• GM crops;
  • regulation;
  • ecology and policy;
  • gene flow;
  • risk assessment


  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References
  • 1
    The current controversy over the possible commercial-scale cultivation of genetically modified crops in Europe presents enormous challenges for the science of ecology. In little more than a decade, ecology has come to occupy centre stage in a very heated public debate.
  • 2
    In outlining the challenge to ecology I discuss some of the key elements in the GM debate, and trace the growing importance of ecological questions within the regulatory and advisory system in the UK. The need to understand, and predict, some of the potential wider environmental impact of genetically modified (GM) crops has both driven changes in the regulations and led directly to farm-scale evaluation of herbicide-tolerant crops.
  • 3
    Several examples illustrate how ecological science is being used, and abused, in the public debate and in the provision of advice to regulators. In particular short-term or laboratory studies identifying possible hazards or impacts often receive widespread media attention but the thorough ecological field-based studies which either evaluate exposure to a hazard or assess fitness over several generations are rarely carried out, or, in the classic case of the impact of Bt maize on the Monarch butterfly, pass almost unnoticed.
  • 4
    It is increasingly important that trained ecologists become involved in the public debate. The challenge of dealing with the problems of variability, complexity and uncertainty, and of developing the necessary predictive tools for risk assessment, bring with it a huge responsibility, not only to be clear about the limitations of our science, but to recognise and acknowledge the boundary between science and informed opinion.


  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References

Secondly, my talk has an intriguing precedent in the Eighth BES Lecture, by John Beringer, in that it is given within a month or two of his retirement from the post, by a former Chair of the Advisory Committee on Releases to the Environment (ACRE) (Beringer 2000). Whilst sharing John's obvious sense of release from the strictures of that job, I want to use this opportunity to present a personal view of what I believe to be the lessons we as ecologists must learn from the GM debate, and to explore the huge challenges it has presented to our science. Needless to say, the views are mine and do not represent those of any official body or of my former employers.

The lecture has three parts. First, by way of background, I will discuss what I believe to have been the major influences in the public debate in Britain about genetically modified (GM) crops. Secondly, in describing the role of science in the advisory and regulatory process, I will provide some examples of ecological studies that illustrate the pitfalls and limitations of that process. Finally, I return to the more generic theme of science and policy making.

Elements in the GM debate

  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References

How did we get to where we are? What has led to the raging controversy over the possible cultivation of GM crops on a commercial scale in Europe? Why, for example, does the owner of my local café feel he must display a notice claiming that ‘All the wheat milled for flour is free from genetic modification’, or the local fish and chip shop have to declare that ‘Potatoes that are grown and consumed in Britain are not genetically modified’? Both are statements of dubious scientific validity (i.e. they are strictly wrong) and have, incidentally, caused wry amusement among scientists, especially plant breeders, in distant parts of the world. (Although it is comforting to be reassured about the safety of the British chip, which must have helped to fur up the arteries of a generation!) To understand the reasons behind such notices, and why the GM debate in the UK is taking place against a backdrop of entrenched opposing views, one must trace some of the events that have brought us to this point.

I should stress that my attempt to recognize what I call ‘elements’ in the GM debate is a strictly amateur exercise. I have no training as a social scientist, and I am certain that more than one member of that discipline will eventually look back on events and provide us with an accurate structured analysis of the times we live in. What I am offering here (and probably in simpler language than my social science colleagues might use) are some insights from the perspective of someone closely involved in the debate and in providing advice to government for more than a decade. To emphasize the lay nature of my effort I have produced a cartoon depicting the elements of the debate as logs that have helped to ‘fuel’ the flames of the GM controversy (Fig. 1).


Figure 1. Elements in the GM debate; see text for explanation.

Download figure to PowerPoint

bse andgovernment science

Most observers agree that the circumstances in which links were drawn between bovine spongiform encephalopathy (BSE) and the fatal human brain disease new variant CJD (Creutzfeldt-Jakob Disease) has had a devastating effect on public confidence in science, especially in the UK. This (and other events such as the outbreaks of Salmonella and E. coli 157) sensitized people to food-related issues and bred distrust of what was (wrongly) seen as government science and the regulatory system into which it fed. ‘Experts’ could clearly get it wrong, tragically wrong. Moreover the probity and independence of these hitherto anonymous experts was open to challenge. I well remember a BBC television programme, ‘Panorama’, that depicted ACRE as a group of cigarette-smoking middle-aged men in business suits huddled around a table in a darkened room (an image that did not go down well with the women on the committee!). The perception of smoke-filled rooms and secret decisions tainted by vested interest was alive and well.

campaigning organizations

Propagation of such a view has suited one or more of the campaigning organizations that form the second ‘plank’ (or log) in my scheme. Well-trusted environmental non-governmental organizations (NGO), notably Friends of the Earth and Greenpeace, have taken a strong broadly anti-GM stance and have been joined in this by organizations ranging from Christian Aid to the Women's Institute. The rooted opposition to genetic modification by the Soil Association has been a major force and continues to influence the debate in Europe about the possibilities of coexistence for organic and GM crops. The reasons why various organizations have adopted the view they have are frequently different, not always clear and not necessarily based on science. As a naive ecologist first encountering the debate many years ago, I could not understand why the various combatants were in the corners they were: why would organic farmers not, for example, welcome genetic protection of crops against pests and disease, helping to eliminate agrochemicals? Of course I have long since learned that the GM debate is about much more than science. It has become what Robin Grove-White has termed ‘iconic’, acting as a distillation point for different belief systems. The GM debate is in many ways a surrogate for a debate about issues such as globalization, multinationalism, who controls the food chain, and what are the ethically acceptable limits to biotechnology.

corporate behaviour

This brings me to a group of factors that I have bunched together under the heading ‘Corporate behaviour’. Among these I include key decisions by multinational biotechnology companies [for example, Monsanto's (1999) decision not to segregate Round-up ready and conventional soybeans in bulk shipments to Europe is regarded by many analysts as the single major trigger to opposition by a public whose choice to avoid GM food was consequently removed] and by British supermarkets (influential among which was Sainsbury's decision to withdraw from sale tomato paste derived from tomatoes modified to delay the ripening process). The huge investment by biotechnology companies that has generated a range of intellectual property issues, including resentment about the apparent ownership of genes and the difficulties of transferring technology to the developing world, should be seen in the context of a trend towards almost total privatization of the plant breeding and biotechnology industry. It is interesting to speculate how different the public debate might be were the genes that have been used to transform crops in public ownership. Or indeed how different it might be if genes for resistance to the herbicides produced by the agrochemical companies had not been among the first to market.

risks and benefits

Whatever the influence on the debate of such factors, it is quite clear that the public largely see the benefits of GM crops to be going to the biotechnology companies, or possibly the farmers, whilst they, on the other hand, appear to be taking all the risks. This separation of risk and benefit, and the fact that the benefits are unclear (e.g. why produce more food in an overfed developed world?) drives a small but I believe significant part of the debate. Interestingly the use of GM technology in the medical sphere, as for example ACRE has found in assessing the risks from clinical trials of GM vaccines, has raised comparatively little public anxiety. As has been demonstrated in another context, we are reasonably happy to expose our tissues to ionizing radiation in the interests of personal health, but unwilling to eat cheese that has been similarly irradiated. I will return to the risk–benefit dichotomy later.

the role of the media

My fifth element is the fourth estate. Whether one believes that the media are the major influence on public opinion or merely reflect it, there is no doubt that they can all recognize a good story and that most are very willing to throw another log on the fire (perhaps especially if it generates more heat than light). and above all else GM has been a good story. It is a story with enough demons, good guys and scary concepts for all tastes (as more than a decade of headlines has demonstrated). However, there are two aspects of the media coverage of the GM debate that have struck me as probably different. The first is that they have reported a very public quarrel between scientists. Arguments that are more usually rehearsed in the relatively obscure pages of scientific journals have been laid bare for public scrutiny. An emphasis on scientific uncertainty and controversy has done little for public confidence in the scientific process, and has included events that can have brought little credit to those involved. For example, public confusion about the work and subsequent treatment of Dr Arpad Pusztai was another important ‘trigger’, more or less coincident with the Monsanto soybean decision, that further inflamed the debate in the late 1990s. A second aspect of media coverage (not unique, but in the context of science journalism surely unusual) is that some sections of the media have themselves become campaigners. One knows what to expect from a particular source and for these sources journalism departing from the editorial line is rare. It was especially disappointing for me to realize that the reaction of particular newspapers to the recent very neutral and balanced report from the GM Science Review Panel (GM Science Review Panel 2003) was actually almost entirely predictable.

plato and public understanding

Mention of the GM Science Review Panel brings me to the final aspect of the debate, which I have labelled ‘Plato and public understanding’ and which obviously requires some explanation. The GM Science Review Panel, comprising 25 scientists from a range of scientific backgrounds, was established as part of the current high-profile public debate about GM crops in the UK. One of the problems about having a public debate about GM that has worried me for some time (e.g. The Times Higher Educational Supplement, 11 September 1998) is that GM is essentially a technology, such as microscopy or vaccination. Yet we debate it as would the students of Plato, concerning ourselves with generalizations, broad ideas and syntheses, rather than the students of Aristotle who would pay attention to case-by-case detail, fact, logic and analysis. One is frequently asked ‘are you (or is he or she) pro-GM or anti-GM?’ which, like the questions ‘are you pro- or anti- microscopy or vaccination?’, is difficult to answer in a platonic, generic way (except perhaps at a cultural or ethical level) and for a scientist difficult to answer without reference to particular detail (wherein of course lies the devil). Despite this, quite remarkably, everyone seems to have a ‘view’ about GM.

If we are to believe some of the surveys of public understanding of the relevant fields of science, these views are often based on very little information. For example, a well-cited survey published in 1999 indicated that, on average, almost 30% of those questioned in 12 developed nations believed that tomatoes did not contain genes until the genetic engineers put them there (and a further 35% did not know) (Hoban 1999). This may not be surprising (why should we expect people to have kept pace with the astonishing developments in genetics in the last 50 years?) but it provides in my view both a crushing indictment of scientists as communicators and educators and a revelation that, in this area, the public may be especially vulnerable to persuasion by those with a vested interest in capturing their support.

And that, roughly speaking, is how we got here: with hugely polarized opinions, megaphone diplomacy, a pretty hostile media, environmental activism, threats of trade wars, and every organization feeling it incumbent upon them to have a view. The many official ‘position’ statements have included contributions from the BES (Shorrocks & Coates 1993) and from the Ecological Society of America (Tiedje et al. 1989): both, you will not be surprised to learn, offering reasoned and balanced perspectives on the issue.

Science in the regulatory process

  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References

the regulatory framework

I want now to turn from the heat of the debate to say something about the role of science in regulation, specifically in the context of the release of genetically modified organisms (GMO) in Europe. To do this I must first describe the regulatory framework. I am aware that this is not exactly riveting material to present in a public lecture series designed to stimulate debate and to be controversial, but it will help to touch on it briefly for the development of my theme about the challenges to ecology.

The release of all GMO, including GM crops, in the European Union (EU) is controlled under an EU-wide European Council Directive. In February 2001 a new Directive (The Deliberate Release Directive 2001/18/EC) was adopted, replacing one that had been in force for 10 years. It included, among other things to do with harmonization and transparency of process, the new, quite explicit, instruction that when assessing the risks of each release the regulatory authorities must consider not only possible direct and immediate effects but also those that may be indirect, delayed, longer term or cumulative. This change in the regulation, as I will argue later, together with the introduction of a requirement for a post-market monitoring plan, was driven almost exclusively by ecological imperatives.

Thus for GM release in the EU there is a body of dedicated legislation, and release and marketing can only take place with the explicit consent of the regulatory authorities; this is in contrast to countries such as the USA, where GM crops are dealt with under existing legislation and, once approved, are deregulated under licence. All types of releases are covered, once the GMO comes out of the contained conditions of the laboratory or glasshouse, from small-scale research and development (R&D) trials (so called ‘Part B’ consents) to applications to ‘place on the market’ (‘Part C’ consents). In the latter case EU member states evaluate the application only after it has been assessed as low risk by the member state to which the application was first made. (Since 1993 the UK has dealt with more than 200 consents for R&D releases under Part B of the Directive and 18 products have been approved for commercial release in the EU; these include four GM maize types, oilseed rape, soya, carnations and chicory, some of which are for import only and several of which will be reassessed under the new legislation.)

At the heart of all applications (referred to as ‘notifications’) by companies or academic research laboratories, is a technical dossier describing the proposed release in detail and upon which the applicants must base their risk assessment. The dossier must address a range of specific questions about the nature of the modification, including a detailed molecular characterization of the genes involved, evidence of their stability and expression, the effect of the modification on the biology of the transformed organism (including its persistence and invasiveness), the possibility and effects of gene flow and hybridization, potential effects on target and non-target organisms, indirect effects resulting from crop management, and a range of other questions depending on the nature, size and site of the release and the need to manage any identified risks. I have described the general features of such dossiers in some detail elsewhere (Gray 2003) and in any case they are publicly available. They are important because they provide the science, the evidence base on which advice is given on whether to issue a consent to release the GMO (and what conditions of monitoring or risk management to attach to that consent).

In the EU that advice is largely the responsibility of scientific advisory committees independent of government (also a difference from some other countries, where ministry officials provide the risk assessment); in the UK this is ACRE. ACRE is an independent statutory committee (i.e. it has legal standing) working in a fairly circumscribed way within the UK legislation that implements the EU Directive. The 12 or 13 members currently include experts in molecular biology, genetics, plant physiology, medical virology, agronomy, sustainable agriculture and ecology. One of the changes during my time with the committee, again reflecting changes in the science base for risk assessment, has been the increase in ACRE members trained as ecologists (covering areas such as weed biology, plant invasions, microbial ecology, herbivore–plant interactions, ecological genetics, biodiversity; several currently I know to be members of BES).

ACRE provides independent advice not only on GMOs but on other deliberate releases (usually for biological control but even recently including organisms as different as beavers and attenuated strains of microbial pathogens as possible vaccines). It provides advice to ministers on relevant new science developments or particular research requirements. This has included the setting up of a number of subgroups to look in detail at specific strategic issues. Significantly these have included in recent years a subgroup on wider biodiversity issues and one on the definition of environmental harm.

Finally, ACRE operates in a fully open and transparent way, with agendas, minutes and their full advice on the Web, the publication of an annual report and the declaration of member's interests. If they ever existed (which in my experience, going back to the late 1980s, they never did in ACRE's case), the smoke-filled rooms I alluded to earlier have been well and truly left behind. Indeed government scientific advisory committees today aspire to operate under guidelines largely devised by a former president of the BES, Bob (now Lord) May.

But I must leave detail of the regulatory framework there in favour of briefly discussing the regulatory process itself. This is not the place, nor am I the person, to tackle the generic limitations of that process. May I recommend the aforementioned GM Science Review Panel report (GM Science Review Panel 2003; also available on the Internet at as a good starting point for a discussion of those limitations and an introduction to the growing literature covering key concepts such as the ‘precautionary principle’ and ‘substantial equivalence’.

Instead I will briefly list what I believe are in practice the main features of the process of assessing potential risks from the release of GM crops.

First, the process is science-based, using both quantitative and qualitative evidence and data, and is comparative, it compares the GMO with its non-modified equivalent under similar conditions (recognizing there is no such thing as absolute risk). Secondly, the approach is precautionary; if there is any doubt one attempts to resolve it, if there are unresolvable uncertainties one attempts to make them explicit. Thirdly, it involves a case-by-case assessment, dealing with each crop–construct combination separately. A step-by-step approach ensures a gradually increasing familiarity with each GM crop as it moves from containment to the glasshouse to the R&D trial through to commercialization. Fourthly, the process is iterative and continuous, capable of responding to new information. The consent holder is obliged to send any new data to the regulatory authorities so that risk assessments can be revisited and if necessary revised or the consent withdrawn. Finally, the process utilizes the classic risk assessment paradigm: envisioning hazards, attempting to calculate the likelihood of those hazards being realized, and, by multiplication, assessing the risks posed by the release. One should note that this process (and ACRE and some other advisory committees that operate it) deals only with risks, not benefits.

the rise of ecology

I come now to the part of my talk I have called ‘The rise of ecology’. This rather dramatic headline is my way of describing a sea change that has taken place in the world of GM risk assessment in the last 5 years, although the gathering force of the current was perceptible long before that.

Throughout most of the 1990s, the process that I have just described to you seemed operationally perfectly adequate for the work of ACRE, and committees linked to it advising on the safety of novel foods and proteins and animal-feeding stuffs. It was, as we have seen, independent, evidence-based, grounded in a well-established peer review system, subject to challenge, dynamic, continuous and evolving; indeed I believe it continues to provide a rigorous basis for risk assessment today. However, especially with rising numbers of applications to grow GM crops on a commercial scale, it became increasingly clear that several of the questions raised by the dossiers were difficult, or impossible, to answer. These were generally questions about the potential environmental impact of releases. Many were, at least initially, in what Phil Dale calls the ‘nice to know’ rather than the ‘need to know’ category: they rarely addressed issues of safety or danger to human health or the environment. They were characterized by requiring ecological expertise and information and, unlike questions about, say, the nature of the inserted DNA or the biology of the transgenic plant, they were not questions about which it was clearly possible to seek immediate clarification from the applicant; one could not ask to see a Southern blot or the data from a field trial. This widening of the ambit of the risk assessment, which arguably drove changes in the regulatory framework, involved an extension of ACRE's remit to consider the potential wider environmental impact of releases (hence the increasing importance of ecology).

Not one to waste words, I will, as they say, use some I prepared earlier. This is from the Chairman's Foreword to ACRE's 1999 Annual Report (ACRE 1999):

perhaps ironically, the widening of the risk assessment is, in my view, unlikely to bring the increase in certainty or provide the yes/no answers that many people undoubtedly want. More and more the expert advisers and the recipients of that advice can expect to meet, and to have to deal with, the uncertainty and variability which characterizes biological systems, especially semi-natural ecosystems. This is not new or unique to the GM debate – we cannot hope to know everything about any new technology at the outset – be that advances in medicine, engineering or agriculture. However we can … acknowledge that uncertainty will always exist but in doing so accept that we do know enough … to be able to confidently take the first precautionary steps.

More and more we will call on the relatively young science of ecology where patterns and processes are well described and frequently understood but where prediction is still developing fast. I am optimistic that the science will serve us well and that much research to help reduce uncertainty is already underway. But … We must learn to operate within the bounds imposed by our understanding of nature, to accept that our calibration of biodiversity is imperfect, and, perhaps most importantly, to be very clear about how we employ such terms as ‘risk’, ‘harm’, ‘impact’ and ‘change’.

This struck a cautious, but I hope essentially optimistic, note. But was my optimism misplaced? Is our science actually serving us well? I want to pick out, in a moment, one or two words from the quotation to explore these questions further, notably ‘variability’ (and complexity), ‘prediction’ and ‘uncertainty’. But I should first mention that the reference in it to ‘biodiversity’ was to the work of ACRE's then newly established ‘wider biodiversity’ subgroup specifically set up to look, not at the direct effects of novel crops, but their less direct effects on the agricultural environment. Chief among these of course were crops (oilseed rape, maize and beet) engineered to be tolerant of wide-spectrum herbicides (glyphosate and glufosinate ammonium). Although, again, the issues were not about safety, the GM legislation was beginning to impact on other environmental legislation, including the pesticide legislation but most notably that deriving from the post-Rio Biodiversity Action Plan.

The widespread cultivation of such crops clearly had the potential to exacerbate the decline in farmland biodiversity. Although other species such as arable weeds are known to have declined equally rapidly, the spectacular declines in UK farmland birds species in the last 20 years (Fuller et al. 1995; Krebs et al. 1999) have been a loud clarion call alerting us to the wider impact of changes in agricultural practice [and explain why the Royal Society for the Protection of Birds (RSPB), together with the government's conservation agencies, were so influential in persuading government to invest in research in this area]. While the proximal causes of population decline may have varied between species (a switch to winter-sown cereals and oilseeds, and the increased use of herbicides and pesticides, being two of the potential contributing factors), the overall cause is generally agreed to have been various aspects of agricultural intensification. If herbicide-tolerant crops lead to further large-scale intensification they clearly have the potential to exacerbate the declines, in a country legally committed to halting and where possible reversing them.

All this resulted in the huge and eagerly anticipated experiment on the environmental impact of herbicide-tolerant crops: the farm-scale evaluations (FSE; Firbank et al. 1999). The timing of this lecture, perhaps thankfully, precludes all but a passing reference to the FSE, but I am convinced that its publication, and the reaction to it, will give us the clearest indicator to date of how well our science is serving us in contributing to the regulatory process. (I should say that you will see some very different accounts from social scientists of why government has spent millions of pounds on the FSE, accounts that emphasize the role of public pressure, distrust, political expediency and other social forces for change.)

Let me return to the question of how well ecology may have met the challenges posed by variability, complexity, the problem of prediction, and uncertainty. On the whole I believe the regulatory process has been well served by the way ecologists have dealt with biological variability. While, in contrast to advising on the release of agrochemicals or pesticides, there are few, if any, standard protocols for assessing risk such as measuring the effects of dosage or exposure times on target and non-target organisms, it has been largely possible to assess the performance of GM plants in field trials using current experimental and mathematical understanding (Firbank et al. 2003; Perry et al. 2003). This is perhaps not surprising given our knowledge of the biology of crops such as oilseed rape, both as crops and feral populations. Additionally some of the important, what you might call ‘back stop’, questions (i.e. would genetic modification produce some dreadful surprises or could we use the existing ecological and genetic paradigms to make predictions?) were answered in the early PROSAMO experiments carried out by Mick Crawley and colleagues (Crawley et al. 1993, 2001). Certainly in 1999, reviewing the risk assessment of a glufosinate-tolerant hybrid oilseed rape for commercial release, I felt (with Alan Raybould) that we had the science to make a confident assessment of risk (Gray & Raybould 1999).

Where we have perhaps been less sure-footed in translating science into regulation has been in the research on gene flow. The production of curves like the one in Fig. 2 has been something of a growth industry in the past decade. Figure 2 actually shows the frequency of double herbicide-resistant oilseed rape hybrids recovered in fields of one herbicide-tolerant variety as a function of distance from a source of pollen from a variety tolerant to a different herbicide. It combines studies from three areas of France. While this is not actually gene flow in the strict population genetic sense, there is no estimate of introgression, it provides a useful indictor of the potential crop-to-crop hybridization (some would say ‘contamination’) in French oilseed rape fields.


Figure 2. Frequency of hybrids as a function of distance from herbicide-tolerant oilseed rape fields (after Champolivier et al. 1999).

Download figure to PowerPoint

Such steeply declining leptokurtic distribution curves have a common shape, but in detail they vary from place to place depending on factors such as the size of the source and recipient populations, the breeding system and pollination mode of the species, and so on. Their generality, along with many years of experience from seed certification schemes, has been used to decide on separation distances between crops to minimize (not prevent) crop to crop gene flow. Their variability has however, caused enormous problems. (For example it is not exactly intuitive to non-biologists, who live in a largely linear world, that by doubling the separation distances you do not halve the risks of cross-pollination.) Individuals tend to emphasize that either (i) most plants mate with themselves or their nearest neighbours (less than 3% hybridization at 10 m) or (ii) rare cross-pollination events beyond a certain threshold may occasionally extend even for several kilometres away from the source (Rieger et al. 2002) (bumping along the floor below 0·05% for 130 m in Fig. 2).

What I think has not been well communicated in this concentration on gene flow is that cross-pollination is only one way in which GM and non-GM seed may be mixed together (others, such as failure to clear out the combine harvester, could be far more significant). Nor in my experience is it always appreciated that physical separation would never be advocated as a strategy to contain genes for which there was any evidence of harm. If gene flow is possible it is assumed it will happen: the probability of exposure in the risk assessment equation is set at 1.

Studies of gene flow between populations of the wild relatives of crop plants, such as those from our laboratory on wild cabbage and sea beet along the Dorset coast have, by using neutral marker genes, provided insights into how populations are structured genetically and how genes move around in species with contrasting biology and pollination modes (Raybould et al. 1996, 1999). But as model systems they probably complicate rather than clarify the situation, hence the value of direct measures of hybridization (and hopefully eventually of introgression) between crops and wild relatives, such as the many studies on gene flow between oilseed rape and wild turnip (these have recently included extensive studies in the UK funded under the BBSRC/NERC Gene Flow research programme; Wilkinson et al. 2003).

This leads me to the issue of the complexity of natural systems, which I believe on the whole has not been handled well in the context of risk assessment and regulation. By ‘complexity’ I am not referring to the fact that innately complex systems can only be understood using multivariate analyses or complex explanations. In fact parsimony and simplification are usually well-honed tools in the ecologist's toolkit. (Mind you, in discussion of risk assessment with politicians and the media one can well understand that Lady Thatcher is alleged to have called for a one-handed economist. As soon as one says ‘on the other hand’ one can notice the interest visibly decline!)

What has concerned me most is the way in which single studies, often preliminary and usually laboratory-based, have been seized on not only by the general media but also by the scientific press. It would be invidious to always lay blame at the authors’ feet, much of the work is perfectly good science and is often clearly presented as preliminary work. But we are now accumulating a long list of headline grabbing papers in this category, ranging from laboratory studies of insects feeding either directly on GM plants or on prey fed GM and non-GM plants (Hilbeck et al. 1998), through apparent increases in out-breeding in transgenic vs. non-transgenic plants (Bergelson, Burrington & Wichmann 1998), to reports of increased fecundity in single experiments in GM vs. non-GM plants (claimed in at least one paper on oilseed rape to be increased fitness, which it is not; Stewart, Raymer & Ramachandran 1997). Many of these studies, which characteristically have identified hazards, have never been followed up by published field-scale studies to measure exposure (in some cases no doubt because regulations covering GM releases preclude such research). Perhaps the best example of what, in a Nature Web debate I somewhat rudely called ‘soundbite science’, is the now famous story of the Monarch butterfly in the USA, where the laboratory study identifying pollen from Bt maize (containing a toxin from the bacterium Bacillus thuringiensis) as a hazard to larvae, caused enormous public outcry (Losey, Rayor & Carter 1999) and was actually followed by an extensive series of field-based and other ecological studies aimed at estimating both the exposure of the insect to that hazard (and hence the actual risk) and the relative magnitude of other risks that Monarchs face, such as predation, pesticides and habitat destruction in their over-wintering sites (Sears et al. 2001). I do not have time to outline the results, which I imagine most of you know about. Or do you? Perhaps unless you have been closely following the scientific debate you may not. While the 4-day experiment in plastic boxes that alerted people to a hazard was front page news, the extensive rigorous long-term research that importantly demonstrated the actual threats to Monarch butterfly populations, if reported at all, might have been found on page 43 along with the Latvian football results! (One hesitates to suggest that this could have been because they implied that Bt maize was probably less hazardous for Monarch butterflies than conventional maize.) Anyway, if you do not know the story (and even if you do) I can recommend an excellent account of the research, and the lessons we can learn from it for risk assessment, in the ACRE Annual Report for 2001 (ACRE 2001; also available on the ACRE Web site,

Moving on, I believe my optimism that the predictive tools would develop and improve was at least partly justified. In the field that I know something about, the effects of genetic modification on plant persistence and invasiveness, we have moved on in less than a decade from an early rejection of the value of so-called Baker traits (the traits characteristic of successful weeds), which were often referred to in the earliest applications but which Mark Williamson and others, for example, showed to be only very weakly predictive (Williamson 1996). The ‘exotic’ or ‘alien’ species model in which GM crops are compared with invasive alien species has also been found to be unhelpful as a predictive tool. More successful has been research based on population dynamics models, which measure the risk of a plant's invasion by its finite rate of increase (λ); these exploit the so-called ‘crop’ model that compares GM crops with conventional crop plants plus the GM trait that may influence fitness. For example, stage projection matrix modelling, in which the effect of changes in individual demographic processes on λ can be assessed, has provided useful insights into what parts of the life history are most sensitive to being altered by a novel trait (Linder & Schmitt 1994, 1995; Bullock 1999). These models, coupled with field experiments, actually allow you to see, for example, if increased seed production (fecundity) may actually result in increased fitness (Gray 2000). Ongoing research is using a combination of experiments and modelling to measure fitness components across habitats and years, under density-dependent conditions, and with and without the presence of and resistance to various types of pathogens.

Weed ecologists have also modelled the potential impact of reduced weed populations on farmland birds. Watkinson et al. (2000) looked at the effects of changes in fat hen population density on skylarks (in a study that incidentally did not escape the banner headline treatment resulting from a generic leap of logic, along the lines that GM herbicide-tolerant crops would kill off all farmland birds!). Arguably the most interesting finding of this work was how dependent the outcome was on the postulated pattern of technology take up by different farms in relation to whether they already had weed-free fields or not. This clearly points to the importance of social and economic considerations in determining the potential environmental impact of GM crops. Are these, I wonder, any easier to predict than ecological changes?

Therefore, in the field of plant ecology at least, I believe, despite some early pessimism (Kareiva, Parker & Pascual 1996), that the combination of experiments (which of course have precision but may lack generality) and models (often the reverse) that characterizes ecology is serving it well in increasing our understanding of, and our ability to predict, the possible environmental consequences of GM crops. Certainly the challenge of predicting the invasiveness of crops engineered to express so-called ‘fitness’ traits is being based on a firm understanding of the variation produced by traditional breeding.

In other areas, notably in predicting more indirect impacts such as those on non-target organisms or on soil biodiversity, progress has been much slower, not least because, as the need for the FSE clearly demonstrated, we frequently lack baseline information about variation in the effects of different conventional crops and farming practices. Mention of lack of information brings me finally and briefly to the issue of ‘uncertainty’.

How well has ecology dealt with uncertainty? Have ecologists in offering advice to regulators honestly owned up to the gaps in their knowledge? I put it to you that not only do we openly acknowledge the gaps in our understanding we appear to be actually proud of them. How many papers in the ecological journals begin with ‘x or y is poorly understood’ or ‘nothing is know about the effects of x on y’? I know such phrases are often a raison d’être for the research that follows or not very thinly veiled requests for more funding, but we do seem to positively wallow in our ignorance. Ecology, one would think, is a science with more gaps than knowledge!

So full marks for not claiming to know more than we do. But with that praise comes a warning. It is important to be clear that this limited understanding of the consequences of changes in complex ecosystems is not unique to GM crops. It is true of most past and current changes in agriculture, including the introduction of novel crops and varieties. We are in a somewhat bizarre situation at present where the effects of all other forms of farming (and food) are often regarded as fully understood, and furthermore carry zero risk. Herbicide-tolerant crops not produced by recombinant DNA technology have been trialled in the UK (and grown around the world) without the thorough regulatory oversight I have described to you for GM crops. My part of Dorset is turning into Iowa as farmers plough up pasture to grow forage maize to feed to cattle kept on concrete. In contrast to the possibility of GM agriculture, such changes and their undoubted impacts on biodiversity appear to be receiving very little attention or research funding. Furthermore, in the highly charged and politically sensitive debate out there, you will be pilloried not only for what you know you do not know but for what you do not even know you do not know: the unknown unknowns. So we must be honest about uncertainty, but careful about its context.

Ecology and policy

  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References

I want finally to make some brief concluding remarks about ecology and policy. When David Walton and I organized the ‘Ecology and government policies’ meeting more than 12 years ago we were concerned that the results of ecological science were not being taken into account by decision-makers. I even remember the angst we felt in the BES at that time that the very word ‘ecology’ was being highjacked and disfigured by politicians. (You may agree the same thing has since happened to the word ‘biodiversity’.) My message to the BES is, not only has ecology arrived as a component of policy and decision-making, it is bang-slap centre stage. As we await the results of perhaps the UK's largest single public-funded ecological experiment on farmland, the science is well and truly in the spotlight. Much is expected of it: I suspect more than it is able to deliver.

Along with this centre-stage position comes a huge challenge, and a huge responsibility, not only to ensure that we continue to deal wisely and with integrity with the problems of biological variability and complexity, with the limitations to prediction, and the issue of uncertainty, but also, I believe, a responsibility to raise our voices when the science is abused, when it is seized on by one side or other (as it will be) in support of the relentless advocacy of a single extreme viewpoint. The temptation to abrogate that responsibility, to leave the heat of the kitchen or hide behind our uncertainties, I have always found to be strong in ecology. We never know enough to form an opinion. We like to travel much further than most along the scientist's road of wanting to know everything about everything before we express a view. But others do not, and their certainty can be both appealing and convincing (although one senses that an increasingly informed and more sceptical public is growing to mistrust their extreme claims). We must not forget that opinion, like policies or any political decisions, are held and made by people: they have social, ethical, cultural and emotional dimensions. Science informs but does not dictate (unlike the five economic tests?). The final decisions on the cultivation of GM crops in Europe and the UK are political decisions. Science will be a part of those decisions but one senses the defining factors will lie elsewhere. But that does not absolve us from getting the science right or failing to emphasize its importance.

I agree wholeheartedly with the recent correspondent (Mark Huxham) to the BES Bulletin (34: 3, p. 8), to whom I referred at the beginning, that we can easily get hung up on what he calls the fact–value divide: the idea that as scientists we can objectively study the natural world without being ethically or politically involved in it. In a debate where, as I said earlier, everyone has a view, where many have a fixed agenda, where we have quite properly moved away from an expert-led, top-down system of advice to government to a more inclusive one in which views may not always be weighed according to the strength of the evidence on which they are based, it is important that on matters that affect our environment informed voices continue to be heard.

In any case what is wrong with having an opinion? The boundary between science and viewpoint, science and rhetoric, is easily recognized. It is perfectly fine to recognize and acknowledge the baggage we bring to our view (in my generation of ecologists this has been largely one of a fascination with natural history and an interest in wildlife conservation). Uncertainty and doubt both make poor long-term bedfellows. In last year's Booker prize-winning novel Life of Pi (2001), Yann Martel put these words into the mouth of his eponymous hero Pi Patel, who had by that time adopted three religions and, in decrying agnosticism, was paying respect to atheists who (he said):

‘Like me, they go as far as the legs of reason will carry them – and then they leap … Surely we are also permitted doubt. But we must move on. To choose doubt as a philosophy of life is akin to choosing immobility as a means of transportation.’

Thank you for your attention.


  1. Top of page
  2. Summary
  3. Introduction
  4. Elements in the GM debate
  5. Science in the regulatory process
  6. Ecology and policy
  7. References
  • ACRE (1999) Advisory Committee on Releases to the Environment. Annual Report No. 6. DETR, London, UK.
  • ACRE (2001) Advisory Committee on Releases to the Environment. Annual Report No. 8. DEFRA, London, UK.
  • Bergelson, J., Burrington, C.B. & Wichmann, G. (1998) Promiscuity in transgenic plants. Nature, 395, 25.
  • Beringer, J.E. (2000) Releasing genetically modified organisms: will any harm outweigh any advantage? Journal of Applied Ecology, 37, 207214.
  • Bullock, J.M. (1999) Using population matrix models to target GMO risk assessment. Aspects of Applied Biology, 53, 205212.
  • Champolivier, J., Gasquez, J., Messian, A. & Richard-Molard, M. (1999) Management of transgenic crops within the cropping system. Gene Flow and Agriculture – Relevance for Transgenic Crops (ed. P.Lutman), pp. 233240. BCPC Symposium Proceedings No. 72. British Crop Protection Council, Farnham, Surrey, UK.
  • Crawley, M.J., Brown, S.L., Hails, R.S., Kohn, D. & Rees, M. (2001) Transgenic crops in natural habitats. Nature, 409, 682683.
  • Crawley, M.J., Hails, R.S., Rees, M., Kohn, D. & Buxton, J. (1993) Ecology of transgenic rape in natural habitats. Nature, 363, 620623.
  • Firbank, L.G., Dewar, A.M., Hill, M.O., May, M.J., Perry, J.N., Rothery, P., Squire, G.R. & Woiwood, I.P. (1999) Farm-scale evaluation of GM crops explained. Nature, 399, 727728.
  • Firbank, L.G., Heard, M.S., Woiwood, I.P., Hawes, C., Haughton, A.J., Champion, G.T., Scott, R.J., Hill, M.O., Dewar, A.M., Squire, G.R., May, M.J., Brooks, D.R., Bohan, D.A., Daniels, R.E., Osborne, J.L., Roy, D.B., Black, H.I.J., Rothery, P. & Perry, J.N. (2003) An introduction to the farm-scale evaluations of genetically modified herbicide-tolerant crops. Journal of Applied Ecology, 40, 216.
  • Fuller, R.J., Gregory, R.D., Gibbons, D.W., Marchant, J.H., Wilson, J.D., Baillie, S.R. & Carter, N. (1995) Population declines and range contractions among lowland farmland birds in Britain. Conservation Biology, 9, 14251441.
  • GM Science Review Panel (2003) GM Science Review. First Report. Defra, London, UK.
  • Gray, A.J. (2000) The transfer of traits to wild relatives. Predicting Field Performance in Crop Protection (ed. L.G.Copping), pp. 165174. BCPC Symposium Proceedings No. 74. British Crop Protection Council, Farnham, Surrey, UK.
  • Gray, A.J. (2003) Risk assessment for LMOs: a European perspective. LMOs and the Environment (ed. C.R.Roseland). Proceedings of an International Conference 27–30 November 2001. Organisation for Economic Co-operation and Development, Paris, France (in press).
  • Gray, A.J. & Raybould, A.F. (1999) Environmental Risks of Herbicide-Tolerant Oilseed Rape – A Review of the PGS Hybrid Oilseed Rape. Genetically Modified Organisms Research Report 15. DETR, London, UK.
  • Hilbeck, A., Baumgartner, M., Fried, P.M. & Bigler, F. (1998) Effects of transgenic Bacillus thuringiensis corn fed prey on the mortality and development time of immature Chrysoperia carnea (Neuroptera, Chrysopidae). Environmental Entomology, 17, 480487.
  • Hoban, T.J. (1999) DNA-free tomatoes? Seed Trade News. Ball Publishing, Batavia, IL.
  • Kareiva, P., Parker, I.M. & Pascual, M. (1996) Can we use experiments and models in predicting the invasiveness of genetically engineered organisms? Ecology, 77, 16701675.
  • Krebs, J.R., Wilson, J.D., Bradbury, R.B. & Siriwardena, G.M. (1999) The second silent spring. Nature, 400, 611612.
  • Linder, C.R. & Schmitt, J. (1994) Assessing the risks of transgene escape through time and crop–wild hybrid persistence. Molecular Ecology, 3, 2330.
  • Linder, C.R. & Schmitt, J. (1995) Potential persistence of escaped transgenes: performance of transgenic oil-modified Brassica seeds and seedlings. Ecological Applications, 5, 10561068.
  • Losey, J.E., Rayor, L.S. & Carter, M.E. (1999) Transgenic pollen harms monarch larvae. Nature, 399, 214.
  • Martel, Y. (2001) Life of Pi. Harcourt Books, San Diego.
  • Perry, J.N., Rothery, P., Clark, S.J., Heard, M.S. & Hawes, C. (2003) Design, analysis and statistical power of the farm-scale evaluations of genetically modified herbicide-tolerant crops. Journal of Applied Ecology, 40, 1731.
  • Raybould, A.F., Goudet, J., Mogg, R.J., Gliddon, C.J. & Gray, A.J. (1996) The genetic structure of a linear population of Beta vulgaris ssp. maritima (sea beet) revealed by isozyme and RFLP analysis. Heredity, 76, 111117.
  • Raybould, A.F., Mogg, R.J., Clarke, R.T., Gliddon, C.J. & Gray, A.J. (1999) Variation and population structure at microsatellite and isozyme loci in wild cabbage (Brassica oleracea L.) in Dorset (UK). Genetic Resources and Crop Evolution, 46, 351360.
  • Rieger, M.A., Lamond, M., Preston, C., Fowles, S.B. & Roush, R.T. (2002) Pollen-mediated movement of herbicide resistance between commercial canola fields. Science, 296, 23862388.
  • Sears, M.K., Hellmich, R.L., Stanley-Horn, D.E., Oberhauser, K.S., Pleasants, J.M., Mattila, H.R., Siegfried, B.D. & Dively, G.P. (2001) Impact of Bt corn pollen on monarch butterfly populations: a risk assessment. Proceedings of the National Academy of Sciences USA, 98, 1193711942.
  • Shorrocks, B. & Coates, D. (1993) The Release of Genetically-Engineered Organisms. British Ecological Society Ecological Issues No. 4. Field Studies Council, Shrewsbury, UK.
  • Stewart, C.N., All, J.N., Raymer, P.L. & Ramachandran, S. (1997) Transgenic insecticidal oilseed rape on the loose. Commercialisation of Transgenic Crops: Risk Benefit and Trade Considerations (eds G.D.McLean, P.M.Waterhouse, G.A.Evans & M.J.Gibbs), pp. 137144. CRC for Bureau of Resource Sciences, Canberra, UK.
  • Tiedje, J.M., Colwell, R.K., Grossman, Y.L., Hodson, R.E., Lenski, R.E., Mack, R.N. & Regal, P.J. (1989) The planned introduction of genetically engineered organisms: ecological considerations and recommendations. Ecology, 70, 298315.
  • Walton, D.W.H. & Gray, A.J. (1991) Ecology and government policies. Trends in Ecology and Evolution, 6, 144145.
  • Watkinson, A.R., Freckleton, R.P., Robinson, R.A. & Sutherland, W.J. (2000) Predicting biodiversity responses to GM herbicide-tolerant crops. Science, 289, 15541556.
  • Wilkinson, M.J., Elliott, L.J., Allainquillaume, J., Shaw, M.W., Norris, C., Welters, R., Alexander, M., Sweet, J. & Mason, D.C. (2003) Hybridisation between Brassica napus and B. rapa on a national scale in the United Kingdom. Science, 302, 457459.
  • Williamson, M. (1996) Biological Invasions. Chapman & Hall, London, UK.