The biology and ecology of an indigenous, non-pest insect on an indigenous host plant were investigated. The study includes a taxonomic description of this new species, an account of its general biology, monitoring of the seasonal fluctuations in numbers of the psyllid and its hymenopterous parasitoids, and a study of various aspects of the host plant, including nitrogen levels, water stress, leaf hardness, and the effect of cutting, in relation to spatial and temporal differences in insect population numbers. Natural enemies, inter- and intra-specific competition had very little determining influence on the population numbers of the psyllid. Temperature and humidity had little direct effect, but may have influenced the population dynamics via its effects on the host plant. Seasonal patterns in psyllid numbers followed fluctuations in nitrogen levels; statistically valid correlations were obtained between leaf nitrogen and psyllid numbers on individual trees at various times. These findings were supported by the results of preliminary laboratory experiments employing different fertilizer treatments. No effect of water stress or leaf hardness was clearly discerned. Cutting of trees altered the characteristics of the subsequent regenerative growth so as to allow massive psyllid infestations to develop, thus showing the tremendous importance of the host plant in determining population levels in this insect. Preliminary investigations of the nature and mechanism of this effect were conducted, and its significance is discussed. The relevance of these findings to modern concepts of regulation in insect populations and to principles of pest management is discussed.

Mealybugs (Hemiptera: Pseudococcidae) cause severe damage to many commercial crops, including grapevine. This is largely because of their ability to transmit various grapevine viral diseases, in particular grapevine leafroll-associated viruses (GLRaVs). Grapevine leafroll is one of the most wide-spread grapevine diseases worldwide. Managing the field-spread of grapevine leafroll disease requires, amongst others, stringent mealybug control. Mealybug monitoring and control methods rely on timely and accurate identification of the species present. However, proper identification of mealybug species is problematic, time-consuming and requires an expert taxonomist. In most cases, only adult females can be reliably identified morphologically. Immature insects, males and damaged specimens cannot be assigned to species. In this study, a molecular method was developed to rapidly and accurately distinguish three mealybug species associated with grapevine, namely the vine mealybug Planococcus ficus (Signoret), the citrus mealybug Planococcus citri (Risso) and the longtailed mealybug Pseudococcus longispinus (Targioni-Tozzetti). During the development of this identification method, a number of tasks were undertaken. Firstly, rapid and reliable DNA extraction methods were tested for mealybug DNA. Two rapid extraction methods were adapted and tested, namely the direct buffer method and the spot-PCR method. These methods reliably extracted DNA even from very small or damaged individuals, and could be performed in 15-20 minutes and three hours, respectively. Secondly, mealybug mitochondrial DNA from the cytochrome c oxidase subunit 1 (CO I) gene was amplified and sequenced. It was found that DNA from the 3’-end of CO I showed minimal intraspecific variation (<1%), but sufficient interspecific variation (7-12%) to clearly delineate species. This region was then used to develop three species-specific forward primers, which were used in conjunction with a common universal reverse primer. These primers were all used in a multiplex PCR to differentially amplify DNA from each of the three species. The primers were designed such that each yielded a DNA product of different length which could be separated by electrophoresis on an agarose gel. In this manner the identity of the species could be determined. The entire identification protocol (including extraction, PCR and electrophoresis) could be completed in approximately four hours. All amplified specimens in a blind trial were correctly identified, regardless of size or condition of the specimen. The protocol is simple enough to be implemented in any molecular laboratory. This represents a considerable improvement over currently available techniques for mealybug identification, and is certain to be of great use in diagnostic identification of mealybugs in vineyards and export consignments. / Dissertation (Magister Scientiae)--University of Pretoria, 2007. / Zoology and Entomology / unrestricted

The inadequacy of arbitrarily classifying interactions between species as antagonistic, neutral or mutualistic has become clear in recent years. Both direct and indirect interactions between species can vary between mutualism and antagonism, depending on the intrinsic and extrinsic contexts of the interaction. This study investigated the characteristics of an ant-plant-homopteran interaction in southern Africa. The polyphagous homopteran Hilda patruelis (Tettigometridae) feeds primarily on the trunk-borne fruiting branches and figs of the Cape fig tree, Ficus sur, and produces honeydew which attracts tending ants. Ten of the sixteen ant species/species groups present on F. sur tended H. patruelis, with Pheidole megacephala the most frequent attendant. Ants attracted to F. sur by H. patruelis honeydew or other liquid food sources also preyed on insects on the tree, including adults of the small agaonid fig wasps whose larvae feed on the ovules in the developing figs. One fig wasp species (Ceratosolen capensis) is also the tree's only pollinator. No benefits to H. patruelis from being tended by ants were detected, either in terms of reduced parasitism, or predation by a lycaenid caterpillar. A P. megacephala colony foraging on a F. sur tree was found to receive a high proportion of its likely energy requirements from the tree, mainly in the form of H. patruelis honeydew, during periods when it was bearing fruit. It is probable that the H. patruelis-P. megacephala interaction constitutes a direct mutualism at times, but that benefits to the homopteran are intermittent or weak. Both H. patruelis and ants benefitted from F. sur, directly or indirectly, through the provision of food (and for some ants, nesting sites). The removal of phloem sap by H. patruelis did not detectably reduce the trees' reproductive output, either in terms of pollinator or viable seed production. The indirect effects of ant and H. patruelis presence on the F. sur trees were on average positive, as ants preyed disproportionately heavily on fig wasp species parasitic on or competing with the pollinator, thus increasing pollinator production. Effects of ant presence on seed production were not investigated, but have been demonstrated as beneficial elsewhere. However, there is great varatlon both in the composition of the wasp fauna arriving to oviposit at different crops, and in ant densities per fig, on several temporal and spatial scales. This results in high variability in the effects of ants on the pollinator and, through it, the tree, from positive to zero and potentially even negative. Despite this conditionality of beneficial outcomes for the tree, the mean effect of ants on the F. sur population studied was to increase pollinator production by up to nearly 20%. This study is among the few to have demonstrated an overall benefit to a plant of having homopteran-tending ants present on it.

Selected insecticides were evaluated to determine effects on the potato leafhopper and its predators. Three types of tests were conducted. First, the common predators of the potato leafhopper in alfalfa were examined to determine their response to the stages of the leafhopper. Second, under field conditions, insecticides were tested to detect effects on the populations of the insects. Finally, LC₅₀ toxicities of selected insecticides were investigated under laboratory conditions. In the predation tests, two studies were conducted, namely, predation on egg stage and predation on nymphal and adult stages of the potato leafhopper. An indirect method of determining egg predation was employed because potato leafhopper eggs are not visible, being concealed in the xylem and phloem tissues. Using two sets of treatments (with and without predators), the potato leafhopper females were allowed to oviposit in predation cages and the suspected predators introduced thereafter. Hatching nymphs were counted and used as indications of the number of eggs laid. Nabis americoferus and Orius insidiosis were found to be predaceous on eggs of the potato leafhopper. In the predation tests, using nymphal and adult stages of the potato leafhopper, the following adult and larval stages of insects were confirmed as predators: Hippodamia convergens, Coccinella novemnotata, and Chrysopa carnea. Two field tests, one in summer of 1977 and one in summer of 1978, were conducted to determine the influence of insecticides on populations of potato leafhoppers and its predators in alfalfa. Samples were taken using a 38 cm sweep net employing a 180° motion at 2 days, 1 week, and 2 weeks after treatment. In both tests, significant control was provided by the chemicals against potato leafhoppers 48 hours following application. A gradual population recovery occurred one week and two weeks after treatment. Most of the entomophagous insects sampled were adversely affected by insecticides. No statistical differences in yield and percent dry matter were obtained with various insecticidal treatments. The acute toxicities of selected insecticides on potato leafhopper, tarnished plant bug, and damsel bug were investigated. The bioassay method used was the immersion technique and toxicity was expressed in LC₅₀. The relative toxicities of Nabis americoferus to 50° Empoasca fabae and Lygus lineolaris were obtained by deriving the selectivity ratios. Selectivity ratios are calculated by dividing the LC₅₀ of the non-target organism divided by the LC₅₀ of the pest. The ratios indicated that most of the insecticides tested were highly selective. Among the organophosphates, azinphosmethyl was the most selective, while among the carbamates, carbaryl was the most selective. / Master of Science

Two ant species, Pheidole megacephala (Fabricius) and Anoplolepis custodiens (Smith) reach pest status in citrus orchards through precipitating outbreaks of homopterous pests. However, predacious ants, including these two ant species, play an important role in pest suppression in agroecosystems and are therefore beneficial to these systems. If A. custodiens and P. megacephala are important natural control agents in citrus, using ant bands to break the mutualism between the ants and the Homoptera as a method of ant control is preferable to poisoning. Ant communities were sampled by pitfall trapping in three experimental subunits of 2-, 4-, 15- and 30-year-old citrus orchards, in the Grahamstown area of the Eastern Cape. In one subunit in each orchard, populations of P. megacephala and A. custodiens were suppressed by poison applications. In a second subunit, trees were banded with trunk barriers so that ants were prevented from foraging in the trees and a third subunit served as the untreated control. Bait pupae of bollworm, false codling moth and fruit fly were planted in bait trays in all of the subunits to investigate predation on these citrus pests in the relative absence of predacious ants and where they were excluded from the trees. Pheidole megacephala dominated exclusively in all of the plots. Community composition did not change dramatically with increasing age of the trees, but species diversity and species abundance did. Rank-abundance curves showed that community diversity was highest in the 2-year-old plots and lowest in the 30-year-old plots. The Simpson and Shannon-Wiener diversity indices and their evenness measures indicated that diversity and equitability were highest in the poisoned subunits and lowest in the banded subunits. Principle component analysis revealed that the poisoned subunits were similar and distinct in species composition, that there was significant monthly variation in species composition and that community stability increases with an increase in orchard age. The presence of P. megacephala was significantly negatively correlated (rs = -0.293; p < 0.001) with pest pupal survival. Pupal survival was significantly higher for bollworm (p < 0.001), FCM (p < 0.001) and fruit fly (p < 0.001) in the poisoned subunits, than in the banded and control subunits. There was a general trend for survivorship to increase with an increase in the age of the trees. A significant difference (p < 0.001) was found between the months in which the trials were carried out. Pupal survival was significantly lower (p < 0.001) for FCM than for bollworm and fruit fly. In citrus orchards, ant communities are organised by ecological processes and interactions and are influenced by methods of ant control. Ant bands are preferable to poisoning as a method of ant control, so that beneficial species are left on the ground to prey on pests that pupate in the soil. Maintaining high ant species diversity in citrus orchards would be beneficial as predation on the pupae was more effective where ant species diversity was higher.

A couve-flor está entre as folhosas mais produzidas na região do Cinturão Verde de São Paulo. A planta apresenta alto valor nutricional para os consumidores e relevante importância econômica e social para os agricultores. Em campos comerciais, têm sido observadas plantas exibindo sintomas de enfezamento e deformação da inflorescência, além da necrose dos vasos condutores. A doença foi denominada de enfezamento e a sua incidência e severidade têm se tornado mais intensas nos últimos anos, afetando a produção e levando ao abandono do cultivo. O quadro sintomatológico levantou a suspeita de que a doença pudesse estar sendo causada por fitoplasmas. Desta forma, este trabalho teve como objetivos detectar e identificar os fitoplasmas associados às plantas doentes de couve-flor, demonstrar a sua patogenicidade, buscar possíveis insetos vetores do agente patogênico e analisar a distribuição espacial da doença no campo. Amostras de plantas coletadas na região do Cinturão Verde ou enviadas de outras localidades tiveram o DNA total extraído e submetido a reações de PCR. A amplificação de fragmentos genômicos de 1,2 kb evidenciaram a presença de fitoplasmas nos tecidos de 66% das plantas sintomáticas. A detecção de fitoplasmas em plantas assintomáticas revelou a ocorrência de infecção latente e que plantas sem sintomas aparentes podem ser portadoras do patógeno. O emprego de PCR com primers específicos, análises convencionais e virtuais de RFLP e análise filogenética permitiram a identificação de fitoplasmas afiliados aos grupos 16SrIII-J, 16SrXIII e 16SrXV-A, com predominância para a ocorrência do grupo 16SrIII. Secções histológicas feitas à mão da região vascular de plantas de couve-flor doentes observadas em microscópio de luz revelaram que a região do floema apresentava-se escurecido e com início de degeneração celular. Fitoplasmas do grupo 16SrXIII foram encontrados em cigarrinhas da espécie Balclutha hebe coletadas em campo cultivado com couve-flor. Estes insetos foram utilizados em experimentos de transmissão e foram capazes de inocular fitoplasmas em plantas sadias de couve-flor e vinca. A transmissão também ocorreu a partir de plantas doentes de couve-flor para plantas sadias de vinca, através do emprego da planta parasita cuscuta. Análise epidemiológica da doença foi feita em dez campos de cultivo de couve-flor, localizados no município de Sorocaba-SP. Análises do índice de dispersão, lei de Taylor modificada e áreas isopatas evidenciaram que plantas de couveflor com sintomas de enfezamento encontraram-se agregadas no campo, e que os focos da doença tinham início nos bordos da cultura. O presente trabalho demonstrou que o enfezamento da couve-flor está associado aos fitoplasmas, os quais podem estar sendo transmitidos por cigarrinhas, sendo que, em termos epidemiológicos, a doença tem padrão agregado de distribuição no campo. / Cauliflower is among the most produced vegetable in the region of the Green Belt of Sao Paulo State. It has high nutritional value and relevant social and economic importance for farmers. In commercial fields, it has been observed plants exhibiting symptoms of stunting, deformation of the inflorescence and conductive vessel necrosis. The disease was named cauliflower stunt and its incidence and severity have become more intense in the past years, the production is being affected and the field is abandoned by the growers. The symptomatology raised the suspect that the disease could be caused by phytoplasmas. Thus, this study aimed to detect and identify phytoplasmas associated to diseased cauliflowers, demonstrate its pathogenicity, look for possible insect vectors of the pathogen and analyse the spatial distributions of disease in the field. Plant samples collected in the region of the green belt or sent from other locations had the total DNA extracted and subjected to PCR reactions. Amplification of genomic fragments of 1.2 kb revealed the presence of phytoplasmas in 66% of the symptomatic plants. The detection of phytoplasmas in asymptomatic plants revealed the occurrence of latent infection and that plants without visible symptoms may be infected. The use of specific primers, conventional and virtual RFLP analysis and phylogenetic analysis allowed the identification of phytoplasmas affiliated to 16SrIII-J, 16SrXIII and 16SrXV-A groups, and predominantly the occurrence of 16SrIII group. Histological sections made by hand of the vascular region of symptomatic cauliflower were observed under light microscope and revealed that the phloem was darkened and with beginning of cellular degeneration. The 16SrXIII phytoplasma was found in Balclutha hebe from cauliflower field. These insects were used in experiments and were able to transmit phytoplasmas to healthy cauliflower and periwinkle. The transmission also occurred from diseased cauliflower to healthy periwinkle through the parasitic plant dodder. Epidemiological analysis of the disease was made in ten plots of cauliflower, located in Sorocaba, SP. Dispersion index, modified Taylors law and isopath areas analysis showed that cauliflower plants with symptoms of stunting were aggregated in field and initial foci were at the edge of the plot. This study demonstrated that cauliflower stunt is associated with phytoplasmas, that may be being transmitted by insects (Homoptera) and, in epidemiological terms, the diseased plants show aggregated pattern of distribution in the field.

Nous avons recherché des protéines toxiques pour les pucerons (Homoptères, insectes piqueurs-suceurs phloémophages) et étudié le mode d'action de certaines d'entre elles. Ce travail constitue une première étape d'un programme de création de plantes résistantes aux pucerons par génie génétique. <br />Les caractéristiques toxicologiques de nombreuses protéines sont évaluées par des tests d'ingestion sur milieux artificiels définis. Des lectines d'origine végétale se liant au mannose présentent des propriétés toxiques intéressantes. Notre étude porte sur la Concanavaline A (lectine de Canavalia ensiformis [L.] DC, ConA) qui est une lectine modèle très étudiée du point de vue biochimique, et la lectine du perce-neige (Galanthus nivalis L., GNA) dont les caractéristiques en font un bon candidat à l'application envisagée. <br />Nous mettons en évidence une variabilité de la toxicité des lectines à mannose chez six espèces de pucerons. La ConA est moins active sur les espèces polyphages. Elle n'est pas phagorépulsive pour notre puceron modèle, Acyrthosiphon pisum (Harris) et agit en quelques heures aux doses moyennes, notamment en inhibant l'ingestion. Une adaptation comportementale à moyen terme (24 h-48 h) est également mise en évidence. Des techniques de marquage révèlent que la cible physiologique primaire de la ConA est la portion antérieure du mésentéron. Elle s'y fixe en très grande quantité. Après liaison aux cellules épithéliales, la lectine induit une hypertrophie de ces cellules et un détachement de leur membrane apicale. Des expériences de compétition lectines/mannosides indiquent que la liaison toxine-épithélium ne semble pas dépendre uniquement d'une interaction sucre-lectine. Nous observons également une forte perturbation du métabolisme des acides aminés des pucerons. Le mode d'action de la ConA et de la GNA diffèrent sensiblement sur ce point. <br />Les premiers tests biologiques effectués sur des pommes de terre transgéniques exprimant de manière constitutive le gène de la GNA sont variables mais prometteurs.

A couve-flor está entre as folhosas mais produzidas na região do Cinturão Verde de São Paulo. A planta apresenta alto valor nutricional para os consumidores e relevante importância econômica e social para os agricultores. Em campos comerciais, têm sido observadas plantas exibindo sintomas de enfezamento e deformação da inflorescência, além da necrose dos vasos condutores. A doença foi denominada de enfezamento e a sua incidência e severidade têm se tornado mais intensas nos últimos anos, afetando a produção e levando ao abandono do cultivo. O quadro sintomatológico levantou a suspeita de que a doença pudesse estar sendo causada por fitoplasmas. Desta forma, este trabalho teve como objetivos detectar e identificar os fitoplasmas associados às plantas doentes de couve-flor, demonstrar a sua patogenicidade, buscar possíveis insetos vetores do agente patogênico e analisar a distribuição espacial da doença no campo. Amostras de plantas coletadas na região do Cinturão Verde ou enviadas de outras localidades tiveram o DNA total extraído e submetido a reações de PCR. A amplificação de fragmentos genômicos de 1,2 kb evidenciaram a presença de fitoplasmas nos tecidos de 66% das plantas sintomáticas. A detecção de fitoplasmas em plantas assintomáticas revelou a ocorrência de infecção latente e que plantas sem sintomas aparentes podem ser portadoras do patógeno. O emprego de PCR com primers específicos, análises convencionais e virtuais de RFLP e análise filogenética permitiram a identificação de fitoplasmas afiliados aos grupos 16SrIII-J, 16SrXIII e 16SrXV-A, com predominância para a ocorrência do grupo 16SrIII. Secções histológicas feitas à mão da região vascular de plantas de couve-flor doentes observadas em microscópio de luz revelaram que a região do floema apresentava-se escurecido e com início de degeneração celular. Fitoplasmas do grupo 16SrXIII foram encontrados em cigarrinhas da espécie Balclutha hebe coletadas em campo cultivado com couve-flor. Estes insetos foram utilizados em experimentos de transmissão e foram capazes de inocular fitoplasmas em plantas sadias de couve-flor e vinca. A transmissão também ocorreu a partir de plantas doentes de couve-flor para plantas sadias de vinca, através do emprego da planta parasita cuscuta. Análise epidemiológica da doença foi feita em dez campos de cultivo de couve-flor, localizados no município de Sorocaba-SP. Análises do índice de dispersão, lei de Taylor modificada e áreas isopatas evidenciaram que plantas de couveflor com sintomas de enfezamento encontraram-se agregadas no campo, e que os focos da doença tinham início nos bordos da cultura. O presente trabalho demonstrou que o enfezamento da couve-flor está associado aos fitoplasmas, os quais podem estar sendo transmitidos por cigarrinhas, sendo que, em termos epidemiológicos, a doença tem padrão agregado de distribuição no campo. / Cauliflower is among the most produced vegetable in the region of the Green Belt of Sao Paulo State. It has high nutritional value and relevant social and economic importance for farmers. In commercial fields, it has been observed plants exhibiting symptoms of stunting, deformation of the inflorescence and conductive vessel necrosis. The disease was named cauliflower stunt and its incidence and severity have become more intense in the past years, the production is being affected and the field is abandoned by the growers. The symptomatology raised the suspect that the disease could be caused by phytoplasmas. Thus, this study aimed to detect and identify phytoplasmas associated to diseased cauliflowers, demonstrate its pathogenicity, look for possible insect vectors of the pathogen and analyse the spatial distributions of disease in the field. Plant samples collected in the region of the green belt or sent from other locations had the total DNA extracted and subjected to PCR reactions. Amplification of genomic fragments of 1.2 kb revealed the presence of phytoplasmas in 66% of the symptomatic plants. The detection of phytoplasmas in asymptomatic plants revealed the occurrence of latent infection and that plants without visible symptoms may be infected. The use of specific primers, conventional and virtual RFLP analysis and phylogenetic analysis allowed the identification of phytoplasmas affiliated to 16SrIII-J, 16SrXIII and 16SrXV-A groups, and predominantly the occurrence of 16SrIII group. Histological sections made by hand of the vascular region of symptomatic cauliflower were observed under light microscope and revealed that the phloem was darkened and with beginning of cellular degeneration. The 16SrXIII phytoplasma was found in Balclutha hebe from cauliflower field. These insects were used in experiments and were able to transmit phytoplasmas to healthy cauliflower and periwinkle. The transmission also occurred from diseased cauliflower to healthy periwinkle through the parasitic plant dodder. Epidemiological analysis of the disease was made in ten plots of cauliflower, located in Sorocaba, SP. Dispersion index, modified Taylors law and isopath areas analysis showed that cauliflower plants with symptoms of stunting were aggregated in field and initial foci were at the edge of the plot. This study demonstrated that cauliflower stunt is associated with phytoplasmas, that may be being transmitted by insects (Homoptera) and, in epidemiological terms, the diseased plants show aggregated pattern of distribution in the field.

Cette thèse porte sur l’analyse des dimensions spatiale et temporelle du fonctionnement et de la diversité taxonomique et fonctionnelle d’une guilde. Au sein d’une guilde, les espèces exploitent un même type de ressources; pour cette raison, les niches écologiques fondamentales des membres d’une guilde sont similaires et une relation de compétition interspécifique se met en place si la ressource partagée est limitante. Quatre questions principales sur le fonctionnement d’une guilde ont été abordées : (1) à quelles échelles spatio-temporelles se structurent la guilde et les ressources exploitées, (2) quelles sont les influences respectives des filtres environnementaux et des interactions interspécifiques sur les traits fonctionnels des membres de la guilde, (3) quelle est la stratégie d’exploitation de ressources d’un membre de la guilde face à une faible disponibilité de ressources et (4) quel est l’impact du contexte climatique sur la structure d’une guilde et de son réseau trophique, et sur le degré de spécialisation écologique des espèces de la guilde sur la ressource. Le modèle biologique de cette étude a été une guilde de parasitoïdes de pucerons de céréales (Hymenoptera : Braconidae : Aphidiinae). Ces parasitoïdes s’attaquent aux pucerons (Homoptera : Aphididae) inféodés aux céréales dans les agroécosystèmes. Les variations d’abondance relative des parasitoïdes et de leurs hôtes ont été importantes aux échelles interrégionale et interannuelle, mais très faibles à l’échelle intra-régionale. La divergence des traits fonctionnels des parasitoïdes s’est maintenue sur trois régions et deux années, et le contexte local a influencé les traits de la guilde dans son ensemble. Pendant la saison hivernale, une période marquée par une faible densité de pucerons, le parasitoïde Aphidius rhopalosiphi a présenté des stratégies contrastées pour maximiser sa fitness en exploitant les hôtes Sitobion avenae et Rhopalosiphum padi, mais une forte spécialisation écologique sur le terrain a été observée en présence d’une espèce compétitrice : Aphidius avenae. Cette présence a pu être corrélée à une hausse des températures hivernales. / This thesis is an analysis of the spatial and temporal dimensions of the functioning and taxonomic and functional diversity of a guild. In a guild, species exploit the same type of resources; consequently, fundamental ecological niches of guild members are similar and an interspecific competitive relationship is established if shared resources are limiting. Four main questions on the functioning of a guild are addressed: (1) the spatiotemporal scales to which guilds the guild and its exploited resources are structures, (2) the respective weight of environmental filtering and interspecific interactions on the configuration of realized niches of guild members, (3) the strategy of resource exploitation in a guild member facing low resource availability and (4) the impact of climatic context on the structure of a guild and its food web, and the degree of specialisation on resources. The biological model chosen in this study was a cereal aphid parasitoid guild (Hymenoptera: Braconidae: Aphidiinae). Those parasitoids attack aphids (Homoptera: Aphididae) in agroecosystems, which are anthropised environments marked by high disturbance rates and in which host resources are variable in density and in quality. Relative abundance variations in parasitoids and in their hosts were considerable at interregional and interannual scales, but weak at the intra-regional scale. Functional trait divergence in parasitoids was maintained across three regions and two years, and regional and annual environmental contexts influenced guild traits as a whole. In winter, a period marked by low aphid density, the parasitoid Aphidius rhopalosiphi presented contrasted strategies to maximise fitness whilst exploiting Sitobion avenae and Rhopalosiphum padi hosts, but a high degree of ecological specialisation in the field was observed in the presence of a competitive species, Aphidius avenae. This presence was correlated to an increase in winter temperatures.