|Abstract:||A large body of research has documented evolutionary change in fish populations as a result of selective harvest, a process known as Fisheries-Induced Evolution (FIE). Much of this research has focused on commercially-harvested marine populations, though recent work has also shown that FIE can also occur in freshwater systems targeted by recreational hook-and-line anglers. For FIE to occur as a result of recreational angling, it is necessary that particular traits (for instance, aspects of an individual fish’s behavior or physiology) are associated with an increased likelihood of capture by anglers, leading to selective harvest that causes directional evolution away from that trait. For researchers and managers to accurately predict the outcomes of FIE in freshwater, it is therefore imperative that selected traits are identified, particularly in heavily fished species. While some work has been done previously in this area, several behavioral and physiological characteristics that could be linked with angling vulnerability have yet to be fully explored.
In this dissertation, I present a series of experiments examining a set of behavioral and physiological traits and their role in driving angling vulnerability in fish. Each experiment utilizes one of two highly sought-after freshwater species, the largemouth bass Micropterus salmoides or the bluegill Lepomis macrochirus. In each experiment, a given set of characteristics is evaluated in individuals of the species in question, and paired with results from actual angling experiments conducted on those same individuals in a naturalistic pond setting.
In chapter 2, I examine the role of boldness, metabolic rate (standard metabolic rate, maximum metabolic rate, and aerobic scope) and stress responsiveness in driving angling vulnerability. A set of largemouth bass from a suite of lines selected for differing vulnerability to angling were assessed for boldness in a standard open-field test. Following this, individuals had blood samples taken before and after an air exposure challenge to assess both baseline and post-stress levels of the primary stress hormone, cortisol. All assessed fish were then stocked into a pond where angling took place. At the conclusion of angling trials, a subset of captured and uncaptured fish were assessed for metabolic rate using intermittent-flow respirometry. Results showed a highly significant association between stress responsiveness and angling vulnerability, specifically that individuals captured by anglers showed a smaller rise in cortisol levels after the air exposure challenge compared to uncaptured fish. Boldness and metabolic rate did not predict angling vulnerability. Because high stress responsiveness has been linked previously to a propensity to freeze in response to challenges (as well as other behavioral traits), selective capture in largemouth bass could lead to evolutionary pressure favoring passive and reactive behavior in exploited systems.
In chapter 3, I examined the role of metabolic phenotype in driving angling vulnerability in bluegill. Similar methods were used to assess metabolic rate as in chapter 2, with an additional examination of anaerobic capacity in the form of excess post-exercise oxygen consumption (EPOC). Fish were first angled, with a subset of fish assessed for metabolic phenotype afterwards. Results showed no difference in metrics of metabolic phenotype (standard and maximal metabolic rates, aerobic scope, EPOC, metabolic recovery time) between captured and uncaptured fish, indicating that, in bluegill, metabolic characteristics are likely not under selective pressure from angling.
Chapter 4 examined the relationships between individual sociability, aggression, and angling vulnerability in bluegill. For this chapter, bluegill were first subjected to angling, with a subset of captured and uncaptured fish then assessed for sociability and aggression in the laboratory. Assessment for sociability consisted of placing an individual bluegill in a large tank divided in half by a transparent barrier that separated the focal fish from a shoal of conspecifics. Sociability was defined as the time spent by the focal fish near the divider, associating with the conspecifics. Following this, focal fish were size-matched and assessed for aggression and dominance in dyadic trials. Results showed a significant effect of time spent near the divider on angling vulnerability, with captured bluegill being more social than uncaptured bluegill. Aggression was not a significant predictor of vulnerability, though a non-significant trend was found whereby captured fish tended to be less aggressive.
While chapter 4 examined bluegill sociability on an individual basis (i.e. each focal fish was examined in isolation), Chapter 5 sought to quantify sociability within the context of interactions within a group of individuals. In addition, swimming performance was assessed for the purpose of determining if this physiological trait was linked with either angling vulnerability or social behavior. For this, groups of 6 individuals were size-matched and placed into a common tank, where they were evaluated for sociability and aggression over three days of observation. Pooled behavior from all three days was then analyzed using methods derived from Social Network Analysis. Each fish was then assessed for swimming performance (critical swimming speed - Ucrit) in a Brett-style swim tunnel before being stocked into a pond for angling. The results showed that, while only fish size predicted whether or not a fish was captured (larger fish were more likely to be caught at least once), more social and less aggressive individuals were found to be the most vulnerable. Specifically, high sociability/low aggression predicted whether an individual was caught multiple times, and also predicted capture order with highly social individuals being captured first. Swimming performance did not predict any aspect of angling vulnerability. These results, combined with the results from chapter 4, indicate that social behavior is indeed a key determinant of angling vulnerability in bluegill, and that angling selection may evolutionarily favor fish that are both more aggressive and less social.
In chapter 6, I examined the role of learning performance and proactivity in driving angling vulnerability in largemouth bass. For this experiment, a set of largemouth bass was assessed for learning performance on an active-avoidance task. For this task, each fish was put into an individual tank that was divided in two by an opaque barrier. The barrier included a small opening for shuttling between sides of the tank. Over a set of trials, an observer first shined a light over the fish, which was followed by chasing with an aquarium net. When fish successfully shuttled to the other side of the tank in response to the light (but before the onset of chasing), this was considered successful learning. From there, each fish was assessed for proactivity in a restraint test, where fish were scored based on the number of attempts each fish made to leap from a container when held out of water. Following angling, it was found that learning performance was significantly linked with angling vulnerability, with high performing individuals being more likely to be captured. Within the framework of “cognitive syndromes”, this result indicates that individuals that learn tasks quickly and are, therefore often prone to mistakes, may be under selective pressure in angled populations of largemouth bass.
Collectively, this research has identified several behavioral and physiological characteristics that drive vulnerability to angling, however the characteristics differed between species. While largemouth bass vulnerability was driven by characteristics broadly related to proactive behavior (rapid learning, low stress responsiveness), for bluegill it was social and unaggressive individuals that were found to be the most vulnerable. Overall, this means that heavily fished populations could experience behavioral evolution as a result of selective capture on these traits, however the traits under selection may differ depending on the species.