A hierarchical examination of genetic diversification: Conservation implications for parrots in the genus Amazona

Historical information embedded within the nuclear and organellar genomes affords the opportunity to investigate evolutionary hypotheses at different levels of resolution. This dissertation utilizes a suite of nuclear and mitochondrial DNA markers within a hierarchical framework to address outstanding questions in Neotropical parrot systematics and conservation.; In order to investigate several hypotheses regarding the evolutionary history of Amazon parrots (Genus Amazona), chapter 2 presents a combined phylogenetic analysis based on data from six molecular data partitions including mitochondrial and nuclear regions. The results demonstrate that Amazona is not monophyletic with respect to the placement of A. xanthops and suggest a Neotropical short-tailed parrot genus as sister to all Amazon parrots. At a finer level, the phylogeny resolves the Greater Antillean endemic Amazona as constituting a monophyletic group while further revealing a paraphyletic history for the extant Amazon species of the Lesser Antilles. The reconstructed phylogeny provides further insights into the mainland sources of Antillean Amazona, reveals areas of taxonomic uncertainty within the genus, and presents historical information that may be included in conservation priority-setting for Amazon parrots.; At the population level, chapter 3 investigates the ability of molecular markers to provide demographic and genetic information to facilitate the goals of ex situ conservation management, as exemplified by the on-island captive breeding program of the endangered St. Vincent Amazon parrot (Amazona guildingii). The degree to which empirical estimates of relatedness among individuals of unknown ancestry may improve upon the assumptions of conventional pedigree-based management is investigated using genotypic data collected at eight microsatellite loci and a suite of marker-based relatedness estimators. The measure of Lynch and Ritland (1999), r_xyLR, is found to explain the highest amount of variation in true relatedness, as revealed through Monte Carlo simulation. Integration of pairwise r_xyLR estimates of founder relatedness into the A. guildingii studbook is demonstrated to minimize mean kinship and maximize gene diversity to a greater extent than those derived from the founder assumption of zero relatedness. Within this framework, an integrated conservation management strategy for the A. guildingii population on St. Vincent is set forth, incorporating all available demographic and genetic information.