L’objectif de cette thèse est de proposer une démarche pour identifier des plantes de service (Pds) utilisables en couverture vivante pour contrôler les adventices en bananeraie. Nous avons d’abord quantifié les effets de compétition réciproque entre le bananier et une couverture contrôlée par fauchage pour Brachiaria decumbens et Cynodon dactylon. Ces couverts ont eu des effets dépressifs sur la nutrition azotée du bananier et la productivité en premier cycle, le second cycle n’étant pas affecté. Dans un deuxième temps nous avons caractérisé au champ, en plantes isolées, la croissance en biomasse de huit légumineuses candidates sous différentes intensités lumineuses. En conditions lumineuses non limitantes, chaque espèce répondait à une dynamique exponentielle, définie par la biomasse 14 jours après semis et le taux de croissance relative (RGR). En conditions d’ombrage limitant, le RGR est réduit et le rapport surface foliaire sur biomasse aérienne croît. Ces réponses permettent de définir quatre stratégies d’adaptation à l’ombrage et d’identifier les Pds adaptées sous bananeraie. Pour trois de ces espèces, sur la base de la réponse de l’efficience photosynthétique, nous proposons un modèle de croissance en fonction de la ressource lumineuse. En peuplement, en présence d’adventices, aucune espèce évaluée n’a pu contrôler les grandes graminées, Stylosanthes guianensis et Centrosema pascuorum ont cependant réduit la biomasse des autres adventices de 60%. En bananeraie, les deux espèces évaluées ont exercé un contrôle sur les adventices mais ont réduit la productivité du premier cycle. Pour autant, ces résultats sont encourageants pour l’utilisation de Pds en bananeraie à des fins de contrôle des adventices. / The objective of this PhD thesis was to provide an approach to identify suitable cover crops (Pds) for weed control use in banana plantations. Firstly, we quantified the reciprocal effects between the banana crop and a grass cover controlled by mowing, for Brachiaria decumbens or Cynodon dactylon. These cover crops have had similar depressive effects on the nitrogen nutrition and on the productivity of the first cycle, but not afterward during the second cycle. Second step, we carried out a field characterization of eight legume cover crops, on isolated plants, under a range of light intensity. In non-limited light growth conditions, the biomass growth corresponded to an exponential dynamic, mathematically defined for each species by the biomass 14 days after sowing, and the relative growth rate (RGR). When affected by the shade, the RGR was reduced and the ratio between Leaf area and aerial plant biomass increased. We defined four adaptive strategies of the cover crops to the shade, what constitute criteria guidelines for cover crop use in banana cropping systems. For three of these species, on the basis of the response of the radiation use efficiency, we are proposing a plant growth model without any limiting factor other than light resource. In stands with the presence of weeds, neither of this cover crops was able to control the tall gramineous, but Stylosanthes guianensis and Centrosema pascuorum reduced the biomass of other weeds by 60%. Finally, in banana plantations, the two evaluated species provided a control of the weeds but reduced the productivity of the first cycle. Our results are comforting for the use of cover crops in banana plantation as a mean for weed control.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Apomixis is an asexual mechanism of reproduction through seed formation with clonal embryos. The high rate of polyembryonic seeds is usually an indicative for the occurrence of sporophytic apomixis and in Bignoniaceae the expression of sporophytic apomixis seems to be related to polyploidy. Handroanthus ochraceus and H. chrysotrichus are two tree species of Bignoniaceae that area morphologically related, which usually leads to identification problems. Handroanthus ochraceus has self-incompatible non-apomictic and monoembryonic populations and self-fertile sporophytic apomictic and polyembryonic populations. H. chrysotrichus has only self-fertile, sporophytic apomictic and polyembryonic populations recognized until now. To understand the genetic and phenotypic consequences of this reproductive mosaic in these populations it were conducted molecular and morphological analysis in the present study. The aims of molecular analyzes were to verify the genetic variability of H. ochraceus populations, compare the genetic diversity parameters investigating the propose of a possible relationship between the reproductive systems and genetic diversity in this species and still understand the possible relationships between apomictic and non-apomictic populations. The ten primers resulted in 104 fragments and the genetic diversity parameters showed a high mean for the populations (P = 66.35%, I = 0.341 and He = 0.227). In this study no clones were detected. The AMOVA analysis showed the higher variation within populations (61%), like allogamous populations. Cluster analysis and Bayesian analysis of genetic assignment determined two distinct groups, one consisting of two non-apomictic populations, Pires do Rio and Biribiri, and other gathering the apomictic populations and the non-apomictic from Uberlândia. The strong relationship between the self-incompatible population of Uberlândia and the apomictic populations indicates separate origins for non-apomictic populations of H. ochraceus. This relationship between self-incompatible population of Uberlândia and self-fertile populations suggest that these non-apomictic individuals may have the same genetic constitution of a possible ancestral population of self-fertile populations. The morphological analysis aimed to find differences between monoembryonic and polyembryonic populations of H. and morphological traits to separate the H. chrysotrichus and H. ochraceus. With this purpose non-apomictic monoembryonic populations and apomictic polyembryonic populations of H. ochraceus and an apomictic polyembryonic population of H. chrysotrichus were analyzed. Qualitative and quantitative morphological characteristics of leaf, flower and fruit were evaluated through univariate and multivariate analyzes. The pollen grain area, the length of peduncles, the width of stigma and the number of extrafloral nectaries in the calyx showed higher values for polyembryonic populations of H. ochraceus so they could be an evidence of polyploidy in H. ochraceus, since these larger measures may be associated with the cell volume increase usually entailed by polyploidy. Several morphological characters separated H. ochraceus from H. chrysotrichus, which can support the identification of individuals and the distinction between the two species. Therefore, the morphometric analysis determined that morphological features can make a distinction between mono and polyembryonic populations of H. ochraceus and the association of this methodology with reproductive biology analysis can be a complement for the determination of ploidies and reproductive system in both species. / A apomixia é um mecanismo de reprodução assexuada através da formação de sementes com embriões clonais. A elevada taxa de sementes poliembriônicas é um indicativo da ocorrência da apomixia esporofítica e, em Bignoniaceae a expressão da apomixia esporofítica parece estar vinculada à poliploidia. Handroanthus ochraceus e H. chrysotrichus são duas espécies arbóreas de Bignoniaceae morfologicamente semelhantes, o que normalmente ocasiona problemas de identificação. Handroanthus ochraceus apresenta populações autoincompatíveis, não apomíticas e monoembriônicas, e populações autoférteis, apomíticas esporofíticas e poliembriônicas. Já H. chrysotrichus apresenta apenas populações autoférteis, apomíticas esporofíticas e poliembriônicas conhecidas. Para compreender as consequências genéticas e fenotípicas deste mosaico reprodutivo foram realizadas análises moleculares e morfológicas. As análises moleculares tiveram por objetivos determinar a variabilidade genética de populações de H. ochraceus, comparar os parâmetros de diversidade genética, averiguar a possível relação entre os sistemas reprodutivos e a diversidade genética na espécie e ainda compreender as possíveis relações entre populações apomíticas e não apomíticas. Os dez primers ISSR utilizados resultaram em 104 bandas e os parâmetros de diversidade genética apresentaram uma média elevada para as populações (P = 66,35%, I = 0,341 e He = 0,227). Não foram detectados clones nas amostras populacionais. O cálculo da AMOVA demonstrou alta variação genética dentro das populações (61%), semelhante a populações alógamas. A análise de agrupamento e a análise Bayesiana de atribuição genética determinaram a formação de dois grupos distintos, um constituído por duas populações não apomíticas, Pires do Rio e Biribiri, e o outro reunindo as populações apomíticas e a não apomítica de Uberlândia. A forte relação entre a população autoincompatível de Uberlândia e as apomíticas indica origens distintas para as populações não apomíticas de H. ochraceus. Esta relação entre a população autoincompatível de Uberlândia e as populações autoférteis permite supor que os indivíduos não apomíticos dessa população apresentem a mesma constituição genética de uma possível população ancestral das populações autoférteis. As análises morfológicas tiveram por objetivo buscar diferenças entre populações mono e poliembriônicas de H. ochraceus e determinar características morfológicas que auxiliam na separação de H. ochraceus e H. chrysotrichus, para isso populações reconhecidamente monoembriônicas não apomíticas e poliembriônicas apomíticas de H. ochraceus e uma população poliembriônica apomítica de H. chrysotrichus foram analisadas. Características qualitativas e quantitativas de folha, flor e fruto foram avaliadas utilizando análises estatísticas univariadas e multivariadas. As características área do grão de pólen, comprimento do pedicelo, largura do estigma e número de nectários extraflorais no cálice apresentaram maiores valores para populações poliembriônicas de H. ochraceus e poderiam ser uma evidencia de poliploidia em H. ochraceus, pois, podem estar associadas ao aumento do volume celular geralmente acarretado pela poliploidia. Diferentes características morfológicas separaram as populações de H. ochraceus da população de H. chrysotrichus, as quais podem auxiliar na identificação de indivíduos e na distinção das espécies. Assim as análises morfométricas determinaram que características morfológicas de alguma forma podem distinguir populações mono e poliembriônicas de H. ochraceus, e em conjunto com análises de biologia reprodutiva podem ser uma metodologia auxiliar na determinação da ploidia e sistema reprodutivo. / Mestre em Genética e Bioquímica

The influence of biological relationships on the morphology of the human skeleton is a current theme in bioarchaeology and forensic research. Whether it is the discovery of kinship relationships in an anonymous cemetery or the question of individual identification based on familial similarity, research is limited by a number of factors. Above all, there is the shortage of osteological assemblages with genealogical documentation. Most such samples also include only a small number of individuals, and thus the number of different degrees of their relatedness is limited. Related to this is the lack or absence of methods suitable for either the exact expression of the degree of morphological similarity between individuals or the statistical evaluation of results. However, for this dissertation an opportunity arose to analyze several osteological assemblages with genealogical documentation from the 17th to 20th centuries, comprising almost one hundred individuals. This represents one of the largest such units in the world to date. The individual samples also include several cases of consanguineous marriages with subsequent generations. The aim of the first part of the dissertation is to propose new methodological approaches to expressing the degree of similarity of individuals based on different types of...