L’évolution convergente des formes de croissance est un phénomène fondamental reliant l’écologie et l’évolution des plantes. Remarquablement illustré dans plusieurs systèmes insulaires, ce phénomène n’a jamais été identifié en Nouvelle-Calédonie, pourtant connue pour la richesse et l’originalité de sa flore. Par une approche combinant architecture des plantes, traits fonctionnels, taxonomie, phylogénie et données environnementales, cette thèse analyse l’histoire évolutive de la monocaulie, une forme de croissance mal connue, en Nouvelle-Calédonie. Les monocaules sont des plantes autoportantes ligneuses dont les fonctions majeures sont assurées par une seule tige apparente. En Nouvelle-Calédonie, elles sont représentées par 182 espèces dicotylédones appartenant à 41 genres et 30 familles et sont gravement menacées d’extinction. L’évolution répétée de la monocaulie en Nouvelle-Calédonie, issue d’au moins 31 événements d’apparition, est l’un des cas les plus remarquables de convergence en milieu insulaire. Dans le genre Atractocarpus, la monocaulie est apparue récemment deux à trois fois via diverses réductions des branches en inflorescences, montrant l’importance des processus hétérochroniques dans l’évolution des formes de croissance. La monocaulie est fortement corrélée à plusieurs traits démontrant des contraintes majeures dans la coordination fonctionnelle. L’évolution de la monocaulie est fortement associée aux forêts denses humides et au substrat ultramafique, et semble avoir contribué à la diversification des lignées par des phénomènes de partitionnement de niche. La remarquable convergence de la monocaulie en Nouvelle-Calédonie peut s’expliquer par quatre hypothèses majeures liées (i) à la structure particulière des forêts denses humides (en lien avec les cyclones) favorisant l’exploration unidirectionnelle de l’espace, (ii) aux contraintes édaphiques liées aux substrats ultramafiques favorisant la paupérisation architecturale, (iii) à l’absence historique de grands brouteurs, auxquels les monocaules sont particulièrement sensibles, et (iv) à la persistance des forêts denses humides lors des épisodes glaciaires (servant de refuges pour ces espèces sensibles) et leur expansion post-glaciaire (fournissant de nombreuses opportunités écologiques). / The convergent evolution in growth habit is a fundamental phenomenon linking plant ecology and evolution. Remarkably illustrated in island biotas, this phenomenon has never been identified in the original and megadiverse New Caledonian biodiversity hotspot. Through an approach combining plant architecture, functional traits, taxonomy, phylogeny and environmental data, this thesis analyses the evolutionary history of the scarcely known monocaulous growth habit in New Caledonia. Monocauls are self-supporting woody plants whose cardinal functions rely on a single visible stem. In New Caledonia, they are represented by 182 dicotyledonous species belonging to 41 genera and 30 families and are critically endangered. The repeated evolution of the monocaulie in New Caledonia, resulting from at least 31 independent events, is one of the most remarkable cases of convergence in insular environments. In the genus Atractocarpus (Rubiaceae), monocauly evolved recently two to three times through branch reductions into inflorescences, emphasizing the importance of heterochronic processes in the evolution of growth habit. Monocauly is strongly correlated with several traits illustrating major constraints in functional coordination. The evolution of monocauly is strongly associated with rainforests and ultramafic substrate, and seems to have contributed to the diversification of lineages by niche partitioning. The remarkable convergence toward monocauly in New Caledonia can be explained by four major hypotheses: (i) the structural features of rainforests (related to cyclone frequency and intensity) favoring unidirectional exploration of space, (ii) the edaphic constraints associated with ultramafic substrate favoring architectural pauperization, (iii) the historical absence of large native browsers to which monocauls are particularly sensitive, and (iv) the persistence of rainforest during – and spread-out after – glacial episodes that served as refugia and further provided ecological opportunities.

Made available in DSpace on 2016-12-08T16:44:42Z (GMT). No. of bitstreams: 1 PGPV13MA118.pdf: 16238 bytes, checksum: 5420ca805eb7e2614dc4d46d421abbfb (MD5) Previous issue date: 2013-03-04 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O objetivo deste trabalho foi comparar os diferentes sistemas de condução: Líder Central (Central Leader), Eixo Alto (Tall Spindle), Eixo Solar (Solaxe) e Eixo Duplo (Bibaum) para a cultura da macieira, em diferentes densidades, em relação ao custo inicial de produção, fenologia, vigor e qualidade das frutas. O experimento foi conduzido em Vacaria, RS durante as safras 2010/2011 e 2011/2012, nas cultivares MaxiGala e Fuji Kiku-8 sobre os portaenxertos de Marubakaido com filtro de M.9 e M.9. Foram avaliados a área da secção do tronco abaixo e acima do ponto de enxertia, o comprimento do entrenó do ramo, o peso de poda no inverno, o índice de intensidade de poda, a área foliar, o número de frutos, a produção estimada por planta, a produtividade estimada, o diâmetro transversal de fruto, a firmeza de polpa, os sólidos solúveis, a fenologia e a viabilidade financeira. Considerando as variáveis que expressam o vigor das plantas conclui-se que, o sistema de condução Eixo Duplo é o que resulta em maior controle do vigor. Por outro lado, o mesmo sistema de condução foi menos produtivo que os demais sistemas de condução. Quanto à qualidade dos frutos o sistema Eixo Duplo apresentou menor tamanho de fruto e sólidos solúveis, porém maior firmeza de frutos. Com relação à fenologia, de modo geral os sistemas de condução sobre M.9 concluem as fenofases antecipadamente. Quanto à avaliação financeira, o custo das mudas foi o item de maior percentagem de participação para todos os sistemas de condução e densidades de plantio. Entre os sistemas de condução o Eixo Duplo é o de maior custo, sendo o valor da muda em torno de 61% dos custos, em função disso o custo inicial foi maior que dos demais sistemas de condução, com média de 11,54% a mais que o Líder Central e o Eixo Alto e 11,46% que o Eixo Solar. Na comparação entre as densidades de plantio, a densidade de 4166 plantas ha-1 para todos os sistemas de condução foi a que apresentou maior custo, devido ao maior número de mudas por hectare

Master of Science / Department of Entomology / David C. Margolies / James R. Nechols / The study of how extrinsic factors affect the foraging efficiency and behavior of predaceous arthropods like Phytoseiulus persimilis is important to understand their various processes of acquiring prey, mates, refuges, oviposition sites, and overcoming obstacles posed by the environment. Many intrinsic and extrinsic factors affect predator foraging efficiency and behavior. One of the most influential extrinsic factors may be the host plant on which herbivorous prey are found. Recent studies suggest that plant architecture plays an important role in tritrophic interactions. In this work, I studied the effects of cucumber plant architecture and prey distribution on the foraging efficiency (prey-finding time and prey-consumption rate) and behavior (time allocated between moving, resting and feeding) of P. persimilis. Plant architecture represented differences in leaf number and size; however, all plants had the same total surface area. Plants with 6 small leaves (ca. 82.98 square cm each) were considered as complex architecture, whereas plants with only 2 large leaves (ca. 240.60 square cm each) were considered as simple. The prey distributions were: prey patch on a single basal leaf (closest leaf to the soil) and prey patch on all leaves. The foraging efficiency was assessed by measuring prey-finding time and prey-consumption rate, whereas the behavior was assessed by conducting observational studies on specific foraging activities. When placed either on the top or at the base of the plant P. persimilis encountered prey more rapidly (interval 0-30 minute) on complex and simple plants with prey patches distributed on all leaves. Differences in prey density (number of prey per leaf) had no effect on the prey-finding time of P. persimilis. The predator consumed more eggs on complex plants with prey patches distributed on all leaves. Phytoseiulus perismilis tended to find prey patches more quickly as well as consume more eggs on leaves close to its release point. Furthermore, the predator was observed to lay more eggs on leaves where it had consumed higher number of prey eggs. The dimensions of individual parts of the plant e.g., stem, petiole and leaf, affected the time allocated by P. persimilis between searching, resting and feeding. The predatory mite spent more time foraging on the stems and petioles of the simple plants whereas on complex plants it spent more time foraging on the leaves.

No description available.

Réalisé en cotutelle avec l'Université Montpellier II / L’étude qualitative et quantitative du mode de développement des plantes envahissantes est actuellement considérée comme une étape clef dans la compréhension des phénomènes d’invasion. L’objectif de ce travail est de préciser les relations qui existent entre la structure architecturale des buissons et leur caractère proliférant. Nous avons sélectionné cinq espèces buissonnantes (Cornus sericea L., Cornaceae ; Prunus virginiana L., Rosaceae ; Rhamnus cathartica L., Rhamnaceae ; Rhus typhina L., Anacardiaceae ; Zanthoxylum americanum Mill., Rutaceae) qui sont connues pour leur aptitude à bloquer la succession végétale sous certaines conditions au Sud du Québec (Canada). L’analyse architecturale a permis chez ces espèces de caractériser les unités structurelles et leurs modifications ontogéniques. Ces modifications ontogéniques doivent être prise en compte afin d’obtenir une description complète de la plasticité phénotypique chez ces espèces. L’analyse des différentes unités structurelles révèle qu’elles ne possèdent pas la même signification fonctionnelle : les niveaux d’organisation les plus grands sont responsables majoritairement des capacités de plasticité phénotypique de la plante et de sa compétition. Ces analyses ont abouti à la définition de trois stratégies architecturales correspondant à des comportements individuels et qui sont également pertinentes pour expliquer la prolifération des populations. / Qualitative and quantitative studies of the pattern of invasive plant development is now considered a key aspect in understanding invasiveness. This work was performed to determine relationships between shrub architectural plasticity and proliferating behaviour. We selected five shrub species (Cornus sericea L., Cornaceae ; Prunus virginiana L., Rosaceae ; Rhamnus cathartica L., Rhamnaceae ; Rhus typhina L., Anacardiaceae ; Zanthoxylum americanum Mill., Rutaceae) known to arrest plant succession under certain conditions in Southern Québec, Canada. Architectural analysis revealed species’ structural units and their ontogenic changes. These ontogenic changes need to be calibrated if a full description of phenotypic plasticity is to be obtained. Analysis of the plant structural units reveals that they are of different functional significance: the higher the level of organization, the greater the capacity for phenotypic plasticity and competition. We defined three architectural strategies related to individual behaviours and which can relevantly explain the population proliferation of shrubs.

La taille est une pratique culturale courante en arboriculture. Pourtant ses effets sur le rendement sont variables et peu prévisibles. Dans cette thèse, nous proposons d'étudier les effets de la taille sur la croissance végétative et la reproduction du manguier (Mangifera indica L.), à l'échelle de l'axe et de l’arbre, afin de mieux comprendre comment s’élabore le rendement suite à la taille. La taille a été caractérisée par deux facteurs : l’intensité de taille, définie à l’échelle de l’arbre comme la quantité de biomasse retirée par volume de canopée, et la sévérité de taille, définie à l’échelle de l’axe comme la profondeur de taille le long de l’axe. A l’échelle de l’axe, l’intensité et la sévérité de taille ont eu des effets sur l’occurrence, l’intensité et la dynamique de la croissance végétative. Nous avons aussi mis en évidence un rôle important du diamètre de l’axe taillé sur l’occurrence et l’intensité de la croissance végétative. La croissance végétative qui résulte de la taille a conduit à l’apparition de nouvelles populations d’axes dont les proportions dépendent de l’intensité de taille des arbres. La floraison de ces différentes populations d’axes diffère en termes d’occurrence, d’intensité et de dynamique et est affectée négativement par l’intensité et la sévérité de taille. A l’échelle de l’arbre, les rendements ont eu tendance à être plus faibles sur les arbres modérément et intensément taillés du fait d’une diminution du taux de floraison et/ou du poids moyen des fruits. Les arbres faiblement taillés ont eu des rendements similaires à ceux des arbres non taillés. Les effets de la taille sur la croissance végétative, aux niveaux structurel et temporel, ont été intégrés dans un modèle structure-fonction d'élaboration du rendement et de la qualité de la mangue, afin de caractériser les effets de la taille à l'échelle plus globale de l'arbre. Ce modèle nous a montré que le nombre d’axes et la surface foliaire retirés au moment de la taille sont largement compensés par la stimulation de la croissance végétative qui a suivi. L’intégration des effets de la taille sur la reproduction dans le modèle permettra de concevoir des itinéraires techniques intégrant les pratiques de taille afin de répondre aux problèmes agronomiques rencontrés sur le manguier, tels que les asynchronismes ou les irrégularités de production. / Pruning is a common practice in tree fruit crops. However, its effects on the yield are variable and hardly predictable. In this thesis, we propose to study the effects of pruning on vegetative growth and reproduction for the mango tree (Mangifera indica L.), at the axis and at the tree scales, in order to better understand how fruit yield built-up following pruning. Pruning was characterized by two factors: pruning intensity, defined at the tree scale as the amount of fresh biomass removed per unit volume of canopy, and pruning severity, defined at the axis scale as the depth of pruning along the axis. At the axis scale, pruning intensity and severity had an effect on the occurrence, intensity and dynamics of vegetative growth. We also highlighted a major role of the diameter of the pruned axis on the occurrence and intensity of vegetative growth. The vegetative growth which results from pruning led to the emergence of new populations of axes whose proportions depend on pruning intensity. Flowering of these different populations of axes differ in terms of occurrence, intensity and dynamics and was affected negatively by pruning intensity and severity. At the tree scale, yields tended to be lower on moderately and intensively pruned trees as a result of a decrease of flowering rate and/or average weight of fruits. Lightly pruned trees had similar yields compared to unpruned trees. The effects of pruning on the vegetative growth, at the structural and temporal levels, were integrated in a functional-structural model of fruit yield and quality build-up. The model has shown that the number of axes and the leaf area removed at the pruning time were widely compensated by the vegetative growth stimulation that followed. The integration of the effects of pruning on reproduction will allow developing crop management techniques integrating pruning techniques in order to respond to agronomic issues related to mango tree as asynchronisms and irregular bearing.

Entre as doenças de maior impacto nos vinhedos paulistas, encontram-se o míldio (Plasmopara viticola) e a ferrugem (Phakopsora euvitis). Os objetivos deste trabalho foram: i. caracterizar a estrutura genética populacional e identificar espécies crípticas de P. viticola; ii. avaliar o progresso temporal de míldio da videira, o efeito da doença na produção das plantas e verificar a contribuição da reprodução sexuada e assexuada de P. viticola na epidemia da doença e; iii. avaliar o progresso temporal de ferrugem da videira. Foram coletados 516 isolados de P. viticola em 11 municípios de São Paulo. Os isolados foram genotipados com 7 marcadores microssatélites. Para a identificação de espécies crípticas de P. viticola, uma subamostra de 130 isolados foi analisada pela técnica CAPS (Cleaved Amplified Polymorphic Sequence), e 94 isolados foram sequenciados para a região ITS1. O progresso temporal do míldio e da ferrugem da videira foram avaliados em 3 ciclos de produção (2014/15, 2015/16 e 2016/17) nos sistemas de condução em espaldeira (ESP), Y descoberto (YD) e Y sob cobertura plástica (YC). A área experimental, localizada em Piracicaba-SP, foi constituída por plantas da variedade Niágara Rosada. As variáveis analisadas foram fenologia, número de folhas por ramo, incidência e severidade das doenças. Para verificar o efeito do míldio na produção das plantas, avaliaram-se a massa média do cacho, a produção por planta, o teor de sólidos solúveis totais (SST) e a acidez titulável (AT) dos frutos. A área abaixo da curva (AAC) foi calculada para o número de folhas por ramo, a incidência e a severidade de míldio e de ferrugem. Os valores de massa do fruto, produção por planta, SST e AT, e os resultados da AAC para as variáveis analisadas, foram submetidos a análise de variância e as médias comparadas pelo teste de Tukey (P<0,05). Na safra 2015/16, 166 isolados de P. viticola foram coletados em três diferentes momentos da epidemia, os quais foram genotipados com os 7 microssatélites mencionados anteriormente. P. viticola clado aestivalis foi a única espécie críptica identificada no Estado de São Paulo. Dentre os 516 isolados analisados, 55 genótipos multilocos (MLG) foram identificados. Quatro MLGs dominantes representaram 65,7% dos genótipos observados. Para o míldio da videira, em 2 das 3 safras avaliadas observou-se desfolha precoce das plantas doentes. A severidade média máxima de míldio observada foi de 14,3% na ESP e 30,9% no YD. Para esses valores de severidade, a doença causou a redução na produção por planta (0,7 a 0,9 kg por planta) e na massa do cacho (67 a 78 g). Em nenhuma das safras houve efeito da doença no teor de SST e na AT. Maiores valores de AAC da severidade de míldio foram verificados no YD em comparação à espaldeira, sugerindo que o sistema em YD favorece o progresso da doença. No YC, a redução da incidência de míldio foi de até 93,7% em relação aos sistemas descobertos, indicando que o sistema em Y associado ao cultivo protegido foi uma ferramenta eficiente no controle da doença. O monitoramento de genótipos de P. viticola na safra 2015/16 revelou que a epidemia foi resultante, predominantemente, de infecções clonais múltiplas causadas por um único genótipo. Queda prematura de folhas em decorrência da epidemia de ferrugem foi observada para os sistemas em YC e YD com valores de severidade inferiores a 5,6%. Em função da desfolha, verificaram-se alterações na fenologia das plantas nos estádios fenológicos finais de desenvolvimento. O progresso da incidência de ferrugem foi mais rápido nos sistemas em YC e YD em relação a ESP. Diferentemente das epidemias de míldio da videira, o uso da cobertura plástica não foi suficiente para o controle satisfatório da ferrugem. / Downy mildew (Plasmopara viticola) and rust (Phakopsora euvitis) are important diseases on vineyards located in São Paulo State. The aims of this study were: i. characterize the population genetic structure and identify cryptic species of P. viticola; ii. evaluate the temporal progress of grapevine downy mildew, the disease effect on yield, and verify the contribution of sexual and asexual reproduction of P. viticola on disease epidemic and; iii. evaluate the temporal progress of grapevine rust. A total of 516 isolates of P. viticola were collected from 11 locations of São Paulo State. The isolates were genotyped with 7 microsatellite markers. To identify cryptic species of P. viticola, a subsample of 130 isolates were analized by Cleaved Amplified Polymorphic Sequence (CAPS), and 94 isolates were sequenced for ITS1 region. Temporal progress of downy mildew and rust were evaluated in 3 growing seasons (2014/15, 2015/16 and 2015/16) on vertical trellis (VT), Y-trellis without plastic cover (YWP), and Y-trellis with plastic cover (YPC). The experimental area, located in Piracicaba-SP, was composed by plants of the variety Niagara Rosada. The variables analyzed were phenology, number of leaves per shoot, disease incidence and disease severity. To verify the effect of downy mildew on yield were evaluated bunch weight, yield per plant, total soluble solids content (TSS) and titratable acidity (TA) of the fruits. Area under the curve (AUC) was calculated for number of leaves per shoot, downy mildew and rust incidence and downy mildew and rust severity. The values of bunch weight, yield per plant, TSS and TA, and AUC results for the analyzed variables, were submitted to variance analysis and the mean values were compared by Tukey test (P<0.05). In the growing season of 2015/16, 166 isolates of P. viticola were collected at three different moments of downy mildew epidemic. The isolates were genotyped with the 7 microsatellites previously mentioned. P. viticola clado aestivalis was the only cryptic species identified in São Paulo State. Among the 516 isolates analyzed, 55 multilocus genotypes (MLG) were identified. Four dominant MLGs represented 65.7% of the observed genotypes. For downy mildew, 2 of the 3 growing seasons evaluated were observed early defoliation of the diseased plants. The maximum average values of disease severity observed were 14.3% in ESP and 30.9% in YD. For these disease severity values, the disease reduced yield per plant (0.7 to 0.9 kg per plant) and bunch weight (67 to 78 g). No effect of downy mildew was verified on TSS and TA in none of the growing seasons. Higher AUC values for downy mildew severity were observed in YWP compared to the VT, suggesting that YWP favors the disease progress. On the YPC, incidence of downy mildew was reduced by up to 93.7% compared to the uncovered systems, indicating that Y-trellis associated with plastic cover was an efficient tool to control the disease. The monitoring of genotypes of P. viticola in the 2015/16 growing season revealed that disease epidemic was drived predominantly by multiple clonal infections caused by a single genotype. Early defoliation due to grape rust epidemic was observed on YWP and YPC with disease severity lower than 5.6%. Due to defoliation caused by rust modification on plant phenology were verified in the final stages of plant development. The progress of rust incidence was faster in YWP and YPC than in VT. Unlikely grape downy mildew epidemics, the use of plastic cover was not sufficient to effectivly control grape rust.

L’étude qualitative et quantitative du mode de développement des plantes envahissantes est actuellement considérée comme une étape clef dans la compréhension des phénomènes d’invasion. L’objectif de ce travail est de préciser les relations qui existent entre la structure architecturale des buissons et leur caractère proliférant. Nous avons sélectionné cinq espèces buissonnantes (Cornus sericea L., Cornaceae ; Prunus virginiana L., Rosaceae ; Rhamnus cathartica L., Rhamnaceae ; Rhus typhina L., Anacardiaceae ; Zanthoxylum americanum Mill., Rutaceae) qui sont connues pour leur aptitude à bloquer la succession végétale sous certaines conditions au Sud du Québec (Canada). L’analyse architecturale a permis chez ces espèces de caractériser les unités structurelles et leurs modifications ontogéniques. Ces modifications ontogéniques doivent être prise en compte afin d’obtenir une description complète de la plasticité phénotypique chez ces espèces. L’analyse des différentes unités structurelles révèle qu’elles ne possèdent pas la même signification fonctionnelle : les niveaux d’organisation les plus grands sont responsables majoritairement des capacités de plasticité phénotypique de la plante et de sa compétition. Ces analyses ont abouti à la définition de trois stratégies architecturales correspondant à des comportements individuels et qui sont également pertinentes pour expliquer la prolifération des populations. / Qualitative and quantitative studies of the pattern of invasive plant development is now considered a key aspect in understanding invasiveness. This work was performed to determine relationships between shrub architectural plasticity and proliferating behaviour. We selected five shrub species (Cornus sericea L., Cornaceae ; Prunus virginiana L., Rosaceae ; Rhamnus cathartica L., Rhamnaceae ; Rhus typhina L., Anacardiaceae ; Zanthoxylum americanum Mill., Rutaceae) known to arrest plant succession under certain conditions in Southern Québec, Canada. Architectural analysis revealed species’ structural units and their ontogenic changes. These ontogenic changes need to be calibrated if a full description of phenotypic plasticity is to be obtained. Analysis of the plant structural units reveals that they are of different functional significance: the higher the level of organization, the greater the capacity for phenotypic plasticity and competition. We defined three architectural strategies related to individual behaviours and which can relevantly explain the population proliferation of shrubs. / Réalisé en cotutelle avec l'Université Montpellier II

La forme, et donc l’architecture d’une plante ornementale en potcomme le rosier buisson détermine sa qualité visuelle et peut êtrecontrôlée via la création variétale et/ou l’utilisation de techniquesculturales comme l’application de facteurs environnementaux.Cependant, ces méthodes sont appliquées empiriquement etprennent rarement en compte l’interaction Génotype x Environnement(GxE). Dans un premier temps, cette étude a montré que, parmiles trois techniques culturales testées sur cinq à huit cultivarsde rosier buisson de formes contrastées, la restriction hydrique modifiait fortement l’architecture avec un effet compactant. Une forteinteraction GxE a également été montrée avec trois réponses architecturales,de faible à forte, pouvant s’expliquer en partie par desconcentrations différentes en cytokinines et acide salicylique entregénotypes.Dans un deuxième temps, l’analyse génétique par cartographiede l’architecture du rosier à partir de deux populationsconnectées en ségrégation et remontantes a montré que la plupartdes caractères architecturaux étudiés étaient contrôlés par un ouplusieurs QTLs à effets faibles dont certains étaient infl uencés parl’environnement et/ou par le fond génétique. L’introgression detels caractères dans d’autres fonds génétiques par exemple seradonc vraisemblablement diffi cile. Cette étude a souligné l’importancede prendre en compte les interactions GxE et QTLxE pourmieux raisonner les programmes de création/sélection, mais aussipour l’utilisation des techniques culturales dans le but de contrôler l / The shape and therefore the architecture of a potted ornamentalplant such as bush rose determines its visual quality. It can becontrolled by plant breeding and/or cultivation techniques suchas the environmental factor application. However, these methodsare used empirically and rarely take into account the Genotypex Environment interaction (GxE). Firstly, among three cultivationtechniques evaluated on fi ve to eight rose bush cultivars withcontrasted shapes, water restriction highly modifi ed the architectureleading to more compact plants. A high GxE interaction was revealedwith three architectural responses ranging from low to high. Theywere partly explained by different cytokinins and salicylic acidconcentrations between genotypes.Secondly, the genetic analysisby mapping of the bush rose architecture based on two connectedrepeat-blooming populations in segregation showed that most ofthe architectural traits were controlled by one or several QTLs withlow effects. Some of them were infl uenced by the environmentand/or by the genetic background. The introgression of such traitsin different genetic backgrounds for example will likely be diffi cult.This study underlined the importance of taking into considerationthe GxE and QTLxE interactions to better orientate breedingprograms but also to better use cultivation techniques in order to control plant architecture.

Contexte et objectifs : Les tentatives pour expliquer les variations des paramètres foliaires fondées uniquement sur les facteurs environnementaux, ce qui conduira à des erreurs importantes si la plante montre des variations ontogénétiques substantielles des propriétés foliaires. Nous avons évalué les variations des 27 traits morpho-anatomiques des feuilles de Dipterocarpus alatus durant six stades de développement architectural et entre les axes à un stade de développement donné sous deux conditions lumineuses différentes.Méthodologies:Une analyse architecturale a été réalisée pour distinguer précisément et objectivement des catégories d'axes et stades de développement de Dipterocarpus alatus. Les feuilles ont été collectées sur l'unité de croissance plus récente et complète sur la partie supérieure du tronc et sur la partie plus externe de la branche du milieu couronne des arbres qui poussent sous deux différentes conditions lumineuses. Vingt-deux traits foliaires ont été mesurées et calculées sur des images de tracés feuilles et de sections transversaux à l'aide de logiciel ImageJ. Densité stomatal a été calculée sur l'impression de vernis à ongles de la surface inférieure des feuilles. Quatre traits de chlorophylle ont été déterminées. Les tests d'ANOVA et Tukey ont été utilisés pour déterminer les différences entre les stades de développement et des axes. La contribution relative lmg a été calculée avec package R relaimpo et comparées avec l'analyse bootstrap pour déterminer le stade ontogénétique, catégorie axe ou l'intensité lumineuse, ce qui est expliqué plus les variations de l'anatomie des feuilles.Résultats:La morphologie et l'anatomie des feuilles diffère fortement durant l'ontogenèse pour les quatre axes en terme d'ordre et de catégorie deux sous des conditions lumineuses. L'effet de l'axe a été présentée au stade B, C, D et F, mais ont tendance à être insignifiante au stade E. Ontogenèse de l'arbre expliqué plus de variations dans la morphologie des feuilles et de l'anatomie que l'intensité lumineuse. Catégorie d'axe expliqué plus de variations dans la morphologie des feuilles, mais pas dans l'anatomie des feuilles que l'ontogenèse de l'arbre. Conclusions:Fortes et significatives variations intra-spécifiques (pendant l'ontogenèse et entre axes) peuvent influencer les variations inter-spécifiques, et donc contestent la validité de la valeur moyenne des traits foliaires entre les espèces. Ontogenèse de l'arbre contribué plus que l'intensité de la lumière dans l'explication de la variabilité dans la morphologie et l'anatomie des feuilles à la fois sur le tronc et deuxième l'ordre l'axe recommande fortement que les études sur les réponses de la morphologie et l'anatomie des feuilles à l'environnement doivent corriger l'effet de l'ontogenèse. / Background and Aims:Attempts to explain variations in leaf parameters based solely on environmental factors, this will lead to significant errors if the plant shows substantial ontogenetic variations in leaf properties. We evaluated variations in 27 morpho-anatomical leaf traits of Dipterocarpus alatus over six architectural development stages and between axes at a given development stage under two different light conditions. Methods:An architectural analysis was conducted to distinguish precisely and objectively axis categories and developmental stages of Dipterocarpus alatus. Leaves were collected on the most recent complete growth unit on the top of the trunk and the outermost part of the middle-crown branch with branchlet and twig from trees growing under two different light conditions. Twenty two leaf traits were measured and calculated using ImageJ on images of leaf tracings and cross-sections. Stomatal density was calculated on nail varnish impression of the leaf lower surface. Four chlorophyll traits were determined. One-way ANOVA and Tukey post-hoc tests were used to determine differences between development stages and axes. The lmg relative contribution was calculated with R package relaimpo and compared with bootstrap analysis to determine ontogenetic stage, axis category or light intensity, which explained more the variations in leaf anatomy.Key results:Leaf morphology and anatomy differed strongly during ontogeny for all four axes in term of order and category under both light conditions. The axis effect was displayed at stages B, C, D and F but tend to be insignificant at stage E. Tree ontogeny explained more variations in leaf morphology and anatomy than light intensity. Axis category explained more variations in leaf morphology but not in leaf anatomy than tree ontogeny. Conclusions: Strong and significant intraspecific variations (during ontogeny and between axes) may influence the interspecific variations, and thus challenge the validity of the mean value of leaf traits between species. Tree ontogeny contributed more than light intensity in explanation of the variability in leaf morphology and anatomy both on the trunk and second axis order strongly recommends that studies on the responses of leaf anatomy to the environment need to correct for the ontogeny effect.