<p>Islands have been the inspiration for some of evolutionary biology's most important advances. This is largely due to the unique properties of islands that promote the differentiation of island species from their mainland counterparts. Rodents are widely distributed across even the most remote islands, a rarity among mammals, making them uniquely suited to study the factors leading to the divergence of insular species. In this dissertation, I use two case studies to examine the morphological and genetic divergences that take place in an insular environment.</p><p>In chapters one and two, I examine how different factors influence insular body size change in rodents. In chapter one, I examine factors influencing the direction of island body size change using classification tree and random forest (CART) analyses. I observe strong consistency in the direction of size change within islands and within species, but little consistency at broader taxonomic scales. Including island and species traits in the CART analyses, I find mainland body mass to be the most important factor influencing size change. Other variables are significant, though their roles seem to be context-dependent.</p><p>In chapter two, I use the distributions of mainland rodent population body sizes to identify `extreme' insular rodent populations and compare traits associated with those populations and their islands with those island populations of a more typical size. I find that althought there is no trend among all insular rodents towards a larger or smaller size, `extreme' populations are more likely to increase in size. Using CART methods, I develop a predictive model for insular size change that identifies resource limitations as the main driver when insular rodent populations become `extremely small'. </p><p>Chapters three and four shift their focus to a single rodent species, the deer mouse <italic>Peromyscus maniculatus</italic>, as they examine the genetic differentiation of deer mice across the California Channel Islands and the nearby mainland. In chapter three, I sequence a region of the mitochondrial control region for individuals from 8 populations across the northern Channel Islands and two mainland sites, and I analyze these sequences by calculating population genetics parameters and creating a Bayesian inference tree and a statistical parsimony haplotype network. All of these analyses reveal significant divergences between island and mainland populations. Among the islands, Santa Barbara and Anacapa islands both display unique genetic signatures, but the other northern islands remain relatively undifferentiated.</p><p>In chapter four, I genotype individuals from the previous chapter at 5 microsatellite loci, I calculate additional population genetics parameters and I utilize a Bayesian clustering algorithm to examine the similarities and differences between nuclear and mitochondrial analyses. I find the nuclear data to be largely congruent with the mitochondrial analyses; there are significant differences between island and mainland populations, and Anacapa Island is significantly differentiated from the other islands. Unlike the previous analyses, Santa Barbara Island is not significantly different from the northern islands, yet San Miguel Island has a unique genetic signature. </p><p>These studies underscore the importance of ecological processes and historical biogeography in the generation of diversity, and they highlight the role of islands as drivers of evolutionary divergence.</p> / Dissertation

No description available.

The influence of human disturbance on the nesting success of the Common eider (Sonateria mollissima) was investigated in the Mingan Archipelago National Park Reserve during summers of 1994 to 19%. Six colonies were used, four located on open islets and two on forested islands. Three experimental treatments were used that entailed varying the frequency and time occurrence of visits in incubation. Every years, each treatment was applied to several eider colonies differing in the density of nesting gulls on the same island. During visits, incubating females were flushed and nest fate recorded. Several variables also were measured during visits describing female quality (assessed by clutch size and laying date) and nest site quality (assessed by nest concealment, distance to shore, vegetation height, and density of nesting gulls on the island). Results showed that frequency of visits did not affect nesting success of eiders, mainly because most observed nest failures occurred after the first visit The time occurrence of the first visit during incubation was of primary importance; eider nests visited early were much more affected by our presence. The nesting success of eiders was related to the density of associated colonies of nesting gulls only when visits occurred during early incubation. Female eiders of greater reproductive quality were those most successful when disturbed regularly. These females were in greater proportion on islands associated with lower density of nesting gulls, i.e., on forested islands rather than on open islets. Within colonies, they were mainly found nesting near shore where nest concealment is sparse. Results showed that visits to eider colonies can greatly affect their reproductive output. Visits during early incubation on open islets had the greatest impact.

Breeding biology and diet of Herring Gulls Larus argentatus were studied on Ile Nue, in the Mingan Archipelago National Park Reserve, in the Gulf of St. Lawrence. Nesting parameters were measured 1994--1996, and fledging success and diet were followed in 1996. Clutch size was low in all three years of the study, and hatching success was low in 1995. Low clutch size and hatching success could not be explained by researcher disturbance, predation, or competition for nest sites, leaving food-stress as the most likely explanation. To reduce the biases associated with determining diet, we used four methods to estimate diet. The 1996 breeding season was divided into three periods using changes in the number of gulf pellets collected from roost sites as an indirect measure of dietary changes; these periods corresponded to the pre-spawning, spawning, and post-spawning periods of capelin Mallotus villosus. (Abstract shortened by UMI.)

No description available.

Mode of access: World Wide Web. / Thesis (Ph. D.)--University of Hawaii at Manoa, 2005. / Includes bibliographical references (leaves 136-146). / Electronic reproduction. / Also available by subscription via World Wide Web / 146 leaves, bound ill. 29 cm

Thesis (M.A.)--University of Hawaii at Manoa, 1993 / Pacific Islands Studies

Thesis (M.A.)--University of Hawaii at Manoa, 2006 / Pacific Islands Studies

Thesis (M.A.)--University of Hawaii at Manoa, 2010 / Pacific Islands Studies

Thesis (M.A.)--University of Hawaii at Manoa, 1983 / Pacific Islands Studies