<p> Habitat loss and fragmentation (hereafter fragmentation) are some of the largest conservation threats today and will increasingly put pressure on species in the future. Maintaining population connectivity helps mitigate the negative effects fragmentation has on vulnerable species, especially those of conservation concern. In this dissertation, I elucidated the movement ecology of black-capped vireos by 1) estimating and characterizing patterns of gene flow, 2) examining the relationships between land cover and connectivity, and 3) simulating how future populations respond to climatic landscape change. I additionally characterized potential biases in family-wise error rate correction across population genetic studies, a correction important for evaluating the genetic structure of a species. I genotyped 343 individuals at 12 microsatellite loci in and around Fort Hood, Texas, which houses the largest and most stable breeding population. To characterize patterns of gene flow among black-capped vireo populations, I analyzed genetic differentiation, migration rates, number of migrants and parentage. Across these independent analyses, I found evidence for asymmetrical movements from Fort Hood to the other central Texas sites consistent with source-sink dynamics and findings from demographic studies. I used gravity models to test the relationships among Euclidean distance, land cover types (water, developed, forest, scrub, open, agriculture and wetlands), brown-headed cowbird control, and genetic similarity. My findings indicate that wetlands, likely via riparian areas, may be acting as corridors among populations. Better understanding of what factors influence connectivity will be crucial for maintaining connectivity when species of conservation concern are threatened by fragmentation.</p><p>

Why do so many organisms invest in multiple, multi-modal, sexually selected traits? Two hypotheses explain the evolution of multiple sexually selected traits: (1) the Redundant Signals hypothesis, which states that multiple traits broadcast overlapping information, either to amplify the signal, or to compensate for unreliable signals, and (2) the Multiple Messages hypothesis, which states that each trait broadcasts different information. I tested the redundant signals and multiple messages hypotheses by assessing multiple color and song traits in a colorful migratory songbird, the chestnut-sided warbler (Dendroica pensylvanica). For each warbler, I measured the area and spectral qualities of three plumage patches: the chestnut stripe, black facial mask, and yellow cap. I also measured several singing performance parameters for each of the two categories of songs sung by males of this species. In addition, I explored how color and song traits inform social and genetic mate choice and how female color relates to female defensive behavior. My results indicate that, in general, males with larger and brighter color patches sang their songs more consistently. However, color traits appear to be related to social mate quality, while song traits appear to be related to genetic mate choice. Males and females paired assortatively by color patch area, and colorful males provisioned nestlings at higher rates, while colorful females produced heavier young. Colorful females also exhibited quieter, more cryptic behavior during threats to their nests, while drab females used chipping displays against potential threats. Male warblers with better singing performance experienced more cuckholdry in their social nests, but also sired more extra-pair reproductive young. I interpret these results as evidence for a partial separation in signal content between the two types of sexual signals used by chestnut-sided warblers, with color indicating social mate quality, and song indicating genetic mate quality. And so, color and song traits in chestnut-sided warblers seem to function as imperfectly redundant signals that may also contain multiple messages.

The Belted Kingfisher is a non-passerine aerial piscivore of relatively recent tropical descent that breeds in earthen burrows along watercourses throughout the temperate zone of North America. In the following chapters I explore the ways in which this suite of life history traits may influence patterns of mortality and fecundity among individuals and thus help to shape the overall behavioral ecology of this species. All kingfishers defending territories in the Connecticut River valley during the winters of 1992–1998 were males. These overwintering males were adults that previously bred in the area and juveniles that dispersed into the area the preceding autumn. Banding and museum-skin data suggest that this pattern of unequal sex distribution during the winter occurs across North America at latitudes north of approximately 40–45°N. Because males and females are equal in body size and social dominance, it appears that advantages associated with year-round breeding-site occupation outweigh the potential costs of overwintering more often for males than for females. One way that males may reduce the potential costs of overwintering at high latitudes is to roost at night in earthen burrows, a habit that I found can reduce a bird's nocturnal thermoregulatory costs by ∼20%. The enhanced probability for burrow-roosting kingfishers to survive climatic extremes may represent an adaptation allowing this species to extend the northerly limit of its non-breeding geographic range. Like many temperate-breeding birds that nest in cavities, nesting success for kingfishers along the Connecticut River was high; females almost invariably laid 7 eggs, few of these eggs failed to hatch (<10%), and few nestlings failed to fledge (<15%). Like many socially monogamous non-passerines, males and females shared all phases of parental care more or less equally, including incubation. Testosterone levels of males peaked during the sexual phase of breeding and remained both high and variable until dropping abruptly once nestlings hatched. Perhaps reflecting differences in post-breeding agendas, the sexual differences in “parental motivation” that I did observe among kingfishers tended to occur late in the nestling period, when males fed the young more often and females more often abandoned.

I studied the spatio-temporal ecology, habitat selection, and survival of tropical hares Lepus flavigularis in Oaxaca, Mexico. Home range size and overlap were estimated to insight into tropical hare's mating behavior and social organization. Habitat selection and survival rates were determined to identify key habitat types and cause-specific mortality for conservation actions. I radio-tracked 51 hares in a savanna of the Isthmus of Tehuantepec, from February 2001 to July 2003. Annual home ranges and core areas of adults averaged 0.56 and 0.09 km2, respectively. Seasonal home ranges of adults varied from 0.22 to 1.11 km2 for females, and from 0.24 to 1.66 km2 for males. Seasonal core areas varied from 0.03 to 0.19 km2 for females, and from 0.02 to 0.20 km 2 for males. Juvenile home range varied from 0.07 to 0.49 km 2 for females, and from 0.11 to 2.64 km2 for males. Home range overlap with more than one individual suggests tropical hares show a polygamous mating behavior, and a non-territorial social organization. Hares selected home ranges with relatively more grassy and sparse shrubby habitats and less dense vegetation. Hares rested during daytime, and favored savanna with bushes of Byrsonima crassifolia that probably provided cover from predators. Hares foraged during crepuscular and nocturnal hours, and favored savanna with scattered trees of Crescentia spp. that allowed visual detection of predators. No effects of sex and season on range size or habitat selection were detected. Annual survival for adults was 0.43, and survival during the wet season (0.56) was lower than during the dry season (0.79), particularly for females. Survival of juvenile females was low during the dry (0.06) and wet (0.15) seasons when compared to juvenile males survival (0.35 and 0.48). Predation was the major cause of mortality with 67% of adult and 94% of juvenile deaths. Induced fires and poaching accounted for 20% and 13% of adult deaths, respectively. Preservation of native vegetation structure in savannas is needed for tropical hare conservation.

Forty-seven mountain and silky cuscuses, and coppery ringtails were radio-tagged at Mt. Stolle, Papua New Guinea, from June 1995 to July 1998, and 15 of these were closely monitored. A total of 4,922 trees of ≥10 cm diameter were measured and identified from 5 randomly chosen hectares. About 70% of the trees were <20 cm in diameter, and <15 m high. Trees of the families Lauraceae, Myrtaceae, and Guttiferae represented >45% of the sample. Bootstrap averages of numbers of trees and species per hectare were 978 and 119 respectively. Daytime searches required significantly more effort to find animals, and provided significantly fewer captured animals per unit effort, than nighttime searches. However, chances of capturing animals once detected were higher during daytime. Drug doses of 9–17 mg Kg−1 immobilized animals within 1 minute, effects lasted 19 minutes, and recovery took an additional 28 minutes. Data for 10 species of Phalangerids and Pseudocheirids from 3 museum collections, and from measurements taken directly from the radio-tagged animals, were used to investigate the degree and nature of sexual dimorphism. Correlates of dimorphism were evaluated in two of the monitored species by analyzing behavioral and ecological data, including behavior budgets for three broad categories (eating, sitting, and walking), time between feeding bouts, home range size, absolute and relative distance traveled per hour, and five denning hole characteristics (type, visibility, height, dbh and height of tree). Sexual dimorphism in which males are larger than females exists in New Guinean medium-sized arboreal marsupial species, but not as commonly as previously reported. Seasonal weight fluctuations in correlation with reproductive status were observed in mountain cuscus females only. Lastly, animals had selective diets and ate mostly superabundant (>10 trees/hectare) species; also, overlap in the diet between the sexes did not differ from random chance, and females had more diverse diets than males. Dens were numerous, but very few were used more than 5 times. Home ranges overlapped very little in animals of the same sex, but little to extensively between sexes. Behavioral data suggest a facultative polygynous mating system in mountain cuscus and coppery ringtail.

To understand how species' distributions vary with landscape and climatic changes, we must first understand the proximate mechanisms responsible for distributional boundaries. A direct physiological link between the Virginia opossum's (Didelphis virginiana) winter energetics and its northern geographical range limit has been hypothesized. However, opossums now occur well beyond the predicted limit. Though opossums commonly occur in the Connecticut River Valley of central Massachusetts, I found in a road-kill survey that opossums are rare in undeveloped areas. These findings suggest that the region is a local distributional edge. Demographic models indicated that northern population persistence depends upon high (0.67–0.81) over-winter survival of subadult females. Revisiting winter energetics, I found that average-weight subadult females should not, under natural conditions, survive most Amherst, Massachusetts winters at these rates. This model explains the failure of opossums in natural areas but not their presence in developed areas. I compared winter temperatures across the region; while urbanized areas had the warmest winter nights (and hence the environment in which opossums have the highest chance of survival), urban temperatures should still have been too cold for sufficient subadult female survival. Radio-monitoring free-ranging opossums, I found that use of anthropogenic resources allowed subadult females to obtain the necessary extra energy required to survive winter. The actual winter survival estimate for subadult urban female opossums was 0.697. This survival rate was high enough to contribute to a growing urban population, because both reproduction and juvenile pre-winter survival were higher than expected. “Rural” animals, which did not use anthropogenic resources, did not persist through the winter. Many individuals shifted to using urban resources before the onset of winter, and the remainder died, primarily of predation in autumn or starvation in winter. Though the demographics for the total sample population indicate a declining population, the opossums in Massachusetts actually could be in a stable “source-sink” dynamic, with populations in urbanized areas growing and supporting declining populations in rural areas. By exploiting anthropogenic resources, Virginia opossums already occur in northeastern North America where they otherwise would not persist, and further northward expansion through an increasingly urbanized landscape is thus expected.

Survival and movement are fundamentally important processes that structure vertebrate populations. For mobile populations, mortality and emigration represent the only two ways that individuals are lost from a population and because each acts in concert with the other, it is neither possible nor practical to estimate or consider one with out estimating or considering the other. When it comes to Atlantic salmon, growth is a third factor that is inextricably linked to the fate of individuals. The individual-based approach used in this work facilitated disentangling these three processes for multiple Atlantic salmon cohorts in two streams in New England. Advancements of this work include: (1) development and testing of stationary passive integrated transponder (PIT) tag detection techniques for tracking movement of stream fish; (2) new methods to incorporate emigration information into capture-mark-recapture models to assist in decoupling emigration from true mortality; (3) elucidating the effect of local movement on estimates of true survival; and (4) examining the effects of size on growth, survival and movement over multiple seasons.

Non-imitative vocal development strategies, such as improvisation and invention, have been poorly documented compared to imitation, but evidence is mounting that they play an important role in the vocal development of some species, and at least a partial role in the vocal development of many other species. Currently the ecology and ontogeny of these alternative vocal development strategies are poorly understood and largely unstudied in all but a handful of species. In Chapter 1, I define improvisation and invention based on what we currently know and discuss some of the underlying developmental steps. I also suggest some potential areas for future research. In Chapter 2 I test the hypothesis that the predominant vocal development strategy is correlated with seasonal movement patterns by comparing song development between migratory Massachusetts and sedentary California red-winged blackbirds (Agelaius phoeniceus). The results of this study support other studies which have found a correlation between vocal development strategy and seasonal movement patterns. Migratory Massachusetts red-winged blackbird males improvise many more songs in their repertoires, and imitate much less from the tutor models, than do the sedentary California males. The songs improvised in the laboratory largely overlapped in structure with songs from wild males. In Chapter 3 I further test the hypothesis that laboratory-developed improvised songs function like normal conspecific songs and are not abnormal byproducts of an artificial learning environment. Using a paired playback design, I compare the response of wild males to playback of both these improvised songs and wild-recorded songs from other populations. The results of these playbacks indicate that, for the most part, wild males regard these laboratory-improvised song types as normal conspecific song. In Chapter 4 I study the ontogeny of improvised songs under social learning conditions and confirm the ability to improvise in migratory Massachusetts Common Grackles (Quiscalus quiscula) and possibly in more sedentary individuals from Florida. Male grackles from Massachusetts all produced better copies of the songs of the adult tutors earlier in development than they did later in development. Common Grackle males appear to be diverging from both the adult tutors and each other so that each male's song occupies a unique acoustic space.

<p> Apex predators are important indicators of ecosystem health, but little is known about the population structure of Bigg&rsquo;s killer whales (<i> Orcinus orca</i>; i.e. &lsquo;transient&rsquo; ecotype) in western Alaska. Currently, all Bigg&rsquo;s killer whales in western Alaska are ascribed to a single broad stock for management under the US Marine Mammal Protection Act. However, recent nuclear microsatellite and mitochondrial DNA analyses indicate that this stock is likely comprised of genetically distinct sub-populations. In accordance with what is known about killer whale vocal dialects in other locations, I sought to evaluate Bigg&rsquo;s killer whale population structure by examining the spatial distribution of group-specific call types in western Alaska. Digital audio recordings were collected from 33 encounters with Bigg&rsquo;s killer whales throughout the Aleutian and Pribilof Islands in the summers of 2001-2007 and 2009-2010. Recorded calls were perceptually classified into discrete types and then quantitatively described using 12 structural and time-frequency measures. Resulting call categories were objectively validated using a random forest approach. A total of 36 call types and subtypes were identified across the entire study area, and regional patterns of call type usage revealed three distinct dialects, each of which corresponding to proposed genetic delineations. I suggest that at least three acoustically and genetically distinct subpopulations are present in western Alaska, and put forth an initial catalog for this area describing the regional vocal repertoires of Bigg&rsquo;s killer whale call types.</p>

<p>Coyote distribution and habitat use, diet and foraging behavior, and space use patterns were investigated in the greater Detroit area of southeastern Michigan. We found evidence of coyotes on 24 of 30 (80%) suburban and 7 of 11 (64%) urban plots. Overall fifty-eight percent of coyote evidence was found within edge habitats, with den sites and tracks the only types of evidence found strictly in interior habitats. Land cover around evidence points included more wooded land cover than expected in suburban areas, suggesting the importance of tree cover for coyote occupancy, and more open space and wooded land cover than expected in urban areas, highlighting their avoidance of heavily populated areas. Coyote diet was assessed through identification of remains of food items recovered in coyote scat. White-tailed deer, eastern cottontail rabbit, and small rodents were the most consumed prey in both urban and suburban areas. Coyote consumption of white-tailed deer biomass was 7.2% greater than expected in suburban areas and 10.0% less than expected in urban areas and the difference was significant (<i>P</i> &lt; 0.004). More white-tailed deer, raccoon, and woodchuck biomass was consumed compared to other studies, likely due to high use of road-kill. In suburban areas, coyote selection for road-killed white-tailed deer was positive regardless of white-tailed deer or rabbit abundance. Coyotes in urban areas used a foraging strategy that incorporated both prey selection and switching, with no strong discernable pattern. Radio-telemetry technology was used to gather relocations of coyotes for analysis of home range and cores areas frequented by coyotes. Smaller home ranges were made up of greater proportions of urban land than natural land cover, although there was variation. Core areas were dominated by relatively large patches of natural land cover and had greater connectivity compared to home range areas. Radio-telemetry data suggested that coyotes were selective in their use of space, avoiding urban land in favor of natural land cover. </p>