Behavioural and ecological implications of ambient acidification on the chemosensory alarm functions in juvenile salmonids

Alarm cues play important roles allowing prey individuals to act with context appropriate responses, increasing their chance to survive predators. One such type of alarm cue is chemical in nature and typically released following mechanical damage to the skin as would occur during a predation event. These damage-released chemical alarm cues are known to mediate local predation risks in many freshwater fish species. Under weakly acidic conditions (pH {598} 6.0) however, individuals exposed to chemical alarm cues have been shown to be impaired in their ability to detect these cues and respond with species-typical alarm behaviour. However, this effect has been demonstrated in only a single field study. In my first chapter, I conducted field observations in nursery streams ranging in pH from 5.71 to 7.49 on two year-classes (young-of-the-year and parr) of wild juvenile Atlantic salmon ( Salmo salar ). I assessed if the detection of these chemical alarm cues was dependent on the ambient pH or if variations in the detection occurred between populations regardless of the acidity levels. Salmon present in any acidified streams did not respond to alarm cues while those in neutral streams exhibited species-typical alarm responses. Secondly, I conducted experiments to further assess whether population or environmental differences was most likely to explain the loss of response to chemical alarm cues observed under acidic conditions. Thus, I conducted a reverse transplant experiment between salmon populations found under neutral (pH range {598} 7.0-7.3) and weakly acidic streams (pH range {598} 5.8-63). I later assessed if five different populations of salmon produced chemical alarm cues enabling consistent antipredator behaviour in a receiving population. My results showed that population differences did not account for the observed difference in alarm response, where ambient acidity created a behavioural impediment to normal chemical alarm function. I then determined at which pH value the loss of alarm function occurs. Using juvenile rainbow trout ( Oncorhynchus mykiss ) in a laboratory study, my results showed that between 6.4 and 6.2 (pH unit), a steep decrease in alarm behaviour occurred despite the introduction of chemical alarm cues suggesting a graded loss of response with increasing acidity. In my fourth chapter, I wanted determine if the learning of a novel odour could occur when paired with chemical alarm cues, under both neutral and acidic conditions. I showed that although it was possible to condition salmon under neutral conditions, no conditioning occurred under acidic conditions. I finally determined under laboratory conditions if a survival cost (increased mortality) from predation exists for juvenile rainbow trout exposed to acidified or neutral alarm cues in the presence of a predatory largemouth bass ( Micropterus salmoides ). Trout exposed to acidified alarm cues had a significantly shorter survival time when compared to trout exposed to neutral alarm cues. Altogether, these results suggest that even subtle chemical changes in ambient acidity may interfere with the use of chemical alarm signal in otherwise pristine conditions