The skin of many fishes contains large epidermal club cells (ECCs) that are known to release chemicals (alarm cues) that warn other fishes of danger. Initial research on ECCs focussed on their role in predator avoidance behaviour, however later research revealed that these cells might also have immune functions. Anthropogenic activities have dramatically increased over the past decades, with the consequence that many organisms simultaneously get exposed to multiple environmental stressors. We have seen considerable reductions in stratospheric ozone with a concomitant increase in global ultraviolet radiation (UVR). Metal pollution associated with industrial activity is also increasing on a global scale. Cadmium (Cd) is one such ubiquitous pollutant which is known to be toxic to organisms at extremely low concentrations. The main goal of my PhD research was to understand how multiple environmental stressors play a role in altering ECC investment and chemically-mediated predator-prey interactions by indirectly elucidating the evolutionary role of ECCs. The first experiment investigated the effects of in vivo ultraviolet radiation (UVR) exposure on ECC investment, physiological stress responses and potency of alarm cues in fathead minnows (Pimephales promelas). Subsequently, I investigated the interactive effects of UVR and/or waterborne cadmium (Cd) exposure using the same end points. I found that minnows exposed to UVR, either in the presence or absence of Cd, showed consistent decrease in ECC investment compared to non-exposed controls. There was a significant increase in cortisol levels of UVR exposed minnows compared to unexposed minnows. However, the combined exposure of UVR and Cd reduced cortisol levels relative to that in UVR only exposure. Surprisingly, there was no difference in the potency of the cues prepared from the skin of UVR and/or Cd exposed or non-exposed fish indicating that UVR and/or Cd exposure combined may have little influence on chemically-mediated predator-prey interactions. In aquatic systems, much of the negative effects of UVR are minimized by dissolved organic carbon (DOC) which is known to attenuate rates of UVR across the water column. In my third study, I investigated if DOC played a role in ameliorating the effects of in vivo UVR exposure on physiological stress and ECC investment in fathead minnows. I used two sources of DOC, a commercial soil based DOC (Sigma Aldrich Humic Acid) and a terrigenous source of DOC (Luther Marsh Natural Organic Matter). I found that fish exposed to UVR, in the presence of either source of DOC, in the presence and absence of UV blocking filter, maintained high ECC investment and reduced cortisol levels compared to fish exposed to UVR only. Studies that have examined factors that influence ECC investment have often been hampered by large variation in baseline levels of ECC. The larger the baseline variation in ECC number, the more difficult it is to elucidate factors responsible for changes in ECC investment. While I did not find this problematic in my work with UVR and Cd, others have failed to find effects in manipulative experiments. Consequently, my fourth study examined between and within variation in ECC investment across multiple sites in Saskatchewan and tried to investigate if holding fish under controlled laboratory conditions for up to 28 days would help reduce variation in ECC investment between and within populations. I found some evidence that I could reduce within population variation in ECC investment through time, but could not reduce among-population variation in mean ECC investment.

Across the animal kingdom, predation is a ubiquitous and omnipresent selective agent for a variety of traits. I aimed to address gaps in our knowledge pertaining to how predation shapes animal behavior. Many species of fish naturally occur in drainages that differ in the density of predators and exhibit obvious population divergence, making them ideal study organisms to investigate predator-driven behavioral evolution. In Chapter 1, I conducted a systematic review of the literature. The purpose of this review was to determine if predation acted as a stronger or weaker selective agent on particular behavioral traits (e.g., foraging, mating, antipredator etc.) across fish. This review showed that predation does not always drive behavior in predictable ways, and that some behavioral traits more consistently diverge than others. It was evident that antipredator behaviors are extremely variable but were typically measured in response to a visual stimulus. Investigations on intraspecific variation pertaining to how fish acquire, process, and respond to information across other sensory modalities are needed. To address this, I focused on a Neotropical fish, Brachyrhaphis rhabdophora (Poeciliidae), from Costa Rica that occur in distinct predation environments. For Chapter 2, I evaluated whether males and females exhibit differential responses to conspecific chemical alarm cues. Chemical alarm cues are released when a prey is injured by a predator and are an honest indicator of risk. It was clear that B. rhabdophora responded to alarm cues, but that males and females sometimes employed different antipredator strategies depending on what predation environment they were from. However, we know that in group-living species, such as B. rhabdophora, risk information can also be acquired indirectly through social cues. There are tradeoffs associated with relying on direct vs. indirect information, and these sources of information may sometimes conflict. For Chapter 3 I considered how B. rhabdophora integrates conflicting information to elicit antipredator behavior. I again exposed fish directly to chemical alarm cues and measured how their antipredator responses changed when visually observing conflicting or reinforcing social information. I found that individuals integrated personal and social information differently based on their evolutionary history with predators. Further, we found evidence that even a single observer fish is able to influence group behavior. Finally, for Chapter 4, I evaluated sex-specific variation in brain size across predation environments. According to the "expensive-tissue hypothesis" there should only be investment in brain tissue when there is sufficient selection for enhanced cognitive abilities. Prey under elevated selection from predators should invest more in cognitive traits to enhance survival, but how sex interplays with this effect is unclear. I found that females had higher relative total brain volumes than males, but males exhibited more variation across predation environments in the relative volumes for certain brain regions. This work as a whole suggests that, yes, evolutionary history matters for a variety of sensory-related traits in B. rhabdophora.