L'électroencéphalographie (EEG) consiste à mesurer l'activité électrique du cerveau à l'aide d'électrodes placées à la surface du cuir chevelu. Cette technique est principalement utilisée pour le diagnostic de l'épilepsie. Certains grapho-éléments comme des ondes lentes et des pointes-ondes peuvent apparaitre sur l'EEG permettant au neurologue de détecter une souffrance épileptique. Malheureusement, cette activité peut être fortement contaminée par des signaux parasites appelés artefacts. Ces artefacts ont pour origines principales l'activité oculaire, l'activité musculaire, le rythme cardiaque et les légers déplacements d'électrodes. Les fréquences des grapho-éléments pathologique recouvrent celles des artefacts et nous avons alors recours à des filtres spatiaux, reposant sur la séparation de sources. Le principe est de déterminer un ensemble de sources d'origines cérébrales et un ensemble de sources d'artefacts. Les sources d'artefacts sont supprimées et les sources cérébrales sont utilisées pour reconstruire le signal. Nous présentons dans cette thèse plusieurs méthodes combinant les filtres spatiaux et les filtres frèquentiels afin d'automatiser le filtrage. Une démarche quantitative a été définie afin de valider ces méthodes, nous permettant de choisir la méthode la plus performante appelée Adaptive Filtering by Optimal Projection (AFOP). Les tests sur des enregistrements cliniques de patient épileptiques, ont montré, selon l'avis du neurologue, l'efficacité d'AFOP sur la plupart des types d'artefacts ainsi que son respect des rythmes cérébraux. / The electroencephalography (EEG) consists in measuring brain electrical activity thanks to electrodes located on the scalp surface. This technique is mainly used for the diagnostic of epilepsy. Sorne grapho-elements like slow waves and spike waves can appear on the EEG, enabling the neurologist to detect an epilepsy pain. Unfortunately, this activity can be highly contaminated by parasitical signals called artifacts. These artifacts have for main origins, the ocular activity, the muscular activity, the cardiac rhythm and tight electrode displacements. The frequencies of pathological grapho-elements recover those of artifacts, and it is then required to use spatial filter which rests on source separation. The principle is to determine a set of cerebral sources and a set of artifacts sources. Artifact sources are then cancelled and the cerebral ones are used to rebuild the signal. This thesis presents several methods using both spatial and frequential filters, making the EEG filtering automated. A quantitative approach of filtering validation is defined, which enables the author to choose the most efficient called Adaptive Filtering by Optimal Projection (AFOP). According to the neurologist, tests on clinical recordings of epileptic patients prove AFOP efficiency on cancelling most of artifact types as well as on respecting cerebn rhythms.

<p> This thesis seeks to describe and explain the spatial pattern of urban residential blight. The empirical analysis is limited to one study area; that of the Chicago metropolitan area. The data are U.S. Census data for census tracts in 1940, 1950, and 1960. Following a review of pertinent literature, an attempt is made to conceptualise the process that generates residential blight. From this conceptual framework, a number of hypotheses are developed concerning the relationship between residential blight and selected socio-economic variables. Other relationships are derived from an interpretation of maps of residential blight in the Chicago area for the different time periods. The hypotheses are tested using such multivariate procedures as principal components analysis, and regression and correlation analysis. The thesis also contains an application of the Blalock-Simon procedure for causal modelling to the Chicago data. The findings of the empirical analysis are related to present knowledge concerning urban residential blight. </p> / Thesis / Doctor of Philosophy (PhD)

This research attempts to extend the application of spatial analysis to the investigation of human agency in social relations. Marcos Llobera's research framework on modelling daily experience and social space showcases great potentials of utilising Geographic Information Systems (GIS) to explore the perception and behaviour of individual agents. By expanding Llobera's work and incorporating Tim Ingold's wayfaring theory to explore the human agency in the context of social relations, this research proposes a new analytical method to investigate social relations through the accumulation of intimate interactions. Exemplified with detailed analysis on two settlements of Kaushi people in Taiwan, the proposed analytical method demonstrates great strength and yields fruitful insights into their social structure and transformation. In addition, this method is particularly instrumental in unravelling specific relations between individuals, as well as between social groups. The application of this method on Kaushi settlements yielded fruitful insights of their social structure and transformation. On the other hand, the side products of this approach can be further employed to investigate the visual structure and movement intensity of a site, as well as to experiment alternative 'what if' scenarios relating to visibility, movement, and interaction. In sum, this research augments the potential of spatial analysis to explore human agency in a social context and lays out a further platform for the investigation of social relations at a settlement scale.

L'objectif de ce travail est de développer une méthode capable de déterminer automatiquement l'état de vigilance chez l'humain. Les applications envisageables sont multiples. Une telle méthode permettrait par exemple de détecter automatiquement toute modification de l'état de vigilance chez des personnes qui doivent rester dans un état de vigilance élevée (par exemple, les pilotes ou les personnels médicaux).Dans ce travail, les signaux électroencéphalographiques (EEG) de 58 sujets dans deux états de vigilance distincts (état de vigilance haut et bas) ont été recueillis à l'aide d'un casque à 58 électrodes posant ainsi un problème de classification binaire. Afin d'envisager une utilisation de ces travaux sur une application du monde réel, il est nécessaire de construire une méthode de prédiction qui ne nécessite qu'un faible nombre de capteurs (électrodes) afin de limiter le temps de pose du casque à électrodes ainsi que son coût. Au cours de ces travaux de thèse, plusieurs approches ont été développées. Une première approche propose d'utiliser un pré-traitement des signaux EEG basé sur l'utilisation d'une décomposition en ondelettes discrète des signaux EEG afin d'extraire les contributions de chaque fréquence dans le signal. Une régression linéaire est alors effectuée sur les contributions de certaines de ces fréquences et la pente de cette régression est conservée. Un algorithme génétique est utilisé afin d'optimiser le choix des fréquences sur lesquelles la régression est réalisée. De plus, cet algorithme génétique permet la sélection d'une unique électrode.Une seconde approche est basée sur l'utilisation du Common Spatial Pattern (CSP). Cette méthode permet de définir des combinaisons linéaires des variables initiales afin d'obtenir des signaux synthétiques utiles pour la tâche de classification. Dans ce travail, un algorithme génétique ainsi que des méthodes de recherche séquentielle ont été proposés afin de sélectionner un sous groupes d'électrodes à conserver lors du calcul du CSP.Enfin, un algorithme de CSP parcimonieux basé sur l'utilisation des travaux existant sur l'analyse en composantes principales parcimonieuse a été développé.Les résultats de chacune des approches sont détaillés et comparés. Ces travaux ont aboutit sur l'obtention d'un modèle permettant de prédire de manière rapide et fiable l'état de vigilance d'un nouvel individu. / The aim of this work is to develop a method able to automatically determine the alertness state of humans. Such a task is relevant to diverse domains, where a person is expected or required to be in a particular state. For instance, pilots, security personnel or medical personnel are expected to be in a highly alert state, and this method could help to confirm this or detect possible problems. In this work, electroencephalographic data (EEG) of 58 subjects in two distinct vigilance states (state of high and low alertness) were collected via a cap with $58$ electrodes. Thus, a binary classification problem is considered. In order to use of this work on a real-world applications, it is necessary to build a prediction method that requires only a small number of sensors (electrodes) in order to minimize the time needed by the cap installation and the cap cost. During this thesis, several approaches have been developed. A first approach involves use of a pre-processing method for EEG signals based on the use of a discrete wavelet decomposition in order to extract the energy of each frequency in the signal. Then, a linear regression is performed on the energies of some of these frequencies and the slope of this regression is retained. A genetic algorithm (GA) is used to optimize the selection of frequencies on which the regression is performed. Moreover, the GA is used to select a single electrode .A second approach is based on the use of the Common Spatial Pattern method (CSP). This method allows to define linear combinations of the original variables to obtain useful synthetic signals for the task classification. In this work, a GA and a sequential search method have been proposed to select a subset of electrode which are keep in the CSP calculation.Finally, a sparse CSP algorithm, based on the use of existing work in the sparse principal component analysis, was developed.The results of the different approaches are detailed and compared. This work allows us to obtaining a reliable model to obtain fast prediction of the alertness of a new individual.

O atropelamento de animais silvestres é considerado como o principal fator antrópico responsável diretamente pela mortalidade de vertebrados terrestres em escala global. Estimativas de mortalidade são fundamentais para avaliar o impacto de rodovias, mas para reduzir o seu viés a remoção de carcaças e a eficiência dos observadores devem ser consideradas. Medidas mitigadoras têm sido implementadas para reduzir a mortalidade da fauna e ampliar a conectividade da paisagem, mas um fator determinante para a sua efetividade é a sua correta localização. Com o objetivo de qualificar o planejamento de medidas mitigadoras, neste trabalho procuramos responder a quatro perguntas: 1) há diferença na remoção e detectabilidade de carcaças entre diferentes grupos taxonômicos? 2) qual a influência da remoção e detectabilidade de carcaças sobre as estimativas de magnitude de mortalidade? 3) a mortalidade se distribui de forma agregada ao longo da rodovia? e 4) a distribuição espacial de atropelamentos de diferentes grupos taxonômicos é similar? Nossos resultados apontam diferenças na taxa de remoção e na detectabilidade de carcaças entre os grupos, além de demonstrar que, ao desconsiderar esses fatores, a magnitude de atropelamentos é subestimada. Ademais, nossos resultados indicam que a distribuição espacial de atropelamentos de mamíferos pode ser utilizada como indicadora da ocorrência de atropelamentos de outros grupos taxonômicos apenas em escalas menos refinadas, exigindo o planejamento de medidas mitigadoras mais amplas. Os resultados aqui apresentados devem ser considerados no monitoramento de animais atropelados e no planejamento de medidas mitigadoras do impacto de rodovias. / Vehicle-wildlife collisions are considered the main human factor responsible directly for vertebrate mortality worldwide. Roadkill estimates are elementary to evaluate road impacts, but carcass removal and searcher efficiency must be considered in order to diminish estimation bias. Mitigation measures have been implemented to reduce wildlife mortality and to increase connectivity, but their correct placement is an important factor defining the effectiveness of these measures. In order to qualify mitigation planning, in this study we aim to answer four main questions: 1) is there difference in carcass removal rates and detectability among different taxonomic groups? 2) do carcass removal and detectability influence mortality magnitude estimates? 3) are roadkills spatially aggregated? and 4) are roadkill spatial distribution of different taxonomic groups similar? Our results show differences in carcass removal and detectability among groups, and demonstrate that mortality magnitude is underestimated when these factors are not considered. Also, our results indicate that mammal roadkill aggregations may be used as a surrogate of roadkill aggregations of other taxonomic groups in larger scales. The results presented here must be considered in roadkill monitoring and in mitigation measures planning.

O atropelamento de animais silvestres é considerado como o principal fator antrópico responsável diretamente pela mortalidade de vertebrados terrestres em escala global. Estimativas de mortalidade são fundamentais para avaliar o impacto de rodovias, mas para reduzir o seu viés a remoção de carcaças e a eficiência dos observadores devem ser consideradas. Medidas mitigadoras têm sido implementadas para reduzir a mortalidade da fauna e ampliar a conectividade da paisagem, mas um fator determinante para a sua efetividade é a sua correta localização. Com o objetivo de qualificar o planejamento de medidas mitigadoras, neste trabalho procuramos responder a quatro perguntas: 1) há diferença na remoção e detectabilidade de carcaças entre diferentes grupos taxonômicos? 2) qual a influência da remoção e detectabilidade de carcaças sobre as estimativas de magnitude de mortalidade? 3) a mortalidade se distribui de forma agregada ao longo da rodovia? e 4) a distribuição espacial de atropelamentos de diferentes grupos taxonômicos é similar? Nossos resultados apontam diferenças na taxa de remoção e na detectabilidade de carcaças entre os grupos, além de demonstrar que, ao desconsiderar esses fatores, a magnitude de atropelamentos é subestimada. Ademais, nossos resultados indicam que a distribuição espacial de atropelamentos de mamíferos pode ser utilizada como indicadora da ocorrência de atropelamentos de outros grupos taxonômicos apenas em escalas menos refinadas, exigindo o planejamento de medidas mitigadoras mais amplas. Os resultados aqui apresentados devem ser considerados no monitoramento de animais atropelados e no planejamento de medidas mitigadoras do impacto de rodovias. / Vehicle-wildlife collisions are considered the main human factor responsible directly for vertebrate mortality worldwide. Roadkill estimates are elementary to evaluate road impacts, but carcass removal and searcher efficiency must be considered in order to diminish estimation bias. Mitigation measures have been implemented to reduce wildlife mortality and to increase connectivity, but their correct placement is an important factor defining the effectiveness of these measures. In order to qualify mitigation planning, in this study we aim to answer four main questions: 1) is there difference in carcass removal rates and detectability among different taxonomic groups? 2) do carcass removal and detectability influence mortality magnitude estimates? 3) are roadkills spatially aggregated? and 4) are roadkill spatial distribution of different taxonomic groups similar? Our results show differences in carcass removal and detectability among groups, and demonstrate that mortality magnitude is underestimated when these factors are not considered. Also, our results indicate that mammal roadkill aggregations may be used as a surrogate of roadkill aggregations of other taxonomic groups in larger scales. The results presented here must be considered in roadkill monitoring and in mitigation measures planning.

O atropelamento de animais silvestres é considerado como o principal fator antrópico responsável diretamente pela mortalidade de vertebrados terrestres em escala global. Estimativas de mortalidade são fundamentais para avaliar o impacto de rodovias, mas para reduzir o seu viés a remoção de carcaças e a eficiência dos observadores devem ser consideradas. Medidas mitigadoras têm sido implementadas para reduzir a mortalidade da fauna e ampliar a conectividade da paisagem, mas um fator determinante para a sua efetividade é a sua correta localização. Com o objetivo de qualificar o planejamento de medidas mitigadoras, neste trabalho procuramos responder a quatro perguntas: 1) há diferença na remoção e detectabilidade de carcaças entre diferentes grupos taxonômicos? 2) qual a influência da remoção e detectabilidade de carcaças sobre as estimativas de magnitude de mortalidade? 3) a mortalidade se distribui de forma agregada ao longo da rodovia? e 4) a distribuição espacial de atropelamentos de diferentes grupos taxonômicos é similar? Nossos resultados apontam diferenças na taxa de remoção e na detectabilidade de carcaças entre os grupos, além de demonstrar que, ao desconsiderar esses fatores, a magnitude de atropelamentos é subestimada. Ademais, nossos resultados indicam que a distribuição espacial de atropelamentos de mamíferos pode ser utilizada como indicadora da ocorrência de atropelamentos de outros grupos taxonômicos apenas em escalas menos refinadas, exigindo o planejamento de medidas mitigadoras mais amplas. Os resultados aqui apresentados devem ser considerados no monitoramento de animais atropelados e no planejamento de medidas mitigadoras do impacto de rodovias. / Vehicle-wildlife collisions are considered the main human factor responsible directly for vertebrate mortality worldwide. Roadkill estimates are elementary to evaluate road impacts, but carcass removal and searcher efficiency must be considered in order to diminish estimation bias. Mitigation measures have been implemented to reduce wildlife mortality and to increase connectivity, but their correct placement is an important factor defining the effectiveness of these measures. In order to qualify mitigation planning, in this study we aim to answer four main questions: 1) is there difference in carcass removal rates and detectability among different taxonomic groups? 2) do carcass removal and detectability influence mortality magnitude estimates? 3) are roadkills spatially aggregated? and 4) are roadkill spatial distribution of different taxonomic groups similar? Our results show differences in carcass removal and detectability among groups, and demonstrate that mortality magnitude is underestimated when these factors are not considered. Also, our results indicate that mammal roadkill aggregations may be used as a surrogate of roadkill aggregations of other taxonomic groups in larger scales. The results presented here must be considered in roadkill monitoring and in mitigation measures planning.

Orientador: Fernando Roberto Martins / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-13T23:57:13Z (GMT). No. of bitstreams: 1 Sims_LigiaPaulillo_M.pdf: 1997370 bytes, checksum: 7862d46e8d35044bd7c95ed32db9713f (MD5) Previous issue date: 2009 / Resumo: O presente estudo investigou a estrutura de estádios ontogenéticos e espacial da espécie Savia dictyocarpa (Phyllanthaceae) na Mata Ribeirão Cachoeira, Campinas, SP. Essa espécie dióica é comum ocupante do dossel de florestas estacionais da região sudeste do Brasil e dada sua importância resolveu-se levantar informações sobre sua história de vida. Para tanto foram feitas duas questões centrais: 1) é possível definir os estádios ontogenéticos de S. dictyocarpa a partir da variação morfométrica e arquitetural que ocorre longo da vida da espécie? 2) qual distribuição espacial dos indivíduos ao longo da ontogenia dessa espécie e quais os fatores ambientais que influem na distribuição? Nos 0,5 hectares amostrados encontramos 194 indivíduos distribuídos em 5 estádios ontogenéticos: infantil, jovem, imaturo, adulto e senil. Não encontramos plântulas ou adultos reprodutivos nas parcelas, mas encontramos algumas plântulas. Portanto a definição do estádio plântula foi feita sem considerações sobre caracteres morfométricos. Os frutos encontrados não possuíam sementes, o que pode ser indicativo de ausência de polinização de flores, seja pela baixa freqüência de indivíduos masculinos ou pela ausência de polinizador. O padrão de ramificação apresentado é próprio do modelo arquitetural de Massart, que é característico de espécies bem adaptadas às condições de pouca luz. Os estádios apresentaram diferenças significativas quanto à altura e diâmetro. Jovem e imaturo apresentaram menor aumento em diâmetro por incremento em altura, ocorrendo o inverso com o estádio adulto. O maior investimento em altura durante os estádios iniciais permite que os indivíduos alcancem rapidamente as boas condições de luminosidade próximas ao dossel, em detrimento da estabilidade mecânica. Encontramos que o investimento em altura da copa por unidade de aumento em diâmetro da copa tem tendência de aumentar ao longo da ontogenia, o que denota diferenciação entre estádios para aumento da eficiência na obtenção de luz. A infestação de lianas parece ter efeito sobre desenvolvimento da copa dos indivíduos adultos, pois indivíduos com lianas têm menor investimento em altura da copa. Todos os estádios apresentaram distribuição aleatória no espaço, sendo que imaturo e jovem possuem tendência de ocupar mesmos locais, possivelmente por necessitarem das mesmas condições ambientais para recrutamento e crescimento. Todas as correlações entre estádios e luz, solo e interferência interespecífica foram espúrias e/ou não significantes. É possível que o padrão aleatório possa ser explicado por características ambientais não consideradas nesse estudo ou por processos estocásticos. A distribuição da espécie pode também ser fortemente afetada pela distribuição de sementes, plântulas e infantes, que inexistiam na área amostrada. Os resultados encontrados sugerem que há variações nas estratégias de ocupação do ambiente e de uso dos recursos ao longo da vida de uma espécie de dossel, no entanto as características ambientais amostradas não foram capazes de explicar como esses caracteres determinam os padrões de recrutamento / Abstract: the present study investigated the ontogenetic and spatial structure of the tree species Savia dictyocarpa (Phyllanthaceae) in the Mata Ribeirão Cachoeira, Campinas municipality, SP. This is a dioecious species that commonly occur on seasonal forest canopy from southeast region of Brazil, and because of its importance we decided to gather information about its life history. To do that, we ask two main questions: 1) Can we define S. dictyocarpa ontogenetic stages using morphometric and architectural variation along the species life? ; 2) Which is the spatial distribution of different ontogenetic stages and which are the environmental factors that influence this distribution? We sampled 0,5 hectares and found 194 individuals distributed in 5 ontogenetic stages: infant, juvenile, immature and adult and senile. We did not found seedlings or reproductive adults in the plots we sampled, but we found some fruits and seedlings on the trail. Thus, seedling stage was described with no consideration of morphometric characteristics. The fruits we found did not have seeds and this fact indicates the absence of flower pollination, which could be caused by low frequency of male individuals or by absence of pollinator. The ramification pattern showed by S. dictyocarpa is correspondent to Massart architectural model, which is common for species that grow under low light conditions. Stages presented significant differences in height and diameter. Juvenile and immature presented smaller diameter increment per unit of height, and adult stage had higher diameter increment per unit of height. This suggests that in early stages a greater investment in height allows then to quickly reach for good light conditions on the forest canopy, disregarding the mechanical stability. We also found that crown length investment per unit of crown width increase along ontogeny, what shows that there is differences between stages to increase the light absorption efficiency. Liana infestation seems to affect adult crown development because individuals with liana have smaller investment in crown length. All stages presented random spatial distribution and juvenile and immature showed the tendency to occupy the same places, probably because they have the same recruitment and growth requirements. All correlations between stages and abiotic and biotic factors (light, soil and interspecific competition) were spurious or nom significant. This random pattern could be explained by environmental characteristics that were not considered here or by stochastic processes. Species distribution could also be strongly affected by seed, seedling or infant distributions, which were not found in the sampled area. These results suggests that there are different environmental occupation and resource use strategies along a canopy species live, however the environmental characteristics sampled here were not able to explain how this strategies determinates recruitment patterns in the space. To understand the absence of seeds and seedlings in the area and to propose conservation and management plans for S. dictyocarpa, reproductive biology and population dynamic studies are required / Mestrado / Ecologia / Mestre em Ecologia

Orientador: Flavio Antonio Maes dos Santos / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-08T20:06:21Z (GMT). No. of bitstreams: 1 Rosseto_Vanessa_M.pdf: 1782142 bytes, checksum: 9d91cab6c070b924e2ff3078c70d69f9 (MD5) Previous issue date: 2007 / Resumo: A coexistência de espécies arbóreas em florestas tropicais pode estar associada com a partição dos recursos, através da utilização diferencial dos recursos ambientais ao longo do tempo e espaço. O objetivo do trabalho foi avaliar quais os possíveis mecanismos relacionados com a coexistência de espécies arbóreas simpátricas, Trichilia sp. (Meliaceae), em três fisionomias florestais em um fragmento urbano de Campinas, SP. Para tanto, foram feitos três questionamentos centrais: (1) Qual a heterogeneidade ambiental de três fisionomias florestais (Floresta Estacional Semidecídua, Floresta Paludosa e Floresta de Maria-Preta)? (2) As espécies de Trichilia apresentam características estruturais divergentes ao longo da ontogenia, que possibilitem a coexistência das espécies? (3) O padrão espacial das espécies de Trichilia varia nas diferentes fisionomias e possíveis diferenças estão associadas com fatores ambientais específicos? Para o estudo foram utilizadas três escalas espaciais: escala fisionômica (comparação dos descritores ambientais entre as fisionomias), escala local (comparação entre áreas dentro de uma mesma fisionomia) e escala microlocal (comparação entre parcelas de uma mesma área). A umidade do solo foi o único fator ambiental divergente entre a Floresta de Maria-Preta (MP) e a Floresta Estacional Semidecídua (FES), enquanto que a Floresta Paludosa (FP) não diferiu. MP foi considerada uma floresta monodominante, porque a espécie arbórea maria-preta, Diatenopteryx sorbifolia (Sapindaceae), representou mais de 50% do número de indivíduos e área basal na comunidade arbórea. Supõe-se que a monodominância esteja associada com características do solo. Na escala local e microlocal observou-se grande heterogeneidade ambiental e, em muitos casos, áreas de diferentes fisionomias foram mais semelhantes entre si, do que áreas em uma mesma fisionomia. Ao longo da ontogenia as espécies divergiram entre si na morfologia (presença de caule descamado em T. claussenii e crescimento clonal em T. elegans) e no tamanho. O padrão espacial das espécies de Trichilia não variou nas diferentes fisionomias, exceto em uma área com maior valor de umidade do solo na FP, com a ocorrência apenas de T. pallida. O padrão espacial de Trichilia não foi associado com a heterogeneidade ambiental, com exceção do alagamento, que aparentemente é um fator restritivo para a maioria das espécies arbóreas. Com isso, sugere-se que a coexistência das espécies de Trichilia analisadas esteja mais relacionada com diferenças na ocupação do estrato vertical que cada espécie ocupa do que com diferenças nas preferências ambientais. Além disso, supõe-se que a elevada heterogeneidade ambiental encontrada em todas as fisionomias, possibilite às espécies explorar uma ampla gama de condições ambientais, que podem estar relacionadas com a não preferência de algumas espécies por um habitat específico / Abstract: Tree species coexistence in tropical forests can be associated with resource partitioning through differential utilization of environmental resources along the time and space. The study evaluated coexistence mechanisms associated with sympatric tree species of Trichilia sp. (Meliaceae), in three physiognomies in an urban fragment in Campinas, SP. We approach three questions: (1) Which is the existing environmental heterogeneity in three different physiognomies of the forest fragment (Semideciduous Forest, Swamp Forest and Maria-Preta Forest)? (2) Have Trichilia species structural characteristics divergent along of the ontogeny that makes possible the species coexistence? (3) Are the spatial patterns of Trichilia species variable in the physiognomies analyzed and are possible differences related to specific environmental factors? We evaluated these questions in three spatial scales: a physiognomy scale (variations of environmental factors among physiognomies), a local scale (variations among places in the same physiognomy) and microlocal scale (variations among plots in the same place). The soil moisture was the only one environmental variable that differed between Maria-Preta Forest (MP) and Semideciduous Forest (FES), while Swamp Forest (FP) not differed. MP was considered a monodominant forest, because a tree species ¿maria-preta¿, Diatenopteryx sorbifolia (Sapindaceae), presenting more than 50% of individual numbers and basal area in the tree community. We presumed that monodominance is associated with soil characteristics. High environmental heterogeneity was observed in local and microlocal scale. In many cases, places of different physiognomies were more similar than places in a same physiognomy. Trichilia species diverged through ontogeny in morphology (presence of scaled off stem in T. claussenii and clonal growth in T. elegans) and in size. Spatial pattern of Trichilia species not varied in the different physiognomies, except for one sample with greater soil moisture in the FP, where only T. pallida occurred. Spatial pattern of Trichilia was not associated with environmental heterogeneity, except for the flooding, that seeming was restrictive to the establishment of the majority of the species. Then we supposed that Trichilia species coexistence is more associated with differences in the vertical strata that each species occupies than with differences in environmental preferences. Furthermore, we suggests that the high environmental heterogeneity encountered found in all physiognomies, makes possible to the species explore an ample gamma of ambient, that can be related with no preference of species for specific habitats / Mestrado / Mestre em Ecologia

The influence of the spatial pattern of foliage removal on regrowth was investigated in the field with a tussock grass, Agropyron desertorum. Tussocks were hand clipped in different spatial patterns that represented extremes of defoliation patterns which could be inflicted by natural herbivores. All defoliated tussocks exhibited increases in specific growth rates following clipping in mid-May. When foliage was removed from the upper portion of the canopy (younger foliage), regrowth rates and season-long aboveground biomass production were less than if the same amount of foliage was removed from low in the canopy (older foliage). The spatial pattern of defoliation also influenced tussock regrowth in a late-May clipping experiment, but differences in the effects of the clipping patterns were associated with the removal of apical meristems rather than with the age or location of foliage removed. Changes in tussock carbon dioxide and water vapor exchange that were associated with changes in growth following mid-May clipping were explored. All clipped tussocks showed increases in integrated daytime carbon dioxide uptake per unit foliage area after defoliation. Differences among treatments in the response of net daytime carbon gain during the first 24 hours after clipping corresponded well with differences in tussock regrowth over a 14-day period following clipping. Increased carbon gain of clipped tussocks was associated with increases in tussock water vapor conductance and intercellular carbon dioxide concentration, and decreases in the ratio of carbon dioxide uptake to water vapor loss. Differences among treatments in daytime carbon gain and regrowth were paralleled by the response of instantaneous rates of light-saturated net photosynthesis for entire tussocks. Defoliation increased the proportion of foliage directly illuminated within the tussock at solar noon. Changes in the fraction of sunlit foliage and the relative amounts of different-aged foliage in tussock canopies were correlated with the responses of light-saturated photosynthesis. Thus, the effects of the spatial pattern of foliage removal on canopy light microclimate and the age of remaining foliage had important implications for carbon gain and regrowth of tussocks following mid-May defoliation.