The pharmaceutical pollutant fluoxetine alters reproductive behaviour in a fish independent of predation risk.

Pharmaceutical pollutants constitute a major threat to wildlife because of their capacity to induce biological effects at low doses. One such pollutant is the antidepressant fluoxetine, which has been detected in surface waters globally at levels that recent studies suggest can alter physiology and behaviour in aquatic organisms. However, wildlife exposed to pharmaceutical contaminants are typically confronted with multiple stressors simultaneously, including predation risk, which is a particularly important natural stressor that can have direct (e.g. mortality) and indirect (e.g. changed prey behaviour) fitness effects. Accordingly, we investigated potential impacts of environmentally realistic fluoxetine exposure on reproductive behaviour in the guppy (Poecilia reticulata) under predation risk. Specifically, we tested whether fluoxetine exposure altered mating behaviour in male and female guppies in the presence of either a predatory spangled perch (Leiopotherapon unicolor) or a non-predatory rainbowfish (Melanotaenia splendida) control. We found that fluoxetine and the presence of a predatory spangled perch did not interact to affect reproductive behaviour. We also found that, independent of a predatory threat, fluoxetine exposure altered male mating strategy, with males in the high treatment conducting significantly more coercive 'sneak' copulations, whereas the number of courtship displays performed was not significantly affected. Moreover, while fluoxetine exposure did not significantly affect the amount of time that males and females spent following one another, we found that females, but not males, followed a potential partner less when in the presence of the predatory fish. Finally, both sexes reacted to the risk of predation by spending less time in close proximity to a predator than a non-predator. In combination, our findings highlight the capacity of fluoxetine to influence processes of sexual selection at field-realistic concentrations and emphasise the importance of considering multiple stressors when assessing impacts of pharmaceutical pollutants on the behaviour of wildlife.