Advancements in molecular sequencing have revealed unexpected cryptic genetic diversity and contrasting evolutionary histories within genes and between genomes of many organisms; often in disagreement with recognized taxonomy. Incongruent patterns between the mitochondrial and nuclear evolutionary history can have several plausible explanations, but widespread systematic conflict inevitably challenges our conceptions of species boundaries when there is discordance between coevolving and coinherited genomes. It is unknown to what degree mitonuclear conflict drives the process of divergence, or how ubiquitous these patterns are across the tree of life. To understand the evolutionary relevance of intergenomic discordance we must identify the conflicting patterns that exist in natural systems by generating robust estimates of the underlying species history, quantify support for alternative hypotheses of lineage formation, and describe patterns of genetic variation present in robust nuclear genomic datasets. Empirically testing correlations between mitonuclear genomic conflict and reduced gene flow at the organism level will contribute toward a better understanding of lineage boundaries and how intergenomic interactions shape the process of divergence. Mitochondrial introgression has been inferred in many salamander systems with limited perspective from nuclear sequence data. Within dusky salamanders (Desmognathus), these patterns have been observed between morphologically and geographically disparate populations. I sequenced regions throughout the nuclear genome to reconstruct species trees, performed population-level analyses testing concordance between the mitochondrial, nuclear datasets, and nuclear genes with mitochondrial functions with the expectation that coevolutionary interactions among genomes are more likely to manifest in these regions. I also estimated migration rates between populations that may have experienced historical mitochondrial introgression to evaluate phylogeographic patterns. Using these data we definitively reject species models in which genetic boundaries are based solely on mitochondrial data, favoring geographic models instead. Furthermore, analyses soundly reject current taxonomic models based on morphological characteristics, suggesting there is greater lineage diversity than is currently recognized. I also used empirical assays of pre-zygotic reproductive mating behavior within and among populations containing diverse mitochondrial lineages to test metrics of reproductive isolation, and to determine if introgression shapes the evolution of complex traits directly influencing rates of divergence. These results may explain incongruent patterns observed between the mitochondrial and nuclear data as a function of inheritance and population dynamics rather than directly functioning to suppress nuclear gene flow. This research builds upon recent studies suggesting that speciation is a highly complex and often non-bifurcating process in which introgression can have a profound and lasting signature on the nuclear evolutionary history. Mechanisms responsible for divergence with gene flow challenge evolutionary biologists to reevaluate our notions and definitions of species boundaries to accommodate seemingly conflicted genomic patterns within and between genomes.

Recent models argue that thermal environments are the major cause of ectotherm life-history clines. However, elevational clines in body size in the mountain dusky salamander Desmognathus ocoee (family Plethodontidae) shift from positive at hatching, to negative at metamorphosis to positive again as adults, and so are not consistent with this explanation. The clinal shift from hatching to metamorphosis was investigated by examining the clinal and seasonal feeding patterns of larval salamanders at high and low elevation sites in rockface and woodland habitats. Repeated cohort sampling was also used to examine clinal and seasonal patterns in body size and to estimate average growth rates. Larval growth in both rockface and woodland habitats was tightly correlated with feeding activity. Although temperature was found to vary between high and low elevation sites, the greatest growth occurred in a cold woodland habitat with a high elevation, and the lowest growth occurred in an adjacent rockface habitat. Because this difference in growth cannot be attributed to thermal differences, we conclude that local food resource levels are the predominant source of local differences in growth. These findings, clinal patterns of variation in other predatory salamanders, and experimental analyses in which both food and temperature are orthogonally manipulated, indicate that general models that single out temperature as the principle cause of ectotherm life-history clines should be viewed with caution.

Understanding the mechanisms that generate shared morphologies across closely related taxa is important when identifying distinct evolutionary lineages using morphological characters. Desmognathus salamanders are an ideal group for testing hypotheses concerning the correlation between morphological similarity and genetic exchange within and among nominal species due to a pattern of high discordance between the two. Phylogeographic hypotheses are tested for populations of the D. quadramaculatus species complex throughout southern Appalachia by combining phylogenetic and population genetic methods with geographical information. Phylogenetic and phylogeographic inferences are then assessed in conjunction with morphological characteristics that have traditionally diagnosed taxonomic entities to understand the genetic basis of shared morphology in this complex, and to assess species boundaries. A history of fragmentation followed by range expansion is suggested as a recurrent pattern that has shaped the current population structure within this complex. The current taxonomy is found to unite populations that share similar morphologies due to parallel evolution rather than ancestry. We suggest revisions in taxonomy that will better reflect the evolutionary history of these lineages. Appreciation of the hidden genetic variation and homoplasious morphological variation often present in and among salamander species can foster the implementation of more appropriate methods for detecting and recognizing the complex history of these organisms.

The introduction of nonnative salmonids into Southern Appalachia may pose a threat to resident salamander populations. In recent years, the stocking and encroachment of rainbow trout (Oncorhynchus mykiss) into headwaters where the black-bellied salamander (Desmognathus quadramaculatus) and brook trout (Salvelinus fontinalis) naturally coexist has raised concerns. In aquatic prey, predator-avoidance responses are primarily influenced through the detection of chemical cues released from predators. The objective of this study was to determine how co-occurrence with a predator influences black-bellied salamander predator recognition behavior. To evaluate this, salamander activity metrics (general activity, number of movements, and latency to move) were recorded before and after exposure to either native trout predator cue (brook), introduced trout predator cue (rainbow), or conditioned tap water (control). Larvae were collected from different streams based on their trout predator assemblage with larvae coming from brook, rainbow, rainbow/brook, and no trout stream reaches. Our results show that larvae that co-occur with trout reduced their activity when exposed to brook trout predator cue, but their response to rainbow trout predator cue depended on their previous co-occurrence. Larvae from areas with only brook trout exhibited a weak predator-avoidance when exposed to rainbow trout predator cue. A follow-up test to determine the influence of alarm cue on predator response in these larvae indicated that the alarm cue enhanced the response to the rainbow trout predator.

Large-scale land uses such as residential wastewater discharge and coal mining practices, particularly surface coal extraction and associated valley fills, are of particular ecological concern in central Appalachia. Identification and quantification of both large-scale land use and microhabitat alterations to ecosystems are a necessary first-step aid in mitigation of negative consequences to biota. In central Appalachian headwater streams absent of fish, salamanders are the dominant, most abundant vertebrate predator providing a significant intermediate trophic role. Stream salamander species are considered to be sensitive to aquatic stressors and environmental alterations with past research having shown linkages among microhabitat parameters, and large-scale land use such as urbanization and logging to salamander abundances. However, there is little information examining these linkages in the coalfields of central Appalachia. In the summer of 2013, I visited 70 sites (sampled three times each) in the southwest Virginia coalfields to collect salamanders and quantify stream and riparian microhabitat parameters. In an information-theoretic framework I compared the effects of microhabitat and large-scale land use on salamander occupancy and abundances. My findings indicate that dusky salamander (Desmognathus spp.) occupancy and abundances are more correlated to microhabitat parameters such as canopy cover than to subwatershed land uses. Brook salamander (Eurycea spp.) occupancy show negative associations to large-scale land uses such as percent recent mining and percent forested. Whereas Eurycea spp. abundances are negatively influenced by suspended sediments, stream bank erosion and stream substrate embeddedness. Management implications of these findings include erosion prevention and control as well as protection and management of riparian habitats. However, quantifying physical environmental quality such as stream and riparian habitat often can be quite difficult, particularly when there are time or fiscal limitations. In order to accurately assess stream and riparian habitat in a time- and cost- effective manner, the U.S. Army Corps of Engineers (USACE) developed a functional condition index (FCI) assessment for streams that measures 11 stream and riparian parameters along with watershed land use to calculate three different scores: a hydrology score, biogeochemical score, and habitat score (Noble et al 2010). Using the salamander data from 2013, I then analyzed the FCI scores using collected occupancy and abundance analyses. Both analyses supported the Habitat FCI score as it had strong correlations with both occupancy and abundance of three Desmognathus spp., and support the use of the USACE protocol for stream and riparian habitat assessment. / Ph. D.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.