ecotoxicology To the Editor — It has been argued that are extremely difficult to control and rearing in the laboratory12,13. These two that we are currently facing the sixth mass standardize, future ERAs would need to examples illustrate solid and feasible ways extinction event1, and the ubiquitous use of be based only on measures of fitness in to measure fitness in the laboratory that chemicals is one major driver for this loss the laboratory. Fitness can be notoriously transcend species boundaries at least of biodiversity2. Ecotoxicology, the difficult to assess per se, especially under for bees, thereby constituting two ideal estimation of environmental risks imposed controlled laboratory conditions, where models of future ERA. The identification by chemicals, is therefore an essential the well-being of an animal is often not of more suitable model species will backbone for adequate mitigation efforts by met. Despite it now being common to have clearly require substantial research, stakeholders. However, if evolutionary guidelines addressing the well-being of especially on the well-being of biology is not taken into account in animals across all vertebrate taxa9, such invertebrates, yet we must not shy away ecotoxicology, we may be asking the wrong practical implementations do not currently from meeting these challenges. questions in order to protect natural exist for invertebrates, in particular those In conclusion, a shift to Darwinian biodiversity. Here, we propose focusing providing essential ecosystem services. ERA appears long overdue if our aim is to on fitness, the essential factor governing Given the major global declines in wild effectively mitigate the role of chemicals wild populations, for ecotoxicological risk pollinating insects, this constitutes a in the ongoing mass extinction of species. assessment (ERA). major knowledge gap. Nevertheless, Policy decisions based on mortality Current ERA schemes appear flawed fitness can be practically incorporated estimates and sublethal effects other than and outdated3 because they emphasize the in future ERAs and must be the gold fitness are probably leading to insufficient mandatory endpoint measurement survival, standard, regardless of how challenging mitigation efforts. In light of limited despite evidence of sublethal effects being it may be. Laboratory procedures to resources, we therefore urge future ERA to far more common and likely having severe estimate fitness indeed already exist focus exclusively on measures of fitness by consequences4. Focusing on longevity only for many species (for example, solitary using multi-generation reproductive studies becomes relevant if it ultimately enhances bees6), which seem appropriate given the in the laboratory. This would be a major fitness. Irrespective of what the stressor fitness of the laboratory-held controls step forward in ERA and enable reliable may be, it therefore seems apparent that any matches respective field estimates. This estimations of the environmental risks reduction of fitness, and not survival alone, value should be maintained over multiple imposed by chemicals. ❐ will have fundamental consequences for all generations to ensure a chemical has no species. Furthermore, false negative results, significant long-term transgenerational Lars Straub, Verena Strobl and such as ecotoxicological studies yielding effect. In some cases, laboratory assays Peter Neumann ✉ negative evidence for survival or even may not allow the estimation of fitness, Institute of Bee Health, Vetsuisse Faculty, indicating enhanced survival5, but disguising for example, for long-lived social insect University of Bern, Bern, Switzerland. a reduced fitness6, are likely to occur. colonies such as honeybees, for which ✉e-mail: peter.neumann@vetsuisse.unibe.ch Indeed, there are ample data showing clear fitness is the number of surviving swarms negative effects of chemicals on fitness, but produced and the number of successfully Published online: 23 April 2020 not on survival7, probably due to a trade-off mating males10. Honeybees are therefore https://doi.org/10.1038/s41559-020-1194-6 between survival and reproduction8. It is not a suitable model species for Darwinian therefore obvious that such a trade-off is ecotoxicology tests. Instead, the model References 1. Barnosky, A. D. et al. Nature 471, 51–57 (2011). paramount for any meaningful ERA. This species for assessments must be selected 2. Gossner, M. M. et al. Nature 540, 266–269 (2016). also holds true for any measurement of based on their suitability for measuring 3. Topping, C. J., Aldrich, A. & Berny, P. Science 367, 360–363 (2020). 4. Neumann, P. Nature 520, 157 (2015). sublethal effects, given they are no more fitness under laboratory conditions, in 5. Guedes, R. N. C. & Cutler, G. C. Pest Manag. Sci. 70, relevant than mortality for the evaluation combination with established fitness data 690–697 (2014). of fitness. Ultimately, any assessment of from the field. These species have to be 6. Sandrock, C. et al. Agric. For. Entomol. 16, 119–128 (2014). a sublethal parameter is only valid if it chosen by stakeholders. One example for 7. Kliot, A. & Ghanim, M. Pest Manag. Sci. 68, 1431–1437 (2012). 8. Harshman, L. G. & Zera, A. J. Trends Ecol. Evol. 22, 80–86 (2007). adequately reflects fitness. Therefore, the practical implementation of Darwinian 9. Bloomsmith, M. A. et al. in Management of Animal Care and conventional ecotoxicology tests are probably ERA is solitary bees, Osmia spp., for Use Programs in Research, Education and Testing 2nd edn misleading for risk assessment schemes which fitness can be measured under (eds Weichbrod, R. H. et al.) Ch. 5 (Taylor & Francis, 2018). 10. Neumann, P. & Blacquière, T. Evol. Appl. 10, 226–230 (2017). and will lead to inadequate policymaking. laboratory conditions6, as females are 11. Phillips, J. K. & Klostermeyer, E. C. J. Kansas Entomol. Soc. 51, A change in mind set must occur in our expected to produce only ~30 offspring 91–108 (1978). community, because nature conservation under favourable conditions11. Likewise, 12. Baron, G. L. et al. Nat. Ecol. Evol. 1, 1308–1316 (2017). 13. Whitehorn, P. R. et al. Science 336, 351–352 (2012). requires an evolutionary approach. the fitness of eusocial bumblebees, Bombus Since field experiments are only used spp., can easily be estimated by taking Competing interests for ‘higher tier’ evaluations, as they advantage of routine multi-generation The authors declare no competing interests.