O Triássico é um período de grande importância na história da vida na terra, durante o qual grandes eventos evolutivos são registrados dentre os vertebrados. Para melhor compreender a transição entre a fauna dominada por terápsidos no início do período para uma fauna dominada por dinossauros no final, tem se intensificado os estudos correlacionando sedimentos triássicos de forma global. Essencial para subsidiar tais correlações, é uma bioestratigrafia regional detalhada. No Brasil, tetrápodes triássicos são conhecidos apenas nos depósitos do Grupo Rosário do Sul, no Rio Grande do Sul. Com base nesses fósseis, muitas propostas bioestratigráficas foram elaboradas, ainda que de forma preliminar. Visto que a busca por novos sítios fossilíferos e descoberta de novos fósseis têm gerado grande acúmulo de dados sobre o Triássico brasileiro, mas poucos trabalhos de catalogação dessa informação foram desenvolvidos, acreditamos que parte das dificuldades no estabelecimento de um zoneamento bioestratigráfico de maior detalhe, neste caso, é decorrente de uma disponibilização ainda inadequada dos dados referentes ao conteúdo fossilífero de cada uma das localidades prospectadas na região. Visando sanar tal obstáculo, o presente trabalho teve como objetivo inicial a elaboração de um banco de dados acerca de tais tetrápodes triássicos, com o resgate das informações já publicadas sobre o tema, além de levantamento in loco de espécimes catalogados em determinadas coleções do Estado. Este banco de dados foi utilizado para testar as propostas bioestratigráficas vigentes através do estudo das similaridades faunísticas, com detalhamento em nível de localidades, através da aplicação de estatística multivariada (análises de cluster) em diferentes graus de abrangência taxonômica. Foram utilizados os índices de similaridade de Jaccard (qualitativo) e de Morisita-Horn (quantitativo) para a construção da matriz de distância, e o método de agrupamento de Ward para geração dos dendogramas. Os agrupamentos resultantes de análises qualitativas foram os que mais coincidiram com as propostas bioestratigráficas prévias. As demais análises, quantitativas, não resultaram em agrupamentos verossímeis, o que não se deve a problemas na utilização do método, mas da utilização de táxons que não são adequados a estudos bioestratigráficos. / The Triassic is a period of great importance in the history of life on Earth, where important evolutionary events occurred among the vertebrate lineages. To better understand the transition between a terapsid dominated fauna of the beginning of the period to one dominated by dinosaurs at its end, studies of the correlation of Triassic sediments at a global scale have been intensifying in recent years. Essential to subsidize such correlations is a detailed regional biostratigraphy. In Brazil, Triassic tetrapods are known only in the deposits of the Rosario do Sul Group, Rio Grande do Sul State. Based on these fossils, many biostratigraphical proposals have been elaborated, but are still considered as preliminary ones. As the search for new fossil sites and new discoveries generate an accumulation of data about the Brazilian Triassic, few works on the allocation of this information have been developed, however, and we believe that part of the difficulty in establishing a more detailed biostratigraphical zonings for this is due to the inadequate availability of data on the fossils of every prospected locality in the region. As a means to resolve this obstacle, the main objective of the present work was the construction of a Database of Triassic tetrapods, along with the information already published on the subject, added to an overview in loco of catalogued specimens of certain scientific collections of the State. This Database was used to test the established biostratigraphical proposals by means of a study on faunistic similarities, with detainment at location level, by use of multivariated statistics (cluster analysis) in different levels of taxonomic amplitude. We used Jaccard (qualitative) and Moristia-Hom (quantitative) similarity indexes to construct the distance matrix, and the Wards method of clustering to generate the dendograms. The resulting qualitive analysis produced clusteres that matched more closely with previous biostratigraphical proposals. The other quantative analysis did not result in stable clusters, but this is not due to methodological problems, however, but due to the use of taxa that are not sucitable for biostratigraphical studies.

O Triássico é um período de grande importância na história da vida na terra, durante o qual grandes eventos evolutivos são registrados dentre os vertebrados. Para melhor compreender a transição entre a fauna dominada por terápsidos no início do período para uma fauna dominada por dinossauros no final, tem se intensificado os estudos correlacionando sedimentos triássicos de forma global. Essencial para subsidiar tais correlações, é uma bioestratigrafia regional detalhada. No Brasil, tetrápodes triássicos são conhecidos apenas nos depósitos do Grupo Rosário do Sul, no Rio Grande do Sul. Com base nesses fósseis, muitas propostas bioestratigráficas foram elaboradas, ainda que de forma preliminar. Visto que a busca por novos sítios fossilíferos e descoberta de novos fósseis têm gerado grande acúmulo de dados sobre o Triássico brasileiro, mas poucos trabalhos de catalogação dessa informação foram desenvolvidos, acreditamos que parte das dificuldades no estabelecimento de um zoneamento bioestratigráfico de maior detalhe, neste caso, é decorrente de uma disponibilização ainda inadequada dos dados referentes ao conteúdo fossilífero de cada uma das localidades prospectadas na região. Visando sanar tal obstáculo, o presente trabalho teve como objetivo inicial a elaboração de um banco de dados acerca de tais tetrápodes triássicos, com o resgate das informações já publicadas sobre o tema, além de levantamento in loco de espécimes catalogados em determinadas coleções do Estado. Este banco de dados foi utilizado para testar as propostas bioestratigráficas vigentes através do estudo das similaridades faunísticas, com detalhamento em nível de localidades, através da aplicação de estatística multivariada (análises de cluster) em diferentes graus de abrangência taxonômica. Foram utilizados os índices de similaridade de Jaccard (qualitativo) e de Morisita-Horn (quantitativo) para a construção da matriz de distância, e o método de agrupamento de Ward para geração dos dendogramas. Os agrupamentos resultantes de análises qualitativas foram os que mais coincidiram com as propostas bioestratigráficas prévias. As demais análises, quantitativas, não resultaram em agrupamentos verossímeis, o que não se deve a problemas na utilização do método, mas da utilização de táxons que não são adequados a estudos bioestratigráficos. / The Triassic is a period of great importance in the history of life on Earth, where important evolutionary events occurred among the vertebrate lineages. To better understand the transition between a terapsid dominated fauna of the beginning of the period to one dominated by dinosaurs at its end, studies of the correlation of Triassic sediments at a global scale have been intensifying in recent years. Essential to subsidize such correlations is a detailed regional biostratigraphy. In Brazil, Triassic tetrapods are known only in the deposits of the Rosario do Sul Group, Rio Grande do Sul State. Based on these fossils, many biostratigraphical proposals have been elaborated, but are still considered as preliminary ones. As the search for new fossil sites and new discoveries generate an accumulation of data about the Brazilian Triassic, few works on the allocation of this information have been developed, however, and we believe that part of the difficulty in establishing a more detailed biostratigraphical zonings for this is due to the inadequate availability of data on the fossils of every prospected locality in the region. As a means to resolve this obstacle, the main objective of the present work was the construction of a Database of Triassic tetrapods, along with the information already published on the subject, added to an overview in loco of catalogued specimens of certain scientific collections of the State. This Database was used to test the established biostratigraphical proposals by means of a study on faunistic similarities, with detainment at location level, by use of multivariated statistics (cluster analysis) in different levels of taxonomic amplitude. We used Jaccard (qualitative) and Moristia-Hom (quantitative) similarity indexes to construct the distance matrix, and the Wards method of clustering to generate the dendograms. The resulting qualitive analysis produced clusteres that matched more closely with previous biostratigraphical proposals. The other quantative analysis did not result in stable clusters, but this is not due to methodological problems, however, but due to the use of taxa that are not sucitable for biostratigraphical studies.

Dry forest is one of the most threatened tropical forests in the world. Human impact has caused its massive transformation but conservation of dry forest has often been neglected across Latin America. In Colombia, less than 10% of the original extension of dry forest remains. This thesis studies the phytogeography of neotropical dry forest and its relevance for conservation using data from 1602 tree species inventories made in dry forests across Latin America and the Caribbean synthesised by The Latin American Seasonally Dry Tropical Forest Floristic Network (DRYFLOR). Clustering and ordination analyses were used to explore the floristic relationships of dry forest across the entire Neotropics, revealing distinctive regional clusters defined by their tree species composition. Colombian dry forests are shown to be part of two wider clusters, one including neighbouring forests in Venezuela and southern Central America, and the second including the inter-Andean dry forests. The high turnover of floristic diversity and endemism within and amongst the main floristic groups demonstrates that to conserve the full species diversity in dry forests across Latin America and the Caribbean will require protecting it simultaneously across multiple regions. A regionally focused study of floristic relationships in the Central American and northern South American dry tropical forest group, using quantitative approaches to conservation prioritization, including a new Conservation Priority Index (CPI), suggests that conservation priority should be placed on the South American Caribbean Coast in the cross-border area of Colombia and Venezuela. This emphasises the need for a biogeographical approach to conservation that cannot be restricted by political borders. Within Colombia, new quantitative floristic data were used to investigate controls of floristic composition in dry forests. Multivariate analyses showed that space related variables explain a larger fraction of the variance of the floristic composition than climatic or edaphic variables. The importance of spatial variables implies that biogeography is a key element in understanding the structure of communities, and that the Andean cordilleras might be acting as geographical barriers isolating these seasonally dry formations. The value of floristic inventory data for assessing the conservation status of tree species using IUCN criteria was assessed in a case of study of the Andean Piedmont dry forest. By combining inventory data from the DRYFLOR database and herbarium records, the number of species for which we have sufficient information to make conservation assessments increases by 16% and the accuracy of predictive species distribution improves for 84% of the species. Together, these results reveal the importance of ecological inventory data as a complementary data source in conservation assessment for dry forest trees in the Neotropics. Finally, the conclusions chapter places these results in the context of conservation planning for Colombian dry forests, including some suggestions for research, policies and actions. These actions include restoration programmes focusing on sustainable harvesting of native dry forest tree species, for example for firewood and other forest resources such as fruits, fibres and medicines. A land use mosaic, including forest fallows and strict conservation areas, may help to guarantee the long-term maintenance dry forest species in Colombia.