Guest blog – Neonic analysis by Jeremy Greenwood

Mark writes: This guest blog, by Prof Jeremy Greenwood, rounds off the story of doubts about Syngenta’s analysis of impacts of neonics on bees. 

Professor Jeremy Greenwood CBE, Hon DSc, CBiol, FRSB is a former Director of the British Trust for Ornithology and is an Honorary Professor attached to the Centre of Ecological and Environmental Monitoring at St Andrews.

I’ll blog about how I see the significance of these results in a general context later today.

 

Prof Greenwood writes:

The neonictinoid thiamethoxam, manufactured by Syngenta, is widely used as an agricultural pesticide. There is extensive evidence under laboratory or semi-laboratory conditions that it damages honeybees and bumble bees but fewer studies have been carried out wholly outside the lab, in the agricultural landscape.  One of them was done by Syngenta (Pilling E, Campbell P, Coulson M, Ruddle N, Tornier I. 2013 A four-year field program investigating long-term effects of repeated exposure of honey bee colonies to flowering crops treated with thiamethoxam. PLoS ONE 8, e77193. doi:10.1371/journal.pone.0077193).  Unfortunately, the experiments in this study were inadequate and their interpretation misleading.

I have blogged here previously (No difference? 4 August 2014) about the Syngenta study but now am able to report on a more thorough examination of its results.

There were two parallel experiments in the study, one on maize, the other on oil-seed rape. In the former, two fields were used in each of three regions, one being treated with thiamethoxam, the other serving as a control (untreated). In the parallel experiment on rape only two regions were used.  Thus the level of replication of the experiments was a mere three (maize) and two (rape).  It was these tiny sample sizes that led Pilling and his colleagues to undertake no statistical analysis of their data.  They mistakenly argued that, because the sample sizes were so small, ‘such an analysis would lack the power to detect anything other than very large treatment effects, and it is clear from a simple inspection of the results that no large treatment effects were present. Therefore a formal statistical analysis … would be potentially misleading’. On the contrary, the truth is that formal statistical analysis is especially important when one is dealing with small sample sizes.

Pilling and his colleagues compounded their error by simply looking over their data and concluding that because the differences they observed between hives associated with treated and with control fields were not large, there was a low risk to honey bees from the use of thiamethoxam. In an interview a Syngenta representative went further, stating ‘We saw absolutely no effect on the honeybee colonies in those trials’. (Eisenstein M. 2015 May 21.  Seeking answers amid a toxic debate.  Nature 521: S51-55).  This is unacceptable.  The fallacy of the argument ‘I can’t see it, so it does not exist’ was established long ago in science. The only safe form of statistical inference is both to estimate the size of an effect and, crucially, to calculate the precision of that estimate (using such things as confidence limits).

Two colleagues in St Andrews and I were given access to the experimental data by Syngenta and we have subjected them to a formal analysis and recently published the results (Robert S. Schick, Jeremy J. D. Greenwood and Stephen T. Buckland. 2017. An experiment on the impact of a neonicotinoid pesticide on honey bees; the value of a formal analysis of the data. Environ Sci Eur (2017) 29:4 DOI 10.1186/s12302-016-0103-8).  What we found was that the confidence limits of the estimates of the effects of thiomethoxam were mostly too wide to reveal anything useful: the results were consistent both with the hypothesis that the effects of thiamethoxam on bees in normal agricultural use were trivial and with the hypothesis that they were large enough to be ecologically or economically important.

There have been a number of other experiments on the effects of neonics on bees out in the agricultural landscape. Of those published so far, all have had so few replicates that the precision of the estimates was too low to reveal even ecologically and economically important effects.  Those who support the continued use of neonics in agriculture have argued that the fact that several studies have not revealed important effects means that there are no such effects.  This is a completely fallacious argument, akin to arguing that if none of six observers, each of them using low-power binoculars, cannot see a distant bird then there is no bird.  Someone needs to bring binoculars of higher power to bear.  Fortunately the Centre for Ecology and Hydrology (CEH) has done so and will shortly be publishing the results of a larger and better-designed study that will lead to substantially more useful results than any of the low-power experiments conducted so far.

 

[registration_form]