Much of ecological theory describes species interactions. These interactions often play an important theoretical role in facilitating coexistence. In particular, rarity in ecological communities, though often observed, provides a significant challenge to theoretical and empirical ecologists alike. I use a plant community model to simulate the effect of stronger negative frequency dependence on the long-term persistence of the rare species in a simulated community. This strong self-limitation produces long persistence times for the rare competitor, which otherwise succumb quickly to stochastic extinction. The results suggest that the mechanism causing species to be rare in this case is the same mechanism allowing those species to persist. To determine if ecological communities generally show the theoretical pattern, I estimate the strength of frequency-dependent population dynamics using species abundance data from 90 communities across a broad range of environments and taxonomic groups. In approximately half of the analyzed communities, rare species showed disproportionately strong negative frequency dependence. In particular, a pattern of increasing frequency dependence with decreasing relative abundance was seen in these communities, signaling the importance of this mechanism for rare species specifically. Insight into the special population dynamics of rare species will inform conservation efforts in response to climate change and other disturbance. Further difficulties in the detection of theoretical patterns in ecological data may be a result of the ecological currency used. Though ecologists typically use abundance data to test theories, energy use is another ecological currency that may be more relevant in some cases. In particular when detecting patterns that are a result of species interactions, the currency used should be the one in which those interactions actually operate. I compare the results of using abundance and energy use to detect two processes with well-defined expectations. The first is a description of population dynamics, the above described relationship between relative abundance and self-limitation. The second, compensatory dynamics, is a description of community-level dynamics. I find that the currency used alters the results, and thus the species-level implications. It does not, however, alter the overall pattern that would have theoretical implications. Results in both currencies support the pattern of strong self-limitation for persistent rare species.

Web page classification is the process of assigning predefined categories to web pages. Empirical evaluations of classifiers such as Support Vector Machines (SVMs), k-Nearest Neighbor (k-NN), and Naïve Bayes (NB), have shown that these algorithms are effective in classifying small segments of web directories. The effectiveness of these algorithms, however, has not been thoroughly investigated on large-scale web page classification of such popular web directories as Yahoo! and LookSmart. Such web directories have hundreds of thousands of categories, deep hierarchies, spindle category and document distributions over the hierarchies, and skewed category distribution over the documents. These statistical properties indicate class imbalance and rarity within the dataset. In hierarchical datasets similar to web directories, expanding the content of each category using the web pages of the child categories helps to decrease the degree of rarity. This process, however, results in the localized overabundance of positive instances especially in the upper level categories of the hierarchy. The class imbalance, rarity and the localized overabundance of positive instances make applying classification algorithms to web directories very difficult and the problem has not been thoroughly studied. To our knowledge, the maximum number of categories ever previously classified on web taxonomies is 246,279 categories of Yahoo! directory using hierarchical SVMs leading to a Macro-F1 of 12% only. We designed a unified framework for the content based classification of imbalanced hierarchical datasets. The complete Yahoo! web directory of 639,671 categories and 4,140,629 web pages is used to setup the experiments. In a hierarchical dataset, the prior probability distribution of the subcategories indicates the presence or absence of class imbalance, rarity and the overabundance of positive instances within the dataset. Based on the prior probability distribution and associated machine learning issues, we partitioned the subcategories of Yahoo! web directory into five mutually exclusive groups. The effectiveness of different data level, algorithmic and architectural solutions to the associated machine learning issues is explored. Later, the best performing classification technologies for a particular prior probability distribution have been identified and integrated into the Yahoo! Web directory classification model. The methodology is evaluated using a DMOZ subset of 17,217 categories and 130,594 web pages and we statistically proved that the methodology of this research works equally well on large and small dataset. The average classifier performance in terms of macro-averaged F1-Measure achieved in this research for Yahoo! web directory and DMOZ subset is 81.02% and 84.85% respectively.

Conserving rare species and protecting biodiversity depends on sound information on the nature of rarity. Rarity is multidimensional, presenting the need for a quantitative classification scheme by which to label species as rare or common. I defined rarity for freshwater fishes based on the range extents, habitat breadths, and site abundance and examined the relationship between these dimensions of rarity and imperilment. Imperiled fishes were most often rare by all three dimensions, whereas undesignated species were most often common by all three dimensions. Next, I examined the effect of sampling intensity on observed rarity of stream fish using different numerical and proportional rarity criteria and found that increasing sampling intensity increased the number of species labelled as rare with proportional criteria but did not affect the number of species labelled as rare with numerical criteria. Additional electrofishing passes within a fixed reach increases the likelihood of detecting rare and endemic species. A tradeoff between information collected and sampling resources should be carefully considered in the context of objectives when sampling for rare species. Finally, I examined the effect of regional and watershed habitat variables, biotic interaction variables, and instream habitat variables, on the rare or common status on 23 North American freshwater fishes. I also compared biological and reproductive traits among species classified into the rarity framework. Rarity was successfully explained in 19 of the 23 species and I found that regional and watershed habitat variables were the most important predictors of rarity. I also found that species large body size, high fecundity, and long age at maturity were generally more common by range extent and site abundance while those species that did not guard nests were more frequently rare by site abundance. These results indicate that large-scale variables can be used to successfully predict species rarity and rare fishes differ in their biology and reproduction from common fishes. / Master of Science

A fingerprint matching algorithm with a novel set of matching parameters based on core points and triangular descriptors is proposed to discover rarity in fingerprints. The algorithm uses a mathematical and statistical approach to discover rare features in fingerprints which provides scientific validation for both ten-print and latent fingerprint evidence. A feature is considered rare if it is statistically uncommon; that is, the rare feature should be unique among N (N>100) randomly sampled prints. A rare feature in a fingerprint has higher discriminatory power when it is identified in a print (latent or otherwise). In the case of latent fingerprint matching, the enhanced discriminatory power from the rare features can help in delivering a confident court judgment. In addition to mining the rare features, a parallel algorithm for fingerprint matching on GPUs is also proposed to reduce the run-time of fingerprint matching on larger databases. Results show that 1) matching algorithm is useful in eliminating false matches. 2) each of the 30 fingerprints randomly selected to mine rare features have a small set of highly distinctive statistically rare features some of whose occurrence is one in 1000 fingerprints. 3) the parallel algorithm implemented on GPUs for larger databases is around 40 times faster than the sequential algorithm. / Master of Science

Submitted by Erika Demachki (erikademachki@gmail.com) on 2014-09-25T18:57:04Z No. of bitstreams: 2 Bruno de Souza Barreto - 2008 - Dissertação.pdf: 2633509 bytes, checksum: 9a1040f27931399225f28b393e640b6e (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-09-26T11:43:42Z (GMT) No. of bitstreams: 2 Bruno de Souza Barreto - 2008 - Dissertação.pdf: 2633509 bytes, checksum: 9a1040f27931399225f28b393e640b6e (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-09-26T11:43:42Z (GMT). No. of bitstreams: 2 Bruno de Souza Barreto - 2008 - Dissertação.pdf: 2633509 bytes, checksum: 9a1040f27931399225f28b393e640b6e (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2008-06-06 / Fundação de Apoio à Pesquisa - FUNAPE / The units of conservation are main strategy adopted by the rulers to guarantee that the negative effects of the socioeconomic development reach less the biodiversity. recent researches have been trying to outline the problem through reserve selection that seek larger representativeness of the goals of conservation (in general some estimador of the biodiversity as the species) in terms of a smaller number of areas as possible. in the last 50 years the cerrado is going by transformations in their natural landscapes due to the progress of the modern agriculture and of the livestock. due to that, the objective of that work was to define priority areas network for conservation of bats in terms of five conservation sceneries differents. the generated nets were restricted by socioeconomic cost, presence of cave and a cost that considers the previous two cost. the results revealed main features of analyses of complementarity: multiple solutions, flexibility of the nets, and influence of the rarity of the species for the definition of groups starting from the pattern of rarity. the costs were important for definition of nets that avoided as much conflicts of conservation as prioritized the species in areas of cave. / As unidades de conservação são principal estratégia adotada pelos governantes para garantir que os efeitos negativos do desenvolvimento socioeconômico atinjam menos a biodiversidade. recentes pesquisas têm tentado contornar o problema através de seleção de reservas que visam maior representatividade das metas de conservação (em geral algum estimador da biodiversidade como as espécies) em termos de um menor número de áreas quanto possível. nos últimos 50 anos o cerrado vem passando por transformações em suas paisagens naturais em decorrência do avanço da agricultura moderna e da pecuária. os morcegos são também ameaçados e em decorrência disso, o objetivo desse trabalho foi definir redes de reservas regionais prioritárias para conservação de morcegos em termos de cinco cenários de conservação diferentes. as redes geradas foram restringidas por custo socioeconomico, custo presença de caverna e custo que podera os dois anteriores. os resultados reelaram as principais caracteísticas de análises de complementaridade: multiplas soluçoes, flexibilidade das redes, influência da raridade das espécies pela definição de grupos a partir do padrão de raridade. os custos foram importantes para definição de redes que evitaram tanto conflitos de conservação quanto priorizaram as espécies em regioes de caverna.

Several aspects of plant-pollinator interactions are presented in the thesis. It contains a review on the open questions of plant-pollinator interactions from single species to complex networks. The following sections document novel results. Firstly, the conservation of complex pollination networks is addressed through the hierarchy of species' importance. In addition, the habitat requirements and interactions of a threatened rare pollinator species are explored. In the following chapters, the results from manipulative approaches applied in the field to plant-pollinator interactions are presented. The effect of pollinator's population decline on pollinators' foraging for pollen is investigated. Moreover, the way plant species loss impact several aspects of pollinator visitation is presented. Lastly, the impact of species removal on plant-pollinator network topology and on species ability of establishing new interactions is investigated.

Background: Over the relatively short period of its evolutionary history, Boechera (Brassicaceae) has undergone rapid radiation that has produced 70+ morphologically distinct, sexual diploids. However, reproductive isolation has moved more slowly than morphological divergence in this group and the diploids appear to hybridize frequently where they coexist. Boechera duchesnensis appears to be the result of hybridization between its putative parents Boechera thompsonii and Boechera formosa. Objectives: The objectives of this study are to (i) analyze and document genetic diversity patterns in the population structure, - including allelic and heterozygosity frequencies - of B. thompsonii and B. formosa in concert with their geographic distribution to determine clustering relationships within these populations, (ii) confirm and expand the morphological characteristics of B. thompsonii and B. formosa, as initially proposed in the literature, including pollen and trichome structure using Scanning Electron Microscopy (SEM) to confirm ploidy level and to determine whether both putative parent species share morphological characteristics with their apomictic diploid offspring, and (iii) use genetic and morphologic evidence to show that B. thompsoii and B. formosa are, in fact, the parents of B. duchesnensis by comparing the genetic diversity patterns, population structure, and morphological characteristics of B. duchesnensis, to those of its proposed putative parents (B. thompsonii and B. formosa) and to confirm that B. duchesnensis shares characteristics of Boechera. Methods: Microsatellite data from 14 loci previously identified in Boechera were used to reexamine the current classifications and taxonomic foundations of three Boechera spp. GenAlEx 6.501 (Peakall and Smouse, 2006, 2012) was used to analyze genetic population structures of two divergent sexual diploids in the genus Boechera: B. thompsonii and B. formosa and to later compare those with the population structure of B. duchesnensis. Geographicaldata were plotted using ArcGIS 10.1 (Esri, 2012) to map heterozygosity distribution. Cluster analysis was run with STRUCTURE 2.3.3 (Pritchard et al., 2000; Falush et al., 2003, 2007) and distribution of allelic diversity and heterozygosity was subsequently compared within each taxon and correlated with geographic distribution characteristics. Resultant data were then compared with B. duchesnensis data to document genetic diversity patterns, population structure, and morphological characteristics. Key Results: Analysis of genetic diversity patterns, allelic distribution of the populations, and heterozygosity of B. thompsonii and B. formosa across their geographic range identified four genetically distinct clusters within B. thompsonii, and one genetically distinct cluster in B. formosa. Allelic frequencies in all four discrete population clusters of B. thompsonii and in one discrete population cluster of B. formosa were close to values found in species on the decline. Reproductive isolation, genetic variability, and allelic frequencies were determined, specimen elevations reported, and morphological characteristics reported in the literature were confirmed and expanded. A codominant genetic analysis performed for 14 different loci for B. duchesnensis against those of its parents showed that B. duchesnensis inherits alleles from both putative parents and confirms B. thompsonii and B. formosa as the parents of B. duchesnensis. Observed levels of heterozygosity of B. thompsonii and B. formosa were lower than expected levels and lower than those of other outcrossing diploids. The mean overall observed heterozygosities for each cluster were determined and documented by geographic location. A substantially higher level of observed heterozygosity in B. duchesnensis (Ho = 0.908) consistent with genetic fixation of a heterozygote and apomixis, supports hybridization as a speciation mechanism and apomixis as a mode of reproduction accounting for genotypic and phenotypic diversity. Morphological characteristics, especially those of pollen and trichomes were confirmed, expanded, and documented with SEM imagery. Discussion: This study provides an analysis of the genetic diversity patterns inherent in the population structure, allelic frequencies, allelic variation among individuals of the rare sexual diploids B. thompsonii, B. formosa, and the apomictic diploid B. duchesnensis in correlation with their geographic distribution. There is an implication of a reproductive barrier, within populations of the same species, that contributes to genetic isolation between clusters. I analyze the tendency of reduced heterozygosity to lead to genetic fixation, reproductive isolation, and how the heightened heterozygosity supports the classification of B. duchesnensis as an apomict. Assessing potential populations that might exist based on similar characteristics could possibly provide inferences about where future research might find similar examples of this hybridization. Reproductive isolation is hypothesized to limit gene flow between identified clusters of B. thompsonii and B. formosa exacerbating low observed heterozygosity levels and low allelic frequency levels. Population studies and cluster analysis have implications for offering future conservation strategies for both taxa.

Whilst the geographical range of species is a fundamental unit of macroecology and a leading predictor of extinction risk, the evolutionary dynamics of species’ ranges remain poorly understood. In some aquatic beetles, thermal niche has been shown to be related to both the relative range size and position of congeners but whether other physiological niche parameters play a role is unknown. Here, immunocompetence and metabolic plasticity were related to biogeography in these insects. Immunocompetence was first compared within a rare-common pair of Hydroporus species, finding species-specific immunity, which was affected by sex and acclimation time in the laboratory, with no clear relationship with rarity. This relationship was explored further in Deronectes species, whilst controlling for sex and acclimation time. Southern, narrow-ranging species had higher phenoloxidase (PO) activity, lower parasite load and antimicrobial peptide (AMP) activity that was stronger against Gram-negative Bacteria but weaker against Gram-positive Bacteria than their wider-ranging counterparts. As both of these studies found that PO activity did not positively correlate with encapsulation or AMP activity as reported in the literature, the pathway was investigated further in Tribolium castaneum. The data showed that the assumptions of the widely-used PO assay were violated, with differential activity between PO-driven reactions in melanin synthesis and different substrates used by larvae and adults. Future work should be wary of using the PO assay as a marker of potential melanin production and take into account the developmental requirements for melanin at different life stages. The relationship between metabolic plasticity and range was then assessed in Deronectes, finding that southerly species had more marked changes in glucose and protein content under elevated temperature stress than their northerly counterparts. Glucose content was also significantly positively correlated to lipid content, indicating that the energetics of species exhibiting differing range sizes may warrant future study. As the results from Hydroporus suggested that there may be trade-offs between immune defence and metabolism, the data on metabolic plasticity and immunocompetence in a sub-sample of Deronectes species were combined with thermal physiology, dispersal ability, body size and phylogenetic relatedness to assess which of these best explained variation in range size and position. Whilst variation in range extent and position were explained in part by thermal physiology, aspects of metabolic plasticity and immunocompetence also appeared to be important. This thesis provides one of the first indications that immunocompetence and metabolic plasticity may be related to geographical range and suggests parameters that may be worthy of exploration in other taxa.

Submitted by JÚLIO HEBER SILVA (julioheber@yahoo.com.br) on 2018-06-26T12:12:21Z No. of bitstreams: 2 Tese - Luciano Fabris Sgarbi - 2018.pdf: 35934499 bytes, checksum: 6fac6e45ed4a58506a9e7aa6a545f73b (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-06-27T11:45:11Z (GMT) No. of bitstreams: 2 Tese - Luciano Fabris Sgarbi - 2018.pdf: 35934499 bytes, checksum: 6fac6e45ed4a58506a9e7aa6a545f73b (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-06-27T11:45:11Z (GMT). No. of bitstreams: 2 Tese - Luciano Fabris Sgarbi - 2018.pdf: 35934499 bytes, checksum: 6fac6e45ed4a58506a9e7aa6a545f73b (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-04-19 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / A recurrent pattern seen in nature is that species vary in abundance among sites, being common in a few sites and rare at the majority of sites. This intraspecific distribution of abundance, in turn, generates an excess of rare species in biological communities. This excess may cause problems in data analyses as datasets include many zeros. In this thesis I studied the causes of species rarity and how the excess of rare species influence the detection of patterns at the community level. In the first chapter I observed that species classified as rare were common in other habitat types, in other sites elsewhere, and in other years. However, among these three factors, the rarity was explained mainly by the habitat type. In the second chapter I observed that species tend to become rarer as they distance themselves (above, below or in both directions) from their optimal elevation. This pattern was quite robust and independent of taxonomic group or geographic region. In the third chapter, I observed that remotion of rare species in general did not have effects on the recovery of community structure patterns. In addition, the effect of that remotion of rare species was similar to that observed for the common species exclusion or random species exclusion. / Um padrão repetidamente observado na natureza é que as espécies variam em abundância entre os locais, sendo que em poucos locais a espécie é abundante, enquanto que na maioria a espécie é rara. Esta distribuição intraespecífica de abundância, por sua vez, gera um padrão recorrente que é o excesso de espécies raras nas comunidades biológicas. Este excesso de espécies raras pode complicar análises ao nível de comunidade visto que produzem conjuntos de dados com muitos zeros. Nesta tese busquei entender quais são as causas para a raridade das espécies e como excesso de espécies raras influência a detecção de padrões ao nível de comunidades. No primeiro capítulo observei que espécies classificadas como raras, foram comuns em outros habitats, em outros locais e em outros anos. No entanto, entre estes três fatores, a raridade foi melhor explicada pelo tipo de habitat em que ela foi amostrada. No segundo capítulo observei que as espécies tendem a se tornar mais raras conforme elas se distanciam (acima, abaixo ou em ambas as direções) de sua elevação ótima. Este padrão foi bastante robusto pois foi independente do grupo taxonômico ou da região geográfica. No terceiro capítulo, observei que a remoção das espécies raras têm pouco efeito na recuperação dos padrões de estrutura das comunidades. Além disso, o efeito da remoção de espécies raras foi similar ao observado quando exclui as espécies comuns, ou exclui as espécies de forma aleatória.

This study is a content analysis of news articles of school shooting incidents that occurred within the United States between 1997 and 2012. This paper was designed to (a) address the current profile of school shooting offenders and offenses, (b) assess a proposed typology of school shootings, (c) consider common case processing characteristics for offenders of school shootings incidents, and (d) address the potential for offender and offense characteristics to affect the amount of media coverage an incident receives. The database of “Major School Shootings in the United States Since 1997” by the Brady Campaign to Prevent Gun Violence was used to compile a sample of 101 incidents in which a single offender committed a school shooting. To the extent possible, media accounts were used to corroborate details of each school shooting incident. Data pertaining to the offender characteristics, case processing characteristics, offense characteristics, characteristics regarding a typology, and media coverage characteristics were examined. The current profile and typology were, for the most part, upheld. Six variables proved to be significantly related to the total amount of media coverage an incident received: mental health history, school-related mass murder type, offender/victim type, total victims injured or killed, region of the U.S., and year of incident occurrence. Of these variables, three remained significant in a regression analysis: the school-related mass murder type, region of the U.S., and year of incident occurrence were predictive of the amount of media coverage an incident received. Implications and limitations of this study are discussed, and directions for future research are suggested.