Using aerobic exercise to evaluate sub-lethal tolerance of acute warming in fishes
Data files
Mar 23, 2020 version files 69.33 KB
Abstract
We investigated whether fatigue from sustained aerobic swimming provides a sublethal endpoint to define tolerance of acute warming in fishes, as an alternative to loss of equilibrium (LOE) during a critical thermal maximum protocol (CTmax). Two species were studied, Nile tilapia Oreochromis niloticus and pacu Piaractus mesopotamicus. Each fish underwent an incremental swim test to determine gait transition speed (UGT), where it first engaged the unsteady anaerobic swimming mode that preceded fatigue. After suitable recovery each fish was swum at 85% of their own UGT and warmed 1°C every 30 min, to identify the temperature at which they fatigued, denoted as CTswim. Fish were also submitted to a standard CTmax, warming at the same rate as CTswim, under static conditions until LOE. All individuals fatigued in CTswim, at a mean temperature approximately 2°C lower than their CTmax. Therefore, if exposed to acute warming in the wild, the ability to perform aerobic metabolic work would be constrained at temperatures significantly below those that directly threatened survival. The collapse in performance at CTswim was preceded by a gait transition qualitatively indistinguishable from that during the incremental swim test. This suggests that fatigue in CTswim was linked to an inability to meet the tissue oxygen demands of exercise plus warming. This is consistent with the oxygen and capacity limited thermal tolerance (OCLTT) hypothesis, regarding the mechanism underlying tolerance of warming in fishes. Overall, fatigue at CTswim provides an ecologically relevant sub-lethal threshold that is more sensitive to extreme events than LOE at CTmax.
Methods
Data were collected by in-vivo experimentation on the two fish species
Usage notes
The data of individual fishes (for example oxygen uptake data) are not reported, these data are calculated values for each fish, organised to test with statistical models such as ANOVA or GLM