Choice of experimental venue matters in ecotoxicology studies: Comparison of a laboratory-based and an outdoor mesocosm experiment

Abstract

The heavy application of pesticides and its potential effects on natural communities has attracted increasing attention to inadvertent impacts of these chemicals. Toxicologists conventionally use laboratory-based tests to assess lethal concentrations of pesticides. However, these tests often do not take into account indirect, interactive and long-term effects, and tend to ignore different rates of disintegration in the laboratory and under natural conditions. Our aim was to investigate the importance of the experimental venue for ecotoxicology tests. We reared tadpoles of the agile frog (Rana dalmatina) in the laboratory and in outdoor mesocosms and exposed them to three initial concentrations of a glyphosate-based herbicide (0, 2 and 6.5mg a.e./l glyphosate), and to the presence or absence of caged predators (dragonfly larvae). The type of experimental venue had a large effect on the outcome: The herbicide was less lethal to tadpoles reared in outdoor mesocosms than in the laboratory. Further, while the herbicide had a negative effect on development time and on body mass in the laboratory, tadpoles exposed to the herbicide in mesocosms were larger at metamorphosis and developed faster in comparison to those reared in the absence of the herbicide. The effect of the herbicide on morphological traits of tadpoles also differed between the two venues. Finally, in the presence of the herbicide, tadpoles tended to be more active and to stay closer to the bottom of laboratory containers, while tadpole behaviour shifted in the opposite direction in outdoor mesocosms. Our results demonstrate major discrepancies between results of a classic laboratory-based ecotoxicity test and outcomes of an experiment performed in outdoor mesocosms. Consequently, the use of standard laboratory tests may have to be reconsidered and their benefits carefully weighed against the difficulties of performing experiments under more natural conditions. Tests validating experimentally estimated impacts of herbicides under natural conditions and studies identifying key factors determining the applicability of experimental results are urgently needed.