L’agriculture intensive et les changements climatiques ont profondément modifié la biodiversité et les systèmes d’interactions trophiques dans les paysages agricoles, entrainant la perte de services écosystémiques tel que le contrôle biologique de ravageurs. La lutte biologique par conservation vise à restaurer la complexité des réseaux trophiques au sein des terres cultivées par l’utilisation de la diversité végétale et la mise en place d’aménagements fonctionnels. Ces travaux de thèse ont cherché à identifier, par des expériences en laboratoire et trois hivers de suivi des populations, comment la diversité végétale et son aménagement pendant la saison hivernale pouvait favoriser les ennemis naturels des pucerons dans les grandes cultures céréalières de l’ouest de la France. Les études en laboratoire ont montré que la diversité floristique cultivée dans les couverts hivernaux devait faire l’objet d’une sélection fonctionnelle. En particulier il est nécessaire de combiner plusieurs espèces, alliant attractivité et profitabilité pour les insectes parasitoïdes de pucerons présents dans la mosaïque agricole durant l’hiver. A l’échelle des parcelles cultivées, il a pu être démontré que l’augmentation des températures hivernales favorisait l’activité et la diversité des insectes parasitoïdes, mais aussi des prédateurs généralistes des communautés d’insectes de la faune du sol. La mise en place d’habitats fonctionnels par la sélection d’une diversité végétale adaptée à ces nouvelles conditions et aux besoins spécifiques de ces agents de bio-contrôle permet alors de promouvoir leur activité. En favorisant ces différentes guildes d’arthropodes pendant cette période de l’année, il pourrait être possible d’augmenter la pression précoce qu’ils exercent sur les populations de pucerons et de limiter le développement de ces dernières au printemps. L’ensemble des résultats obtenus suggère que la saison hivernale représente une période favorable au développement de méthodes durables pour la lutte biologique par conservation dans les régions au climat tempéré. / Intensive agriculture and global warming induced deep modifications in the agricultural landscapes’ biodiversity and trophic system structures, resulting in the loss of several ecosystem services such as pest biological control. Conservation biological control aims to restore trophic systems’ stability within arable lands by using plant diversity and creating functional habitats for natural enemies. This thesis aimed at identifying how vegetal diversity and its management during the winter season could favor aphids’ natural enemies in cereal intensive agricultural landscapes from the west part of France. Laboratory research and a follow-up of populations over three consecutive winters led to significant results, further developed in this thesis. The experimental laboratory research showed that the selection of the flower diversity used in sown winter cover crops should take into account their functional traits. It is indeed necessary to use a variety of species, thus combining attractiveness and profitability for the aphid’s parasites present in the agricultural mosaic during wintertime. At the field scale, studies demonstrated that warmer winter conditions are also favoring the activity of generalist ground dwelling predators, as well as that of parasitoids. The implementation of functional habitats, through the selection of an adequate plant diversity better adapted to the changing climates and to the aphids’ natural enemies’ needs, allowed to support the ground predators’ activities. Consequently, their increased activity should contribute to the early regulation of aphid populations, allowing to decrease the intensity of their population dynamics during the growing seasons. The results obtained in this thesis suggest that winter constitutes a favorable period/season for the development of new methods within the conservation biological control framework/biological control framework for conservation in temperate climatic areas.

High-biomass cover cropping enhances marketable yields in organic production of vegetables, linked to the improvement of soil quality and weed control. Although, during transition from chemical to organic cover-cropping production, especially with no-till systems, reduction of nitrogen availability to the main crop and increase in weed and pest pressure may occur. In 2004-2005, summer and fall broccoli (Brassica oleraceae L. Botrytis Group) crops were grown in twin rows on permanent (controlled traffic) raised beds (185-cm wide). Before broccoli transplanting, high-biomass cover crops were grown in specific bed areas. Legumes (Vicia villosa, Glycine max L., and Crotolaria juncea L.) on bed tops (grow zones) and grass species (Secale Cereale L., Setaria italica L., and Sorghum bicolor X S. bicolor var. Sudanese) in the alleyways (bed shoulders and bottoms). Experimental treatments were tillage (conventional, CT; and no-tillage, NT), farmscaping (with and without), and nitrogen sidedressing (with and without, applied 3 weeks after transplanting as a mixture of sodium nitrate - 22 kg N ha-1 - and feathermeal - 44 kg N ha-1). Weeds were managed by mechanical cultivation in CT and a spot weeding by hand in NT treatments. High numbers of beneficial insects (Cotesia glomerata, Cotesia orobenae and Diadegma insulare) kept the primary insect pest population (Pieris rapae, Evergestis rimosalis (Guenee), and Plutella xylostella) at a pest to predator ratio below 4:1. Although the excellent insect pest control was attributed to the farmscape plantings, pest level and crop yields were not significantly affected by farmscaping (likely due to the close proximity of the farmscaped plots (10-50m from non-farmscaped)). Broccoli yield averaged 62% higher in fall than summer (12.1 vs. 7.5 t ha-1) likely due to cool weather conditions during broccoli head development (October), increasing head size, uniformity, and marketability. In 2004, broccoli yield in CT plots was either equal or slightly higher than NT (9.5 vs. 9.0 t ha-1). However, in 2005, broccoli yield in CT plots was significantly higher in both spring (8.8 vs. 6.8 t ha-1; p = 0.0258) and fall crops (13.5 vs. 12.3 t ha-1 with p = 0.0484). Nitrogen sidedressing improved yield in all plots (9.8. vs. 12.7 t ha-1) and particularly in NT (8.6 vs. 12.1 t ha-1), indicating that availability and/or synchrony of nitrogen was a limiting factor. Incorporating high-N legume residues in the grow zones resulted in a lesser N response in CT. / Master of Science

L'intérêt des prédateurs généralistes dans la lutte biologique par conservation est peu étudié en verger de pommiers. Cette thèse visait à déterminer le rôle de prédateurs généralistes invertébrés (principalement carabes et araignées) et vertébrés (mésanges et chauves souris) dans le contrôle de trois ravageurs clés du pommier et de définir les éléments paysagers qui leur sont favorables. Pour cela, nous avons effectué des suivis de populations dans des vergers en agriculture biologique et nous les avons associés à des mesures de prédation des ravageurs par analyse moléculaire des contenus stomacaux et faeces des prédateurs à l'aide d'amorces spécifiques développées à cet effet. Nous montrons que les ravageurs sont consommés dès leur apparition dans les vergers, notamment les pucerons par les araignées de la frondaison, ce qui suggère que ces prédateurs ont un potentiel de régulation important. De plus une complémentarité temporelle semble exister entre araignées et carabes dans le contrôle des tordeuses. L'influence des éléments paysagers intra et extra verger diffère selon les taxa de prédateurs considérés. Ce dernier résultat suggère des compromis dans les aménagements à réaliser pour favoriser leur action et la nécessité d'une meilleure compréhension de leur écologie. / Generalist predators have been merely studied in conservation biological control studies in apple orchards. This work aimed to determine the potential of invertebrate (arthropods, mainly carabids and spiders) and vertebrate (great tits and bats) predators at suppressing three key apple pests and then to devise which landscape features are beneficial to them. For this purpose, we monitored pest and predator populations in organic apple orchards of southeastern France and related their abundance and diversity to measures of pest predation using molecular gut content and faeces analysis with specific primers designed for this purpose. We found that pests were consumed as soon as they arrived in the orchard, in particular aphid by canopy spiders, indicating that these predators are likely to be efficient. We also observed some temporal complementarity in predation of tortricids by spiders and carabids. Finally, the influence of landscape features within and surrounding the orchards was different for the different studied taxa suggesting the need for compromises in orchard management for enhancing natural enemy populations and for a better consideration of the ecology of these predators.

Made available in DSpace on 2016-04-28T14:58:35Z (GMT). No. of bitstreams: 1 2008- Andre Luis Santos Resende01.pdf: 7392122 bytes, checksum: fd01e5b7d3b70d5a684bf2fec040cc46 (MD5) Previous issue date: 2008-02-22 / Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico / The agroecological approaches of vegetable organic production presumed the design of more diversified production units over time and space, where the intercropped systems are included. One of these approaches is that the persistence, the abundance and the diversity of natural enemies that act in the insect-pest biological control are stimulated. In Brazil, despite of informal reports, such as personal communication, little has been recorded about the effects of intercropped systems in populations of insect pests and natural enemies. In this context, the present study has as objectives to determine the diversity and the community structure of the ladybirds (Coleoptera: Coccinellidae) associated with aphids in kale cultivated intercropped with coriander, and to evaluate the potential of coriander to provide the vital resources for survivor and reproduction of ladybirds, as well as evaluate the crop performance of this intercropped system, in agroecologically-based organic production system. The study was conducted in two consecutive years (2006 and 2007) in the experimental farm of integrated crop-livestock production called of Sistema Integrado de Produ??o Agroecol?gica (SIPA) or Fazendinha Agroecol?gica km 47, localized in the municipality of Seropedica, Rio de Janeiro, Brazil, where the experiments were carried out and involved the intercropped system of kale (Brassica oleraceae var. acephala D.C.) with coriander (Coriandrum sativum L.) in relation to kale in the monocrop system, in order to determine the diversity of ladybirds by collection of specimens through removal sampling and yellow sticky cards, as well as characterize the structure of their community by the faunistic analysis. The performance of intercropped system was also evaluated, determining the crop parameters of associated crops. A total of 25 ladybird species are present in the SIPA, but Hyperaspis (Hyperaspis) festiva Mulsant, Scymnus (Pullus) sp.3, Cycloneda sanguinea (L.) and Eriopis connexa Germar were more frequent species. There was no infestation by aphids in kale intercropped with coriander, which was used by the ladybirds as food site (resource of pollen, nectar and alternatives preys), oviposition site, refuge for larvae, pupae and adults, and mating site. As for the cropping parameters, the Land Equivalent Ratio (LER) for the intercropped system of kale with coriander, taking into account the fresh mass yield, were superior in 92% (coriander harvested with 55 days after the sowing) and 85% (coriander allowed for flowering) in relation to monocrop system. The intercropped system of kale with coriander in the arrangement of one row of kale and four paralleled rows of coriander, harvesting the two central rows at 55 days after sowing and allowed the two rows in the border of the seedbed for flowering, do not cause increases of kale yield, but was effective in respect to land effective use and as strategy of ladybirds conservation in diversified production systems. / As abordagens agroecol?gicas de produ??o org?nica vegetal pressup?em o desenho das unidades de produ??o mais diversificado, no tempo e no espa?o, onde se inclui os cons?rcios de culturas, tamb?m conhecidos como policultivos. Um dos princ?pios dessa abordagem ? que atrav?s da diversifica??o dos cultivos, estimula-se a persist?ncia, a abund?ncia e a diversidade de inimigos naturais que atuam no controle biol?gico de insetos-pragas. No Brasil, apesar dos relatos informais, do tipo comunica??o pessoal, pouco tem sido registrado a respeito dos efeitos dos policultivos nas popula??es de insetos-pragas e inimigos naturais. Nesse contexto, o presente estudo teve por objetivos determinar a diversidade e a estrutura da comunidade de joaninhas predadoras (Coleoptera: Coccinellidae) associados a pulg?es em couve, cultivada em cons?rcio com coentro, al?m de avaliar o potencial do coentro como provedor de recursos vitais para sobreviv?ncia e reprodu??o das joaninhas, assim como o desempenho fitot?cnico desse cons?rcio, em sistema de produ??o org?nica conduzido em bases agroecol?gicas. O estudo foi realizado em dois anos consecutivos (2006 e 2007) na unidade experimental de produ??o integrada lavoura-pecu?ria denominada de Sistema Integrado de Produ??o Agroecol?gica (SIPA) ou Fazendinha Agroecol?gica km 47, localizada em Serop?dica (RJ), realizando experimentos que envolveram o cons?rcio couve (Brassica oleraceae var. acephala D.C.) com coentro (Coriandrum sativum L.) em compara??o com o monocultivo de couve, para determinar a diversidade de joaninhas por meio da coleta de indiv?duos atrav?s amostragens por remo??o e placas amarelas adesivas, bem como caracterizar a estrutura de sua comunidade por meio da an?lise faun?stica. O desempenho do cons?rcio foi tamb?m avaliado, determindo os par?metros fitot?cnicos das culturas associadas. Um total de 25 esp?cies de joaninhas est? presente no SIPA, mas Hyperaspis (Hyperaspis) festiva Mulsant, Scymnus (Pullus) sp.3, Cycloneda sanguinea (L.) e Eriopis connexa Germar foram as esp?cies mais frequentes na comunidade. N?o houve infesta??o por pulg?es na couve consorciada com coentro, o qual foi usado pelas joaninhas como s?tio de alimenta??o (fonte de p?len, n?ctar e presas alternativas), s?tio de oviposi??o, abrigo para larvas, pupas e adultos, al?m de s?tio de acasalamento. Quanto aos par?metros fitot?cnicos, o ?ndice de equival?ncia de ?rea (IEA) para os cons?rcios couve e coentro, considerando as produtividades de massa fresca, foram superiores em 92% (coentro colhido aos 55 dias ap?s a semeadura) e 85% (coentro deixado para florescer) em rela??o aos cultivos solteiros. O cons?rcio couve-coentro no arranjo de uma linha central de couve e quatro linhas paralelas de coentro, colhendo-se as duas centrais aos 55 dias ap?s a semeadura e deixando as duas linhas na borda do canteiro para florescer, n?o causa aumentos de produtividade da couve, mas mostra-se eficiente em termo do uso eficiente da terra e como estrat?gia de conserva??o de joaninhas nos sistemas de produ??o diversificados.

L'expansion des zones agricoles a conduit à la perte de biodiversité due à la réduction des habitats naturels et semi-naturels dans les paysages agricoles. Avec l'augmentation de la production agricole dans le monde, des techniques écologiquement rationnelles sont de plus en plus discutées qui permettent une gestion durable des habitats environnants. Les effets de ces habitats sur la population des insectes nuisibles et de leurs ennemis naturels sont encore mal connus. L'objectif de cette thèse était de comprendre les effets des environnements naturels et semi-naturels sur la population des insectes nuisibles et des ennemis naturels dans les régions tropicales (Brésil) et tempérées (France). Au Brésil (chapitre II), on a évalué l'effet de la distance des fragments sur la population de fourmis prédatrices et omnivores dans la canne à sucre. Les résultats montrent que la richesse en espèces diminue avec la distance des fragments forestiers et que la prédominance des espèces Dorymyrmex bruneus et Pheidole oxyops augmente. Des espèces de fourmis colonisant les champs de canne à sucre ont également été trouvées dans des fragments forestiers, ce qui suggère que ces derniers habitats sont des refuges pour les espèces de fourmis prédatrices pendant les périodes de perturbation comme la récolte de la canne à sucre ou le travail du sol. Cela a été confirmé par des différences plus fortes dans les communautés de fourmis après la récolte de la canne à sucre (saison sèche) que quatre mois plus tard / (saison des pluies) quand l'absence de perturbation a permis la re-lonosiation par les fourmis. Il y avait aussi une différence dans la richesse des espèces de fourmis entre les différents types de fragments (vallées fluviales et plaines). En France, on a évalué l'effet des bandes de fleurs sauvages, de la végétation spontanée et des bandes d'herbe sur la communauté des ennemis naturels et la régulation du puceron de la pomme rosâtre Dysaphis plantaginaea (chapitre III). En ce qui concerne les principaux ennemis naturels, nos résultats ont montré une densité plus élevée de hoverflies par rapport à d'autres types de bande, mais aucune différence pour les coccinelles (coccinelidae). Cependant, aucune différence de densité naturelle de l'ennemi n'a été observée à l'intérieur des vergers. Le nombre de pucerons était plus élevé près des marges, ce qui suggère que la colonisation à partir des bandes marginales peut contrecarrer les effets régulateurs positifs des ennemis naturels. Les effets positifs de la végétation de la marge de bande sur la régulation des ravageurs de la pomme nécessitent un mouvement d'ennemis naturels dans le verger. Nous avons testé les mouvements des prédateurs généralistes en étiquetant les bandes de marge avec des protéines d'oeufs et nous avons vérifié si elles se nourrissaient d'insectes nuisibles en utilisant l'analyse de marqueurs génétiques des gènes de papillon (Cydia pomonella) à l'intérieur des prédateurs (chapitre IV). Les résultats ont montré que peu de personnes se déplaçaient des marges de champ dans le verger. Cependant, 25% des prédateurs capturés se nourrissaient de C. pomonella indiquant encore un haut niveau de régulation naturelle. En conclusion, le mouvement limité des prédateurs à partir des bandes de marges dans les vergers peut expliquer l'absence de différences entre les traitements de bandes dans la régulation des ravageurs du verger (pucerons). Les habitats naturels et semi-naturels peuvent contribuer à la lutte contre les ravageurs en marge des cultures, mais dans les vergers de pommiers cet effet a fortement diminué avec la distance. / The expansion of agricultural areas has led to the loss of biodiversity due to the reduction of natural and semi-natural habitats in agricultural landscapes. With the increase of agricultural production in the world, environmentally sound techniques are increasingly discussed that allow a sustainable management of surrounding habitats. The effects of these habitats on the population of insect pests and their natural enemies are still poorly understood. The objective of this thesis was to understand the effects of natural and semi-natural environments on the population of insect pests and natural enemies in tropical (Brazil) and temperate (France) regions. In Brazil (Chapter II) the effect of the distance of fragments on the population of predatory and omnivorous ants in sugarcane was evaluated. The results showed that the species richness decrease with distance from forest fragments and that the dominance of the species Dorymyrmex bruneus and Pheidole oxyops increase. Ant species colonizing sugarcane fields were also found in forest fragments suggesting that the latter habitats are shelters for predatory ant species during periods of disturbance such as sugarcane harvest or soil tillage. This was confirmed by stronger differences in ant communities after sugarcane harvest (dry season) than four months later /(rainy season) when absence of disturbance allowed re-colonization by ants. There was also a difference in the richness of ant species between different fragment types (river valleys and plateaus). In France, the effect of wildflower strips, spontaneous vegetation and grass strips on the community of natural enemies and the regulation of the rosy apple aphid Dysaphis plantaginaea (Chapter III) were evaluated. Concerning major natural enemies, our results showed a higher density of hoverflies compared with other strip types but no difference for ladybirds (coccinelidae). No differences in natural enemy densities were observed inside orchards. Aphid number was higher close to the margins suggesting that colonization from margin strips may counteract positive regulatory effects of natural enemies. Positive effects of strip margin vegetation on regulation of apple pests require a movement of natural enemies into the orchard. We tested the movements of generalist predators by labelling margin strips with egg protein and we checked whether they fed on pest insects using genetic marker analysis of codling (Cydia pomonella) moth genes inside predators (chapter IV). The results showed that few individuals were moving from the field margins into the orchard. However, 25% of the captured predators fed on C. pomonella still indicating a high level of natural regulation. In conclusion, the limited movement of predators from margin strips into orchards may explain the absence of differences between strip treatments in orchard pest regulation (aphids). Natural and semi-natural habitats can contribute to pest control at the margins of crops, but in apple orchards this effect strongly decreased with distance.

Organic mulches and cover crops mulched in situ were assessed for their effects on B. cinerea primary inoculum and disease levels in inflorescences at flowering and/or bunches at harvest. Organic mulches were used to enhance biological degradation of vine debris to reduce levels of B. cinerea primary inoculum the following season. Four mulch types (anaerobically and aerobically fermented marc (grape pressings), inter-row grass clippings and shredded office paper) were applied under ten-year-old Riesling vines in a ten-replicate randomized block design in New Zealand over two consecutive years. Plastic mesh bags, each containing naturally infected vine debris, were placed under vines on bare ground (control) and at the soil-mulch interface, in winter (July) 2003 and 2004. In each year, half the bags were recovered at flowering (December) and the remainder at leaf plucking (February), for assessment of B. cinerea sporulation from the vine debris and debris degradation rate. Bait lamina probes, which measure soil biological activity, were placed in the soil-mulch interface three weeks before each of the two bag-recovery dates in both years and were then removed and assessed at the same times as were the bags. All mulches led to a reduction in B. cinerea sporulation. This reduction was significantly correlated with elevated rates of vine debris decomposition and increased soil biological activity. Over both years, compared with the controls, all treatments gave a 3-20-fold reduction in B. cinerea sporulation, a 1.6-2.6-fold increase in vine debris degradation and in the two marc and the paper treatments, a 1.8-4-fold increase in activity of soil organisms. The mulches also altered vine characteristics and elevated their resistance to B. cinerea through changes to the soil environment. Functional soil biological activity, as measured by Biolog Ecoplates and bait lamina probes, was increased 2-4 times in the two marc and paper treatments, compared with the control, an effect relating to the elevated soil moisture and reduced temperature fluctuations under these mulches. Soil nutrient levels and the C:N ratios were also affected in these treatments. The mulched paper lowered vine canopy density by up to 1.4 times that of the other treatments, an effect which probably led to elevated light penetration into the canopy and consequent increased canopy temperature, photosynthesis and lowered canopy humidity. These changes to soil and vine characteristics increased grape skin strength by up to 10% in the paper treatment and sugar concentrations by 1.2-1.4 °Brix in the two marc and paper treatments. The severity of B. cinerea infections in the anaerobic marc, aerobic marc and paper treatments were reduced to 12%, 3% and 2.2% of the control, respectively, in field assessments averaged over two consecutive harvests. Cover crops mulched in situ had similar effects to those of the organic mulches, increasing soil biological activity and reducing B. cinerea primary inoculum and the severity of B. cinerea infection in grapes at harvest (2006). Inter-row phacelia and ryegrass were mulched in winter 2005 and compared with a bare ground control, under 10-year-old Chardonnay vines in a ten-replicate randomized block design. Functional soil biological activity increased by 1.5-4.5 times in the two cover crop treatments compared with the control, an effect possibly related to elevated soil moisture in these treatments. This increase in soil moisture and soil biological activity increased vine debris degradation, reduced B. cinerea primary inoculum on the debris and decreased B. cinerea severity at flowering (December 2005) and harvest (April 2006). These results show the potential of organic mulches and cover crops mulched in situ to enhance soil ecosystem services and improve the sustainability of viticultural practices.

Trichogramma carverae Oatman and Pinto is mass-released for biological control of the leafroller pest, light brown apple moth, Epiphyas postvittana (Walker) in Australian vineyards. Parasitoid performance can, however, be constrained by a lack of suitable adult food and no information is available on the effect of nectar on the parasitism and longevity of T. carverae. To address this, the effect of alyssum, Lobularia maritima (L.) flowers on E. postvittana parasitism was studied in a vineyard experiment with and without releases of T. carverae. Egg parasitoid activity was assessed with E. postvittana egg �sentinel cards� and no parasitism was recorded in plots without T. carverae releases. Where T. carverae were released, there was no significant enhancement of parasitism by the presence of L. maritima flowers. Three hypotheses were subsequently tested to account for the lack of an effect: (i) T. carverae does not benefit from L. maritima nectar, (ii) T. carverae was feeding on nectar from other flowering plants (weeds) present in the vineyard, (iii) T. carverae was feeding on sugars from ripe grapes. A growth-cabinet experiment using potted L. maritima plants with and without flowers did not support hypothesis one. No parasitism was recorded after day two for T. carverae caged without flowers whilst parasitism occured until day eight in the presence of flowers. A laboratory experiment with common vineyard weeds (Trifolium repens, Hypochoeris radicata, Echium plantagineum) as well as L. maritima did not support hypothesis one but gave partial support to hypothesis two. Survival of T. carverae was enhanced to a small but statistically significant extent in vials with intact flowers of L. maritima, white clover (T. repens) and catsear (H. radicata) but not in vials with flowering shoots of these species from which flowers and flowering buds had been removed. Paterson�s curse (E. plantagineum) flowers had no effect on T. carverae survival. In a laboratory study, punctured grapes significantly enhanced T. carverae survival compared with a treatment without grapes, supporting hypothesis three. Trichogramma carverae performance in the field experiment was probably also constrained by relatively cool and wet weather. Further work on the enhancement of T. carverae efficacy by L. maritima and other carbohydrate sources is warranted. Greenhouse and field experiments were conducted to investigate whether T. carverae benefit from different groundcover plant species. Ten T. carverae adults (<24h after eclosion) were caged with different groundcover species and a control with no plant materials. Epiphyas postvittana egg sentinel cards were used to measure parasitism and longevity was recorded visually. Survival and realised parasitism of T. carverae was significantly higher in L. maritima than in Brassica juncea, Coriandrum sativum, shoots of these species from which flowers had been removed and nil control treatments. A similar experiment with Fagopyrum esculentum (with- and without-flowers) and a control treatment showed that survival was significantly higher in intact F. esculentum than in without-flower and control treatments. There was no significant treatment effect on parasitism in the early stages of that experiment, though parasitism was recorded in the presence of F. esculentum flowers for 12 days, compared with 6 days in other treatments. Higher parasitism was observed in intact Borago officinalis than in the flowerless shoot, water only and no plant material control treatments in a third experiment. There was no significant treatment effect on parasitism. Fitted exponential curves for survival data differed significantly in curvature in the first, second and third experiments but the slope was a non-significant parameter in the second and third experiments. In a second series of laboratory experiments, one male and one female T. carverae were caged with groundcover species to investigate male and female longevity and daily fecundity. Both male and female longevity in F. esculentum and L. maritima treatments were significantly higher than on shoots of these species from which flowers had been removed, and than in the control treatments. Daily fecundity was significantly greater in the intact L. maritima treatment than in all other treatments. Fitted exponential curves for daily fecundity differed significantly in position and slope but not in curvature. There was no significant treatment effect on longevity or parasitism when a male and female were caged with intact B. juncea, B. officinalis or without-flower of these species, nor in the treatment with no plant materials. No parasitism was observed in a survey of naturally occurring egg parasitoids on two sites close to Orange and Canowindra in New South Wales, illustrating the importance of mass releases of T. carverae in biological control of E. postvittana. In an experiment on the Canowindra site, parasitism was significantly higher on day one and day two after T. carverae release when with-flower treatments were compared with without-flower treatments. Parasitism was significantly higher in the F. esculentum treatment than in C. sativum, L. maritima, vegetation without-flowers and control treatments on these dates. On day five, parasitism was higher in C. sativum than in all other treatments. There was no significant increase in parasitism in a second experiment conducted on the Orange site. Coriandrum sativum, F. esculentum and L. maritima appear to be suitable adult food sources for T. carverae and offer some scope for habitat manipulation in vineyards The adults of many parasitoid species require nectar for optimal fitness but very little is known about flower recognition. Flight cage experiments showed that the adults of T. carverae benefited from L. maritima bearing white flowers to a greater extent than was the case for light pink, dark pink or purple flowered cultivars, despite all cultivars producing nectar. Survival and realised parasitism on non-white flowers were no greater than when the parasitoids were caged on L. maritima shoots from which flowers had been removed. The possibility that differences between L. maritima cultivars were due to factors other than flower colour, such as nectar quality, was excluded by dyeing white L. maritima flowers by placing the roots of the plants in 5% food dye (blue or pink) solution. Survival of T. carverae was lower on dyed L. maritima flowers than on undyed white flowers. Mixing the same dyes with honey in a third experiment conducted in the dark showed that the low level of feeding on dyed flowers was unlikely to be the result of olfactory or gustatory cues. Flower colour appears, therefore, to be a critical factor in the choice of plants used to enhance biological control, and is likely to also be a factor in the role parasitoids play in structuring invertebrate communities. Provision of nectar producing plants to increase the effectiveness of biological control is one aspect of habitat manipulation, but care needs to be taken to avoid the use of plant species that may benefit pest species. Greenhouse experiments were conducted to investigate whether the adult E. postvittana and larvae benefit from nectar producing groundcover species. Newly emerged E. postvittana adults were caged with different groundcover species and a honey-based artificial adult diet. The longevity of male and female E. postvittana when caged with shoots of borage (B. officinalis) and buckwheat (F. esculentum) bearing flowers was as long as when fed a honey-based artificial diet. This effect was not evident when caged with shoots of these plants from which flowers had been removed. Longevity was significantly lower than in the artificial diet treatment when caged with coriander (C. sativum) or alyssum (L. maritima) irrespective of whether flowers were present or not. There was no significant treatment effect on the lifetime fecundity of E. postvittana. A second experiment with mustard (B. juncea) (with- and without-flowers), water only and honey-based artificial adult diet showed no significant treatment effects on the longevity of male and female E. postvittana or on the lifetime fecundity of E. postvittana. The anomalous lack of a difference between the water and honey-based diet treatments precludes making conclusions on the value of B. juncea for E. postvittana. Two greenhouse experiments were conducted to evaluate the effects of groundcover species on the larval development of E. postvittana. In the first experiment, larval mortality was significantly higher in C. sativum, and L. maritima than in B. juncea, B. officinalis and white clover (T. repens) a known host of E. postvittana. Coriandrum sativum and L. maritima extended the larval period. In B. juncea and B. officinalis, mortality did not differ from that in T. repens. In F. esculentum, larval mortality was significantly higher than in T. repens. A short larval period was observed on B. juncea, B. officinalis and F. esculentum. Fitted exponential curves for larval mortality differed significantly in curvature between plant treatments. Similarly, successful pupation was significantly lower in C. sativum, F. esculentum and L. maritima than in T. repens. The percentage of successful pupation in B. juncea and B. officinalis did not differ from F. esculentum and T. repens. Fitted exponential curves for pupation differed significantly in curvature. A similar trend was observed in a second experiment with potted plants. The overall results suggest that C. sativum and L. maritima denied benefit to E. postvittana adults and larvae, so could be planted as vineyard groundcover with minimal risk of exacerbating this pest. Overall results suggest that T. carverae require nutrients to reach their full reproductive potential and flowers provide such nutrients. Lobularia maritima and C. sativum may be considered �selective food plants� for T. carverae whereas F. esculentum appears to be a �non-selective food plant�; both T. carverae and E. postvittana benefited from it. Fruits such as grapes can be used as food resources in habitat manipulation and this merits further research. This result also suggests that within species flower colour is an important factor for flower selection in habitat manipulation.

Trichogramma carverae Oatman and Pinto is mass-released for biological control of the leafroller pest, light brown apple moth, Epiphyas postvittana (Walker) in Australian vineyards. Parasitoid performance can, however, be constrained by a lack of suitable adult food and no information is available on the effect of nectar on the parasitism and longevity of T. carverae. To address this, the effect of alyssum, Lobularia maritima (L.) flowers on E. postvittana parasitism was studied in a vineyard experiment with and without releases of T. carverae. Egg parasitoid activity was assessed with E. postvittana egg �sentinel cards� and no parasitism was recorded in plots without T. carverae releases. Where T. carverae were released, there was no significant enhancement of parasitism by the presence of L. maritima flowers. Three hypotheses were subsequently tested to account for the lack of an effect: (i) T. carverae does not benefit from L. maritima nectar, (ii) T. carverae was feeding on nectar from other flowering plants (weeds) present in the vineyard, (iii) T. carverae was feeding on sugars from ripe grapes. A growth-cabinet experiment using potted L. maritima plants with and without flowers did not support hypothesis one. No parasitism was recorded after day two for T. carverae caged without flowers whilst parasitism occured until day eight in the presence of flowers. A laboratory experiment with common vineyard weeds (Trifolium repens, Hypochoeris radicata, Echium plantagineum) as well as L. maritima did not support hypothesis one but gave partial support to hypothesis two. Survival of T. carverae was enhanced to a small but statistically significant extent in vials with intact flowers of L. maritima, white clover (T. repens) and catsear (H. radicata) but not in vials with flowering shoots of these species from which flowers and flowering buds had been removed. Paterson�s curse (E. plantagineum) flowers had no effect on T. carverae survival. In a laboratory study, punctured grapes significantly enhanced T. carverae survival compared with a treatment without grapes, supporting hypothesis three. Trichogramma carverae performance in the field experiment was probably also constrained by relatively cool and wet weather. Further work on the enhancement of T. carverae efficacy by L. maritima and other carbohydrate sources is warranted. Greenhouse and field experiments were conducted to investigate whether T. carverae benefit from different groundcover plant species. Ten T. carverae adults (<24h after eclosion) were caged with different groundcover species and a control with no plant materials. Epiphyas postvittana egg sentinel cards were used to measure parasitism and longevity was recorded visually. Survival and realised parasitism of T. carverae was significantly higher in L. maritima than in Brassica juncea, Coriandrum sativum, shoots of these species from which flowers had been removed and nil control treatments. A similar experiment with Fagopyrum esculentum (with- and without-flowers) and a control treatment showed that survival was significantly higher in intact F. esculentum than in without-flower and control treatments. There was no significant treatment effect on parasitism in the early stages of that experiment, though parasitism was recorded in the presence of F. esculentum flowers for 12 days, compared with 6 days in other treatments. Higher parasitism was observed in intact Borago officinalis than in the flowerless shoot, water only and no plant material control treatments in a third experiment. There was no significant treatment effect on parasitism. Fitted exponential curves for survival data differed significantly in curvature in the first, second and third experiments but the slope was a non-significant parameter in the second and third experiments. In a second series of laboratory experiments, one male and one female T. carverae were caged with groundcover species to investigate male and female longevity and daily fecundity. Both male and female longevity in F. esculentum and L. maritima treatments were significantly higher than on shoots of these species from which flowers had been removed, and than in the control treatments. Daily fecundity was significantly greater in the intact L. maritima treatment than in all other treatments. Fitted exponential curves for daily fecundity differed significantly in position and slope but not in curvature. There was no significant treatment effect on longevity or parasitism when a male and female were caged with intact B. juncea, B. officinalis or without-flower of these species, nor in the treatment with no plant materials. No parasitism was observed in a survey of naturally occurring egg parasitoids on two sites close to Orange and Canowindra in New South Wales, illustrating the importance of mass releases of T. carverae in biological control of E. postvittana. In an experiment on the Canowindra site, parasitism was significantly higher on day one and day two after T. carverae release when with-flower treatments were compared with without-flower treatments. Parasitism was significantly higher in the F. esculentum treatment than in C. sativum, L. maritima, vegetation without-flowers and control treatments on these dates. On day five, parasitism was higher in C. sativum than in all other treatments. There was no significant increase in parasitism in a second experiment conducted on the Orange site. Coriandrum sativum, F. esculentum and L. maritima appear to be suitable adult food sources for T. carverae and offer some scope for habitat manipulation in vineyards The adults of many parasitoid species require nectar for optimal fitness but very little is known about flower recognition. Flight cage experiments showed that the adults of T. carverae benefited from L. maritima bearing white flowers to a greater extent than was the case for light pink, dark pink or purple flowered cultivars, despite all cultivars producing nectar. Survival and realised parasitism on non-white flowers were no greater than when the parasitoids were caged on L. maritima shoots from which flowers had been removed. The possibility that differences between L. maritima cultivars were due to factors other than flower colour, such as nectar quality, was excluded by dyeing white L. maritima flowers by placing the roots of the plants in 5% food dye (blue or pink) solution. Survival of T. carverae was lower on dyed L. maritima flowers than on undyed white flowers. Mixing the same dyes with honey in a third experiment conducted in the dark showed that the low level of feeding on dyed flowers was unlikely to be the result of olfactory or gustatory cues. Flower colour appears, therefore, to be a critical factor in the choice of plants used to enhance biological control, and is likely to also be a factor in the role parasitoids play in structuring invertebrate communities. Provision of nectar producing plants to increase the effectiveness of biological control is one aspect of habitat manipulation, but care needs to be taken to avoid the use of plant species that may benefit pest species. Greenhouse experiments were conducted to investigate whether the adult E. postvittana and larvae benefit from nectar producing groundcover species. Newly emerged E. postvittana adults were caged with different groundcover species and a honey-based artificial adult diet. The longevity of male and female E. postvittana when caged with shoots of borage (B. officinalis) and buckwheat (F. esculentum) bearing flowers was as long as when fed a honey-based artificial diet. This effect was not evident when caged with shoots of these plants from which flowers had been removed. Longevity was significantly lower than in the artificial diet treatment when caged with coriander (C. sativum) or alyssum (L. maritima) irrespective of whether flowers were present or not. There was no significant treatment effect on the lifetime fecundity of E. postvittana. A second experiment with mustard (B. juncea) (with- and without-flowers), water only and honey-based artificial adult diet showed no significant treatment effects on the longevity of male and female E. postvittana or on the lifetime fecundity of E. postvittana. The anomalous lack of a difference between the water and honey-based diet treatments precludes making conclusions on the value of B. juncea for E. postvittana. Two greenhouse experiments were conducted to evaluate the effects of groundcover species on the larval development of E. postvittana. In the first experiment, larval mortality was significantly higher in C. sativum, and L. maritima than in B. juncea, B. officinalis and white clover (T. repens) a known host of E. postvittana. Coriandrum sativum and L. maritima extended the larval period. In B. juncea and B. officinalis, mortality did not differ from that in T. repens. In F. esculentum, larval mortality was significantly higher than in T. repens. A short larval period was observed on B. juncea, B. officinalis and F. esculentum. Fitted exponential curves for larval mortality differed significantly in curvature between plant treatments. Similarly, successful pupation was significantly lower in C. sativum, F. esculentum and L. maritima than in T. repens. The percentage of successful pupation in B. juncea and B. officinalis did not differ from F. esculentum and T. repens. Fitted exponential curves for pupation differed significantly in curvature. A similar trend was observed in a second experiment with potted plants. The overall results suggest that C. sativum and L. maritima denied benefit to E. postvittana adults and larvae, so could be planted as vineyard groundcover with minimal risk of exacerbating this pest. Overall results suggest that T. carverae require nutrients to reach their full reproductive potential and flowers provide such nutrients. Lobularia maritima and C. sativum may be considered �selective food plants� for T. carverae whereas F. esculentum appears to be a �non-selective food plant�; both T. carverae and E. postvittana benefited from it. Fruits such as grapes can be used as food resources in habitat manipulation and this merits further research. This result also suggests that within species flower colour is an important factor for flower selection in habitat manipulation.

Le développement d’une agriculture moins dépendante des pesticides repose en partie sur la maximisation des services écosystémiques. L’objectif principal de cette thèse était de quantifier le service de régulation des insectes ravageurs par leurs ennemis naturels, en conditions expérimentales et en situation de production. Le travail a porté sur les brassicacées légumières soumises aux attaques de ravageurs aériens et souterrains (pucerons et mouche du chou). Nous avons évalué leur nuisibilité pour quantifier les dégâts et dommages qu’ils peuvent générer. Nos résultats confirment la moindre nuisibilité des pucerons par rapport à celle de la mouche du chou. Pour celle-ci, la nuisibilité sur légume-fleur s’exprime non seulement par la mortalité des plants mais aussi par un effet sur leur développement.Nous avons également montré que les prédateurs épigés réduisaient significativement les populations des deux ravageurs, ce qui, pour la mouche du chou, conduit à une baisse conséquente de la mortalité des plants. Enfin, simplifier le travail du sol au printemps n’a pas permis de favoriser les prédateurs hivernant dans la parcelle, ni d’augmenter le service de régulation à court terme. Cependant les prédateurs précoces sont déterminants et doivent être favorisés par des pratiques intra- et péri-parcellaires ciblées. Les connaissances acquises dans ce travail démontrent le potentiel de la lutte biologique par conservation pour une gestion agro-écologique efficace des insectes ravageurs. / Developping farming systems that use less pesticides is partly based on the intensification of ecosystem services. The main aim of this thesis was to quantify the insect pest regulation service provided by their natural enemies, both in experimental plots and in farmers fields. We worked on brassicaceous vegetables that are confronted to pest attacks on above- and belowground organs (aphids and the cabbage root fly). We assessed their harmfulness, in order to quantify the injuries and damage these pests can inflict to plants. Our results confirm that aphids are less harmful than the cabbage root fly. On flower vegetables, the latter not only leads to plant mortality but can also slow plant development.We also showed that ground dwelling predators significantly reduce the populations of both pests, leading, for the cabbage root fly, to a strong decrease in plant mortality rate. Finally, simplifying soil tillage in the spring did not favor ground dwelling predators overwintering inside the field and neither did it increase natural regulation service in the short term. However, early ground dwelling predators are fundamental and must be favored by tailored practices both inside and around the fields. The knowledge gained in the present thesis show the potential of conservation biological control for the agroecological management of insect pest populations.

Le pou rouge Dermanyssus gallinae est un acarien d’importance sanitaire et économique majeure en élevage de poules pondeuses partout dans le monde mais paradoxalement le fonctionnement de son écosystème a été relativement peu étudié. Hématophage strict, D. gallinae n’est pas un parasite typique, il entretient une relation lâche mais obligatoire avec son hôte. Confiné dans l’environnement proche de son hôte (nids, fientes sèches, litière, fumier, . . .) D. gallinae est fortement soumis aux interactions avec les autres organismes qui peuplent ces microhabitats. Ce travail a pour objectif d’apporter des éléments pertinents de caractérisation de cet écosystème singulier mal connu pour, au delà de la production de connaissance, alimenter une réflexion sur le développement de stratégies de gestion intégrée innovantes afin de s’affranchir de la stricte lutte chimique qui reste à l’heure actuelle le moyen de luttele plus largement utilisé. L’étude des interactions chimiques entre le pou rouge et son hôte a permis de mieux identifier les sources de stimulus et mieux comprendre les réponses de D. gallinae aux odeurs de poule ainsi que de déterminer dans quelle mesure il était possible d’interférer entre l’hôte et son microprédateur au moment de l’étape de repérage. La caractérisation de la structure des communautés d’arthropodes partageant les mêmes microhabitats que le pou rouge a montré que D. gallinae côtoyait plusieurs espèces d’acariens et insectes prédateurs d’arthropodes qui sont potentiellement ses ennemis naturels. Il a également été mis en évidence que le bâtiment d’élevage constituait une unité propre en termes de composition et de structure des communautés d’arthropodes. Les différences significatives de fréquence et d’occurrencedes différentes espèces prédatrices entre bâtiments, ainsi que leur capacitéde propagation depuis l’environnement extérieur constituent des indicateurs prometteurs pour le développement de la lutte biologique par conservation contre le pou rouge. Des questions transversales quant aux effets croisés, synergiques ou antagonistes, entre la manipulation des odeurs de l’hôte et la mise à contribution des processus écologiques impliquant les arthropodes non hématophages sont posées en vue d’une démarche intégrant ces outils ainsi que les autres moyens de contrôle disponibles dans un contexte de protection biologique intégrée du pou rouge. / The Poultry Red Mite (PRM) Dermanyssus gallinae is a mite of major sanitary andeconomic importance for the egg production industry worldwide but paradoxicallythe functioning of its ecosystem has been little studied. The objective of this work was to provide relevant elements for the characterization of this singular ecosystem that is poorly known, in order to, beyond the production of knowledge, contribute to the development of innovative strategies of integrated management. This is expected to allow going further than the strict chemical control which is still to date the most widely used means of control. The study of the chemical interactions between the PRM and its host allowed to better identify sources of stimulus and to better understand D. gallinae’s responses to chicken odors as well as to determine to what extent it was possible to interfere between the host and its micropredator during the first phase of the host location. The characterization of the structure of arthropod communities sharing the same microhabitats as does PRM showed that D. gallinae was associated with several species of mites and predatory insects of arthropods which could potentially be its natural enemies. It was also demonstrated that the livestock building was a specific unit in terms of the composition and structure ofthe arthropod communities. Significant differences in the frequency and occurrence of predatory species between buildings and their propagation capacity from the external environment are promising indicators for the development of conservation biological control (CBC) against PRM. A reflection on the possible synergistic and antagonistic cross-effects between the manipulation of host odors and the stimulation of ecological processes involving non-haematophagous arthropods is carried out. This is expected to participate in the development of an approach integrating these tools as well as other available means of control in a context of integrated biological protection of PRM.