The invasive cactus moth, Cactoblastis cactorum, poses a threat to opunitoid cacti species of North America. The following work contains four separate studies investigating C. cactorum host plant preference and performance, predation and parastitism of C. cactorum, effects of C. cactorum on local Opuntia populations, and associational effects of host and non-host plants on C. cactorum and native Opuntia-feeding herbivores. We found that, among southwestern and Mexican opuntioid taxa, moths preferred O. engelmannii var. linguiformis and var. engelmannii for oviposition, while Consolea rubescens and O. streptacantha were superior larval hosts. Oviposition was best predicted by number of cladodes and degree of spininess; epidermal toughness was a significant predictor of most larval fitness parameters. In general, oviposition preference was not correlated with larval performance. A lack of co-evolutionary history between C. cactorum and North American opuntioid species may help explain this disconnect. We placed irradiated C. cactorum eggsticks and pupae on Opuntia plants in the field to test for predation. We found evidence of predation, most likely from ants, on ~16% of eggsticks and ~18% of pupae. Predation rates, ant abundance, and cladode growth were higher, and C. cactorum damage lower, on Opuntia located near the extrafloral nectar-producing legume Chamaecrista fasciculata. We attribute these associational effects to the ability of C. fasciculata to attract ants to its extrafloral nectar. Over the course of six years, ~78% of nearly 600 tagged Opuntia were attacked by C. cactorum at least once and ~76% of the plants survived. Two separate studies found that Opuntia stricta was more likely to be attacked by C. cactorum than O. humifusa; we also found that O. stricta was more likely to die following an attack. A plant's odds of survival decreased as C. cactorum attack frequency increased, but plants that did survive had positive growth rates, regardless of attack status. We did not find significant evidence of associational effects for O. humifusa and O. stricta, either for C. cactorum or native Opuntia-feeding herbivores. It could be that present herbivore densities are low enough, and host plants plentiful enough, to avoid mechanisms that usually lead to associational resistance or susceptibility. Overall, our results suggest that the presence and spread of C. cactorum should be taken seriously, especially for rare opuntioids and the Opuntia-rich deserts of North America. However, for more common opuntioid host species, there may be enough resistant or tolerant individuals, and sufficient top down control through ant predation, for populations to persist at current C. cactorum densities. We acknowledge that information on Opuntia reproduction and recruitment rates are needed to confirm this suggestion, and see this as an excellent opportunity for future research

Indirect ecological effects such as associational resistance and resource competition have the potential to affect ecological interactions and influence the structure of ecological communities. Although resource competition is commonly studied, the effects of associational resistance are not as evident if studies are not designed to detect them. Additionally, the relative strengths of different ecological mechanisms ought to be measured in studies, rather than the strength of singular mechanisms. This permits proper attribution of causes and effects in community structure and detection of higher order interactions in a way that naïve reductionism will not. In a series of experiments, I looked at the effects of large-scale addition and removal of Borrichia frutescens on associational resistance of Iva frutescens to the gallformer Asphondylia borrichiae in order to test the mechanism and strength of associational resistance in the system. Additionally, I measured the effects of relative host abundance and interpatch distance of hosts on associational resistance. Finally, I looked at the effect of the presence of stemborers competing with the gall former for host plant resources on parasitism rate and parasitoid guild composition. I found evidence for a strong effect of associational resistance by natural enemies over short distances, although the phenomenon is likely of minor importance in comparison to other factors known to affect gall former population density, such as environmental effects and host plant genotype. Competitors also had a weak effect, reducing mean gall diameter, but not significantly altering total parasitism rate. However, the presence of stemborer competitors did slightly alter the composition of the parasitoid guild. In sum, both associational resistance and competition from stemborers have detectable effects on A. borrichiae, albeit weak ones. Environmental factors, such as soil nitrogen content, are likely much stronger determinants of gall former population size.

Les peuplements forestiers mélangés subissent moins de dégâts d’insectes herbivores que les peuplements purs, du fait d’une diminution de l’accessibilité à la ressource ou/et d’une amélioration des conditions de survie de leurs ennemis naturels. Pour vérifier ces hypothèses, nous avons étudié un insecte ravageur, la processionnaire du pin (Thaumetopoea pityocampa), et ses insectes parasitoïdes dans le contexte de monoculture de pin maritime (Pinus pinaster) de la forêt des Landes de Gascogne, où persistent localement des zones de milieu ouvert (parefeux, coupes rases) et des boisements d’essences feuillues (haies en bordure de peuplement de pin, ripisylves, îlots). Nous avons pu démontrer que :(1) La colonisation des parcelles de pin par la processionnaire est limitée parla présence de haies de feuillus en lisière de parcelle. En effet, les feuillus jouent un rôle de barrière physique, entravant la détection visuelle des pins par la femelle de processionnaire lorsqu’elle recherche un site d’oviposition.(2) La longévité des principaux parasitoïdes, spécialiste et généraliste, des oeufs de processionnaire est favorisée par la consommation de miel lat produit pardes pucerons du chêne. Cela permet notamment à l’espèce généraliste, qui émergedeux mois avant la processionnaire, de prolonger sa présence dans le milieu et doncd’augmenter sa probabilité de parasiter des pontes de processionnaire.(3) Les chrysalides de processionnaire du pin survivent mieux dans le sol desmilieux ouverts que sous couvert forestier (de pin ou de feuillus), du fait d’unetempérature et d’une humidité plus élevées. L’association de pins et de milieuxouverts favorise la processionnaire par complémentation des habitats, tandis que laprésence de feuillus peut représenter un piège écologique pour les chenilles aumoment de l’enfouissement.(4) À l’échelle du paysage, les peuplements de pin maritime au centre depaysages hétérogènes sont moins infestés que dans les paysages de monoculture.De plus, les niveaux d’infestation de la processionnaire diminuent lorsque laproportion de feuillus dans le paysage environnant augmente.Ces résultats sont interprétés en fonction des mécanismes écologiquesexpliquant la relation entre diversité et herbivorie. Des possibilités de transfert vers lagestion forestière de la forêt des Landes de Gascogne sont proposées, ainsi que desperspectives en termes de recherche scientifique. / Mixed forests are less prone to insect damage than pure forests because ofreduced host accessibility and/or improved control by natural enemies. To test thesehypotheses, we have studied the ecology of the pine processionary moth (PPM)(Thaumetopoea pityocampa) and its parasitoid, in a monoculture of maritime pine(Pinus pinaster) plantations, the Landes de Gascogne forest. There, open areas(firebreaks, clear cuts) and patches of broadleaved woodlands (hedgerows, riparianforest, natural forest remnants) still persist locally. In this study we have shown that:(1) Pine stand colonization by PPM was limited by the presence ofbroadleaved hedgerows at stand edge. Broadleaved trees formed physical barriersdisrupting the visual detection of pine trees by PPM females when searching for anoviposition site.(2) The longevity of the two main PPM egg parasitoids increased whenspecimen were fed with honeydew produced by oak aphids. The generalist species,which emerges two months before PPM, could benefit from this longer lifespan tooverlap its host emergence.(3) PPM pupae survived better in the soil of open areas than under forestcovers (pine or broadleaved trees), because of higher temperature and humidity. Theassociation between pine stands and open areas benefits PPM through habitatcomplementation, whereas the presence of broadleaved trees may act as anecological trap for PPM caterpillars.(4) Maritime pine stands within heterogeneous landscapes exhibited lowerPPM infestations than similar stands within pine monocultures. PPM infestation levelsdecreased with increasing percent broadleaved area in the surrounding landscape.These results are discussed according to the ecological mechanisms whichmay explain the relationship between insect herbivory and tree species diversity.Perspectives for improved PPM management in the Landes de Gascogne forest, andfor further scientific research are proposed.

This document reports the results of 4 studies of subalpine ecosystem ecology, describing ways that spatial heterogeneity in soils and plant communities mediate ecosystem responses to environmental change. Ecosystem responses to environmental change are also mediated by regional climate patterns and interannual variability in weather. In the first chapter we report the results of an experiment to test for the mediating effects of associational resistance in a forest community that experienced wide-spread beetle kill. We found that Engelmann spruce were more likely to survive a beetle outbreak when growing in low densities (host dilution) and not through other types of associational resistance that relate to higher tree-species richness or greater phylogenetic diversity of the forest community. In the second chapter we report the effects of early snowmelt on soil moisture in subalpine meadow and aspen communities. We found that soil organic matter, soil texture, and forest cover mediated the effects of early snowmelt and were more important drivers of growing-season soil moisture than was snow-free date. In the third chapter we report the effect of early snowmelt on growth and seed production of early-season and midsummer herbaceous species. We found that the primary effect that snowmelt timing had on plant growth was through its effect on species distribution. Changes in the timing of snowmelt had limited effect on the growth, flowering, and seed count of species after they were established. In the final chapter, we report the effect of early snowmelt on soil respiration, microbial biomass, dissolved organic carbon and soil organic carbon. We found that early snowmelt resulted in warmer soil temperatures compared to neighboring snow-cover plots, and that microbial biomass and soil respiration showed no signs of a snowmelt legacy effect during the growing season. Soil organic carbon in rapid and slow-turnover pools was affected more by plant community than by snowmelt timing, and the primary drivers of soil respiration during the snow-free period were first soil organic matter and second soil temperature. Taken together, this dissertation reports our findings that subalpine ecosystems are resilient to environmental change in part because organisms in these systems are adapted to environmental conditions that are highly variable between sites, seasons, and years.

The cabbage seedpod weevil, Ceutorhynchus obstrictus (Marsham) (Coleoptera: Curculionidae), is an important pest of brassicaceous oilseed crops, especially canola (Brassica napus L. and Brassica rapa L.) in North America and Europe. Application of foliar insecticide is the only method currently employed to control C. obstrictus populations; because this approach is environmentally unsustainable, alternatives including host plant resistance have been explored. White mustard, Sinapis alba L., is resistant to C. obstrictus and was chosen as a potential source of resistance for B. napus oilseed. Interspecific crosses of S. alba x B. napus have produced several lines that are resistant to C. obstrictus feeding and oviposition and yield fewer, lighter-weight weevil larvae that take longer to develop. I investigated potential mechanisms of this resistance, including assessing differences in visual and olfactory cues among resistant and susceptible genotypes, and antixenosis and antibiosis. Determining effects of visual cues associated with host plant resistance required investigation of weevil vision. Deployment strategies for resistant germplasm were assessed to evaluate incorporation of susceptible refugia to promote long-term durability of resistance traits. Results reported in Chapter 2 indicate that the C. obstrictus visual system is apparently trichromatic and incorporates receptors with response maxima near 350, 450, and 550 nm. Modelling indicated that UV light alone reduced weevil responses but the interaction of yellow and UV light increased responses at a threshold reflectance level of UV. Results reported in Chapter 3 indicated that differences in yellow and UV reflectance among host plant flowers influence host selection in C. obstrictus. Results described in Chapter 4 determine differential attraction to the odours of S. alba and B. napus and among resistant and susceptible accessions. Inferences of the identities of glucosinolates found in varying amounts among susceptible and resistant genotypes suggested that 2-phenylethyl glucosinolate influenced attractiveness. Results described in Chapter 5 indicate differences in adult feeding and oviposition preferences among resistant and susceptible genotypes. Oocyte development, larval biomass and larval development time varied among weevils feeding on resistant and susceptible genotypes. Based on results of Chapter 4, 1-methoxy-3-indolylmethyl glucosinolate was implicated as contributing to antixenosis and antibiosis resistance. Results reported in Chapter 6 describe effects of mixed plots of resistant and susceptible genotypes on weevil spatial distribution and oviposition. These results are consistent with associational resistance and attributed to reduced apparency of susceptible plants in mixtures and antixenosis resistance associated with resistant germplasm. / Plant Science

No description available.

Les relations entre plantes et herbivores occupent une part importante des travaux concernant le rôle de la biodiversité dans le fonctionnement des écosystèmes. Les effets de la diversité des plantes sur les herbivores et la phytophagie se répartissent le long d’un continuum entre résistance et susceptibilité par association, ces deux concepts décrivant respectivement la réduction ou au contraire l’augmentation des dégâts causés par les herbivores avec la diversité des plantes dans les communautés végétales. Si de nombreux travaux menés dans différents écosystèmes suggèrent que la résistance par association est plus fréquente que la susceptibilité par association, les mécanismes sous jacents restent discutés et semblent dépendre de la plus ou moins grande spécialisation des herbivores et de l’identité des espèces végétales associées. En particulier deux questions se posent : (1) quel est l’effet de la diversité des plantes sur la diversité des insectes, et principalement sur les insectes herbivores ? (2) Quelle composante de la diversité des plantes explique le mieux la résistance et la susceptibilité par association ? La méta-analyse présentée dans cette thèse montre que la diversité des animaux augmente avec la diversité des plantes, quels que soient les écosystèmes et les taxons considérés. Dans le cas des Arthropodes, la richesse spécifique des herbivores est mieux corrélée à la richesse spécifique des plantes que celle des prédateurs. De plus, la diversité des herbivores augmente plus rapidement que la diversité des prédateurs avec la diversité des plantes. Ces observations soulèvent un paradoxe : si la diversité des herbivores augmente avec celle des plantes, comment expliquer les nombreux cas où la phytophagie diminue (résistance par association) ? Cette question a fait l’objet d’une approche expérimentale centrée sur le chêne pédonculé et fondée sur la manipulation de la diversité spécifique et génétique des essences forestières. Il en ressort que ces deux niveaux de diversité ont des effets contrastés sur la phytophagie. Les dégâts causés par les insectes herbivores généralistes tendent à augmenter avec la diversité génétique chez leur hôte (susceptibilité par association), alors que les spécialistes ne répondent pas à la diversité génétique chez le chêne. Au contraire, la diversité spécifique est à l’origine d’une résistance par association pour les herbivores spécialistes mais elle est sans effet sur les généralistes. Un résultat original émerge des travaux présentés dans cette thèse : les effets de la diversité des plantes sur les insectes herbivores et leurs dégâts s’exprimeraient principalement à l’échelle locale – celle d’un arbre et de ses voisins immédiats – au travers de « l’apparence » des arbres, une variable conditionnée par leur taille et celle de leurs voisins. La prise en compte de la spécialisation des insectes herbivores et des différentes composantes de la diversité des arbres contribue donc à une meilleure compréhension des mécanismes impliqués dans la résistance et la susceptibilité par association des forêts mélangées. / Functional and structural consequences of plant diversity on insect communities and herbivoryThe relationship between biodiversity and ecosystem functioning remains a key issue in ecology. In particular it has been suggested that plant species diversity can trigger plant-insect interactions, ranging from associational resistance (higher herbivory in monocultures than in more diverse plant communities) to associational susceptibility (higher herbivory in plant mixtures). Although it is now widely acknowledged that associational resistance is more frequent than associational susceptibility, the underlying mechanisms are still debated and seem to depend on herbivore specialization and plant species identity. To contribute to this debate we addressed two questions: (1) What are the consequences of plant diversity for insect herbivore diversity? (2) Which component of plant diversity better accounts for herbivory? On the basis of a meta-analysis we show in this report that animal diversity significantly increase with increasing plant diversity, regardless of habitats or taxa. In arthropods, herbivores species richness increases more steeply with plant richness than predator richness. This result raises an apparent paradox: if herbivore diversity is positively correlated with plant diversity, why herbivory decreases with increasing plant diversity (associational resistance)? To solve this paradox we developed an experimental approach focused on pedunculate oak and based on the manipulation of tree genetic and species diversity. We showed that generalist and specialist herbivores respond differently to the two levels of tree diversity. Herbivory by generalist herbivores significantly increased with oak genetic diversity (i.e. associational susceptibility) while specialist herbivores did not respond to the genetic diversity of their host tree. Conversely, tree species diversity triggered associational resistance to specialist insect herbivores but had no effect on generalists.  The novelty of this works also lies in the identification of a new mechanism that may account for associational resistance in tree species mixtures: host apparency. More apparent oak trees – those that are taller than their neighbours – are more prone to insect damage. Increasing tree species diversity comes together with higher probability of associating faster growing species (such as pine or birch) than can hide neighbouring oaks, reducing their apparency and then their susceptibility. Our findings suggest that accounting for the diet breadth of insect herbivores and the different components of tree diversity can contribute to a better understanding of the mechanisms leading to associational resistance or susceptibility in mixed forests.

Qui du phénomène de résistance (RA) ou de susceptibilité (SA) par association est prépondérant en milieu forestier ? En d’autres termes, la défoliation par les insectes herbivores est-elle moins (RA) ou plus (SA) importante sur des arbres hôtes situés en mélange comparés à des monocultures ? A cause des contraintes méthodologiques, les connaissances sur la relation diversité - résistance en forêts matures restent faibles. Pour répondre à cette question, nous avons utilisé le dispositif exploratoire du projet FunDivEUROPE. Sur un réseau de 209 parcelles forestières sélectionnées le long de deux gradients orthogonaux de richesse spécifique (d’une à cinq essences en mélange) et de latitude (de la forêt méditerranéenne à la boréale), nous avons évalué la défoliation des houppiers des arbres. A l’aide d’un échantillon constitué de onze essences différentes, nous avons démontré pour la première fois un patron global de diminution de l’herbivorie (RA) en forêts matures à travers l’Europe. Nous avons également comparé l’herbivorie des insectes dans des placettes constituées de chênes purs ou mélangés à d’autres espèces, placées en lisière ou à l’intérieur de petits bois présents dans les vallées et coteaux de Gascogne. Nous avons observé moins de dégâts foliaires sur les chênes entourés par des voisins hétérospécifiques, et nous avons montré que la magnitude de cette résistance (RA) était plus importante en lisière qu’à l’intérieur des bois. Enfin, nous avons testé cette hypothèse de RA sur une espèce d’herbivore invasif en Italie, le cynips du châtaignier. Basée sur la même approche méthodologique, notre étude a montré de plus faibles dégâts de galles sur les châtaigniers lorsque la richesse spécifique de la parcelle était élevée. Notre travail fournit donc de nouvelles preuves supportant l’hypothèse de RA à travers trois contributions originales : i) la RA existe en forêt mature, ii) quelle que soit l’espèce d’herbivore, iii) y compris sur des espèces exotiques. Ces résultats devraient avoir d’importantes implications pour la gestion des insectes herbivores en forêt puisque le maintien et l’amélioration de la diversité des essences représentent un outil prometteur pour prévenir les dégâts d’insectes. / Whether increasing forest diversity should result in less insect damage (Associational Resistance, AR) or more damage (Associational Susceptibility, AS) is still debated. Moreover little is known about the diversity - resistance relationships in mature forests due to methodological constraints. To answer this question we used the FundivEUROPE exploratory platform. In this network of 209 mature forest plots sampled along two orthogonal gradients of tree species richness (from one to five species mixtures) and latitude (from the Mediterranean to the boreal forest biomes), we assessed insect defoliation in tree crown. Focusing on eleven broadleaved species, we could for the first time demonstrate a global pattern of reduced defoliation (AR) in mature forests across Europe. We replicated the comparison of insect herbivory in pure vs. mixed plots of oak trees, at the edge or within forest patches of the valleys and hills of Gascony. Here we found significantly less damage on oaks surrounded by heterospecific neighbors, and showed that the magnitude of AR was larger at forest edge than in forest interiors. Last we tested the AR hypothesis with an invasive alien species, the Asian chestnut gall wasp, in Italy. Based on the same methodological approach, our study showed lower gall damage on chestnuts when tree species richness was higher in mixed mature stands. Our work therefore provide new evidences to support the AR hypothesis with three original contributions: i) AR does apply to mature forests, ii) irrespective of the identity of insect herbivores, iii) including exotic species. These findings may have important implications for pest management in forest since the maintenance or improvement of tree species diversity represent a promising tool to prevent insect damage.

Les herbivores consomment les tissus végétaux et affectent la survie et la croissance des végétaux. Les ennemis naturels des herbivores peuvent réguler leurs populations et indirectement diminuer les dommages sur la végétation. Les relations entre plantes, herbivores et prédateurs sont liées aux caractéristiques intrinsèques des espèces végétales mais aussi influencées par la communauté végétale environnante. L'hypothèse de résistance par association stipule qu'une communauté végétale diversifiée entraîne une diminution des dégâts sur une plante cible par dilution, répulsion des herbivores et/ou favorisation des ennemis naturels. Inversement, les herbivores peuvent se concentrer sur la plante-cible dans une communauté végétale diversifiée (susceptibilité par association). L'objectif de cette thèse est d'évaluer 1) l'influence de la composition de la communauté végétale sur les herbivores et leurs dommages sur une plante-cible, 2) l'impact de la régulation des herbivores par les prédateurs et 3) les interactions entre herbivores, communauté végétale et prédateurs, et les conséquences pour les intensités d'herbivorie sur la plante-cible. Nous nous sommes focalisés sur les dommages causés par les insectes herbivores sur des plantules d'essences forestières feuillues (Betula pendula, Quercus robur et Q. ilex). La variété importante de communautés végétales, au niveau inter (essences dominantes) et intraparcellaire (présence, structure et composition), nous a permis de tester les effets du voisinage végétal sur les insectes herbivores et leurs dommages sur les plantules cibles. Les oiseaux insectivores sont les principaux prédateurs d'insectes et nous avons cherché à estimer les effets de leur exclusion sur les insectes et les conséquences pour les plantules, et ceci le long des gradients de composition du voisinage végétal. Enfin, le degré de spécialisation (spécialiste/généraliste) et le mode de vie (exo/endophyte) ont été caractérisés car ils conditionnent les réponses des insectes herbivores à la végétation voisine ainsi que la prédation avienne. Le voisinage végétal des plantules-cibles modifie la charge en herbivores et les dommages engendrés. Ces réponses sont liées à la spécialisation de l'insecte herbivore : la colonisation des plantules par les espèces spécialistes (mineuses de feuilles) est négativement affectée par la structure de la végétation locale (présence, recouvrement arbustif). Les dommages d'insectes généralistes augmentent avec la richesse spécifique des peuplements et dans les parcelles dominées par des espèces conspécifiques. Les abondances des insectes externes et leurs dommages sont régulés par les oiseaux insectivores mais cet effet indirect dépend de l'espèce de plantule-cible et de la composition de la communauté végétale. Les effets de cascade trophique sont les plus intenses dans les parcelles de pin maritime où les plantules constituent une ressource nouvelle pour les niveaux trophiques supérieurs. Plus localement, le voisinage végétal modifie la sensibilité des arthropodes à la prédation: les oiseaux insectivores bénéficient aux plantules lorsque les proies sont plus accessibles (végétation environnante supprimée).Pour conclure, une des principales richesses de ces résultats vient de la mise en évidence d'interactions très fortes existant entre les effets des oiseaux insectivores et la végétation environnante. Les effets indirects des prédateurs et directs de la végétation environnante sont très liés au degré de spécialisation des insectes phytophages et à leur mode de vie. / Herbivory is thought to depend on intrinsic plant resistance traits and negatively affect survival and growth of plants. Predators can depress herbivore populations and thereby indirectly limit the consumption of primary producers. The surrounding vegetation is also expected to modify the interactions between a focal plant, its herbivores and their natural enemies. The surrounding vegetation may disrupt the colonisation and the consumption of a focal plant by insect herbivores, providing associational resistance. Increasing diversity of neighbouring vegetation can decrease resource availability or enhance herbivores control by predators. By contrast, associational susceptibility may occur with higher levels of herbivory in a focal plant in more diverse plant communities. We tested 1) the effect of the neighbouring vegetation on the levels of herbivory on seedlings, 2) the indirect impact of predators on herbivores and subsequent herbivory, and 3) the interactions between surrounding vegetation, focal plants, herbivores and predators. We examined the damage caused by insect herbivores on three native broadleaved species. We then compared the importance of taxonomic similarity between seedlings and canopy tree species and the structure and composition of surrounding vegetation at both large- and small-scales. Birds are vertebrate predators likely to limit damage to plants and facilitate plant growth by consuming herbivorous insects. We estimated their effects by excluding them from focal plants, and compared their interactive effects with vegetation diversity or removal. The effects of surrounding vegetation and avian predation were also disentangled on concealed- vs. external-feeding guilds (predation sensitivity) and specialist- vs. generalist-insects.Surrounding vegetation of focal seedlings influences the abundance of insect herbivores and damage. Specialist insects (leaf miners) are affected by the understorey vegetation close to oak seedlings, decreasing with its presence or structural diversity. The damage caused by generalist insects depend on the large-scale composition of plant communities, and increase with tree species richness and cover or on conspecific forest habitats. Exclusion of insectivorous birds affects insect herbivory in a species-specific manner, and also greatly varies with forest habitats and presence of vegetation around focal plants. The indirect effect of bird predation on leaf damage is observed on seedlings beneath noncongeneric canopy trees. The removal of local surrounding vegetation affects the top-down effect of insectivorous birds on insect herbivores, according to their sensitivity to predation and the accessibility of preys. To conclude, our studies provide experimental evidences of interactive effects between bird predation and neighbouring vegetation on insect herbivores and levels of herbivory. These indirect effects are strongly related to the specialisation of herbivores and to their sensitivity to bird predation.