Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier



Open Access Dissertation

Document Type


Degree Name

Doctor of Philosophy (PhD)

Degree Program

Organismic and Evolutionary Biology

Year Degree Awarded


Month Degree Awarded


First Advisor

Jeff Podos

Subject Categories

Ecology and Evolutionary Biology | Life Sciences


In many animal species, communication can enable individuals to resolve conflict without the high potential costs involved in direct fighting. During contests, animals may exchange information about their aggressive motivational state. A central question throughout the study of animal communication research has been whether animal signals convey reliable information, and this question has been particularly relevant to communication during conflicts where the evolutionary interests of competitors directly oppose. Deceptive signaling of aggressive motivation would be highly favored by natural selection because it could allow individuals to gain access to resources they might not gain through direct combat. However, selection should also favor signal recipients to respond only to informative signals, and therefore only reliable signals are expected to be maintained over evolutionary time. Due to methodological limitations, there has been a lack of empirical research that has appropriately assessed the reliability of aggressive signals. In my dissertation, I seek to identify animal signals with potentially aggressive content, and to assess whether such signals do indeed reliably convey information about aggressive motivation. To do this, I combined both observational and experimental approaches focusing on vocal signaling in a clade of songbirds, the New World warblers, and in particular the black-throated blue warbler (Setophaga caerulescens). First, I asked how black-throated blue warblers use their vocal repertoires across different social contexts, and especially across contexts that vary in intensity of conflict, with the goal of identifying potentially important vocal signal features involved in aggressive escalation. I found that black-throated blue warbler songs fall into two acoustically distinct categories of song types, called type I and type II songs. The use of type II songs relative to type I songs increased with increasing intensity of agonistic interactions, especially during early stages of aggressive escalation. I also found that low-amplitude versions of songs (soft songs) were strongly associated with close range vocal interactions and were frequently produced just prior to fights. These results suggest that soft songs might convey information about a highly aggressive state, and that type II songs play a role in aggressive escalation. Next, I applied a recently developed experimental approach to ask whether signal features reliably predict subsequent aggressive behavior such as an attack by the signal sender. I conducted experimental trials in which subjects were provoked by playback of rivals’ signals, their vocal responses were documented, and then they were presented with a taxidermic model that could be attacked. I found that the use of soft song was an extremely reliable predictor of whether birds would subsequently attack the model. I then extended this approach to ask whether sequences of signals used during aggressive escalation might convey increasing levels of aggressive motivation. I simulated interactions that gradually increased in intensity by presenting subjects with two sequential and increasing levels of threat. I found that the use of type II song in response to the low threat level predicted later use of soft song in response to the high threat level, and that soft song, in turn, predicted attack of the model. These results provide evidence that animal signals not only reliably convey motivation to attack, but can also convey motivation to escalate to more intense stages of signaling. Next, I asked how generally these patterns of aggressive signaling might apply to other warbler species, and how aggressive signals might vary across species in this clade. I conducted experimental trials to identify signals that reliably predict future attack in two additional warbler species, the ovenbird (Seiurus aurocapillus), and the American redstart (Setophaga ruticilla), and compared findings to those obtained for black-throated blue warblers. I found that in all three species, the use of soft song reliably predicted future attack. Additionally, in ovenbirds, use of a non-song call, and in American redstarts the use of shortened songs and postural displays, were also reliable predictors of attack. These findings show an underlying commonality in reliable aggressive signaling among the three species, as well as species-specific diversity, and provide insights into processes of aggressive signal evolution. Finally, I conclude my dissertation by providing evidence that when aggressive interactions in songbirds do escalate to combat, such fighting can be costly and potentially fatal. I describe an observation where a male black-capped chickadee (Poecile atricapillus) appeared to kill another male during a competitive interaction. The two opponents were both high-ranking males, consistent with the prediction that mortal combat is most likely to occur between evenly matched individuals.