Standardize or Diversify Experimental Conditions in Ecotoxicology? A Case Study on Herbicide Toxicity to Larvae of Two Anuran Amphibians.

Abstract

Despite a steeply increasing number of ecotoxicological studies on the effects of pesticides on nontarget organisms, studies assessing the adequacy and reliability of different experimental approaches have remained scarce. We scrutinized effects of a glyphosate-based herbicide on larvae of two European anuran amphibians by estimating species-specific LC50 values, assessing how an additional stress factor may influence outcomes, and investigating whether replicate experiments yielded qualitatively the same results. We exposed Rana dalmatina and Bufo bufo tadpoles to two predator treatments (no predator vs. predator chemical cues) combined with varying herbicide concentrations, repeated the experiment with a subset of the experimental treatments and partly with slight modifications 1week later and assessed survival. Our results indicated that the herbicide was moderately toxic to tadpoles. The presence of predator chemical cues did not affect the lethality of the herbicide in either species. The estimated sensitivity of R. dalmatina tadpoles varied considerably across experiments, whereas in case of B. bufo LC50 values remained very similar. Our results suggest that differences in the experimental setup may often have no influence on the measured effects of pesticides, whereas replicated experiments can deliver widely differing results in other cases, perhaps depending on the studied species, the population origin of the tested individuals, or the test conditions. This draws attention to the suggestion that strict standardization may not deliver widely applicable insights into the toxicity of contaminants and, instead, intentionally introducing variation into the design of ecotoxicological experiments and replicating entire experiments may prove highly beneficial.