All locomotion produces sound and flight is no exception. In owls, flight sounds are quieted by three wing and feather features: the leading-edge comb, a modified barb structure that projects dorsally from the front edge of the outermost primary feather (P10), the velvety dorsal surface of flight feathers, and the fringed vane of flight and tail feathers. There are two hypotheses for the evolution of quieting features: stealth and self-masking. Under the stealth hypothesis, we predict quiet flight evolved to aid owls in sneaking up on prey. Under the self-masking hypothesis, we predict quiet flight evolved to aid acoustic hunters in locating prey. To test these hypotheses, we investigated the relationship between leading-edge comb morphology and ecology (Chapter 1), tested the function of the dorsal velvet in Barn Owls (Chapter 2), and investigated the function of quieting features in Lesser Nighthawks, another nocturnal bird (Chapter 3). In Chapter 1, We used phylogenetic generalized least squares (pgls) to test the correlation between comb morphology and the stealth or self-masking scores. We found comb morphology to be correlated with both stealth and self-masking (pgls; DF = 66 test-statistic = -3.92; P-value = 0.0002). In Chapter 2, impairing the dorsal velvet of 10 feathers on 13 barn owls increased broadband sound production and the upstroke increased more than the downstroke, such that the upstroke of manipulated birds was louder than the downstroke, supporting the frictional noise hypothesis. Finally, In Chapter 3, we found Lesser Nighthawks initiated pursuit of prey at a greater distance than the audible detection distance of insects (0.5 m) both when hunting on the ground (1.1 ± 0.2 m; P-value = 0.02) and on the wing (2.5 ± 0.4 m; P-value <0.0001), suggesting they use visual cues to initiate pursuit of prey under these conditions. This dissertation contributes to a research area which has received little consideration: the role of sound and hearing in predation.

Here I examine vocal learning in the Costa’s hummingbird. I used the standard isolation experiment technique utilized for decades to study vocal learning in songbirds. Young male Costa’s hummingbirds were raised in isolation, with playback of adult song, and with playback in the presence of an adult model. All bird vocalizations were recorded to examine song ontogeny. I found that the necessary conditions required for vocal learning of a tutor song to take place in a lab setting included the presence of an adult model during tutoring and the tutor song must be Costa’s like. I developed a timeline of song ontogeny events including information about the onset of singing, and the stages of song development. Another experiment examined the timing of the sensitive phase by tutoring Costa’s for 20 hours in one of three different 30-day tutoring periods: Early (35-65 days post hatch (dph)), Mid (75-105 dph), and Late (115-145 dph). Results indicated that Costa’s can learn some vocal information across all three tutoring periods but that the amount of tutoring during one tutoring period was insufficient for production of normal Costa’s song. Finally, I examined open-ended vocal learning by exposing previously isolated birds to other Costa’s hummingbirds and found that songs change after exposure, even after two years of isolation.

ABSTRACT OF THE DISSERTATION

Courtship Behavior and Mechanisms of Isolation Across a Hummingbird Hybrid Zone

by

Brian Myers

Doctor of Philosophy, Graduate Program in Evolutionary Biology

University of California, Riverside and San Diego State University, September 2020

Dr. Kevin Burns and Dr. Christopher Clark, Co-Chairpersons

In my dissertation, I focused on the discovery and evolutionary maintenance of a hybrid zone between migratory Allen’s (Selasphorus sasin) and Rufous (Selasphorus rufus) Hummingbird centered in southern Oregon. Via incorporation of morphological, genomic, and behavioral data of 304 birds, my dissertation examines how natural selection, sexual selection, and biogeographic processes have affected the diversification and interactions of Allen’s and Rufous Hummingbird.

In Chapter 1, I used Linear Discriminant Function Analysis (LDA), cline analysis, and a hybrid index to reveal the geographic extent of the Allen’s and Rufous Hummingbird hybrid zone, classify individuals as parents or hybrids, and investigate how selection acts on a suite of behavioral and morphological characters. The contact zone extends north into the breeding range of Rufous, and south into the range of Allen’s. Using geographic cline analysis, I found a role of sexual selection in shaping species barriers.

In Chapter 2, I performed behavioral sequence analysis on Allen’s, Rufous, and their hybrids to assign discrete phenotypes based on the sequences of their courtship displays. For the first time in ethology, I implemented k-mer analysis to identify how many behavioral phenotypes are present across my sampled individuals. I also described novel and transgressive courtship behaviors across the contact zone and found that hybrids are more variable than parental species. At least within Allen’s and Rufous Hummingbird, the modular nature of displays shows that courtship might play a role in diversification between these two species, and possibly within Selasphorus.

In Chapter 3, using whole-genome data, I found evidence of gene flow from Rufous Hummingbird across the range of migratory Allen’s, evidence of intergradation between non-migratory and migratory Allen’s, and high gene flow across the entire non-migratory Allen’s, migratory Allen’s, and Rufous Hummingbird species complex. Further, I found evidence of selection on the Z chromosome in migratory Allen’s and Rufous Hummingbird and estimated divergence dates of Allen’s and Rufous to be further back in time than previously reported.