| Popis: |
This dissertation aims to investigate the influence of morphological, physiological, and natural history traits on the evolution of lizards and to use this information to predict conservation status in data deficient species. My dissertation was divided in three research chapters which aimed to answer the following ecological, evolutionary and conservation questions: Chapter 2) which are the important traits to predict species into open or forested environments? Chapter 3) how Pleistocene climatic fluctuations affected morphological variability within species? Chapter 4) what traits can be used to predict the conservation status of data deficient squamate species around the world?In Chapter 2, I used species occurrence (i.e., open, forested or both environments), species traits, and a phylogenetic information for Brazilian lizards to i) identify the most common colonization direction (from open to forested, or from forested to open environments), ii) test if the diversification rates were affected by the different environments, and iii) determine the most important traits to predict the colonization and survival of lizards in one or the other environment type. I found that 80% of the transitions were from the forested to the open habitats. Next, I tested 12 evolutionary models using a Hidden Geographic State Speciation and Extinction analysis and found that the different environments influenced the diversification rates. Finally, I created a predictive model using phylogenetic relatedness, species traits and occurrence in open or forested habitats. I identified four species traits as important in predicting species occurrence, in addition to phylogenetic relatedness. The approaches I used in this chapter should be able to help the identification of phenotypic transitions between different environments in regions not explored here and reveal traits to be used to understand the genomic basis of adaptation.In Chapter 3, I collected individual-level morphometric measurements for body size, hind limb size and head depth, length, and width for five species of Neotropical lizards and used these data to investigate two spatial hypotheses. In the first hypothesis, niche variation, I predicted increase of the morphometric variability of the head measurements as the populations get more distant from climatically stable areas. In the second hypothesis, spatial sorting, I expected the hind limb size to increase as the individuals get more distant from the climatically stable areas. To test these two hypotheses, I built niche models for five species of lizard for three time slices: present, Mid-Holocene, and Last Glacial Maximum. Then, I combined the resultant maps to create a niche suitability map and select the 95% percentile of the most suitable areas through time (i.e., climatically stable areas). Finally, I used linear models to test the two hypotheses, after removing the effects of body size. My findings did not support either of the hypotheses (niche variation and spatial sorting).In Chapter 4, I gathered occurrence and distribution data, and trait data for lizards and amphisbaenians around the globe. I combined these data with bioclimatic data and a machine learning approach to i) identify important variables to predict if species are threatened or non- threatened; ii) use the identified important variables and the trained model to predict conservation status of data deficient species. With an accuracy of 84%, the model indicated occurrence area (or species distribution) as the most important variable to predict species conservation status. Among the data deficient species, more than 50% were predicted to be threatened by my trained model. The classification of those species as threatened have the purpose to bring attention to them in a way they can be prioritized for data collection and future conservation assessment. |
