Grégory DANIEL
Interaction between dispersal and behavioural syndromes – empirical approch in a fragmented
population of passerine birds
Thesis supervised in co-direction by :
In University of Lyon 1 (France) : Blandine DOLIGEZ, Chargée de Recherche au CNRS In University of Uppsala (Sweden) : Lars Gustafsson, Professor in EBC
15 décembre 2015
Jury's Members :
Ingrid AHNESJÖ (Professeur en Écologie Animale, Université d'Uppsala) - Examinateur Dominique ALLAINE (Professeur des Universités, Université Lyon 1) – Examinateur Thierry BOULINIER (Directeur de Recherche, CNRS) - Rapporteur
Blandine DOLIGEZ (Chargée de Recherche, CNRS) – Directrice de Thèse
Arnaud GRÉGOIRE (Maître de Conférences, Université de Montpellier) - Examinateur Lars GUSTAFSSON (Professeur en Biologie, Université d'Uppsala) – Co-Directeur de la Thèse Anne LOISON (Directrice de Recherche, CNRS) – Rapporteur
UNIVERSITE CLAUDE BERNARD - LYON 1
Président de l’Université
Vice-président du Conseil d’Administration
Vice-président du Conseil des Etudes et de la Vie Universitaire Vice-président du Conseil Scientifique
Directeur Général des Services
M. François-Noël GILLY
M. le Professeur Hamda BEN HADID M. le Professeur Philippe LALLE M. le Professeur Germain GILLET M. Alain HELLEU
COMPOSANTES SANTE
Faculté de Médecine Lyon Est – Claude Bernard
Faculté de Médecine et de Maïeutique Lyon Sud – Charles Mérieux Faculté d’Odontologie
Institut des Sciences Pharmaceutiques et Biologiques Institut des Sciences et Techniques de la Réadaptation
Département de formation et Centre de Recherche en Biologie Humaine
Directeur : M. le Professeur J. ETIENNE Directeur : Mme la Professeure C. BURILLON Directeur : M. le Professeur D. BOURGEOIS Directeur : Mme la Professeure C. VINCIGUERRA Directeur : M. le Professeur Y. MATILLON
Directeur : Mme. la Professeure A-M. SCHOTT
COMPOSANTES ET DEPARTEMENTS DE SCIENCES ET TECHNOLOGIE
Faculté des Sciences et Technologies Département Biologie
Département Chimie Biochimie Département GEP
Département Informatique Département Mathématiques Département Mécanique Département Physique
UFR Sciences et Techniques des Activités Physiques et Sportives Observatoire des Sciences de l’Univers de Lyon
Polytech Lyon
Ecole Supérieure de Chimie Physique Electronique Institut Universitaire de Technologie de Lyon 1 Ecole Supérieure du Professorat et de l’Education Institut de Science Financière et d'Assurances
Directeur : M. F. DE MARCHI
Directeur : M. le Professeur F. FLEURY Directeur : Mme Caroline FELIX Directeur : M. Hassan HAMMOURI
Directeur : M. le Professeur S. AKKOUCHE Directeur : M. le Professeur Georges TOMANOV Directeur : M. le Professeur H. BEN HADID Directeur : M. Jean-Claude PLENET Directeur : M. Y.VANPOULLE Directeur : M. B. GUIDERDONI Directeur : M. P. FOURNIER Directeur : M. G. PIGNAULT
Directeur : M. le Professeur C. VITON
Directeur : M. le Professeur A. MOUGNIOTTE Directeur : M. N. LEBOISNE
Interaction entre dispersion et syndromes comportementaux – approche empirique ans une population fragmenté de passereaux.
Résumé
La dispersion est un trait d'histoire de vie clé pour les processus écologiques et évolutifs dans les populations naturelles. Les dernières recherches se sont notamment focalisées sur les corrélations entre traits comportementaux et la dispersion, ceci afin de mettre en évidence des syndromes comportementaux de dispersion, tout en démontrant la base génétique de la dispersion. Les dispersants ne seraient donc pas une part aléatoire d'une population, mais des individus montrant des stratégies particulières qui augmenteraient leur chances de succès.
Cette thèse s'est orientée vers trois objectifs de recherche majeurs. Le premier est la mise en évidence d’une base génétique de la probabilité de disperser dans une population fragmentée de gobe-mouches à collier Ficedula albicollis. Les résultats nous ont montré, au- delà de l'estimation de la base génétique de la dispersion, une distribution spatiale non aléatoire de l’apparentement dans cette population, qui pourrait être dû à des effets génétiques sur les règles de décision de choix de l’habitat. Le deuxième s'intéresse à la corrélation phénotypique et génétique entre le comportement de dispersion natale et le comportement de défense du nid, chez le martinet alpin Tachymarpis melba. Nous avons montré que la dispersion natale et le comportement de défense du nid sont négativement corrélés au niveau phénotypique mais aussi génétique dans ces populations. Enfin, le troisième nous à conduit à tester l’existence de syndromes comportementaux de dispersion, c’est-à-dire si les dispersants présentent un profil comportemental particulier, leur permettant en particulier de coloniser de nouveaux sites, chez le gobe-mouche à collier.
Mot-Clés : dispersion, syndromes comportementaux, héritabilité, traits de personnalité, néophobie, agressivité, prise de risque, gobemouche à collier, Ficedulla albicollis.
Summary
Dispersal is a key like history trait for ecological and evolutionary processes in wild population. The last researching particularly focused on the correlation between behavioural trait and dispersal, in order to emphasize the existence of behavioural syndromes of dispersal, and on the estimation of the genetic basis of the dispersal behaviour. Dispersant individuals could not be a random part of the population, but individuals showing particular strategies, that help them to succeed in their dispersal attempt.
This thesis has three main aims of research. The first is to show a genetic basis of the dispersal propensity in a fragmented population of collared flycatchers (Ficedulla albicollis).
We shown not only the genetic bases of the dispersal, but also a non-random spatial distribution of relationship between individuals in this population, that might be due to genetic effects on the decision rules of habitat choice in this population. The second aim concerns phenotypic and genetic correlation between the natal dispersal and a behavioural trait, the nest-defense behaviour, in the alpin swift (Tachymarpis melba). We shown that natal dispersal and nest-defense behaviour are negatively correlated at a phenotypic level, but also at a genetic level in theses populations. Finally, the third aim attempt to test the existence of behavioural syndrome of dispersal, that is if dispersant individuals have a particular behavioural profile, which enable them to colonize new sites, in the collared flycatcher.
Keywords : dispersal, behavioural syndromes, heritability, personality traits, neophobia, aggressiveness, risk-taking behaviour, collared flycatcher, Ficedulla albicollis.
Address for Thesis Work :
In France:
Université de Lyon, F-69000, Lyon, France Université Lyon 1, F-69622, Villeurbanne, France.
CNRS, UMR5558, Laboratoire de Biométrie et Biologie Evolutive, F-69622, Villeurbanne, France.
In Sweden:
Uppsala University
Institutionen för ekologi och genetik, Zooekologi Evolutionsbiologiskt Centrum, EBC, Norbyv. 18D
SE-752 36 UPPSALA, Sweden
Remerciements
Il est souvent dit que le travail de thèse peut être long, parfois fastidieux et éreintant durant certaines périodes. Il arrive que certaines thèses le soient encore plus et s'apparentent alors à une traversée du désert. Mais au final, j'en retiendrai que c'est une aventure menée par la curiosité, très enrichissante professionnellement, mais aussi personnellement. Je tiens donc à remercier toutes celles et ceux qui m'ont permis d'arriver jusqu'au bout.
Tout d'abord, je remercie mes directeurs de thèse Lars GUSTAFSSON (Suède) et Blandine DOLIGEZ (France) pour m'avoir donné l'opportunité de faire cette thèse. J'insisterai sur la patience, la compréhension, la disponibilité, l'écoute, les coups de fouet aussi (au sens figuré bien sûr), de Blandine, sans qui cette thèse n'aurait pas aboutie, du moins, pas en 2015.
Malgré son emploi du temps de ministre, inhérent à son travail et à son investissement dans celui-ci, elle a su être présente, et je l'en remercie vraiment. Et je la remercie encore plus pour tout le temps qu'elle m'a accordé lors des dernières semaines d'analyses et de rédaction, temps qu'elle a parfois pris sur ce qui aurait dû être des temps de repos. Merci.
Je remercie mes rapporteurs - Thierry BOULINIER, Anne LOISON et Denis RÉALE - pour leur compréhension et leur patience. Et je remercie plus généralement tous les chercheurs qui se sont investis dans la réalisation de cette thèse, les membres de mon jury – en sus de ceux déjà cités : Dominique ALLAINE (examinateur), Ingrid AHNESJÖ (examinateur) et Arnaud GRÉGOIRE (examinateur) - et les membres de mon comité de thèse - Julien COTE et Jean-Paul LÉNA.
Je remercie aussi la Région Rhône Alpes pour mon financement de thèse, et la bourse CMIRA. De même que je remercie l'Evolutionsbiologiskt centrum pour le financement de mon année de thèse en Suède.
Je continue avec toutes les personnes qui m'ont aidé quand je fus dans un trou sans fond durant une très longue période. Je remercie tout d'abord Blandine BONNE, qui est à mes côtés depuis longtemps déjà, qui a fait preuve d'énormément de patience et d'écoute, et qui m'a épaulé tout au long de cette thèse, et m'a permis de garder espoir. Je remercie aussi les personnes qui m'ont aidé à réfléchir et m'ont donné les outils pour poursuivre mon travail
(vive les schémas heuristiques!) : ma coach et amie, Cécile FOISSEY, mais aussi Catherine BOITEUX. Toutes deux ont participé activement à mon long rétablissement psychologique.
Et je leur en serai toujours redevable, même si elles diront aisément que j'ai été le premier acteur de mon « retour ».
Je remercie aussi ma famille qui n'a pas cessé d'être présente, même de loin et pour qui, en partie, j'ai terminé cette thèse. Que voulez-vous ? Tenter de rendre ses parents fiers de vous n'est pas forcément passé de mode. Je remercie aussi ma sœur, qui par son flegme, m'a toujours réconforté avec l'idée que quoiqu'il arrive, la vie continuerait.
Je remercie évidemment aussi tous mes amis de la région lyonnaise avec qui j'ai passé de super bons moments au LBBE et en dehors : Adrien, Emilie, Marine, Paf, Aurélie pour n'en citer que quelques-uns. Je remercie aussi Charlotte et Bertrand, toujours présents pour déguster l'une des meilleures pizzas de Villeurbanne ou se faire une cession ornitho pour aller trouver nos amis les piafs. Merci encore aux amis récents de la Haute-Marne qui ont bien pensé à moi lors des derniers jours de galère. Et un merci très particulier à mon suédois préféré, Emil, qui m'a ravi les oreilles avec sa Nyckelharpa et enchanté mon palet avec les dégustation de vin maisons et de whisky de tout horizon, sans parler des soirées saunas-bières.
Je remercie aussi tous les étudiants qui sont venus sur le terrain et qui m'ont aidé à récolter les données comportementales, et tous les étudiants qui ont participé à l'analyse des fichiers audio et vidéos, car cela prend un temps fou et demande aussi de la minutie. Merci pour votre travail, sans lequel, les données comportementales utilisées dans cette thèse et dans les manuscrits à venir, n'auraient pas pu être aussi nombreuses.
Merci également aux membres du secrétariat et du pôle informatique du LBBE, sans qui il serait parfois difficile d'avancer. J'en profite aussi pour remercier Matts BJÖRKLUND de l'Evolutionsbiologiskt centrum de l'Université d'Uppsala
Table of Contents
General Introduction 9
Materiel & Methods 14
Chapter 1 Heritability of between-patch dispersal propensity in a
passerine bird: a role for individual habitat selection rules? 25
Chapter 2 Negative phenotypic and genetic correlation between natal dispersal propensity and nest defence behaviour in a wild bird 59
Chapter 3 Do colonizing individuals show specific behavioural profiles?
A small-scale experiment testing for dispersal behavioural syndromes in a natural bird population
79
General Discussion and Perspectives
99
Bibliography 104
Appendixes 110
Appendix 1 : Food supplementation mitigates differences in nest-defence behaviour between dispersing and non-dispersing individuals in a passerine bird
111
Appendix 2 : Doligez, B., Daniel, G., Warin, P., Pärt, T., Gustafsson, L., Réale, D. (2012). Estimation and comparison of heritability and parent–offspring resemblance in dispersal probability from capture–recapture data using different methods: the Collared Flycatcher as a case study. Journal of Ornithology, 152(S2), 539–554. doi:10.1007/s10336-010-0643-4
129
General Introduction
Dispersal is defined as the movement of an individual from its birth site – natal dispersal - or previous breeding site to a new breeding area –breeding dispersal. (Greenwood and Harvey 1982). Dispersal is one of the key life-history trait in evolutionary and ecological processes in natural populations. Dispersal has consequences on population dynamics, gene flows and the distribution of species across both space and time (Clobert et al. 2001, 2009;
Ronce 2007). In a world of global climatic changes and local habitat degradations (Kokko and López-Sepulcre 2006), dispersal is increasingly discussed in studies for its importance in conservatory biology,. Indeed, dispersal is one of the main ways for individuals to escape environmental changes and their deleterious consequences on the animals’ survival and/or reproduction. Whether the aim is conservating species or understanding of the consequences of this trait on populations, dispersal has started to be under scrutiny both from a theoretical and empirical point of view. This thesis is a part of this scientific work to understanding dispersal.
Dispersal: proximate and ultimate causes.
A dispersal event comprehends three different stages. First comes emigration, which is the decision to leave the current site. Second, transience takes place the movement of the individual between the current patch and the new one,. And finally, immigration, the settlement of the individual in a new site (Clobert et al. 2009; Ronce 2007). During each of these stages, many factors can influence the dispersal behaviour (Bowler and Benton 2005;
Clobert et al. 2009). Firstly ultimate factors i.e. factors that can induce the evolution of the dispersal within a meta-population are considered. They are mostly identified by theoretical models (reviewed in Bowler & Benton 2005). They can be spatial-temporal variations in the habitat (McPeek and Holt 1992), social interactions between conspecific or heterospecific individuals. These contacts can be positive (cooperation) or negative (competition for food, or predation). Kin interaction like inbreeding avoidance, or cooperation between individuals which are sharing a part of their genes (Lambin et al. 2001) is also a factor
Secondly, proximate factors are examined, which are the elements influencing an individual’s choice to disperse. The external factors shape condition-dependant dispersal.
They often are the mirror of ultimate causes. For example, an individual’s decision to disperse can be triggered,, by environmental abiotic factors (Dries Bonte et al. 2008; Massot et al.
2002); the density of conspecifics at the site, that can entail a competition between individuals
for resources; the predation/parasitism pressure; or, when cue recognition exists between kin, the avoidance of kin or the search of it in order to cooperate with them.
For a long time dispersal has been considered highly flexible, allowing the individuals to react in an optimal way to external cues. However, an increasing number of studies focusing on the individual’s capacity of dispersal based on their phenotypic traits, (i.e. their internal state) it became obvious that dispersal is also phenotypic-dependent. These phenotypic traits based on differences in dispersal behaviour are of different nature:
morphological, behavioural, physiological and in life-history traits. For instance, in insects and plants, and less commonly in vertebrates, several studies have shown that dispersant and non-dispersant individuals present different morphs. This difference can be obvious such as the presence of winged and apterous individuals within a same population of insects, or more subtle like a difference in body condition or body size (Belthoff and Dufty 1998; O’Riain et al. 1996). Physiological differences have also been shown like corticosteroid (Belthoff and Dufty 1998) or testosterone (Holekamp 2003) levels being more important in dispersant individuals than in non-dispersant ones. Finally, individuals can show consistent differences in their behaviour, such as aggressiveness, exploration, activity or proneness to take risk(reviewed in Cote et al. 2010). For instance, the most active individuals can be the most dispersant too (e.g. O’Riain et al. 1996; Belthoff & Dufty 1998; Bonte et al. 2004).
Concerning aggressiveness, some studies have shown that dispersant individuals can be more aggressive than non-dispersant ones (e.g. Trefilov et al. 2000; Duckworth & Badyaev 2007), but others showed the opposite result (e.g. Guerra & Pollack 2010).
Amongst these phenotypic traits correlated to dispersal, morphological factors were often used in studies to show the genetic determinism of the dispersal (Roff and Fairbairn 2001) and gave us the first elements of response regarding the genetic basis of dispersal.
Dispersal : a behaviour partly genetically determined.
For a trait to evolve, it has to fulfill three criteria : i) to be variable within individuals or within populations, ii) to impact the fitness if individuals (survival and/or reproduction), iii) to have a genetic basis, that is to present a certain amount of heritability.
Regarding the first criterion, several studies have shown the high level of variability of the dispersal behaviour between individuals or between populations (Matthysen et al. 2005).
Concerning the link between dispersal and individual fitness, research studies didn't conclude on a clear relation. Some have shown a positive relation between dispersal and fitness, and others a negative relation (reviewed in Belichon et al. 1996). However, even if the link
between dispersal and individual fitness is not clear, the existence of such a link has been frequently demonstrated. Indeed, in their study of 2008, Doligez & Pärt concluded that amongst the 133 studies (the most recent in 2008) in which a component of individual fitness was compared between dispersant and non-dispersant individuals, 65 have shown the existence of such link.
To test the existence of a genetic basis of dispersal, different kind of experiments were conducted for example line selection on a trait which increases dispersal capacity of individuals. This kind of experiment has been run mainly on insects and plants (Roff and Fairbairn 2001). but also vertebrates species(mostly on birds) Another type of studies has been led to show the heritability of the dispersal behaviour, the within-sibling resemblance in dispersal distance or dispersal propensity (e. g. Newton & Marquiss 1983; Ims 1989; Léna et al. 1998; Matthysen et al. 2005; Sharp et al. 2008). However, these two kinds of studies don’t let us to estimate a heritability level of dispersal. The use of parents-offsprings regression and then animal model, allows us to do so. Parents-offsprings regression gave us the first estimation of the heritability of the dispersal behaviour (Doligez and Pärt 2008). Nevertheless this regression is based solely on resemblance in dispersal between parents and their offsprings, using only a small part of the pedigree of the population. On the contrary,
“Animal model” are mixed models that used all the pedigree of the population, allowing to split up the total phenotypic variance of a trait into additive genetic variance and environmental variance (Kruuk 2004; Wilson et al. 2010). In such models, all relationships between individuals in a population are considered to estimate the heritability of a trait, and not only relationship between parents and offsprings. Moreover, by considering all relationships, this kind of analysis takes consanguinity into account within the population.
And finally, these models can allow us to estimate variation of the dispersal trait which is due to other factors, like environmental effects.
Finally, a few studies have shown a direct link between dispersal behaviour and genes.
Haag et al. (2005) have shown the relationship between the allelic frequencies of a gene coding for a metabolic enzyme (pgi) and the flight metabolic rate, both factors that are important in a newly installed and isolated population of Glanville fritillary butterfly. Trefilov et al. (2000) have highlighted the link between the polymorphism of a promotor region of a gene coding for a serotonin transporter with aggressiveness in a rhesus macaque population.
Finally Sinervo & Clobert (2003) have demonstrated that males of side-blotched lizards behave differently in their dispersal movements depending on their color morphs which is a trait coded by OBY locus. Therefore, it seems that dispersal is genetically based.
Cost of dispersal, and meanings to reduce it
Dispersal can be a way for individuals to increase their fitness by escaping environmental constraints which reduce their survival or their reproduction in their current habitat. Nevertheless, it is also a behaviour that can entail important energetic costs and mortality risk. Indeed, during transience within non-suitable habitat for the species, individuals can be exposed to more predation or a lack of available food (Yoder et al. 2004).
However, selection on dispersal could have favoured interactions between dispersal and traits that reduce costs associated with dispersal, during its emigration, transience or immigration stages. One of the possible interactions between dispersal and a phenotypic trait could be through a behavioral trait. The personality trait concept is relatively recent and is defined as behavioural differences between individuals that are constant in the time and across context (Gosling 2001; Sih et al. 2004; Réale et al. 2007). For instance, an individual that is more aggressive than another one in a specific context, stays more aggressive than this same individual in a different context. Several example of consistent behavioural differences between dispersal have been shown: in aggressiveness (Verbeek et al. 1996), in exploration (Dingemanse et al. 2002), in boldness (Fraser et al. 2001; Bize et al. 2012), in activity (Sih et al. 2003) or in social behaviour.
Personality traits and dispersal
A few of these behaviours have been genetically correlated with dispersal. For instance, in a study published in 2007 Duckworth & Badyaev showed that the colonisation of a new breeding site in western bluebird had been facilitated by the aggression levels of the new settlers. They also demonstrated that, despite being a heritable factor in western bluebird, these aggression levels dropped in the years following the settling event. Another example can be found in populations of great tit: in 2013 Korsten et al. underlined the strong genetic link (rG= 0,9 ± 0,40 s.e.) between level of exploratory behaviours and how far individuals would disperse.
These examples highlight the fact that there could be a function integration between dispersal and another trait – a behaviour for example, which translates into a genetic correlation between dispersal and behaviours. The individuals who decide to disperse are therefore not a random sample of the population but actually a subgroup presenting a specific behavioural profile regarding dispersal: a behavioural dispersal syndrome.
Objectives of this thesis
This thesis focused on three main research objectives. The first one concentrates on putting in evidence the genetic component at play in the probability of dispersing in a fragmented population of collared flycatcher Fideculla albicollis. This population has been followed since 1080, and individuals could be identified by their rings: we therefore had access to a large number of dispersal events – either natal or reproductive, as well as the social pedigree, which allowed us to know how individuals were related. Using mixed models of quantitative genetic (“animal model”), we estimated how heritable natal and global dispersions (natal and reproductive dispersions) are, while also keeping into account other sources of variances, in particular the influence of the environment such as temporal or spatial variances.
The second objective aims at uncovering the phenotypic and genetic correlation between natal dispersion and a behavioural trait (the nest defence behaviour) in alpine swift Tachymarpis melba. Using the same types of animal models in two different populations for which we have natal dispersion data and measures of levels of nest defence, we tested the correlation between natal dispersion and nest defence behaviour at the phenotypical and genetic levels.
Finally, the third objective consists in unravelling the existence of behavioural syndromes of dispersion in collared flycatcher i.e. whether the dispersing individuals present a specific behavioural profile which allows them to colonise new sites more easily. We set up new breeding areas near existing sites in our study population, and measured aggressiveness, neophobia and boldness in new settlers, during the colonisation event and two years later once the new population installed. We compared the scores of the individuals who colonised the new sites with the ones of the birds who didn’t as to test the existence of a behavioural profile associated with colonisation. We will present here the preliminary results.
Materiel and Methods
Study population and site
This research work has been conducted on a population of collared flycatcher (Fidedulla albicollis Temm.), little passerine bird (c.a. 13 g) from muscicapidae family. It's a migratory bird spending the winter period south of Sahara desert in Africa. Its breeding area goes mainly from the East of France to Russia. We worked on an insular population, separated from the main breeding area, on the Swedish island of Gotland which is in south of Baltic sea (57°10'N, 18°20E) (Gustafsson 1989; Pärt and Gustafsson 1989).
The study site is localized in the southern part of the island, and it is composed of forests, mainly with oaks, suitable for reproduction of collared flycatcher, and of fields and coniferous forests non-suitable for reproduction of our hole-nesting bird. The formers are separated from each other by the lasts (Gustafsson and Nilsson 1985; Gustafsson 1987). So, the site is fragmented and a long-term monitoring has been conducted on 15 forest patches since 1980 (see figure 1). In each forest, nest boxes were installed at height of 1.50 meters approximately. The nest box model is easily accepted by flycatcher to breed. These nest boxes are distant from 20 meters to 50 meters, from each other. And every monitored patches contains 20 to 150 nest boxes, depending of their area. Other species which breed in cavities too are in competition with the collared flycatcher, for breeding site. The average proportion of nest boxes occupied is 40% by collared flycatchers, 25% to 30% for great tit (Parus major) and 10% to 15% for blue tits (Cyanistes caeruleus). These two species are also competitors of the collared flycatcher, regarding food resource needed for the brood feeding. Chicks eat mainly caterpillars and other larvae. Because of patches are saturated of nest boxes and because of the very low number of natural cavities, almost all breeding pairs in monitored patches are monitored during the successive years.
Population monitoring and individual identification
Males are the firsts to arrive on breeding site, around the end of April. Females follow just behind. After the first cues of male presence in the study site (song of males), every nest box has been regularly visited (every 2 or 3 days), during all the breeding period. This monitoring allowed us to watch precisely the nest building, the egg laying, the egg hatching, the chicks feeding and the chicks fledgling. So, breeding data, like clutch size and date, the chicks' condition and their number, and the number of fledglings individual, were easily
obtained. Every individual born or bred in the study site were followed during all their lifetime by an individual identification, ringing with an aluminium ring (from the Museum of Natural History of Stockholm). Chicks were ringed in the nest when they were 8 days old.
They fly out of the nest when they were 16-17 days old. Parents were captured and identified during different period of the reproduction: females were captured since incubation period, whereas males were captured during the chicks feeding (females were captured a second time during this same period). The capture at incubation time was made by hand capture of the female when it is on the nest, incubating, or with a wire trap installed in the nest box, on the entrance of the nest box. This type of trap, light and thinly frightening for the female, cannot be used during the feeding period because of the presence of chicks which could trigger the trap easily. During this period, a flap trap was used. Captured adults were either already ringed and quickly identified, or they were not. These lasts were immigrants in our study population and were 35% of breeding adults approximately, each year (Doligez et al. 2004).
These individuals were ringed to identify them after.
Each year, a part of the flycatcher population was not identified for 4 possible reasons:
i) they were non-breeders, ii) they bred in a natural cavity, iii) their breeding attempt had failed early in the breeding season, iv) male of the breeding pair was polygynous.
Indeed, the collared flycatcher is a facultative polygynous species. In our population, it has been estimated 50% of males attempt to have a second territory to attract a second female, and among them, only 10% succeed (Gustafsson 1989). However, when they succeed, they don't put a lot effort into the breeding with the secondary female, including the chicks feeding, except if the secondary nest is closed to the former one, and the two breeding attempts are staggered in time. Consequently, these males are difficult to catch (Doligez et al. 1999).
In addition of this polygynous phenomenon, some level of extra-pair paternity has been observed in our population: 15% of chicks within 30% of nests. This is in the normal limits found in passerine bird population (Sheldon and Ellegren 1999).
Behavioural measurements
Three behaviours have been measured in the study population between 2011 and 2014:
aggressiveness, neophobia and risk taking behaviour. Each behavioural test has been conducted at different stages of the reproduction of flycatchers pairs.
Figure 1: Map of the study site in the southern part of the island of Gotland. In red, the 15 long-term monitored patches (AL, TU, AN, FP, FO, FK, GR, BJ, BS, OJ, RO-PS, BO, RUW, RUE, FA). In blue, the 9 new plots which were installed in 2011 (RM, TB, BK, KT, SB, RN, OL) and in 2013 (RS, SK).
Aggressiveness
Aggressiveness tests have been done when breeding pair was constructing the nest, when individuals, specially male, have to defend their territory against potential competitor for this same territory. Aggressiveness against a conspecific individual decrease after the beginning of incubation (Kral and Bicik 1989; Kral 1996; Kral et al. 1996), whereas aggressiveness against a heterospecific individual, like the great tit, lasts all the breeding period (Král and Bičík 1992). To test both of aggressive behaviours (against conspecific and heterospecific), two kinds of dummy were used: flycatcher pair dummy and one male great tit. For tests with flycatcher dummies, we used one male dummy and one female dummy in order to elicit response of both members of the breeding pair. Dummies are made with grey clay, painted to have similar colour of feathering than birds they represent (see figure 2).
The great tit dummy was installed on the hole entrance of the nest box, whereas, for flycatchers dummies, one was installed on the hole entrance of the nest box, and the other on the roof of the nest box. The position of each flycatcher dummy (male/female) was defined randomly for every test. To the visual stimuli, a sound stimuli was added: a playback of male song associated with the dummy species. The audio broadcast, composed with a small MP3 player and small loudspeakers, was hidden under the nest box (see figure 3). Because we had to do a lot of tests during a small period of time (100 tests per day at the maximum), five sets of each species were used. We also used five different male song playbacks of each species.
The repartition of each dummy set (one male flycatcher, one female flycatcher and one male great tit) was defined per patch and per day, in such way, one set of dummies was not used two days successively in the same patch, or by a the same observer.
Figure 2: Pictures of painting clay dummies used for aggressiveness tests: a) dummy of male great tit; b) dummies of flycatcher pair, female (above) and male (below).
A same nest was tested four times during the nest building period. For the first day, the dummy species was determined randomly. The second day, it was the unused species in the first day, that we took for the test. The third day, we did nothing in order to avoid the familiarization of the breeding pairs with dummies or the test itself. The fourth and fifth day, one test was done, and the dummy species was determined in the same way than the two first testing days (see figure 4). During the test, the observer was hidden under a camouflage net at a distance between 5 meters to 10 meters from the nest box, 7 meters on average, depending on the density of the vegetation. The goal was to optimize the sight of ongoing actions around and on the nest box, and, in the same time, to avoid a response of the breeding pair due to the presence of the hidden observer. The different tests that were conducted, shown that birds didn't detect the observer under the camouflage net, as long as this last was wearing dark clothes, and was not speaking too loud (At several times, individuals of different bird species went to a branch distant from less than one meter of the hidden observer).
A pre-test phase has been conducted on the 30 first nest of flycatcher in 2011 to determine the optimal length of aggressiveness test. Following these pre-tests, the length of
Figure 3: Pictures of the different experiment designs installed on nest boxes during behavioural test: a) aggressiveness test, with a male flycatcher dummy on the roof of the nest box, a female flycatcher dummy on the entrance hole, and the audio broadcast system hidden under the nest box; b) the risk-taking test, with a stuffed red squirrel fixed on the entrance hole of the nest box; and c) neophobia test, with the new object (hockey player) fixed at the right of the entrance hole.
the test was fixed to 15 minutes. This was a sufficient last to catch the maximal response of each birds in the dummy presence.
During the test, the observer recorded on a voice recorder: i) each location change of each individual of the breeding pair, ii) the distance from the nest box where individual were (on the nest, less than 2 meters, between 2 meters and 5 meters, between 5 meters and 10 meters, more than 10 meters), iii) the alarm calling and their frequencies (one call isolated or continuous alarming), iv) hovering over the nest or dummies, v) direct attacks on dummies, by skimming or pecking, and vi) the presence of other individuals (flycatcher, tits or others species) which could have influenced the behavioural response of the breeding pair. When the test ended, the observer took his test material (voice recorder, camouflage net), removed the dummies and the audio broadcast from the nest box. He walked away quickly to disturb birds as less as possible with a visible human presence.
Neophobia
Neophobia tests took place during the chicks feeding, when they were 5 days old.
During this period, the need in food by chicks is the most important, and parents must put the maximal effort into the chicks feeding. Neophobia was tested once per nest, and the behaviour
Figure 4: Here, we present every possible combinations of successive dummies used (great tit or flycatcher pair) in time for the 4 tests of aggressiveness which were run on each nest. At day 1 and day 4, dummy species is determined randomly, whereas at day 2 and day 5, it was the dummy species unused the day before.
was recorded with a video camera. Testing neophobia of individuals is testing their fear from a new object in a common environment (Garamszegi et al. 2009 ; Greenberg and Mettke- Hofmann 2001 ; Mettke-Hofmann et al. 2002). Here, a small blue and red figurine of a hockey player was used. It measures around 7 cm in height (see figure 3). Because of the possible variation in food provisioning between individuals and the variation in response to the novel object, the test was run in two phases. During the first phase, normal behaviour of the breeding pair was recorded, without the novel object. This phase served us as a comparing basis with the second phase. During this second phase of the test, the novel object was installed on the nest box. The video camera was put as distant as possible from the next box, while optimizing the image quality, between 6 meters and 10 meters from the nest box generally. The image framing focused on the nest box, this last one occupying two-third of the image, in order to have the possibility to clearly differentiate the male (black and white) from the female (brown).The observer started the recording, then he faked the installation of the figurine in order to standardize the time spent near the nest box between the two phases of the test. Finally, he walked away quickly from the nest box (more than 100 meters). Since one hour passed, the observer came back, controlled the video camera (battery level, possible modification of the image framing), did needed adjustments, and pin the figurine on the nest box to be placed at the right side of the entrance hole. He walked away during one hour again, then came back to stop the recording and took off the figurine and the recording material.
When we viewed the video recordings, the noticed behaviours are, for each phase; i) the time of bird returning to the nest, ii) the entry occurrences in the nest box, iii) the number of hovering over the figurine or the nest box, iv) the number of attacks on the figurine.
Risk taking behaviour
The test was run when chicks were near to fly out of the nest and were 13 and 14 days old. At this advanced stage of the reproduction, parents have a lot to loose if the breeding attempt fail. Indeed, they already put lot of effort to raise their brood. The risk-taking behaviour was observed in a context of nest-defense with the presence of a nest predator on the nest box.
To simulate the presence of a nest predator, a stuffed red squirrel or a stuffed great spotted woodpecker was fixed on the entrance hole of the nest box (see figure 3). To do all tests, two specimens of each nest predator species were used The species and the number of the specimen used for a test were determined per patch and per day, in such way that every nest predator specimen were used an equivalent time in each patch and were not used 2 days
successively in the same patch. Doing like this, we tried not only to minimize the possibility that birds of a same patch become used to a particular nest predator specimen, but also to randomize the use of each nest predator specimen in each patch. And if a red squirrel specimen was used for the first day of the test, so a great spotted woodpecker specimen was used for the second day of the test, and inversely.
The stuffed great spotted woodpeckers were severely damaged by birds attacks during the first year of test. So, they are removed from the specimen used the next years and replaced by 3 new stuffed red squirrels. After 2011, risk-taking tests has been conducted only with stuffed red squirrel, and the repartition per patch and per day followed the same randomization method did in 2011 when we had both red squirrels and great spotted woodpeckers specimens.
Pre-tests have been conducted in 2011 on the 20 first nests which reach the necessary stage in our population. They last 15 minutes. During these test, we had strong reactions of breeding pairs, very quickly after the observer went under his camouflage net (sometimes, few seconds after). Therefore, the time length of the test was reduced to 5 minutes after the arrival of every member of the breeding pair (male and female), with a maximal length of 10 minutes. Pursuing the test beyond this time can be traumatizing for presents birds, including chicks under the nest, because of the possible high strength of parental response (personal observations). To avoid a premature exit of chicks out of the nest box, the entrance hole was blocked. However in 2012, it was a good year for the flycatcher reproduction. The 13 days old chicks were more advanced in their development than usual. And we observed some decreasing of the brood between the first and the second test, let us think the departure of some chicks from the nest. After this observation, and because preliminary statistical tests running on data of the past year shown that the individual behaviour was repeatable between the two tests, it was decided to run only test per nest, to avoid possible early departure of chicks from the nest.
Behavioural response was recorded on a voice recorder, by direct observation of birds, under a camouflage net, at a distance of 5 meters to 10 meters from the nest box, depending of of the vegetation density nearby. The observer begun by preparing his material in order to start the test as quick as possible after fixing the nest predator specimen on the entrance hole of the nest. When the stuffed nest predator was fixed, and the observer was under the camouflage net, the test started. Different behaviours were recorded by the observed: i) the distance from the nest where the individual was (on the nest, less than 2 meters, between 2 meters and 5 meters, between 5 meters and 10 meters, more than 10 meters), ii) each
movement of individuals, iii) hovering over the nest-predator specimen, iv) pecking and skimming on the nest predator specimen.
At the end of the test, the observer put in his bag all the recording material and took the nest predator specimen off the nest box, as quickly as possible, and walk away from the nest box, and avoided to do two successive test on nest which were distant of less 50 meters between each other.
Colonization experiment
In order to test the hypothesis of the existence of a colonization behavioural profile in our population, we ran a colonization experiment. During the beginning of the breeding season, new nest boxes were installed in suitable patches for flycatcher reproduction which were empty of nest boxes before that. These new patches are at the direct proximity (less than 1 kilometre) of the long-term monitored study site (“old patches” hereafter). Nest boxes are the same than those in old patches. They were fixed at the same density and the same height than those in the old patches. The new patches have similar environmental characteristics (tree densities and species) with those in the old patches, with a mix of patches of high tree density and of open patches (änge).
In 2011, 8 new patches were installed (see figure 1). At the end of the breeding season, it appeared that one new patch had a lot of flycatchers density throughout the reproduction period (visual observations or alarm callings and songs), that was not representative with the number of breeding pairs in our nest boxes. This patch had a lot of old trees, and we deducted it offered a lot of natural cavities, more than in old patches, for the flycatcher reproduction.
We decided to remove it at the end of the breeding season, and behavioural tests on nest boxes installed in this patch were discarded from analysis. In 2012, to increase our sample size, 3 new patches were installed. Among these ones, no breeding pairs was observed in one patch.
It was also discarded from the analysis. Finally, between 2011 and 2012, 9 new patches, with 20 to 60 nest boxes each, were installed and included in the analysis. The population of collared flycatchers in their patches were monitored and behavioural tests were conducted on nest box of breeding pairs in the same way than in old patches.
Nest boxes were installed before the arrival of the first flycatchers. After fixing the nest box on a tree, the hole entrance was blocked for every nest box, in order to prevent the installations of tits pairs which would have given social information on the patch to flycatcher, and would have reduced the number of available nest box for the colonization by
flycatchers. In the two days after the first sign of flycatcher arrival on the breeding site (by direct observations or hearing male song), nest boxes were unblocked.
Study on an alpine swift population: quick presentation of the study site and species.
The study of the second chapter was conducted on the Swiss population of alpine swift (Tachymarptis melba L.), in collaboration with Pierre BIZE from the university of Aberdeen (United-Kingdom) who managed the monitoring of this population during years when data have been collected. We are going to present quickly the study population and site, and the measurement of the nest-defense behaviour. You will be able to find complementary information in chapter 2 of this thesis and the cited publications therein.
The alpine swift is a migratory bird, breeding in colonies on rock cliffs or high buildings. During the breeding season, its area is limited to the southern Europe and to the south west of Asia. Adults come back in their previous breeding site and do only one clutch per year, with 1 to 4 eggs (Bize et al. 2006, 2012). They are monogamous and parental investment in the reproduction is the same for both male and female of the breeding pair.
Youngs are sexually mature at 2 to 3 years old (Tettamanti et al. 2012), and they breed for the first time mainly in their birth site.
The study population is in Switzerland, in two colonies distant of 21 km from each other, in clock towers of Bienne (47°10’N, 7°12’E) and of Solothurn (47°12’N, 7°32’E), with 50 to 100 breeding pairs per year approximately. A monitoring was done regularly during the breeding season, every year. Parents were caught during incubation or during chicks feeding.
They are individually identified by a numbered ring, measured, and ringed if they were previously not. Chicks were ringed in the nest when they are 10 days old.
The natal dispersal is determined in a binary way, that is individuals ringed as chicks in a nest within a study colony, and bred for the first time in this same colony are defined as non-dispersant individuals (so called philopatric). Inversely, those which didn't breed in the same colony than their birth colony are defined as dispersant individuals. As all adults are philopatric in our population, immigrants, that are unringed, in the 2 colonies were considered as dispersant individuals for the natal dispersal in our analysis.
Between 2013 and 2014, during the adult captures, the nest-defense behaviour of breeding pairs was measured by only one observer, following a shyness-boldness gradient, with 5 levels from 0 to 5, with a step of 0.5. This measure is the result of the reaction of each individual facing a human approach and hand capture (Bize et al. 2012; Patrick et al. 2013).
The minimal score (0) was given to individuals which flushed since they saw the observer
approaching, whereas the maximal score (2) was given to indidivuals which moved towards observer, flapped their wings and clawed the hand of the observer during hand capture (details in chapter 2 and in Bize et al. 2012).
