The Impact of Differential Cue Reliability on Behavioral Predictability Within and Across Generations

Abstract

How does an animal determine where to forage, how to find mates, and when to be vigilant for predators? Cues provide information about the environment, enabling animals to make decisions and can result in the evolution of predictable, adaptive behaviors. However, cues are not always reliable, and it is difficult to predict how an animal will respond when a cue may provide inconsistent, incorrect, or novel information. Understanding the relationship between cue reliability and behavioral predictability is of increasing concern, as global change is expected to reduce cue reliability and introduce novel cues to naïve populations. Using Trinidadian guppy (Poecilia reticulata) and threespine stickleback (Gasterosteus aculeatus) fish, my dissertation examines how changes in cue reliability impact animal behavior, informing our ability to predict behavioral decisions. It is expected that multiple sensory cues about the same ecological stressor should provide greater reliability than cues found in isolation. However, these cues often vary across space and time, suggesting that individuals may receive limited or conflicting information that can be passed down to offspring. For my first chapter, I tested whether different sensory cues of an evolutionarily known predator play a role in phenotype expression within and across generations in Trinidadian guppies. I found that males in the parent generation that did not to emerge into a novel environment (more risk-averse) exhibited stronger changes in activity following exposure to any predator cue compared to parents that emerged. In offspring, I found changes in both morphology and behavior. Offspring of parents exposed to olfactory cues alone were smaller in length compared to offspring of other treatments, while risk-averse males of parents exposed to visual or combined cues showed increased levels of activity. My findings suggest that a complex relationship exists between cue reliability, sensory cue-type, sex, and risk aversion, and that this relationship has consequences for subsequent generations. In response to altered home ranges and species introductions, there is an increasing likelihood of prey populations becoming exposed to novel predators. As a result, there is a need to determine whether animals can use their evolutionary experience with reliable predator cues to respond appropriately to novel predators. The predator similarity hypothesis predicts that if novel predators possess similar cues to known predators, it is expected that prey should exhibit similar anti-predator behaviors because these cues have been reliable in the past. By contrast, if novel predators possess cues are different from known cues, prey should exhibit either no response or an ineffective response due to a lack of threat association. For my second chapter, I tested the predatory similarity hypothesis using a gradient of visual predator similarity to a known predator. I exposed parenting threespine stickleback males to control (stick), known (trout), similar (northern pike), or novel (rubber duck) intruder models and measured changes in paternal care. I found that parents exhibited a generalized reduction in fanning to any territory intrusion, regardless of the intruder’s visual similarity to a known predator. Instead, males fell into two response groups: those that responded strongly to a territory intrusion and those that did not. My results suggest that predator similarity is not directly correlated with behavioral change and that risk perception, rather than cue reliability, at the individual level may be a stronger predictor of responses during an invasive species encounter. If theory predicts that prey will exhibit similar anti-predator behavior when novel cues resemble known cues, then this suggests that the mechanisms responsible for behavioral outcomes (respond/don’t respond) are the same. My third chapter expands on the predator similarity hypothesis by investigating the neural mechanisms driving predator identification and anti-predator behavior in threespine stickleback. Here, I exposed territorial stickleback to either a known (trout), similar (pike), or novel (rubber duck) model and recorded changes in movement behavior. Following assays, I collected brain tissue and examined neural activity using immunohistochemistry and RNA-seq methods. Like my second chapter, stickleback exhibited a generalized reduction in movement behavior that was not dependent on treatment. I observed no differences in neural activity in the telencephalon or diencephalon but found differences in RNA expression between trout and pike exposed fish. These results provide additional support suggesting that stickleback exhibit a generalized anti-predator response regardless of cue reliability but that subtle within-modality differences in visual trait processing may be categorized in different ways. Together, this suggests that different response pathways may exist at the molecular level, providing a potential avenue for behavioral change in the presence of novel invasive predators. Finally, cue reliability studies often occur in laboratory conditions where few cues are introduced and manipulated, but whether rapid shifts in reliability impact organisms in natural conditions are difficult to assess. Aquatic habitat fragmentation offers a unique ecological scenario where populations experience periods of isolation that can impact both biotic and abiotic cues. Utilizing west-coast drought conditions in northern California, my last chapter examines how shifting cue reliability following fragmentation events alter aggression and parental care in threepine stickleback, as well as subsequent risk-taking behavior in offspring. Following the identification of parenting males in connected and fragmented (pooled) areas, I collected parenting data and exposed males to a control, a familiar conspecific intruder (neighbor), and an unfamiliar conspecific intruder (stranger) to collect aggression data. I then measured risk-taking behavior in offspring in each community type using a scototaxis assay. In parents, I found that males in connected sites tailored their aggression based on male identity. However, males found in pools exhibited the same levels of aggression towards both neighbors and strangers and exhibited a general reduction in parenting behavior. Additionally, offspring found in pools exhibited a reduction in risk-taking behavior. Together, my findings indicate that shifts in cue reliability impact behavioral phenotypes in field conditions. My dissertation highlights the importance of empirically testing established theory in decision-making processes within and across generations. As human induced rapid environmental change is expected to persist, my work identifies potential factors driving behavioral outcomes, which will inform our ability to broadly predict population-level behavior and persistence.