Extended exposure to elevated temperature affects escape response behaviour in coral reef fishes

Introduction

Predator–prey interactions are one of the key drivers structuring communities (Lima & Dill, 1990). Predation pressure can alter prey morphology, regulate population size, and control biodiversity (Sih et al., 1985). Similarly, antipredator responses of prey reduce the efficacy of predators and enhance survival. Factors that influence predator attack success or prey escape performance may alter the outcome of these interactions, potentially causing a cascade of trophic effects. Increased temperature has already been shown to affect antipredator responses in terrestrial ectotherms (Cury de Barros et al., 2010), freshwater fishes (Weetman, Atkinson & Chubb, 1998), and marine species (Allan et al., 2015). Understanding how changes in environmental temperature can affect key ecological processes can be useful when predicting the composition of future communities.

Rising environmental temperatures can substantially impact the behaviour and locomotor performance components of escape responses in fish (Domenici, 2010). Previous studies have shown that temperature can influence locomotor performance through physiological mechanisms like muscle performance (Johnston, Fleming & Crockford, 1990) and enzyme activity (Johnson & Bennett, 1995). Additionally, temperature can affect behavioural components of escape responses, including responsiveness (Preuss & Faber, 2003) and directionality (Szabo et al., 2008). The effect of environmental temperature will depend on both the level of thermal change compared to the normal thermal conditions, and the length of time thermal change is experienced (Munday, McCormick & Nilsson, 2012). With sufficient exposure time, plasticity can occur resulting in changes to the phenotype, which potentially reduce the negative impacts of rising temperature (Angilletta, 2009). Exposure during the early stages of life is believed to unlock greater plasticity in traits (West-Eberhard, 2003), and this has been found to be true for coral reef fishes (Donelson et al., 2011). While many studies have investigated the effects of short-term exposure to elevated temperatures in adult fishes, to date few studies have examined how juvenile fishes respond to long-term exposure to near-future temperature increases.

The interaction between a predator and its prey is an intricate sequence that, at its end, leads to a predator striking at a prey and the prey being captured or escaping (Lima & Dill, 1990). Many amphibians and fishes undertake a C-start escape response to a startling stimulus (Bullock, 1984) involving a sudden acceleration away from the direction of the perceived threat (Domenici & Blake, 1997). C-start responses are controlled by two large neurons called Mauthner cells, one for each lateral half of the body (Eaton, Bombardieri & Meyer, 1977). When a stimulus is detected, the corresponding Mauthner cell triggers the escape response resulting in a rapid contraction of white muscle fibres. At first, the fish bends into a C-shape (stage 1), then contralateral contraction releases this tension which propels the fish to flee the stimulus in the opposite direction (stage 2) (Eaton, Bombardieri & Meyer, 1977). The initiation of C-starts are neurally controlled with maximum performance believed to be limited by muscle power output (Wakeling & Johnston, 1998). However, fish rarely perform at their maximal limits (Webb, 1986) and modulation of performance during an escape response has been shown to occur in different behavioural contexts, showing there is potential for optimisation (Korn & Faber, 2005; Domenici, 2010; Ramasamy, Allan & McCormick, 2015). This supports the economic model that individuals balance the energetic cost of escaping with the threat of predation (Ydenberc & Dill, 1986). How long-term exposure to elevated temperature can alter behavioural components of an escape response is still unknown.

The present study examines whether short-term or long-term exposure to elevated temperature regimes affect the fast-start escape response in two tropical damselfishes. Using a repeatable drop stimulus (i.e., a stimulus that was consistently replicated), we compared escape responses of juveniles from current-day controls (29 °C) to individuals exposed to 50 and 100 year projections (30 and 31 °C; (Collins et al., 2013) for either 4d or 90d. Specifically, we tested: (1) the effects of short-term exposure to elevated temperature as an indication of thermal sensitivity and (2) how short-term exposure differed to long-term exposure at elevated temperature as an indication of plasticity. The inclusion of two species was to explore the species-specific effect that temperature had on escape responses.

Methods

Results

There was no significance in the logistic regression model for responsiveness in P. amboinensis (χ²(4) = 4.18, p = 0.38; Fig. 1A). At control, 84.62% of individuals responded with a C-start, 7.69% showed no reaction, and 7.69% displayed avoidance behaviour. Exposure of four days to either 30 or 31 °C had very little effect on responsiveness. However, for individuals that did not display a C-start, all showed no reaction to the stimulus and none displayed avoidance behaviour. After 90 days, fish were 2.19 times less likely (71.43%) to perform a C-start at 30 °C and 3.14 times less likely (63.64%) at 31 °C. Additionally, avoidance behaviour was returned (14.29%) and divided non C-start individuals in half with those showing no reaction (14.29%). There was significance in the regression model for directionality in P. amboinensis (χ²(4) = 11.23, p = 0.024; Fig. 1B). The model explained 28.70% (Nagelkerke R²) of the variance in directionality of C-starts and correctly classified 67.60% of cases. Compared to controls, fish exposed to 30 °C for 4 days were 5 times more likely (33.33%) to turn towards the stimulus and 6.67 times more likely (40%) after 90 days. At 31 °C, fish were 25 times more likely (71.43%) to turn towards the stimulus, which was the same for both exposure durations. Temperature nor exposure duration had any significant effect on response latency (F_4,42 = 1.32, p = 0.27, partial η² = 0.053; Fig. 2A), maximum swim speed (F_4,41 = 0.6, p = 0.66 partial η² = 0.056; Fig. 2B), or response distance (F_4,42 = 0.65, p = 0.62, partial η² = 0.058; Fig. 2C) for P. amboinensis.

For P. moluccensis, logistic regression model for responsiveness was statistically significant (χ²(4) = 16.58, p = 0.002; Fig. 1A). The model explained 28% (Nagelkerke R²) of the variance in proportion of individuals that elicited a C-start and correctly classified 80% of cases. At control, 82.61% of fish displayed a C-start. Of the fish that did not C-start, most showed avoidance behaviour (13.04%) and no reaction (4.35%) least of all. Responsiveness rose to absolute (i.e., 100% of individuals performed a C-start) after four days of exposure to either 30 or 31 °C. When 90 day exposure was compared to controls, individuals were 1.45 times less likely (76.47%) to perform a C-start at 30 °C and 3.55 times less likely (57.14%) at 31 °C. For individuals that did not display a C-start, there was an increase in no reaction up to 38% at 31 °C (from 4.35% at control). There was no significance in the regression model for directionality (χ²(4) = 5.98, p = 0.20; Fig. 1B). Fish with four days of exposure to 30 °C were 2.8 times more likely (50%) to turn towards the stimulus compared to controls and 4.2 time more likely (60%) at 31 °C. However, this was reduced after prolonged exposure where directionality was very similar to control (23.08% and 25% for 30 and 31 °C, respectively). There was no significant effect of temperature and exposure duration on response latency (F_4,60 = 1.23, p = 0.30, partial η² = 0.052; Fig. 2A), maximum swim speed (F_4,60 = 0.19, p = 0.94, partial η² = 0.013; Fig. 2B), nor response distance (F_4,60 = 0.85, p = 0.49, partial η² = 0.054; Fig. 2C) for P. moluccensis.

Discussion

Studying how escape responses are affected by elevated temperature can be important when making predictions on the ability of prey to escape predators under climate change conditions (Gilman et al., 2010). Supporting our predictions, we found short-term exposure to elevated temperature negatively impacted directionality for both species. However, there was little effect on responsiveness, response latency, maximum, swim speed, and escape distance. After long-term exposure, directionality in P. moluccensis followed our predictions by reducing the number of turns towards the stimulus and returning to control levels. For P. moluccensis, long-term exposure also altered responsiveness with fish responding less often with a C-start response at elevated temperatures. Such alterations in the C-start responsiveness to predators is likely to change prey survival, while the differential effects of temperature on the two closely related species suggests that future temperature elevation may affect community composition through differential survival.

The ability for prey to perceive and avoid a predatory threat is essential to survival. Directionality in escape responses is important as turning towards a stimulus would increase exposure time to a predator (Domenici, 2010). Short-term exposure to elevated water temperature negatively impacted directionality by increasing number of turns towards the stimulus. Previous study on goldfish also found directionality was impaired at increased temperatures compared to controls (Szabo et al., 2008). Altered directionality may have been due to an impairment of sensorimotor control (Foreman & Eaton, 1993). While not significant, short-term exposure to elevated temperatures increased C-start responsiveness in P. moluccensis from 80% to 100%, and reduced avoidance reactions to 0% for both species. Generally, proportion of C-starts has been shown to increase with exposure to temperature both above (Szabo et al., 2008) and below control treatments (Preuss & Faber, 2003), suggesting responsiveness may increase with thermal stress, at either extreme. Hyperactivity under stress can be beneficial to survival, as a fish that does not react during a predator attack will most likely get eaten. However, this strategy can risk expending additional energy on a false or non-threatening stimulus.

Extended exposure to new environmental conditions can alter the physiological and behavioural response of individuals (Angilletta, 2009). We found directionality in P. moluccensis showed signs of beneficial plasticity after long-term exposure by returning to control levels. Long-term exposure may have reduced the negative impact of sensorimotor control brought on by thermal stress. Physiological plasticity to similar temperatures (+2 °C) in this species has also been shown to occur for aerobic metabolism (Grenchik, Donelson & Munday, 2013) potentially demonstrating that P. moluccensis could exhibit beneficial plasticity to future projected temperatures in a range of traits. The proportion of C-starts reduced while avoidance behaviour increased with long-term exposure in both species. Decreasing the proportion of C-starts deviates further from control levels, therefore cannot be considered beneficial. However, future sea temperature increases are expected to exceed the thermal optimum for many reef fish (Tewksbury, Huey & Deutsch, 2008), forcing species to adopt energy saving strategies (Ydenberc & Dill, 1986). Reducing the proportion of C-starts and replacing with avoidance behaviour may help compensate for the costs of increasing thermal conditions.

While other studies have shown that elevated temperature affects the locomotor performance in fast-starts (Webb & Zhang, 1994; Beddow, Leeuwen & Johnston, 1995), we found projected increased sea temperatures had no effect on maximum swim speed or escape distances in our species, regardless of exposure duration. Similarly, we found no effect on response latency with temperature and exposure duration. The lack of change in performance at these temperatures indicates once a C-start was initiated, the physical ability to escape is maintained. This suggests future warming may have little effect on mortality rates. Although promising, this may be problematic because metabolic demands for fish increase with rising ambient temperature (Nilsson, Östlund-Nilsson & Munday, 2010; Grenchik, Donelson & Munday, 2013) and C-starts are high energy manoeuvres (Jayne & Lauder, 1993). If an individual’s energy budget is limited, then more resources must be allocated to the rising cost of metabolic demands and potentially leave less available for other activities such as escape behaviour (Careau et al., 2008). Thus, the behavioural modulation of reducing C-start responsiveness after long-term exposure observed, may act as a means of balancing the increasing cost of C-starts at higher temperature.

To conclude, both species showed negative effects of increased temperature after short-term, though the impact was greater for P. amboinensis. After long-term exposure, P. moluccensis adjusted to temperature with beneficial plasticity and energy saving strategies. These results show the C-start escape response possesses thermal plasticity and extended exposure can induce beneficial changes. The difference in response by species matches a previous study showing P. moluccensis possessed a higher capacity for thermal plasticity in aggressive interactions (Warren et al., 2016). Changes in escape behaviour will ultimately affect mortality rates in prey and the species-specific response suggests some will better cope with future temperatures. This study is important as it illustrates the specific contribution that elevated temperature will have on individual performance. However, comprehensive climate change scenarios will also include responses to changes in ocean chemistry (CO₂ levels and pH). Furthermore, precise measurement of escape success will require future study with only the prey, only the predator, and both exposed to elevated temperature as predators will also likely have specific responses to thermal conditions. Future studies may also consider a longitudinal approach with several populations from differing sites to consider adaptation to local thermal environments. Studies investigating the factors that influence predator–prey interactions, like this one, are crucial because these interactions determine population sizes and replenishment, leading to changes in community structure.

Supplemental Information