• Lawrence, Michael J.
  • Godin, Jean-Guy J.
  • Zolderdo, Aaron J.
  • Cooke, Steven J.


Stressed fish have been shown to have higher predator-induced mortality than unstressed conspecifics, suggesting a role for the hypothalamic–pituitary–interrenal axis in modifying risk-taking behaviors. Yet, there is also evidence of behavioral resiliency in the face of chronic stressors. Here, we tested the behavioral resiliency hypothesis, which posits that animals can maintain consistent behavioral phenotypes in the face of significant physiological challenges. We determined whether chronic plasma cortisol elevation promotes risk-taking behaviors in a model teleost fish, the pumpkinseed sunfish (Lepomis gibbosus). Experimental fish were implanted with cocoa butter either as a sham or with cortisol. At 48 h post-implantation, the behavior of individual focal fish was tested in an experimental arena comprising of a simulated physical refuge, an open zone containing a constrained conspecific shoal, and a compartment containing either a model of a northern pike (Esox lucius) paired with corresponding pike olfactory cues in lake water or no pike model (control) paired with sham lake water cues only. The fish were assayed individually for their refuge utilization, shoaling tendency, and general activity. Neither cortisol treatment nor predation-risk treatment influenced any of these behaviors. This suggests that sunfish, in the context of our experiment, were behaviorally resilient to the physiological effects of chronic plasma cortisol elevation and in the face of an apparent threat of predation. Our results thus provide support for the behavioral resiliency hypothesis in fish under both physiological and ecological stressors. We posit that behavioral resiliency is an evolutionary adaptation ensuring appropriate responses to environmental conditions.


Fish collection and implantation procedures

Juvenile pumpkinseed sunfish (mean ± SE mass = 8.7 ± 0.2 g; total length = 81.8 ± 0.5 mm; N = 125) were captured haphazardly using a seine net in the nearshore waters of shallow weedy bays in Lake Opinicon, Ontario, Canada (44°55′90″N, 76°32′80″W) during August 2017 (Ontario Ministry of Natural Resources permit #1086180). Seining was used as the primary collection method to ensure an unbiased sample of behavioral phenotypes in the population (Wilson et al. 2011; Gutowsky et al. 2017). Following capture, fish were immediately transported to the nearby Queen’s University Biological Station (QUBS; Chaffey’s Lock, Ontario, Canada) in a well-aerated cooler and were transferred to a large, indoor flow-through tank containing lake water (∼212 L; >90% O2 saturation, 23.7 ± 0.1°C), where they were held for 24 h prior to experimental manipulation. A subset (N = 40) of the fish captured were retained for use as stimulus conspecifics for the assessment of shoaling tendency in the behavioral experiment (see below) and were not implanted with cocoa butter. These fish were held in a separate tank (∼406 L) and were kept under similar holding conditions to the focal test fish. These stimulus fish were released into the lake upon completion of the study.

We captured eight Northern pike (Esox lucius Linnaeus; 549.6 ± 22.4 mm; range 490–650 mm), to generate predator (i.e., pike) olfactory cues used in our behavioral experiment (see below), using rod-and-reel angling techniques including trolling and bait casting (see Lawrence et al. [2018c] for more details). Upon capture, pike were transported quickly back to the QUBS and held in large outdoor tanks (∼940 L) with flow-through lake water. All pike were eventually live-released back into the lake following this study. Both pumpkinseed and pike were not fed at any time while in captivity. By the time of the onset of behavioral testing, all fish were fasted for a total of 72 h (Fig. 1). This was done to standardize individual hunger status, thereby preventing any potential confounding effects of hunger state in mediating risk-taking behaviors (e.g., Smith 1981; Dill and Fraser 1984; Gotceitas and Godin 1991). Our study conformed to the guidelines for the use and care of experimental animals of the Canadian Council on Animal Care and received prior approval of the Carleton University Animal Care Committee (AUPs #104262 and 104281).