LYSANDRE LANDRY ROLE DES FACTEURS ÉCOLOGIQUES DANS L'ÉVOLUTION PARALLÈLE DU GRAND COREGONE (COREGONUS CLUPEAFORMIS MiTCHILL.) Mémoire présenté à la Faculté des études supérieures de l'Université Laval dans le cadre du programme de maîtrise en biologie pour l'obtention du grade de Maître es Sciences (M. Se.) DÉPARTEMENT DE BIOLOGIE FACULTÉ DES SCIENCES ET DE GÉNIE UNIVERSITÉ LAVAL QUÉBEC JUIN 2006 Lysandre Landry, 2006 RÉSUMÉ La sélection naturelle est un des facteurs dominants lors de l'évolution répétée et parallèle observée notamment chez les poissons des régions tempérées du Nord. Ainsi, au sein des populations de Coregonus clupeaformis, l'écotype nain s'est différencié plusieurs fois à partir d'un écotype normal, notamment dans six lacs du bassin versant de la rivière St-Jean (Maine, USA et Québec, CA). Cette étude porte sur les caractéristiques de l'habitat ainsi que sur la disponibilité des ressources et leurs variations temporelles influençant le niveau de différenciation atteint entre les écotypes de ces populations. Les résultats appuient les hypothèses théoriques faisant intervenir le rôle des fluctuations environnementales (biomasse et structure des ressources) dans la variation des pressions sélectives et le rôle de la compétition pour les ressources (disponibilités réduites) dans l'établissement et le maintien d'un degré de différenciation. Le rôle de d'autres interactions trophiques comme les relations prédateurs-proies semble aussi impliqué dans ce mécanisme de divergence. ABSTRACT Natural sélection is one of the dominant factors in repeated parallel évolution observed notably in north temperate fish populations. In Coregonus clupeaformis populations located in six lakes of the St-John's.river watershed (Maine, USA and Québec, CA), a dwarf ecotype has differentiated from a normal ecotype many times. Hère we study habitat and resource features, it temporal variations in relation to the observed level of differentiation between écotypes of thèse populations. The results support the hypothèses that environmental fluctuation (zooplankton community structure and biomass) cause variation in sélective pressure and that the resource compétition (availability of habitat and resources) influence the establishment and persistence of a degree of differentiation. Trophic interactions such as predator-prey relationships seem also to be involved in giving rise to this divergence. 11 AVANT-PROPOS De l'exaltation de mes premiers pas dans le laboratoire du docteur Louis Bernatchez, jusqu'à la fin de ce très long et cahoteux chemin, j'ai rencontré une multitude de personnes merveilleuses et généreuses avec qui j'ai grandi et sans qui je n'y serais jamais arrivée. Je veux d'abord remercier Louis Bernatchez pour sa patience et sa confiance en moi ainsi que Sean M. Rogers et David Campbell; les premiers qui m'ont fait vibrer par leur passion pour la biologie évolutive. Ensuite et non moindrement, merci aux personnes qui ont fait et qui font encore partie de l'équipe du projet corégone, sans oublier Vicky Albert sans qui je n'aurais jamais fait partie de ce projet. Vive la plus belle «gang» évolutive au monde!!! Je veux remercier aussi à mon co-directeur Warwick F. Vincent pour son soutien et ses encouragements ainsi que les membres de son laboratoire : Sébastien Roy, Leira Rétamai, Marie-Josée Martineau et Milla Rautio pour l'aide en limnologie. Toute ma gratitude à Gesche Winkler qui m'a soutenue, rassurée et aidée dans le choix des meilleurs protocoles pour traiter les données d'invertébrés lacustres, pour les discussions écologiques et plus encore. Merci à Julian J. Dodson et aux gens de son laboratoire qui ont toujours été là quand j'en avais besoin. Merci à Robert St-Laurent sans qui les gens qui m'ont aidé et accueilli sur le terrain n'aurait jamais été aussi heureux, mais aussi pour son aide avec le matériel de terrain et avec l'échantillonnage, et pour les discussions passionnées qui m'ont permis de continuer. Merci à la contribution généreuse des biologistes du «Maine Department of Inland Fisheries and Wildlife»; M. David Basley et ses heureux acolytes Derrick Cote et Frank O. Frost, qui m'ont permis d'obtenir des invertébrés à observer et des échantillons à analyser. Je veux remercier aussi les aubergistes du Lac de l'Est pour nous avoir accueilli et permis d'utiliser leur espace de vie pour les observations et analyses sur le terrain. Merci à mes joyeux compagnons de terrain, de tri et d'identification : Indi Noël, Robin Martel, Julie Breton, Justin Roy, Guillaume Côté et Tanguy Castric qui ont survécu à mes angoisses benthiques et bien sûr, à Lucie Papillon pour sa compréhension et son sens de l'organisation. 111 Merci aux «cracs» des statistiques, des mathématiques et de la réflexion logique: Pierre Duchesne, Julie Béliveau, Sophie Baillargeon, Scott McCairns, Mathieu Cusson, Milla Rautio et David Véliz. Je veux remercier particulièrement Guillaume Côté, Mélissa Lieutenant-Gosselin, Scott McCairns, Julien April et Anne-Marie Gale ainsi que ma petite soeur Mylène et mon père Jacques Landry pour leur patience et leur contribution généreuse à l'amélioration de ce mémoire. Finalement, je veux remercier encore Sylvain Gutjahr, Suzy Tremblay, Guillaume Côté, Mélissa Lieutenant-Gosselin, Scott McCairns, Marie-France Barrette, Véronique Thériault, Pierre-Philippe Dupont, Vincent Bourret, Vanessa Viera, Sean Rogers, David Campbell, David Véliz, Catherine Lippe, Dylan J. Fraser, Mélanie Dionne, Dany Garant, Mathieu Cusson, Catherine Potvin, Marc Ringuette et Frédéric Lecomte pour les discussions scientifiques et personnelles passionnantes, pour leurs conseils et leur confiance en moi. Je veux remercier ma mère Paule Croteau, mes sœurs Mylène et Marie-Christine, mon père Jacques Landry et Kerstin Bellmann pour le soutien et l'affection. L'article qui compose le cœur de ce mémoire a été soumit au journal Evolution. Le travail de laboratoire et l'analyse des données ont été réalisés par L. Landry sous la supervision du directeur, L. Bernatchez et du co-directeur, W. F. Vincent. Une partie du travail de laboratoire a été fait avec l'aide de R. Martel. La rédaction a été effectuée avec l'aide et les critiques de L. Bernatchez et de W. F. Vincent. De plus, j'ai également participé à une étude sur la spécialisation alimentaire des écotypes nains et normaux du Grand Corégone à l'aide d'isotopes stables. Le travail de laboratoire et l'échantillonnage ont été réalisés par N. Pelletier et l'analyse des données par L. Landry. La rédaction de ces travaux sera effectuée par L. Landry sous la supervision du directeur L. Bernatchez. J'ai aussi effectué des travaux sur la faune benthique dont le travail au laboratoire a été fait avec l'aide de Julie Breton et de Indi Noël. Je ferai l'analyse de ces données et je participerai à la rédaction d'une publication sur celles-ci avec l'aide de L. Bernatchez. Finalement, L. Landry a été soutenu financièrement par le groupe de recherches inter-universitaires, Québec-Océan. «Les choses que je sens et que j'imagine ne sont peut-être rien du tout hors de moi et en ellesmêmes». Descartes, Deuxième méditation V TABLE DES MATIÈRES RÉSUMÉ i ABSTRACT i AVANT-PROPOS n TABLE DES MATIÈRES v LISTE DES TABLEAUX vu LISTE DES FIGURES INTRODUCTION GÉNÉRALE La radiation adaptative La sélection naturelle et la divergence populationnelle Compétition L'opportunité écologique Niche Chance et contingence historique Impératifs à la divergence ou à la différenciation des populations Mécanisme potentiel de la divergence populationnelle Les facteurs stochastiques versus déterministes Les lacs et les poissons des régions tempérées du nord Relation phénotype-environnement Opportunité écologique Le grand corégone Relation phénotype-environnement Opportunité écologique Objectif principal CHAPITRE 1 THE RÔLE OF HABITAT AND PREY COMMUNITY IN THE PARALLEL EVOLUTION OF LAKE WHITEFISH DWARF ECOTYPES Résumé Abstract Introduction Materials and Methods Study Sites Limnological Variables Zooplanktonic Variables Statistical Analyses Results Limnological Différences Between Lake Groups Resource Availability and Spatio-Temporal Stability Resource Stucture and Stability VII 1 2 3 3 4 4 5 6 7 7 8 8 9 11 11 12 12 14 15 16 17 20 20 20 22 23 25 25 26 26 VI Discussion Parallelism in Abiotic Features and Levels of Ecotypic Differentiation Parallelism in Prey Availability and Ecotypic Differentiation Acknowledgments CONCLUSION GÉNÉRALE La sélection naturelle L'évolution en parallèle Prémisses à la différenciation des populations Le paysage de sélection Fluctuation des pressions de sélection Les grands corégones Objectifs et résultats Réduction des ressources disponibles et différenciation populationnelle élevée Variation temporelle et différenciation populationnelle faible Spécialisation et paysage de sélection Interaction écologique et paysage de sélection Nuance et hypothèses alternatives Paysage de sélection Composition taxonomique Maintenance versus promotion de la différenciation BIBLIOGRAPHIE 27 28 30 33 44 45 45 45 46 46 47 47 48 48 49 49 50 50 51 51 53 Vil LISTE DES TABLEAUX TABLE 1. Characteristics of the six study lakes with high (A) or low (B) differentiation of dwarf and normal Lake whitefish ecotypes 34 2. Dominant taxa in each size class (0.3 mm to up to 2.5 mm) of lake groups A and B and for each month (June and August) 36 TABLE LISTE DES FIGURES 1. Location map of sampling sites in the St.-John river watershed in southeasthem Québec (Canada) and northeasthern Maine (United States) 38 FIGURE 2. Oxygen profiles of each lake nested by lake groups A and B in June and August 2002 and 2003 39 FIGURE 3. Principal component analysis of abiotic factors explaining différences between lake group A and B 40 FIGURE FIGURE4. Top panels: Comparisons of biomass between months in lake group A and B. Bottom panels: Comparisons of biomass by month (June and August) between lake groups 41 5. Comparison of size class (mm) of zooplanktonic communities in terms of biomass and density between lake groups FIGURE 42 6. Relative biomass and density of zooplankton by taxon in lake group A and B for each month 43 FIGURE INTRODUCTION GÉNÉRALE La compréhension des mécanismes qui maintiennent et génèrent la biodiversité est au cœur des préoccupations de la recherche fondamentale et appliquée actuelle. D'ailleurs, la Radiation adaptative récemment revue par Schluter en 2000 est une théorie évolutive et écologique permettant d'acquérir des connaissances au sujet du maintien et de la promotion de la divergence populationnelle menant ultimement à la formation d'espèce (sensu Mayr 1942; voir Schluter 1996; Taylor 1999). Bien que selon cette théorie le contexte écologique joue un rôle aussi prépondérant que l'histoire évolutive, peu d'études se sont penchées sur la caractérisation et la quantification des facteurs écologiques en milieu naturel. Dans cette étude, cette question sera abordée en utilisant comme modèle une espèce de poisson des régions tempérées du nord; le grand corégone (Coregonus clupeaformis Mitchill.), ayant divergé en deux écotypes à plusieurs reprises, sous un patron parallèle, dans plusieurs lacs du bassin versant de la rivière St-Jean (Maine, États-Unis et Québec, Canada) (Chouinard étal. 1996;Bernatchezera/. 1999). LA RADIATION ADAPTATIVE La radiation adaptative étudie les mécanismes écologiques de spéciation découlant de divergences populationnelles. Elle se définit par la diversification rapide d'une lignée évolutive en plusieurs espèces, suite à la différenciation de traits reliés à l'exploitation de différents habitats et/ou de ressources disponibles (revu par Schluter 2000). Cette divergence populationnelle est observée notamment chez plusieurs groupes taxonomiques du règne animal tel que les oiseaux (Grant et Grant 2002; Benkmann 2003), les poissons (revu par Schluter et McPhail 1993; Robinson et Wilson 1994), les lézards (Losos et al 1998) mais aussi les bactéries en laboratoire (Rainey et Travisano 1998; MacLean et Bell 2003). Selon cette théorie, trois processus écologiques sont responsables de l'origine et de la maintenance de la divergence populationnelle : 1) la sélection naturelle divergente, 2) la compétition pour les ressources par le déplacement de caractère ou le relâchement des caractères, et l'opportunité écologique, et enfin, 3) la spéciation écologique (revu par Schluter 1996; Schluter 2000). Cette introduction portera sur les deux premiers processus, mais l'emphase sera mise sur l'opportunité écologique puisqu'il s'agit du thème central du chapitre de ce mémoire. LA SÉLECTION NA TURELLE ET LA DIVERGENCE POPULA TIONNELLE La sélection naturelle dite divergente a lieu lorsque chacun des points sur le gradient phénotypique d'une ressource impose aux espèces consommatrices des pressions de sélections distinctes favorisant des combinaisons de traits spécifiques pour leurs exploitations efficaces (Schluter 1996). Premièrement, cette sélection a lieu dans la population s'il y a une variation phénotypique sur les traits reliés à l'exploitation de la ressource sur laquelle il y a aussi une variation phénotypique. Par exemple, une variété de largeur de bec existant pour l'exploitation d'une variété de taille de graine disponible (Boag et Grant 1981; Smith 1993; Benkman et al. 2003). En se référant à la métaphore du paysage de sélection de Simpson (1953); représenté par des collines de fitness élevés et des vallées de fitness plus bas, les différentes opportunités écologiques peuvent représenter plusieurs pics sélectifs potentiellement adaptatifs (i.e. : le génotype qui exploite cet opportunité à un plus haut fitness que les génotypes qui exploitent d'autres opportunités (sensu William 1966; Kawecki et Ebert 2004)). La hauteur du pic correspond par exemple à la disponibilité des graines et l'étendue correspond à la taille des graines. Deuxièmement, il doit aussi y avoir un avantage à l'exploitation d'un pic de ressource pour certain individu de la population possédant le ou les traits relié(s) à son exploitation efficace. Pour ce dernier cas, il est possible d'établir un paysage adaptatif illustrant cette situation. Ainsi, dans le paysage adaptatif correspondant, l'étendue du paysage de pics représente le phénotype des individus et la hauteur représente le fitness correspondant dans l'environnement étudié (Simpson 1953). Finalement, il y a sélection naturelle seulement si le, ou les traits reliés à l'exploitation efficace ont une composante héréditaire. Ces trois facteurs sont des prérequis pour que la sélection naturelle ait lieu (Darwin 1859; revu par Harder 2001). COMPÉTITION La compétition pour les ressources peut se manifester de deux façons : par déplacement des caractères ou par le relâchement des caractères. La première propose que lorsque deux espèces très similaires entrent en contact dans un environnement où les ressources sont limitées, les deux espèces seront fortement en compétition. Cette forte compétition amènera une des deux espèces à disparaître par exclusion compétitive ou à utiliser une ressource alternative (Brown et Wilson 1956; revus par Schluter et McPhail 1992; Grant 1994; Robinson et Wilson 1994; Losos 2000; Day et Young 2004) qui avec le temps favorisera un déplacement de caractère par sélection naturelle divergente (Schluter et McPhail 1992; revus par Grant 1994; Robinson et Wilson 1994; Losos 2000). Le relâchement de caractère est sensiblement similaire; il se produit lorsqu'une espèce qui se trouvait en sympatrie avec une espèce similaire se retrouve en allopatrie (revus par Robinson et Wilson 1994). Ainsi, il se produira un relâchement de caractère causé par la compétition intraspécifïque pour la ressource anciennement utilisée (Smith and Skùlason 1996; Bolnick 2001; 2004), si une ressource alternative est présente. Enfin, la compétition est vue comme le principal agent promoteur de la différenciation ou de la divergence des populations (Simpson 1953; Schluter 2000). L'OPPORTUNITÉ ÉCOLOGIQUE Cette ressource alternative et disponible est appelée opportunité écologique (Simpson 1953). Brièvement, l'opportunité écologique est la chance de faire face à une niche «ouverte», ou d'être en contact avec une ressource alternative et sous-utilisée (Schluter 2000). Niche Une niche «ouverte» est traditionnellement définie par le concept de Grinell (1917-1924 dans Schluter 2000) et est caractérisée par l'ensemble des caractéristiques environnementales utilisables pour une espèce dans un habitat donné possédant des caractéristiques biotiques et abiotiques particulières (Schluter 2000). Puisque la niche ne peut pas être en elle-même ouverte parce qu'elle fait intervenir les caractéristiques de l'espèce à laquelle elle appartient et les caractéristiques de l'environnement dans lequel elle habitent, une alternative à l'appellation «niche ouverte» serait de décrire la niche en associant les définitions de «Grinell (1917-1924)-d'Hutchinson (1952)» et «d'Elton (1927)MacArthur et Levin (1967)», comme suggéré par Leibold (1995) (voir aussi Robinson et Wilson 1994). Ainsi, Leibold (1995) définit la niche écologique d'une espèce comme étant le résultat de compromis effectués par des espèces qui interagissent dans un milieu pour assurer leur coexistence. La source du compromis est l'ensemble des conditions abiotiques et biotiques ainsi que les interactions entre toutes ces composantes qui permettent à l'espèce de maintenir un fitness absolu égal à zéro. En somme, la différence entre ces deux définitions réside dans le fait que les caractéristiques environnementales disponibles pour une espèce sont le résultat de sa propre coexistence avec les autres organismes qui habitent dans un milieu possédant des caractéristiques propres qui peuvent, elles aussi, être modifiées par ses habitants. Peu importe la définition de niche exacte employée, on conçoit qu'il puisse exister dans un nouvel environnement ou qu'il puisse se libérer dans son environnement une nouvelle «opportunité» d'exploitation, autrement dit, une niche «ouverte». Chance et contingence historique Par la nature de la définition de niche, il est maintenant clair que la chance ou «l'opportunité» d'avoir accès à une nouvelle ressource ou à un nouvel habitat fait intervenir l'histoire évolutive individuelle et l'histoire évolutive de la communauté dans laquelle l'espèce se retrouve. L'histoire évolutive individuelle, ou populationnelle, façonne au fil du temps l'architecture génétique des individus qui expriment chacun un certain phenotype et procure une certaine variation dans le phenotype à l'intérieur d'une population (Gould et Lewontin 1979). L'histoire évolutive individuelle détermine, en partie, le niveau de flexibilité de l'expression d'un ou de phénotypes dans différents environnements. Ce niveau de flexibilité favorise ou limite l'exploitation de nouvelles ressources ou de nouveaux habitats par les individus et donc, par la population. Parallèlement, la chance fait aussi intervenir l'histoire évolutive de la communauté. Cette dernière fait référence à la sommation des histoires évolutives et donc, de l'expression phenotypique, à un moment donné, de l'ensemble de la communauté dans laquelle l'espèce étudiée coexiste. Cela revient à dire que l'histoire évolutive de la communauté façonne, maintient ou modifie ce qui est accessible à l'espèce dans un environnement, à un moment donné (ex. contingence historique; Gould et Lewontin 1979; Robinson et Wilson 1994; Tyerman et al. 2005; Day et Young 2004). Conséquemment, la chance semble jouer un rôle essentiel en établissant les contraintes évolutives et donc aussi les caractéristiques de la niche. Ces dernières jouent donc un rôle important en déterminant la nature de la divergence possible, c'est-à-dire la nature de la sélection agissant sur les espèces (Travisano et al. 1995; Rainey et Travisano 1998; Losos et al. 1998; Taylor et McPhail 2000; Pigeon et al. 1997; Langerhans et DeWitt 2004). Plus simplement, l'opportunité écologique est constituée des ressources et des caractéristiques de l'environnement qui imposent certaines contraintes à l'atteinte d'un fitness plus grand ou égal à zéro. Celles-ci permettent (ou limitent) la divergence dans le contexte de la théorie de la radiation adaptative. Elle est donc vue comme l'agent régulateur responsable de l'étendue de la divergence ou de la différenciation des populations (Schluter 2000) puisque responsable de la sélection naturelle divergente. IMPÉRATIFS À LA DIVERGENCE OU À LA DIFFÉRENCIATION DES POPULATIONS Pour qu'il y ait divergence populationnelle, la sélection naturelle doit respecter certaines particularités en plus des prérequis mentionnés précédemment. D'abord, il doit y avoir une certaine variation sur un trait phénotypique reliée à des bases génétiques différentes et héritables, et non à la variation d'expression d'un seul génotype dans un environnement possédant des caractéristiques particulières. Ainsi, la variation dans l'expression phénotypique d'un génotype reliée aux changements de l'environnement (ou plasticité phénotypique) doit être associée à des coûts et des contraintes (revu par Kawecki et Ebert 2004). Ensuite, il doit y avoir une faible variation temporelle dans la disponibilité de l'opportunité écologique pour qu'en tout temps, il soit préférable pour un individu adapté à un pic dans le paysage sélectif d'exploiter ce pic (ex. avantage des individus spécialistes : Futuyma et Moreno 1988; revus par Kawecki et Ebert 2004). Une forte variation de l'opportunité écologique sur un pic ou sur plusieurs pics permet un fitness plus élevé des individus adaptés à plusieurs pics (ex. avantage des individus généralistes) défavorisant ainsi la divergence populationnelle. Aussi, les couples doivent aussi se former et se reproduire à l'intérieur d'un même pic adaptatif; assortative mating (revus par Kawecki et Ebert 2004). C'est-à-dire qu'un flux de gènes élevé entre les individus d'une population soumise à différentes forces sélectives empêcherait la divergence en deux sous-populations chacune adaptée à un pic sélectif (revus par Kawecki et Ebert 2004; mais voir St-Laurent et Bernatchez 2003; Hendry et Taylor 2004). Il doit y avoir aussi, en plus de ce flux de gènes restreint, une sélection forte contre les individus adaptés à un pic sélectif pour l'exploitation d'un deuxième et moyenne contre les individus adaptés aux deux pics. Une sélection trop forte sur un pic par rapport à un autre empêcherait la divergence en limitant l'exploitation éventuelle de pics alternatifs ou d'environnements potentiellement alternatifs (Kawecki et Ebert 2004). Enfin, les différences de qualité et de taille des habitats doivent être minime pour qu'il puisse y avoir divergence populationnelle. En ce sens, un mécanisme de divergence populationnelle suggérant l'influence de la compétition interspécifique à d'ailleurs été étudié dans des populations au laboratoire. MÉCANISME POTENTIEL DE LA DIVERGENCE POPULA TIONNELLE En effet, différentes théories et études ont proposé que la réduction de la compétition interspécifique dans l'environnement où habitent une population ou dans un environnement nouvellement colonisé par une espèce permettrait exploitation de ressources alternatives et donc la divergence populationnelle (Smith et Skùlason 1996; Rainey et Travisano 1998). D'autre part, l'augmentation de la compétition interspécifique pour une ressource anciennement utilisée qui deviendrait limitante serait aussi un agent promoteur de la différenciation dans un environnement possédant un gradient de ressource disponible (Bonick 2001; 2004). Ainsi, la réduction du fitness causée par l'exploitation massive d'une ressource jadis adaptative pourrait amener à l'exploitation d'une ressource alternative qui est initialement moins avantageuse mais dont les coût reliés à son utilisation deviennent peu à peu moins élevés que l'utilisation de la ressource initialement exploitée (ex. compétition intraspécifique élevée ou fréquence dépendante). Un milieu donc où deux ou plusieurs ressources sont de qualité et de taille similaire. La sélection favoriserait ainsi certains individus avantagés par cette ressource alternative et donc, à plus long terme, la spécialisation de la population, par l'augmentation de l'efficacité d'utilisation de cette seconde ressource (Bolnick 2001; 2004). LES FACTEURS STOCHASTIQUES VERSUS DÉTERMINISTES Bien que les facteurs déterministes comme la sélection naturelle semblent les plus importants dans la divergence des populations, plusieurs études ont mis en évidence, le potentiel des facteurs stochastiques tels que la contingence historique à expliquer ces phénomènes. Par contre, de multiples observations réalisées en milieu naturel portant sur 8 des événements de divergence populationnelle en parallèle potentiellement reliés à l'exploitation de ressources similaires pointent les facteurs déterministes comme moteur le plus probable. Enfin, même dans un cadre stochastique, la force de la sélection pourrait avoir un effet sur le degré de divergence (Travisano et al. 1995) entre les différentes populations ayant divergé en parallèle. LES LACS ET LES POISSONS DES RÉGIONS TEMPÉRÉES DU NORD Les paires d'espèces sympatriques limnétiques-benthiques des lacs des régions tempérées du Nord sont des exemples de divergences populationnelles répétées, indépendantes et parallèles chez les vertébrés. Elles ont été observées chez plusieurs familles de poissons tels les salmonidés (Griffiths 1994; Bernatchez 2004), les gastrosteidés (Schluter et McPhail 1992), les centrarchidés (Robinson et al. 2000) et les osmeridés (Taylor et Bentzen 1993). Dans le cas des paires d'épinoches à trois épines (Gasterosteus aculeatus) et de grands corégones (Coregonus clupeaformis) d'Amérique du Nord la colonisation des lacs s'est effectuée suite au retrait des glaciers du Winconsinien et le scénario de la double colonisation ne peut être écarté dans presque tous les cas (Taylor et McPhail 2000; Lu et al. 2001). Cependant, une partie de la divergence phénotypique qui a eu lieu ou qui se maintient actuellement s'est déroulée (ou se déroule) en sympatrie. Relation phénotype-environnement La divergence chez ces paires «d'espèces» est reconnue comme étant liée à l'utilisation de l'environnement pélagique par un écotype et de l'environnement épibenthique par l'autre. L'environnement pélagique est caractérisé par la disponibilité de proies de plus petites tailles et la présence de prédateurs alors que l'environnement épibenthique présente des proies plus grosses (Robinson et Wilson 1994; Kahilainen et al. 2004). Ces caractéristiques de l'environnement utilisé sont reliées à des caractéristiques morphologiques, (ex. : le nombre de branchicténie et leur longueur, Schluter et McPhail 1993; Bernatchez et al. 1999; Amundsen et al. 2004a; 2004b), comportementales (ex. : la position dans la colonne d'eau et l'activité natatoire, Rogers et al. 2002) et physiologiques (ex. : la croissance, Rogers et Bernatchez 2005 et expression différentielle de gènes, Derome et al. 2006). De plus, chez les paires d'espèces d'épinoches à trois épines et de grands corégones, un degré de différenciation morphologique et/ou génétique propre à chacun des lacs a été observé (Lu et Bernatchez 1999; Taylor et McPhail 2001). L'intensité de la sélection naturelle divergente causée par les opportunités écologiques semble être la principale cause de la variation dans la divergence parallèle en sympatrie de ces paires d'espèces (Lu et al. 2001). Opportunité écologique Les causes environnementales de la divergence des paires d'espèces des lacs des régions tempérées du nord semblent être la pauvreté en espèces et l'hétérogénéité des habitats disponibles dans les lacs (Hindar et Jonsson 1982; Taylor 1999). Les écosystèmes dits pauvres sont associés à des événements de perturbations et d'extinctions dus aux multiples événements de glaciations survenus en Amérique du Nord (Taylor 1999). On retrouve donc une faible diversité en espèces de poissons réduisant les interactions compétitrices et prédatrices. De plus, un gradient physique élevé peut être observé à l'intérieur de chacun des lacs (ex. profondeur, température, développement des bordures, substrats) et associé à une variation dans les espèces de proies ou la taille des proies retrouvées dans l'habitat (Taylor 1999). On observe aussi une variation au niveau de la morphométrie, de la productivité et de la physico-chimie entre les lacs pouvant résulter en des compositions différentes en espèces et en abondance de proies, ainsi qu'en des différences dans leur cycle saisonnier (Griffith 1994; Robinson et al. 2001; Nosil et Reimchen 2005; Hindar et Jonsson 1982; Malmquist et al. 1992). Ces différences au niveau des proies pourraient agir comme des pressions de sélection différentes sur les communautés de poissons puisque chacune des espèces a un contenu calorique différent (Oison et al. 2003) et est potentiellement manipulée et assimilée de façon différente (revu par Futuyma et Moreno 1988) leur donnant des avantages différents au niveau énergétique. Les différences au niveau des facteurs physico-chimiques pourraient aussi influencer les dynamiques évolutives. En effet, la majorité des lacs des régions tempérées du Nord sont dimictiques c'est-à-dire qu'ils présentent deux périodes de mélanges et deux périodes de stratification thermique par année (Wetzel 2001). Ainsi, durant la saison estivale la stratification amène l'observation de température différente en fonction de la profondeur 10 influençant différemment les caractéristiques abiotiques et biotiques rencontrés selon les strates (i.e. : épilimnion, métalimnion et hypolimnion) et ainsi les poissons s'alimentant plus ou moins profondément dans la colonne d'eau sont soumis à différentes caractéristiques de leur environnement. De plus, ces différentes caractéristiques peuvent aussi être influencées par le niveau trophique des lacs (ex. : eutrophe, oligotrophe, dystrophe) permettant d'émettre l'hypothèse voulant que différentes caractéristiques de la limnologie des lacs puissent influencer les communautés qui y vivent (ex. : Carpenter et al. 1985; Tessier et Woodruff 2000) et donc les pressions de sélection qui agissent potentiellement sur le niveau de différenciation atteint entre les populations. En somme, deux environnements peuvent grossièrement subdiviser les lacs: le pélagique et le benthique. Cette subdivision constitue une cause potentielle de l'évolution répétée des formes naines pélagiques et «normales» épibenthiques (Robinson et Wilson 1994). À un niveau plus fin, les différentes variations dans les facteurs écologiques pourraient être la cause des variations dans les pressions sélectives qui agissent sur les poissons. À leur tour, les variations dans les pressions de sélection pourraient expliquer le degré de divergence génétique et morphologique observé entre les écotypes à l'intérieur des populations. Cependant, les lacs des régions tempérées dans lesquels vivent les diverses formes sympatriques de poissons ont fait l'objet de peu d'études quant aux facteurs écologiques incluant la disponibilité des ressources ou «opportunité écologique». Par contre, Schluter et McPhail (1992) ont fait la quantification des ressources et de sa fluctuation saisonnière en utilisant les contenus stomacaux. De plus, Robinson et ses collaborateurs (2000) ont aussi fait la quantification des ressources écologiques par les contenus stomacaux et la caractérisation de certaines particularités de l'environnement qui serait potentiellement reliées à la divergence populationnelle. Néanmoins, même si quelques études suggèrent que l'opportunité écologique joue un rôle clé dans la divergence des populations sympatriques de poissons (Taylor et McPhail 2000; Lu et al. 2001), aucune a traité directement des ressources disponibles dans l'environnement, de leurs variations temporelles et des caractéristiques abiotiques globales différenciant des environnement possédant des populations ayant divergé en parallèle de façon répétée. 11 LE GRAND CORÉGONE Le grand corégone (Coregonus clupeaformis Mitchill.) est l'un des modèles animal permettant d'étudier le rôle de l'opportunité écologique lors de divergence populationnelle. L'évolution répétée de l'écotype nain à partir de l'écotype normal dans plusieurs lacs d'Amérique du Nord, et plus précisément dans les lacs d'origine post-glacière du bassin versant de la rivière St-Jean, porte à postuler en faveur de l'importance de la sélection naturelle divergente par le biais de l'opportunité écologique pour l'initiation et le maintien du degré de divergence entre les écotypes de grand corégone en sympatrie (Lu et Bernatchez 1999; Bernatchez et al. 1999). Relation phénotype-environnement L'écotype nain et l'écotype normal du grand corégone sont reconnus pour leur utilisation quasi exclusive de l'environnement pélagique et épibenthique respectivement (Bernatchez et al. 1999). Ces écotypes montrent aussi une association phénotype-environnement à ces deux environnements (Bernatchez et al. 1999; Rogers et al. 2002). Par exemple, les nains ont une plus petite taille à l'âge adulte, un taux de croissance plus faible et mature sexuellement en plus bas âge comparativement à l'écotype normal (revus par Scott et Crosmann 1974; Chouinard et al. 1996). Ces caractéristiques de l'écotype nain sont d'ailleurs reliées à celles qui caractérisent les espèces de poissons planctonivores (voir Robinson et Wilson 1994). De plus, lors des expériences en laboratoire en conditions contrôlées, l'écotype nain maintient une position plus élevée dans la colonne d'eau et nage plus activement par rapport à l'écotype normal (Rogers et al. 2002). Au niveau morphologique, le trait qui est le plus discriminant entre les écotypes est la possession d'une plus grande quantité de branchicténies chez les nains. L'augmentation en nombre de ces dernières semblent aider à la rétention des proies de plus petite taille (Budy et al. 2005; revu par Schluter et McPhail 1993; Robinson et Wilson 1994; mais voir Langeland et Nost 1995) en régulant la dynamique des fluides à l'intérieur de la cavité buccale (Sanderson et al. 2001; Brainerd 2001). L'augmentation de l'efficacité de capture des petites proies pourrait favoriser l'accroissement de l'efficacité de conversion énergétique (Whiteley et al., en préparation; Robinson et Wilson 1994; Lu et Bernatchez 1999; pour un point de vue inverse voir Trudel et al. 2001). De plus, Bernatchez (2004) a démontré que le nombre de 12 branchicténies est sujet à la sélection directionnelle chez les nains mais pas chez les normaux. Cela permet d'émettre l'hypothèse voulant que la disponibilité des proies planctoniques, la structure de taille disponible chez les proies et leur composition taxonomique, soient responsables de la force et de la direction de la sélection naturelle divergente chez les populations de grand corégones. Opportunité écologique Un autre argument en faveur d'une divergence causée par l'opportunité écologique provient du fait que les nains sont toujours retrouvés en sympatrie avec les normaux dans les lacs où le cisco de lac, C. artedii, est absent. Ce dernier représente un compétiteur direct potentiel qui limiterait l'exploitation de la niche planctonique (Davis et Todd 1998; Trudel et al. 2001). Bien que d'après la littérature l'habitat pélagique est similaire dans les différents lacs, une corrélation négative entre le degré de chevauchement de la diète des nains par rapport au normaux et l'étendue du flux de gènes entre les ecotypes de différents lacs est observée (Bernatchez et al. 1999). De plus, une corrélation négative entre l'étendue de la spécialisation morphologique entre les ecotypes et l'étendue du flux de gènes diffèrent entre les lacs (Lu et Bernatchez 1999). Conséquemment, on peut émettre l'hypothèse que les différents lacs représentent différents degrés d'opportunité écologique résultant d'un nombre possible de facteurs tels que les caractéristiques physico-chimiques des lacs, la disponibilité des ressources et leurs stabilités. Ces facteurs détermineraient la force et la direction des forces sélectives (Bernatchez et al. 1999; Lu et al. 2001). OBJECTIF PRINCIPAL Cette étude propose d'approfondir nos connaissances sur la différence dans la disponibilité et la stabilité des caractéristiques biotiques et abiotiques présente entre deux groupes de trois lacs du bassin versant de la rivière St-Jean. Le premier groupe de lacs présente un degré élevé de divergence morphologique et génétique entre les ecotypes nains et normaux de grand corégone et le deuxième groupe un degré de différenciation faible (Lu et Bernatchez 1999). L'objectif principal est de vérifier l'hypothèse selon laquelle le degré de divergence observé actuellement entre les ecotypes nains et normales est causé par les particularités de l'opportunité écologique dans chacun des lacs. J'émets aussi la prédiction 13 qu'une plus grande opportunité écologique ainsi qu'une faible variabilité temporelle de celle-ci, soient liées à un degré de divergence plus élevé. CHAPITRE 1 THE ROLE OF HABITAT AND PREY COMMUNITY IN THE PARALLEL EVOLUTION OF LAKE WHITEFISH DWARF ECOTYPES 15 RÉSUMÉ Les populations sympatriques de poissons habitant les régions tempérées du nord permettent d'étudier les processus de divergence populationnelle et de radiation adaptative. En dépit de l'intérêt considérable porté à l'étude de ses systèmes, l'identification des caractéristiques abiotiques et biotiques de l'habitat et des ressources ainsi que leur stabilité temporelle a été négligée. L'identification de ces caractéristiques pourrait permettre une meilleure compréhension des processus responsable du degré de différenciation écologique et morphologiques observé dans les cas répétés de divergence populationnelle. Le grand corégone, Coregonus clupeaformis (Mitchill.), est un organisme idéal pour obtenir ce type d'information puisque l'évolution parallèle et sympatrique de l'écotype nain à partir de l'écotype normal s'est produite dans plusieurs lacs post-glacières. Dans cette étude, 19 caractéristiques biotiques et abiotiques ont été comparées entre deux groupes de trois lacs où l'on peut y observer un degré élevé ou faible de différenciation morphologique et génétique entre les écotypes. Les résultats montrent que les lacs, possédant les populations les plus différenciées, sont caractérisés par une diminution plus prononcée de l'oxygène retrouvé dans l'hypolimnion. Ces lacs montrent aussi une variation saisonnière plus faible de la biomasse des proies zooplanctoniques que les lacs possédant des écotypes sympatriques moins différenciés. Enfin, les lacs possédant des écotypes sympatriques plus différenciés montrent une biomasse des proies zooplanctoniques plus faible au mois d'août et une distribution de leurs tailles plus étroite. Ces résultats indiquent qu'une différentiation plus élevée entre les écotypes se produit dans les lacs où la disponibilité de l'habitat est réduite, ce qui suggère une augmentation de la compétition pour les ressources. La variation saisonnière plus élevée dans les ressources ainsi que la distribution des tailles de la communauté des proies zooplanctoniques pourraient être des composantes déterminantes du paysage sélectif influençant la divergence des traits phénotypiques tels que les branchicténies entre les écotypes. Cette étude a identifié un lien entre le métabolisme des lacs, les interactions trophiques et la structure de l'opportunité écologique qui pourrait être déterminant dans l'évolution des populations de grands corégones. Ces observations pourraient représenter un pas dans l'établissement d'un lien entre les théories écologiques fondamentales telles que les cascades trophiques et les théories évolutives fondamentales telles que la radiation adaptative. 16 ABSTRACT Sympatric fish populations observed in many north temperate lakes are among the best models to study the processes of population divergence and adaptive radiation. Despite the considérable research interest that has been devoted to such Systems, little is known about the environmental conditions that favour ecological and morphological divergence. In this study we examined the biotic and abiotic properties of postglacial lakes in which Lake whitefish, Coregonus clupeaformis (Mitchill.), occurs as a derived dwarf ecotype in sympatry with an ancestral normal ecotype. We compared 19 limnological variables between two groups of lakes with either high or low levels of morphological and genetic differentiation between the dwarf and normal ecotype. Oxygen was the primary variable separating the two lake groups, with the most divergent populations observed in the lakes with the greatest hypolimnetic oxygen depletion. At a lower level of significance, thèse lakes also had less monthly variation of prey biomass, lower zooplankton densities and a narrower distribution of the prey length in the plankton food resources. Thèse results indicate that the highest differentiation between sympatric ecotypes occurs in lakes with reduced habitat availability that could increase compétition for resources. They also suggest that temporal stability of resources during the growing season and the lengthdistribution of prey are additional components of the sélective landscape that influence phenotypic divergence between ecotypes. This study therefore revealed that parallelism in the extent of phenotypic divergence among sympatric fish population is associated with parallelism in adaptive landscape in ternis of différences in limnological characteristics, as well as availability and structure of the zooplanktonic community. Studies of this type represent a first step in 'evolutionary limnology' towards bridging the gap between aquatic ecological (e.g., trophic cascades), and evolutionary théories (e.g., adaptive radiation). 17 INTRODUCTION Divergent natural sélection for the exploitation of différent resources and habitats is considered a key process in population divergence and adaptive radiation (Smith and Skùlason 1996; Schluter 2000). This gênerai hypothesis has been supported by many empirical studies that hâve documented parallel phenotypic évolution in independent lineages under similar ecological conditions (Schluter and McPhail 1993; Smith and Skùlason 1996; Ruzzante et al. 2003; Chiba 2004). Divergent natural sélection can only arise if there is an alternate ecological opportunity (sensu Simpson 1953) associated with the availability of "open niches"; i.e., underused habitats and resources which can fulfill the requirements for a species to complète its life cycle (Grinell 1917, 1924 in Schluter 2000). Although chance and history also play a rôle (Travisano et al. 1995), biotic and abiotic properties of the environment are thought to be the primary forces shaping phenotypic divergence (Travisano et al. 1995; Losos et al. 1998; MacLean and Bell 2003; Langerhans and DeWitt 2004). Thèse niche characteristics and their temporal variation will likely détermine the strength of sélection acting on phenotypic évolution (Kawecki and Ebert 2004). High variability in sélective régimes due to fluctuations in temporal ecological opportunity may occur through changes in abiotic or biotic conditions in the environment (Grant and Grant 2002; Hjelm and Johansson 2003; Svanbâck and Persson 2004). This, in turn, may constrain divergence because potential changes in fïtness through the exploitation of alternative ecological resources would not occur. Progress in understanding divergent phenotypic évolution is best achieved by the study of young lineages in which rapid phenotypic divergence and reproductive isolation are still in action, and that lend themselves to comparative studies (Schluter 2000). Such idéal conditions for rapid divergence likely prevailed in north temperate lakes that hâve been cyclically affected by glacial expansions and retreats during the Pleistocene, perhaps as many as 20 times (Martinson et al. 1987). Members of the north temperate freshwater and anadromous fish fauna share several attributes that corroborate the predicted effect of postglacial ecological opportunity in priming rapid divergence, and make them putative candidates for studying the process of population divergence and adaptive radiation. Of particular interest is the occurrence of sympatric forms that predominate in salmonid fishes, 18 but that also exist in phylogenetically remote families, including Gasterosteidae, Osmeridae, Centrarchidae and Catostomidae (reviewed in Robinson and Wilson 1994; Smith and Skùlason 1996; Taylor 1999; Robinson and Schluter 2000). Despite the considérable interest that has been devoted to studying such Systems, little attention has been paid to documenting the "limnological landscape", which may be the key towards understanding the origin of ecological and morphological différences in diverging freshwater fish populations (Bernatchez et al. 1999; Lu et al. 2001). Lake whitefish, Coregonus clupeaformis (Mitchill.), is well-suited for studying the rôle of ecological opportunity in population divergence. The parallel évolution of a dwarf ecotype derived from an ancestral normal ecotype has been documented in several postglacial lakes of the St. John River basin (Maine, USA), where dwarf whitefish are only found in sympatry with a normal population (Chouinard et al. 1996; Pigeon et al. 1997; Lu and Bernatchez 1999). Dwarf and normal sympatric ecotypes mainly use pelagic and epibenthic habitat, respectively (Bernatchez et al. 1999), and thèse forms also exhibit phenotypeenvironment associations with their respective niches (Bernatchez et al. 1999; Rogers et al. 2002; Rogers and Bernatchez 2005; Derôme et al. 2006). The dwarf forms hâve smaller size at maturity, slower growth rates and an earlier âge at maturation relative to normal ecotype (Rogers et al. 2002), as is typical for planktivorous fishes in gênerai (Robinson and Wilson 1994). Both ecotypes also differ in their swimming behavior, whereby the dwarf ecotype typically swims more actively and prefers higher positioning in the water column in expérimental laboratory conditions (Rogers et al. 2002). At the morphological level, the most discriminating trait between ecotypes is the gill-raker apparatus (Lu and Bernatchez 1999), where the dwarf ecotype typically has more numerous and less separate gill-rakers than the normal population, probably to facilitate rétention of smaller, planktonic prey (Robinson and Wilson 1994; Sanderson et al. 2001; Kahilainen et al. 2004; Budy et al. 2005). It has also been suggested that divergent sélection has played an important rôle in driving divergence in gill-raker characteristics (Rogers et al. 2002), and that sélection has acted more strongly on the dwarf relative to normal ecotype (Bernatchez 2004). Most North American temperate lakes where cases of parallel phenotypic évolution hâve been reported in fishes typically show poor fish species diversity (review by Taylor 1999; 19 Schluter 2000). For example, dwarf whitefish strictly occur in sympatry with the normal ecotype, and in the absence of the Lake cisco, C. artedii, a direct competitor species that may limit exploitation of the planktonic niche (Davis and Todd 1998; Trudel et al. 2001; Bernatchez 2004). However, the absence of Lake cisco cannot explain in itself the amonglake variation in niche utilization and level of divergence between dwarf and normal whitefish ecotypes. For instance, Bernatchez et al. (1999) documented a négative corrélation between the degree of planktonic/benthic diet overlap and the level of gène flow between them, as well as a négative corrélation between the extent of gène flow and the morphological specialization between ecotypes in différent lakes (Lu and Bernatchez 1999). Indeed, thèse observations raise the hypothesis that différent lakes represent différent degrees of, or simply différent, ecological opportunity, due to variation in lakespecific physico-chemical characteristics (Griffiths 1994; Hindar and Jonsson 1982; reviewed by Taylor 1999), resource (trophic and spatial) availability (reviewed by Smith and Skùlason 1996) and stability (reviewed by Kawecki and Ebert 2004). Ail of thèse factors could potentially affect the strength and direction of the sélective forces (e.g., Robinson et al. 2000; Nosil and Reimchen 2005) acting on the dwarf ecotype. To our knowledge, however, no study on fïsh sympatric populations has assessed the relative importance of temporal variation of spécifie abiotic and biotic environmental factors in explaining patterns of fish population divergence in replicate adaptive radiations. In the présent study, we first postulated that the observed variability in the extent of morphological differentiation observed between sympatric dwarf and normal whitefish ecotypes could be linked to limnological characteristics distinguishing lake groups. Secondly, we tested the hypothesis that différences in patterns of divergence observed in the gill-raker apparatus between sympatric dwarf and normal ecotypes is associated with lake-to-lake variation in ecological opportunity and/or stability, whereby higher levels of divergence between ecotypes would be positively associated with a more pronounced and temporally more stable availability of alternative ecological opportunity. Thirdly, we tested the hypothesis that planktonic community structure (size distribution and taxonomic composition) and stability could be reflective of the sélective landscape acting on the dwarf ecotype in différent lakes. We addressed thèse hypothèses by comparing 19 biotic and abiotic characteristics of two groups of lakes located within the same watershed. Ail of the 20 lakes had sympatric dwarf and normal populations, and the groups were separated according to either high or low levels of morphological and genetic differentiation. MA TERIALS AND METHODS Study Sites Six dimictic lakes in the S t.-John River watershed (Fig. 1) were selected for this study on the basis of prior work in the région. As previously documented by Lu and Bernatchez (1999), three contained Lake whitefish with highly divergent, sympatric populations of the two ecotypes (group A lakes: Cliff Lake (CL), Indian Pond (IP) and Webster Lake (WL)) and three had a low level of differentiation (group B lakes: Crescent Pond (CP), Témiscouata Lake (TL) and East Lake (EL)). Lake cisco (C. artediî) is absent from ail six lakes, but ail harbor healthy populations of two whitefish predators: Lake trout (Salvelinus namaycush) and Burbot (Lota Lota; except in CP). Five sampling stations were randomly selected in each lake between the depths of 5 m and 25 m, where both whitefish ecotypes are most commonly found. Both forms are known to avoid warm waters of the littoral zone ( < 5 m) during the growing season (Bernatchez, unpublished observations). Sampling stations were located using a géographie positioning system (GPS) and bathymétrie maps and were sampled at two différent times of the year (16-26 June, 20-28 August) in 2003. Limnological Variables For the lakes in Maine (USA), altitude, fetch, volume (V), surface area, mean depth (Zmoy), maximal depth (Zmax) (Wetzel 2001) were obtained from topographical and bathymétrie maps and from information available on the PEARL website (http://pearl.spatial.maine.edu). For the lakes in Québec, altitude and Zmax were interpreted from maps, whereas volumes, surfaces and Zmoy were obtained from bathymétrie maps imported as images in a vector-based Géographie Information System (GIS) software package Arc GIS 8.3 and georeferenced on a 1:50000 Arc GIS map (National Topographie Data Base (NTDB), Natural Resources Canada (NRC). Using the perimeter and area fonction of this software, data were obtained for each 6.1 m (20 ft) depth interval. Lake volume was then estimated by the method of summation of the frusta of a séries of truncated cônes of the strata (Wetzel 2001). Shoreline development (DL) of each 21 lake was calculated following Wetzel (2001) and with data obtained from perimeter values calculated in Arc GIS software. Chemical and physical characteristics of ail lakes were measured at the deeper station of the lake, when the Zmax was less than 25 m, for both June and August with a 5 L Go-flow bottle. For total phosphorus (TP), colored dissolved organic matter (CDOM), Chlorophyll a (Chl a) and dissolved inorganic carbon (DIC), bottles were washed and samples were preserved following the NLET (National, Laboratory for Environmental Testing, Burlington (Ontario, Canada)) standard protocols. Water transparency, or Secchi depth (ZSecchi), was measured using a 20-cm Secchi disk (Wetzel and Likens 1979). Température and oxygen were measured every 0.5 m with an oxygen meter (YSI model 58). Oxygen concentrations were measured in mg L"1 and transformed to percent saturation (%) following Wetzel and Likens (1979). Température and oxygen profiles were obtained at the same stations in 2002 and 2003. Oxygen saturation (02-hypolimnetic) and water température values (Tmax-z) at the bottom of the lakes in August were the two variables considered for more detailed statistical analyses. The epilimnion and metalimnion of each lake were defined as in Hutchinson (1957). Two water samples were taken in each stratum for Chl a, CDOM, DIC and TP analyses. Following sampling, Chl a samples were filtered onto 25 mm Whatman glass fiber filters (grade GF/F) and preserved in cryogénie vials in liquid nitrogen until laboratory analysis. In the laboratory, Chl a analysis were performed following Nusch (1980) with a Cary Eclipse spectrofluorometer. Prior to sample analyses, a standard curve was generated using a standard Chl a solution {Anacystis nidulans, Sigma Biochemical Co.). Concentrations were calculated using Jeffrey and Welschmeyer's (1997) équation. CID samples were analyzed by Gran titration on the evening of the date of sampling with the method as in Wetzel and Likens (1979); pH was measured with a Hanna instruments pH-meter (HI 8424 Microcomputer). TP samples were acidified to pH 2 with 1M H2SO4 the same day of sampling and were analyzed at the NLET laboratory (Environment Canada, Burlington, ON). CDOM samples were stored in the dark at 4°C and then filtered through 0.2 um cellulose acétate membrane prior to absorbance measurements at 320 nm with a Ultrospec 2000 (Pharmacia Biotech) spectrophotometer (Williamson et al. 1999; Belzile et al. 2002). Mean values of thèse limnological variables were calculated for 22 each lake from ail data collected throughout the summer, for both the epilimnion and metalimnion. Zooplanktonic Variables Zooplankton were collected at each station, in June and August of 2003, using a 12 cm Wisconsin plankton net with 80 \im. mesh size. Vertical hauls, at approximately 1 m s"1, were made starting at 0.5 m from the bottom to the surface. S amples were preserved in 95 % ethanol until laboratory analyses (Rodriguez et al. 1993). The depth of the vertical hauls was used to calculate the total water volume filtered at each lake. In the laboratory, samples were filtered onto a 300 iam sieve. Ail individuals were counted and identified to genus or family except for copepods, which were assigned as either calanoid, cyclopoid, harparticoid or nauplii catégories (Thorp and Covich 1991). For each sample, individuals of a given taxonomic group were pooled in an Eppendorf tube and preserved in 95 % ethanol until length measurements were made. Twenty individuals of each taxon in each sample were measured following Culver et al. (1985) to the nearest 0.025 mm using either a dissecting microscope with an ocular micrometer. Thèse samples were then divided into 0.1 mm length classes spanning a range from 0.3 mm to 2.5 mm. The organisms that measured 2.5 mm or more were put in the same length class because they represented less than 1 % of ail the biomass and density in a sampling station. Absolute numbers of individuals of each length class were transformed to density (mean number per class) by multiplying the frequency of individuals of a given taxon in a length class by the number of individuals per liter of this taxon at a given station. Dry weight (DW) of each measured individual was calculated using the following lengthweight régressions: Laçasse and Magnan (1991) for Bosminidae (named as taxon 5), Ceriodaphnia spp. (taxon 17), Daphnia spp. (taxon 6), Sididae (taxon 7), Polyphemus pediculus (taxon 9), Leptodora kindii (taxon 13) and Holopedium spp. (taxon 4) ; Dumont et al. (1975) for calanoids (taxon 1), cyclopoids (taxon 2) and Rotifera (taxon 11); Culver et al. (1985) for copepoda nauplii (taxon 3), Johnston and Cunjak (1998) for Chaoboridae (taxon 10) and Chironomidae (taxon 19); Meyer (1989) for Ephemeroptera (taxon 15), Coleoptera (taxon 12) and Hydracarina (taxon 8); and Rahklola et al. (1998) for Asplanchna spp. (taxon 16). The total biomass (g DW L'1) was estimated following the 23 same procédure as for density. For Annelida (taxon 14) biomass was not estimated and so not included in the biomass estimation. Statistical Analyses Limnological différences between lake groups AU limnological variables were included in the analyses to avoid the loss of information (Rodriguez et al. 1993). Logarithmic transformations log (x + 1) were applied to ail environmental variables except pH in order to reduce skew distribution (Weckstrom et al. 1997) and to better approximate assumptions of normality and homoscedasticity. Sélection of the most informative variables for discriminating lake groups A and B was made using a stepwise ANOVA (STEPDISC procédure, a =0,05). Statistical différences between the two lake groups for each selected variable were further assessed by Student t-tests. AH of thèse statistical analyses were completed using SAS (v. 8.2, SAS Institute). Principal component analsysis (PCA) was then performed using PRIMER-E (v. 5, PRIMER-E Ltd., Plymouth, UK) to evaluate similarities and différences among lakes and aposteriori différences between the two lake groups, after removal of corrélations between variables (Weckstrom et al. 1997). Prey availability and variation in density and biomass Lake group différences in resource availability, both in terms of density and biomass, were analyzed using a two factor (month and lake group) mixed procédure (Littell et al. 2004). Individual lakes represented the basic sampling unit for this statistical test. Mean density and biomass values were calculated over the five stations surveyed in each lake, which was nested in its respective lake group and then considered as a random factor. Spatial variability in biomass and density were measured as the coefficient of variation (CV = 100 * (standard déviation / mean)) for each lake, where stations were the sampling units (see for temporal stability; Caldeira et al. 2005). A mixed procédure with CV oï biomass and density as response variables was then performed as above in order to test for différences between the two lake groups. Biomass and density values were log-transformed to satisfy the assumptions of normality and homocedasticity of error terms. 24 Availability and spatio-temporal variation inprey diversity and size A repeated measure ANOVA using a mixed procédure was applied to evaluate lake group différences for différent classes of prey length and between months. Although repeated measures models are more typically used for a response variable measured repeatedly over time in an expérimental unit, they can also be applied to variables distributed through space (Little et al. 2001). In this analysis, the response variables were density and biomass, the expérimental unit was lake (random components) nested within lake group (treatments that are fixed component) and subject was the lake by month interaction. For each lake, individuals of ail taxa were summed for a given month within a size class. The 23 response measurements were taken at each of the five stations and were averaged prior to analysis. Since the response measurements for the 23 size classes were measured on the same community in each lake, the independence assumption on the error ternis for the ANOVA was likely violated (Littell et al. 2001). Consequently, the error term was assumed as normally distributed with constant variance, and the dependence expressed with a covariance structure of size class. Log-transformed values of biomass and density were analyzed to satisfy the normal distribution and homocedasticity of error terms assumption. The appropriate covariance structure was determined to be first order auto régressive (AR (1)) for density and first order auto régressive heterogeneous (ARH (1)) for biomass over ail covariance structures examined using the AIC and SBC (Littell et al. 2001). Ail analyses were performed using procédures available in SAS. Spatial vanability in community biomass was measured as the CV of biomass per size class, measured over ail five stations in each lake. The CV of community biomass can be expressed as: V [(Zvariance + Icovariance) / Sbiomass] where the denominator is the average community biomass of the five stations (Caldeira et al. 2005). Hère, total variance (Evariance) was the sum of the variances of community biomass for ail size class over the five stations of each lake. Total covariance (Icovariance) was the sum of covariances between ail pair of size class over the five stations of each lake. The above équation provided a tool for understanding how thèse three components (total variance, total covariance and total biomass) affect the spatial vanability of biomass production of the communities and how spatial vanability varies with size class. High total variance and 25 covariance indicate high spatial variability (Caldeira et al. 2005). The same mixed procédure was applied for measuring the CVof biomass and density as response variables. A similarity matrix was used, ANOSIM, to test for différences between zooplanktonic assemblages in terms of species composition relative to biomass and density. The resulting R statistic provided the estimated différences of mean rank between and within assemblages. Thèse assemblages were considered to be différent when similarities were greater within than those observed between. R values were interpreted as showing a strong différence at R >0.75, an intermediate différence at R >0.5, a small différence at R >0.25, and a random grouping at R =0. A similarity of percentages routine, SIMPER, was used to assess the contributions of différent taxa to observed assemblage dissimilarities. ANOSIM and SIMPER were performed using the PRIMER-E v.5 program. Calculation of dominant taxa of each lake group and by month for biomass and density response variable were performed in order to provide global estimâtes of community structure (Winkler et al. 2005). RESULTS Limnological Différences Between Lake Groups The limnological characteristics of the six study lakes are summarized in Table l.The stepwise ANOVA revealed significant limnological différences différences between lake groups and the strongest discriminator was hypolimnetic oxygen saturation in August (O2hypolimnetic; F = 204.49, P = 0.0001). Additional variables in decreasing order of explanation were: mean depth (Zmoy; F = 15.55, P =0.029), dissolved organic matter (CDOM; F = 25.50, P =0.037) and total phosphorus (TP; F = 504.27, P = 0.028). However, Student t-tests performed between iake groups on the individual variables revealed a significant différence for O2-hypolimnetic only (F = 204.49, P = 0.0001). Lakes in group A (high divergence between Lake whitefish ecotypes) were characterized by more pronounced oxygen depletion with increasing depth than lakes in group B during August, that is, during the late summer period of maximum stratification (Fig. 2). Moreover, the ANOVA with three factors (lake groups, months, years) showed no significant différence for oxygen saturation (O2-hypolimnetic) between years 2002 and 2003 in the month-by- 26 year-by-lake group interaction (F = 0.06, P - 0.815) where lakes was nested within the two lake groups. The différence between months was significant in the mixed model of monthby-lake group interaction (F - 31.02; P - 0.0001) (Fig. 2). Although thèse trends were not significant (P >0.05), group B lakes were characterized by higher mean depth, lower CDOM and lower total phosphorus compared to group A. PCA analysis also showed that hypolimnetic oxygen saturation was the best variable explaining différences between lake groups (Fig. 3). Resource Availability and Spatio-Temporal Stability No significant différence was observed between lake groups A and B, either in terms of variation of total biomass and density between months (F =5.45, P = 0.080 and F = 3.45, P = 0.137 respectively) or between lake groups only (months pooled) (F = 2.62, P = 0.181; F = 0.02, P = 0.897). We therefore performed an additional ANOVA for each variable separately (total biomass and density, log-transformed) in order to test for différences in variation between months within each lake group, and for différences in variation between lake groups within each month. For the first analysis, lake group B showed a significant différence between months where the total biomass variable (F = 15.13, P = 0.0177) was higher in August than June. (Fig. 4). For the second analysis, a highly significant différence in total biomass was observed between lake groups in August, where total biomass was higher in group B lakes (F = 33.16, P - 0.0045; Fig. 4). There was no significant différence in spatial heterogeneity in terms of the coefficients of variation between the two lake groups, either for biomass (F = 0.20, P = 0.676) or density (F = 0.0, P = 0.957). Resource Stucture and Stability A highly significant différence between the two lake groups was observed in size structure of planktonic prey communities for both biomass (F = 2.97, p <0.0001) and density (F = 1.96, P = 0.009; Fig. 5). The lake groups differed for the size classes 1.6-2.0 mm in terms of both biomass (p <0.05) and density (p <0.001). For smaller size classes, a significant différence was only observed for biomass of the 1.1 mm size class (P = 0.020). Overall, group B lakes, were characterized both by a higher biomass and density of larger prey compared to group A lakes. Moreover, the range of prey size distribution was broader in B 27 lakes relative to A. No significant temporal variation in size structure was observed since complète interactions in the model were not significant. Similarly, no significant spatial heterogeneity in ternis of coefficient of variation was observed in community size structure for both indépendant variables (total biomass, F = 0.05, P = 0.835 and total density, F = 0.27, P = 0.632). Calanoid copepods were the dominant prey type in ternis of total density and total biomass in both lake groups and at ail times. For both years, the second most abundant taxa were cyclopoid in June and Daphnia sp. in August (Fig. 6). Thèse taxa also predominated in size classes that differed significantly between lake groups (Fig. 5; Table 2). Species contribution to similarity, (using SIMPER and ANOSIM) was assessed in ternis of withinassemblage and between assemblage similarity, both between lake groups and between months in each lake group. The calculated ANOSIM R (and P in %) where - 0.078 (75.8 %) and 0.096 (19.3 %) for density and biomass respectively, indicating the absence of significant différence of taxonomic diversity between lake groups. When comparing sampling months within lake group, ANOSIM R (and P in %) were - 0.037 (80 %) and 0.259 (20 %) for density (A and B lake groups) and - 0.111 (60 %) and 0.63 (10 %) for biomass (A and B lake groups). Thus, a significant différence between month zooplanktonic assemblages for both density and biomass was observed for B type (low divergence) but not A type (high divergence) lakes. Ail SIMPER analyses showed that calanoids accounted for ail dissimilarities between assemblages as well as similarities within assemblage. DISCUSSION Compétition has been considered as the key ecological déterminant of population divergence (Robinson and Wilson 1994). Namely, interspecific compétition between closely related species may impede differentiation. In Coregonus, there is strong évidence that the présence of Lake cisco, C. artedii hampered the évolution of the dwarf Lake whitefish ecotype (Davis and Todd 1998; Trudel et al. 2001; Bernatchez 2004). Yet, the absence of this competitor has not systematically resulted in phenotypic differentiation and population divergence for occupying the pelagic niche by the dwarf ecotype (McPhail 1994 in Bernatchez et al. 1999; but also Schluter and McPhail 1992; Robinson et al. 2000). 28 Thèse latter observations suggest that factors other than a réduction of interspecific compétition are involved in the process of population divergence in whitefish. Random processes are difficult to rule out as a possible explanation. However, a pattern of repeated parallel phenotypic évolution in many lakes suggests that divergence is unlikely to occur in the absence of deterministic processes, which could include intraspecific compétition (Bolnick 2001, 2004) and/or the présence of similar ecological pressures (Smith and Skùlason 1996; Schluter 2000). Below, we address the hypothesis that différences in abiotic and biotic characteristics we observed between lakes translate into différent ecological landscapes, which in tum may influence the potential for compétitive interactions and the level of divergence observed between whitefish sympatric ecotypes. Parallelism in Abiotic Features and Levels ofEcotypic Differentiation A given Iake may vary in its physical and chemical gradients (Hindar and Jonsson 1982; Taylor 1999), and also differ from other similarly structured lakes in terms of trophic levels, morphometric parameters and other limnological features (Griffiths 1994; Wetzel 2001). Most studies hâve used a candidate approach to identify abiotic variables that could be correlated with levels of differentiation or diversification (Griffith 1994; Robinson et al. 2000; Nosil and Reimchen 2005). To our knowledge, the présent study is the first to be based on a "global variable searching" approach in order to identify the most déterminant environmental factors that may influence differentiation or the potential for compétitive interactions between sympatric forms of freshwater fishes. Such a multivariate approach allows the identification of factors that discriminate lacustrine différences separately or together, with ecotypic differentiation, and the ability to illustrate what proportion of the effect that can be attributed to each factor (Rodriguez and Magnan 1995). Hère, we identified several abiotic characteristics potentially associated with ecological opportunity and distinguishing two groups of lakes lakes harboring sympatric dwarf and normal whitefish ecotypes that evolved in parallel in each Iake and showing high and low level of phenotypic and genetic divergence, respectively (Lu and Bernatchez 1999). We observed that the three lakes harboring the most divergent sympatric populations (Cliff L., Webster L., Indian P.) were characterized by lower mean depth, higher dissolved organic matter (CDOM) and a higher total phosphorus concentration compared to those lakes harboring 29 less differentiated populations (Crescent P., East L., Témiscouata L.). However, the primary variable differentiating the two lake groups was oxygen: the Iakes harboring the most divergent populations were characterized by significantly more pronounced oxygen depletion in the hypolimnion relative to the other three Iakes during the period of highest observed thermal stratification. The observed importance of hypolimnetic oxygen depletion is particularly interesting since this is the most integrative variable that reflects primary production and microbial activities relative to lake morphometry throughout the growing season (Hutchinson 1957). It also has major implications for habitat availability. Oxygen depletion in August would likely reduce the water volume available for whitefish populations since, as for any salmonid species, coregonines need well-oxygenated water (Scott and Crossman 1974). Limnetic ecotypes of the European whitefish (C. lavaretus) are known to preferentially inhabit the lake bottom in daytime (Kahilainen et al. 2004; Mayr 2002 in Kahilainen et al. 2005) to escape predators. Previous fish studies hâve also documented a possible association between habitat size and the level of phenotypic differentiation. For instance, Robinson et al. (2000) reported that phenotypic polymorphism in the pumkinseed sunfish (Leptomis gibbosus) was more pronounced in Iakes characterized by smaller volume of open water habitat. It is noteworthy that, as in our study, total lake volume was not in itself a factor significantly linked to the observed level of phenotypic differentiation. However, an aposteriori analysis revealed that the réduction in the proportion of available habitat between 5 to 25 m depth in August relative to June, was significant only for Iakes harbouring the most differentiated sympatric whitefish forms (P <0.0001). However, the volume of available habitat was not différent between lake groups (P — 0.230). In summary, the observed parallelism observed in the extent of phenotypic divergence was paralleled by différences in limnological characteristics of the Iakes. Namely, the occurrence of the most differentiated forms of C. clupeaformis in Iakes showing the lowest oxygen saturation suggests that hypolimnetic oxygen depletion reduces habitat availability that in turn increases the potential for both intraspecific (intra-ecotypic) and/or interspecific (inter-ecotypic) compétitive interactions. 30 Parallelism in Prey Availability and Ecotypic Differentiation Just as beak morphology in relation to seed size exploitation is a key feature in the phenotypic differentiation of finches (Boag and Grant 1981; Smith 1993; Grant and Grant 2002; Benkmann 2003), many studies on the adaptive divergence of northern temperate particulate feeding fishes hâve identified gill raker characteristics as a key trait between diverging populations (Schluter and McPhail 1993; Robinson and Wilson 1994; Lu and Bernatchez 1999). Gill raker numbers and spacing is believed to détermine the efficiency of the rétention of small prey (Budy et al. 1995; Schluter and McPhail 1993; Robinson and Wilson 1994) by potentially regulating fluid dynamics in the buccal cavity (Sanderson et al. 2001). In Coregonus, Kahilainen et al. (2004; 2005) reported a link between gill-raker numbers, the length distribution of taxa preyed upon, and the available taxa in one given lake. Others hâve also associated gill raker characteristics and taxa found in stomach contents or in the environment, again in a single lake (see Hindar and Jonsson 1982; Malmquist et al. 1992; Kahilainen et al. 2003). To our knowledge, however, this study is the first to provide évidence that parallelism in the extent of phenotypic divergence among sympatric fish populationis is paralleled by différences in availability and structure of the zooplanktonic community (but see Robinson et al. 2000). Namely, our results revealed a monthly variation of biomass and taxonomic composition in lake groups exhibiting the lowest ecotypic differentiation only. The fact that sympatric ecotypes with the lowest level of differentiation occur in lakes characterized by more pronounced seasonal variation in zooplankton availability and in the planktonic taxa arrangement, potentially conferring differential energetic advantage (Oison et al. 2003), corroborâtes the hypothesis that temporal fluctuation in trophic resources availability is likely to favor less specialist phenotypes (Hjelm and Johansson 2003; Kawecki and Ebert 2004; Hindar and Jonsson 1992). This also brings support to our working hypothesis that temporal variation in resource, and potentially variation in sélective régime in thèse lakes may contribute to maintaining a lower level of differentiation between sympatric ecotypes, which tend to be phenotypically less specialized in those environments (Lu and Bernatchez 1999). 31 We also observed that lakes harboring the most divergent sympatric populations were characterized by lower prey biomass on average. Hère, the lower availability of planktonic prey and the réduction of habitat availability in August could contribute to increase the potential for compétitive interactions. For instance, Hindar and Jonsson (1992) showed that food limitation could promote ségrégation in type of exploitation, which in turn may potentially play a rôle in the intensity of population differentiation by selectively favoring the more specialized planktivorous individuals within a population in environments where food is always in limitative condition (see Bolnick 2001; 2004). The présent study also revealed that lakes harboring Lake whitefish populations with the highest level of ecotypic differentiation were also characterized by a narrower range of planktonic prey length potentially available to the dwarf ecotype. Hère, a narrower and smaller length distribution of the planktonic invertebrate community could resuit from a higher prédation rate by planktivorous species (Brook and Dodson 1965), such as the dwarf whitefish ecotype. Such a zooplanktonic structure has also previously been interpreted as évidence for resource limitation, that may increase the potential for compétition and potentially resuit in a higher relative fitness of individuals with a higher efficiency in planktonic prey rétention (Werner 1986 in Magnan 1988). This in turn, could possibly resuit in a higher level of differentiation in gill raker characteristics between sympatric dwarf and normal whitefish populations. As for the rôle of seed size in beak differentiation in finches (Boag and Grant 1981; Grant and Grant 2002; Benkmann 2003), variation in availability of resources and distribution of lengths of the planktonic prey community during the growing season are important aspects of the sélective landscape that could influence the observed level of divergence in gill rakers between sympatric whitefish ecotypes. In this way, a lower availability in range of prey size could increase sélective pressures for a more eficient exploitation of the zooplanktonic prey and thus resuit in a more pronounced specialization characterized by a higher number of gill raker numbers, as well as narrower spacing between rakers. Indeed, increase in length and number of gill rakers hâve been previously associated to an increase in planktonivorous feeding behavior (see Magnan 1988). An increased number of gillrakers would be predicted to evolve if they resulted in reduced mean spacing of gill rakers and an increased feeding efficiency on smaller prey (Trudel et al. 2001, Bernatchez 2004). 32 This hypothesis is currently being tested by means of bioenergetic modelling (Whiteley et al. in préparation). Admitedly, an évaluation of the size structure of the zoobenthic prey, as well as a characterization of the overlap between the planktonic and the benthic prey over many years will be required to more rigorously define the sélective landscape and its rôle in shaping patterns of differentiation observed in each lake between whitefish ecotypes (Benkmann 2003). Hère, we hypothetize that in lakes harbouring the most differentiated sympatric populations, the landscape of prey size distribution between zooplankton and zoobenthos will be more discontinuous then in lakes harbouring less differentiated sympatric forms. This, in turn would allow to test the hypothesis that limitation in benthic trophic resources may also promote population differentiation by increasing the advantage of exploiting planktonic prey as an alternate resource (Wilson and Turellli 1986 in Bolnick 2001). Moreover, knowledge on the ecological processes driving whitefish population divergence would be improved by comparing the limnology of ail lakes studied hère with other neighbouring ones where only the ancestral normal whitefish occurs. To conclude, this study illustrâtes the usefulness of comparing ecological features across many environments in improving our understanding of the rôle of adaptive landscape on population divergence. Hère, this approach revealed that parallelism in the extent of phenotypic divergence among sympatric fish population is associated with parallelism in adaptive landscape, both in ternis of différences in limnological characteristics, as well as availability and structure of the zooplanktonic community. This allowed to establish a link between lake metabolism and trophic interactions. Such a link produces particular niche opportunities that may be déterminant in the évolution of Lake whitefish populations, and possibly of other fish species complexes. Evolution is conspicuously lacking as a thème in limnological textbooks (Livingstone 2002) yet hère is a natural convergence between evolutionary and limnological perspectives on aquatic ecosystems. Studies of the type undertaken hère may represent a first step towards bridging the gap between evolutionary (e.g. adaptive radiation; Schluter 2000) and aquatic ecological théories (e.g. trophic cascades; Carpenter et al. 1985). 33 ACKNOWLEDGMENTS We are grateful to David Basley, Derrick Cote and Frank 0 . Frost of the Department of Inland fisheries and wildlife (Maine), as well as R. Saint-Laurent for field sampling, constructive exchange, and great cooking. We also thank R. Martel and S. Roy for assistance in the laboratory, and S. Baillargeon for assistance with statistical analyses. We also thank G. Winkler, J. Landry, S. McCairns, G. Côté, M. Lieutenant-Gosselin, A. Whiteley, A.-M. Gale, M. Dionne, S.M. Rogers, DJ. Fraser, M. Cusson, M. Rautio, L. Rétamai, P. Duchesne and J J . Dodson for their very constructive inputs. Research was supported by from grants from the Natural Sciences and Engineering Research Council of Canada (NSERC), Canada Research Chair in conservation genetics of aquatic resources to L.B and the Canada Research Chair in aquatic ecosystem studies to W.F.V. L.L. was financially supported by Québec-Océan (Groupe interinstitutionnel de recherches océanographiques du Québec). 34 TABLE 1. Characteristics of the six study lakes with high (A) or low (B) differentiation of dwarf and normal Lake whitefîsh ecotypes. Abiotic variables describe limnological characteristics of each lake (see Materials and Methods for détails on abbreviations). Biotic variables relate to planktonic invertebrates communities in each lake. Numbers identify taxa présent in each lake as well as the dominant taxon in terms of biomass and density. (see Materials and Methods for détails on taxonomic identification by numbers). Data for morphological and genetic differentiation between Lake whitefîsh ecotypes in each lake are from Lu and Bernatchez (1999). Data type Variable Month CL Lake group A IP WL CP Lake group B TL EL 8.04 3.99 1.32 6.65 17.44 17.61 6.00 11.15 2.09 10.68 8.30 19.81 5.33 2.05 294.13 46.40 69.25 7.83 21.94 1.44 6.48 9.02 9.37 6.00 2851 4.89 7.81 12.10 15.85 5.83 2.00 286.51 46.26 69.29 9.17 6.23 1.26 6.29 9.91 23.71 3.22 13.19 2.00 5.63 8.00 13.41 6.59 2.00 27158 46.16 69.08 54.24 5.53 1.03 6.47 6.62 10.18 4.75 9.25 1.25 5.94 7.30 20.73 7.38 1.84 370.64 46.42 69.61 86.04 6.88 1.76 7.45 17.76 11.28 5.25 2511.26 79.84 14.06 12.40 74.68 38.93 3.35 148.80 47.60 68.75 76.05 4.93 0.84 6.81 5.81 20.94 4.25 132.88 8.85 6.36 6.60 30.48 16.92 2.17 320.10 47.18 6955 Limnological O2-hypolimnetic (%) TP (Hg L'1) Qûa pH CID CDOM Z-Secchi Volume Surface Fetch Tmax-z Zmax Zmoy DL Altitude Latitude Longitude (^g L'1) (mg C L 1 ) (m"1) (m) (10W) (10W) (103m) (°Q (m) (m) (m) (m) 0 0 35 TABLE 1. Continuée. Invertebrate Taxa l-8;10-12 ;16;17 l-10;15-18 l-7;9-ll;13; 16;17 l-10;16;17 l-3;5-10;13 ;16;17 1-7;13;14 June August 1 (54.6) 1 (78.2) 1(81.0) 6 (36.0) 1 (51.4) 1 (39.2) 1 (83.9) 6 (39.2) 1 (59.5) 1 (59.1) 1 (58.6) 1 (57.4) June August 16 (42.6) 1 (53.8) 1 (83.6) 6 (42.8) 1 (50.0) 1(43.0) 1 (79.1) 1 (27.8) 1 (43.1) 1(41.2) 1 (58.5) 1 (45.7) 0.3-3.6 0.3-2.3 0.3-3.6 0.3-5.3 0.3-5.8 0.3-3.4 June August 0.75 (0.03) 0.79 (0.04) 0.63 (0.04) 0.86 (0.09) 0.68 (0.10) 0.67 (0.02) 0.65 (0.06) 0.81 (0.10) 0.96 (0.08) 0.88 (0.05) 0.85 (0.04) 1.02(0.14) 5 biomass mg DW m" (std) June J mg DW m (std) August 63.5 (16.0) 10.1 (3.1) 18.3 (6.8) 10.2 (4.0) 12.6 (4.8) 17.7 (6.6) 17.9 (4.9) 35.7(16.1) 26.5 (8.3) 39.8 (7.8) 18.6 (10.6) 37.2 (15.9) 15.0(2.8) 2.9(1.0) 6.4 (2.9) 3.8(1.5) 3.3(1.1) 6.1 (2.0) 5.2(1.6) 8.0 (3.0) 3.4(1.2) 6.7(1.3) 3.6(1.5) 5.4 (32) 3.003 2.089 1.435 0.974 0.796 0217 0.256 0.392 0.140 0.188 0.084 0.039 0.058 0.079 0.041 0.085 0.020 0.019 Dominant taxa biomass taxon(%) taxon(%) density taxon(%) taxon(%) Range size mm Mean size mm (std) mm (std) Total 3 3 density 10 ind m (std) j 10 ind m"J (std) fish morphology Différence fish genetic 0 Rst eigenvalue June August 36 TABLE 2. Dominant taxa in each size class (0.3 mm to up to 2.5 mm) for lake groups A and B and for each month (June and August). Panel Dominant taxa are given in terms of relative biomass (above) and relative density (below). Taxa explaining significant différences between lake groups in terms of size classes (see Fig. 5) are bolded. "."; size class not represented. Lake group Length class (mm) 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 + Taxa A-June Biomass ( %) 5 5 16 1 1 1 1 1 2 1 1 1 6 1 1 88.96 64.02 42.60 56.05 51.93 79.72 79.05 41.52 46.64 46.46 59.83 69.09 50.60 79.41 92.33 6 12 19 100.00 100.00 100.00 t Taxa A-August Biomass (%) Taxa 5 5 1 1 1 1 1 2 2 6 6 6 6 6 6 6 6 6 10 6 10 52.74 45.46 45.19 61.68 61.00 74.48 48.70 27.13 50.04 67.56 94.48 83.45 67.33 99.62 97.77 98.49 92.49 98.25 100,0 91.72 100.00 5 1 1 1 1 1 1 1 1 1 1 1 6 1 6 6 1 6 1 6 6 52.80 66.76 69.95 81.78 83.87 80.89 77.53 47.86 59.42 73.62 70.21 74.76 59.35 61.79 49.93 55.01 72.50 51.81 73.47 100.00 100.00 10 88.87 6 100.00 t • B-June Biomass ( %) m Taxa 5 5 5 5 1 1 1 1 1 6 1 1 6 6 6 6 6 1 1 6 6 6 10 B-August Biomass ( %) 77.19 50.74 40.74 28.19 31.55 86.22 87.08 66.20 47.38 48.97 73.52 61.06 80.90 61.29 72.33 81.77 66.32 53.26 70.18 99.29 98.80 98.90 96.64 37 TABLE 2. Continuée!. Lake group Lenght class (mm) 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 + Taxa A-June Density(%) 5 5 16 1 16 1 1 2 2 6 6 6 6 1 1 . 73.65 43.69 40.42 51.29 47.16 69.49 69.28 33.84 38.57 47.52 66.64 65.77 79.93 51.70 78.96 6 12 19 100.00 100.00 100.00 Taxa A-August Density(%) Taxa 1 1 6 6 6 6 6 6 6 6 6 6 6 6 10 6 10 40.79 34.89 43.53 57.90 53.55 56.92 42.55 60.52 62.85 90.38 97.32 95.51 88.91 98.29 88.88 91.79 72.73 91.70 100.00 74.49 100.00 3 1 1 1 1 1 1 2 1 1 1 1 6 6 6 6 6 6 1 6 6 58.05 77.84 78.14 79.18 79.77 75.40 65.80 45.36 47.92 60.63 57.99 56.16 76.41 65.41 71.00 78.52 52.62 75.31 51.88 100.00 100.00 10 93.45 6 86.18 5 1 1 1 m • B-June Density(%) t Taxa 5 5 2 17 17 1 1 1 6 6 6 6 6 6 6 6 6 6 6 6 6 6 10 B-August Density(%) 49.17 26.87 23.60 24.87 30.14 81.55 80.80 48.43 47.24 72.90 49.28 60.33 91.91 82.30 87.96 91.94 83.67 67.21 53.71 97.54 94.10 93.62 96.87 38 Québec (Canada) Temiscouata Lake FIGURE 1. Location map of sampling sites in the St.-John river watershed in southeastem Québec (Canada) and northeastern Maine (United States); Temiscouata Lake (TL), East Lake (EL), Crescent Pond (CP), Cliff Lake (CL), Indian Pond (IP) and Webster Lake (WL) The graphs illustrate phenotypic différences between dwarf (bold Unes) and normal (thin Unes) obtained from the discriminant function analysis of eight meristic traits (modified from Lu and Bernatchez 1999). Most of the variation was explained by gill raker number. Arrows represent the centroids of the discriminant analysis in each lake. 2002 Oxygen concentration (%) 50 100 150 0 50 150 0 2003 .10 100 FIGURE 2. Oxygen profiles of each lake nested by lake groups A and B in June and August 2002 and 2003. ANOVAs showed significant différence between lake groups by month interaction (F =0.0001). 40 oo -2 -2 -1 0 PC 1(48,7%) FIGURE 3. Principal component analysis of abiotic factors explaining différences between lake group A (high differentiation between ecotypes; triangles) and B (low differentiation between ecotypes; circles). The proportions (in percent) of the variance explained by the principal component are given within parenthèses. Open symbols represent mean and SD for each lake group. 41 B a June August g June August Month June August o m 70 60504030 20100- T 0.722 I-L I 1 ^ 1 Lake group HGURE 4. Top panels: compansons of biomass between months in lake group A (high differentiation between whitefish ecotypes) and B (low differentiation between ecotypes). Bottom panels: comparison of biomass by month (June and August) between lake groups. P values of the t-test are given in each panel. Histograms and error bars are for means and SDs. 42 Size class (mm) FIGURE 5. Comparison of size class (mm) of zooplanktonic communities in terms of biomass and density between lake groups. White bars; Lake group A (high differentiation between whitefish ecotypes), Black bars; lake group B (low differentiation). * P < 0.01, ** P < 0.005 and *** P < 0.001 for mixed ANOVAs where the repeated factor was size class (mm) and the random factor was lake nested within lake group. Back transformations of the explanatory variables and their SE are given. 43 June August 1.61.1 10.3 • 1 14.8 ,l •H ^353 M 54.3 10F F un 1.81.5 11.6// B 7.6^ 138.2 9.8^ 14.7^ 5^C—y 18.1 Biomass (%) B I H OU M B D S Calanoida Cyclopoda Nauplii (Copepoda) Holopedium spp. Bosminidae Daphnia spp. Sididae Hydracarina Asplanchna spp. Ceriodaphnia spp. Others FIGURE 6. Relative biomass and density of zooplankton by taxon in lake group A (high differentiation between whitefish ecotypes) and B (low differentiation) for each month. Values represent the percent contribution of each taxon for a month in a lake group. CONCLUSION GÉNÉRALE 45 LA SÉLECTION NATURELLE La sélection naturelle est le point central de la théorie de la radiation adaptative (Schluter 2000). Elle agit sur les traits héréditaires reliés à l'exploitation des ressources et/ou d'un habitat, où les multiples associations trait(s)-ressource(s) (ou habitat) confèrent un fitness différent. L'exemple classique est celui des pinsons des îles Galapagos qui se sont différenciés à des traits morphologiques (ex. les traits du bec) reliés à l'exploitation de ressources (ex. taille des graines) (Boag et Grant 1981; Smith 1993; Benkman 2003). Les facteurs environnementaux imposent des pressions de sélection sur les traits reliés à leurs exploitations et lors d'événements de divergence populationnelle, ils pourraient être responsable du niveau de différenciation atteint entre les populations. L'ÉVOLUTION EN PARALLÈLE Les espèces ayant subi une évolution répétée et parallèle permettent d'étudier le rôle de facteurs écologiques dans la divergence (ex. les lézards; Losos et al. 1998; les poissons; Schluter et McPhail 1993), car 1) l'histoire évolutive des populations et les phénomènes stochastiques à eux seuls rendent improbable la répétition de l'évolution des populations en parallèle qui semble plus probable en présence de facteur déterministe comme la sélection naturelle et 2) puisque l'évolution en parallèle se produit dans des environnements similaires, les variations entre les milieux semblent responsables des variations observées dans les niveaux de divergence (Schluter 1996; Lu et al. 2001; MacLean et Bell 2003; Tyerman et al. 2005). L'étude des systèmes caractérisés par des lignées évolutivement jeunes où il s'est produit une divergence phénotypique rapide, de façon répétitive et parallèle, et où l'isolation reproductive est encore en formation, tels les poissons des lacs tempérée du Nord, semble être de bons candidats pour acquérir des connaissances sur les processus de divergence populationnelle (Schluter 2000). PRÉMISSES À LA DIFFÉRENCIATION DES POPULATIONS À ce jour, les connaissances acquises dans ce domaine ont permis d'identifier des conditions probables sous lesquelles la divergence populationnelle peut se produire. 46 Le paysage de sélection Premièrement, les observations en nature et en laboratoire ont montré que la différenciation phénotypique s'effectue en lien avec l'hétérogénéité structurale sur un gradient de l'habitat ou des ressources exploitées (Rainey et Travisano 1998; Boag et Grant 1981; Benkman 2003). Par exemple, chez les pinsons, c'est la variation dans l'habitat de différentes disponibilités de graines sur un gradient de taille qui réfère à l'hétérogénéité spatiale disponible (Abbott et al. 1977; Boag et Grant 1981; Smith 1993; Benkman 2003). Ainsi, les différentes disponibilités de tailles des graines sont exploitées grâces à des variations dans la morphologie du bec. Ces gradients de disponibilité de graines en fonction de leurs tailles peuvent être vus comme un paysage de sélection où chacune des hautes disponibilités pour une taille de graines sur le gradient permet à un individu possédant un certain trait relié à son exploitation d'atteindre un fitness plus élevé (Schluter et Grant 1992; Benkman 2003). Du point de vue des poissons des lacs des régions tempérées du Nord, il semble que ce soit les caractéristiques des branchicténies, c'est-à-dire leurs longueurs, leurs nombres et leurs espacements qui augmenteraient l'efficacité de filtration et de rétention de taille de proies benthiques ou planctoniques disponibles (Schluter et McPhail 1993; Robinson et Wilson 1994; Sanderson et al. 2001). Ainsi, la présente étude illustre que la structure de taille, la disponibilité et l'avantage d'exploitation que confère différents arrangements taxonomiques des proies pourraient être perçues comme des facteurs écologiques qui exercent des pressions de sélection sur les traits reliés à l'exploitation des ressources au même titre que les caractéristiques structurales des graines provenant de différents arrangements taxonomiques de plantes (Abbott et al. 1977). Fluctuation des pressions de sélection Deuxièmement, la faible variabilité temporelle des pressions sélectives semble capitale pour la promotion et la maintenance de la différenciation des populations (Hjelm et Johansson 2003; Svanbâck et Persson 2004). Une grande variabilité dans les pressions sélectives et donc dans la disponibilité de certaines caractéristiques du gradient de ressources ou de l'habitat favoriserait les individus généralistes ou phénotypiquement plastique sur les traits reliés à l'exploitation des ressources plutôt que spécialistes ou ayant des traits fixés indépendants de l'environnement dans lequel les individus vivent (Futuyma 47 et Moreno 1988; revu par Kawechi et Ebert 2004). D'ailleurs, Hjelm et Johansson (2003) ont démontré, à la suite d'une étude sur cinq ans, que les variations dans la disponibilité des proies étaient associées à une plasticité phénotypique de la structure morphologique des branchicténies. En ce sens, il est possible que des fluctuations dans la disponibilité des ressources au cours d'une seule saison de croissance permettent d'identifier des facteurs qui pourraient être la source du degré de divergence observé entre les écotypes d'une population. LES GRANDSCORÉGONES Dans cette étude, le grand corégone provenant de six lacs du bassin versant de la rivière StJean a été utilisé comme modèle puisqu'à l'intérieur de ses populations l'ampleur de la différenciation génétique et morphologique observée entre les paires sympatriques d'écotypes nains et normales est différente. Au sein de ces populations, les branchicténies sont reconnues comme étant le trait morphologique discriminant le plus les écotypes (Lu et Bernatchez 1999). D'ailleurs, une étude a permis de montrer que les branchicténies étaient aussi soumis à la sélection directionnelle mais seulement chez l'écotype nain (Bernatchez 2004). Ces informations, ajoutées aux observations des contenus stomacaux renseignant sur la diète naine et normale respectivement et principalement constitué de zooplancton et de zoobenthos (Bernatchez et al. 1999), ont permis d'envisager que les caractéristiques des ressources zooplanctoniques était potentiellement les principaux facteurs écologiques responsables des pressions de sélection sur les branchicténies. Conséquemment, les communautés zooplanctoniques pourraient constituer l'opportunité écologique alternative ou le principal paysage de sélection chez les populations de grands corégones. Les lacs étant reconnu pour leurs caractéristiques similaires, telle la présence d'un habitat planctonique et épibenthique pour les poissons ainsi que d'un gradient physique et chimique (voir introduction), ont aussi permis d'envisager que certaines caractéristiques de l'habitat pouvaient favoriser ou limiter la différentiation des populations. OBJECTIFS ET RÉSULTATS Cette étude portait sur la caractérisation abiotique et biotique et la variation temporelle dans ses caractéristiques dans des lacs possédant une population naine et une normale de grand 48 corégone vivant en sympatrie. L'hypothèse principale était de vérifier si les différences, dans les patrons de différenciations observés dans la structure morphologique des branchicténies chez les nains et les normaux, étaient associées à une variation dans les caractéristiques de l'opportunité écologique, soit les caractéristiques abiotiques et biotiques observées. Réduction des ressources disponibles et différenciation populationnelle élevée En somme, cette étude a permis de montrer que les caractéristiques de l'opportunité écologique étaient différentes dans les différents lacs. Premièrement, les lacs possédant des populations avec une différence élevée entre les écotypes montraient une diminution des ressources zooplanctoniques et de l'oxygène dans l'hypolimnion au mois d'août. Ces résultats ont permis d'émettre l'hypothèse voulant que cette diminution de l'oxygène dans l'hypolimnion puissent résulter en une diminution de l'habitat disponible pour les grands corégones à ce moment. Cette diminution de l'habitat et des ressources pourrait affecter la compétition pour les ressources entre les planctonivores et ainsi appuyer l'hypothèse voulant que la compétition pourrait être potentiellement responsable de la promotion de la divergence populationnelle. D'ailleurs, une étude similaire sur la compétition pour les ressources entre les écotypes de crapets soleil a, elle aussi, établi une corrélation entre l'augmentation du niveau de divergence phénotypique entre les écotypes et la réduction du volume de l'habitat disponible (Robinson et al. 2000). Par contre, dans la présente étude, aucune observation de la diversité en espèces de poissons potentiellement compétitives et de leurs densités n'a été effectuée dans chacun des lacs et aucun relevé antérieur a été fait ou est disponible pour tous les lacs. Ainsi, il est impossible d'émettre plus qu'une hypothèse concernant une pression de compétition potentiellement plus élevée dans les lacs où la divergence est élevée entre les écotypes par rapport aux lacs où la divergence est plus faible. Variation temporelle et différenciation populationnelle faible Deuxièmement, les lacs possédant des populations avec une différence faible entre les écotypes ont montré une variation temporelle dans la biomasse des proies zooplanctoniques plus élevée ainsi qu'une fluctuation plus élevée dans les arrangements taxonomiques au 49 niveau de la densité et de la biomasse des proies. Ces résultats semblent aller de pair avec l'hypothèse voulant que la fluctuation dans la disponibilité des ressources et ainsi des pressions de sélection pouvait favoriser les individus généralistes plutôt que spécialistes, ou la plasticité phénotypique (Kawecki et Ebert 2004). Chez le grand corégone, les études au laboratoire, en conditions contrôlées, ont montré que le nombre de branchicténies avait une base génétique (Rogers et al. 2002). Ces expériences n'ont pas été effectuées sur toutes les populations des lacs de cette étude. Cependant, en admettant que ces résultats soient véridiques pour tous les lacs, les individus ayant des caractéristiques intermédiaires pour l'exploitation des proies disponibles pourraient être favorisés dans un habitat où les ressources varient en disponibilité au niveau temporel. Spécialisation et paysage de sélection Finalement, une distribution de taille des proies disponibles plus étroite et comprenant de plus petite proie a été retrouvée dans les lacs où la différenciation entre les écotypes observés est élevée. Cette distribution pourrait représenter un paysage de sélection car Bernatchez (2004) a démontré que le nombre de branchicténies est sous l'effet de la sélection chez les nains. Par contre, pour le reconnaître, il faudrait aussi effectuer une étude sur l'augmentation de l'efficacité de conversion énergétique avec l'augmentation du nombre de branchicténie (Whiteley et al. en préparation) et accepter l'hypothèse voulant que l'augmentation du nombre de branchicténies facilite effectivement la rétention de proies de plus petite taille (voir Sanderson et al. 2001). La confirmation de ces hypothèses indiquerait que les nains sont soumis à une pression de sélection plus forte dans les lacs où la différenciation est élevée et que celle-ci favorise les individus plus spécialisés pour l'exploitation de petites proies. À l'inverse, la distribution plus étendue de tailles de proies disponibles comprenant de plus grosses proies favoriserait les individus plus généralistes, ce qui expliquerait probablement le patron de divergence observé dans les lacs où la différenciation entre les écotypes est faible. Interaction écologique et paysage de sélection Ce qui est intéressant dans ce concept de paysage de sélection potentielle est que plusieurs études, dont celle réalisée par Brooks et Dodson (1965), ont observé un patron similaire de 50 la distribution de taille des proies zooplanctoniques dans des lacs où la prédation sur les communauté zooplanctonique est élevée comparativement à ceux ou une pression de prédation faible a été observé. Ceci a donc permis d'avoir une plus grande confiance en l'idée que la compétition serait effectivement plus élevée dans les lacs possédant de plus petite taille de proies zooplanctoniques et que cette compétition imposerait des pressions sur les communautés de proies. En somme, un effet rétro-actif (prédateurs-proies) pourrait augmenter la pression de sélection sur les individus moins spécialisés sur les proies de petites tailles. Cet effet laisse aussi supposer, par la théorie des cascades trophiques (Carpenter et al. 1985; Tessier et Woodruff 2002), que des particularités de l'environnement abiotiques de chacun des milieux permettraient de favoriser des événements de co-évolutions différentes entre les grands corégones et leurs proies dans des environnements possédant des aspects différents et ainsi amener l'évolution de degré de différenciation différent. NUANCE ET HYPOTHÈSES ALTERNA TIVES Paysage de sélection Un des aspects à envisager pour soutenir les observations de cette étude serait premièrement de considérer le paysage complet «de sélection» pour les populations de grand corégone dans chacun des lacs, c'est-à-dire d'effectuer la quantification et la caractérisation de la niche épibenthique. Plusieurs études ont montré que la limitation de la niche la plus utilisée et donc l'augmentation de la compétition intraspécifique pour les ressources pouvait résulter en une divergence populationnelle pour l'exploitation d'une ressource alternative (Bolnick 2001; 2004). Dans le cas présent, une limitation de la niche épibenthique plus prononcée dans les lacs où la différenciation entre les écotypes est élevée pourrait appuyer cette hypothèse. D'un autre côté, il est possible que la différence entre le niveau de différenciation entre les écotypes des différents lacs réside dans la différence entre la distribution de l'étendue des tailles des proies zooépibenthiques et zooplanctoniques dans chacun des lacs. En fait, il est plausible qu'une distribution plus similaire des tailles du zooplancton et du zoobenthos limite l'étendue de la différenciation observée entre les écotypes et qu'à l'inverse une distribution de tailles des proies plus différentes entre les deux habitats favorise la différenciation. De plus, bien que ce paysage 51 ait été établi pour les pinsons dans deux études (Schluter et Grant 1992; Benkman 2003) aucune étude à ce jour n'a envisagé de le faire pour mieux comprendre les facteurs qui influencent le degré de divergence inter-écotypique dans des populations naturelles. Composition taxonomique Un des aspects intéressants de cette étude est que la variable biomasse semble montrer des différences significatives entre les groupes de lacs (soit des lacs possédant une différence, entre les écotypes, élevées ou faibles) contrairement à la variable de densité de proies. Il y a par contre, une exception, c'est lorsqu'on observe la distribution des tailles des proies. Ces résultats semblent anodins a priori, mais sachant que la biomasse est calculée à l'aide de régression longueur-poids qui font référence à la composition taxonomique, il est plausible que tous les résultats appuyant la distinction entre les lacs au niveau de la disponibilité des ressources soient potentiellement reliés à une composante taxonomique. De plus, l'exception concernant la distribution des tailles des proies peut être expliquée par la relation entre la composition taxonomique des communautés zooplanctoniques et leurs compositions en taille des proies (Rodriguez et Magnan 1993). La composition taxonomique peut aussi influencer la quantité d'énergie qu'obtient un prédateur lorsqu'il s'alimente sur un taxon en particulier. En effet, le temps de manipulation de la proie et son contenu calorique sont des exemples d'éléments qui font intervenir l'avantage qu'a un prédateur dans un certain environnement plutôt qu'à un autre et/ou à un moment donné (voir aussi : Pike 1984; Futuyma et Moreno 1988). C'est pourquoi, il serait intéressant d'effectuer une étude plus approfondie de l'alimentation des grands corégones dans les différents lacs à l'étude et de reconstruire un paysage de sélection à l'aide des proies sur lesquelles les grands corégones s'alimentent et à l'aide des contenus caloriques des proies plutôt qu'avec leur biomasse. Ces dernières informations permettraient potentiellement de voir un paysage de sélection plus précis que celui dépeint dans cette étude. Maintenance versus promotion de la différenciation Cette étude porte sur les facteurs qui maintiennent actuellement la différenciation morphologique observée dans l'étude de Lu et Bernatchez (1999). Ces facteurs sont 52 probablement liés à la présence d'une opportunité écologique dans chacun des lacs de l'étude puisque dans chacun de ceux-ci il y eu divergence. Il est donc possible que ces facteurs soient impliqués dans les événements ayant promu la divergence dans le passé. Par ailleurs, la comparaison avec des lacs où aucune divergence s'est produite en l'absence de cisco de lac pourrait être envisagée pour répondre à cette question et ainsi mieux caractériser les facteurs constituant de l'opportunité écologique. De plus, bien que cette étude suggère des facteurs qui pourraient être impliqués dans l'obtention des deux niveaux de différenciation entre les écotypes, elle n'explique pas pourquoi dans les lacs différenciés les normaux sont encore présents bien que l'oxygène dans l'hypolimnion soit réduit à la fin de l'été, ou encore pourquoi il existe deux écotypes dans les lacs ou il y a des fluctuations saisonnières dans la disponibilité des ressources. En ce sens, il est vrai que la majorité des variables abiotiques observées entre les groupes de lacs étaient similaires et qu'il puisse s'agir de composantes déterminantes dans l'établissement des caractéristiques environnementales requises pour l'observation d'événements de divergence populationnelle chez les poissons et donc de l'opportunité écologique. En somme, 1) la caractérisation du paysage de sélection complète incluant les aspects spécifiques de spécialisation alimentaires observées au niveau des diètes des grands corégones et 2) des études similaires à celle-ci dans des lacs où il n'y a pas eu de divergence populationnelle même en l'absence de cisco de lac, pourraient permettre de mieux comprendre le rôle des facteurs écologiques dans les processus d'évolution de la divergence populationnelle chez les grands corégones. 53 BIBLIOGRAPHIE Abbott, I., I. K. Abbott, and P. R. Grant. 1977. Comparative ecology of Galapagos ground finches (Geospiza Gould) : évaluation of the importance of floristic diversity and interspecific competiton. Ecol. Monogr. 47 :151-184. Amundsen, P.-A., T. Bonn, and G. H. Vaga. 2004a. Gill raker morphology and feeding ecology of two sympatric morphs of European whitefish (Coregonus lavaretus). Ann. Zool.. Fennici 41:291-300. Amundsen, P.-A., R. Knudsen, A. Klemetsen and R. Kristoffersen. 2004b. Ressource compétition and interactive ségrégation between sumpatric whitefish morphs. Ann. Zool.. Fennici 41:301-307. Belzile, C , J. A. E. Gibson, and W. F. Vincent. 2002. Colored dissolved organic matter and dissolved organic carbon exclusion from lake ice: Implications for irradiance transmission and carbon cycling. Limnol. Oceanogr. 47:1283-1293. Benkman, C. 2003. Divergent sélection drives the adaptive radiation of crossbills. Evolution 57:1176-1181. Bernatchez, L. 2004, Ecological theory of adaptive radiation: an empirical assessment from coregonine fishes (Salmoniformes): Evolution Illuminated : Salmon and Their Relatives. Oxford University Press: New York. Bernatchez, L., A. Chouinard, and G. Lu. 1999. Integrating molecular genetics and ecology in studies of adaptive radiation: whitefish, Coregonus sp., as a case study. Biol. J. Linn. Soc. 68:173-194. Boag, P. T. and P. R. Grant. 1981. Intense natural sélection in a popualtion of Dawin's ground finches (Geospizinae) in the Galapagos. Science 214:82-85 Bolnick, D. I. 2001. Intraspecific compétition favours niche width expansion in Drosophila melanogaster. Nature 410:463-466. . 2004. Can intraspecific compétition drive disruptive sélection? An expérimental test in natural populations of sticklebacks. Evolution 58:608-618. Brainerd, E. L. 2001. Caught in the crossflow. Nature 412:388. Brooks, J. L., and S. I. Dodson. 1965. Prédation, body size, and composition of plankton. Science 150:28-35. 54 Budy, P., T. Haddix, and R. Schneidervin. 2005. Zooplankton size sélection relative to gill raker spacing in Rainbow trout. T. Am. Fish. Soc. 134:1228-1235. Caldeira, M. C , A. Hector, M. Loreau, and J. S. Pereira. 2005. Species richness, temporal variability and résistance of biomass production in a Mediterranean grassland. Oikos 110:115-123. Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985. Cascading trophic interactions and lake productivity. BioScience 35:634-639. Chiba, S. 2004. Ecological and morphological patterns in communities of land snails of the genus Mandarina from the Bonin Islands. J. Evolution. Biol. 17:131-143. Chouinard, A., D. Pigeon, and L. Bernatchez. 1996. Lack of specialization in trophic morphology between genetically differentiated dwarf and normal forms of Lake whitefish (Coregonus clupeaformis Mitchill.) in Lac de lEst, Québec. Can. J. Zool. 74:1989-1998 Culver, D. A., M. M. Boucherie, D. J. Bean, and J. W. Fletcher. 1985. Biomass of freshwater crustacean zooplankton from length-weight régressions. Can. J. Fish. Aquat.Sci. 42:1380-1390. Davis, B.M. and T. N. Todd. 1998. Compétition between larval Lake herring (Coregonus artediï) and Lake whitefish (Coregonus clupeaformis) for zooplankton. Can. J. Fish. Aquat. Sci. 55(5):1140-1148 Day, T. and K. A. Young. 2004. Competitve and facilitative evolutionary diversification. BioScience. 54(2):101-109 Derôme, N., P. Duchesne and L. Bernatchez. 2006. Parallelism in gène transcription among sympatric Lake whitefish ecotypes (Coregonus clupeaformis Mitchill.). Molecular Ecology 15: inpress Dieckmann, U., and M. Doebeli. 1999. On the origin of species by sympatric speciation. Nature 400:354-357. Dumont, H. J., I. Van de Velde, and S. Dumont. 1975. The dry weight estimate of biomass in a sélection of Cladocera, Copepoda, and Rotifera from the plankton, periphyton and benthos of continental waters. Oecologia 19:75-97. Futuyma, D. J. and G. Moreno. 1988. The évolution of ecological specialization. Annu. Rev. Ecol. Syst. 19:207-233 55 Grant, P. R. 1994. Ecological character displacement. Science. 266(5186) :746-747. Grant, P. R., and R. B. Grant. 2002. Unpredictable évolution in a 30-year study of Darwin's finches. Science 296:707-711. Griffiths, D. 1994. The size structure of lacustrine Arctic charr (Pisces, Salmonidae) Populations. Biol. J. Linn. Soc. 51:337-357. Harder, L. D. 2001. Mode and Tempo of species diversification. Am. J. Bot. 88:17071710. Hendry , A. P., and E. B. Taylor. 2004. How much of the variation in adaptive divergence can be explained by gène flow: an évaluation using lake/stream stickleback. Evolution 58:2319-2331. Hindar, K., and B. Jonsson. 1982. Habitat and food ségrégation of dwarf and normal Arctic charr (Salvelinus alpinus) from Vangsvatnet Lake, Western Norway. Can. J. Fish. Aquat.Sci. 39:1030-1045. Hjelm, J., and F. Johansson. 2003. Temporal variation in feeding morphology and sizestructured population dynamics in fishes. Proc. R. Soc. Lond. B 270:1407-1412. Hutchinson, G. E. 1957. A treatise on limnology. Wiley: New York. USA. Jeffrey, S. W., and N. A. Welschmeyer. 1997, Spectrophotometric and fluorometric équations in common use in oceanography: Phytoplankton pigments in oceanography: guidelines to modem methods. UNESCO: Paris, France. Johnston, T. A., and R. A. Cunjak. 1999. Dry mass-length relationships for benthic insects: a review with new data from Catamaran Brook, New Brunswick, Canada. Freshwater Biol. 41:653-674. Kahilainen, K., E. Alajârvi, and H. Lehtonen. 2005. Planktivory and diet-overlap of densely rakered whitefish (Coregonus lavaretus (L.)) in a subarctic lake. Ecol. Freshw. Fish 14:50-58. Kahilainen, K ., T. Malinen, A.Tuomaala and H. Lehtonen. 2004. Diel and seasonal habitat and food ségrégation of three sympatric Coregonus lavaretus forms in a subarctic lake. J. Fish Biol. 64(2):418-434 Kahilainen, K., H. Lehtonen, and K. Kononen. 2003. Conséquence of habitat ségrégation to growth rate of two sparsely rakered whitefish {Coregonus lavaretus (L.)) forms in a subarctic take. Ecol. Freshw. Fish 12:275-285. Kawecki, T. J., and D. Ebert. 2004. Conceptual issues in local adaptation. Ecol. Lett. 7:1225-1241. Laçasse, S., and P. Magnan. 1992. Biotic and abiotic déterminants of the diet of Brook trout, Salvelinus fontinalis, in lakes of the Laurentian Shield. Can. J. Fish. Aquat. Sci. 49:1001-1009. Langeland, A., and T. Nost. 1995. Gill raker structure and sélective prédation on zooplankton by particulate feeding fish. J. Fish Biol. 47:719-732. Langerhans, B. R., and T. J. Dewitt. 2004. Shared and unique features of evolutionary diversification. Am. Nat. 164. Leibold, M. A. 1995. The niche concept revisited - mechanistic models and community context. Ecology 76:1371-1382. Littell, R. C , G. A. Milliken, W. W. Stroup, and R. D. Wolfinger. 2004, SAS System for Mixed Models. SAS Institute Inc: Cary, North Carolina. Livingstone, D. 2002. Teaching limnology in the 21 st century: two new texts. Limnol. Oceanogr. Bull. 11(1): 1-3. Losos, J. B., 2000. Ecological character displacement and the study of adaptation. P. Nat. Acad. Sci. USA. 97(11):5693-5695. Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and L. Rodriguez-Schettino. 1998. Contingency and determinism in replicated adaptive radiations of island lizards. Science 279:2115-2118. Lu, G., and L. Bernatchez. 1999. Correlated trophic specialization and genetic divergence in sympatric Lake whitefish ecotypes {Coregonus clupeaformis): support for the ecological speciation hypothesis. Evolution 53:1491-1505. Lu, G., D. J. Basley and L. Bernatchez. 2001. Contrasting patterns of mitochondrial DNA and microsatellite introgressive hybndization between lineages of Lake whitefish (Coregonus clupeaformis); relevance for speciation. Mol. Fxol. 10(4):965-985 MacLean, R., and G. Bell. 2003. Divergent évolution during an expérimental adaptive radiation. Proc. R. Soc. Lond. B 270:1645-1650. Magnan, P. 1988. Interactions between Brook charr, Salvalinus fontinalis, and non salmonid species-ecological shift, morphological shift, and their impact on zooplankton communities. Can. J. Fish. Aquat. Sci. 45(6):999-1099 57 Malmquist, H. J., S. S. Snorrason, S. Skulason, B. Jonsson, O. T. Sandlund, and P. M. Jonasson. 1992. Diet différenciation in polymorphic Arctic charr in Thingvallavatn, Iceland. J. Anim. Ecol. 61:21-35. Martinson, D. G., N. G. Pistas, J. D. Hays, J. Imbrie, T. C. Moore and N. J. Shackleton. 1987. Age, dating and orbital theory of the Ice Ages: development of a high resolution 0-300,000 year chronostratigraphy. Quaternary Res. 27:1-29. Meyer, E. 1989. The relationship between body length parameters and dry mass in running water invertebrates. Arch. Hydrobiol. 117:191-203. Nosil, P., and T. E. Reimchen. 2005. Ecological opportunity and levels of morphological variance within freshwater stickleback populations. Biol. J. Linn. Soc. 86:297-308. Nusch, E. A. 1980. Comparison of différent methods for chlorophyll and phaeopigments détermination. Arch. Hydrobiol. Beih. Ergebn. Limnol. 17:14-36 Oison, N. W., C. P. Paukert, D. W. Willis and J. A. Klammer. 2003. Prey sélection and diets of bluegill, Lepomis macrochirus with differing population characteristics in two Nebraska natural lakes. Fish. Mana. Ecol. 10(1): 31-40 Pigeon, D., A. Chouinard, and L. Bernatchez. 1997. Multiple modes of speciation involved in the parallel évolution of sympatric morphotypes of Lake whitefish {Coregonus clupeaformis, Salmonidae). Evolution 51:196-205. Pike, G.H. 1984. Optimal foraging theory-A critical review. Annu. Rev. Ecol. Syst. 15:523575 Rahkola, M., J. Karjalainen, and V. A. Avinsky. 1998. Individual weight estimâtes of zooplankton based length-weight régressions in Lake Ladoga and Saimaa Lake System. Nordic J. Freshw. Res. 74. Rainey, P. B. and M. Travisano. 1998. Adaptive radiation in a heterogeneous environment. Nature. 394 (6688):69-72 Robinson, B. W., and D. S. Wilson. 1994. Character Release and Displacement in Fishes a Neglected Literature. Am. Nat. 144:596-627. Robinson, B.W., and D. Schluter. 2000. Natural sélection and the évolution of adaptive genetic variation in northern freshwater fishes. Pp. 65-94 in, T. A. Mousseau, B. Sinervo, and J. A. Endler (eds.), Adaptive genetic variation in the wild. Oxford University Press: New York. USA. 5S Robinson, B. W., D. S. Wilson, and A. S. Margosian. 2000. A pluralisme analysis of character release in Pumpkinseed sunfish (Lepomis gibbosus). Ecology 81:27992812. Rodriguez, M. A., and P. Magnan. 1993. Community structure of lacustrine macrobenthos do taxon-based and size-based approaches yield similar insights. Can. J. Fish. Aquat. Sci. 50:800-815. . 1995. Application of multivariate analyses in studies of the organization and structure of fish and invertebrate communities. Aquat. Sci. 57:199-216. Rodriguez, M. A., P. Magnan, and S. Laçasse. 1993. Fish species composition and lake abiotic variables in relation to the abundance and size structure of cladoceran zooplankton. Can. J. Fish. Aquat. Sci. 50:638-647. Rogers S. M. and L. Bernatchez 2005. Integrating QTL mapping and génome scans towards the characterization of candidate loci under parallel sélection in the Lake whitefish (Coregonus clupeaformis) Mol. Ecol. 14(2):351-361 Rogers, S. M., V. Gagnon, and L. Bernatchez. 2002. Genetically based phenotypeenvironment association for swimming behavior in Lake whitefish ecotypes (Coregonus clupeaformis Mitchill). Evolution 56:2322-2329. Ruzzante, D., S. Walde, V. Cussac, P. Macchi, M. Alonso, and M. Battini. 2003. Resource polymorphism in a Patagonian fish Percichthys trucha (Percichthyidae): phenotypic évidence for interlake pattern variation. Biol. J. Linn. Soc. 78:497-515. Sanderson, S. L., A. Y. Cheer, J. S. Goodrich, J. D. Graziano, and W. T. Callan. 2001. Crossfow filltration in suspension-feeding fishes. Nature 412:440-441. Schluter, D. 1996. Ecological causes of adaptive radiation. Am. Nat.l48:S40-S64. . 2000, The Ecolôgy of Adaptive Radiation, Oxford University Press: Oxford, U.K. Schluter, D., and P. R. Grant. 1984. Déterminants of morphological patterns in communities of Darwin's finches. Am. Nat. 123:175-196. Schluter, D. & McPhail, J. D. 1992 Ecological character displacement and speciation in sticklebacks. Am. Nat. 140:85-108. Schluter, D., and J. D. McPhail. 1993. Character displacement and replicate adaptive radiation. Trends Ecol. Evol. 8:197-200. 59 Scott, W. B. and E. J. Crossmann. 1974. Poissons d'eau douce du Canada. Ministère de l'environnement, Service des pêches et des sciences de la mer: Ottawa, Canada. Simpson, G. G. 1953, The Major Features of Evolution. Simon and Schuster: New York, USA. Smith, T. B. 1993. Disruptive sélection and the genetic basis of bill size polymorphism in the African finch Pyrenestes .Nature.363:618-620 Smith, T. B., and S. Skùlason. 1996. Evolutionary significance of resource polymorphisms in fishes, amphibians, and birds. Annu. Rev. Ecol. Syst. 27:111-133. Saint-Laurent R., M. Legault, L. Bernatchez. 2003. Divergent sélection maintains adaptive differentiation despite high gène flow between sympatric Rainbow smelt ecotypes (Osmerus mordax Mitchill). Mol. Ecol.. 12: 315-330. Svanbâck, R., and L. Persson. 2004. Individual diet specialization, niche width and population dynamics: implication for trophic polymorphisms. J. Anim. Ecol. 73:973-982. Taylor, E. B. 1999. Species pairs of north temperate freshwater fishes: Evolution, taxonomy, and conservation. Reviews in Fish Biology and Fisheries 9:299-324. Taylor, E.B., and Bentzen, P. 1993. Evidence for multiple origins and sympatric divergence in trophic ecotypes of smelt (Osmerus) in northeastern North America. Evolution 47: 813-832 Taylor, E. B., J. D. McPhail. 2000. Historical contingency and ecological determinism interact to prime speciation in sticklebacks, Gasterosteus. Proc. R. Soc. Lond. B. 267:2375-2384. Tessier, A. J., and P. Woodruff. 2002. Cryptic trophic cascade along a gradient of lake size. Ecology 83:1263-1270. Thorp, J. H., and A. P. Covich. 1991, Ecology and Classification of North American Freshwater Invertebrates. Académie Press, Inc.: London. Travisiano, M., J. A. Mongold, A. F. Bennett, and R. E. Lenski. 1995. Expérimental tests of the rôles of adaptation, chance, and history in évolution. Science 267:87-90. Trudel, M., A. Tremblay, R. Schetagne, and J. B. Rasmussen. 2001. Why are dwarf fish so small? An energetic analysis of polymorphism in Lake whitefish (Coregonus clupeaformis). Can. J. Fish. Aquat. Sci. 58:394-405. 60 Tyerman, J., N. Havard, G. Saxer, M. Travisano, and M. Doebeli. 2005. Unparallel diversification in bacterial microcosms. Proc. R. Soc. Lond. B 272:1393-1398. Vamosi, S. M., and D. Schluter. 2002. Impacts of trout prédation on fitness of sympatric sticklebacks and their hybrids. Proc. R. Soc. Lond. B 269:923-930. Weckstrom, J., A. Korhola, and T. Blom. 1997. The relationship between diatoms and water température in thirty subarctic Fennoscandian lakes. Arctic Alpine Res. 29:75-92. Wetzel, R. G. 2001, Limnology, Saunders: Philadelphia, USA. Wetzel, R. G., and G. E. Likens. 1979, Limnological Analyses. Saunders: Philadelphia, USA. Williamson, C. E., D. P. Morris, M. L. Pace, and O. G. Oison. 1999. Dissolved organic carbon and nutrients as regulator of lake ecosystems: résurrection of a more integrated paradigm. Limnoi. Oceanogr. 44:795-803. Winkler, G., P. Sirois, L. E. Johnson, and J. J. Dodson. 2005. Invasion of an estuarine transition zone by Dreissena polymorpha veligers had no détectable effect on zooplankton community structure. Can. J. Fish. Aquat. Sci. 62:578-592.