Netbiome Frag&Binv
Fragmentation & biological invasions

Consequences of forest fragmentation and conditions for biological invasions:
the case of Caribbean birds

Transformation rate of natural ecosystems due to human activities has recently increased. Habitat fragmentation and biological invasions are major threats on biodiversity, as both are main processes responsible for populations and species declines.

Despite a large body of literature focusing on the impact of fragmentation or biological invasions on species abundance and diversity, changes in ecological and evolutionary processes due to these two global changes remain poorly understood.

This project aims to assess effects of fragmentation on several attributes of individuals/populations in a set of bird species showing a gradual specialization on forest habitat :

  • genetic diversity, due to demographic changes (smaller and more isolated populations)
  • phenotypic quality of individuals (i.e. morphological, ornamental, immunological, physiological stress)
  • host-parasite interactions

In addition, we will test two recent hypotheses explaining the success of biological invasions (the so-called enemy release hypothesis and the hypothesis of different immune defense strategies in invaders), and investigate some of their consequences for native species.

This project will be conducted in four territories: French Guiana, Guadeloupe, Martinique, and Montserrat. If these territories host a high endemic biodiversity, they go through real problems of forest loss and fragmentation as well as species introductions or invasions, mainly due to human activities and demographic growth.

In addition, to contribute to a better knowledge of ecological and evolutionary consequences of habitat fragmentation and biological invasions, the integrated approach of this project will produce valuable results for decision rules in nature and wildlife management frameworks.

The collaboration of scientific partners and partners concretely involved in local conservation plans ensures the integration of research and wildlife management.



Human activities have been transforming much of natural ecosystems in the world for a long time, but this process has recently accelerated. As habitat modifications due to human activities may be associated with changes in both interspecific interactions (competition, predation, parasitism ...) and abiotic factors, direction and strength of natural and sexual selection may be deeply altered. Therefore, there is particular urgency to investigate how these anthropogenic habitat changes affect biodiversity, not only to preserve it but also to predict how evolutionary trajectories of species will be affected.

Why studying the forest ecosystem ?

Forests have been the habitats the most affected by destruction, fragmentation or degradation. If temperate forests and woodlands are now regenerating in many areas after severe destruction and overexploitation in the past, tropical and subtropical forests are disappearing at an alarming rate. The interest in these biomes is also justified by the fact that they host the higher terrestrial biodiversity.

Anse Couleuvre, Martinique

Why using birds as a biological model ?

Although birds are well studied models in temperate areas, ecological and evolutionary processes have been poorly studied in tropical birds. For instance, little is known about dispersal and movement responses to forest fragmentation in tropics. Moreover, it is sometimes assumed that birds are less affected by fragmentation at local geographic scales than other organisms because of their dispersal abilities. However, some elements suggest that fragmenting a landscape can have a greater impact on bird species persistence than previously thought, particularly in sub-tropical and tropical areas.

Merle à col blanc

Why these territories ?

The Caribbean region is one of the 25 biodiversity hotspots and French Guiana belongs to one of the three major tropical wilderness areas. On the one side, these regions are simultaneously hosting a very high diversity of species but on the other side they have been subjects to severe losses of native habitats (notably rainforest) caused by activities of a dense and growing human population.
The high level of endemism, which is characterizing these territories, not only emphasizes the originality of local faunas and floras but also provide respective countries and territories with a critical responsibility to preserve biodiversity.

Fragment forestier isolé par des cultures

In the three island territories, landscape has firstly been deeply modified by agriculture development after colonization by Europeans (production of sugar, bananas, coffee, tobacco…). In Guadeloupe and Martinique, the second factor of land modification is the local demographic expansion (for instance, Martinique is currently the French region with the highest density after the Parisian region). In addition to (and probably in interaction with) habitat modification, native fauna and flora have suffered from the introduction/invasion of numerous species acting as competitors or predators.

In French Guiana, many habitats quoted as important for biodiversity at the regional scale (included forest habitats) are located within the coastal area. However, more than 80% of the population inhabits this coastal area, and demographic models predict that population will be multiplied by two over the next 20 years (INSEE predictions), especially in this part of the region. Most studies in French Guiana have focused on the Amazonian rainforest (which benefits from protection rules), but little is known on the effects of fragmentation in the coastal forest while it is the most perturbed area.


This project is divided in five workpackages (WP):


WP1 : Fragmentation and population genetic diversity

As fragmentation basically leads to smaller and more isolated forest patches, it should bring forest species to smaller populations exchanging fewer individuals. This change in demographic processes is expected to affect genetic diversity, which is a crucial feature of natural populations. Indeed, numerous studies showed that populations with lower genetic variability have a reduced evolutionary potential and experience a higher extinction risk. Studying population genetic structure will be an indirect way of assessing effects of fragmentation on demography of birds. We will appraise how landscape structure affects dispersal patterns.

Génotypage d un individu

WP2 : Fragmentation and phenotypic variability

WP3 : Fragmentation and host-parasite interactions

WP4 : Biological invasion and ecological immunology

WP5 : Ecological specialization and response to fragmentation


This project is based on integrated methodologies (population genetics, morphometrics, immunology, spatial analyses…) and both descriptive and experimental approaches are used.

For each species, the descriptive approach requires sampling individuals in several populations from different environments (i.e. large/small and connected/isolated forest patches). Birds are captured using mist nets. They are ringed and then released after data/samples collection.

Filet de capture
Mist net
Oiseau pris dans un filet
Bird captured in a mist net
Camp de travail
Fieldwork camp
Pose d'une bague
Bird ringing

Mesure de la longueur du tarse
Measurement of tarsus length
Mesure de la longueur de l'aile
Measurement of wing length
Prise de sang
Taking of blood sample
Photographie pour mesure de couleur
Photography for colour measurement

Indeed, several morphological traits are measured (weight, wing length, tarsus length…) and a blood sample is collected for each individual. For some individuals/species, pictures of bare parts and/or plumage are taken, and a few feathers are collected.

All samples collected in the field are analyzed in the laboratory to produce several data.

Genetic data
Données génétiques
Genotyping of an individual
Sexage moléculaire
Molecular sexing

Plaque hémagglutination
Haemagglutination test
Prevalence and load of parasites
Détection de Plasmodium sur frottis
Blood smear and detection of Plasmodium
Quality of colour traits
Mesure de couleur sur photographie
Colour measurement from a photography

The experimental approach concerns the workpackage 4 (role of host immunity in the outcome of biological invasions). Birds are captured using mist nets and they are kept in cages for a few days before release. Basically, inflammatory response is measured after an intraperitoneal injection of lipopolysaccharides (LPS) from E. coli membranes diluted in Phosphate Buffer Saline (PBS). This technique is known to trigger a systematic and mild inflammation which can be quantified through the production of an acute phase protein (haptoglobin). Immune response of challenged individuals is then compared between native and invaded areas.

Species studied

The species studied in the project are listed in the following table.

French names Scientific names English names
(French Guiana)
Campylotpère à ventre gris Campylopterus largipennis Gray-breasted Sabrewing
Dryade à queue fourchue Thalurania furcata Fork-tailed Woodnymph
Ermite à brins blancs Phaethornis superciliosus Long-tailed Hermit
Grimpar bec-en-coin Glyphorynchus spirurus Wedge-billed Woodcreeper
Manakin auréolé Pipra aureola Crimson-hooded Manakin
Manakin à tête d'or Pipra erythrocephala Golden-headed Manakin
Manakin casse-noisette Manacus manacus White-bearded Manakin
Manakin tijé Chiroxiphia pareola Blue-backed Manakin
Merle à col blanc Turdus albicollis White-necked Thrush
Myrmidon à flancs blancs Myrmotherula axillaris White-flanked Antwren
Pipromorphe de McConnell Mionectes macconnelli McConnell's Flycatcher
Colibri à tête bleue Cyanophaia bicolor Blue Headed Hummingbird
Colibri madère Eulampis jugularis Purple-throated Carib
Colombe à croissants Geotrygon mystacea Bridled Quail-Dove
Colombe rouviolette Geotrygon montana Ruddy Quail-Dove
Elénie siffleuse Elaenia martinica Caribbean Elaenia
Grive à pieds jaunes Turdus lherminieri Forest Thrush
Merle à lunettes Turdus nudigenis Bare-eyed Thrush
Moqueur corrosol Margarops fuscatus Pearly-eyed Thrasher
Moqueur grivotte Margarops fuscus Scaly-breasted Thrasher
Moucherolle gobemouche Contopus latirostris Lesser Antillean Pewee
Paruline caféiette Dendroica plumbea Plumbeous Warbler
Saltator gros-bec Saltator albicollis Lesser Antillean Saltator
Solitaire siffleur Myadestes genibarbis Rufous-throated Solitaire
Sporophile rouge-gorge Loxigilla noctis Lesser Antillean Bullfinch
Sucrier à ventre jaune Coereba flaveola Bananaquit
Trembleur brun Cinclocerthia ruficauda Brown Trembler
Trembleur gris Cinclocerthia gutturalis Grey Trembler
Tyran janeau Myiarchus oberi Lesser Antillean flycatcher
Vacher luisant Molothrus bonariensis Shiny Cowbird
Viréo à moustaches Vireo altiloquus Black-whiskered Vireo

Résults of the project Frag&Binv

Dispersal of a tropical bird negatively affected by forest fragmentation at a small spatial scale

Published by Stéphane Garnier - 19-06-2017

Landscape genetic analyses revealed that forest fragmentation in Guadeloupe Island negatively affects dispersal in the Plumbeous warbler, an endemic bird of the Lesser Antilles.

Cette étude a consisté à mettre en relation les patrons de structuration génétique des populations d'une espèce aviaire en Guadeloupe avec plusieurs facteurs potentiels de structuration : l'occupation des sols, le réseau routier, et l'altitude. Une approche d'optimisation a permis de montrer que l'occupation des sols est le facteur principal influençant la connectivité fonctionnelle des populations. De plus, les secteurs urbains semblent offrir la plus grande résistance à la dispersion de cette espèce spécialiste de l'habitat forestier, tandis que les terres agricoles sont associées à une résistance intermédiaire entre secteurs forestiers et urbains. Cette étude vient d'être acceptée dans la revue Molecular Ecology.

Référence de l'article : Khimoun A, Peterman W, Eraud C, Faivre B, Navarro N, Garnier S (2017) Landscape genetic analyses reveal fine-scale effects of forest fragmentation in an insular tropical bird. Molecular Ecology, sous presse.



Résumé de l'article

Dans le cadre de la génétique du paysage, la modélisation de surfaces de résistance est un outil puissant pour tester, de manière explicite, différentes hypothèses quant aux effets du paysage sur les flux de gènes. Afin d’évaluer les conséquences de la fragmentation de la forêt tropicale sur la connectivité des populations d’une espèce d’oiseau forestier, nous avons optimisé des surfaces de résistance sans a priori en utilisant deux types de distances écologiques : la distance de moindre cout (LCP) et la distance de résistance (IBR). Nous avons implémenté une procédure en 2 étapes afin de i) définir de manière objective la résolution thématique (niveau de détail utilisé pour décrire une variable de paysage) et l’étendue spatiale du paysage (surface à l’intérieur de la fenêtre de paysage) et de ii) tester le rôle relatif de plusieurs traits du paysage (altitude, routes, occupation des sols) dans la différenciation génétique de la Paruline caféiette (Setophaga plumbea).

Nous avons détecté une réduction du flux de gènes à petite échelle spatiale, principalement induite par le type d’occupation du sol, avec un impact négatif de la matrice non-forestière sur la connectivité fonctionnelle du paysage. Toutefois, les différentes composantes de la matrice paysagère n’affectent pas la connectivité des populations de la même manière et leur conductivité augmente avec leur similarité structurelle par rapport à l’habitat forestier : les surfaces ouvertes, telles que les zones urbaines et les prairies, sont associées aux plus fortes valeurs de résistance, alors que les surfaces agricoles présentent des valeurs de résistance intermédiaires. Nos résultats ont révélé une meilleure performance de IBR comparé à LPC pour expliquer les flux de gènes, ceci reflétant des mouvements de dispersion sub-optimaux au travers de ce paysage anthropisé. Ce dernier résultat remet en question l’utilisation courante des LCP pour définir les corridors d’habitat et plaide pour une plus large utilisation de la modélisation issue de la théorie des circuits dans ce contexte. Enfin, nos résultats soulignent la nécessité de définir, de manière objective, les différentes échelles du paysage (i.e. résolution thématique et étendue du paysage) et mettent en lumière de potentiels écueils liés à la paramétrisation des surfaces de résistance.


Relative performances of different microsatellites (EST versus genomic) depends on the level of divergence studied

Published by Stéphane Garnier - 29-01-2017

Despite their lower polymorphism, microsatellites (SSRs) developed from expressed sequence tags (EST-SSRs) seem to have a higher power to detect weak genetic structure compared to genomic SSRs.

This study compares relative performances of two categories of microsatellites (microsatellites developed from expressed sequence tags (EST-SSRs) and genomic microsatellites) in detecting genetic differentiation in three bird species from the Lesser Antilles that have contrasted levels of genetic structure. It shows that these relative performances depend on the level of differentiation that is looked for. The most surprising result is the higher power of EST-SSRs than genomic microsatellites to detect a weak genetic structure, despite their lower level of allelic variability. This result challenges the common practice consisting in choosing the most polymorphic loci to study population genetic structure. This study also attests that EST-SSRs are valuable molecular markers for conservation genetic studies, particularly in species where development of specific genomic microsatellites remains delicate. This study has just been published in Molecular Ecology Resources.

Publication reference:

Khimoun A, Ollivier A, Faivre B, Garnier S (2017) Level of genetic differentiation affects relative performances of expressed sequence tag and genomic SSRs. Molecular Ecology Resources, 17: 893-903.




Microsatellites, also called simple sequence repeats (SSRs), are markers of choice to estimate relevant parameters for conservation genetics, such as migration rates, effective population size and kinship. Cross-amplification of SSRs is the simplest way to obtain sets of markers, and highly conserved SSRs have recently been developed from expressed sequence tags (EST) to improve SSR cross-species utility. As EST-SSRs are located in coding regions, the higher stability of their flanking regions reduces the frequency of null alleles and improves cross-species amplification. However, EST-SSRs have generally less allelic variability than genomic SSRs, potentially leading to differences in estimates of population genetic parameters such as genetic differentiation. To assess the potential of EST-SSRs in studies of within-species genetic diversity, we compared the relative performance of EST- and genomic SSRs following a multispecies approach on passerine birds. We tested whether patterns and levels of genetic diversity within and between populations assessed from EST- and from genomic SSRs are congruent, and we investigated how the relative efficiency of EST- and genomic SSRs is influenced by levels of differentiation. EST- and genomic SSRs ensured comparable inferences of population genetic structure in cases of strong genetic differentiation, and genomic SSRs performed slightly better than EST-SSRs when differentiation is moderate. However and interestingly, EST-SSRs had a higher power to detect weak genetic structure compared to genomic SSRs. Our study attests that EST-SSRs may be valuable molecular markers for conservation genetic studies in taxa such as birds, where the development of genomic SSRs is impeded by their low frequency.






The success of biological invasions: fewer parasites and a lower inflammatory immune response

Published by Stéphane Garnier - 20-08-2016

Birds from invasive populations have fewer parasites and a lower inflammatory immune reponse than birds from native populations.

A biogeographical approach comparing populations of Spectacled Thrush (a neotropical species originated from South America that recently colonized the Lesser Antilles) between its native area and the geographic area recently colonized provided results that support two hypothesis proposed to explain the success in biological invasions: the Enemy Release Hypothesis (ERH) and the hypothesis of different immune defense strategies in invaders. First, populations from the recently colonized area have a lower prevalence of blood parasites than native populations. Second, the inflammatory response (the immune component that is the most costly) is lower in birds from the expansion range than in their native counterparts. These results have just been published in Ecology and Evolution.


Publication reference:

Bailly J, Garnier S, Khimoun A, Arnoux E, Eraud C, Goret J-Y, Luglia T, Gaucher P, Faivre B (2016) Reduced inflammation in expanding populations of a neotropical bird species. Ecology and Evolution sous presse.





The loss of regulating agents such as parasites is among the most important changes in biotic interactions experienced by populations established in newly colonized areas. Under a relaxed parasite pressure, individuals investing less in costly immune mechanisms might experience a selective advantage and become successful colonizers as they re-allocate resources to other fitness related traits. Accordingly, a refinement of the evolution of increased competitive ability (EICA) hypothesis proposed that immunity of invasive populations has evolved toward a reduced investment in innate immunity, the most costly component of immunity, and an increased humoral immunity that is less costly. Biogeographical approaches comparing populations between native and expansion ranges are particularly relevant in exploring this issue, but remain very scarce. We conducted a biogeographical comparison between populations of Spectacled Thrush (Turdus nudigenis) from the native area (South America) and from the expansion range (Caribbean islands). First, we compared haemosporidian prevalence and circulating haptoglobin (an acute-phase protein produced during inflammation). Second, we challenged captive birds from both ranges with Escherichia coli lipopolysaccharides (LPS) and measured postchallenge haptoglobin production and body mass change. Birds from the expansion range showed lower haemosporidian prevalence and lower levels of haptoglobin than birds from the native range. In addition, the inflammation elicited by LPS injection and its associated cost in terms of body mass loss were lower in birds from the expansion range than in birds from the native range. In accordance with the enemy release hypothesis, our results suggest that range expansion is associated with a reduced infection risk. Our study also supports the hypothesis that individuals from newly established populations have evolved mechanisms to dampen the inflammatory response and are in accordance with one prediction of the refined EICA hypothesis, proposed to understand biological invasions.


Landscape structure in Guadeloupe and dispersal in an endemic forest bird

Published by Stéphane Garnier - 25-07-2016

Landscape genetics analyses revealed that open habitats and urban areas in Guadeloupe are strong barriers to dispersal in a bird species endemic from this island and Dominica: the Plumbeous warbler.

Ces résultats ont été présentés par Aurélie Khimoun lors du colloque Island Biology 2016 qui s'est tenu aux Açores du 18 au 22 juillet. Il s'agissait de la seconde conférence internationale sur l'Evolution, l'Ecologie et la Conservation des milieux insulaires. Ci-dessous le résumé de cette communication.




Résumé de la présentation

 La fragmentation des habitats est une menace sévère pour la biodiversité puisqu’elle peut entrainer la diminution de la taille et de la connectivité des populations, conduisant à une augmentation de la dérive génétique et une diminution des flux de gènes. En conséquence, on s’attend à observer une diminution de la diversité génétique au sein des populations et une augmentation de la différenciation génétique entre les populations, pouvant entrainer la modification des trajectoires évolutives des espèces et conduire à des extinctions locales. Comprendre comment la structure du paysage influence la dispersion des individus et les flux de gènes entre populations est donc une étape cruciale pour inférer les conséquences de la fragmentation des habitats et proposer des directives de gestion. Dans cette étude, nous avons suivi une approche de génétique du paysage, qui vise à lier les patrons de différentiation génétique des populations à des traits particuliers du paysage par l’intermédiaire de surfaces de résistance où des valeurs de résistance-au-mouvement des individus sont assignées aux différents types d’occupation des sols. Nous nous sommes focalisés sur une espèce d’oiseau spécialiste de l’habitat forestier, la Paruline caféiette, qui est endémique des iles de la Guadeloupe et de la Dominique. Nous avons utilisé une méthode d’optimisation de modèle pour déterminer les traits du paysage qui favorisent/contraignent les flux de gènes chez cette espèce en Guadeloupe. La connectivité fonctionnelle du paysage a été révélée avec des données microsatellites et une mesure de distance écologique a été calculée afin de décrire la connectivité structurelle du paysage. Les données génétiques ont révélé une structuration génétique entre et au sein des deux parties l’ile (Grande-Terre et Basse-Terre). Nos résultats montrent que la distance de moindre coût décrit mieux les données de différentiation génétique que la distance Euclidienne seule, attestant de l’effet de la structure du paysage sur les flux de gènes. Concernant les valeurs de résistance assignées aux différents types d’occupation des sols, les plus faibles valeurs de résistance étaient assignées aux surfaces forestières, agricoles, et aux surfaces en eau, alors que les plus fortes valeurs de résistance étaient assignées aux pairies et surfaces urbaines. Enfin, les surfaces de résistance optimisées ont révélé que l’isthme connectant Basse-Terre et Grande-Terre ainsi que les environs de la ville de Pointe-à-pitre constituent une barrière relativement forte aux flux de gènes chez la Paruline caféiette.

Forest fragmentation in the Lesser Antilles and parasite diversity in an insular bird species

Published by Stéphane Garnier - 25-07-2016

Habitat fragmentation influences host-parasite interactions and it seems to increase the infection risk.

Molecular analyses allowed the detection and the identification of parasites involved in avian malaria in several populations of an endemic species of the Lesser Antilles: the Lesser Antillean Bullfinch. These analyses were performed in Martinique and Guadeloupe, and they revealed that the proportion of infected individuals was higher in populations living in fragmented forests than in preserved forests. Antón David Pérez-Rodríguez gave a talk to present these results in the meeting Island Biology 2016, held in Azores (18-22 July). This meeting was the II International Conference on Island Evolution, Ecology, and Conservation. See the summary below.




Variation in pathogen pressure is often highly ranked among the many negative effects of forest fragmentation on wild animal populations, yet its effects seem to be context-dependent, so that more studies are needed to define a general pattern concerning the effect of fragmentation on infection risk. We set out to study if forest fragmentation could modify patterns of prevalence and richness of avian hemosporidian parasites infecting an endemic species of the Lesser Antilles: the Lesser Antillean Bullfinch Loxigilla noctis. Between 2012 and 2014, a total of 1,098 bullfinches were sampled for parasites in heterogeneous forest fragments on Guadeloupe and Martinique. We employed molecular methods to characterize parasite diversity, and we discovered that two parasite lineages (Plasmodium ICTCAY01 and Haemoproteus LOXPOR01) comprised the bulk of all hemosporidian infections. Parasite prevalence varied markedly between islands, but there was a consistent trend linking higher prevalence with fragmented landscapes. Remarkably, this link was little affected by climate, despite the sampled fragments spanning a wide range of rainfall regimes. Our work contributes to our understanding of how forest fragmentation may determine parasite pressures on tropical islands, which on the one hand are home to many endemic, often endangered avian species; and on the other hand are frequently under the threat of habitat loss by anthropogenic causes.


Tropical birds' response to forest fragmentation depends on their ecological specialization.

Published by Stéphane Garnier - 10-06-2016

A comparative approach on several species of tropical bird showed that forest specialist species are the most sensitive to fragmentation of this habitat.

Des analyses de génétique des populations ont été effectuées en Guadeloupe sur huit espèces d'oiseaux tropicaux. Parallèlement, leur niveau de spécialisation à la forêt a été estimé empiriquement à l'aide d'un millier de points d'écoute. Les résultats obtenus montrent que plus une espèce est spécialiste de la forêt, plus elle est affectée par la fragmentation de son habitat, notamment en termes de rupture de connectivité des populations. Si ce résultat est conforme aux prédictions initiales, il demeure une démonstration empirique assez unique à cette échelle spatiale pour des oiseaux. Cette étude vient d'être publiée dans la revue Molecular Ecology.


Référence de l'article :

Khimoun A, Eraud C, Ollivier A, Arnoux E, Rocheteau V, Bely M, Lefol E, Delpuech M, Carpentier M-L, Leblond G, Levesque A, Charbonnel A, Faivre B, Garnier S (2016) Habitat specialization predicts genetic response to fragmentation in tropical birds. Molecular Ecology 25: 3831-3844.



Résumé de l'article

La fragmentation des habitats est l’une des principales menaces pour la biodiversité, pouvant entrainer des changements de structuration génétique des populations, avec des modifications ultimes du potentiel évolutif des espèces et des extinctions locales. Néanmoins, la fragmentation n’a pas le même impact sur toutes les espèces et identifier les traits écologiques qui sont liés à la sensibilité des espèces à la fragmentation pourrait aider à prioriser les efforts de conservation. Malgré les liens théoriques existants entre écologie des espèces et risque d’extinction, les études comparatives testant explicitement l’hypothèse que certains traits écologiques sous-tendent la structuration génétique des espèces sont rares. Ici, nous avons suivi une approche comparative sur huit espèces d’oiseaux, coexistant à travers le même paysage fragmenté. Pour chaque espèce, nous avons quantifié les niveaux relatifs de spécialisation forestière et de différentiation génétique entre les populations. Afin de tester le lien entre spécialisation et sensibilité à la fragmentation forestière, nous avons évalué les réponses des espèces à la fragmentation en comparant les niveaux de différentiation génétique entre paysages de forêt continue et fragmentée. Nos résultats ont révélé une structuration génétique significative et conséquente à une échelle spatiale réduite pour des organismes mobiles tels que les oiseaux. De plus, nous avons montré que les espèces spécialistes de la forêt sont plus sensibles à la fragmentation que les espèces généralistes. Enfin, nos résultats suggèrent qu’un indice de spécialisation d’habitat, même assez simple, peut être un bon prédicteur des conséquences génétiques et démographiques de la fragmentation des habitats, fournissant ainsi un outil quantitatif fiable pour la biologie de la conservation.


Contrasted patterns of genetic diversity across several bird species in the Lesser Antilles

Published by Stéphane Garnier - 12-02-2016

Patterns of genetic differentiation between three islands of the Lesser Antilles show important contrasts across eight bird species, and these patterns are not always consistent with the boundaries of subspecies described in these islands.

Population genetic analyses have been performed in three islands of the Lesser Antilles in order to compare the structure of intra-specific genetic diversity between eight bird species. Patterns of differentiation not only change from one species to another but also were sometimes inconsistent with the geographical boundaries of subspecies described in these islands. These results have been published in Genetica.

Reference: Khimoun A, Arnoux E, Martel G, Pot A, Eraud C, Condé B, Loubon M, Théron F, Covas R, Faivre B, Garnier S (2016) Contrasted patterns of genetic differentiation across eight bird species in the Lesser Antilles. Genetica 144: 125-138.




Archipelagoes are considered as “natural laboratories” for studying processes that shape the distribution of diversity. The Lesser Antilles provide a favorable geographical context for divergence to occur. However, although morphological subspecies have been described across this archipelago in numerous avian species, the potential for the Lesser Antilles in driving intra-specific genetic divergence in highly mobile organisms such as birds remains understudied. Here, we assessed level of intra-specific genetic diversity and differentiation between three islands of the Lesser Antilles (Guadeloupe, Dominica and Martinique) using a multi-species approach on eight bird species. For each species, we built a set of microsatellite markers from cross-species amplifications. Significant patterns of inter-island and/or within-island genetic differentiation were detected in all species. However, levels of intra-specific genetic differentiation among the eight bird species were not always consistent with the boundaries of subspecies previously described in the sampled islands. These results suggest different histories of colonization/expansion and/or different species-specific ecological traits affecting gene flow, advocating for multi-species studies of historical and contemporary factors shaping the distribution of diversity on islands.

Influence of landscape structure in Guadeloupe on gene flow in an endemic bird species

Published by Stéphane Garnier - 24-10-2015

Landscape genetics analyses revealed how landscape structure in Guadeloupe affects populations of a bird species, the Plumbeous warbler, which is endemic from Guadeloupe and Dominica islands. Open habitats and urban areas seem to create barriers to dispersal in this forest specialist species.

These results have been presented in the Spatial data meetings for interdisciplinary sciences, organized in Dijon (October 23-24). See the summary below.


Influence of landscape structure on gene flow in a forest specialist bird species

(Khimoun A, Navarro N, Laffont R, Faivre B & Garnier S)


Habitat fragmentation is defined as the conversion of a formerly continuous habitat into smaller and isolated patches and results in a pattern of landscape spatial heterogeneity. Fragmentation is a severe threat to biodiversity as it may reduce local population size and connectivity, leading to higher genetic drift and decreased gene flow. Reduced within-population genetic diversity and increased genetic structure among previously connected populations are then expected, with ultimate modifications of species evolutionary trajectory and potential local extinctions. Thus, understanding how landscape structure influences individual dispersal and gene flow between populations is a crucial step to infer consequences of fragmentation and ultimately to propose guidelines for land settlement.

Genetic data can be used to estimated gene flow between populations and thus to infer the functional connectivity of a landscape for a particular species. Most population genetic studies have tended to characterize discrete subpopulations and to correlate their geographic boundaries with landscape features to characterize putative movement barriers. Here, we used a landscape genetic approach, where patterns of genetic connectivity can be correlated with landscape features by building resistance surfaces that assign different resistance-to-movement values to different landscape feature. We focused on a forest specialist bird species, the Plumbeous Warbler, which is endemic from the Guadeloupe and Dominica islands. We followed an approach of model optimization to infer how land cover types (i.e. forest, urbanized or agricultural surfaces, etc…) may constrain/promote gene flow in the Guadeloupean fragmented landscape. Landscape functional connectivity was assessed from microsatellite data and two different ecological distances, the least-cost distance and the resistance distance, were computed to describe the landscape structural connectivity. Least-cost and resistance distances better described the genetic structure than the Euclidian distance, attesting for the effect of landscape structure on gene flow. Regarding the resistance values of the different land cover classes, both distances provided equivalent results with lowest resistance values attributed to the forest, agriculture and water surfaces, the highest resistance values attributed to pastures and urban areas (open areas) and intermediate resistance values attributed to banana and sugar cane crops. The optimized landscape resistance surface (see Figure) revealed that the Isthmus connecting Basse-Terre and Grande-Terre as well as the surrounding of Pointe-à-Pitre constitute a relatively strong barrier to the Plumbeous Warbler gene flow. These results are preliminary and additional analysis need to be performed, for example, regarding the influence of other landscape feature (e.g. altitude, road network).




Intriguing differentiation in the forest thrush in Guadeloupe

Published by Stéphane Garnier - 23-06-2014

A study of population structure in the forest thrush revealed a strong morphological and genetic differentiation between three islands hosting this species. More interestingly, we found two entities clearly differentiated from a genetic point of view within the same small island, namely Guadeloupe.


Flying abilities of birds are usually expected to provide them with high dispersal potential. In other words, birds should be able to disperse between close populations (in the absence of physical barrier to dispersal). This idea has recently been challenged by works revealing the existence of behavioral barriers to gene flow. This is also what is suggested by results recently obtained in a bird species endemic to the Lesser Antilles, the forest thrush. We observed a very strong genetic differentiation within the same island and between populations about 10 kilometers apart. Guadeloupian populations of the forest thrush consist of two distinct gene pools (one in Basse-Terre and the other one in Grande-Terre) with no (or very few) genetic exchanges. These results have been published in Heredity.


Reference: Arnoux E, Eraud C, Navarro N, Tougard C, Thomas A, Cavallo F, Vetter N, Faivre B, Garnier S (2013) Morphology and genetics reveal an intriguing pattern of differentiation at a very small geographic scale in a bird species, the forest thrush Turdus lherminieri. Heredity, 113: 514-525.




Mobile organisms are expected to show population differentiation only over fairly large geographical distances. However, there is growing evidence of discrepancy between dispersal potential and realized gene flow. Here we report an intriguing pattern of differentiation at a very small spatial scale in the forest thrush (Turdus lherminieri), a bird species endemic to the Lesser Antilles. Analysis of 331 individuals from 17 sampling sites distributed over three islands revealed a clear morphological and genetic differentiation between these islands isolated by 40–50 km. More surprisingly, we found that the phenotypic divergence between the two geographic zones of the island of Guadeloupe was associated with a very strong genetic differentiation (Fst from 0.073–0.153), making this pattern a remarkable case in birds given the very small spatial scale considered. Molecular data (mitochondrial control region sequences and microsatellite genotypes) suggest that this strong differentiation could have occurred in situ, although alternative hypotheses cannot be fully discarded. This study suggests that the ongoing habitat fragmentation, especially in tropical forests, may have a deeper impact than previously thought on avian populations.


Publication of an article in Journal of Ornithology

Published by Stéphane Garnier - 21-05-2013

A strong phenotypic differentiation has been shown at a small spatial scale (within Guadeloupe Island) in the forest thrush, a little known species with a high heritage value.

Reference: Arnoux E, Eraud C, Thomas A, Cavallo F, Garnier S, Faivre B (2013) Phenotypic variation of Forest Thrushes Turdus lherminieri in Guadeloupe: evidence for geographic differentiation at fine spatial scale. Journal of Ornithology, 154: 977-985.


Spatial differentiation in avian models has been extensively studied at a coarse scale for both theoretical and conservation purposes. Yet, studies at a finer spatial scale are also particularly relevant in birds because their dispersal ability may be much more reduced than expected. In the Forest Thrush Turdus lherminieri, we studied morphological characters commonly used to assess differentiation because they mirror both demographic and selective processes. The Forest Thrush is an endemic and vulnerable Antillean bird species which has dramatically declined in the last 15 years, and whose population functioning and structure remain unknown. We compared birds from 11 sites in Guadeloupe, which were distributed from 2 to 42 km apart over the two main geographic zones of the island (i.e. Grande-Terre and Basse-Terre). Using two synthetic descriptors (for body size and feather size), we detected a strong micro-geographic differentiation between Forest thrush populations for the body-size descriptor but not for the feather-size descriptor. Both males and females were significantly larger in Basse-Terre than in Grande-Terre despite the fine geographic scale. Several hypotheses could explain these results: (i) geographic isolation and differentiation caused by (1) absence of gene flow, (2) phenotypic plasticity, or (3) divergence with gene flow. Although further investigation is needed to identify the exact process generating phenotypic divergence, our study provides a first highlight to the high local variability of this species.