Insects and Insecticides

Written evidence submitted by Dr Christopher Connolly, University of Dundee


1. Pesticides are screened for safety on the basis of their ability to kill individual bees (LD50) but no consideration is given to sub-lethal toxicity.

2. The LD50 is determined for individual bees, not whole colonies.

3. Sub-lethal toxicity does not, necessarily, mean the death of the individual bee.

4. Sub-lethal toxicity may induce a vulnerability to other insults such as disease.

5. Many pesticides target the insect brain.

6. Sub-lethal toxicity in bees may lead to a dysfunction in the brain.

7. Many pesticides are used prophylactically by farmers and in combinations that are not reported.

8. Pesticides can act together by disrupting related targets.

9. All chemicals, be they medical therapeutics or pesticides, exert off-target activity. How this works is unpredictable and need to be tested empirically.

10. Lab tests versus ‘realistic’ field studies.


1. The level of pesticide required to kill a bee is important, but misses the real toxicity of compounds. Chemicals may cause chronic damage to insect pollinators (possibly even humans!) if exposed acutely (eg. Asbestos exposure in humans) or chronically (eg. Alcohol/smoking or therapeutic drugs like valium in humans). In both human cases, toxicity is only evident after long periods. Delayed toxicity has now been demonstrated in bumblebees (Whitehorn et al 2012, Gill et al 2012), where pesticide effects require many weeks.

2. For the social insect such as the bees, ants and wasps, it is the colony that is the breeding unit and so it is this that is most important. I accept that it is not reasonable to use whole colonies of honeybees for toxicity studies as this would be prohibitively expensive and flawed by their interaction with a complex environment that cannot be controlled.

3. Nevertheless, in the case of the social insects, individual weaknesses (non-lethal) may have a direct impact on the entire colony and poisons may even be taken back to the colony where they are stored (Mullin et al 2010) and fed to their developing young. As the neonicotinoids are based on nicotine, it is possible that the developmental toxic effects, observed in the human foetus of a smoking mother, predicts similar developmental deficits of bee larvae fed neonicotinoid contaminated food. Societal breakdown could occur at multiple levels, such as, learning (to be efficient in sourcing food), communication (sharing information regarding food resource availability/colony condition), navigation (negotiating their way in the environment)(Henry et al 2012), reproduction (queen only) and behaviour (colony dynamics).

4. Bees (or other pollinators) weakened by pesticide exposure may be more vulnerable to other threats such as disease or mite infestation. In fact the combined toxicity of a pesticide along with a disease is a common strategy of "Integrated Pest Management" as recommended by WHO to tackle malaria (using a fungus with Permethrin), cattle ticks (fungus plus deltamethrin) and maize rootworm (nematode plus tefluthrin). So, it is likely that such interactions occur in our pollinators that are facing multiple chemical and disease stresses. In support of this hypothesis, this possibility is starting to be reported (Alaux et al 2010, Aufauvre 2012, Vidau 2011, Pettis et al 2012, Wu 2012). The mechanistic basis for this is unknown.

5. We know that many pesticides target the insect brain, making the social insects more vulnerable to their exposure. The brain is a plastic structure that relies on changes to drive higher cognitive function, mood and social behaviour.

6. Dysfunction of the brain may not cause gross morphological changes. In fact, dysfunction is more likely to result in subtle changes to the structure and function of synapses (sites of information transfer between neurons and the sites of learning). Synapses can learn to become stronger, or weaker, and so directly impact the efficiency of information flow in that particular circuit. Disturbing this ‘plasticity’ can lead to alterations in their learning ability and/or affect mood/social interactions.

7. Pesticides are now used as preventative measures, in the absence of any threat to the crop (or pets – eg. Worming). Therefore, the risk to the environment and human health is much greater than necessary. We should not be killing all insects (and so the local ecosystem), only those that have become a problem. In fact, the situation is even worse as the information on what pesticides have been applied (and where and when) is not available. Therefore, should particular pesticide combinations be dangerous, we could never learn from such mistakes. Suppose 10% of local inhabitants are exposed to a cancer-causing combination of pesticides. Ten years later we may (or may not) identify a link with the local environment but would not have access to the information required to make that link. However, if the local use of pesticides were available, bioinformaticians/epidemiologists could correlate local bee losses (we saw a 5% overwintering failure in the west of Scotland and a 20% loss in the east, Fife was particularly bad) with local pesticide use. The identity of the farmers could easily be kept confidential as it is the correlation of pesticide use to pollinator losses that is important. Achieving this important policy change would have a major impact and could fast track scientific research by targeting it to potential causes of the pollinator declines. Such information may also inform on the causes of the many idiopathic, chronic human diseases like the neurodegenerative disease and Irritable Bowel Syndrome in humans.

8. Pesticides can work together at target sites to enhance toxicity. We have tested this hypothesis in our ongoing research programme "An investigation into the synergistic impact of sublethal exposure to industrial chemicals on the learning capacity and performance of bees" (funded by the IPI), with respect to the cholinergic synapse that is targeted by pesticides that; A. Alter the release of acetylcholine (eg. λ-cyhalothrin and τ-fluvalinate). B. Inhibit the removal of excess acetylcholine (eg. Chlorpyrifos and coumaphos). C. Directly stimulate the excitatory acetylcholine receptors (neonicotinoids). Together, chemicals targeting these sites are likely to work in concert to increase the neural deficits or lower the dose required to perturb the neural pathway. Our studies have shown interactions between imidacloprid and coumaphos, at both the level of brain activity (Dundee - manuscript under review, Palmer et al) and learning (Newcastle - manuscript under review, Williamson et al) in the honeybee, or with imidacloprid and λ-cyhalothrin on bumblebee colony performance (Gill et al 2012). Similarly, interactions between coumaphos and τ-fluvalinate have been shown to enhance toxicity to honeybees (Johnson et al 2009). Interactions at other synapses are also likely, as well as interaction at other sites (eg. Gut function or chemical detoxification).

9. In addition to the consequences of toxicity due to pesticide effects at target sites, significant off-target activity is also common. This is also true for therapeutic drugs where their use is determined according to their side effects. For pesticides, it is well known that many of the fungicides are much more toxic than anticipated, exhibiting unexpected synergy with other chemicals (Pilling et al 1995). We are, using in vitro models, researching a particular fungicide that appears to interact with cholinergic therapeutic agents used medicinally to treat Alzheimer’s disease patients and women treated for bladder weakness (unpublished data – MRC grant application under review).

10. With respect to the criticism of the validity of all lab studies, past and future, in preference for the more relevant field studies, I consider this claim totally unprofessional and lacking all scientific credibility. Laboratory studies are the cornerstone of all therapeutic drug discovery as they provide a mechanistic description of events that can be controlled and tested experimentally. These studies identify real and quantified threats. In contrast, field studies are performed in a particular context with an uncontrolled surrounding area. What may be found at one site could be irrelevant to that found at another site. This is especially important given the multiple stresses to which our pollinators are exposed and the likelihood that multiple threats contribute to the pollinator decline. It is true that a laboratory based mechanistic explanation does not confirm that these effects are largely responsible for the pollinator decline. This will require countrywide bioinformatics once we know what pesticides have been used. An isolated field study has limited value.

How do we proceed to put in place more appropriate testing regimes? In the absence of knowledge regarding local pesticide use this will be difficult and should not be permitted. Nevertheless, more interaction of DEFRA with university laboratories is essential to determine these new risks. Key disciplines, such as pharmacology and neuroscience must be included in the assessment process (this is seriously lacking at present). All new compounds should be subjected to these higher standards (sub-lethal and chronic toxicity on both honeybees and bumblebees) before they are released for use. This will require the companies paying (indirectly to avoid any undue influence) for the independent university study.

In summary, we are playing ‘Environmental Ker-Plunk’, using pesticides to remove insect species (possibly also higher species) and we don’t know which species will be lost and how many other species will collapse with them. Eventually, the entire ecosystem will collapse unless we monitor and regulate pesticide use appropriately. With the growing world population, with increasing appetites, we have to learn to live with pesticides, not just ignore them.

26 October 2012

Prepared 5th November 2012