ReportsEcology

Country-specific effects of neonicotinoid pesticides on honey bees and wild bees

See allHide authors and affiliations

Science  30 Jun 2017:
Vol. 356, Issue 6345, pp. 1393-1395
DOI: 10.1126/science.aaa1190

eLetters is an online forum for ongoing peer review. Submission of eLetters are open to all. eLetters are not edited, proofread, or indexed.  Please read our Terms of Service before submitting your own eLetter.

Compose eLetter

Plain text

  • Plain text
    No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g. higgs-boson@gmail.com
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Statement of Competing Interests
CAPTCHA

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Image CAPTCHA
Enter the characters shown in the image.

Vertical Tabs

  • RE: Unconvincing Statistics and Improper Controls
    • Ben A Woodcock, Ecologist, Centre for Ecology & Hydrology
    • Other Contributors:
      • Matt Heard, Ecologist, Centre for Ecology & Hydrology
      • Ricahrd Shore, Ecotoxicologist, Centre for Ecology & Hydrology
      • Peter Henrys, Statistician, Centre for Ecology & Hydrology
      • James Bullock, Ecologist, Centre for Ecology & Hydrology
      • Richard Pywell, Ecologist, Centre for Ecology & Hydrology

    Our paper (‘Country-specific effects of neonicotinoid pesticides on honeybees and wild bees’) represents the largest field scale assessment of the impacts of neonicotinoids (clothianidin and thiamethoxam) on honeybees and wild bees. In the following e-letter we respond to issues raised in an e-letter posted on this site by Pierroz et al. in relation to this publication. All these issues were considered by us in designing the study and writing the paper.

    Pierroz e-letter - Point 1: ‘After publication of “Country-specific effects of neonicotinoid pesticides on honey bees and wild bees”, by Woodcock et al. 2017 (Science 356 (6345): 1393-1395; DOI: 10.1126/science.aaa1190), we took umbrage with the statistical analyses used in this study. The study assessed 258 statistical endpoints and found significant results in only sixteen. Considering the 95% confidence interval, 6.2% of results yielding statistical significance should be attributable to random chance, especially since the Bonferroni Correction was not used’.

    Response.
    • Our focus in the paper was to report key biological endpoints (recognised within pollinator ecology and by risk assessors such as EFSA (EFSA 2014)) that are established to predict the viability of honeybees and wild bees following exposure to neonicotinoids. For honeybees we present 9 statistical endpoints in the paper that focus on measures of colony strength and the overwintering success of hives, with a further 5 statistical...

    Show More
    Competing Interests: We are the authors of the paper in question. All conflicts of interest (i.e. funding from Bayer and Syngenta) are defined there.
  • Unconvincing Statistics and Improper Controls
    • Grady Pierroz, Ph.D. Candidate, University of California, Berkeley
    • Other Contributors:
      • Peggy G. Lemaux, Extension Specialist/Principal Investigator, University of California, Berkeley

    After publication of “Country-specific effects of neonicotinoid pesticides on honey bees and wild bees”, we took umbrage with the statistical analyses used in this study. The study assessed 258 statistical endpoints and found significant results in only sixteen. Considering the 95% confidence interval, 6.2% of results yielding statistical significance should be attributable to random chance, especially since the Bonferroni Correction was not used. The significant results did not even show a clear trend, as only nine of the sixteen indicated negative effects of neonicotinoids on bee health, while the remaining seven indicated positive effects. In no way did the data presented convincingly lead to the conclusion that neonicotinoids have an overall negative effect on honey bee health.

    In addition, supplemental materials revealed data that clearly invalidate the study’s methods. Within the supplement are results from two different quality control experiments, which assessed the measurable amount of neonicotinoids in the fields used in this study. These two quality control experiments led to wildly different findings, and both showed that the experiments had major problems. One quality control check indicates that the control fields had neonicotinoid concentrations at levels equal to or higher than the experimental fields. Data from the other check shows that the thiomethaxam-treated field had no detectable levels of thiomethaxam, while the control field had a detect...

    Show More
    Competing Interests: None declared.
  • RE: Problems with data provided on statistical tests:
    • Ben A. Woodcock, Ecologist (co-author of paper in questions), NERC Centre for Ecology & Hydrology
    • Other Contributors:
      • Peter A. Henrys, Statistician (coauthor of paper in question), NERC Centre for Ecology & Hydrology

    Clarification of the test statistics: Thanks for the author of this e-letter for raising this we cans see some clarification is needed. As stated throughout the paper and supplementary materials, the tests conducted were based on the likelihood ratio test (LRT). The use of LRT was suitable across the analyses and allowed some level of consistency throughout the paper. For each LRT performed in the paper, we have stated that the Chi-Squared distribution was used to assess overall significance, we also state the degrees of freedom for this test, the estimated deviance and the corresponding p-value. This was quoted as " “χ2(6)=1.47, P=0.01” The key aspect here is that the deviance has been reported in all cases, not the Chi-squared test statistic as assumed by the authors. For distributions where the dispersion is known (e.g. binomial, poisson, negative binomial) the deviance and the chi squared test statistic (from a LRT) are equivalent and hence interpreting what we quote as a test statistic is absolutely fine. For distributions where the dispersion is estimated (e.g. Gaussian, quasipoisson, quasibinomial) the two are not equivalent and treating the deviance as a chi squared test statistic will lead to inconsistencies. In these cases the deviance should be scaled by the estimated dispersion to achieve the Chi squared test statistic. Therefore, the values quoted in the text should be interpreted as deviance measures rather than test statistics. We acknowledge that ou...

    Show More
    Competing Interests: We are the coauthors of the paper in question and competing interests (i.e. our funding source form Bayer and Syngenta) are defined in that manuscript.
  • Multiple stressors, neonicotinoid insecticides and bee declines
    • Joachim R. de Miranda, Researcher/Professor, Swedish University of Agricultural Sciences, Uppsala, Sweden
    • Other Contributors:
      • Francesco Nazzi, Professor/Researcher, University of Udine, Udine, Italy

    The paper by Woodcock et al. (1), recently published on Science, is the latest in a long series trying to assess if neonicotinoid insecticides are harmful to bees under realistic field conditions. It follows a number of studies carried out under more controlled conditions which clearly showed that compounds from this family of insecticides have several sublethal effects on bees, affecting navigation, immunity and reproduction (2-4). In fact, despite the bigger scale of the present study, its conclusions do not seem to be different from those attained previously, in that negative effects were noted on some bees, under certain conditions, whereas other bees, under the same conditions, did not seem to suffer adverse effects (5).
    This study will therefore most likely be coopted to bolster partisan agendas on both sides of the neonicotinoid debate: that these chemicals are inherently dangerous to pollinators and should be banned, or that the real-world consequences for pollinators of neonicotinoid use in agriculture remain unproven and are (therefore) insufficient to off-set its incremental benefits relative to alternative pest control methods. Both claims are grounded in reasoned arguments and a resolution requires a decision of how much damage society is willing to accept, and to what benefit. Neonicotinoids are highly effective at killing insects and it is disingenuous to pretend that just pollinating insects are somehow exempted, even when this cannot be demonstrated...

    Show More
    Competing Interests: None declared.
  • Problems with data provided on statistical tests:

    When examining the detailed data provided in the Supplementary File (Tables S4B,S4C,S5B), it seems that p-values published for the CHI2 statistics are sometimes not the good ones. Maybe mistyping.
    For example: Table S4C: CHI2 (6)=1.47, p<0.01;CHI2(4)=0.85, p=0.01; CHI2(4)=6.65, p=0.03.

    Competing Interests: None declared.

Navigate This Article