Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2016-04-15T20:28:21Z No. of bitstreams: 1 LarynneDantasDeSenna_DISSERT.pdf: 2775430 bytes, checksum: 1f765d0818e89af5941a1275fe961a99 (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2016-04-19T23:40:44Z (GMT) No. of bitstreams: 1 LarynneDantasDeSenna_DISSERT.pdf: 2775430 bytes, checksum: 1f765d0818e89af5941a1275fe961a99 (MD5) / Made available in DSpace on 2016-04-19T23:40:44Z (GMT). No. of bitstreams: 1 LarynneDantasDeSenna_DISSERT.pdf: 2775430 bytes, checksum: 1f765d0818e89af5941a1275fe961a99 (MD5) Previous issue date: 2015-06-02 / P ara contribuir no desempenho das pol?ticas e estrat?gias formuladas por comit?s de bacia hidrogr?fica , ?ndices v?m sendo criados na expectativa de expressar as m?ltiplas dimens?es dos recursos h?dricos em u ma forma facilmente interpret?vel. O uso do ?ndice de Pobreza H?drica (WPI) est? se difundindo mundialmente, sendo o mesmo formado pela combina??o dos sub?ndices Recurso, Acesso, Capacidade, Uso e Ambiente. A lgumas cr?ticas quanto ? f orma??o do WPI foram surgindo , de ntre elas destaca - s e a atribui??o de pesos dos sub ?ndices, feita por um processo arbitr?rio atribuindo subjetividade ao crit?rio de sele??o. Ao envolver an?lise estat?stica, quando se considera as caracter?sticas das vari?veis geradas pe la An?lise de Componentes Principais (ACP), verifica - se que a mesma ? capaz de solucionar esse problema . O objetivo deste trabalho ? comparar os resultados do WPI original com o ?ndice gerado atrav?s da An?lise de Componentes Principais (ACP), par a a indic a??o dos pesos dos sub ?ndices aplic?veis na bacia hidrogr?fica do Rio Serid? (RN e PB) . Conclui - se que o uso da An?lise de Componentes Principais na atribui??o do s pesos do ?ndice de Pobreza H?drica permitiu identificar que os sub - ?ndices Recurso, Acesso e Ambiente s?o os mais representativos para a bacia hidrogr?fica do Rio Serid? , e que este novo ?ndice, o WPI?, apresentou faixas de valores mais abra ngentes, permitindo identificar mais facilmente as disparidades entre os munic?pios. Al?m disso, a avalia? ? o dos sub ?ndices na ?rea de estudo tem grande potencial de informar ao tomador de decis?o na gest?o dos recursos h?dricos, as localidades mais cr?ticas e que merecem maiores investimentos nos aspectos analisados, j? que o ?ndice em si n?o permite captar essa informa??o. / To contribute in the performance of policies and strategies formulated by development agencies, indexes have been created in anticipation of expressing the multiple dimensions of water resources in an easily interpretable form. Use of Hydro Poverty Index ( WPI) is spreading worldwide , with the same formed by the combination of sub - indices Resource, access, capacity , use and environment. S ome critics a s to its formation have emerged, a mong them stands out the allo cation of weights of sub - indexes , made by an arbitrary process attributing subjectivity to the selection criteria. By involving statistical analysis, when considering the characteristics of the variables generated by the Principal Component Analysis (PCA), it turns out that it is able to solve this problem. The objective of this study is to compare the results of the original WPI with content generated by Principal Com ponent Analysis (PCA) for the indicati on of the weights of sub - indec es applicable in the Serid? River hydrographic Basin . We conclude that the use of Principal Component Analysis in the allocation of weights of Water Poverty Index has identified the sub - indices Resources, Access and Environment are the most representative for the river basin Serid? , and that this new index, WPI' , presented the most comprehensive ranges of values , allowing more easily identify disparities among municipalities. In addition, t he evaluation of the sub - indec es in the study area has great potential to inform the decision - maker in the management of water resources, the most critical locations and deserve greater investments in the aspects analyzed, as the index itself can not cap ture this information.

Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2016-05-03T22:20:29Z No. of bitstreams: 1 RildeciMedeiros_TESE.pdf: 2108807 bytes, checksum: b6e2efbc281920f2413849709ddf66ac (MD5) / Approved for entry into archive by clediane guedes (clediane@bczm.ufrn.br) on 2016-08-11T19:52:02Z (GMT) No. of bitstreams: 1 RildeciMedeiros_TESE.pdf: 2108807 bytes, checksum: b6e2efbc281920f2413849709ddf66ac (MD5) / Made available in DSpace on 2016-08-11T19:52:02Z (GMT). No. of bitstreams: 1 RildeciMedeiros_TESE.pdf: 2108807 bytes, checksum: b6e2efbc281920f2413849709ddf66ac (MD5) Previous issue date: 2015-06-25 / Esta tese prop?e a descri??o, a interpreta??o e a an?lise da estrutura composicional de resumos de disserta??es e teses, tanto no que se refere aos mecanismos lingu?sticos que evidenciam zonas textuais de diferentes sequ?ncias tipol?gicas como no que concerne ao plano de texto. Com efeito, o problema de pesquisa configurou-se a partir da no??o de estrutura composicional (sequ?ncias e planos de textos), como um dos n?veis ou planos da an?lise textual, conforme construtos te?ricos empreendidos por Jean-Michel Adam (2011a). O objetivo principal de estudo foi reconhecer como se d? a estrutura composicional do resumo de disserta??es e teses tendo em vista unidades textuais e a organiza??o global dessa categoria de texto. A hip?tese de trabalho ? a de que determinadas categorias informacionais da composi??o textual do resumo s?o necess?rias para o processo de representa??o do texto original e sua produ??o de sentido. Para tanto, este estudo tem como base te?rica e metodol?gica a Lingu?stica Textual (LT) e, sobretudo, a An?lise Textual dos Discursos (ATD), uma vez que buscamos compreender a estrutura organizacional do resumo tanto do ponto de vista lingu?stico quanto textual. Tal estrutura envolve a planifica??o textual do resumo, no seu prop?sito comunicativo, ou seja, o de divulga??o cient?fica, na sua forma textual t?pica. Assim, do ponto de vista do referencial te?rico e metodol?gico, o desenvolvimento deste estudo teve como base os pressupostos te?ricos e descritivosda ATD (ADAM, 2011a, 2012; PASSEGGI et al., 2010) e, ainda, da LT (BEAUGRANDE; DRESSLER, (2012 [1981]); COSERIU; LAMAS (2010); MARCUSCHI, 2009 [1983]; F?VERO; KOCH, 1994;KOCH, 2006; BENTES, 2004; BENTES; LEITE, 2010), situando-o no campo dos estudos do texto. Quanto ? metodologia, trata-se de um estudo emp?rico, de ordem documental e de base qualitativa, com abordagem descritiva e interpretativista. Numa perspectiva emp?rica, buscamos compreender o problema pertinente ? composi??o textual do resumo escrito, visando elucid?-lo ? luz de aportes te?ricos e metodol?gicos das teorias mencionadas. O corpus de an?lise foi composto de sete resumos destinados ? uma coleta sistem?tica de dados. Esses textos,produzidos no per?odo de 2004 a 2011,foram selecionados de disserta??es de mestrado e de teses de doutorado, na vers?o eletr?nica, de cursos de p?s-gradua??o da Universidade Federal do Rio Grande do Norte. A partir do estado da arte, tornou-se evidente a flutua??o terminol?gica do conceito de resumo. Os resultados encontrados revelam que os resumos do corpus de an?lise, em geral, apresentam uma heterogeneidade tipol?gica. Entretanto, o plano de texto ? fixo. Por fim, os conhecimentos novos advindos desta pesquisa contribu?ram tanto para a compreens?o da estrutura composicional de resumos como para sua produ??o. / This doctoral dissertationproposes the description, interpretation and analysis of the compositional structure of thesis and dissertation abstracts, with regard to the linguistic mechanisms that evidence text zones of different typological sequences, such as those of the text plan. Along these lines, the research problem was developed from the notion of compositional structure (sequences and text plans), as one of the levels or plans of text analysis, according to the theoretical framework proposed by Jean-Michel Adam (2011a). The main objective of this study was to recognize how the compositional structure, of thesis and dissertation abstracts, is achieved, with respect to text units and the global organization of this text category. The hypothesis posed in this research posits that specific informational text composition categories of abstracts are necessary to process the representation of the original text and the way in which it makes its meaning. Subsequently, this study is based on the theoretical and methodological framework of Text Linguistics (TL) and, above all, Textual Discourse Analysis (TDA), as we endeavor to understand the organizational structure of abstracts from both a linguistic and textual perspective. This structure involves the text plan of abstracts, with respect to their communicative purpose, i.e, the sharing of scientific information in its standard textual form. Thus, the development of this study, from a theoretical and methodological perspective, is based on the theoretical and descriptive premises from TDA (ADAM, 2011a, 2012; PASSEGGI et al., 2010), and also from TL (BEAUGRANDE; DRESSLER, (2012 [1981]); COSERIU; LAMAS (2010); MARCUSCHI, 2009 [1983]; F?VERO; KOCH, 1994;KOCH, 2006; BENTES, 2004; BENTES; LEITE, 2010), within the field of text studies. The methodology of this study relies on empirical, documental research, which is qualitative, and adopts a descriptive and interpretive approach. From the empirical perspective, our objective is to understand the problems pertaining to the textual composition of abstracts, aiming to elucidate them in light of the theoretical and methodological framework previously mentioned. The corpus of the analysis is comprised of seven abstracts designated for systematic data collection. These texts, written between 2004 and 2011,were selected from Master?s theses and Doctoral dissertations in their electronic version, from the graduate program at the Federal University of Rio Grande do Norte. A thorough review of the literature reveals a clear fluctuation in the terminology of the concept, ?abstract?. The results of the analysis revealed that the abstracts, which comprise the corpus of analysis in this study, in general, present typological heterogeneity, while the text plan remains fixed. Finally, the new knowledge gained in this research contributes both to the understanding of the compositional structure of abstracts as well as their production.

O presente estudo apresenta indícios de um aprender como ato a partir de um trajetar da pesquisadora nos ambientes escolares e, em especial, em turmas do segundo ano do Ensino Fundamental, numa escola municipal localizada no Vale do Rio dos Sinos, Rio Grande do Sul, Brasil. Com uma proposta vinculada ao projeto Civitas – LELIC/PPGEDU/UFRGS, as tessituras do escrito dialogam com a produção do filósofo russo Mikhail Bakhtin, tendo como objeto de análise os enunciados produzidos pelos sujeitos atuantes na pesquisa (alunos/as; professorapesquisadora). Nessa perspectiva, a metodologia utilizada, singularmente, se apresenta como um ato ético-estético que se entrelaça, acolhe e enuncia modos singulares de ver, conceber e sentir o processo de aprendizagem e, mais precisamente, ao ato de aprender, revelando o inusitado de novas vizinhanças: e, neste sentido, aponta para as possibilidades da invenção fugindo do hábito para leituras originais do que se dá “às vistas” e “às escutas”, aos acabamentos estéticos provisórios possibilitados pelos excedentes de visão, também, como plano de produção de si numa arquitetônica coletiva de produção do conhecimento. Uma tessitura ao mesmo tempo ética (pelo encontro com o outro) e estética (na busca de lhe dar, à experiência, um acabamento) pelos quais os sujeitos situados se relacionam e produzem a sala de aula como lugar coletivo. / This study presents evidence of learning as an act from the researcher's transit in school environments and in particular in the second year classes of an elementary school, a public school located in the Vale do Rio dos Sinos, Rio Grande do Sul, Brazil. With a proposal linked to the Civitas project - LELIC / PPGEDU / UFRGS, the interwoven of the written dialogues with the production of the Russian philosopher Mikhail Bakhtin, which has as object of analysis the utterances produced by the acting subjects in this research (students; the teacher-researcher). From this perspective, the methodology used is uniquely presented as an ethic-aesthetic act that intertwines, welcomes and sets unique ways of seeing, and feeling the learning process development, and more specifically the act of revealing the unusual learning of new neighborhoods. And in this sense it points to possibilities of the invention escaping from the current, original readings of what gives "the sights" and "the listenings" the provisional aesthetic completion made possible by the exceding vision as well as a production plan in itself, an architectural collective production of knowledge. A fabric that is at the same time ethic (in the meeting with the other) and aesthetic (in the quest of giving the other the experience, a completion) in which situated subjects relate and produce a classroom as a collective place.

O tema central desta pesquisa é o ensino e a aprendizagem do cuidado de enfermagem. O objetivo foi conhecer as representações e as práticas predominantes no processo atual de ensino e aprendizagem do cuidado, na percepção de alunos e professores dos cursos de Enfermagem, em universidades gaúchas. A presente investigação foi realizada em três universidades de diferentes regiões do Estado do Rio Grande do Sul. A metodologia utilizada foi a história oral e para análise das narrativas utilizou-se a análise temática. Ao trilhar os caminhos da construção desta dissertação, foram estabelecidos limites para a organização e apresentação dos resultados, uma vez que as possibilidades de desdobramentos eram múltiplas desta forma, para o desenvolvimento da versão final foram desenvolvidos três capítulos. Por fim, apresento algumas considerações finais, apontando estratégias e caminhos através de um ensino que é possível, desde que os seres ensinantes e aprendentes estejam dispostos a mesclar as representações e práticas de seus saberes e fazeres, na construção de uma ação em que haja vidas comprometidas a cuidar de outras vidas. / The central theme of this research is teaching and learning nursing care. The aim was to study the representations and practices prevalent in the current process of teaching and learning care, the perception of students and teachers of nursing courses in universities gaúchas. This research was conducted at three universities in different regions of the State of Rio Grande do Sul. Methodology used was oral history and analysis of the narratives used the thematic analysis. To tread the path of construction of this thesis, limits to the organization and presentation of the results have been established, since the possibilities of multiple deployments were thus for the development of the final three chapters were developed. Finally, I present some concluding remarks, pointing strategies and paths through an education that is possible, since the ensinantes beings and learners are willing to merge the representations and practices of their knowledge and practices in the construction of an action where lives committed to caring for other lives.

Submitted by Thiago Oliveira (thiago.oliveira@ifpb.edu.br) on 2016-11-24T19:31:49Z No. of bitstreams: 1 Emannuel Diego Gon?alves de Freitas - SEGMENTA??O DE REGI?ES DE AVC ISQU?MICOS EM IMAGENS DE TC POR MEIO DA CLASSIFICA??O DE TEXTURA.pdf: 682546 bytes, checksum: c6407fae8a68c0cf9693200aba271e15 (MD5) / Approved for entry into archive by Thiago Oliveira (thiago.oliveira@ifpb.edu.br) on 2016-11-24T19:33:07Z (GMT) No. of bitstreams: 1 Emannuel Diego Gon?alves de Freitas - SEGMENTA??O DE REGI?ES DE AVC ISQU?MICOS EM IMAGENS DE TC POR MEIO DA CLASSIFICA??O DE TEXTURA.pdf: 682546 bytes, checksum: c6407fae8a68c0cf9693200aba271e15 (MD5) / Made available in DSpace on 2016-11-24T19:33:07Z (GMT). No. of bitstreams: 1 Emannuel Diego Gon?alves de Freitas - SEGMENTA??O DE REGI?ES DE AVC ISQU?MICOS EM IMAGENS DE TC POR MEIO DA CLASSIFICA??O DE TEXTURA.pdf: 682546 bytes, checksum: c6407fae8a68c0cf9693200aba271e15 (MD5) Previous issue date: 2016-07 / ResumoImagens de Tomogra a Computadorizada (TC) s?o utilizadas como instrumentos deaux?lio ao diagn?stico m?dico, p ossibilitando a detec??o preco ce, b em como o acompanhamento,de diversas do en?as tais como o Acidente Vascular Cerebral (AVC), cuja an?lise da imagem docr?nio ? realizada p elo m?dico de forma visual, o qual lo caliza a les?o, diferencia o AVC isqu?micodo hemorr?gico e realiza a demarca??o manualmente. T?cnicas de an?lise computacionaldessas imagens, p o dem ampliar a quantidade de informa??o adquirida e contribuir para odiagn?stico correto em um pro cedimento m?dico. Tal raz?o motiva o desenvolvimento de sistemascomputadorizados para o aux?lio ? detec??o e ao diagn?stico (Computer-Aided Detection andDiagnosis- CADDx) para do en?as, empregando t?cnicas de Pro cessamento Digital de Imagense Vis?o Computacional. Neste contexto, s?o utilizados v?rios m?to dos com a nalidade desegmentar, reconhecer e identi car detalhes da regi?o de interesse na imagem de TC analisada.Esta p esquisa centraliza seus esfor?os em estab elecer um m?to do de segmenta??o da regi?o das?reas isqu?micas, no acidente vascular cerebral, em imagens de Tomogra a Computadorizada.Ao nal do pro cesso de segmenta??o, um contorno ? delineado automaticamente em voltada regi?o segmentada sem a necessidade de uma interven??o humana. Para tanto, ? feita aclassi ca??o de informa?? es de textura extra?das da imagem, obtidas com os descritores deHaralick. Duas investiga?? es foram realizadas: a determina??o do melhor conjunto de descritoresde Harlick usados como medidas de textura e an?lise de desemp enho dos descritores selecionadospara segmenta??o. Em ambas foram empregadas Redes Neurais Arti ciais MLP (MultilayerPerceptron) para a classi ca??o de ?reas de textura com e sem Acidente Vascular CerebralIsqu?mico (AVCi). Para a an?lise dos descritores de Haralick foram feitos testes de signi c?nciaestat?stica e testes de classi ca??o com a RNA para 1, 2 e 3 descritores de Haralick combinadosentre si. Os resultados obtidos foram comparados com o padr?o ouro, dado p ela segmenta??omanual das regi? es isqu?micas. Nos testes realizados o sistema obteve b ons resultados com oconjunto de descritores Correla??o, Homogeneidade e Soma das M?dias, que mostrou-se capazde promover a classi ca??o necess?ria para segmentar a regi?o cerebral atingida p elo AVC comprecis?o.

4 mal im Jahr erscheinende Zeitschrift über aktuelle Themen der TU Chemnitz

Background: This research was designed to understand factors influencing Motor Vehicle Crashes (MVCs) on or near the Hualapai Tribe’s reservation in northwestern Arizona. The goal is to enable the Tribe to develop and implement a locally relevant MVC intervention program. The specific aims were to: i) compile and analyze 2010-2016 MVC data from Hualapai and federal sources to assess distributions of frequency, rates, high risk-locales, causes, days, times, age and sex of the drivers, in addition to assessing related conditions; ii) document local perceptions of environmental, social, and behavioral barriers to safe driving practices, and knowledge of MVC risk factors and existing tribal laws; and iii) provide evidence-based recommendations using the results from quantitative and qualitative data analyses. Approach/Methods: The research process applied a Community-Based Participatory Research approach and mixed methods using: a) secondary data analyses of records from six tribal programs and Indian Health Service (IHS), and b) qualitative analyses of data from two focus groups and field documents. Sensitivity analyses were conducted of IHS and Tribal Driving Under the Influence (DUI) data, using the 2010 U.S. Census as the denominator after adjusting for an undercount of up to 25%. Descriptive statistics, Fisher’s Exact Test, and linear and logistic regressions were used to examine significance. MVCs per 10 miles per year were estimated for State Route 66 and Diamond Bar Road/Grand Canyon West on or near the reservation due to higher numbers of crashes. Statistical process control charts, especially g-charts monitoring time between events, were plotted to examine the stability in the number of MVCs over time for each road. NVivo11Pro© was used to code and analyze the focus group data, guided by both inductive and deductive theories. Results: Driver’s seatbelt use in the Hualapai community increased from 2010 to 2012 (p < .0005), and reported DUIs decreased from 2010 to 2016 (p = .027). Similarly, car/booster seat use improved from 2014 to 2015 (p < .0005). Two hundred and fifty (N = 250) MVC-related injuries were registered at IHS facilities for Hualapai community members between 2006 and 2015. The highest rate, 22.4%, was observed in 15- to 24-year-olds, followed by 21.6% in 55- to 64-year-olds. For several combinations of numerators and denominators, sensitivity analysis of the IHS data shows a clear disparity between the Tribe’s MVC rate compared to the 2008 U.S. rate of 771.4 nonfatal injuries per 100,000 persons and the national goal for Healthy People 2020 (694.3 nonfatal injuries/100,000 persons). The major themes emerged from qualitative analyses of the focus groups were: i) unsafe traffic infrastructures, ii) DUI, iii) repeated DUI offenders driving on the reservation, and iv) a perception of lenient tribal traffic laws and enforcement on the reservation. Using study results, five Public Service Announcements were co-developed to inspire community-members to continue the trends noted from 2010-2016 and be aware of continued risks. Conclusions: MVCs are a multidimensional issue needing communitywide awareness of the range of risk factors. An intervention that addresses human and structural risks requires an alliance of tribal programs and external partners (e.g., IHS, university, federal, and state). Recommendations: Local recommendations include providing school- and institution-based education about alcohol/alcoholism and DUI consequences, and ongoing culturally and locally relevant communitywide education through the local newsletter and radio station.

The central original contribution to knowledge proposed by this thesis is the setting forth of a conceptualisation of ethical theory specifically in relation to design, with a focus on visual communication design. Building on earlier work by design theorist Clive Dilnot in the area of design ethics and on philosopher Giorgio Agamben’s formulation of the philosophical concept of potentiality, a way of thinking about the relationship between design and ethics is proposed which concludes that design is in fact always inherently ethical. However, this conception of ethical design purposefully leaves questions of the qualification of good and bad unresolved, stating only that the ethical is the prerequisite condition in which both good and bad become possibilities. Design’s significantly unethical capability to suppress and anaesthetise individuals’ ethical experience is highlighted through a proposal of a process of an/aesth/ethics. Observation of the relationship between design and ethics in the real world through a series of interviews demonstrates something of the complexity of design’s relationship with ethics and the diverse range of positions, beliefs, attitudes and paradoxes abounding within the design profession when it comes to addressing the question of “good” design practice. Six “sites” of ethics within contemporary design discourse are introduced and discussed. The ethicality of design practices in relation to these sites are then analysed through the lens of the proposed ethical framework: identifying strengths, weaknesses and potentials within these observed strategies. The way of thinking about ethical design proposed here demonstrates potential in contributing to designers’ ability to critically consider the ethicality of their own practices. From this foundation they may be better equipped to begin addressing the question of the qualification of the “goodness” of design. In conclusion, proposals are made for how this framework could be practically developed and used to support and encourage ethical design in the real world.

Previous job interview studies have found that evaluations of English-as-an-additional language (L+) candidates related less to demonstrated qualifications and more to matches or mismatches in communicative expectations. Candidates’ pragmatic skillfulness can affect interviewers’ perceptions of their competence, and by extension, their hireability. Despite the importance of pragmatics to interview success, few studies have looked at the efficacy of pragmatics training. To address this gap, a mixed-methods study was ncarried out with L+ English university students and professional interviewers. Two training types – pragmatics-focused feedback (n = 9) and feedback plus a pragmatics lesson (n = 9) – were compared to a control (n = 9). A second focus was to understand the factors that influenced the nine interviewers’ evaluations. To this end, the interviewers engaged in a video-stimulated recall session. The resulting data were coded thematically. Finally, the interviewers’ communication was analyzed using an Interviewer Actions instrument and qualitative analysis. Results showed that both experimental groups significantly outperformed the control group, which provides an endorsement of pragmatics training for L+ candidates. A second finding was that language ability themes were most prevalent in interviewer comments. This reveals a self-referential emphasis on the candidates’ talk as the primary source of competency judgments, which disadvantages L+ speakers. The Interviewer Action scores, supported by candidate evaluations and comments, indicated that engaged and supportive interviewer communication was most favourably received by the candidates. However, the qualitative analysis highlighted the challenge for interviewers in engaging with candidates while maintaining neutrality vis-à-vis responses. With increasingly diverse candidate pools, interviewers must upgrade their communication skills to make confident judgments about all interviewees. / Graduate

HIV/AIDS threaten more than 40 million people worldwide and more than 5 million in South Africa alone. There is no cure for the disease yet, and novel drugs need to be discovered to make any progress in combating the disease. Twelve extracts from indigenous South African plants were analysed, of which one, Elaeodendron croceum, showed exceptionally good inhibition of transcription factors and a recombinant HIV strain in the HeLa-TAT-Luc and MT-2 VSV-pseudotyped recombinant virus assays. The pure compound isolated from this extract seemed to be the most toxic of all the samples, with toxicity of only 25% at a concentration of 100 _g/ml. When the concentration is increased, the toxicity increased slowly from 15% at a concentration of 0.195 _g/ml until it reached 25% toxicity at a concentration of 100 _g/ml. The active concentration of the compound against HIV is much lower at 100 ng/ml with an inhibition of approximately 90% of the recombinant virus. The therapeutic index of 250 makes it a promising possibility to be studied further for the compound to be used as a drug. The semi-purified extract and the pure compound were tested for its toxicity on VERO cells. The semi-purified extract had no toxicity up to a concentration of 50 _g/ml and the pure compound had toxicity of 20 % up to a concentration of 25_g/ml. The active concentration of 100 ng/ml for the VSV-Pseudotype assay is much lower than the start of toxicity at 25 _g/ml, and leaves a margin of activity before the toxicity level is reached. Both the extract and pure compound shows promising results in vitro to be developed into a medicine to be used against HIV, but need more research on the effects in vivo. Using an extract is easier, cheaper and faster than isolating a pure compound from the extract. It might also be possible that the extract could be prepared as a tea and its use could be very accessible. / Thesis (PhD (Plant Physiology))--University of Pretoria, 2007. / Plant Science / unrestricted