Duffy DL, Henry-Freeman JL, Ketterson ED (2007) Diet condition affects 523 an attractive white plumage pattern in dark-eyed juncos (Junco hyemalis). Behav Ecol 524 Sociobiol 61:1391–1399 525 526 Møller AP, Pomiankowski A (1993) Why have birds got multiple sexual ornaments? Behav Ecol Sociobiol 32:167–176 527 Partan SR, Marler, P (1999) Communication Goes Multimodal. Science 283:1272–1273 528 Partan SR, Marler P (2005) Issues in the classification of multimodal communication signals. 529 530 531 Am Nat 166:231–245 Poole KG, Murphy CG (2007) Preferences of female barking treefrogs, Hyla gratiosa, for larger males: univariate and composite tests. Anim Behav 73:513–524 532 Raffel TR, LeGros RP, Love BC, Rohr JR, Hudson PJ (2009) Parasite age-intensity 533 relationships in red-spotted newts: Does immune memory influence salamander disease 534 dynamics? Int J Parasitol 39:231–241 535 Roulin A, Almasi B, Meichtry-Stier S, Jenni L (2011) Eumelanin- and phenomelanin-based 536 colour advertise resistance to oxidative stress in opposite ways. J Evol Biol 24:2241– 537 2247 22 538 539 540 541 542 543 Schulte-Hostedde AI, Zinner B, Millar JS, Hickling GJ (2005) Restitution of mass-size residuals: validating body condition indices. Ecology 86:155–163 Taylor LA, Clark DL, McGraw KJ (2011) Condition dependence of male display coloration in a jumping spider (Habronattus pyrrithrix). Behav Ecol Sociobiol 65:1133–1146 Todd BD (2007) Parasites Lost? An overlooked hypothesis for the evolution of alternative reproductive strategies in amphibians. Am Nat 170:793–799 544 Vásquez T, Pfennig K (2007) Looking on the bright side: females prefer coloration indicative 545 of male size and condition in the sexually dichromatic spadefoot toad, Scaphiopus 546 couchii. Behav Ecol Sociobiol 62:127–135 547 548 549 Weir LK, Grant JWA (2010) Courtship rate signals fertility in an externally fertilizing fish. Biol Lett 6:727–731 Zahavi A (1975) Mate selection-a selection for a handicap. J Theor Biol 53:205–214 23 Table 1 Relationships between the development of morphological sexual traits for the laboratory experiment. (Δ trait = traitend – traitstart, df = degree of freedom, Sum Sq = sum of the squares, SVL = Snout-Vent-Length, Experimental treatment = injection of LPS or PBS). Response Δ Filament Δ Crest Δ Hind-foot-web Predictor Δ Hind-foot-web Δ SVL Experimental treatment Δ Hind-foot-web : Δ SVL Δ Hind-foot-web : Experimental treatment Residuals Sum Sq 31.64 15.06 0.98 5.77 2.02 133.53 Df 1 1 2 1 2 115 F 27.51 12.9 0.38 4.98 0.87 P <0.001 <0.001 0.69 0.03 0.42 Δ Crest Δ SVL Experimental treatment Δ Crest : Δ SVL Δ Crest : Experimental treatment Residuals 70.84 2.78 0.94 2.75 1.14 110.55 1 1 2 1 2 115 73.7 2.89 0.5 2.86 0.59 <0.001 0.09 0.61 0.09 0.56 Δ Filament Δ SVL Experimental treatment Δ Filament : Δ SVL Δ Filament : Experimental treatment Residuals 17634.23 1680.99 74.89 315.27 1618.44 25724.19 1 1 2 1 2 115 79.56 7.94 0.15 1.85 3.61 <0.001 0.0057 0.86 0.18 0.03 Δ Hind-foot-web Δ SVL Experimental treatment Δ Hind-foot-web : Δ SVL Δ Hind-foot-web : Experimental treatment Residuals 7101.06 5041.50 8.14 444.64 1304.68 33147.96 1 1 2 1 2 115 24.51 18.29 0.05 1.85 2.4 <0.001 <0.001 0.95 0.18 0.096 Δ Filament Δ SVL Experimental treatment Δ Filament : Δ SVL Δ Filament : Experimental treatment Residuals 181.02 1.70 1.73 1.81 29.42 865.78 1 1 2 1 2 115 21.71 0.31 0.09 0.31 1.78 <0.001 0.58 0.91 0.58 0.17 Δ Filament Δ SVL Experimental treatment Δ Filament : Δ SVL Δ Filament : Experimental treatment Residuals 163.23 1.05 0.72 0.01 28.22 888.2326 1 1 2 1 2 115 19.76 0.15 0.07 0.13 1.37 <0.001 0.70 0.93 0.72 0.26 Table 2 Relationships between the development of morphological sexual traits and the variation in Body Condition Index (BCI). (Δ trait = traitend – traitstart, df = degree of freedom, Sum Sq = sum of the squares, Experimental treatment = injection of LPS or PBS). Response Δ Filament Predictor Δ BCI Experimental treatment Δ BCI : Experimental treatment Residuals Sum Sq 9.44 4.71 2.26 172.59 Df 1 2 2 117 F 6.46 2.13 0.81 P 0.012 0.12 0.45 Δ Crest Δ BCI Experimental treatment Δ BCI : Experimental treatment Residuals 10438.80 592.90 699.04 35317.25 1 2 2 117 17.40 34.73 0.93 1.17 <0.001 0.40 0.31 Δ Hind-foot-web Δ BCI Experimental treatment Δ BCI : Experimental treatment Residuals 185.09 3.66 2.98 889.73 1 2 2 117 14.86 21.48 0.49 0.09 <0.001 0.61 0.91 Figures Fig. 1 Positive relationships between the development of morphological sexual trait (Δ trait = traitend – traitstart) and Δ Body Condition Index (Δ BCI). Positive relations with BCI were found for Δ filament, Δ crest and Δ hind-foot-web after controlling for experimental treatment and tank identity (Δ filament: F1,117 = 6.46, P = 0.012; Δ crest: F1,117 = 17.4, P < 0.001; Δ hind-foot-web: F1,117 = 14.86, P < 0.001). Fig. 2 Relationships between courtship activity and two morphological traits. SVL measured at the end of the experiment was negatively correlated to courtship activity (F1,52 = 6.62, P = 0.013) and marginally positively correlated to filament length after controlling for SVL (F1,52 = 3.39, P = 0.071). 1 0 -1 -3 -2 Δ Filament length 2 3 Figure 1 -0.05 0.00 0.05 0.10 -20 -60 -40 Δ Crest 0 20 Δ Body Condition Index -0.05 0.00 0.05 0.10 5 0 -5 Δ hind-foot-web 10 Δ Body Condition Index -0.05 0.00 0.05 Δ Body Condition Index 0.10 25 20 15 0 5 10 15 10 5 0 Behavioral activity 20 25 Figure 2 33 34 35 36 SVL (mm) 37 38 39 3 4 5 6 7 Filament length (mm) 8 Fish Amphibian Amphibian Perca flavescens Bufo marinus Bufo marinus 15 15 1 0.126 12.6 0.01 1 Fish 5 Insect Salmo salar 100 1.2 Insect 20 20 3 10 20 1 65 - Insect 1 20.8 5 Insect Allonemobius socius 0.5 Insect Insect Tenebrio molitor Cyphoderris strepitans Euoniticellus intermedius Gyrllus campestris Gryllodes sigillatus Clade Species - - 0.3 0.1 0.005 0.01 0.001 0.01 0.001 0.005 132-237 100-300 50-200 7.92 0.25 1.2 0.05 0.77 - 0.12 Concentration Quantity Dose per mg/kg injected Weight (g) (mg/ml) animal (ml) weight Table S1 Use of LPS in the literature in a range of different species groups. Supplementary material C C B B A A A A A A hormonal behavioral fever metabolic rate E 0127:B8 E 0127:B8 mx response sexual signal, paternal resources sexual signal, mating success reproductive investment sexual signal, longevity spermatophore production prophylaxis Effect E 055:B5 E 0128:88 - S L6136 S L6136 S L6136 S L6136 E L8274 Injection Serotype method Sherman and Stephens 1998 Salinas et al. 2004 Haukenes et al. 2011 Sherman et al. 1991 Kerr et al. 2010 Jacot et al. 2004 Leman et al. 2009 Reaney and Knell 2010 Fedorka and Mousseau 2007 Moret and SivaJothy 2003 Reference Streptopelia decaocto Zonotrichia leucophrys Pavo cristatus Passer domesticus Taeniopygia guttata Taeniopygia guttata Gallus Gallus Bufo paracnemis Bufo paracnemis Xenopus laevis Lissotriton helveticus Ctenophorus fordi Iguana iguana Anolis carolinensis Podarcis hispanica Bufo marinus - 211.6 2116 0.08 - Reptile Reptile Reptile Reptile 0.1 Bird 1 0.26 Brid Bird 1 0.1 Bird Bird 0.1 Bird - 2.5 25 0.14 0.7 Amphibian Bird 2.5 - Amphibian 1 - 1 0.88 0.58 0.3 0.5 to 5 2.5 5 0.2 2 0.02 0.2 2 10 - Amphibian 0.2 2 - 1 Amphibian Amphibian 0.1 0.2 4 0.15 0.1 0.1 0.1 0.05 0.25 1 - 0.02 - - - - 26 - 4000 17 17 30 - - 3.8-4.6 85 3–5 1.4 200-300 150-250 150-250 45-400 - B D B B B B B B - B B B C C C E 055:B5 E 055:B5 E 055:B5 E 055:B5 E 055:B5 E 0111B4 E 0111:B4 E 055 : B5 E 055:B5 E 0111B4 E L3129 E 055:B5 - E 0111:B4 E 0111:B4 0111:B4 immune responses Sell et al. 2003 Deen and Hutchison 2001 Merchant et al. 2008 López et al. 2009 Uller et al. 2006 Bonneaud et al. 2003 Alonso-Alvarez sexual signal et al. 2004 body weight, Bertrand et al. oxidative damage 2006 Loyau et al. sexual signal 2005 Eraud et al. survival 2009 hormone, Owen-Ashley et activity, song, al. 2006 activity and fitness brain activity sexual signal hypothermia reproductive investment fever or hypothermia sexual signal This study Zou et al. 2000 behavioral fever behavioral fever Llewellyn et al. 2011 Bicego and Branco 2002 Bicego et al. 2002 feeding rate, activity Mammal Mammal Mus musculus Mus musculus 0,1 - 0,5 0.18 0.4 0.5 0.67 0.8 4.65 4 0.05 0.2 0.02 1 0.2 0.05 0.1 25-30 - 20 1500 28.6 16-27 25 Leg B wing B F D E Velando et al. 2006 Losdat et al. 2011 Aubert et al. 1997 Raghavendra et al. 2000 Durazzo et al. 2008 reproductive success nest building, fever hyperalgesia E 0111:B4 sperm Munoz et al. 2010 Owen-Ashley and Wingfield 2006 Toomey et al. 2010 song, territorial defense retinal carotenoid E 08:B28 E 055:B5 E 055:B5 E 055:B5 S L7261 E 055:B5 song, body weight, hormone fever Phodopus Mammal 0.5 0.1 NA S L-6511 food intake sungorus Rattus E 0127: exploratory Mammal 0.05 0.1 B Rico et al. 2010 norvegicus B8 behavior A = In pleural membrane between two abdominal segments, B = Intraperitoneal, C = Dorsal lymphatic sac, D = Intramuscular, E = Intravenous, F = Subcutaneous. Bird Parus major 0.0286 Bird 0.1 2 Bird Bird 1 Bird Sula nebouxii gambelii Melospiza melodia morphna Carpodacus mexicanus Zonotrichia leucophrys oriantha r 0.40 0.61 0.39 95% CI 0.25, 0.54 0.49, 0.71 0.23, 0.53 P <0.001 <0.001 <0.001 205:3513–3518 Bicego KC, Branco LGS (2002) Discrete electrolytic lesion of the preoptic area prevents LPS-induced behavioral fever in toads. J Exp Biol Functional Ecol 20:1022–1027 Bertrand S, Criscuolo F, Faivre B, Sorci G (2006) Immune activation increases susceptibility to oxidative tissue damage in Zebra Finches. mice. Brain Behav Immun 11:107–118 Aubert A, Goodall G, Dantzer R, Gheusi G (1997) Differential effects of lipopolysaccharide on pup retrieving and nest building in lactating carotenoids and immune activation on sexual signals an antioxidant activity. Am Nat 164:651–659 Alonso-Alvarez C, Bertrand S, Devevey G, Gaillard M, Prost J, Faivre B, Sorci G (2004) An experimental test of the dose-dependent effect of References Pairs of morphological traits Δ filament length ‒ Δ hind-foot-web size Δ filament length ‒ Δ crest size Δ hind-foot-web size ‒ Δ crest size coefficient, CI = confidence interval) Table S2 Pearson correlations between pairs of development of morphological sexual traits (Δ trait= traitend – traitstart, r = Pearson’s correlation 58:2280 Jacot A, Scheuber H, Brinkhof MWG (2004) Costs of an induced immune response on sexual display and longevity in field crickets. Evolution intraperitoneal injection of lipopolysaccharide. Fish Physiol Biochem 37:425–432 Haukenes AH, Barton BA, Renner, KJ (2011) Plasma cortisol and hypothalamic monoamine responses in yellow perch Perca flavescens after 18:231–235 Fedorka KM, Mousseau TA (2006) Immune system activation affects male sexual signal and reproductive potential in crickets. Behav Ecol Eraud C, Jacquet A, Faivre B (2009) Survival cost of an early immune soliciting in nature. Evolution 63:1036–1043 (Phodopus sungorus). Behav Process 79:195–198 Durazzo A, Proud K, Demas GE (2008) Experimentally induced sickness decreases food intake, but not hoarding, in Siberian hamsters iguana. J Therm Biol 26:55–63 Deen CM, Hutchison VH (2001) Effect of lipopolysaccharides and acclimatation temperature on induced behavioral fever in juvenile Iguana 161:367–379 Bonneaud C, Mazuc J, Gonzalez G, Haussy C, Chastel O, Faivre B, Sorci G (2003) Assessing the cost of mounting an immune response. Am Nat 93:512–516 Bicego KC, Steiner AA, Antunes-Rodrigues J, Branco LGS (2002) Indomethacin impairs LPS-induced behavioral fever in toads. J Appl Physiol Munoz NE, Blumstein DT, Foufopoulos J (2010) Immune system activation affects song and territorial defense. Behav Ecol 21:788–793 Soc Lond B 270:2475–2480 Moret Y, Siva-Jothy MT (2003) Adaptive innate immunity? Responsive-mode prophylaxis in the mealworm beetle, Tenebrio molitor. Proc R (Anolis carolinensis). Vet Immunol Immunopathol 125:176–181 Merchant M, Fleury L, Rutherford R, Paulissen M (2008) Effects of bacterial lipopolysaccharide on thermoregulation in green anole lizards Behav Ecol Sociobiol 58:552–557 Loyau A, Saint Jalme M, Cagniant C, Sorci G (2005) Multiple sexual advertisements honestly reflect health status in peacocks (Pavo cristatus). Losdat S, Richner H, Blount JD, Helfenstein F (2011) Immune activation reduces sperm quality in the great tit. Plos One 6:e22221 96:65–69 López P, Gabirot M, Martín J 2008. Immune activation affects chemical sexual ornaments of male Iberian wall lizards. Naturwissenschaften marinus): field and laboratory studies. Physiol Biochem Zool 84:77–86 Llewellyn D, Brown GP, Thompson MB, Shine R (2011) Behavioral responses to immune-system activation in an Anuran (the Cane toad, Bufo mating to male sagebrush crickets. J Evol Biol 22:163–171 Leman JC, Weddle CB, Gershman SN, Kerr AM, Ower GD, St John JM, Vogel LA, Sakaluk SK (2009) Lovesick: immunological costs of crickets. Behav Ecol 21: 647–654 Kerr AM, Gershman SN, Sakaluk SK (2010) Experimentally induced spermatophore production and immune responses reveal a trade-off in Sherman E, Stephens A (1998) Fever and metabolic rate in the toad Bufo marinus. J Therm Biol 23:49–52 Sherman E, Baldwin L, Fernandez G, Deurell E (1991) Fever and thermal tolerance in the toad Bufo marinus. J Therm Biol 16:297–301 Physiol Behav 78:679-688 Sell KM, Crowe SF, Kent S (2003) Lipopolysaccharide induces biochemical alterations in chicks trained on the passive avoidance learning task. salmon (Salmo salar L.) parr and the effect of temperature on the kinetics of Mx responses. Fish Shellfish Immunol 17 :159–170 Salinas I, Lockhart K, Bowden TJ, Collet B, Secombes CJ, Ellis AE (2004) An assessment of immunostimulants as Mx inducers in Atlantic Behav Brain Research 215:102–109 Rico JLR, Ferraz DB, Ramalho-Pinto FJ, Morato S (2010) Neonatal exposure to LPS leads to heightened exploratory activity in adolescent rats. 21:367–372 Reaney LT, Knell RJ (2010) Immune activation but not male quality affects female current reproductive investment in a dung beetle. Behav Ecol mice. Eur J Pharmacol 395:15–21 Raghavendra V, Agrewala JN, Kulkarni SK (2000) Melatonin reversal of lipopolysacharides-induced thermal and behavioral hyperalgesia in morphna). J Exp Biol 209:3062–3070. Owen-Ashley NT, Wingfield JC (2006) Seasonal modulation of sickness behavior in free-living northwestern song sparrows (Melospiza melodia lipopolysaccharide in captive and free-living white-crowned sparrows (Zonotrichia leucophrys gambelii). Horm Behav 49:15–29 Owen-Ashley NT, Turner M, Hahn TP, Wingfield JC (2006) Hormonal, behavioral, and thermoregulatory responses to bacterial and analysis of expression in vivo and in vitro. Immunogenetics 51:332–338 Zou J, Bird S, Minter R, Horton J, Cunningham C, Secombes CJ (2000) Molecular cloning of the gene for interleukin-1β from Xenopus laevis hypothesis. Proc R Soc Lond B 273:1443–1448 Velando A, Drummond H, Torres R (2006) Senescent birds redouble reproductive effort when ill: confirmation of the terminal investment Uller T, Isaksson C, Olsson M (2006) Immune challenge reduces reproductive output and growth in a lizard. Functional Ecol 20:873–879 J Exp Biol 213:1709–1716 Toomey MB, Butler MW, McGraw KJ (2010) Immune-system activation depletes retinal carotenoids in house finches (Carpodacus mexicanus). ARTICLE 5 Impacts of varying environmental conditions on sexual traits and consequences for mate choice in the palmate newt. Cornuau JH, Sibeaux A, Tourat A, Schmeller DS, Loyau A. En préparation 1 Impacts of varying environmental conditions on sexual traits and 2 consequences for mate choice in the palmate newt 3 4 5 Jérémie H. Cornuaua, Adélaïde Sibeauxa, Audrey Tourata, Dirk S. Schmellera,b and Adeline 6 Loyaua,b 7 a 8 09200 Saint Girons, France Station d’Ecologie Expérimentale du CNRS à Moulis, 2-4 Route du CNRS, Moulis Village, 9 10 b 11 Permoser Strasse 15, 04138 Leipzig, Germany, both senior authors Helmholtz Center of Environmental Research – UFZ, Department of Conservation Biology, 12 13 Address correspondence to J. Cornuau. Email: [email protected]. 14 15 16 17 18 19 20 21 22 23 24 25 1 26 Keywords: multiple signals; sexual traits; condition dependence; trade-off; GEI 27 28 Primary Research Article 29 Changing environments may substantially alter population dynamics when they impact on 30 sexual traits, their expression and information content. During mate choice, females 31 commonly assess multiple, sexual traits. Theory predicts that females’ preferences for 32 multiple sexual traits have evolved because the reliability of these sexual traits varies 33 differently according to the environment. Morphological sexual traits, more static and less 34 labile may provide information about past condition but may not be reliable when the 35 environment changes. In contrast, behavioral sexual traits are more dynamic and more labile 36 and may provide information about current condition. Here we tested these predictions using 37 a full factorial design experiment of food limitation over long and short time periods in the 38 palmate newt (Lissotriton helveticus). We created stable and unstable environments. We 39 found that morphological sexual traits were little labile. Therefore, they may indicate the 40 bearer's genetic condition or long-term viability, rather than current condition. In contrast, 41 some behavioral traits (quivers and breathing) appeared dynamic, potentially signaling the 42 current male condition. However, courtship display remained fixed over the time scale of the 43 experiment and was the main determinant of female mate choice. Finally, we found some 44 plasticity in female preference for morphological traits depending on environmental stability. 45 Our study shows that environmental change can have profound impacts on the expression of 46 morphological and behavioral sexual traits and on female preferences. It further underscores 47 the importance to pay attention to behavioral cues to understand the impact of global change 48 on population viability. 49 2 50 INTRODUCTION 51 It is often predicted that past and current anthropogenic influences on environment will 52 render environmental conditions more variable and unpredictable (Meyers & Bull, 2002; 53 Kussell & Leibler, 2005; Reed et al., 2010). Directional environmental changes, such as 54 higher temperatures due to global warming may be predictable and individuals and 55 populations may adapt to such changes (e.g. Winder & Schindler, 2004; Vatka et al., 2011). 56 However, when environmental changes are irregular and unpredictable, it is much more 57 difficult to predict how individuals will deal with such variations and whether their 58 morphological and behavioral responses will be adaptive (Kussell & Leibler, 2005; 59 Ghalambor et al., 2007; Reed et al., 2010; Hof et al., 2011). It is further suggested that 60 environmental changes can affect evolutionary mechanisms such as natural or sexual selection 61 (Candolin & Heuschele, 2008; Candolin & Wong 2012). Therefore, some models predict that, 62 under certain conditions, environmental changes affect the expression of both male sexual 63 traits and female mating preference, which may lead to a fundamental breakdown in the 64 mechanism of sexual selection and therefore modify population faith (Greenfield & 65 Rodriguez, 2004; Higginson & Reader, 2009; Kokko & Heubel, 2008; Ingleby et al., 2010). 66 67 It has been recently proposed that natural variations in environmental pressures can be 68 finely reflected in male sexual traits and that female can adaptively modify their mate 69 preference so that the best mates are chosen in every environment (Chaine & Lyon, 2008; 70 Bro-Jørgensen, 2010, Cornwallis & Uller 2010). According to Bro-Jørgensen (2010), females 71 can do so using multiple sexual traits (i.e., the “fluctuating environments hypothesis”). This 72 hypothesis not only could resolve an issue puzzling biologists since decades (Møller & 73 Pomiankowski, 1993; Candolin, 2003; Bro-Jørgensen, 2010), but also shed light on how 74 adaptive female mating preferences can be maintained in changing environments (Chaine & 3 75 Lyon, 2008). The fluctuating environments hypothesis assumes that environmental variation 76 in time and/or space could have strong effects on signaler condition (i.e. male condition), 77 signal transmission or signal reliability (Mills et al., 2007; Higginson & Reader, 2009; 78 Rundus et al., 2010; Tolle & Wagner, 2011). For example, male body condition is a dynamic 79 individual characteristic that can rapidly vary according to either time and/or environment 80 (Lailvaux & Kasumovic, 2011). Bro-Jørgensen (2010) assumes that labile (or flexible) sexual 81 traits are evolutionary shaped to reliably signal male condition for the choosy female even 82 when facing environmental variability. In contrast, the less labile sexual traits could be 83 particularly misleading when individual condition changed due to environmental change, but 84 may be more informative in a stable environment than labile traits (Bro-Jørgensen, 2010). 85 Using both labile and fixed traits may therefore provide to females reliable information about 86 male quality in any environment encountered and/or at any time scale, thus enabling adaptive 87 female mate choice in heterogeneous and fluctuating environmental conditions (Bro- 88 Jørgensen, 2010). 89 90 Behavioral and morphological sexual traits are probably the best examples of a 91 difference in trait lability. This difference is linked to the physiological delay that exists 92 between the production or maintenance of the trait and the moment of the female’s evaluation 93 (Butler & McGraw 2011). For example in Himalayan warbler Phylloscopus humei feather 94 development seems reflected the period of feather growth and not current male quality 95 (Scordato et al., 2012, see also Naguib & Nemitz, 2007). Contrary to morphological sexual 96 traits, behavioral ones may respond rapidly to changes in male condition and thus are 97 generally expected to provide more accurate information about male condition at the time of 98 assessment (Loyau et al., 2005; Leman et al., 2009; Chargé et al., 2010). In the Houbara 99 bustard Chlamydotis undulata undalata, a change in male health status has a rapid effect on 4 100 sperm quality and may therefore be rapidly assessed by females via a change in the expression 101 of behavioral sexual traits (Chargé et al., 2010; Loyau & Lacroix, 2010). 102 103 Amphibians are strongly impacted by global change and are considered the most 104 threatened vertebrate taxa worldwide (Stuart et al., 2004; Alford, 2011). As global change is 105 expected to affect freshwater ecosystems at several trophic levels (Woodward et al., 2010), 106 we chose to manipulate environmental quality through food availability in male diet. Here, we 107 tested the predictions of the fluctuating environment hypothesis on the palmate newt 108 (Lissotriton helveticus), an amphibian species with strong sexual dimorphism involving both 109 morphological and behavioral sexual traits that are driven by female preference (Haerty et al., 110 2007; Cornuau et al., 2012). We hypothesized that (1) morphological sexual traits are more 111 static (less labile), provide information about past environmental conditions and unreliable 112 signal male condition in unstable/varying environments (2) behavioral sexual traits are more 113 dynamic (more labile) and provide reliable information about current condition even in 114 unstable/varying environments, (3) females use both static and labile sexual traits to obtain 115 reliable information about male condition and to adaptively choose their mate, thus female 116 mate choice strategy is not the same under stable and variable environmental conditions 117 because unpredictable environmental variation alters trait reliability. Generally, our results 118 contribute to the understanding of how global change can affect reproduction and population 119 dynamics of an amphibian species by impacting on morphological and behavioral sexual 120 traits. 121 122 123 5 124 MATERIAL AND METHODS 125 126 male palmate newts and 40 females were collected by dip netting from one pond 126 near Caumont (N43.02630 E1.07247, Ariège, France) at a very early point of the breeding 127 season (February 07th and March 01st 2011). Individuals were in their early aquatic stage with 128 male sexual traits only at the very start of their development. Newts were brought and 129 maintained in captivity at the Station d’Ecologie Expérimentale du CNRS à Moulis, France. 130 Unisex groups of 10 individuals were placed in opaque tanks (52×33.5×29.5 cm) with 10 131 liters of water, plants collected in their native environment, and a clay brick perforated with 132 holes to ensure shelter. Newts were individually marked by subcutaneous injection of Visible 133 Implant Elastomers (VIE, Northwest Marine Technology, Washington, Shaw Island, WA, 134 USA) by injecting one of 4 colors at the base of the four legs. They were kept at 18°C (±1) 135 under fluorescent tubes (ReptiSun 2.0, ZooMed) to simulate natural light with a 12 hours 136 light/ 12 hours dark cycle. 137 138 Diet treatment 139 To test the condition dependence of sexual traits and their reliability under stable versus 140 unstable environmental conditions, we used a diet treatment for males (Cotton et al., 2004). 141 We chose to perform “food” and “no food” treatments after a preliminary experiment that 142 showed no mortality in male newts due to a lack of food during a 4-week-period (Green, 143 1991, data not shown). After one week of acclimatization the 126 males were randomly 144 assigned to one of four dietary treatments (Fig. 1). Each treatment consisted of a long 145 treatment (four weeks) of either high- (H) or low-diet (L), followed by a short treatment (one 146 week) of H- or L-diet. Our experimental setup resulted in a full factorial design of long and 147 short treatments for male diet: HH, HL, LH, and LL. Therefore, HH and LL males 148 experienced a stable, yet different environment, whereas HL and LH males experienced a 6 149 varying environment. Males under H-diet were fed with 100 mg of chironom larvae per 150 individual every two days complemented with tubifex and daphnia every week, which 151 corresponds to an ad libitum diet. Males under L-diet were not fed. During the long and short 152 male diet treatments, females were fed ad libitum with living chironom larvae, tubifex worms 153 and daphnia, therefore experiencing a stable environment. 154 155 In order to understand the impact of laboratory treatment on male sexual traits on 156 longer time scales we conducted an additional experiment where male newts were place in 157 semi-natural tank (Fig. 1). After these 5 weeks of long and short treatments and of low or high 158 diet simulating stable and variable environmental conditions, we divided males into two 159 groups of 63 males each and placed each group of males with 20 females in an outdoor semi- 160 natural tank (1000 L) filled up with water and covered with a net to prevent predation. Newts 161 fed ad libitum on prey that was naturally growing in these tanks. Prey availability was likely 162 intermediate between the H and L diets. After 5 weeks in outdoor tanks, males were caught 163 and the morphological traits were measured (see below). Seventeen males were not re-caught 164 due to an unknown cause (predation by dragonfly larvae or snake, natural death). Missing 165 males were homogenously distributed among treatments (4/32 for HH, 4/31 for HL, 6/32 for 166 LH and 3/31 for LL), and these males did not differ from re-caught males in any 167 morphological trait measured before the release in outdoor tanks (all P > 0.20). We released 168 all individuals in the population of origin at the end of the experiment. 169 170 Morphological data 171 Morphological sexual traits of males were measured four times (n=126 for the 172 laboratory experiment and 109 at the end of the outdoor experiment), (i) at the start of the diet 173 treatment (t0), (ii) after the long diet treatment (t4 = 4 weeks later), (iii) immediately after 7 174 behavioral measurements at the end of the laboratory treatment (t5 = 5 weeks after t0), and (iv) 175 after 5 weeks in outdoor tanks (t10 = 10 weeks after t0). We recorded body mass (accuracy: 176 0.01 g) and SVL (Snout-Vent-Length, from the snout tip to the posterior end of the cloaca) 177 with a caliper (accuracy: 0.01mm). We further calculated a body condition index (BCI) as the 178 residual of the linear regression of the cube root of body mass on SVL (Baker, 1992). The 179 BCI reflects the lipid content in newts and has high accuracy in amphibians (Denoël et al., 180 2002; Báncilá et al., 2010). The three morphological traits of the palmate newts measured 181 included crest, hind foot web and the tail filament (Cornuau et al. 2012, Fig. S1). The crest is 182 usually small in palmate newts and we chose to use tail area as an index of crest development 183 (Green, 1991). All measures were realized by taking a photo on a millimeter paper 184 background with a digital camera, with subsequent processing in IMAGEJ v.1.28 185 (http://rsb.info.nih.gov/ij/). We averaged the size of left and right hind foot web. The crest, 186 hind foot web and filament length were highly positively correlated at t4, t5 and t10 and these 187 correlations were true for each of the four diet groups (all P < 0.001). Therefore, we 188 synthesized these traits in a single variable by performing a Principal Component Analysis 189 (PCA). The first principal component (hereafter called PC1) explained 66.65% of the variance 190 and was positively correlated to crest area (r = 0.85), hind foot web area (r = 0.85) and 191 filament length (r = 0.74). All the morphological variables measured in our study were 192 randomly distributed amongst the different treatment groups (MANOVA, Wilkinson test, start 193 of the long treatment: F6,119 = 1.68, P = 0.13, start of the short treatment: F6,119 = 0.67, P = 194 0.67). 195 196 Behavioral and mate choice data 197 We used 40 females and 120 males for behavioral tests. Each female was used three 198 times to reduce the number of females necessary for the experiment, always separated by a 8 199 delay of two days (we took a female effect into account in our model, see below). After diet 200 treatment (long treatment plus short treatment, Fig. 1), one female and one male were 201 randomly chosen and placed in an experimental tank (26×33.5×29.5 cm). The female was 202 placed in an opaque PVC-cylinder previously disposed in the center of the experimental tank. 203 The male was placed close to, but outside of, the PVC-cylinder. Experimental tanks contained 204 2 cm of large grained sand and 8 cm of de-chlorinated water. After 5 min of acclimatization, 205 the opaque cylinder was removed and individuals could interact freely during 40 min to allow 206 for observations of sexual behaviors (see also Cornuau et al. 2012). We recorded the male 207 behavior using a video camera placed above the experimental tank. The videos were analyzed 208 with the software The Observer 7.0 (Noldus Information Technology). Newts generally 209 performed several distinct sexual behaviors (for details see Table S1; see also Halliday, 1974; 210 Wells, 2007), of which we focused on display, quiver and breathing. Display activity 211 corresponds to the time spent (in seconds) in courtship (fan, whip and wave), quiver 212 corresponds to the male offering to deposit a spermatophore to the female by slowly quivering 213 his tail, while a breathing event corresponds to surfacing of newts to breathe (Table S1). We 214 also recorded the number of sperm mass transfers to females. For each of the 120 males, we 215 evaluated the access to reproduction as 0 if the male did not transfer any spermatophore to the 216 female or 1 if he transferred at least one spermatophore. Finally, we calculated the number of 217 male offers declined by females (i.e. male failures), by subtracting the number of quivers by 218 the number of spermatophores taken by the female. 219 220 Statistical Analyses 221 Data were analyzed using R version 2.13.1 (http://cran.r-project.org/). The effect of 222 diet treatments (long treatment, short treatment and their interaction) and environmental 223 stability (HH and LL against HL and LH) on the BCI trajectory and morphological sexual 9 224 traits (PC1) was assessed by repeated measure analysis of variance (ANOVA). The test was 225 done on data collected at the end of the laboratory treatment (t5) and at the end of the outdoor 226 experiment (t10). Males were maintained in groups so we nested the tank identity within the 227 diet treatment. Sexual traits showed some degree of allometric scaling with SVL, which was 228 corrected of by including SVL as a covariate in the models (Cotton et al., 2004). 229 230 We tested the effect of the diet treatment (long, short and the interaction between long 231 and short diet) on display, quiver, breath, access to reproduction and number of failures using 232 general linear mixed models (GLMMs). We conducted additional models to test the effect of 233 environment stability (HH and LL against HL and LH) on behavioral measure. In order to 234 assess the effect of morphological and behavioral sexual trait on female mate choice, we 235 conducted a model with PC1, display and quiver as independent variable and access to 236 reproduction as dependant variable. In order to assess if the used of morphological sexual trait 237 and behavioral sexual trait depend of the environment stability, we tested the interaction 238 between factor of interest and environment stability on the access to reproduction. In our 239 experiment females were used three times. To control for any effect of the female identity we 240 included female as a random effects. We detected any effect of female order (1st, 2nd, or 3rd 241 male) or female reproductive status (reproduction with no, 1 or 2 males) on display, quiver, 242 breath, access to reproduction or number of failure (all P > 0.05). However to avoid any 243 potential effect of this variable on our result we included female order and female 244 reproductive status as random effects in all the models. 245 246 On the whole, we assessed the effect of independent variable on dependant variable 247 by testing whether the removal of an independent variables or their interaction caused a 248 significant decrease in the model fit (Crawley 2007, Zuur et al. 2009). The statistical 10 249 significance was assessed by comparing a full model with reduce models not containing factor 250 of interest. The p-values corresponded to the decrease in deviance when the respective 251 variable or their interaction was removed. The statistical significance of the factor of interest 252 was based on log likelihood tests. Log likelihood tests follow a χ2 distribution and the degrees 253 of freedom always differ by one. We used maximum-likelihood to fit the models. 254 255 256 RESULTS 257 Impacts of diet treatments on body condition index 258 The long and short treatments affected BCI variation over the time course of the 259 experiment (Table 1). The BCI of the males assigned to the HH-diet remained constant during 260 the laboratory experiment t0-t5 (long plus short treatments) but decreased after the outdoor 261 experiment t5-t10 (Fig. 2a). The BCI of the males assigned to the HL-diet remained constant 262 during long treatment but decreased during the short treatment (Fig. 2a). In contrast, males 263 assigned to the LH-diet showed a strong decrease of their BCI already during the long 264 treatment but an increase during the short treatment (Fig. 2a). Finally the BCI of the males 265 assigned to the LL-diet decreased both during the long and the short treatments and increased 266 during the outdoor experiment (Fig. 2a). We measured a significant effect of both long and 267 short treatments on BCI at t5 (long treatment: F1,122 = 67.393, P < 0.001, short treatment: F1,122 268 = 56.687, P < 0.001, long×short treatments: F1,122 = 13.895, P = 0.001). In contrast both long 269 and short treatments did not influence BCI measured at t10 (long treatment: F1,105 = 1.621, P = 270 0.206, short treatment: F1,105 = 0.720, P = 0.398, long×short treatment: F1,105 = 1.184, P = 271 0.279). 272 273 Lability of morphological sexual traits 11 274 The variation of the PC1 synthesizing the morphological traits was significantly 275 affected by the diet during the long treatment but not the short one (Table 1). Generally, PC1 276 was linked to the amount of food provided to the individuals. The PC1 of HH and HL males 277 strongly increased during the long treatment, remained constant during the short treatment and 278 decreased during the outdoor experiment (Fig. 2b). PC1 of LH and LL males remained 279 constant during both long and short treatments but increased during the outdoor experiment 280 (Fig. 2b). 281 PC1 at t5 (end of the laboratory experiment), was significantly affected by the long 282 treatment but not the short treatment (long treatment: F1,121 = 101.382, P < 0.001, short 283 treatment: F1,121 = 0.963, P = 0.328, long×short treatments: F1,122 = 0.364, P = 0.547, SVL: 284 F1,121 = 68.891, P < 0.001). Globally we found a positive correlation between PC1 and BCI 285 measured at t5 (BCI: F1,124 = 27.615, P < 0.001). The relationship between PC1 and BCI 286 varied with environmental stability (F1,122 = 4.062, P = 0.046), being significant only in stable 287 environments (stable: F1,61 = 36.880, P < 0.001, varying: F1,61 = 0.612, P = 0.437). PC1 at t5 288 was positively correlated to BCI at t0, but this relationship depended on the diet group during 289 the long treatment (BCI: F1,122 = 10.252, P = 0.002, BCI×long diet: F1,122 = 6.111, P = 0.015, 290 Fig. 3). The significance was driven by the males assigned to the L-diet during the long 291 treatment (L-diet: F1,122 = 10.598, P = 0.002, H-diet: F1,122 = 0.052, P = 0.820, Fig. 3). 292 At t10 (end of the outdoor experiment) we did not find a significant effect of neither 293 long or short treatment on PC1 (F1,104 = 2.029, P = 0.157, F1,121 = 0.002, P = 0.962, 294 respectively; long×short treatments: F1,122 = 0.079, P = 0.780, SVL: F1,121 = 28.965, P < 295 0.001). However, we found a positive correlation between PC1 and BCI (F1,107 = 11.289, P = 296 0.001, Fig. 4). This relationship was not affected by the experimental outdoor tanks 297 (BCI×tanks: F1,105 = 0.034, P = 0.854), environmental stability (BCI×environment: F1,105 = 298 0.652, P = 0.421), nor treatments (BCI×long treatment: F1,105 = 0.174, P = 0.678; BCI×short 12 299 treatment: F1,105 = 0.698, P = 0.405). We also did not find a relationship between PC1 at t10 300 and BCI measured at t0, t4 or t5 (all P > 0.110). 301 302 Lability of behavioral sexual traits 303 Both short and long diet treatments had no impact on male display activity (Table 2, 304 Fig. 5a). Display activity was not related to male BCI (χ21 = 0.946, P = 0.331) and there was 305 no difference between males experiencing stable and varying environments (χ21 = 0.771, P = 306 0.380). However, we found a significant effect of female body condition on the time spend in 307 display (χ21 = 4.936, P = 0.026). There was evidence that the number of quivers was 308 significantly affected by the short diet treatment (Table 2, Fig. 5b). Males assigned to the L- 309 diet during the short treatment made more quivers compared to males assigned to the H-diet 310 (Fig. 5b). The number of quivers was largely explained by male BCI (χ21 = 4.279, P = 0.038) 311 but not by environmental stability (χ21 = 0.428, P = 0.513). Breathing appeared to be 312 negatively affected by the short treatment, with males on the L-diet breathing less frequently 313 than males on the H-diet (Table 2). The numbers of breathing events was positively 314 influenced by male BCI (χ21 = 6.234, P = 0.0125) but was not explained by environmental 315 stability (χ21 = 0.461, P = 0.497). 316 317 Impacts of diet treatments on male mating success 318 Neither the long or short diet treatment had an impact on male access to reproduction, 319 while the short treatment had an impact on the number of failures (Table 3). Males assigned to 320 the L-diet during the short treatment were less successful in the transition between quiver and 321 sperm mass transfer to females than H-diet males, experiencing more failures (Table 3). We 322 found that access to reproduction was not influenced by PC1 at t5 (χ21 = 1.258, P = 0.262). In 323 contrast display activity and the number of quivers had a strong impact on access to 13 324 reproduction (χ21 = 11.160, P = 0.001; χ21 = 14.236, P < 0.001, Fig. 6), and this result was not 325 influencing by environments stability (χ21 = 0.054, P = 0.816; χ21 = 1.526, P = 0.217). The 326 probability to access to reproduction was positively explained by the courtship activity and 327 the number of quivers (Fig. 6). Neither male BCI (χ21 = 0.021, P = 0.89) nor environmental 328 stability (χ21 = 0.723, P = 0.395) significantly affected the access to reproduction. However, 329 access to reproduction was affected by the interaction between environmental stability and 330 PC1 (χ21 = 6.72, P = 0.010, Fig. 7). Indeed morphological sexual traits did not contribute to 331 mating success in stable environments (χ21 = 1.234, P = 0.267, Fig. 7), while females 332 preferred males with less developed morphological sexual traits in varying environments (χ21 333 = 4.170, P = 0.030, Fig. 7). 334 335 14 336 337 338 DISCUSSION 339 and behavioral sexual traits in the palmate newt under stable and varying environmental 340 conditions. We tested the predictions of the fluctuating environment hypothesis and found that 341 both morphological and behavioral sexual traits expressed by males varied in their lability. 342 Our study suggests that female mate choice may be driven by several sexual traits that varied 343 in their lability and their information-content: a fixed trait (display activity) unrelated to 344 environmental conditions, a very labile trait (quiver frequency) that signals male current 345 condition, and traits of a higher variation in the degree of lability (morphological sexual traits) 346 that would reliably signal male current condition in stable environments and past condition in 347 an unstable environment. Display activity may inform females about male intrinsic quality 348 (genetic or long-term quality) whatever the environment experienced by both sexes, quiver 349 frequency may signal male reproductive motivation and current male condition, 350 morphological traits may inform about male mid-term condition and are only reliable in rather 351 stable environments. Our results therefore support the fluctuating environment hypothesis 352 (Bro-Jørgensen 2010) as during mate choice females of the palmate newt have access to 353 multiple signals that varied in their lability. Our study showed that the signal reliability of 354 morphological sexual traits could be compromised when environmental fluctuations are rapid. Here, we investigated the information content and variation of multiple morphological 355 356 As expected, we found that the morphological sexual traits (filament length, hind foot 357 web area and crest size) showed a low degree of lability, because they were affected by the 358 food restrictions only during the long treatment (4 weeks). Further, males in better condition 359 at the beginning of the experiment (t0, before the long treatment) were able to develop 360 morphological traits of higher quality (e.g. longer filaments), suggesting that fat reserves 361 maintained over the terrestrial winter season could be crucial during the breeding season. 15 362 These findings confirm that the production of morphological sexual traits, on a yearly basis, is 363 an expensive physiological process linked to male body condition. However, we found that 364 the relationship between the expression of male morphological sexual traits and body 365 condition varied with environmental stability. It means that changing environmental 366 conditions may weaken the reliability of morphological sexual traits (Robinson et al., 2007; 367 2012). Indeed, in a rapidly changing environment, morphological sexual traits may accurately 368 reflect past male condition but are poor predictors of current condition. In that case, females 369 who base their choice only on morphological sexual traits would not obtain reliable 370 information on male condition at the exact time of assessment and may perform maladapted 371 mate choice (Bro-Jørgensen, 2010). 372 373 We expected behavioral sexual traits to be more labile than morphological sexual 374 traits. However, not all behavioral sexual traits were impacted the same way by varying 375 environmental conditions. Firstly, contrary to our expectations, courtship display activity 376 (including fan, wave and whip) remained fixed over the time scale of the experiment, despite 377 varying environmental conditions. This suggests that display activity is not condition- 378 dependent, which is supported by a previous study in the Plethodontid salamander 379 Desmognathus ochrophaeus showing that metabolic costs of display were low (Bennett & 380 Houck, 1983). While display activity is fixed, its assessment during mate choice cannot 381 provide reliable information on current male condition. However, if display activity is 382 genetically determined (Drayton et al., 2010) and female preference for such a trait is 383 heritable, mating with high displaying males will pay-off in any environment the offspring 384 will experience (Fisher 1930). Therefore, the use of a fixed, unreliable trait might be a good 385 mate choice strategy in a varying environment. 386 16 387 Secondly, we found that some male behavioral traits (the number of quivers, the 388 number of breath events and the number of male failures in the transition between quiver and 389 spermatophore transfer) were more labile than display activity. These behaviors were labile 390 over a short term. Surprisingly, the males who experienced the low diet during the short 391 treatment made more quivers than the males who experienced the high diet, but were less 392 successful in the transition between quiver and sperm mass transfer (=access to reproduction) 393 to females. As a quiver represents a male’s offer to deposit a spermatophore to the female, a 394 food shortage may increase male motivation to reproduce, maybe as a terminal effort (Sadd et 395 al., 2006; Saaristo et al., 2010). However, in our experiment, females appear to be able to 396 detect such a motivation at least to some extent, as the higher motivation of L-diet males did 397 not result in a higher mating success compared to H-diet males. 398 399 Plasticity in female preferences during mate choice can involve the direction of the 400 preference for a given trait. In this case, females always evaluate the same trait(s) but prefer 401 males with the more developed traits in some environments and males with the less developed 402 traits in others (Chaine & Lyon, 2008; Bro-Jørgensen, 2010). Plasticity in female preference 403 can also affect the strength of the preference, so that females assess different traits in different 404 environments, or prioritize some traits over others depending on the circumstances (Cotton et 405 al., 2006; Chaine & Lyon, 2008). Here, we found some plasticity in female preference for 406 male morphological sexual traits, depending on the circumstances. Female mate choice was 407 not driven by male morphological sexual traits in stable environments, while there was a 408 negative association between female choice and morphological sexual traits in varying 409 environments. In addition, in a previous study in the palmate newt, we found that females 410 prefer males with more developed morphological sexual traits (Cornuau et al., 2012; see also 411 Haerty et al., 2007). Hence, when the environmental variations cannot be reliably tracked by 17 412 labile sexual traits, females may prefer using a fixed trait and may only use labile traits as 413 additional source of information. Our results further suggest that the degree of lability might 414 contain important information for female mate choice. In our experiment, medium labile traits 415 (morphological traits) and very labile traits (behavioral trait quivers) were used differently in 416 different environments and determined reproductive success of males. 417 418 In conclusion, we were able to show that the access to reproduction of males vary with 419 fluctuating environmental conditions. That was especially the case when environment 420 changed in quick alterations. Our study is one of few showing the impact of varying 421 environments on reproduction experimentally. The interactions environment x reproductive 422 strategy appear to be complex, but need to be better understood for a wide range of species to 423 allow predicting population size and dynamics for conservation-relevant species. Such 424 knowledge is key in situations where population-level estimates of reproductive outcome and 425 survival are not enough to correctly predict the population dynamics of the focal species. 426 Pinpointing those situations where information on individual behaviors is needed will help to 427 direct efforts towards gathering relevant information for the management of focal species. Our 428 study has shown that when stress was unpredictable, it was much more difficult to predict 429 how individuals will deal with such variations and whether their morphological and 430 behavioral responses will be adaptive (Kussell & Leibler, 2005; Ghalambor et al., 2007; Reed 431 et al., 2010; Hof et al., 2011). We have also shown that a varying environment changed the 432 link between different traits, thus changing the information reliability, which could lead to 433 maladaptive behavior and increased stress. Therefore, it is important to further analyze the 434 variability of the investment in the expression of costly sexual signals that convey honest 435 information about their quality and that will be of interest to individuals of the other sex 436 seeking prospective mates in stable and variable environments. 18 437 Ethical standards The experiments comply with the current laws of the country in which 438 they were performed. 439 440 REFERENCES 441 Alford RA (2011) Bleak future for amphibians. Nature, 480, 461–462. 442 Baker JMR (1992) Body condition and tail height in great crested newts, Triturus cristatus. 443 Animal Behaviour, 43, 157–159. 444 Báncilá RI, Hartel T, Plaiasu R, Smets J, Cogalniceanu D (2010) Comparing three body 445 condition indices in amphibians: a case study of yellow-bellied toad Bombina variegate. 446 Amphibia Reptilia, 31, 558–562. 447 448 449 450 Bennett AF, Houck LD (1983) The energetic cost of courtship and aggression in a plethodontid salamander. Ecology, 64, 979–983. Bro–Jørgensen J (2010) Dynamics of multiple signaling systems: animal communication in a world in flux. Trends in Ecology and Evolution, 25, 292–300. 451 Butler MW, McGraw KJ. (2011) Past or present? Relative contributions of developmental and 452 adult condition to adult immune function and coloration in mallard ducks (Anas 453 platyrhynchos). Journal of Comparative Physiology B, 181, 551–563. 454 Candolin U (2003) The use of multiple cues in mate choice. Biological Reviews, 78, 575–595. 455 Candolin U, Heuschele J (2008) Is sexual selection beneficial during adaptation to 456 environmental change? Trends in Ecology and Evolution, 23, 446–452. 457 Candolin U, Wong BBM. 2012. Sexual selection in changing environments: consequences for 458 individuals and populations. pp201-215. In Behavioural responses to a changing world 459 mechanisms and consequences. Edited by Candolin U, Wong BBM; Oxford University 460 Press, Oxford. 19 461 462 Chaine AS, Lyon BE (2008) Adaptative plasticity in female mate choice dampens sexual selection on male ornaments in the lark bunting. Science, 319, 459–426. 463 Chargé R, Saint Jalme M, Lacroix F, Cadet A, Sorci G (2010) Male health status, signalled by 464 courtship display, reveals ejaculate condition and hatching success in a lekking species. 465 Journal of Animal Ecology, 79, 843–850. 466 467 468 469 Cornuau JH, Rat M, Schmeller DS, Loyau A (2012) Multiple signals in the palmate newt: ornaments help when courting. Behavioral Ecology and Sociobiology, 66, 1045-1055. Cornwallis CK, Uller T (2010) Towards an evolutionary ecology of sexual traits. Trends in Ecology and Evolution, 25, 145-152. 470 Cotton S, Fowler K, Pomiankowski A (2004) Do sexual ornaments demonstrate heightened 471 condition-dependent expression as predict by the handicap hypothesis? Proceedings of 472 the Royal Society of London, Series B, 271, 771–783. 473 474 Cotton S, Small J, Pomiankowski A (2006) Sexual selection and condition-dependent mate preferences. Current Biology, 16, 755–765. 475 Crawley MJ (2007) The R book. Chichester: John Wiley & Sons Ltd. 476 Denoël M, Hervant F, Schabetsberger R, Joly P (2002) Short- and long-term advantages of an 477 alternative ontogenetic pathway. Biological Journal of the Linnean Society, 77, 105– 478 112. 479 Drayton JM, Milner RNC, Junt J, Jennions MD (2010) Inbreeding and advertisement calling 480 in the cricket Telegryllus commodus: laboratory and field experiments. Evolution, 64, 481 3069–3083. 482 Fisher RA. 1930. The Genetical Theory of Natural Selection. (Clarendon Press, Oxford). 483 Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non-adaptive 484 phenotypic plasticity and the potential for contemporary adaptation in new 485 environments. Functional Ecology, 21, 394–407. 20 486 Green AJ (1991) Large male crests, an honest indicator of condition, are preferred by female 487 smooth newts, Triturus vulgaris (Salamandridae) at the spermatophore transfer stage. 488 Animal Behaviour, 41, 367–369. 489 490 Greenfield MD, Rodriguez RL (2004) Genotype-environment interaction ans the reliability of mating signals. Animal Behaviour, 68, 1461–1468. 491 Haerty W, Gentilhomme E, Secondi J (2007) Female preference for a male sexual trait 492 uncorrelated with male body size in the palmate newt (Triturus helveticus). Behaviour, 493 144, 797–814. 494 495 Halliday TR. 1974. Sexual behaviour of the smooth newt, Triturus vulgaris (Urodela, Salamandridae). Journal of Herpetology. 8, 277–292. 496 Higginson AD, Reader T (2009) Environmental heterogeneity, genotype-by-environment 497 interactions and the reliability of sexual traits as indicators of mate condition. 498 Proceedings of the Royal Society of London, Series B, 276, 1153–1159. 499 500 501 502 503 504 505 506 Hof C, Levinsky I, Araújo MB, Rahbek C (2011) Rethinking species’ ability to cope with rapid climate change. Global Change Biology, 17, 2987–2990. Ingleby FC, Hunt J, Hosken DJ (2010) The role of genotype-by-environment interactions in sexual selection. Journal of Evolutionary Biology, 23, 2031–2045. Kokko H, Heubel K (2008) Condition-dependance, genotype-by-environment interactions and the lek paradox. Genetica, 134, 55–62. Kussell E, Leibler S (2005) Phenotypic diversity, population growth, and information in fluctuating environments. Science, 309, 2075–2078. 507 Lailvaux SP, Kasumovic MM (2011) Defining individual condition over lifetimes and 508 selective contexts. Proceedings of the Royal Society of London, Series B, 278, 321–328. 21 509 Leman JC, Weddle CB, Gershman SN, Kerr AM, Ower GD, St John JM, Vogel LA, Sakaluk 510 SK (2009) Lovesick: immunological costs of mating to male sagebrush crickets. 511 Journal of Evolutionary Biology, 22, 163–171. 512 Loyau A, Lacroix F (2010) Watching sexy displays improves hatching success and offspring 513 growth through maternal allocation. Proceedings of the Royal Society of London, Series 514 B, 277, 3453–3460. 515 Loyau A, Saint Jalme M, Cagniant C, Sorci G (2005) Multiple sexual advertisements honestly 516 reflect health status in peacocks (Pavo cristatus). Behavioral Ecology and Sociobiology, 517 58, 552–557. 518 519 Meyers LA, Bull JJ (2002) Fighting change with change: adaptive variation in an uncertain world. Trends in Ecology and Evolution, 17, 551–557. 520 Mills SC, Alatalo RV, Koskela E, Mappes J, Mappes T, Oksanen TA (2007) Signal reliability 521 compromised by genotype-by-environment interaction and potential mechanisms for its 522 preservation. Evolution, 61, 1748–1757. 523 524 Møller AP, Pomiankowski A. (1993) Why have birds got multiple sexual ornaments? Behavioral Ecology and Sociobiology, 32, 167–176. 525 Naguib M, Nemitz A (2007) Living with the past: nutritional stress in juvenile males has 526 immediate effects on their plumage ornaments and on adult attractiveness in zebra 527 finches. Plos One, 2, e902. 528 Reed TE, Waples RS, Schindler DE, Hard JJ, Kinnison MT (2010) Phenotypic plasticity and 529 population viability: the importance of environmental predictability. Proceedings of the 530 Royal Society of London, Series B, 277, 3391–3400. 531 Robinson MR, Pilkington JG, Clutton-Brock TH, Pemberton JM, Kruuk LEB (2007) 532 Environmental heterogeneity generates fluctuating selection on a secondary sexual trait. 533 Current Biology, 18, 751–757. 22 534 535 Robinson MR, van Doorn GS, Gustafsson L, Qvarnström A (2012) Environment-dependent selection on mate choice in a natural population of birds. Ecology Letters, 15, 611–618. 536 Rundus AS, Sullivan-Beckers L, Wilgers DJ, Hebets EA (2010) Females are choosier in the 537 dark: environment-dependent reliance on courtship components and its impact on 538 fitness. Evolution, 65, 268–282. 539 Saaristo M, Craft JA, Lehtonen KK, Lindström K (2010) An endocrine disrupting chemical 540 changes courtship and parental care in the sand gody. Aquatic Toxicology, 97, 285–292. 541 Sadd B, Holman L, Armitage H, Lock F, Marland R, Siva-Jothy MT (2006) Modulation of 542 sexual signaling by immune challenged male mealworm beetles (Tenebrio molitor, L.): 543 evidence for terminal investment and dishonestly. Journal of Evolutionary Biology, 19, 544 321–325. 545 546 Scordato ESC, Bontrager AL, Price TD. (2012) Cross-generational effects of climate change on expression of a sexually selected trait. Current Biology, 22, 78–82. 547 Stuart SN, Chanson JS, Cox NA, Young BE, Rodrigues ASL, Fischman DL, Waller RW 548 (2004) Status and trends of amphibian declines and extinctions worldwide. Science, 549 306, 1783-1786. 550 Tolle AE, Wagner WE (2011) Costly signals in a field cricket can indicate high or low 551 condition direct benefits depending upon the environment. Evolution, 65, 283–294. 552 Vatka E, Orell M, Rytkönen S (2011) Warming climate advances breeding and improves 553 synchrony of food demand and food availability in a boreal passerine. Global Change 554 Biology, 17, 3002–3009. 555 Wells KD (2007) The ecology and behavior of amphibians. The University of Chicago Press. 556 Winder M, Schindler DE (2004) Climatic effects on the phenology of lake processes. Global 557 Change Biology, 10, 1844–1856. 23 558 Woodward G, Perkins DM, Brown LE (2010) Climate change and freshwater ecosystems: 559 impacts across multiple levels of organization. Philosophical transactions of the royal 560 society B-biological sciences, 365, 2093–2106. 561 562 Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM. (2009) Mixed effects models and extensions in ecology with R. Springer Science 563 564 24 565 TABLES 566 Table 1. Results of the repeated measures analysis of variance on variation of BCI and 567 morphological sexual traits PC1 (filament length, hind foot web and crest). Significant 568 interactions are highlighted in bold. 569 BCI F P PC1 F P Between subjects Long treatment Short treatment Long×Short treatments SVL 48.732 2.644 3.339 <0.001 37.526 <0.001 0.107 0.015 0.903 0.071 0.705 0.44 71.236 <0.001 Within subjects Time Time×Long treatment Time×Short treatment Time×Short×Long treatments SVL 103.632 50.596 24.775 4.446 <0.001 <0.001 <0.001 0.004 570 571 572 25 43.371 21.470 0.314 0.159 81.098 <0.001 <0.001 0.815 0.924 <0.001 573 Table 2. Summary results of the effect of diet treatment on on behavioral measurements 574 (Display: time spent in courtship, Quiver: number of quivers, Breath: number of breathing 575 events). Bold values represent significant parameters. 576 Dependant variable Display Independent variable Long treatment Short treatment Long×Short treatments χ2 0.178 0.348 0.816 Df 1 1 1 P 0.673 0.555 0.366 Quiver Long treatment Short treatment Long×Short treatments 0.043 8.518 0.257 1 1 1 0.836 0.003 0.612 Breath Long treatment Short treatment Long×Short treatments 2.926 10.194 0.673 1 1 1 0.087 0.001 0.412 577 578 579 26 580 Table 3. Summary results of the effect of diet treatment on reproductive success 581 (Morphological sexual traits = index of morphological sexual traits, Failures = number of 582 quivers minus number of successful spermatophore transfers). Bold values represent 583 significant parameters. 584 Dependant variable Access to Reproduction Independent variable Long treatment Short treatment Long×Short treatments χ2 0.008 0.211 0.725 Df 1 1 1 P 0.926 0.650 0.395 Failures Long treatment Short treatment Long×Short treatments 0.016 7.681 0.787 1 1 1 0.898 0.005 0.375 27 585 FIG. LEGENDS 586 Fig. 1. Experimental design composes by one laboratory experiment and one outdoor 587 experiment. During laboratory experiment males newt was assigned one of four dietary 588 treatments. Each treatment consisted of a long treatment (four weeks) of either high- (H) or 589 low-diet (L), followed by a short treatment (one week) of H- or L-diet. Our experimental 590 setup resulted in a full factorial design of long and short treatments for male diet: HH, HL, 591 LH, and LL. Morphological measure was carry out at t0, t4, t5 and t10 (M = morphological 592 measure). Behavioral measure was carry out at t5 (B = behavioural measure). During outdoor 593 experiment all males was placed in semi natural outdoor tanks during five weeks before 594 release in their initial pond. 595 596 Fig. 2 Variation in (a) BCI (Body Condition Index) and (b) PC1 (index of morphological 597 sexual traits including crest, hind foot web and filament length) over the experiment 598 according to the diet treatment (mean ± SE). H-diet during long and short treatments (HH), 599 black bars; H-diet during long treatment and L-diet during short treatment (HL), dark-gray 600 bars; L-diet during long treatment and H-diet during short treatment (LH), light-gray bars; L- 601 diet during long and short treatments (LL), white bars. The laboratory experiment was 602 performed between t0 and t5. The outdoor experiment was performed between t5 and t10. 603 604 Fig. 3 Relationship between PC1 at t5 (index of morphological sexual traits) and BCI at t0 605 (Body Condition Index). Males assigned to high diet during long treatment: open circle and 606 dashed line. Males assigned to low diet during long treatment: closed circles and continuous 607 line). 608 28 609 Fig. 4 Relationship between PC1 (index of morphological sexual traits) and BCI (Body 610 Condition Index) at t10. 611 612 Fig. 5 Effect of the diet treatment on (a) the time spent displaying (mean ± SE) and (b) the 613 number of quiver H-diet, black bars; L-diet, white bars. 614 615 Fig. 6 Variation of (a) the time spent displaying (mean ± SE) and (b) the number of quiver 616 according to the access or not to reproduction. 617 618 Fig. 7 Variation of the PC1 (index of morphological sexual traits) according to the access or 619 not to reproduction and the environmental stability. Access to reproduction, black bars; no 620 access to reproduction, white bars. 621 622 623 624 625 626 627 628 29 Laboratory experiment M t0 Outdoor experiment M Long diet four weeks t4 B Short diet one week M t5 M five weeks t10 High diet Sampling Low diet natural diet High diet Release High diet Low diet Low diet 629 630 Fig. 1. Experimental design composes by one laboratory experiment and one outdoor 631 experiment. During laboratory experiment males newt was assigned one of four dietary 632 treatments. Each treatment consisted of a long treatment (four weeks) of either high- (H) or 633 low-diet (L), followed by a short treatment (one week) of H- or L-diet. Our experimental 634 setup resulted in a full factorial design of long and short treatments for male diet: HH, HL, 635 LH, and LL. Morphological measure was carry out at t0, t4, t5 and t10 (M = morphological 636 measure). Behavioral measure was carry out at t5 (B = behavioural measure). During outdoor 637 experiment all males was placed in semi natural outdoor tanks during five weeks before 638 release in their initial pond. 639 640 641 642 643 644 645 646 30 0.06 + (a) Body Condition Index 0.04 0.02 0 -0.02 -0.04 - -0.06 -0.08 647 + (b) t0 t4 t5 t 10 t4 t5 t 10 1.6 1.2 PC1 0.8 0.4 0 -0.4 - -0.8 -1.2 648 t0 649 Fig. 2 Variation in (a) BCI (Body Condition Index) and (b) PC1 (index of morphological 650 sexual traits including crest, hind foot web and filament length) over the experiment 651 according to the diet treatment (mean ± SE). H-diet during long and short treatments (HH), 652 black bars; H-diet during long treatment and L-diet during short treatment (HL), dark-gray 653 bars; L-diet during long treatment and H-diet during short treatment (LH), light-gray bars; L- 654 diet during long and short treatments (LL), white bars. The laboratory experiment was 655 performed between t0 and t5. The outdoor experiment was performed between t5 and t10. 656 31 + 5.5 4.5 3.5 PC1 at t5 2.5 1.5 0.5 -0.5 -1.5 -2.5 - -3.5 -4.5 -0.04 -0.02 - 0 0.02 0.04 BCI at t0 0.06 0.08 + 657 658 Fig. 3 Relationship between PC1 at t5 (index of morphological sexual traits) and BCI at t0 659 (Body Condition Index). Males assigned to high diet during long treatment: open circle and 660 dashed line. Males assigned to low diet during long treatment: closed circles and continuous 661 line). 662 663 32 + 4 3 2 PC1 at t10 1 0 -1 -2 - -3 -4 -5 -0.15 -0.1 - -0.05 0 0.05 0.1 0.15 + BCI at t10 664 665 Fig. 4 Relationship between PC1 (index of morphological sexual traits) and BCI (Body 666 Condition Index) at t10. 667 668 33 Time spent displaying (s) (a) 800 NS NS N=60 N=60 750 N=60 N=60 700 650 669 (b) short diet Short treatment NS ** 13 12 Number of quivers long diet Long treatment 10 N=60 N=60 N=60 N=60 8 6 4 2 0 670 Long treatment Short treatment 671 Fig. 5 Effect of the diet treatment on (a) the time spent displaying (mean ± SE) and (b) the 672 number of quiver H-diet, black bars; L-diet, white bars. 673 674 675 34 Time spent displaying (s) (a) 1000 900 *** N=67 800 700 600 N=53 500 Access access 676 (b) 13 No access no access *** Number of quivers 12 10 N=53 N=67 8 6 4 2 0 Access No access 677 678 Fig. 6 Variation of (a) the time spent displaying (mean ± SE) and (b) the number of quiver 679 according to the access or not to reproduction (boxplot). 680 681 682 683 35 1.2 + * 0.8 PC1 at t5 0.4 NS 0 -0.4 - -0.8 -1.2 684 Stable environment Varying environment 685 Fig. 7 Variation of the PC1 (index of morphological sexual traits) according to the access or 686 not to reproduction and the environmental stability. Access to reproduction, black bars; no 687 access to reproduction, white bars. 688 689 690 691 692 693 694 695 696 697 698 699 700 701 36 702 SUPPLEMENTARY MATERIAL 703 Table S1. Short description of palmate newt behavior during breeding season. step of behavior orientation static and retreat display (here name display) behavior sniff follow fan wave whip spermatophore transfer quiver touch tail deposition taking spermatophore breath breath description male smells the female cloaca male follows the female male folds its tail against its flank and the distal portion of the tail vibrates rapidly male held its tail to be presented with a full view for female. male brought back its tail against its flank with a extremely quickly movement after passing the female, male quivers its tail. This behavior represent a spermatophore transfer request that the female can accept or refuse according if it touch or not its tail (name touch tail) female touch the tail of male signaling that it accepts spermatophore transfer male deposits one spermatophore on the floor female following male's tail, moves on spermatophore which will stick on its cloaca male goes up to the surface to breathe. This behavior can occurs during each steps describe above 704 Bold characters represent behavior used in our study. Fan, wave and whip were synthesized 705 and name display in this paper. After Halliday 1974; Halliday 1975; Denoël 1999; Wells 706 2007. 707 708 37 1 2 3 4 709 710 Figure S1. Pictures of morphological sexual traits during breeding season with natural 711 variation. 1-Male newt; 2-Crest; 3-Hind fit web; 4-Filament. All pictures were taken by JHC. 38 712 REFERENCES 713 Denoël M. 1999. Le comportement social des urodèles. Cahiers d’Ethologie. 19, 221–258. 714 Halliday TR. 1974. Sexual behaviour of the smooth newt, Triturus vulgaris (Urodela, 715 716 717 718 719 Salamandridae). Journal of Herpetology. 8, 277–292. Halliday TR. 1975. An observational and experimental study of sexual behaviour in the smooth newt, Triturus vulgaris (Amphibia: Salamandridae). Anim Behav. 23, 291–322. Wells KD. 2007. The ecology and behavior of amphibians. Chicago and London. The University of Chicago Press. 720 39 SUMMARY: Multiple signals and mate choice in the palmate newts. The aim of this PhD is the study of multiples signals in mate choice in the palmate newts Lissotriton helveticus. In a first part, we have made a review about multiples signals theory (article 1). In a second part we have study the effect of ornaments on female mate choice (article 2 et 3). In a third part we have study the trade-off between ornament and male’s quality (article 4 et 5). Finally we have investigated the effect of environment heterogeneity on multiples signals and male success (article 5). Globally our result shows that access to reproduction is explain by multiples signals that need to be study in a dynamic framework. Our results have important implication about sexual selection, population dynamic, global change and emergent diseases. AUTEUR : Jérémie CORNUAU TITRE : Signaux multiples et choix du partenaire chez le triton palmé Lissotriton helveticus DIRECTEUR DE THÈSE : Adeline Loyau, Dirk Schmeller LIEU ET DATE DE SOUNENANCE : Station d’Écologie Expérimentale du CNRS à Moulis, 04-12-12 RÉSUMÉ : L’objectif général de cette thèse est l’étude des signaux multiples dans le cadre du choix du partenaire sexuel chez le triton palmé Lissotriton helveticus. Premièrement nous avons réalisé une synthèse sur les signaux multiples dans le cadre du choix du partenaire (article 1). Deuxièmement nous avons étudié l’importance des ornements du mâle sur le choix des femelles (article 2 et 3). Troisièmement nous avons étudié le rôle des signaux multiples en tant que signaux de qualité individuelle (article 4 et 5). Quatrièmement nous avons étudié l’effet de l’hétérogénéité de l’environnement sur l’honnêteté des signaux multiples et l’accès à la reproduction des mâles (article 5). Ces résultats ont des applications aussi bien théoriques que pratiques sur la compréhension de la sélection sexuelle, la dynamique des populations, l’effet des changements globaux et l’impact des maladies émergentes. MOTS-CLÉS : Signaux multiples, Sélection sexuelle, Choix du partenaire, Condition dépendance, Changements globaux, Triton palmé, Lissotriton helveticus DISCIPLINE ADMINISTRATIVE : Sciences et Vie de la Terre, Spécialité Écologie Comportementale Station d’Écologie Expérimentale du CNRS, 2,4 Route du CNRS, 09200, Moulis, France